This action might not be possible to undo. Are you sure you want to continue?
https://www.scribd.com/doc/61419610/DeepeshBook
08/02/2011
text
original
in Image Processing
Joachim Weickert
B.G. Teubner Stuttgart
Anisotropic Diﬀusion
in Image Processing
Joachim Weickert
Department of Computer Science
University of Copenhagen
Copenhagen, Denmark
B.G. Teubner Stuttgart 1998
Dr. rer. nat. Joachim Weickert
Born in 1965 in Ludwigshafen/Germany. Studies in mathematics, physics and com
puter science at the University of Kaiserslautern. 1987 B.Sc. in physics and indus
trial mathematics. 1991 M.Sc. in industrial mathematics. 1996 Ph.D. in mathe
matics. Postdoctoral researcher at the Image Sciences Institute at Utrecht Uni
versity from 2/96 to 3/97. Since then visiting assistant research professor at the
Department of Computer Science, Copenhagen University.
The cover image shows a thresholded nonwoven fabric image which was processed
by applying a coherenceenhancing anisotropic diﬀusion ﬁlter (see Section 5.2 for
more details). The goal was to visualize the quality relevant adjacent ﬁbre struc
tures, socalled stripes. The displayed equations describe the basic structure of
nonlinear diﬀusion ﬁltering in the continuous, semidiscrete, and fully discrete set
ting. Their theoretical foundations are treated in Chapters 2–4.
c ( Copyright 2008 by Joachim Weickert.
All rights reserved. No part of this book may be reproduced by any means, or
transmitted, or translated into a machine language without the written permission
of the author.
This book had been published by B. G. Teubner (Stuttgart) in 1998 and went out
of print in 2001. The copyright has been returned to the author in 2008. In the
current version a few typos and other errors have been corrected.
To my parents Gerda and Norbert
Preface
Partial diﬀerential equations (PDEs) have led to an entire new ﬁeld in image
processing and computer vision. Hundreds of publications have appeared in the last
decade, and PDEbased methods have played a central role at several conferences
and workshops.
The success of these techniques is not really surprising, since PDEs have proved
their usefulness in areas such as physics and engineering sciences for a very long
time. In image processing and computer vision, they oﬀer several advantages:
• Deep mathematical results with respect to wellposedness are available, such
that stable algorithms can be found. PDEbased methods are one of the
mathematically bestfounded techniques in image processing.
• They allow a reinterpretation of several classical methods under a novel uni
fying framework. This includes many wellknown techniques such as Gaussian
convolution, median ﬁltering, dilation or erosion.
• This understanding has also led to the discovery of new methods. They
can oﬀer more invariances than classical techniques, or describe novel ways
of shape simpliﬁcation, structure preserving ﬁltering, and enhancement of
coherent linelike structures.
• The PDE formulation is genuinely continuous. Thus, their approximations
aim to be independent of the underlying grid and may reveal good rotational
invariance.
PDEbased image processing techniques are mainly used for smoothing and
restoration purposes. Many evolution equations for restoring images can be de
rived as gradient descent methods for minimizing a suitable energy functional, and
the restored image is given by the steadystate of this process. Typical PDE tech
niques for image smoothing regard the original image as initial state of a parabolic
(diﬀusionlike) process, and extract ﬁltered versions from its temporal evolution.
The whole evolution can be regarded as a socalled scalespace, an embedding of
the original image into a family of subsequently simpler, more global representa
tions of it. Since this introduces a hierarchy into the image structures, one can use
a scalespace representation for extracting semantically important information.
One of the two goals of this book is to give an overview of the stateoftheart of
PDEbased methods for image enhancement and smoothing. Emphasis is put on a
v
vi PREFACE
uniﬁed description of the underlying ideas, theoretical results, numerical approxi
mations, generalizations and applications, but also historical remarks and pointers
to open questions can be found. Although being concise, this part covers a broad
spectrum: it includes for instance an early Japanese scalespace axiomatic, the
Mumford–Shah functional for image segmentation, continuousscale morphology,
active contour models and shock ﬁlters. Many references are given which point the
reader to useful original literature for a task at hand.
The second goal of this book is to present an indepth treatment of an interest
ing class of parabolic equations which may bridge the gap between scalespace and
restoration ideas: nonlinear diﬀusion ﬁlters. Methods of this type have been pro
posed for the ﬁrst time by Perona and Malik in 1987 [326]. In order to smooth an
image and to simultaneously enhance important features such as edges, they apply
a diﬀusion process whose diﬀusivity is steered by derivatives of the evolving image.
These ﬁlters are diﬃcult to analyse mathematically, as they may act locally like
a backward diﬀusion process. This gives rise to wellposedness questions. On the
other hand, nonlinear diﬀusion ﬁlters are frequently applied with very impressive
results; so there appears the need for a theoretical foundation.
We shall develop results in this direction by investigating a general class of
nonlinear diﬀusion processes. This class comprises linear diﬀusion ﬁlters as well as
spatial regularizations of the Perona–Malik process, but it also allows processes
which replace the scalar diﬀusivity by a diﬀusion tensor. Thus, the diﬀusive ﬂux
does not have to be parallel to the grey value gradient: the ﬁlters may become
anisotropic. Anisotropic diﬀusion ﬁlters can outperform isotropic ones with respect
to certain applications such as denoising of highly degraded edges or enhancing
coherent ﬂowlike images by closing interrupted onedimensional structures. In or
der to establish wellposedness and scalespace properties for this class, we shall
investigate existence, uniqueness, stability, maximum–minimum principles, Lya
punov functionals, and invariances. The proofs present mathematical results from
the nonlinear analysis of partial diﬀerential equations.
Since digital images are always sampled on a pixel grid, it is necessary to know
if the results for the continuous framework carry over to the practically relevant
discrete setting. These questions are an important topic of the present book as
well. A general characterization of semidiscrete and fully discrete ﬁlters, which
reveal similar properties as their continuous diﬀusion counterparts, is presented. It
leads to a semidiscrete and fully discrete scalespace theory for nonlinear diﬀusion
processes. Mathematically, this comes down to the study of nonlinear systems of
ordinary diﬀerential equations and the theory of nonnegative matrices.
Organization of the book. Image processing and computer vision are inter
disciplinary areas, where researchers, practitioners and students may have a very
diﬀerent scientiﬁc background and diﬀering intentions. As a consequence, I have
tried to keep this book as selfcontained as possible, and to include various aspects
PREFACE vii
such that it should contain interesting material for many readers. The prerequisites
are kept to a minimum and can be found in standard textbooks on image process
ing [163], matrix analysis [407], functional analysis [9, 58, 7], ordinary diﬀerential
equations [56, 412], partial diﬀerential equations [185] and their numerical aspects
[293, 286]. The book is organized as follows:
Chapter 1 surveys the fundamental ideas behind PDEbased smoothing and
restoration methods. This general overview sketches their theoretical properties,
numerical methods, applications and generalizations. The discussed methods in
clude linear and nonlinear diﬀusion ﬁltering, coupled diﬀusion–reaction methods,
PDE analogues of classical morphological processes, Euclidean and aﬃne invariant
curve evolutions, and total variation methods.
The subsequent three chapters explore a theoretical framework for anisotropic
diﬀusion ﬁltering. Chapter 2 presents a general model for the continuous setting
where the diﬀusion tensor depends on the structure tensor (interest operator,
secondmoment matrix), a generalization of the Gaussiansmoothed gradient al
lowing a more sophisticated description of local image structure. Existence and
uniqueness are discussed, and stability and an extremum principle are proved.
Scalespace properties are investigated with respect to invariances and information
reducing qualities resulting from associated Lyapunov functionals.
Chapter 3 establishes conditions under which comparable wellposedness and
scalespace results can be proved for the semidiscrete framework. This case takes
into account the spatial discretization which is characteristic for digital images,
but it keeps the scalespace idea of using a continuous scale parameter. It leads
to nonlinear systems of ordinary diﬀerential equations. We shall investigate under
which conditions it is possible to get consistent approximations of the continuous
anisotropic ﬁlter class which satisfy the abovementioned requirements.
In practice, scalespaces can only be calculated for a ﬁnite number of scales,
though. This corresponds to the fully discrete case which is treated in Chapter
4. The investigated discrete ﬁlter class comes down to solving linear systems of
equations which may arise from semiimplicit time discretizations of the semidis
crete ﬁlters. We shall see that many numerical schemes share typical features
with their semidiscrete counterparts, for instance wellposedness results, extremum
principles, Lyapunov functionals, and convergence to a constant steadystate. This
chapter also shows how one can design eﬃcient numerical methods which are in
accordance with the fully discrete scalespace framework and which are based on
an additive operator splitting (AOS).
Chapter 5 is devoted to practical topics such as ﬁlter design, examples and ap
plications of anisotropic diﬀusion ﬁltering. Speciﬁc models are proposed which are
tailored towards smoothing with edge enhancement and multiscale enhancement
of coherent structures. Their qualities are illustrated using images arising from
computer aided quality control and medical applications, but also ﬁngerprint im
viii PREFACE
ages and impressionistic paintings shall be processed. The results are juxtaposed
to related methods from Chapter 1.
Finally, Chapter 6 concludes the book by giving a summary and discussing
possible future perspectives for nonlinear diﬀusion ﬁltering.
Acknowledgments. In writing this book I have been helped and inﬂuenced
by many people, and it is a pleasure to take this opportunity to express my grat
itude to them. The present book is an extended and revised version of my Ph.D.
thesis [416], which was written at the Department of Mathematics at the Uni
versity of Kaiserslautern, Germany. Helmut Neunzert, head of the Laboratory of
Technomathematics, drew my interest to diﬀusion processes in image processing,
and he provided the possibility to carry out this work at his laboratory. I also
thank him and the other editors of the ECMI Series as well as Teubner Verlag for
their interest in publishing this work.
Pierre–Louis Lions (CEREMADE, University Paris IX) invited me to the CERE
MADE, one of the birthplaces of many important ideas in this ﬁeld. He also gave
me the honour to present my results as an invited speaker at the EMS Confer
ence Multiscale Analysis in Image Processing (Lunteren, The Netherlands, October
1994) to an international audience, and he acted as a referee for the Ph.D. thesis.
After the defence of my thesis in Kaiserslautern, I joined the TGV (“tools
for geometry in vision”) group at Utrecht University Hospital for 14 months. In
this young and dynamic group I had the possibility to learn a lot about medical
image analysis, and to experience Bart ter Haar Romeny’s enthusiasm for scale
space. During that time I also met Atsushi Imiya (Chiba University, Japan) at
a workshop in Dagstuhl (Germany). He introduced me into the fascinating world
of early Japanese scalespace research conducted by Taizo Iijima decades before
scalespace became popular in America and Europe.
In the meantime I am with the computer vision group of Peter Johansen and
Jens Arnspang (DIKU, Copenhagen University). The discussions and collabora
tions with the members of this group increased my interest in scalespace related
deep structure analysis and information theory. In the latter ﬁeld I share many
common interests with Jon Sporring.
The proofreading of this book was done by Martin Reißel and Andrea Bechtold
(Kaiserslautern). Martin Reißel undertook the hard job of checking the whole
manuscript for its mathematical correctness, and Andrea Bechtold was a great help
in all kinds of diﬃculties with the English language. Also Robert Maas (Utrecht
University Hospital) contributed several useful hints.
This work has been funded by Stiftung Volkswagenwerk, Stiftung Rheinland–
Pfalz f¨ ur Innovation, the Real World Computing Partnership, the Danish Research
Council, and the EU–TMR Research Network VIRGO.
Copenhagen, October 1997 Joachim Weickert
Contents
1 Image smoothing and restoration by PDEs 1
1.1 Physical background of diﬀusion processes . . . . . . . . . . . . . . 2
1.2 Linear diﬀusion ﬁltering . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Relations to Gaussian smoothing . . . . . . . . . . . . . . . 3
1.2.2 Scalespace properties . . . . . . . . . . . . . . . . . . . . . 6
1.2.3 Numerical aspects . . . . . . . . . . . . . . . . . . . . . . . . 10
1.2.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2.5 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.6 Generalizations . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.3 Nonlinear diﬀusion ﬁltering . . . . . . . . . . . . . . . . . . . . . . 14
1.3.1 The Perona–Malik model . . . . . . . . . . . . . . . . . . . . 15
1.3.2 Regularized nonlinear models . . . . . . . . . . . . . . . . . 20
1.3.3 Anisotropic nonlinear models . . . . . . . . . . . . . . . . . 22
1.3.4 Generalizations . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.3.5 Numerical aspects . . . . . . . . . . . . . . . . . . . . . . . . 25
1.3.6 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.4 Methods of diﬀusion–reaction type . . . . . . . . . . . . . . . . . . 27
1.4.1 Single diﬀusion–reaction equations . . . . . . . . . . . . . . 27
1.4.2 Coupled systems of diﬀusion–reaction equations . . . . . . . 29
1.5 Classic morphological processes . . . . . . . . . . . . . . . . . . . . 31
1.5.1 Binary and greyscale morphology . . . . . . . . . . . . . . . 31
1.5.2 Basic operations . . . . . . . . . . . . . . . . . . . . . . . . 32
1.5.3 Continuousscale morphology . . . . . . . . . . . . . . . . . 32
1.5.4 Theoretical results . . . . . . . . . . . . . . . . . . . . . . . 34
1.5.5 Scalespace properties . . . . . . . . . . . . . . . . . . . . . 34
1.5.6 Generalizations . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.5.7 Numerical aspects . . . . . . . . . . . . . . . . . . . . . . . . 36
1.5.8 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
1.6 Curvaturebased morphological processes . . . . . . . . . . . . . . . 37
1.6.1 Meancurvature ﬁltering . . . . . . . . . . . . . . . . . . . . 37
1.6.2 Aﬃne invariant ﬁltering . . . . . . . . . . . . . . . . . . . . 40
1.6.3 Generalizations . . . . . . . . . . . . . . . . . . . . . . . . . 42
1.6.4 Numerical aspects . . . . . . . . . . . . . . . . . . . . . . . . 43
ix
x CONTENTS
1.6.5 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
1.6.6 Active contour models . . . . . . . . . . . . . . . . . . . . . 46
1.7 Total variation methods . . . . . . . . . . . . . . . . . . . . . . . . 49
1.7.1 TVpreserving methods . . . . . . . . . . . . . . . . . . . . . 50
1.7.2 TVminimizing methods . . . . . . . . . . . . . . . . . . . . 50
1.8 Conclusions and further scope of the book . . . . . . . . . . . . . . 53
2 Continuous diﬀusion ﬁltering 55
2.1 Basic ﬁlter structure . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.2 The structure tensor . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.3 Theoretical results . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.4 Scalespace properties . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.4.1 Invariances . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.4.2 Informationreducing properties . . . . . . . . . . . . . . . . 65
3 Semidiscrete diﬀusion ﬁltering 75
3.1 The general model . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.2 Theoretical results . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.3 Scalespace properties . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.4 Relation to continuous models . . . . . . . . . . . . . . . . . . . . . 86
3.4.1 Isotropic case . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.4.2 Anisotropic case . . . . . . . . . . . . . . . . . . . . . . . . . 88
4 Discrete diﬀusion ﬁltering 97
4.1 The general model . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.2 Theoretical results . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.3 Scalespace properties . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.4 Relation to semidiscrete models . . . . . . . . . . . . . . . . . . . . 102
4.4.1 Semiimplicit schemes . . . . . . . . . . . . . . . . . . . . . 102
4.4.2 AOS schemes . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5 Examples and applications 113
5.1 Edgeenhancing diﬀusion . . . . . . . . . . . . . . . . . . . . . . . . 114
5.1.1 Filter design . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.1.2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.2 Coherenceenhancing diﬀusion . . . . . . . . . . . . . . . . . . . . . 127
5.2.1 Filter design . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.2.2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
6 Conclusions and perspectives 135
Bibliography 139
Index 165
Chapter 1
Image smoothing and restoration
by PDEs
PDEbased methods appear in a large variety of image processing and computer
vision areas ranging from shapefromshading and histogramme modiﬁcation to
optic ﬂow and stereo vision.
This chapter reviews their main application, namely the smoothing and restora
tion of images. It is written in an informal style and refers to a large amount of
original literature, where proofs and full mathematical details can be found.
The goal is to make the reader sensitive to the similarities, diﬀerences, advan
tages and shortcomings of these techniques, and to point out the main results and
open problems in this rapidly evolving area.
For each class of methods the basic ideas are explained and their theoretical
background, numerical aspects, generalizations, and applications are discussed.
Many of these ideas are borrowed from physical phenomena such as wave prop
agation or transport of heat and mass. Nevertheless, also gas dynamics, crack
propagation, grassﬁre ﬂow, the study of salinity proﬁles in oceanography, or mech
anisms of the retina and the brain are closely related to some of these approaches.
Although a detailed discussion of these connections would be far beyond the scope
of this work, they are mentioned wherever they appear, in order to allow the
interested reader to pursue these ideas. Also many historical notes are added.
The outline of this chapter is as follows: We start with reviewing the physi
cal ideas behind diﬀusion processes. This helps us to better understand the next
sections which are concerned with the properties of linear and nonlinear diﬀusion
ﬁlters in image processing. The subsequent study of image enhancement methods
of diﬀusion–reaction type relates diﬀusion ﬁlters to variational image restoration
techniques. After that we investigate morphological ﬁlters, a topic which looks at
ﬁrst glance fairly diﬀerent to the diﬀusion approach. Nevertheless, it reveals some
interesting relations when it is interpreted within a PDE framework. This becomes
1
2 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
especially evident when considering curvaturebased morphological PDEs. Finally
we shall discuss total variation image restoration techniques which permit discon
tinuous solutions. The last section summarizes the advantages and shortcomings
of the main methods and gives an outline of the questions we are concerned with
in the subsequent chapters.
1.1 Physical background of diﬀusion processes
Most people have an intuitive impression of diﬀusion as a physical process that
equilibrates concentration diﬀerences without creating or destroying mass. This
physical observation can be easily cast in a mathematical formulation.
The equilibration property is expressed by Fick’s law:
j = −D ∇u. (1.1)
This equation states that a concentration gradient ∇u causes a ﬂux j which aims
to compensate for this gradient. The relation between ∇u and j is described by
the diﬀusion tensor D, a positive deﬁnite symmetric matrix. The case where j and
∇u are parallel is called isotropic. Then we may replace the diﬀusion tensor by a
positive scalarvalued diﬀusivity g. In the general anisotropic case, j and ∇u are
not parallel.
The observation that diﬀusion does only transport mass without destroying it
or creating new mass is expressed by the continuity equation
∂
t
u = −div j (1.2)
where t denotes the time.
If we plug in Fick’s law into the continuity equation we end up with the diﬀusion
equation
∂
t
u = div (D ∇u). (1.3)
This equation appears in many physical transport processes. In the context of
heat transfer it is called heat equation. In image processing we may identify the
concentration with the grey value at a certain location. If the diﬀusion tensor is
constant over the whole image domain, one speaks of homogeneous diﬀusion, and
a spacedependent ﬁltering is called inhomogeneous. Often the diﬀusion tensor is a
function of the diﬀerential structure of the evolving image itself. Such a feedback
leads to nonlinear diﬀusion ﬁlters. Diﬀusion which does not depend on the evolving
image is called linear.
Sometimes the computer vision literature deviates from the preceding nota
tions: It can happen that homogeneous ﬁltering is named isotropic, and inhomo
geneous blurring is called anisotropic, even if it uses a scalarvalued diﬀusivity
instead of a diﬀusion tensor.
1.2 LINEAR DIFFUSION FILTERING 3
1.2 Linear diﬀusion ﬁltering
The simplest and best investigated PDE method for smoothing images is to apply
a linear diﬀusion process. We shall focus on the relation between linear diﬀusion
ﬁltering and the convolution with a Gaussian, analyse its smoothing properties for
the image as well as its derivatives, and review the fundamental properties of the
Gaussian scalespace induced by linear diﬀusion ﬁltering. Afterwards a survey on
discrete aspects is given and applications and limitations of the linear diﬀusion
paradigm are discussed. The section is concluded by sketching two linear general
izations which can incorporate apriori knowledge: aﬃne Gaussian scalespace and
directed diﬀusion processes.
1.2.1 Relations to Gaussian smoothing
Gaussian smoothing
Let a greyscale image f be represented by a realvalued mapping f ∈ L
1
(IR
2
). A
widelyused way to smooth f is by calculating the convolution
(K
σ
∗f)(x) :=
IR
2
K
σ
(x−y) f(y) dy (1.4)
where K
σ
denotes the twodimensional Gaussian of width (standard deviation)
σ>0 :
K
σ
(x) :=
1
2πσ
2
exp
−
[x[
2
2σ
2
. (1.5)
There are several reasons for the excellent smoothing properties of this method:
First we observe that since K
σ
∈ C
∞
(IR
2
) we get K
σ
∗f ∈ C
∞
(IR
2
), even if f is
only absolutely integrable.
Next, let us investigate the behaviour in the frequency domain. When deﬁning
the Fourier transformation T by
(Tf)(ω) :=
IR
2
f(x) exp(−i'ω, x`) dx (1.6)
we obtain by the convolution theorem that
(T(K
σ
∗f)) (ω) = (TK
σ
)(ω) (Tf)(ω). (1.7)
Since the Fourier transform of a Gaussian is again Gaussianshaped,
(TK
σ
)(ω) = exp
−
[ω[
2
2/σ
2
, (1.8)
4 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
we observe that (1.4) is a lowpass ﬁlter that attenuates high frequencies in a
monotone way.
Interestingly, the smoothing behaviour can also be understood in the context
of a PDE interpretation.
Equivalence to linear diﬀusion ﬁltering
It is a classical result (cf. e.g. [331, pp. 267–271] and [185, pp. 43–56]) that for any
bounded f ∈C(IR
2
) the linear diﬀusion process
∂
t
u = ∆u, (1.9)
u(x, 0) = f(x) (1.10)
possesses the solution
u(x, t) =
f(x) (t = 0)
(K
√
2t
∗ f)(x) (t > 0).
(1.11)
This solution is unique, provided we restrict ourselves to functions satisfying
[u(x, t)[ ≤ M exp (a[x[
2
) (M, a > 0). (1.12)
It depends continuously on the initial image f with respect to  . 
L
∞
(IR
2
)
, and it
fulﬁls the maximum–minimum principle
inf
IR
2
f ≤ u(x, t) ≤ sup
IR
2
f on IR
2
[0, ∞). (1.13)
From (1.11) we observe that the time t is related to the spatial width σ =
√
2t of
the Gaussian. Hence, smoothing structures of order σ requires to stop the diﬀusion
process at time
T =
1
2
σ
2
. (1.14)
Figure 5.2 (b) and 5.3 (c) in Chapter 5 illustrate the eﬀect of linear diﬀusion
ﬁltering.
Gaussian derivatives
In order to understand the structure of an image we have to analyse grey value
ﬂuctuations within a neighbourhood of each image point, that is to say, we need
information about its derivatives. However, diﬀerentiation is illposed
1
, as small
perturbations in the original image can lead to arbitrarily large ﬂuctuations in
the derivatives. Hence, the need for regularization methods arises. A thorough
1
A problem is called wellposed, if it has a unique solution which depends continuously on
the input data and parameters. If one of these conditions is violated, it is called illposed.
1.2 LINEAR DIFFUSION FILTERING 5
treatment of this mathematical theory can be found in the books of Tikhonov and
Arsenin [402], Louis [266] and Engl et al. [128].
One possibility to regularize is to convolve the image with a Gaussian prior to
diﬀerentiation [404]. By the equality
∂
n
x
1
∂
m
x
2
(K
σ
∗f) = K
σ
∗ (∂
n
x
1
∂
m
x
2
f) = (∂
n
x
1
∂
m
x
2
K
σ
) ∗ f (1.15)
for suﬃciently smooth f, we observe that all derivatives undergo the same Gaussian
smoothing process as the image itself and this process is equivalent to convolving
the image with derivatives of a Gaussian.
Replacing derivatives by these Gaussian derivatives has a strong regularizing
eﬀect. This property has been used to stabilize illposed problems like deblurring
images by solving the heat equation backwards in time
2
[141, 177]. Moreover, Gaus
sian derivatives can be combined to socalled diﬀerential invariants, expressions
that are invariant under transformations such as rotations, for instance [∇K
σ
∗u[
or ∆K
σ
∗u.
Diﬀerential invariants are useful for the detection of features such as edges,
ridges, junctions, and blobs; see [256] for an overview. To illustrate this, we focus
on two applications for detecting edges.
A frequently used method is the Canny edge detector [69]. It is based on calcu
lating the ﬁrst derivatives of the Gaussiansmoothed image. After applying sophis
ticated thinning and linking mechanisms (nonmaxima suppression and hysteresis
thresholding), edges are identiﬁed as locations where the gradient magnitude has a
maximum. This method is often acknowledged to be the best linear edge detector,
and it has become a standard in edge detection.
Another important edge detector is the Marr–Hildreth operator [278], which
uses the LaplacianofGaussian (LoG) ∆K
σ
as convolution kernel. Edges of f are
identiﬁed as zerocrossings of ∆K
σ
∗f. This needs no further postprocessing and
always gives closed contours. There are indications that LoGs and especially their
approximation by diﬀerencesofGaussians (DoGs) play an important role in the
visual system of mammals, see [278] and the references therein. Young developed
this theory further by presenting evidence that the receptive ﬁelds in primate eyes
are shaped like the sum of a Gaussian and its Laplacian [449], and Koenderink
and van Doorn suggested the set of Gaussian derivatives as a general model for
the visual system [242].
If one investigates the temporal evolution of the zerocrossings of an image ﬁl
tered by linear diﬀusion, one observes an interesting phenomenon: When increasing
the smoothing scale σ, no new zerocrossings are created which cannot be traced
back to ﬁner scales [439]. This evolution property is called causality [240]. It is
2
Of course, solutions of the regularization can only approximate the solution of the original
problem (if it exists). In practice, increasing the order of applied Gaussian derivatives or reducing
the kernel size will ﬁnally deteriorate the results of deblurring.
6 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
closely connected to the maximum–minimum principle of certain parabolic opera
tors [189]. Attempts to reconstruct the original image from the temporal evolution
of the zerocrossings of the Laplacian have been carried out by Hummel and Mo
niot [190]. They concluded, however, that this is practically unstable unless very
much additional information is provided.
In the western world the evolution property of the zerocrossings was the key
investigation which has inspired Witkin to the socalled scalespace concept [439].
This shall be discussed next.
1.2.2 Scalespace properties
The general scalespace concept
It is a wellknown fact that images usually contain structures at a large variety of
scales. In those cases where it is not clear in advance which is the right scale for the
depicted information it is desirable to have an image representation at multiple
scales. Moreover, by comparing the structures at diﬀerent scales, one obtains a
hierarchy of image structures which eases a subsequent image interpretation.
A scalespace is an image representation at a continuum of scales, embedding
the image f into a family ¦T
t
f[t ≥ 0¦ of gradually simpliﬁed versions of it, provided
that it fulﬁls certain requirements
3
. Most of these properties can be classiﬁed as
architectural, smoothing (informationreducing) or invariance requirements [12].
An important architectural assumption is recursivity, i.e. for t = 0, the scale
space representation gives the original image f, and the ﬁltering may be split into
a sequence of ﬁlter banks:
T
0
f = f, (1.16)
T
t+s
f = T
t
(T
s
f) ∀ s, t ≥ 0. (1.17)
This property is very often referred to as the semigroup property. Other architec
tural principles comprise for instance regularity properties of T
t
and local behaviour
as t tends to 0.
Smoothing properties and information reduction arise from the wish that the
transformation should not create artifacts when passing from ﬁne to coarse rep
resentation. Thus, at a coarse scale, we should not have additional structures
which are caused by the ﬁltering method itself and not by underlying structures
at ﬁner scales. This simpliﬁcation property is speciﬁed by numerous authors in
diﬀerent ways, using concepts such as no creation of new level curves (causality)
[240, 450, 189, 255], nonenhancement of local extrema [30, 257], decreasing number
3
Recently it has also been proposed to extend the scalespace concept to scale–imprecision
space by taking into account the imprecision of the measurement device [171].
1.2 LINEAR DIFFUSION FILTERING 7
of local extrema [255], maximum loss of ﬁgure impression [196], Tikhonov regular
ization [302, 303], maximum–minimum principle [189, 328], positivity [324, 138],
preservation of positivity [191, 193, 320], comparison principle [12], and Lyapunov
functionals [415, 429]. Especially in the linear setting, many of these properties are
equivalent or closely related; see [426] for more details.
We may regard an image as a representative of an equivalence class containing
all images that depict the same object. Two images of this class diﬀer e.g. by grey
level shifts, translations and rotations or even more complicated transformations
such as aﬃne mappings. This makes the requirement plausible that the scalespace
analysis should be invariant to as many of these transformations as possible, in
order to analyse only the depicted object [196, 16].
The pioneering work of Alvarez, Guichard, Lions and Morel [12] shows that
every scalespace fulﬁlling some fairly natural architectural, informationreducing
and invariance axioms is governed by a PDE with the original image as initial
condition. Thus, PDEs are the suitable framework for scalespaces.
Often these requirements are supplemented with an additional assumption
which is equivalent to the superposition principle, namely linearity:
T
t
(af +bg) = a T
t
f +b T
t
g ∀ t ≥ 0, ∀ a, b ∈ IR. (1.18)
As we shall see below, imposing linearity restricts the scalespace idea to essentially
one representative.
Gaussian scalespace
The historically ﬁrst and best investigated scalespace is the Gaussian scalespace,
which is obtained via convolution with Gaussians of increasing variance, or – equiv
alently – by linear diﬀusion ﬁltering according to (1.9), (1.10).
Usually a 1983 paper by Witkin [439] or a 1980 report by Stansﬁeld [392]
are regarded as the ﬁrst references to the linear scalespace idea. Recent work
by Weickert, Ishikawa and Imiya [426, 427], however, shows that scalespace is
more than 20 years older: An axiomatic derivation of 1D Gaussian scalespace
has already been presented by Taizo Iijima in a technical paper from 1959 [191]
followed by a journal version in 1962 [192]. Both papers are written in Japanese.
In [192] Iijima considers an observation transformation Φ which depends on
a scale parameter σ and which transforms the original image f(x) into a blurred
version
4
Φ[f(x
′
), x, σ]. This class of blurring transformations is called boke (defo
cusing). He assumes that it has the structure
Φ[f(x
′
), x, σ] =
∞
−∞
φ¦f(x
′
), x, x
′
, σ¦ dx
′
, (1.19)
4
The variable x
′
serves as a dummy variable.
8 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
and that it should satisfy ﬁve conditions:
(I) Linearity (with respect to multiplications):
If the intensity of a pattern becomes A times its original intensity, then the
same should happen to the observed pattern:
Φ[Af(x
′
), x, σ] = AΦ[f(x
′
), x, σ]. (1.20)
(II) Translation invariance:
Filtering a translated image is the same as translating the ﬁltered image:
Φ[f(x
′
−a), x, σ] = Φ[f(x
′
), x−a, σ]. (1.21)
(III) Scale invariance:
If a pattern is spatially enlarged by some factor λ, then there exists a σ
′
=
σ
′
(σ, λ) such that
Φ[f(x
′
/λ), x, σ] = Φ[f(x
′
), x/λ, σ
′
]. (1.22)
(IV) (Generalized) semigroup property:
If f is observed under a parameter σ
1
and this observation is observed un
der a parameter σ
2
, then this is equivalent to observing f under a suitable
parameter σ
3
= σ
3
(σ
1
, σ
2
):
Φ
Φ[f(x
′′
), x
′
, σ
1
], x, σ
2
= Φ[f(x
′′
), x, σ
3
]. (1.23)
(V) Preservation of positivity:
If the original image is positive, then the observed image is positive as well:
Φ[f(x
′
), x, σ] > 0 ∀ f(x
′
) > 0, ∀ σ > 0. (1.24)
Under these requirements Iijima derives in a very systematic way that
Φ[f(x
′
), x, σ] =
1
2
√
πσ
∞
−∞
f(x
′
) exp
−(x −x
′
)
2
4σ
2
dx
′
. (1.25)
Thus, Φ[f(x
′
), x, σ] is just the convolution between f and a Gaussian with standard
deviation σ
√
2.
This has been the starting point of an entire world of linear scalespace research
in Japan, which is basically unknown in the western world. Japanese scalespace
theory was wellembedded in a general framework for pattern recognition, feature
extraction and object classiﬁcation [195, 197, 200, 320], and many results have
1.2 LINEAR DIFFUSION FILTERING 9
been established earlier than in the western world. Apart from their historical
merits, these Japanese results reveal many interesting qualities which should induce
everyone who is interested in scalespace theory to have a closer look at them. More
details can be found in [426, 427] as well as in some English scalespace papers
by Iijima such as [195, 197]. In particular, the latter ones show that there is no
justiﬁcation to deny Iijima’s pioneering role in linear scalespace theory because of
language reasons.
Table 1.1: Overview of continuous Gaussian scalespace axiomatics (I1 = Iijima
[191, 192], I2 = Iijima [193, 194], I3 = Iijima [196], O = Otsu [320], K = Koenderink
[240], Y = Yuille/Poggio [450], B = Babaud et al. [30], L1 = Lindeberg [255], F1 =
Florack et al. [140], A = Alvarez et al. [12], P = Pauwels et al. [324], N = Nielsen
et al. [303], L2 = Lindeberg [257], F2 = Florack [138]).
I1 I2 I3 O K Y B L1 F1 A P N L2 F2
convolution kernel • • • • • • • • • • •
semigroup property • • • • • • • • •
locality •
regularity • • • • • • • •
inﬁnetes. generator •
max. loss principle •
causality • • • • •
nonnegativity • • • • • •
Tikhonov regulariz. •
aver. grey level invar. • • • • • •
ﬂat kernel for t →∞ • •
isometry invariance • • • • • • • • • • •
homogen. & isotropy •
separability • •
scale invariance • • • • • • • •
valid for dimension 1 2 2 2 1,2 1,2 1 1 > 1 N 1,2 N N N
Table 1.1 presents an overview of the current Japanese and western Gaussian
scalespace axiomatics (see [426, 427] for detailed explanations). All of these ax
iomatics use explicitly or implicitly
5
a linearity assumption. We observe that –
despite the fact that many axiomatics reveal similar ﬁrst principles – not two of
them are identical. Each of the 14 axiomatics conﬁrms and enhances the evidence
that the others give: that Gaussian scalespace is unique within a linear framework.
A detailed treatment of Gaussian scalespace theory can be found in two
Japanese monographs by Iijima [197, 198], as well as in English books by Lin
deberg [256], Florack [139], and ter Haar Romeny [176]. A collection edited by
5
Often it is assumed that the ﬁlter is a convolution integral. This is equivalent to linearity
and translation invariance.
10 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
Sporring, Nielsen, Florack and Johansen [389] gives an excellent overview of the
various aspects of this theory, and additional material is presented in [211]. Many
relations between Gaussian scalespace and regularization theory have been elab
orated by Nielsen [302], and readers who wish to analyse linear and nonlinear
scalespace concepts in terms of diﬀerential and integral geometry can ﬁnd a lot
of material in the thesis of Salden [351].
1.2.3 Numerical aspects
The preceding theory is entirely continuous. However, in practical problems, the
image is sampled at the nodes (pixels) of a ﬁxed equidistant grid. Thus, the diﬀu
sion ﬁlter has to be discretized.
By virtue of the equivalence of solving the linear diﬀusion equation and con
volving with a Gaussian, we can either approximate the convolution process or the
diﬀusion equation.
When restricting the image to a ﬁnite domain and applying the Fast Fourier
Transformation (FFT), convolution in the spatial domain can be reduced to mul
tiplication in the frequency domain, cf. (1.7). This proceeding requires a ﬁxed
computational eﬀort of order N log N, which depends only on the pixel number
N, but not on the kernel size σ. For large kernels this is faster than most spatial
techniques. Especially for small kernels, however, aliasing eﬀects in the Fourier
domain may create oscillations and over and undershoots [178].
One eﬃcient possibility to approximate Gaussian convolution in the spatial do
main consists of applying recursive ﬁlters [109, 448]. More frequently the Gaussian
kernel is just sampled and truncated at some multiple of its standard deviation
σ. Factorizing a higherdimensional Gaussian into onedimensional Gaussians re
duces the computational eﬀort to O(Nσ). Convolution with a truncated Gaussian,
however, reveals the drawback that it does not preserve the semigroup property of
the continuous Gaussian scalespace [255].
Lindeberg [255] has established a linear scalespace theory for the semidiscrete
6
case. His results are in accordance with those of Norman [312], who proposed in
1960 that the discrete analogue of the Gaussian kernel should be given in terms of
modiﬁed Bessel functions of integer order. Since this scalespace family arises nat
urally from a semidiscretized version of the diﬀusion equation, it has been argued
that approximating the diﬀusion equation should be preferred to discretizing the
convolution integral [255].
Recently, interesting semidiscrete and fully discrete linear scalespace formula
tions have been established utilizing stochastic principles:
˚
Astr¨om and Heyden [27]
study a framework based on stationary random ﬁelds, while the theory by Salden
et al. [353] exploits the relations between diﬀusion and Markov processes.
6
By semidiscrete we mean discrete in space and continuous in time throughout this work.
1.2 LINEAR DIFFUSION FILTERING 11
Among the numerous numerical possibilities to approximate the linear diﬀusion
equation, ﬁnite diﬀerence (FD) schemes dominate the ﬁeld. Apart from some im
plicit approaches [166, 67, 68] allowing realizations as a recursive ﬁlter [14, 10, 451],
explicit schemes are mainly used. A very eﬃcient approximation of the Gaus
sian scalespace results from applying multigrid ideas. The Gaussian pyramid [64]
has the computational complexity O(N) and gives a multilevel representation at
ﬁnitely many scales of diﬀerent resolution. By subsequently smoothing the image
with an explicit scheme for the diﬀusion equation and restricting the result to a
coarser grid, one obtains a simpliﬁed image representation at the next coarser grid.
Due to their simplicity and eﬃciency, pyramid decompositions have become very
popular and have been integrated into commercially available hardware [70, 214].
Pyramids are not invariant under translations, however, and sometimes it is ar
gued that they are undersampled and that the pyramid levels should be closer
7
.
These are the reasons why some people regard pyramids rather as predecessors of
the scalespace idea than as a numerical approximation
8
.
1.2.4 Applications
Due to its equivalence to convolution with a Gaussian, linear diﬀusion ﬁltering has
been applied in numerous ﬁelds of image processing and computer vision. It can
be found in almost every standard textbook in these ﬁelds.
Less frequent are applications which exploit the evolution of an image under
Gaussian scalespace. This deep structure analysis [240] provides useful information
for extracting semantic information from an image, for instance
• for ﬁnding the most relevant scales (scale selection, focusofattention). This
may be done by searching for extrema of (nonlinear) combinations of normal
ized Gaussian derivatives [256] or by analysing information theoretic mea
sures such as the entropy [208, 388] or generalized entropies [390] over scales.
• for multiscale segmentation of images [172, 254, 256, 313, 408]. The idea is
to identify segments at coarse scales and to link backwards to the original
image in order to improve the localization.
In recent years also applications of Gaussian scalespace to stereo, optic ﬂow
and image sequences have become an active research ﬁeld [139, 215, 241, 258, 259,
302, 306, 441]. Several scalespace applications are summarized in a survey paper
by ter Haar Romeny [175].
7
Of course, multiresolution techniques such as pyramids or discrete wavelet transforms [92,
106] are just designed to have few or no redundancies, while scalespace analysis intends to
extract semantical information by tracing signals through a continuum of scales.
8
Historically, this is incorrect: Iijima’s scalespace work [191] is much older than multigrid
ideas in image processing.
12 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
Interesting results arise when one studies linear scalespace on a sphere [236,
353]: while the diﬀusion equation remains the correct concept, Gaussian kernels are
of no use anymore: appropriate kernels have to be expressed in terms of Legendre
functions [236]. This and other results [12, 255] indicate that the PDE formulation
of linear scalespace in terms of a diﬀusion equation is more natural and has a
larger generalization potential than convolution with Gaussians.
1.2.5 Limitations
In spite of several properties that make linear diﬀusion ﬁltering unique and easy
to handle, it reveals some drawbacks as well:
(a) An obvious disadvantage of Gaussian smoothing is the fact that it does not
only smooth noise, but also blurs important features such as edges and, thus,
makes them harder to identify. Since Gaussian smoothing is designed to be
completely uncommitted, it cannot take into account any apriori informa
tion on structures which are worth being preserved (or even enhanced).
(b) Linear diﬀusion ﬁltering dislocates edges when moving from ﬁner to coarser
scales, see e.g. Witkin [439]. So structures which are identiﬁed at a coarse
scale do not give the right location and have to be traced back to the original
image [439, 38, 165]. In practice, relating dislocated information obtained at
diﬀerent scales is diﬃcult and bifurcations may give rise to instabilities. These
coarsetoﬁne tracking diﬃculties are generally denoted as the correspondence
problem.
(c) Some smoothing properties of Gaussian scalespace do not carry over from
the 1D case to higher dimensions: A closed zerocrossing contour can split
into two as the scale increases [450], and it is generally not true that the
number of local extrema is nonincreasing, see [254, 255] for illustrative coun
terexamples. A deep mathematical analysis of such phenomena has been
carried out by Damon [105] and Rieger [342]. It turned out that the pairwise
creation of an extremum and a saddle point is not an exception, but happens
generically.
Regarding (b) and (c), much eﬀorts have been spent in order to understand
the deep structure in Gaussian scalespace, for instance by analysing its toppoints
[210]. There is some evidence that these points, where the gradient vanishes and
the Hessian does not have full rank, carry essential image information [212]. Part
III of the book edited by Sporring et al. [389] and the references therein give an
overview of the stateoftheart in deep structure analysis.
Due to the uniqueness of Gaussian scalespace within a linear framework we
know that any modiﬁcation in order to overcome the problems (a)–(c) will either
1.2 LINEAR DIFFUSION FILTERING 13
renounce linearity or some scalespace properties. We shall see that appropriate
methods to avoid the shortcomings (a) and (b) are nonlinear diﬀusion processes,
while (c) requires morphological equations [206, 207, 218].
1.2.6 Generalizations
Before we turn our attention to nonlinear processes, let us ﬁrst investigate two
linear modiﬁcations which have been introduced in order to address the problems
(a) and (b) from the previous section.
Aﬃne Gaussian scalespace
A straightforward generalization of Gaussian scalespace results from renouncing
invariance under rotations. This leads to the aﬃne Gaussian scalespace
u(x, t) :=
IR
2
1
4π
det(D
t
)
exp
−
(x−y)
⊤
D
−1
t
(x−y)
4
f(y) dy (1.26)
where D
t
:= tD, t > 0, and D ∈ IR
2×2
is symmetric positive deﬁnite
9
. For a ﬁxed
matrix D, calculating the convolution integral (1.26) is equivalent to solving a
linear anisotropic diﬀusion problem with D as diﬀusion tensor:
∂
t
u = div (D ∇u), (1.27)
u(x, 0) = f(x). (1.28)
In [427] it is shown that aﬃne Gaussian scalespace has been axiomatically derived
by Iijima in 1962 [193, 194]. He named u(x, t) the generalized ﬁgure of f, and (1.27)
the basic equation of ﬁgure [196]. In 1971 this concept was realized in hardware in
the optical character reader ASPET/71 [199, 200]. The scalespace part has been
regarded as the reason for its reliability and robustness.
In 1992 Nitzberg and Shiota [310] proposed to adapt the Gaussian kernel shape
to the structure of the original image. By chosing D in (1.26) as a function of the
structure tensor (cf. Section 2.2) of f, they combined nonlinear shape adaptation
with linear smoothing. Later on similar ideas have been developed in [259, 443].
It should be noted that shapeadapted Gaussian smoothing with a spatially
varying D is no longer equivalent to a diﬀusion process of type (1.27). In practice
this can be experienced by the fact that shapeadaptation of Gaussian smoothing
does not preserve the average grey level, while the divergence formulation ensures
that this is still possible for nonuniform diﬀusion ﬁltering; see Section 1.1. Also in
this case the diﬀusion equation seems to be more general. If one wants to relate
9
Isotropic Gaussian scalespace can be recovered using the unit matrix for D.
14 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
shapeadapted Gaussian smoothing to a PDE, one has to carry out sophisticated
scaling limits [310].
Noniterative shapeadapted Gaussian smoothing diﬀers from nonlinear aniso
tropic diﬀusion ﬁltering by the fact that the latter one introduces a feedback into
the process: it adapts the diﬀusion tensor in (1.27) to the diﬀerential structure of
the ﬁltered image instead of the original image. Such concepts will be investigated
in Section 1.3.3 and in the remaining chapters of this book.
Directed diﬀusion
Another method for incorporating apriori knowledge into a linear diﬀusion process
is suggested by Illner and Neunzert [202]. Provided we are given some background
information in form of a smooth image b, they show that under some technical
requirements and suitable boundary conditions the classical solution u of
∂
t
u = b ∆u −u ∆b, (1.29)
u(x, 0) = f(x) (1.30)
converges to b along a path where the relative entropy with respect to b increases in
a monotone way. Numerical experiments have been carried out by Giuliani [159],
and an analysis in terms of nonsmooth b and weak solutions is due to Illner and
Tie [203].
Such a directed diﬀusion process requires to specify an entire image as back
ground information in advance; in many applications it would be desirable to
include a priori knowledge in a less speciﬁc way, e.g. by prescribing that features
within a certain contrast and scale range are considered to be semantically impor
tant and processed diﬀerently. Such demands can be satisﬁed by nonlinear diﬀusion
ﬁlters.
1.3 Nonlinear diﬀusion ﬁltering
Adaptive smoothing methods are based on the idea of applying a process which
itself depends on local properties of the image. Although this concept is well
known in the image processing community (see [349] and the references therein
for an overview), a corresponding PDE formulation was ﬁrst given by Perona and
Malik [326] in 1987. We shall discuss this model in detail, especially its illposedness
aspects. This gives rise to study regularizations. These techniques can be extended
to anisotropic processes which make use of an adapted diﬀusion tensor instead of
a scalar diﬀusivity.
1.3 NONLINEAR DIFFUSION FILTERING 15
1.3.1 The Perona–Malik model
Basic idea
Perona and Malik propose a nonlinear diﬀusion method for avoiding the blurring
and localization problems of linear diﬀusion ﬁltering [326, 328]. They apply an
inhomogeneous process that reduces the diﬀusivity at those locations which have
a larger likelihood to be edges. This likelihood is measured by [∇u[
2
. The Perona–
Malik ﬁlter is based on the equation
∂
t
u = div (g([∇u[
2
) ∇u). (1.31)
and it uses diﬀusivities such as
g(s
2
) =
1
1 +s
2
/λ
2
(λ > 0). (1.32)
Although Perona and Malik name their ﬁlter anisotropic, it should be noted that
– in our terminology – it would be regarded as an isotropic model, since it utilizes
a scalarvalued diﬀusivity and not a diﬀusion tensor.
Interestingly, there exists a relation between (1.31) and the neural dynamics of
brightness perception: In 1984 Cohen and Grossberg [94] proposed a model of the
primary visual cortex with similar inhibition eﬀects as in the Perona–Malik model.
The experiments of Perona and Malik were visually very impressive: edges
remained stable over a very long time. It was demonstrated [328] that edge de
tection based on this process clearly outperforms the linear Canny edge detector,
even without applying nonmaxima suppression and hysteresis thresholding. This
is due to the fact that diﬀusion and edge detection interact in one single process
instead of being treated as two independent processes which are to be applied
subsequently. Moreover, there is another reason for the impressive behaviour at
edges, which we shall discuss next.
Edge enhancement
To study the behaviour of the Perona–Malik ﬁlter at edges, let us for a moment
restrict ourselves to the onedimensional case. This simpliﬁes the notation and
illustrates the main behaviour since near a straight edge a twodimensional image
approximates a function of one variable.
For the diﬀusivity (1.32) it follows that the ﬂux function Φ(s) := sg(s
2
) satisﬁes
Φ
′
(s) ≥ 0 for [s[ ≤ λ, and Φ
′
(s) < 0 for [s[ > λ, see Figure 1.1. Since (1.31) can
be rewritten as
∂
t
u = Φ
′
(u
x
)u
xx
, (1.33)
we observe that – in spite of its nonnegative diﬀusivity – the Perona–Malik model
is of forward parabolic type for [u
x
[ ≤λ, and of backward parabolic type for [u
x
[ >λ.
16 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
0
1
0 lambda
diffusivity
0
0.6
0 lambda
flux function
Figure 1.1: (a) Left: Diﬀusivity g(s
2
)=
1
1+s
2
/λ
2
. (b) Right: Flux function
Φ(s)=
s
1+s
2
/λ
2
.
Hence, λ plays the role of a contrast parameter separating forward (low contrast)
from backward (high contrast) diﬀusion areas.
It is not hard to verify that the Perona–Malik ﬁlter increases the slope at
inﬂection points of edges within a backward area: If there exists a suﬃciently
smooth solution u it satisﬁes
∂
t
(u
2
x
) = 2u
x
∂
x
(u
t
) = 2Φ
′′
(u
x
)u
x
u
2
xx
+ 2Φ
′
(u
x
)u
x
u
xxx
. (1.34)
A location x
0
where u
2
x
is maximal at some time t is characterized by u
x
u
xx
=0
and u
x
u
xxx
≤0. Therefore,
(∂
t
(u
2
x
)) (x
0
, t) ≥ 0 for [u
x
(x
0
, t)[ > λ (1.35)
with strict inequality for u
x
u
xxx
<0.
In the twodimensional case, (1.33) is replaced by [12]
∂
t
u = Φ
′
(∇u)u
ηη
+g([∇u[
2
)u
ξξ
(1.36)
where the gauge coordinates ξ and η denote the directions perpendicular and paral
lel to ∇u, respectively. Hence, we have forward diﬀusion along isophotes (i.e. lines
of constant grey value) combined with forward–backward diﬀusion along ﬂowlines
(lines of maximal grey value variation).
We observe that the forward–backward diﬀusion behaviour is not only restricted
to the special diﬀusivity (1.32), it appears for all diﬀusivities g(s
2
) whose rapid
decay causes nonmonotone ﬂux functions Φ(s) = sg(s
2
). Overviews of several
common diﬀusivities for the Perona–Malik model can be found in [43, 343], and
a family of diﬀusivities with diﬀerent decay rates is investigated in [36]. Rapidly
decreasing diﬀusivities are explicitly intended in the Perona–Malik method as they
1.3 NONLINEAR DIFFUSION FILTERING 17
give the desirable result of blurring small ﬂuctuations and sharpening edges. There
fore, they are the main reason for the visually impressive results of this restoration
technique.
It is evident that the “optimal” value for the contrast parameter λ has to depend
on the problem. Several proposals have been made to facilitate such a choice in
practice, for instance adapting it to a speciﬁed quantile in the cumulative gradient
histogramme [328], using statistical properties of a training set of regions which
are considered as ﬂat [444], or estimating it by means of the local image geometry
[270].
Illposedness
Unfortunately, forward–backward equations of Perona–Malik type cause some the
oretical problems. Although there is no general theory for nonlinear parabolic
processes, there exist certain frameworks which allow to establish wellposedness
results for a large class of equations. Let us recall three examples:
• Let S(N) denote the set of symmetric N N matrices and Hess(u) the
Hessian of u. Classical diﬀerential inequality techniques [411] based on the
Nagumo–Westphal lemma require that the underlying nonlinear evolution
equation
∂
t
u = F(t, x, u, ∇u, Hess(u)) (1.37)
satisﬁes the monotony property
F(t, x, r, p, Y ) ≥ F(t, x, r, p, X) (1.38)
for all X, Y ∈S(2) where Y −X is positive semideﬁnite.
• The same requirement is needed for applying the theory of viscosity solutions.
A detailed introduction into this framework can be found in a paper by
Crandall, Ishii and Lions [103].
• Let H be a Hilbert space with scalar product (., .) and A : H →H . In order
to apply the concept of maximal monotone operators [57] to the problem
du
dt
+Au = 0, (1.39)
u(0) = f (1.40)
one has to ensure that A is monotone, i.e.
(Au−Av, u−v) ≥ 0 ∀ u, v ∈ H. (1.41)
18 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
We observe that the nonmonotone ﬂux function of the Perona–Malik process im
plies that neither (1.38) is satisﬁed nor A deﬁned by Au := −div (g([∇u[
2
) ∇u)
is monotone. Therefore, none of these frameworks is applicable to ensure well
posedness results.
One reason why people became pessimistic about the wellposedness of the
Perona–Malik equation was a result by H¨ollig [187]. He constructed a forward–
backward diﬀusion process which can have inﬁnitely many solutions. Although this
process was diﬀerent from the Perona–Malik process, one was warned what can
happen. In 1994 the general conjecture was that the Perona–Malik ﬁlter might have
weak solutions, but one should neither expect uniqueness nor stability [329]. In the
meantime several theoretical results are available which provide some insights into
the actual degree of illposedness of the Perona–Malik ﬁlter.
Kawohl and Kutev [222] proved that the Perona–Malik process does not have
global (weak) C
1
solutions for intial data that involve backward diﬀusion. The exis
tence of local C
1
solutions remained unproven. If they exist, however, Kawohl and
Kutev showed that these solutions are unique and satisfy a maximumminimum
principle. Moreover, under special assumptions on the initial data, it was possible
to establish a comparison principle.
Kichenassamy [224, 225] proposed a notion of generalized solutions, which are
piecewise linear and contain jumps, and he showed that an analysis of their moving
and merging gives similar eﬀects to those one can observe in practice.
Results of You et al. [446] give evidence that the Perona–Malik process is
unstable with respect to perturbations of the initial image. They showed that the
energy functional leading to the Perona–Malik process as steepest descent method
has an inﬁnite number of global minima which are dense in the image space. Each
of these minima corresponds to a piecewise constant image, and slightly diﬀerent
initial images may end up in diﬀerent minima for t →∞.
Interestingly, forward–backward diﬀusion equations of Perona–Malik type are
not as unnatural as they look at ﬁrst glance: besides their importance in computer
vision they have been proposed as a mathematical model for heat and mass transfer
in a stably stratiﬁed turbulent shear ﬂow. Such a model is used to explain the
evolution of stepwise constant temperature or salinity proﬁles in the ocean. Related
equations also play a role in population dynamics and viscoelasiticity, see [35] and
the references therein.
Numerically, the mainly observable instability is the socalled staircasing eﬀect,
where a sigmoid edge evolves into piecewise linear segments which are separated
by jumps. It has already been observed by Posmentier in 1977 [333]. He used an
equation of Perona–Malik type for numerical simulations of the salinity proﬁles
in oceans. Starting from a smoothly increasing initial distribution he reported the
creation of perturbations which led to a stepwise constant proﬁle after some time.
1.3 NONLINEAR DIFFUSION FILTERING 19
In image processing, numerical studies of the staircasing eﬀect have been carried
out by Nitzberg and Shiota [310], Fr¨ohlich and Weickert [148], and Benhamouda
[36]. All results point in the same direction: the number of created plateaus de
pends strongly on the regularizing eﬀect of the discretization. Finer discretizations
are less regularizing and lead to more “stairs”. Weickert and Benhamouda [425]
showed that the regularizing eﬀect of a standard ﬁnite diﬀerence discretization is
suﬃcient to turn the Perona–Malik ﬁlter into a wellposed initial value problem
for a nonlinear system of ordinary diﬀerential equations. Its global solution satis
ﬁes a maximum–minimum principle and converges to a constant steadystate. The
theoretical framework for this analysis will be presented in Chapter 3.
There exists also a discrete explanation why staircasing is essentially the only
observable instability: In 1D, standard FD discretizations are monotonicity pre
serving, which guarantees that no additional oscillations occur during the evolu
tion. This has been shown by Dzu Magaziewa [123] in the semidiscrete case and
by Benhamouda [36, 425] in the fully discrete case with an explicit time discretiza
tion. Further contributions to the explanation and avoidance of staircasing can be
found in [4, 36, 98, 225, 438].
Scalespace interpretation
Perona and Malik renounced the assumption of Koenderink’s linear scalespace
axiomatic [240] that the smoothing should treat all spatial points and scale levels
equally. Instead of this, they required that region boundaries should be sharp and
should coincide with the semantically meaningful boundaries at each resolution
level (immediate localization), and that intraregion smoothing should be preferred
to interregion smoothing (piecewise smoothing). These properties are of signiﬁcant
practical interest, as they guarantee that structures can be detected easily and
correspondence problems can be neglected. Experiments demonstrated that the
Perona–Malik ﬁlter satisﬁes these requirements fairly well [328].
In order to establish a smoothing scalespace property for this nonlinear dif
fusion process, a natural way would be to prove a maximum–minimum principle,
provided one knows that there exists a suﬃciently smooth solution. Since the ex
istence question used to be the bottleneck in the past, the ﬁrst proof is due to
Kawohl and Kutev who established an extremum principle for their local weak C
1
solution to the Perona–Malik ﬁlter [222]. Of course, this is only partly satisfying,
since in scalespace theory one is interested in having an extremum principle for
the entire time interval [0, ∞).
Nevertheless, also other attempts to apply scalespace frameworks to the Perona–
Malik process have not been more successful yet:
• Salden [350], Florack [143] and Eberly [124] proposed to carry over the linear
scalespace theory to the nonlinear case by considering nonlinear diﬀusion
20 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
processes which result from special rescalings of the linear one. Unfortunately,
the Perona–Malik ﬁlter turned out not to belong to this class [143].
• Alvarez, Guichard, Lions and Morel [12] have developed a nonlinear scale
space axiomatic which comprises the linear scalespace theory as well as
nonlinear morphological processes (which we will discuss in 1.5 and 1.6).
Their smoothing axiom is a monotony assumption (comparison principle)
requiring that the scalespace is orderpreserving:
f ≤ g =⇒ T
t
f ≤ T
t
g ∀ t ≥ 0. (1.42)
This property is closely related to a maximum–minimum principle and to
L
∞
stability of the solution [12, 261]. However, the Perona–Malik model
does not ﬁt into this framework, because its local weak solution satisﬁes a
comparison principle only for some ﬁnite time, but not for all t > 0; see [222].
1.3.2 Regularized nonlinear models
It has already been mentioned that numerical schemes may provide implicit reg
ularizations which stabilize the Perona–Malik process [425]. Hence, it has been
suggested to introduce the regularization directly into the continuous equation in
order to become more independent of the numerical implementation [81, 310].
Since the dynamics of the solution may critically depend on the sort of regu
larization, one should adjust the regularization to the desired goal of the forward–
backward heat equation [35]. One can apply spatial or temporal regularization
(and of course, a combination of both). Below we shall discuss three examples
which illustrate the variety of possibilities and their tailoring towards a speciﬁc
task.
(a) The ﬁrst spatial regularization attempt is probably due to Posmentier who
observed numerically the stabilizing eﬀect of averaging the gradient within
the diﬀusivity [333].
A mathematically sound formulation of this idea is given by Catt´e, Lions,
Morel and Coll [81]. By replacing the diﬀusivity g([∇u[
2
) of the Perona–
Malik model by a Gaussiansmoothed version g([∇u
σ
[
2
) with u
σ
:= K
σ
∗u
they end up with
∂
t
u = div (g([∇u
σ
[
2
) ∇u). (1.43)
In [81] existence, uniqueness and regularity of a solution for σ> 0have been
established.
This process has been analysed and modiﬁed in many ways: Whitaker and
Pizer [438] have suggested that the regularization parameter σ should be
1.3 NONLINEAR DIFFUSION FILTERING 21
a decreasing function in t, and Li and Chen [252] have proposed to subse
quently decrease the contrast parameter λ. A detailed study of the inﬂuence
of the parameters in a regularized Perona–Malik model has been carried out
by Benhamouda [36]. Kaˇcur and Mikula [217] have investigated a modiﬁca
tion which allows to diﬀuse diﬀerently in diﬀerent grey value ranges. Spatial
regularizations of the Perona–Malik process leading to anisotropic diﬀusion
equations have been proposed by Weickert [413, 415] and will be described
in 1.3.3. Torkamani–Azar and Tait [403] suggest to replace the Gaussian
convolution by the exponential ﬁlter of Shen and Castan
10
[381].
In Chapter 2 we shall see that spatial regularizations lead to wellposed scale
spaces with a large class of Lyapunov functionals which guarantee that the
solution converges to a constant steadystate.
From a practical point of view, spatial regularizations oﬀer the advantage
that they make the ﬁlter insensitive to noise at scales smaller than σ. There
fore, when regarding (1.43) as an image restoration equation, it exhibits
besides the contrast parameter λ an additional noise scale σ. This avoids a
shortcoming of the genuine Perona–Malik process which misinterprets strong
oscillations due to noise as edges which should be preserved or even enhanced.
Examples for spatially regularized nonlinear diﬀusion ﬁltering can be found
in Figure 5.2 (c) and 5.4 (a),(b).
(b) P.L. Lions proved in a private communication to Mumford that the one
dimensional process
∂
t
u = ∂
x
(g(v) ∂
x
u), (1.44)
∂
t
v =
1
τ
([∂
x
u[
2
−v) (1.45)
leads to a wellposed ﬁlter (cf. [329]). We observe that v is intended as a
timedelay regularization of [∂
x
u[
2
where the parameter τ > 0 determines
the delay. These equations arise as a special case of the spatiotemporal
regularizations of Nitzberg and Shiota [310] when neglecting any spatial reg
ularization. Mumford conjectures that this model gives piecewise constant
steadystates. In this case, the steadystate solution would solve a segmen
tation problem.
(c) In the context of shear ﬂows, Barenblatt et al. [35] regularized the one
dimensional forward–backward heat equation by considering the thirdorder
equation
∂
t
u = ∂
x
(Φ(u
x
)) +τ∂
xt
(Ψ(u
x
)) (1.46)
10
This renounces invariance under rotation.
22 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
where Ψ is strictly increasing and uniformly bounded in IR, and [Φ
′
(s)[ =
O(Ψ
′
(s)) as s → ±∞. This regularization was physically motivated by in
troducing a relaxation time τ into the diﬀusivity.
For the corresponding initial boundary value problem with homogeneous
Neumann boundary conditions they proved the existence of a unique gen
eralized solution. They also showed that smooth solutions may become dis
continuous within ﬁnite time, before they ﬁnally converge to a piecewise
constant steadystate.
These examples demonstrate that regularization is much more than stabilizing
an illposed process: Regularization is modeling. Appropriately chosen regulariza
tions create the desired ﬁlter features. We observe that spatial regularizations are
closer to scalespace ideas while temporal regularization are more related to image
restoration and segmentation, since they may lead to nontrivial steadystates.
1.3.3 Anisotropic nonlinear models
All nonlinear diﬀusion ﬁlters that we have investigated so far utilize a scalarvalued
diﬀusivity g which is adapted to the underlying image structure. Therefore, they
are isotropic and the ﬂux j = −g∇u is always parallel to ∇u. Nevertheless, in
certain applications it would be desirable to bias the ﬂux towards the orientation
of interesting features. These requirements cannot be satisﬁed by a scalar diﬀu
sivity anymore, a diﬀusion tensor leading to anisotropic diﬀusion ﬁlters has to be
introduced.
First anisotropic ideas in image processing date back to Graham [167] in 1962,
followed by Newman and Dirilten [300], Lev, Zucker and Rosenfeld [250], and
Nagao and Matsuyama [297]. They used convolution masks that depended on
the underlying image structure. Related statistical approaches were proposed by
Knutsson, Wilson and Granlund [237]. These ideas have been further developed
by Nitzberg and Shiota [310], Lindeberg and G˚arding [259], and Yang et al. [443].
Their suggestion to use shapeadapted Gaussian masks has been discussed in Sec
tion 1.2.6.
Anisotropic diﬀusion ﬁlters usually apply spatial regularization strategies
11
. A
general theoretical framework for spatially regularized anisotropic diﬀusion ﬁlters
will be presented in the remaining chapters of this book.
Below we study two representatives of anisotropic diﬀusion processes. The ﬁrst
one oﬀers advantages at noisy edges, whereas the second one is welladapted to the
processing of onedimensional features. They are called edgeenhancing anisotropic
diﬀusion and coherenceenhancing anisotropic diﬀusion, respectively.
11
An exception is the timedelay regularization of Cottet and ElAyyadi [100, 101].
1.3 NONLINEAR DIFFUSION FILTERING 23
(a) Anisotropic regularization of the Perona–Malik process
In the interior of a segment the nonlinear isotropic diﬀusion equation (1.43)
behaves almost like the linear diﬀusion ﬁlter (1.9), but at edges diﬀusion
is inhibited. Therefore, noise at edges cannot be eliminated successfully by
this process. To overcome this problem, a desirable method should prefer
diﬀusion along edges to diﬀusion perpendicular to them.
Anisotropic models do not only take into account the modulus of the edge
detector ∇u
σ
, but also its direction. To this end, we construct the orthonor
mal system of eigenvectors v
1
, v
2
of the diﬀusion tensor D such that they
reﬂect the estimated edge structure:
v
1
 ∇u
σ
, v
2
⊥ ∇u
σ
. (1.47)
In order to prefer smoothing along the edge to smoothing across it, Weickert
[415] proposed to choose the corresponding eigenvalues λ
1
and λ
2
as
λ
1
(∇u
σ
) := g([∇u
σ
[
2
), (1.48)
λ
2
(∇u
σ
) := 1. (1.49)
Section 5.1 presents several examples where this process is applied to test
images.
In general, ∇u does not coincide with one of the eigenvectors of D as long
as σ>0. Hence, this model behaves really anisotropic. If we let the regular
ization parameter σ tend to 0, we end up with the isotropic Perona–Malik
process.
Another anisotropic model which can be regarded as a regularization of an
isotropic nonlinear diﬀusion ﬁlter has been described in [413].
(b) Anisotropic models for smoothing onedimensional objects
A second motivation for introducing anisotropy into diﬀusion processes arises
from the wish to process onedimensional features such as linelike structures.
To this end, Cottet and Germain [102] constructed a diﬀusion tensor with
eigenvectors as in (1.47) and corresponding eigenvalues
λ
1
(∇u
σ
) := 0, (1.50)
λ
2
(∇u
σ
) :=
η[∇u
σ
[
2
1 + ([∇u
σ
[/σ)
2
(η > 0). (1.51)
This is a process diﬀusing solely perpendicular to ∇u
σ
. For σ → 0, we
observe that ∇u becomes an eigenvector of D with corresponding eigenvalue
0. Therefore, the process stops completely. In this sense, it is not intended as
an anisotropic regularization of the Perona–Malik equation. Wellposedness
24 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
results for the Cottet–Germain ﬁlter comprise an existence proof for weak
solutions.
Since the Cottet–Germain model diﬀuses only in one direction, it is clear
that its result depends very much on the smoothing direction. For enhancing
parallel linelike structures, one can improve this model when replacing ∇u
⊥
σ
by a more robust descriptor of local orientation, the structure tensor (cf.
Section 2.2). This leads to coherenceenhancing anisotropic diﬀusion [418],
which shall be discussed in Section 5.2, where also many examples can be
found.
1.3.4 Generalizations
Higher dimensions. It is easily seen that many of the previous results can
be generalized to higher dimensions. This may be useful when considering e.g.
medical image sequences from computerized tomography (CT) or magnetic reso
nance imaging (MRI), or when applying diﬀusion ﬁlters to the postprocessing of
ﬂuctuating higherdimensional numerical data. The ﬁrst threedimensional non
linear diﬀusion ﬁlters have been investigated by Gerig et al. [155] in the isotropic
case and by Rambaux and Gar¸ con [339] in the anisotropic case. A generalization
of coherenceenhancing anisotropic diﬀusion to higher dimensions is proposed in
[428], and S´anchez–Ortiz et al. [355] describe nonlinear diﬀusion ﬁltering of 3D
image sequences by treating them as 4D data sets.
More sophisticated structure descriptors. The edge detector ∇u
σ
en
ables us to adapt the diﬀusion to magnitude and direction of edges, but it can
neither distinguish between edges and corners nor does it give a reliable measure
of local orientation. As a remedy, one can steer the smoothing process by more
advanced structure descriptors such as higherorder derivatives [127] or tensor
valued expressions of ﬁrstorder derivatives [414, 418]. The theoretical analysis in
the present work shall comprise the second possibility. It has also been proposed
to replace ∇u
σ
by a Bayesian classiﬁcation result in feature space [26].
Vectorvalued models. Vectorvalued images can arise either from devices
measuring multiple physical properties or from a feature analysis of one single
image. Examples for the ﬁrst category are colour images, multispectral Landsat
exposures and multispin echo MR images, whereas representatives of the second
class are given by statistical moments or the jet space induced by the image itself
and its partial derivatives up to a given order. Feature vectors play an important
role for tasks like texture segmentation.
1.3 NONLINEAR DIFFUSION FILTERING 25
The simplest idea how to apply diﬀusion ﬁltering to multichannel images would
be to diﬀuse all channels separately and independently from each other. This leads
to the undesirable eﬀect that edges may be formed at diﬀerent locations for each
channel. In order to avoid this, one should use a common diﬀusivity which combines
information from all channels. Such isotropic vectorvalued diﬀusion models were
studied by Gerig et al. [155, 156] and Whitaker [433, 434] in the context of medical
imagery. Extensions to anisotropic vectorvalued models with a common tensor
valued structure descriptor for all channels have been investigated by Weickert
[422].
1.3.5 Numerical aspects
For nonlinear diﬀusion ﬁltering numerous numerical methods have been applied:
Finite element techniques are described in [367, 391, 34, 216]. B¨ansch and
Mikula reported a signiﬁcant speedup by supplementing them with an adaptive
mesh coarsening [34]. Neural network approximations to nonlinear diﬀusion ﬁlters
are investigated by Cottet [100, 99] and Fischl and Schwartz [137]. Perona and
Malik [327] propose hardware realizations by means of analogue VLSI networks
with nonlinear resistors. A very detailed VLSI proposal has been developed by
Gijbels et al. [158].
In [148] three schemes for a spatially regularized 1D Perona–Malik ﬁlter are
compared: a wavelet method of Petrov–Galerkin type, a pseudospectral method
and a ﬁnitediﬀerence scheme. It turned out that all results became fairly similar,
when the regularization parameter σ was suﬃciently large. Since the computational
eﬀort is of a comparable order of magnitude, it seems to be a matter of taste which
scheme is preferred.
Most implementations of nonlinear diﬀusion ﬁlters are based on ﬁnite diﬀer
ence methods, since they are easy to handle and the pixel structure of digital
images already provides a natural discretization on a ﬁxed rectangular grid. Ex
plicit schemes are the most simple to code and, therefore, they are used almost
exclusively. Due to their local behaviour, they are wellsuited for parallel architec
tures. Nevertheless, they suﬀer from the fact that fairly small time step sizes are
needed in order to ensure stability. Semiimplicit schemes – which approximate the
diﬀusivity or the diﬀusion tensor in an explicit way and the rest implicitly – are
considered in [81]. They possess much better stability properties. A fast multigrid
technique using a pyramid algorithm for the Perona–Malik ﬁlter has been studied
by Acton et al. [5, 4]; see also [349] for related ideas.
While the preceding techniques are focusing on approximating a continuous
equation, it is often desirable to have a genuinely discrete theory which guarantees
that an algorithm exactly reveals the same qualitative properties as its continuous
counterpart. Such a framework is presented in [420, 421], both for the semidiscrete
26 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
Table 1.2: Requirements for continuous, semidiscrete and fully dis
crete nonlinear diﬀusion scalespace.
requirement continuous semidiscrete discrete
u
t
= div (D∇u)
du
dt
= A(u)u u
0
= f
u(t = 0) = f u(0) = f u
k+1
= Q(u
k
)u
k
'D∇u, n` = 0
smoothness D ∈ C
∞
A Lipschitz Q continuous
continuous
symmetry D symmetric A symmetric Q symmetric
conservation div form; column sums column sums
reﬂective b.c. are 0 are 1
nonnega positive nonnegative nonnegative
tivity semideﬁnite oﬀdiagonals elements
connectivity uniformly pos. irreducible irreducible;
deﬁnite pos. diagonal
and for the fully discrete case. A detailed treatment of this theory can be found
in Chapter 3 and 4, respectively. Table 1.2 gives an overview of the requirements
which are needed in order to prove wellposedness, average grey value invariance,
causality in terms of an extremum principle and Lyapunov functionals, and con
vergence to a constant steadystate [423].
We observe that the requirements belong to ﬁve categories: smoothness, sym
metry, conservation, nonnegativity and connectivity requirements. These criteria
are easy to check for many discretizations. In particular, it turns out that suitable
explicit and semiimplicit ﬁnite diﬀerence discretizations of many discussed models
create discrete scalespaces. The discrete nonlinear scalespace concept has also led
to the development of fast novel schemes, which are based on an additive operator
splitting (AOS) [424, 430]. Under typical accuracy requirements, they are about
10 times more eﬃcient than the widely used explicit schemes, and a speedup by
another order of magnitude can be achieved by a parallel implementation [431]. A
general framework for AOS schemes will be presented in Section 4.4.2.
1.3.6 Applications
Nonlinear diﬀusion ﬁlters have been applied for postprocessing ﬂuctuating data
[269, 415], for visualizing qualityrelevant features in computer aided quality con
trol [299, 413, 418], and for enhancing textures such as ﬁngerprints [418]. They have
proved to be useful for improving subsampling [144] and line detection [156, 418],
for blind image restoration [445], for scalespace based segmentation algorithms
1.4 METHODS OF DIFFUSION–REACTION TYPE 27
[307, 308], for segmentation of textures [433, 437] and remotely sensed data [6, 5],
and for target tracking in infrared images [65]. Most applications, however, are
concerned with the ﬁltering of medical images [26, 28, 29, 155, 244, 248, 264, 270,
308, 321, 355, 386, 393, 431, 434, 437, 444]. Some of these applications will be
investigated in more detail in Chapter 5.
Besides such speciﬁc problem solutions, nonlinear diﬀusion ﬁlters can be found
in commercial software packages such as the medical visualization tool Analyze.
12
1.4 Methods of diﬀusion–reaction type
This section investigates variational frameworks, in which diﬀusion–reaction equa
tions or coupled systems of them are interpreted as steepest descent minimizers of
suitable energy functionals. This idea connects diﬀusion methods to edge detection
and segmentation ideas.
Besides the variational interpretation there exist other interesting theoretical
frameworks for diﬀusion ﬁlters such as the Markov random ﬁeld and mean ﬁeld
annealing context [152, 153, 247, 251, 328, 387], robust statistics [41], and deter
ministic interactive particle models [279]. Their discussion, however, would lead us
beyond the scope of this book.
1.4.1 Single diﬀusion–reaction equations
Nordstr¨om [311] has suggested to obtain a reconstruction u of a degraded image
f by minimizing the energy functional
E
f
(u, w) :=
Ω
β(u−f)
2
+ w[∇u[
2
+ λ
2
(w−lnw)
dx. (1.52)
The parameters β and λ are positive weights and w : Ω →[0, 1] gives a fuzzy edge
representation: in the interior of a region, w approaches 1 while at edges, w is close
to 0 (as we shall see below).
The ﬁrst summand of E punishes deviations of u from f (deviation cost), the
second term detects unsmoothness of u within each region (stabilizing cost), and
the last one measures the extend of edges (edge cost). Cost terms of these three
types are typical for variational image restoration methods.
The corresponding Euler equations to this energy functional are given by
0 = β(u−f) − div (w∇u), (1.53)
0 = λ
2
(1−
1
w
) + [∇u[
2
, (1.54)
12
Analyze is a registered trademark of Mayo Medical Ventures, 200 First Street SW, Rochester,
MN 55905, U.S.A.
28 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
equipped with a homogeneous Neumann boundary condition for u.
Solving (1.54) for w gives
w =
1
1 +[∇u[
2
/λ
2
. (1.55)
We recognize that w is identical with the Perona–Malik diﬀusivity g([∇u[
2
) in
troduced in (1.32). Therefore, (1.53) can be regarded as the steadystate equation
of
∂
t
u = div (g([∇u[
2
) ∇u) +β(f −u). (1.56)
This equation can also be obtained directly as the descent method of the functional
F
f
(u) :=
Ω
β(u−f)
2
+ λ
2
ln
1+
∇u
2
λ
2
dx. (1.57)
The diﬀusion–reaction equation (1.56) consists of the Perona–Malik process
with an additional bias term β(f −u). One of Nordstr¨om’s motivations for intro
ducing this term was to free the user from the diﬃculty of specifying an appropriate
stopping time for the Perona–Malik process.
However, it is evident that the Nordstr¨om model just shifts the problem of
specifying a stopping time T to the problem of determining β. So it seems to
be a matter of taste which formulation is preferred. People interested in image
restoration usually prefer the reaction term, while for scalespace researchers it is
more natural to have a constant steadystate as the simplest image representation.
Nordstr¨om’s method may suﬀer from the same illposedness problems as the
underlying Perona–Malik equation, and it is not hard to verify that the energy
functional (1.57) is nonconvex. Therefore, it can possess numerous local minima,
and the process (1.56) with f as initial condition does not necessarily converge to
a global minimum. Similar diﬃculties may also arise in other diﬀusion–reaction
models, where convergence results have not yet been established [152, 186].
A popular possibility to avoid these illposedness and convergence problems is
to renounce edgeenhancing diﬀusivities in order to end up with (nonquadratic)
convex functionals [43, 88, 110, 367, 391]. In this case the frameworks of convex
optimization and monotone operators are applicable, ensuring wellposedness and
stability of a standard ﬁniteelement approximation [367].
Diﬀusion–reaction approaches have been applied to edge detection [367, 391],
to the restoration of inverse scattering images [263], to SPECT [88] and vascular
reconstruction in medical imaging [102, 325], and to optic ﬂow [368, 111] and
stereo problems [343]. They can be extended to vectorvalued images [369] and
to cornerpreserving smoothing of curves [136, 323]. Diﬀusion–reaction methods
with constant diﬀusivities have also been used for local contrast normalization in
images [330].
1.4 METHODS OF DIFFUSION–REACTION TYPE 29
1.4.2 Coupled systems of diﬀusion–reaction equations
Mumford and Shah [295, 296] have proposed to obtain a segmented image u from
f by minimizing the functional
E
f
(u, K) = β
Ω
(u−f)
2
dx +
Ω\K
[∇u[
2
dx + α[K[ (1.58)
with nonnegative parameters α and β. The discontinuity set K consists of the
edges, and its onedimensional Hausdorﬀ measure [K[ gives the total edge length.
Like the Nordstr¨om functional (1.52), this expression consists of three cost terms:
the ﬁrst one is the deviation cost, the second one gives the stabilizing cost, and
the third one represents the edge cost.
The Mumford–Shah functional can be regarded as a continuous version of the
Markov random ﬁeld method of Geman and Geman [154] and the weak membrane
model of Blake and Zisserman [42]. Related approaches are also used to model
materials with two phases and a free interface.
The fact that (1.58) leads to a free discontinuity problem causes many challeng
ing theoretical questions [249]. The book of Morel and Solimini [292] covers a very
detailed analysis of this functional. Although the existence of a global minimizer
with a closed edge set K has been established [108, 17], uniqueness is in general
not true [292, pp. 197–198]. Regularity results for K in terms of (at least) C
1
arcs
have recently been obtained [18, 19, 20, 48, 49, 107].
The concept of energy functionals for segmenting images oﬀers the practical
advantage that it provides a framework for comparing the quality of two seg
mentations. On the other hand, (1.58) exhibits also some shortcomings, e.g. the
problem that sigmoidlike edges produce multiple segmentation boundaries (over
segmentation, staircasing eﬀect) [377]. Another drawback results from the fact that
the Mumford–Shah functional allows only singularities which are typical for mini
mal surfaces: Corners or Tjunctions are not possible and segments meet at triple
points with 120
o
angle [296]. In order to avoid such problems, modiﬁcations of the
Mumford–Shah functional have been proposed by Shah [379]. An aﬃne invariant
generalization of (1.58) is investigated in [32, 31] and applied to aﬃne invariant
texture segmentation [31, 33], and a Mumford–Shah functional for curves can be
found in [323].
Since many algorithms in image processing can be restated as versions of the
Mumford–Shah functional [292] and since it is a prototype of a free discontinuity
problem it is instructive to study this variational problem in more detail.
Numerical complications arise from the fact that the Mumford–Shah functional
has numerous local minima. Global minimizers such as the simulated annealing
method used by Geman and Geman [154] are extremely slow. Hence, one searches
for fast (suboptimal) deterministic strategies, e.g. pyramidal regiongrowing algo
rithms [3, 239].
30 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
Another important class of numerical methods is based on the idea to approx
imate the discontinuity set K by a smooth function w, which is close to 0 near
edges of u and which approximates 1 elsewhere.
We may for instance study the functional
F
f
(u, w) :=
Ω
β(u−f)
2
+ w
2
[∇u[
2
+ α
c[∇w[
2
+
(1−w)
2
4c
dx (1.59)
with a positive parameter c specifying the “edge width”. Ambrosio and Tortorelli
proved that this functional converges to the Mumford–Shah functional for c → 0
(in the sense of Γconvergence, see [22] for more details).
Minimizing F
f
corresponds to the gradient descent equations
∂
t
u = div (w
2
∇u) + β(f −u), (1.60)
∂
t
w = c∆w −
w
α
[∇u[
2
+
(1−w)
4c
(1.61)
with homogeneous Neumann boundary conditions. Equations of this type are in
vestigated by Richardson and Mitter [341]. Since (1.60) resembles the Nordstr¨om
process (1.56), similar problems can arise: The functional F
f
is not jointly convex
in u and v, so it may have many local minima and a gradient descent algorithm
may get trapped in a poor local minimum. Wellposedness results for this system
have not been obtained up to now, but a maximum–minimum principle and a local
stability proof have been established.
Another diﬀusion–reaction system is studied by Shah [375, 376]. He replaces
the functional (1.58) by two coupled convex energy functionals and applies gradi
ent descent. This results in ﬁnding an equilibrium state between two competing
processes. Experiments indicate that it converges to a stable solution. Proesmans
et al. [337, 336] observed that this solution looks fairly blurred since the equations
contain diﬀusion terms such as ∆u. They obtained pronounced edges by replacing
such a term by its Perona–Malik counterpart div (g([∇u[
2
) ∇u). Related equations
are also studied in [398]. Of course, this approach gives rise to the same theoretical
questions as (1.60), (1.61).
The system of Richardson and Mitter is used for edge detection [341]. Shah in
vestigates diﬀusion–reaction systems for matching stereo images [378], while Proes
mans et al. apply coupled diﬀusionreaction equations to image sequence analysis,
vectorvalued images and stereo vision [336, 338]. Their ﬁnite diﬀerence algorithms
run on a parallel transputer network.
It should also be mentioned that there exist reaction–diﬀusion systems which
have been applied to image restoration [334, 335, 382], texture generation [406, 440]
and halftoning [382], and which are not connected to Perona–Malik or Mumford–
Shah ideas. They are based on Turing’s pattern formation model [405].
1.5 CLASSIC MORPHOLOGICAL PROCESSES 31
1.5 Classic morphological processes
Morphology is an approach to image analysis based on shapes. Its mathematical
formalization goes back to the group around Matheron and Serra, both working
at ENS des Mines de Paris in Fontainebleau. The theory had ﬁrst been developed
for binary images, afterwards it was extended to greyscale images by regarding
level sets as shapes. Its applications cover biology, medical diagnostics, histology,
quality control, radar and remote sensing, science of material, mineralogy, and
many others.
Morphology is usually described in terms of algebraic set theory, see e.g. [280,
371, 181, 184] for an overview. Nevertheless, PDE formulations for classic morpho
logical processes have been discovered recently by Brockett and Maragos [60], van
den Boomgaard [50], Arehart et al. [25] and Alvarez et al. [12].
This section surveys the basic ideas and elementary operations of binary and
greyscale morphology, presents its PDE representations for images and curves,
and summarizes the results concerning wellposedness and scalespace properties.
Afterwards numerical aspects of the PDE formulation of these processes are dis
cussed, and generalizations are sketched which comprise the morphological equiv
alent of Gaussian scalespace.
1.5.1 Binary and greyscale morphology
Binary morphology considers shapes (silhouettes), i.e. closed sets X⊂IR
2
whose
boundaries are Jordan curves [16]. Henceforth, we identify a shape X with its
characteristic function
χ(x) :=
1 if x∈X,
0 else.
(1.62)
Binary morphological operations aﬀect only the boundary curve of the shape and,
therefore, they can be viewed as curve or shape deformations.
Greyscale morphology generalizes these ideas [274] by decomposing an image
f into its level sets ¦X
a
f, a∈IR¦, where
X
a
f := ¦x ∈ IR
2
, f(x) ≥ a¦. (1.63)
A binary morphological operation A can be extended to some greyscale image f
by deﬁning
X
a
(Af) := A(X
a
f) ∀ a ∈ IR. (1.64)
We observe that for this type of morphological operations only greylevel sets
matter. As a consequence, they are invariant under monotone greylevel rescalings.
This morphological invariance (greyscale invariance) is characteristic for all meth
ods we shall study in Section 1.5 and 1.6, except for 1.5.6 and some modiﬁcations
32 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
in 1.6.5. It is a very desirable property in all cases where brightness changes of the
illumination occur or where one wants to be independent of the speciﬁc contrast
range of the camera. On the other hand, for applications like edge detection or
image restoration, contrast provides important information which should be taken
into account. Moreover, in some cases isolines may give inadequate information
about the depicted physical object boundaries.
1.5.2 Basic operations
Classic morphology analyses a shape by matching it with a socalled structuring
element, a bounded set B⊂IR
2
. Typical shapes for B are discs, squares, or ellipses.
The two basic morphological operations, dilation and erosion with a structuring
element B, are deﬁned for a greyscale image f ∈L
∞
(IR
2
) by [60]
dilation: (f ⊕B) (x) := sup ¦f(x−y), y ∈B¦, (1.65)
erosion: (f ⊖B) (x) := inf ¦f(x+y), y ∈B¦. (1.66)
These names can be easily motivated when considering a shape in a binary image
and a discshaped structuring element. In this case dilation blows up its boundaries,
while erosion shrinks them.
Dilation and erosion form the basis for constructing other morphological pro
cesses, for instance opening and closing:
opening: (f ◦ B) (x) := ((f ⊖B) ⊕B) (x), (1.67)
closing: (f • B) (x) := ((f ⊕B) ⊖B) (x). (1.68)
In the preceding shape interpretation opening smoothes the shape by breaking nar
row isthmuses and eliminating small islands, while closing smoothes by eliminating
small holes, fusing narrow breaks and ﬁlling gaps at the contours [181].
1.5.3 Continuousscale morphology
Let us consider a convex structuring element tB with a scaling parameter t >0.
Then, calculating u(t) = f ⊕tB and u(t) = f ⊖tB, respectively
13
, can be shown
to be equivalent to solving
∂
t
u(x, t) = sup
y∈B
'y, ∇u(x, t)`, (1.69)
∂
t
u(x, t) = inf
y∈B
'y, ∇u(x, t)`. (1.70)
with f as initial condition [12, 360].
13
Henceforth, we frequently use the simpliﬁed notation u(t) instead of u(., t)
1.5 CLASSIC MORPHOLOGICAL PROCESSES 33
By choosing e.g. B := ¦y ∈ IR
2
, [y[ ≤ 1¦ one obtains
∂
t
u = [∇u[, (1.71)
∂
t
u = −[∇u[. (1.72)
The solution u(t) is the dilation (resp. erosion) of f with a disc of radius t and
centre 0 as structuring element. Figure 5.5 (a) presents the temporal evolution of
a test image under such a continuousscale dilation.
Connection to curve evolution
Morphological PDEs such as (1.71) or (1.72) are closely related to shape and
curve evolutions. This can be illustrated by considering a smooth Jordan curve
C : [0, 2π] [0, ∞) →IR
2
,
C(p, t) =
x
1
(p, t)
x
2
(p, t)
(1.73)
where p is the parametrization and t is the evolution parameter. We assume that
C evolves in outer normal direction n with speed β, which may be a function of
its curvature κ :=
det(Cp,Cpp)
Cp
3
:
∂
t
C = β(κ) n, (1.74)
C(p, 0) = C
0
(p). (1.75)
One can embed the curve C(p, t) in an image u(x, t) in such a way that C is just
a level curve of u. The corresponding evolution for u is given by [319, 362, 16]
∂
t
u = β(curv(u)) [∇u[. (1.76)
where the curvature of u is
curv(u) := div
∇u
[∇u[
. (1.77)
Sometimes the image evolution (1.76) is called the Eulerian formulation of the
curve evolution (1.74), because it is written in terms of a ﬁxed coordinate system.
We observe that (1.71) and (1.72) correspond to the simple cases β = ±1.
Hence, they describe the curve evolutions
∂
t
C = ±n. (1.78)
This equation moves level sets in normal direction with constant speed. Such a
process is also named grassﬁre ﬂow or prairie ﬂow. It is closely related to the
Huygens principle for wave propagation [25]. Its importance for shape analysis
in biological vision has already been pointed out in the sixties by Blum [47]. He
simulated grassﬁre ﬂow by a selfconstructed opticomechanical device.
34 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
1.5.4 Theoretical results
Equations such as (1.78) may develop singularities and intersections even for
smooth initial data. Hence, concepts of jump conditions, entropy solutions, and
shocks have to be applied to this shape evolution [228].
A suitable framework for the image evolution equation (1.76) is provided by
the theory of viscosity solutions [103]. The advantage of this analysis is that it
allows us to treat shapes with singularities such as corners, where the classical
solution concept does not apply, but a unique weak solution in the viscosity sense
still exists.
It can be shown [90, 129, 103], that for an initial value
f ∈ BUC(IR
2
) := ¦ϕ∈L
∞
(IR
2
) [ ϕ is uniformly continuous on IR
2
¦ (1.79)
the equations (1.69),(1.70) possess a unique global viscosity solution u(x, t) which
fulﬁls the maximum–minimum principle
inf
IR
2
f ≤ u(x, t) ≤ sup
IR
2
f on IR
2
[0, ∞). (1.80)
Moreover, it is L
∞
stable: for two diﬀerent initial images f, g the corresponding
solutions u(t), v(t) satisfy
u(t)−v(t)
L
∞
(IR
2
)
≤ f −g
L
∞
(IR
2
)
. (1.81)
1.5.5 Scalespace properties
Brockett and Maragos [60] pointed out that the convexity of B is suﬃcient to
ensure the semigroup property of the corresponding dilations and erosions. This
establishes an important architectural scalespace property.
Similar results have been found by van den Boomgaard and Smeulders [53].
Moreover, they conjecture a causality property where singularities play a role sim
ilar to zerocrossings in Gaussian scalespace.
Jackway and Deriche [206, 207] prove a causality theorem for the dilation–
erosion scalespace, which is also based on local extrema instead of zerocrossings.
They establish that under erosion the number of local minima is decreasing, while
dilation reduces the number of local maxima. The location of these extrema is
preserved during their whole lifetime.
A complete scalespace interpretation is due to Alvarez, Guichard, Lions and
Morel [12]: They prove that under three architectural assumptions (semigroup
property, locality and regularity), one smoothing axiom (comparison principle) and
additional invariance requirements (greylevel shift invariance, invariance under ro
tations and translations, morphological invariance), a twodimensional scalespace
1.5 CLASSIC MORPHOLOGICAL PROCESSES 35
equation has the following form:
∂
t
u = [∇u[ F(t, curv(u)) (1.82)
Clearly, dilation and erosion belong to the class (1.82), thus being good candi
dates for morphological scalespaces. Indeed, in [12] it is shown that the converse
is true as well: all axioms that lead to (1.82) are fulﬁlled.
14
1.5.6 Generalizations
It is possible to extend morphology with a structuring element to morphology with
nonﬂat structuring functions. In this case we have to renounce invariance under
monotone grey level transformations, but we gain an interesting insight into a
process which has very much in common with Gaussian scalespace.
A dilation of an image f with a structuring function b : IR
2
→IR is deﬁned as
(f ⊕b) (x) := sup
y∈IR
2
¦f(x−y) +b(y)¦. (1.83)
This is a generalization of deﬁnition (1.65), since one can recover dilation with a
structuring element B by considering the ﬂat structuring function
b(x) :=
0 (x∈B),
−∞ (x∈B).
(1.84)
Van den Boomgaard [50, 51] and Jackway [206] proposed to dilate an image f(x)
with quadratic structuring functions of type
b(x, t) = −
[x[
2
4t
(t > 0). (1.85)
It can be shown [50, 53] that the result u(x, t) is a weak solution of
∂
t
u = [∇u[
2
, (1.86)
u(x, 0) = f(x). (1.87)
The temporal evolution of a test image under this process is illustrated in Figure
5.5 (b).
In analogy to the fact that Gaussiantype functions k(x, t) = a exp(
x
2
4t
) are the
only rotationally symmetric kernels which are separable with respect to convolu
tion, van den Boomgaard proves that the quadratic structuring functions b(x, t)
are the only rotationally invariant structuring functions which are separable with
respect to dilation [50, 51].
14
Invariance under rotations is only satisﬁed for a disc centered in 0 as structuring element.
36 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
A useful tool for understanding this similarity and many other analogies be
tween morphology and linear systems theory is the slope transform. This general
ization of the Legendre transform is the morphological equivalent of the Fourier
transform. It has been discovered simultaneously by Dorst and van den Boomgaard
[119] and by Maragos [275] in slightly diﬀering formulations.
The close relation between (1.86) and Gaussian scalespace has also triggered
Florack and Maas [142] to study a oneparameter family of isomorphisms of linear
diﬀusion which reveals (1.86) as limiting case.
1.5.7 Numerical aspects
Dilations or erosions with quadratic structuring functions are separable and, thus,
they can be implemented very eﬃciently by applying onedimensional operations.
A fast algorithm is described by van den Boomgaard [51].
For morphological operations with ﬂat structuring elements, the situation is
more complicated. Schemes for dilation or erosion which are based on curve evo
lution turn out be be diﬃcult to handle: they require prohibitive small time steps,
and suﬀer from the problem of coping with singularities and topological changes
[319, 25, 360].
For this reason it is useful to discretize the corresponding image evolution equa
tions. The widelyused Osher–Sethian schemes [319] are based on the idea to derive
numerical methods for such equations from techniques for hyperbolic conservation
laws. Overviews of these level set approaches and their various applications can be
found in [372, 374].
To illustrate the basic idea with a simple example, let us restrict ourselves to the
onedimensional dilation equation ∂
t
u=[∂
x
u[. A ﬁrstorder upwind Osher–Sethian
scheme for this process is given by
u
n+1
i
−u
n
i
τ
=
min
u
n
i
−u
n
i−1
h
, 0
2
+
max
u
n
i+1
−u
n
i
h
, 0
2
, (1.88)
where h is the pixel size, τ is the time step size, and u
n
i
denotes a discrete
approximation of u(ih, nτ).
Level set methods possess two advantages over classical settheoretic schemes
for dilation/erosion [25, 360, 218, 66]:
(a) They give excellent results for nondigitally scalable structuring elements
whose shapes cannot be represented correctly on a discrete grid, for instance
discs or ellipses.
(b) The time t plays the role of a continuous scale parameter. Therefore, the
size of a structuring element need not be multiples of the pixel size, and it
is possible to get results with subpixel accuracy.
1.6 CURVATUREBASED MORPHOLOGICAL PROCESSES 37
However, they reveal also two disadvantages:
(a) They are slower than settheoretic morphological schemes.
(b) Dissipative eﬀects such as blurring of discontinuities occur.
To address the ﬁrst problem, speedup techniques for shape evolution have been
proposed which use only points close to the curve at every time step [8, 435,
373]. Blurring of discontinuities can be minimized by applying shockcapturing
techniques such as highorder ENO
15
schemes [395, 385].
1.5.8 Applications
Continuousscale morphology has been applied to shapefromshading problems,
gridless image halftoning, distance transformations, and skeletonization. Applica
tions outside the ﬁeld of image processing and computer vision include for instance
shape oﬀsets in CAD and path planning in robotics. Overviews and suitable ref
erences can be found in [233, 276].
1.6 Curvaturebased morphological processes
Besides providing a useful reinterpretation of classic continuousscale morphology,
the PDE approach has led to the discovery of new morphological operators. These
processes are curvaturebased, and – although they cannot be written in conserva
tion form – they reveal interesting relations to diﬀusion processes. Two important
representatives of this class are mean curvature motion and its aﬃne invariant
counterpart. In this subsection we shall discuss these PDEs, possible generaliza
tions, numerical aspects, and applications.
1.6.1 Meancurvature ﬁltering
In order to motivate our ﬁrst curvaturebased morphological PDE, let us recall
that the linear diﬀusion equation (1.9) can be rewritten as
∂
t
u = ∂
ηη
u +∂
ξξ
u, (1.89)
where the unit vectors η and ξ are parallel and perpendicular to ∇u, respectively.
The ﬁrst term on the righthand side of (1.89) describes smoothing along the
ﬂowlines, while the second one smoothes along isophotes. When we want to smooth
15
ENO means essentially nonoscillatory. By adapting the stencil for derivative approximations
to the local smoothness of the solution, ENO schemes obtain both highorder accuracy in smooth
regions and sharp shock transitions [183].
38 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
the image anisotropically along its isophotes, we can neglect the ﬁrst term and end
up with the problem
∂
t
u = ∂
ξξ
u, (1.90)
u(x, 0) = f(x). (1.91)
By straightforward calculations one veriﬁes that (1.90) can also be written as
∂
t
u =
u
2
x
2
u
x
1
x
1
−2u
x
1
u
x
2
u
x
1
x
2
+u
2
x
1
u
x
2
x
2
u
2
x
1
+u
2
x
2
(1.92)
= ∆u −
1
[∇u[
2
'∇u, Hess(u)∇u` (1.93)
= [∇u[ curv(u). (1.94)
Since curv(u)=div
∇u
∇u
is the curvature of u (mean curvature for dimensions ≥
3), equation (1.94) is named (mean) curvature motion (MCM). The corresponding
curve evolution
∂
t
C(p, t) = κ(p, t) n(p, t) (1.95)
shows that (1.90) propagates isophotes in inner normal direction with a velocity
that is given by their curvature κ=
det(Cp,Cpp)
Cp
3
.
Processes of this type have ﬁrst been studied by Brakke in 1978 [54]. They
arise in ﬂame propagation, crystal growth, the derivation of minimal surfaces, grid
generation, and many other applications; see [372, 374] and the references therein
for an overview. The importance of MCM in image processing became only recently
clear: As nicely explained in a paper by Guichard and Morel [173], mean curvature
motion can be regarded as the limit process when classic morphological operators
such as median ﬁltering are iteratively applied.
Figure 5.5 (c) and 5.11 (a) present examples for mean curvature ﬁltering. Equa
tion (1.95) is also called geometric heat equation or Euclidean shortening ﬂow. The
subsequent discussions shall clarify these names.
Intrinsic heat ﬂow
Interestingly, there exists a further connection between linear diﬀusion and motion
by curvature. Let v(p, t) denote the Euclidean arclength of C(p, t), i.e.
v(p, t) :=
p
0
[C
ρ
(ρ, t)[ dρ, (1.96)
where C
ρ
:= ∂
ρ
C. The Euclidean arclength is characterized by [C
v
[ = 1. It is
invariant under Euclidean transformations, i.e. mappings
x → Rx +b (1.97)
1.6 CURVATUREBASED MORPHOLOGICAL PROCESSES 39
where R∈IR
2×2
denotes a rotation matrix and b ∈IR
2
is a translation vector.
Since it is wellknown from diﬀerential geometry (see e.g. [71], p. 14) that
κ(p, t) n(p, t) = ∂
vv
C(p, t), (1.98)
we recognize that curvature motion can be regarded as Euclidean invariant diﬀu
sion of isophotes:
∂
t
C(p, t) = ∂
vv
C(p, t). (1.99)
This geometric heat equation is intrinsic, since it is independent of the curve para
metrization. However, the reader should be aware of the fact that – although this
equation looks like a linear onedimensional heat equation – it is in fact nonlinear,
since the arclength v is again a function of the curve.
Theoretical results
For the evolution of a smooth curve under its curvature, it has been shown in
[188, 151, 169] that a smooth solution exists for some ﬁnite time interval [0, T).
A convex curve remains convex, a nonconvex one becomes convex and, for t →T,
the curve shrinks to a circular point, i.e. a point with a circle as limiting shape.
Moreover, since under all ﬂows C
t
=C
qq
the Euclidean arclength parametrization
q(p):=v(p) is the fastest way to shrink the Euclidean perimeter
[C
p
[ dp, equation
(1.99) is called Euclidean shortening ﬂow [150]. The time for shrinking a circle of
radius σ to a point is given by
T =
1
2
σ
2
. (1.100)
In analogy to the dilation/erosion case it can be shown that, for an initial image
f ∈BUC(IR
2
), equation (1.90) has a unique viscosity solution which is L
∞
stable
and satisﬁes a maximum–minimum principle [12].
Scalespace interpretation
A shape scalespace interpretation for curve evolution under Euclidean heat ﬂow is
studied by Kimia and Siddiqi [227]. It is based on results of Evans and Spruck [129].
They establish the semigroup property as architectural quality, and smoothing
properties follow from the fact that the total curvature decreases. Moreover, the
number of extrema and inﬂection points of the curvature is nonincreasing.
As an image evolution, MCM belongs to the class of morphological scalespaces
which satisfy the general axioms of Alvarez, Guichard, Lions and Morel [12], that
have been mentioned in 1.5.5.
When studying the evolution of isophotes under MCM, it can be shown that,
if one isophote is enclosed by another, this ordering is preserved [129, 227]. Such a
shape inclusion principle implies in connection with (1.100) that it takes the time
T =
1
2
σ
2
to remove all isophotes within a circle of radius σ. This shows that the
40 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
relation between temporal and spatial scale for MCM is the same as for linear
diﬀusion ﬁltering (cf. (1.14)).
Moreover, two level sets cannot move closer to one another than they were
initially [129, 227]. Hence, contrast cannot be enhanced. This property is charac
teristic for all scalespaces of the Alvarez–Guichard–Lions–Morel axiomatic and
distinguishes them from nonlinear diﬀusion ﬁlters.
1.6.2 Aﬃne invariant ﬁltering
Motivation
Although Euclidean invariant smoothing methods are suﬃcient in many applica
tions, there exist certain problems which also require invariance with respect to
aﬃne transformations. A (full) aﬃne transformation is a mapping
x → Ax +b (1.101)
where b ∈ IR
2
denotes a translation vector and the matrix A ∈ IR
2×2
is invertible.
Aﬃne transformations arise as shape distortions of planar objects when being
observed from a large distance under diﬀerent angles.
Aﬃne invariant intrinsic diﬀusion
In analogy to the Euclidean invariant heat ﬂow, Sapiro and Tannenbaum [362,
363] constructed an aﬃne invariant ﬂow by replacing the Euclidean arclength
v(p, t) in (1.99) by an “arclength” s(p, t) that is invariant with respect to aﬃne
transformations with det(A) = 1.
Such an aﬃne arclength was proposed by Blaschke [44, pp. 12–15] in 1923. It
is characterized by det(C
s
, C
ss
)=1, and it can be calculated as
s(p, t) :=
p
0
det
C
ρ
(ρ, t), C
ρρ
(ρ, t)
1
3
dρ. (1.102)
By virtue of
∂
ss
C(p, t) =
κ(p, t)
1
3
n(p, t) (1.103)
we obtain the aﬃne invariant heat ﬂow
∂
t
C(p, t) =
κ(p, t)
1
3
n(p, t), (1.104)
C(p, 0) = C
0
(p). (1.105)
1.6 CURVATUREBASED MORPHOLOGICAL PROCESSES 41
Aﬃne invariant image evolution
When regarding the curve C(p, t) as a levelline of an image u(x, t), we end up
with the evolution equation
∂
t
u = [∇u[
curv(u)
1
3
(1.106)
=
u
2
x
2
u
x
1
x
1
−2u
x
1
u
x
2
u
x
1
x
2
+u
2
x
1
u
x
2
x
2
1
3
(1.107)
= [∇u[
2
3
u
1
3
ξξ
, (1.108)
where ξ is the direction perpendicular to ∇u. The temporal evolution of an image
under such an evolution resembles mean curvature motion; see Fig. 5.6(a).
Besides the name aﬃne invariant heat ﬂow, this equation is also called aﬃne
shortening ﬂow, aﬃne morphological scalespace (AMSS), and fundamental equa
tion in image processing.
This image evolution equation has been discovered independently of and si
multaneously with the curve evolution approach of Sapiro and Tannenbaum by
Alvarez, Guichard, Lions and Morel [12] via an axiomatic scalespace approach.
After having mentioned some theoretical results, we shall brieﬂy sketch this rea
soning below.
Theoretical results
The curve evolution properties of aﬃne invariant heat ﬂow can be shown to be the
same as in the Euclidean invariant case, with three exceptions [363]:
(a) Closed curves shrink to points with an ellipse as limiting shape (elliptical
points).
(b) The name aﬃne shortening ﬂow reﬂects the fact that, under all ﬂows C
t
=
C
qq
, the aﬃne arclength parametrization q(p):=s(p) is the fastest way to
shrink the aﬃne perimeter
L(t) :=
det
C
p
(p, t), C
pp
(p, t)
1
3
dp. (1.109)
(c) The time for shrinking a circle of radius σ to a point is
T =
3
4
σ
4
3
. (1.110)
For the image evolution equation (1.106) we have the same results as for MCM
and dilation/erosion concerning wellposedness of a viscosity solution which satis
ﬁes a maximum–minimum principle [12].
42 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
Scalespace properties
Alvarez, Guichard, Lions and Morel [12] proved that (1.106) is unique (up to tem
poral rescalings) when imposing on the scalespace axioms for (1.82) an additional
aﬃne invariance axiom:
For every invertible A∈IR
2×2
and for all t ≥0, there exists a rescaled
time t
′
(t, A)≥0, such that
T
t
(Af) = A(T
t
′ f) ∀ f ∈ BUC(IR
2
). (1.111)
For this reason they call the AMSS equation also fundamental equation in image
analysis. Simpliﬁcations of this axiomatic and related axioms for shape scalespaces
can be found in [16].
The scalespace reasoning of Sapiro and Tannenbaum investigates properties
of the curve evolution, see [362] and the references therein. Based on results of
[229, 24] they point out that the Euclidean absolute curvature decreases as well
as the number of extrema and inﬂection points of curvature. Moreover, a shape
inclusion principle holds.
1.6.3 Generalizations
In order to analyse planar shapes in a way that does not depend on their location in
IR
3
, one requires a multiscale analysis which is invariant under a general projective
mapping
(x
1
, x
2
)
⊤
→
a
11
x
1
+a
21
x
2
+a
31
a
13
x
1
+a
23
x
2
+a
33
,
a
12
x
1
+a
22
x
2
+a
32
a
13
x
1
+a
23
x
2
+a
33
⊤
(1.112)
with A=(a
ij
) ∈ IR
3×3
and det A=1. Research in this direction has been carried
out by Faugeras and Keriven [130, 131, 133], Bruckstein and Shaked [62], Olver
et al. [314], and Dibos [112, 113]. It turns out that intrinsic heatequationlike
formulations for the projective group are more complicated than the Euclidean
and aﬃne invariant ones, and that there is some evidence that they do not reveal
the same smoothing scalespace properties [314]. A study of heat ﬂows which are
invariant under subgroups of the projective group can be found in [314, 113].
An intrinsic heat ﬂow for images painted on surfaces has been investigated by
Kimmel [232]. It is invariant to the bending (isometric mapping) of the surface.
This geodesic curvature ﬂow and other evolutions, both for scalar and vectorvalued
images, can be regarded as steepest descent methods of energy functionals which
have been proposed by Polyakov in the context of string theory [235].
Euclidean and aﬃne invariant curve evolutions can also be modiﬁed in order
to obtain area or lengthpreserving equations [188, 150, 352, 366].
1.6 CURVATUREBASED MORPHOLOGICAL PROCESSES 43
Recently it has been suggested to modify AMSS such that any evolution at T
or Xjunctions of isophotes is inhibited [75]. Such a ﬁltering intends to simplify
the image while preserving its “semantical atoms” in the sense of Kanizsa [219].
Generalizations of MCM or AMSS to 3D images are investigated in [91, 78, 298,
316]. In this case two principal curvatures occur. Since they can have diﬀerent sign,
the question arises of how to combine them to a process which simpliﬁes arbitrary
surfaces without creating singularities. In contrast to the 2D situation, topological
changes such as the splitting of conconvex structures may occur. This problem is
reminescent of the creation of new extrema in diﬀusion scalespaces when going
from 1D to 2D.
Aﬃne scalespace axiomatics for image sequences (movies) have been estab
lished in [12, 287, 288], and possibilities to generalize the axiomatic of Alvarez,
Guichard, Lions and Morel to colour images are discussed in [82].
1.6.4 Numerical aspects
Due to the equivalence between curve evolution and morphological image pro
cessing PDEs, we have three main classes of numerical methods: curve evolution
schemes, settheoretic morphological schemes and approximation schemes for the
Eulerian formulation. A comparison of diﬀerent methods of these classes can be
found in [95].
Curve evolution schemes are investigated by Mokhtarian and Mackworth [291],
Bruckstein et al. [61], Cohignac et al. [95], Merriman et al. [285], Ruuth [347], and
Moisan [289]. In [61] discrete analogues of MCM and AMSS for the evolution of
planar polygons are introduced. In complete analogy to the behaviour of the con
tinuous equations, convergence to polygones, whose corners belong to circles and
ellipses, respectively, is established. Related discrete curve evolutions are analysed
in [135, 359]. Curve evolution schemes can reveal perfect aﬃne invariance.
Convergent settheoretic morphological schemes for MCM and AMSS have been
proposed by Catt´e et al. [80, 79]. On the one hand, they are very fast and they
are entirely invariant under monotone greyscale transformations, on the other
hand, it is diﬃcult to ﬁnd consistent approximations on a pixel grid which have
good rotational invariance. This is essentially the same tradeoﬀ as for circular
structuring elements in classical settheoretic morphology, cf. 1.5.7.
Most direct approximations of morphological image evolution equations are
based on the Osher–Sethian schemes [319, 372, 374]. In the case of MCM or AMSS,
this leads to an explicit ﬁnite diﬀerence method which approximates the spatial
derivatives by central diﬀerences. Diﬀerent variants of these schemes have been
proposed in order to get better rotational invariance, higher stability or less dissi
pative eﬀects [10, 15, 95, 267, 362]. A comparative evaluation of these approaches
has been carried out by Lucido et al. [267]. Niessen et al. [305, 304] approximate
44 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
the spatial derivatives by Gaussian derivatives which are calculated in the Fourier
domain.
Concerning stability one observes that all these explicit schemes can violate a
discrete maximum–minimum principle and require small time steps to be experi
mentally stable. For AMSS an additional constraint appears: the behaviour of this
equation is highly nonlocal, since aﬃne invariance implies that circles are equiva
lent to ellipses of any arbitrary eccentricity. If one wants to have a good numerical
approximation of aﬃne invariance, one has to decrease the temporal step size sig
niﬁcantly below the step size of experimental stability [16, 218]. Using Gaussian
derivatives for MCM or AMSS permits larger time steps for large kernels [304], but
their calculation in the Fourier domain is computationally expensive and aliasing
may lead to oscillatory solutions, cf. 1.2.3.
One way to achieve unconditional L
∞
stability for MCM is to approximate
u
ξξ
by suitable linear combinations of onedimensional secondorder derivatives
along grid directions and to apply an implicit ﬁnite diﬀerence scheme [95, 13, 146].
Schemes of this type, however, renounce consistency with the original equation as
well as rotational invariance: round shapes evolve into polygonal structures.
A consistent semiimplicit approximation of MCM which discretizes the ﬁrst
order spatial derivatives explicitly and the secondorder derivatives implicitly has
been proposed by Alvarez [10]. In order to solve the resulting linear system of
equations he applies symmetric Gauß–Seidel iterations.
Nicolet and Sp¨ uhler [301] investigate a consistent fully implicit scheme for MCM
leading to a nonlinear system of equations. It is solved by means of nonlinear Gauß–
Seidel iterations. Comparing it with the explicit scheme they report a tradeoﬀ
between the larger time step size and the higher computational eﬀort per step.
An inherent problem of all ﬁnite diﬀerence schemes for morphological image
evolutions are their dissipative eﬀects which create additional blurring of discon
tinuities. As a remedy, one can decompose the image into binary level sets, map
them into Lipschitzcontinuous images by applying a distance transformation, and
run a ﬁnite diﬀerence method on them. Afterwards one extracts the processed level
sets as the zerolevel sets of the evolved images, and assembles the ﬁnal image by
superimposing all evolved level sets [75]. The natural price one has to pay for the
excellent results is a fairly high computational eﬀort.
A software package which contains implementations of MCM, AMSS and many
other modern techniques such as wavelets, MumfordShah segmentation, and ac
tive contour models (cf. 1.6.6) is available under the name MegaWave2.
16
16
MegaWave2 has been developed by Jacques Froment and Sylvain Parrino, CEREMADE,
University Paris IX, 75775 Paris Cedex 16, France. More information can be found under
http://www.ceremade.dauphine.fr.
1.6 CURVATUREBASED MORPHOLOGICAL PROCESSES 45
1.6.5 Applications
The special invariances of AMSS are useful for shape recognition tasks [12, 96, 97,
132], for corner detection [15], and for texture discrimination [265, 16]. MCM and
AMSS have also been applied to denoising [305, 304] and blob detection [157, 301]
in medical images and to terrain matching [268]. The potential of MCM for shape
segmentation [290] has even been used for classifying chrysanthemum leaves [1].
If one aims to use these equations for image restoration one usually modiﬁes
them by multiplying them with a term that reduces smoothing at highcontrast
edges [13, 364, 365, 304]; see also Fig. 5.6 (b). Adapting them to tasks such as tex
ture enhancement requires more sophisticated and more global feature descriptors
than the gradient, for instance an analysis by means of Gabor functions [72, 234].
Introducing a reaction term as in [102] allows to attract the solution to speciﬁed
grey values which can be used as quantization levels [11]. Another modiﬁcation
results from omitting the factor [∇u[ in the mean curvature motion (1.94), see e.g.
[126, 127]. This corresponds to nonlinear diﬀusion ﬁlters and a restoration method
by total variation minimization [345] which shall be described in 1.7.2.
In order to improve images, MCM or AMSS have also been combined with other
processes such as linear diﬀusion [13], shock ﬁltering ([14], cf. 1.7.1) or global PDEs
for histogramme enhancement [358].
Malladi and Sethian [272] propose to replace MCM by a technique in which
the motion of level curves is based on either min(κ, 0) or max(κ, 0), depending on
the average grey value within a certain neighbourhood. This socalled min–max
ﬂow produces a restored image as steadystate solution and reveals good denoising
properties. Wellposedness results are not known up to now, since the theory of
viscosity solutions is no longer applicable.
All these preceding modiﬁcations are at the expense of renouncing the mor
phological invariance of the genuine operators (and also aﬃne invariance in the
case of [364, 365, 304], unless an “aﬃne invariant gradient” [231, 315] is used). If
one wants to stay within the morphological framework one can combine diﬀerent
morphological processes, for instance MCM and dilation/erosion. This leads to a
process which is useful for analysing components of shape [228, 230, 383, 384, 453],
and which is called entropy scalespace or reaction–diﬀusion scalespace.
Recently Steiner et al. proposed a method for caricaturelike shape exaggera
tion [394]. They evolved the boundary curve by means of a backward Euclidean
shortening ﬂow with a stabilizing bias term as in (1.56).
It is interesting to note that, already in 1965, Gabor – the inventor of optical
holography and the socalled Gabor functions – proposed a deblurring algorithm
based on combining MCM with backward smoothing along ﬂowlines [149, 260].
This longtime forgotten method is similar to the Perona–Malik process (1.36) for
large image gradients.
46 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
1.6.6 Active contour models
One of the main applications of curve evolution ideas appears in the context of
active contour models (deformable models). 2D versions are also called snakes,
while 3D active models are sometimes named active surfaces or active blobs. Ac
tive contour models can be used to search image features in an interactive way.
Especially for assisting the segmentation of medical images they have become very
popular [282]: The expert user gives a good initial guess of an interesting contour
(organ, tumour, ...), which will be carried the rest of the way by some energy
minimization. Apart from these practical merits, snake models incorporate many
ideas ranging from energy minimization over curve evolution to the Perona–Malik
ﬁlter and diﬀusion–reaction models. It is therefore not surprising that they play an
important role in several generalization and uniﬁcation attempts [356, 380, 452].
Explicit snakes
Kass, Witkin and Terzopoulos proposed the ﬁrst active contour model in a journal
paper in 1988 [221]. Their snakes can be thought of as an energyminimizing spline,
which is attracted by image features such as edges. For this reason, the energy
functional consists of two parts: an internal energy fraction which controls the
smoothness of the result, and an external energy term attracting the result to
salient image features.
Such a snake is represented by a curve C(s) = (x
1
(s), x
2
(s))
⊤
which minimizes
the functional
E(C(s)) =
C(s)
α
2
[C
s
(s)[
2
+
β
2
[C
ss
(s)[
2
−γ [∇f(C(s))[
2
ds. (1.113)
The ﬁrst summand is a membrane term causing the curve to shrink, the second
one is a rigidity term which encourages a straight contour, and the third term
pushes the contour to high gradients of the image f. We observe that terms 1
and 2 describe the internal energy, while the third one represents the external
(image) energy. The nonnegative parameters α, β and γ serve as weights of these
expressions. Since this model makes direct use of the snake contour, it is also called
an explicit model.
Minimizing the functional (1.113) by steepest descent gives
∂C
∂t
= αC
ss
−βC
ssss
−γ∇([∇f[
2
), (1.114)
which can be approximated by ﬁnite diﬀerences. A 3D version of such an active
contour model is presented in [401].
Usually, the result will depend on the choice of the initial curve and a good
segmentation requires an initial curve which is close to the ﬁnal segment. The main
1.6 CURVATUREBASED MORPHOLOGICAL PROCESSES 47
disadvantage of the preceding model is its topological rigidity: a classical explicit
snake cannot split in order to segment several objects simultaneously
17
. In practice,
it is also sometimes not easy to ﬁnd a good balance between the parameters α, β
and γ.
Implicit snakes
In 1993 some of the inherent diﬃculties of explicit snakes have been solved by
replacing them by a socalled implicit model. It was discovered by Caselles, Catt´e,
Coll and Dibos [73], and later on by Malladi, Sethian and Vemuri [273].
The idea behind implicit snakes is to embed the initial curve C
0
(s) as a zero
level curve into a function u
0
: IR
2
→ IR, for instance by using the distance
transformation. Then u
0
is evolved under a PDE which includes knowledge about
the original image f:
∂
t
u = g([∇f
σ
[
2
) [∇u[
div
∇u
[∇u[
+ ν
. (1.115)
This evolution is stopped at some time T, when the process does hardly alter
anymore, and the ﬁnal contour C is extracted as the zero level curve of u(x, T).
The terms in (1.115) have the following meaning:
• [∇u[ div (∇u/[∇u[) is the curvature term of MCM which smoothes level sets;
see (1.94).
• ν[∇u[ describes motion in normal direction, i.e. dilation or erosion depending
on the sign of ν. This socalled balloon force [93] is required for pushing a
level set into concave regions, a compensation for the property of MCM to
create convex regions.
• g is a stopping function such as the Perona–Malik diﬀusivity (1.32): it be
comes small for large [∇f
σ
[ = [∇K
σ
∗f[. Hence, it slows down the snake as
soon as it approaches signiﬁcant edges of the original image f.
For this model Caselles et al. could prove wellposedness in the viscosity sense
[73]. Whitaker and Chen developed similar implicit active contour models for 3D
images [436, 435], and Caselles and Coll investigated related approaches for image
sequences [74].
An advantage of implicit snake models is their topological ﬂexibility: The con
tour may split. This allows simultaneous segmentation of multiple objects. More
over, they use essentially only one remaining parameter, the balloon force ν. On
17
Recently McInerley and Terzopoulos have proposed modiﬁed explicit deformable models
which allow topological changes [281, 283].
48 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
the other hand, the process does not really stop at the desired result, it only slows
down, and it is diﬃcult to interpret implicit snakes in terms of energy minimiza
tion. In order to address the initialization problem, Tek and Kimia use implicit
active contours starting from randomly chosen seed points, both in 2D [399] and
in 3D [400]. They name this technique reaction–diﬀusion bubbles.
Geodesic snakes
Geodesic snakes make a synthesis of explicit and implicit snake ideas. They have
been proposed simultaneously by Caselles, Kimmel and Sapiro [76] and by Kichenas
samy, Kumar, Olver, Tannenbaum and Yezzi [226]. These snakes replace the con
tour energy (1.113) by
E(C) =
C(s)
α
2
[C
s
(s)[
2
−γ g([∇f
σ
(C)[
2
)
ds (1.116)
where g denotes again a Perona–Malik diﬀusivity of type (1.32). Under some addi
tional assumptions they derive that minimizing (1.116) is equivalent to searching
min
C
C(s)
g
[∇f
σ
(C(s))[
2
[C
s
(s)[ ds. (1.117)
This is nothing else than ﬁnding a curve of minimal distance (geodesic) with respect
to some imageinduced metric. Embedding the initial curve as a level set of some
image u
0
and applying a descent method to the corresponding EulerLagrange
equation leads to the image evolution PDE
∂
t
u = [∇u[ div
g([∇f
σ
[
2
)
∇u
[∇u[
. (1.118)
This active contour model is parameterfree, but often a speed term νg([∇f
σ
[
2
)[∇u[
is added to achieve faster and more stable attraction to edges. A theoretical analysis
of geodesic snakes concerning existence, uniqueness and stability of a viscosity
solution can be found in [76, 226], and extensions to 3D images are studied in [77,
226, 271]. Recently also an aﬃne invariant analogue to geodesic active contours has
been proposed [315]. Techniques which can be related to geodesic active contours
have also been used for solving 3D vision problems such as stereo [134] and motion
analysis [322].
Selfsnakes
The properties of geodesic snakes induced Sapiro to use a related technique for
image enhancement [357]: he replaced g([∇f
σ
[
2
) in (1.118) by g([∇u
σ
[
2
). Then the
1.7 TOTAL VARIATION METHODS 49
snake becomes a selfsnake no longer underlying external image forces. For σ = 0
this gives
∂
t
u = [∇u[ div
g([∇u[
2
)
∇u
[∇u[
(1.119)
= g([∇u[
2
) u
ξξ
+∇
⊤
g([∇u[
2
)
∇u (1.120)
with ξ ⊥ ∇u. Although this equation cannot be cast in divergence form, we observe
striking similarities with the Perona–Malik process from Section 1.3.1: the latter
can be written as
∂
t
u = g([∇u[
2
) ∆u +∇
⊤
g([∇u[
2
)
∇u. (1.121)
Thus, it cannot be excluded that a selfsnake without spatial regularization reveals
the same illposedness problems as the Perona–Malik ﬁlter [356]. For σ > 0, how
ever, Chen, Vemuri and Wong [89] established existence and stability of a unique
viscosity solution to a modiﬁed selfsnake process. Their model contains a reaction
term which inhibits smoothing at edges and keeps the ﬁltered image u close to the
original image f; cf. [380]. The restored image is given by the steadystate of
∂
t
u = [∇u[ div
g([∇u
σ
[
2
)
∇u
[∇u[
+β [∇u[ (f −u) (β > 0). (1.122)
The temporal evolution of a regularized selfsnake without reaction term is de
picted in Fig. 5.6(c). Generalizations of selfsnakes to vectorvalued images [357,
361] can be obtained using Di Zenzo’s ﬁrst fundamental form for colour images
[114]; see also [414, 422] for related ideas.
1.7 Total variation methods
Inspired by observations from ﬂuid dynamics where the total variation (TV)
TV (u) :=
Ω
[∇u[ dx (1.123)
plays an important role for shock calculations, one may ask if it is possible to apply
related ideas to image processing. This would be useful to restore discontinuities
such as edges.
Below we shall focus on two important TVbased image restoration techniques
which have been pioneered by Osher and Rudin: TVpreserving methods and tech
niques which are TVminimizing subject to certain constraints.
18
18
Another image enhancement method that is close in spirit is due to Eidelman, Grossmann
and Friedman [125]. It maps the image grey values to gas dynamical parameters and solves the
compressible Euler equations using shockcapturing total variation diminishing (TVD) techniques
based on Godunov’s method.
50 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
1.7.1 TVpreserving methods
In 1990 Osher and Rudin have proposed to restore blurred images by shock ﬁltering
[317]. These ﬁlters calculate the restored image as the steadystate solution of the
problem
∂
t
u = −[∇u[ F(L(u)), (1.124)
u(x, 0) = f(x). (1.125)
Here, sgn(F(u)) = sgn(u), and L(u) is a secondorder elliptic operator whose zero
crossings correspond to edges, e.g. the Laplacian L(u) = ∆u or the secondorder
directional derivative L(u)=u
ηη
with η ∇u.
By means of our knowledge from morphological processes, we recognize that
this ﬁlter aims to produce a ﬂow ﬁeld that is directed from the interior of a region
towards its edges where it develops shocks. Thus, the goal is to obtain a piecewise
constant steadystate solution with discontinuities only at the edges of the initial
image.
It has been shown that a onedimensional version of this ﬁlter preserves the
total variation and satisﬁes a maximum–minimum principle, both in the continuous
and discrete case. For the twodimensional case not many theoretical results are
available except for a discrete maximum–minimum principle.
Recently van den Boomgaard [52] pointed out that the 1D version of (1.124)
with F(u) := sgn(u) arises as the PDE formulation of a classical image enhance
ment algorithm by Kramer and Bruckner [243]. Kramer and Bruckner proved in
1975 that their Ndimensional discrete scheme converges after a ﬁnite number of
iterations to a state where each point is a local extremum.
Osher and Rudin have also proposed another TVpreserving deblurring tech
nique [318]. It solves the linear diﬀusion equation backwards in time under the
regularizing constraint that the total variation remains constant. This stabiliza
tion can be realized by keeping local extrema ﬁxed during the whole evolution.
From a practical point of view, TVpreserving methods suﬀer from the problem
that ﬂuctuations due to noise do also create shocks. For this reason, Alvarez and
Mazorra [14] replace the operator L(u) =u
ηη
in (1.124) by a Gaussiansmoothed
version L(K
σ
∗u) = K
σ
∗u
ηη
and supplement the resulting equation with a noise
eliminating mean curvature process. They prove that their semiimplicit ﬁnite
diﬀerence scheme has a unique solution which satisﬁes a maximum–minimum prin
ciple.
1.7.2 TVminimizing methods
Total variation is good for quantifying the simplicity of an image since it mea
sures oscillations without unduly punishing discontinuities. For this reason, blocky
1.7 TOTAL VARIATION METHODS 51
images (consisting only of a few almost piecewise constant segments) reveal very
small total variation.
In order to restore noisy blocky images, Rudin, Osher and Fatemi [345] have
proposed to minimize the total variation under constraints which reﬂect assump
tions about noise.
19
To ﬁx ideas, let us study an example. Given an image f with additive noise of
zero mean and known variance σ
2
, we seek a restoration u satisfying
min
u
Ω
[∇u[ dx (1.126)
subject to
Ω
(u−f)
2
dx = σ
2
, (1.127)
Ω
u dx =
Ω
f dx. (1.128)
In order to solve this constrained variational problem, PDE methods can be
applied: A solution of (1.126)–(1.128) veriﬁes necessarily the Euler equation
div
∇u
[∇u[
−µ −λ(u−f) = 0 (1.129)
with homogeneous Neumann boundary conditions. The (unknown) Lagrange mul
tipliers µ and λ have to be determined in such a way that the constraints are
fulﬁlled. Interestingly, (1.129) looks similar to the steadystate equation of the
diﬀusion–reaction equation (1.56), but – in contrast to TV approaches – equa
tion (1.56) is not intended to satisfy the noise constraint exactly [346]. Moreover,
the divergence term in (1.129) is identical with the curvature, which relates TV
minimizing techniques to MCM.
In [345] a gradient descent method is proposed to solve (1.129). It uses an
explicit ﬁnite diﬀerence scheme with central and onesided spatial diﬀerences and
adapts the Lagrange multiplier by means of the gradient projection method of
Rosen.
One may also reformulate the constrained TV minimization as an uncon
strained problem [83]: The penalized least square problem
min
u
1
2
u−f
2
L
2
(Ω)
+α
Ω
[∇u[ dx
(1.130)
is equivalent to the constrained TV minimization, if α is related to the Lagrange
multiplier λ via α =
1
λ
.
19
Related ideas have also been developed by Geman and Reynolds [153].
52 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
In recent years, problems of type (1.130) have attracted much interest from
mathematicians working on inverse problems, optimization, or numerical analysis
[2, 84, 85, 87, 115, 117, 118, 204, 205, 253, 409]. To overcome the problem that
the total variation integral contains the nondiﬀerentiable argument [∇u[, one ap
plies regularization strategies or techniques from nonsmooth optimization. Much
research is done in order to ﬁnd eﬃcient numerical methods for which convergence
can be established.
TVminimizing methods have been generalized in diﬀerent ways:
• The constrained TVminimization idea is frequently adapted to other con
straints such as blur, noise with blur, or other types of noise [262, 344, 346,
116, 253, 410, 83]. Lions et al. [262] and Dobson and Santosa [115] have
shown the existence of BV(Ω)solutions for problems of this type. Recently,
Chambolle and Lions [83] have extended the existence proof to noncompact
operators (which comprises also the situation without blur), and they have
established uniqueness.
• The tendency of TVminimizing to create piecewise constant structures can
cause undesired eﬀects such as the creation of staircases at sigmoidlike edges
[116, 83]. As a remedy, it has been proposed to minimize the L
1
norm of
expressions containing also higherorder derivatives [344, 83]. Another pos
sibility is to consider the constrained minimization of
B(u) :=
Ω
[∇u[
p(∇u)
dx, (1.131)
where p([∇u[) decreases from 2 to 1, as [∇u[ ranges from 0 to ∞; see [46].
• TVminimizing methods have also been used for estimating discontinuous
blurring kernels such as motion or outoffocus blur from a degraded image.
This leads to TVbased blind deconvolution algorithms [86].
• They have been applied to colour images [45], where the generalized TV
norm is chosen as the l
2
norm of the TV norms of the separate channels.
• Strong and Chan have identiﬁed the parameter α in (1.130) as a scale para
meter [396]. By adapting α to the local image structure, they establish re
lations between TVminimizing methods and nonlinear diﬀusion techniques
[397].
Total variation methods have been applied to restoring images of military rele
vance [345, 346, 262, 253], to improving material from criminal and civil investiga
tions as court evidence [344], and to enhancing pictures from confocal microscopy
[409] and tomography [115, 396]. They are useful for enhancing reconstruction al
gorithms for inverse scattering problems [37], and the idea of L
1
norm minimization
has also led to improved optic ﬂow algorithms [245].
1.7 TOTAL VARIATION METHODS 53
1.8 Conclusions and further scope of the book
Although we have seen that there exists a large variety of PDEbased scalespace
and image restoration methods which oﬀer many advantages, we have also become
aware of some limitations. They shall serve as a motivation for the theory which
will be explored in the subsequent chapters.
Linear diﬀusion and morphological scalespaces are wellposed and have a solid
axiomatic foundation. On the other hand, for some applications, they possess the
undesirable property that they do not permit contrast enhancement and that they
may blur and delocalize structures.
Pure restoration methods such as diﬀusion–reaction equations or TVbased
techniques do allow contrast enhancement and lead to stable structures but can
suﬀer from theoretical or practical problems, for instance unsolved wellposedness
questions or the search for eﬃcient minimizers of nonconvex or nondiﬀerentiable
functionals. Moreover, most imageenhancing PDE methods focus on edge detec
tion and segmentation problems. Other interesting image restoration topics have
found less attention.
For both scalespace and restoration methods many questions concerning their
discrete realizations are still open: discrete scalespace results are frequently miss
ing, minimization algorithms can get trapped in a poor local minimum, or the use
of explicit schemes causes restrictive step size limitations.
The goal of the subsequent chapters is to develop a theory for nonlinear aniso
tropic diﬀusion ﬁlters which addresses some of the abovementioned shortcomings.
In particular, we shall see that anisotropic nonlinear diﬀusion processes can share
many advantages of the scalespace and the image enhancement world. A scale
space interpretation is presented which does not exclude contrast enhancement,
and wellposedness results are established. Both scalespace and wellposedness
properties carry over from the continuous to the semidiscrete and discrete setting.
The latter comprises for instance semiimplicit techniques for which unconditional
stability in the L
∞
norm is proved. The general framework, for which the results
hold, includes also linear and isotropic nonlinear diﬀusion ﬁlters. Finally, speciﬁc
anisotropic models are presented which permit applications beyond segmentation
and edge enhancement tasks, for instance enhancement of coherent ﬂowlike struc
tures in textures.
54 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
Chapter 2
Continuous diﬀusion ﬁltering
This chapter presents a general continuous model for anisotropic diﬀusion ﬁlters,
analyses its theoretical properties and gives a scalespace interpretation. To this
end, we adapt the diﬀusion process to the structure tensor, a wellknown tool
for analysing local orientation. Under fairly weak assumptions on the class of ﬁl
ters, it is possible to establish wellposedness and regularity results and to prove a
maximum–minimum principle. Since the proof does not require any monotony as
sumption it applies also to contrastenhancing diﬀusion processes. After sketching
invariances of the resulting scalespace, we focus on analysing its smoothing prop
erties. We shall see that, besides the extremum principle, a large class of associated
Lyapunov functionals plays an important role in this context [414, 415].
2.1 Basic ﬁlter structure
Let us consider a rectangular image domain Ω := (0, a
1
) (0, a
2
) with boundary
Γ := ∂Ω and let an image be represented by a mapping f ∈ L
∞
(Ω). The class
of anisotropic diﬀusion ﬁlters we are concerned with is represented by the initial
boundary value problem
∂
t
u = div (D ∇u) on Ω (0, ∞), (2.1)
u(x, 0) = f(x) on Ω, (2.2)
'D∇u, n` = 0 on Γ (0, ∞). (2.3)
Hereby, n denotes the outer normal and '., .` the Euclidean scalar product on IR
2
.
In order to adapt the diﬀusion tensor D ∈ IR
2×2
to the local image structure, one
would usually let it depend on the edge estimator ∇u
σ
(cf. 1.3.3), where
u
σ
(x, t) := (K
σ
∗ ˜ u(., t)) (x) (σ > 0) (2.4)
and ˜ u denotes an extension of u from Ω to IR
2
, which may be obtained by mirroring
at Γ (cf. [81]).
55
56 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
However, we shall choose a more general structure descriptor which comprises
the edge detector ∇u
σ
, but also allows to extract more information. This will be
presented next.
2.2 The structure tensor
In order to identify features such as corners or to measure the local coherence of
structures, we need methods which take into account how the orientation of the
(smoothed) gradient changes within the vicinity of any investigated point.
The structure tensor – also called interest operator, scatter matrix or (windowed
second) moment tensor – is an important representative of this class. Matrices of
this type are useful for many diﬀerent tasks, for instance for analysing ﬂowlike
textures [340, 31], corners and Tjunctions [145, 182, 310, 309], shape cues [256,
pp. 349–382] and spatio–temporal image sequences [209, pp. 147–153], [168, pp.
219–258]. Related approaches can also be found in [39, 40, 220]. Let us focus on
some aspects which are of importance in our case.
In order to study orientations instead of directions, we have to identify gra
dients which diﬀer only by their sign: they share the same orientation, but point
in opposite directions. To this end, we reconsider the vectorvalued structure de
scriptor ∇u
σ
within a matrix framework. The matrix J
0
resulting from the tensor
product
J
0
(∇u
σ
) := ∇u
σ
⊗∇u
σ
:= ∇u
σ
∇u
T
σ
(2.5)
has an orthonormal basis of eigenvectors v
1
, v
2
with v
1
 ∇u
σ
and v
2
⊥ ∇u
σ
.
The corresponding eigenvalues [∇u
σ
[
2
and 0 give just the contrast (the squared
gradient) in the eigendirections.
Averaging this orientation information can be accomplished by convolving
J
0
(∇u
σ
) componentwise with a Gaussian K
ρ
. This gives the structure tensor
J
ρ
(∇u
σ
) := K
ρ
∗ (∇u
σ
⊗∇u
σ
) (ρ ≥ 0). (2.6)
It is not hard to verify that the symmetric matrix J
ρ
=
j
11
j
12
j
12
j
22
is positive semidef
inite and possesses orthonormal eigenvectors v
1
, v
2
with
v
1

¸
2j
12
j
22
−j
11
+
(j
11
−j
22
)
2
+ 4j
2
12
¸
. (2.7)
The corresponding eigenvalues µ
1
and µ
2
are given by
µ
1,2
=
1
2
j
11
+j
22
±
(j
11
−j
22
)
2
+ 4j
2
12
, (2.8)
where the + sign belongs to µ
1
. As they integrate the variation of the grey values
within a neighbourhood of size O(ρ), they describe the average contrast in the
2.3. THEORETICAL RESULTS 57
eigendirections. Thus, the integration scale ρ should reﬂect the characteristic win
dow size over which the orientation is to be analysed. Presmoothing in order to
obtain ∇u
σ
makes the structure tensor insensitive to noise and irrelevant details
of scales smaller than O(σ). The parameter σ is called local scale or noise scale.
By virtue of µ
1
≥µ
2
≥0, we observe that v
1
is the orientation with the highest
grey value ﬂuctuations, and v
2
gives the preferred local orientation, the coherence
direction. Furthermore, µ
1
and µ
2
can be used as descriptors of local structure:
Constant areas are characterized by µ
1
= µ
2
= 0, straight edges give µ
1
≫µ
2
= 0,
corners can be identiﬁed by µ
1
≥ µ
2
≫0, and the expression
(µ
1
−µ
2
)
2
= (j
11
−j
22
)
2
+ 4j
2
12
(2.9)
becomes large for anisotropic structures. It is a measure of the local coherence.
An example which illustrates the advantages of the structure tensor for analysing
coherent patterns can be found in Figure 5.10 (d); see Section 5.2.
2.3 Theoretical results
In order to discuss wellposedness results, let us ﬁrst recall some useful notations.
Let H
1
(Ω) be the Sobolev space of functions u(x) ∈ L
2
(Ω) with all distributional
derivatives of ﬁrst order being in L
2
(Ω). We equip H
1
(Ω) with the norm
u
H
1
(Ω)
:=
u
2
L
2
(Ω)
+
2
¸
i=1
∂
x
i
u
2
L
2
(Ω)
1/2
(2.10)
and identify it with its dual space. Let L
2
(0, T; H
1
(Ω)) be the space of functions
u, strongly measurable on [0, T] with range in H
1
(Ω) (for the Lebesgue measure
dt on [0, T]) such that
u
L
2
(0,T;H
1
(Ω))
:=
¸
T
0
u(t)
2
H
1
(Ω)
dt
¸
1/2
< ∞. (2.11)
In a similar way, C([0, T]; L
2
(Ω)) is deﬁned as the space of continuous functions
u : [0, T] →L
2
(Ω) supplemented with the norm
u
C([0,T];L
2
(Ω))
:= max
[0,T]
u(t)
L
2
(Ω)
. (2.12)
As usual, we denote by C
p
(X, Y ) the set of C
p
mappings from X to Y .
Now we can give a precise formulation of the problem we are concerned with.
We need the following prerequisites:
58 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
Assume that f ∈L
∞
(Ω), ρ≥0, and σ, T >0.
Let a:=ess inf
Ω
f, b := ess sup
Ω
f, and consider the problem
∂
t
u = div (D(J
ρ
(∇u
σ
)) ∇u) on Ω (0, T],
u(x, 0) = f(x) on Ω,
'D(J
ρ
(∇u
σ
))∇u, n` = 0 on Γ (0, T],
where the diﬀusion tensor D = (d
ij
) satisﬁes the following
properties:
(C1) Smoothness:
D ∈ C
∞
(IR
2×2
, IR
2×2
).
(C2) Symmetry:
d
12
(J)=d
21
(J) for all symmetric matrices J ∈IR
2×2
.
(C3) Uniform positive deﬁniteness:
For all w∈L
∞
(Ω, IR
2
) with [w(x)[ ≤ K on
¯
Ω, there
exists a positive lower bound ν(K) for the eigenvalues
of D(J
ρ
(w)).
(P
c
)
Under these assumptions the following theorem, which generalizes and extends
results from [81, 414], can be proved.
Theorem 1 (Wellposedness,
1
regularity, extremum principle)
The problem (P
c
) has a unique solution u(x, t) in the distributional sense which
satisﬁes
u ∈ C([0, T]; L
2
(Ω)) ∩ L
2
(0, T; H
1
(Ω)), (2.13)
∂
t
u ∈ L
2
(0, T; H
1
(Ω)). (2.14)
Moreover, u∈C
∞
(
¯
Ω(0, T]). This solution depends continuously on f with respect
to  . 
L
2
(Ω)
, and it fulﬁls the extremum principle
a ≤ u(x, t) ≤ b on Ω (0, T]. (2.15)
1
For a complete wellposedness proof one also has to establish stability with respect to per
turbations of the diﬀusion equation. This topic will not be addressed here.
2.3. THEORETICAL RESULTS 59
Proof:
(a) Existence, uniqueness and regularity
Existence, uniqueness and regularity are straightforward anisotropic exten
sions of the proof for the isotropic case studied by Catt´e, Lions, Morel and
Coll [81]. Therefore, we just sketch the basic ideas of this proof.
Existence can be proved using Schauder’s ﬁxed point theorem. One considers
the solution U(w) of a distributional linear version of (P
c
) where D depends
on some function w instead of u. Then one shows that U is a weakly contin
uous mapping from a nonempty, convex and weakly compact subset W
0
of
W(0, T) :=
w ∈ L
2
(0, T; H
1
(Ω)),
dw
dt
∈ L
2
(0, T; H
1
(Ω))
¸
into itself. Since
W(0, T) is contained in L
2
(0, T; L
2
(Ω)), with compact inclusion, U reveals a
ﬁxed point u ∈ W
0
, i.e. u = U(u).
Smoothness follows from classical bootstrap arguments and the general the
ory of parabolic equations [246]. Since u(t) ∈ H
1
(Ω) for all t > 0, one deduces
that u(t) ∈ H
2
(Ω) for all t > 0. By iterating, one can establish that u is a
strong solution of (P
c
) and u ∈ C
∞
((0, T]
¯
Ω).
The basic idea of the uniqueness proof consists of using energy estimates for
the diﬀerence of two solutions, such that the Gronwall–Bellman inequality
can be applied. Then, uniqueness follows from the fact that both solutions
start with the same initial values.
Finally an iterative linear scheme is investigated, whose solution is shown to
converge in C([0, T]; L
2
(Ω)) to the strong solution of (P
c
).
(b) Extremum principle
In order to prove a maximum–minimum principle, we utilize Stampacchia’s
truncation method (cf. [58], p. 211).
We restrict ourselves to proving only the maximum principle. The minimum
principle follows from the maximum principle when being applied to the
initial datum −f.
Let G ∈ C
1
(IR) be a function with G(s) = 0 on (−∞, 0] and 0 < G
′
(s) ≤ C
on (0, ∞) for some constant C. Now, we deﬁne
H(s) :=
s
0
G(σ) dσ, s ∈ IR,
ϕ(t) :=
Ω
H(u(x, t) −b) dx, t ∈ [0, T].
By the Cauchy–Schwarz inequality, we have
Ω
[G(u(x, t)−b) ∂
t
u(x, t)[ dx ≤ C u(t)−b
L
2
(Ω)
∂
t
u(t)
L
2
(Ω)
60 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
and by virtue of (2.13), (2.14) we know that the righthand side of this
estimate exists. Therefore, ϕ is diﬀerentiable for t > 0, and we get
dϕ
dt
=
Ω
G(u−b) ∂
t
u dx
=
Ω
G(u−b) div (D(J
ρ
(∇u
σ
)) ∇u) dx
=
Γ
G(u−b) 'D(J
ρ
(∇u
σ
)) ∇u, n`
. .. .
=0
dS
−
Ω
G
′
(u−b)
. .. .
≥0
'∇u, D(J
ρ
(∇u
σ
)) ∇u`
. .. .
≥0
dx
≤ 0. (2.16)
By means of H(s) ≤
C
2
s
2
, we have
0 ≤ ϕ(t) ≤
Ω
H(u(x, t)−f(x)) dx ≤
C
2
u(t)−f
2
L
2
(Ω)
. (2.17)
Since u ∈ C([0, T]; L
2
(Ω)), the righthand side of (2.17) tends to 0 = ϕ(0)
for t →0
+
which proves the continuity of ϕ(t) in 0. Now from
ϕ ∈ C[0, T], ϕ(0) = 0, ϕ ≥ 0 on [0, T]
and (2.16), it follows that
ϕ ≡ 0 on [0, T].
Hence, for all t ∈ [0, T], we obtain u(x, t)−b ≤ 0 almost everywhere (a.e.) on
Ω. Due to the smoothness of u for t > 0, we ﬁnally end up with the assertion
u(x, t) ≤ b on
¯
Ω (0, T].
(c) Continuous dependence on the initial image
Let f, h ∈ L
∞
(Ω) be two initial values and u, w the corresponding solutions.
In the same way as in the uniqueness proof in [81], one shows that there
exists some constant c > 0 such that
d
dt
u(t)−w(t)
2
L
2
(Ω)
≤ c ∇u(t)
2
L
2
(Ω)
u(t)−w(t)
2
L
2
(Ω)
.
Applying the Gronwall–Bellman lemma [57, pp. 156–137] yields
u(t)−w(t)
2
L
2
(Ω)
≤ f −h
2
L
2
(Ω)
exp
¸
c
t
0
∇u(s)
2
L
2
(Ω)
ds
¸
.
2.3. THEORETICAL RESULTS 61
By means of the extremum principle we know that u is bounded on
¯
Ω[0, T].
Thus, ∇u
σ
is also bounded, and prerequisite (C3) implies that there exists
some constant ν = ν(σ, f
L
∞
(Ω)
) > 0, such that
t
0
∇u(s)
2
L
2
(Ω)
ds
≤
T
0
∇u(s)
2
L
2
(Ω)
ds
≤
1
ν
T
0
Ω
'∇u(x, s), D(J
ρ
(∇u
σ
(x, s)))∇u(x, s)` dx
ds
=
1
ν
T
0
Ω
u(x, s) div
D(J
ρ
(∇u
σ
(x, s)))∇u(x, s)
dx
ds
≤
1
ν
T
0
u(s)
L
2
(Ω)
∂
t
u(s)
L
2
(Ω)
ds
≤
1
ν
u
L
2
(0,T;H
1
(Ω))
∂
t
u
L
2
(0,T;H
1
(Ω))
.
By virtue of (2.13), (2.14), we know that the righthand of this estimate
exists. Now, let ǫ > 0 and choose
δ := ǫ exp
−c
2ν
u
L
2
(0,T;H
1
(Ω))
∂
t
u
L
2
(0,T;H
1
(Ω))
.
Then for f −h
L
2
(Ω)
< δ, the preceding results imply
u(t)−w(t)
L
2
(Ω)
< ǫ ∀ t ∈ [0, T],
which proves the continuous dependence on the initial data. 2
Remarks:
(a) We observe a strong smoothing eﬀect which is characteristic for many dif
fusion processes: under fairly weak assumptions on the initial image (f ∈
L
∞
(Ω)) we obtain an inﬁnitely often diﬀerentiable solution for arbitrary small
positive times. More restrictive requirements – for instance f ∈ BUC(IR
2
)
in order to apply the theory of viscosity solutions – are not necessary in our
case.
(b) Moreover, our proof does not require any monotony assumption. This has the
advantage that contrastenhancing processes are permitted as well. Chapter
5 will illustrate this by presenting examples where contrast is enhanced.
62 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
(c) The continuous dependence of the solution on the initial image has signif
icant practical impact as it ensures stability with respect to perturbations
of the original image. This is of importance when considering stereo image
pairs, spatiotemporal image sequences or slices from medical CT or MRI
sequences, since we know that similar images remain similar after ﬁltering.
2
(d) The extremum principle oﬀers the practical advantage that, if we start for
instance with an image within the range [0, 255], we will never obtain results
with grey value such as 257. It is also closely related to smoothing scalespace
properties, as we shall see in 2.4.2.
(e) The wellposedness results are essentially based on the fact that the regu
larization by convolution with a Gaussian allows to estimate ∇u
σ

L
∞
(Ω)
by
u
L
∞
(Ω)
. This property is responsible for the uniform positive deﬁniteness
of the diﬀusion tensor.
2.4 Scalespace properties
Let us now investigate scalespace properties of the class (P
c
) and juxtapose the re
sults to other scalespaces. To this end, we shall not focus on further investigations
of architectural requirements like recursivity, regularity and locality, as these qual
ities do not distinguish nonlinear diﬀusion scalespaces from other ones. We start
with brieﬂy discussing invariances. Afterwards, we turn to a more crucial task,
namely the question in which sense our evolution equation – which may allow con
trast enhancement – can still be considered as a smoothing, informationreducing
image transformation.
2.4.1 Invariances
Let u(x, t) be the unique solution of (P
c
) and deﬁne the scalespace operator T
t
by
T
t
f := u(t), t ≥ 0, (2.18)
where u(t) := u(., t).
The properties we discuss now illustrate that an invariance of T
t
with respect to
some image transformation P is characterized by the fact that T
t
and P commute.
Much of the terminology used below is borrowed from [12].
2
This does not contradict contrast enhancement: In the case of two similar images, where
one leads to contrast enhancement and the other not, the regularization damps the enhancement
process in such a way that both images do not diﬀer much after ﬁltering.
2.4. SCALESPACE PROPERTIES 63
Grey level shift invariance
Since the diﬀusion tensor is only a function of J
ρ
(∇u
σ
), but not of u, we may shift
the grey level range by an arbitrary constant C, and the ﬁltered images will also
be shifted by the same constant. Moreover, a constant function is not aﬀected by
diﬀusion ﬁltering. Therefore, we have
T
t
(0) = 0, (2.19)
T
t
(f +C) = T
t
(f) +C ∀ t ≥ 0. (2.20)
Reverse contrast invariance
From D(J
ρ
(−∇u
σ
)) = D(J
ρ
(∇u
σ
)), it follows that
T
t
(−f) = −T
t
(f) ∀ t ≥ 0. (2.21)
This property is not fulﬁlled by classical morphological scalespace equations like
dilation and erosion. When reversing the contrast, the role of dilation and erosion
has to be exchanged as well.
Average grey level invariance
Average grey level invariance is a further property in which diﬀusion scalespaces
diﬀer from morphological scalespaces. In general, the evolution PDEs of the latter
ones are not of divergence form and do not preserve the mean grey value. A con
stant average grey level is essential for scalespace based segmentation algorithms
such as the hyperstack [307, 408]. It is also a desirable quality for applications
in medical imaging where grey values measure physical qualities of the depicted
object, for instance proton densities in MR images.
Proposition 1 (Conservation of average grey value).
The average grey level
µ :=
1
[Ω[
Ω
f(x) dx (2.22)
is not aﬀected by nonlinear diﬀusion ﬁltering:
1
[Ω[
Ω
T
t
f dx = µ ∀ t > 0. (2.23)
Proof:
Deﬁne I(t) :=
Ω
u(x, t) dx for all t ≥ 0. Then the Cauchy–Schwarz inequality
64 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
implies
[I(t)−I(0)[ =
Ω
u(x, t)−f(x)
dx
≤ [Ω[
1/2
u(t)−f
L
2
(Ω)
.
Since u ∈ C([0, T]; L
2
(Ω)), the preceding inequality gives the continuity of I(t) in
0.
For t > 0, Theorem 1, the divergence theorem and the boundary conditions yield
dI
dt
=
Ω
∂
t
u dx =
Γ
'D(J
ρ
(∇u
σ
))∇u, n` dS = 0.
Hence, I(t) must be constant for all t ≥ 0. 2
Average grey level invariance may be described by commuting operators, when
introducing an averaging operator M : L
1
(Ω) →L
1
(Ω) which maps f to a constant
image with the same mean grey level:
(Mf)(y) :=
1
[Ω[
Ω
f(x) dx ∀ y ∈ Ω. (2.24)
Then Proposition 1 and grey level shift invariance imply that the order of M and
T
t
can be exchanged:
M(T
t
f) = T
t
(Mf) ∀ t ≥ 0. (2.25)
When studying diﬀusion ﬁltering as a pure initial value problem in the domain
IR
2
, it also makes sense to investigate Euclidean transformations of an image. This
leads us to translation and isometry invariance.
Translation invariance
Deﬁne a translation τ
h
by (τ
h
f)(x) := f(x +h). Then diﬀusion ﬁltering fulﬁls
T
t
(τ
h
f) = τ
h
(T
t
f) ∀ t ≥ 0. (2.26)
This is a consequence of the fact that the diﬀusion tensor depends on J
ρ
(∇u
σ
)
solely, but not explicitly on x.
Isometry invariance
Let R ∈ IR
2×2
be an orthogonal transformation. If we apply R to f by deﬁning
Rf(x) := f(Rx), then the eigenvalues of the diﬀusion tensor are unaltered and
any eigenvector v is transformed into Rv. Thus, it makes no diﬀerence whether
the orthogonal transformation is applied before or after diﬀusion ﬁltering:
T
t
(Rf) = R(T
t
f) ∀ t ≥ 0. (2.27)
2.4. SCALESPACE PROPERTIES 65
2.4.2 Informationreducing properties
Nonenhancement of local extrema
Koenderink [240] required that a scalespace evolution should not create new level
curves when increasing the scale parameter. If this is satisﬁed, isointensity linking
through the scales is possible and a structure at a coarse scale can (in principle) be
traced back to the original image (causality). For this reason, he imposed that at
spatial extrema with nonvanishing determinant of the Hessian isophotes in scale
space are upwards convex. He showed that this constraint can be written as
sgn(∂
t
u) = sgn(∆u). (2.28)
A suﬃcient condition for the causality equation (2.28) to hold is requiring that
local extrema with positive or negative deﬁnite Hessians are not enhanced: an
extremum in ξ at time θ satisﬁes ∂
t
u > 0 if ξ is a minimum, and ∂
t
u < 0 if ξ
is a maximum. This implication is easily seen: In the ﬁrst case, for instance, the
eigenvalues η
1
, η
2
of the Hessian Hess(u) are positive. Thus,
∆u = trace(Hess(u)) = η
1
+η
2
> 0, (2.29)
giving immediately the causality requirement (2.28).
Nonenhancement of local extrema has ﬁrst been used by Babaud et al. [30] in
the context of linear diﬀusion ﬁltering. However, it is also satisﬁed by nonlinear
diﬀusion scalespaces, as we shall see now.
3
Theorem 2 (Nonenhancement of local extrema).
Let u be the unique solution of (P
c
) and consider some θ > 0. Suppose that ξ ∈ Ω
is a local extremum of u(., θ) with nonvanishing Hessian. Then,
∂
t
u(ξ, θ) < 0, if ξ is a local maximum, (2.30)
∂
t
u(ξ, θ) > 0, if ξ is a local minimum. (2.31)
Proof:
Let D(J
ρ
(∇u
σ
)) =: (d
ij
(J
ρ
(∇u
σ
))). Then we have
∂
t
u =
2
¸
i=1
2
¸
j=1
∂
x
i
d
ij
(J
ρ
(∇u
σ
))
∂
x
j
u +
2
¸
i=1
2
¸
j=1
d
ij
(J
ρ
(∇u
σ
)) ∂
x
i
x
j
u. (2.32)
Since ∇u(ξ, θ) = 0 and ∂
x
i
d
ij
(J
ρ
(∇u
σ
(ξ, θ))) is bounded, the ﬁrst term of the
righthand side of (2.32) vanishes in (ξ, θ).
3
As in the linear diﬀusion case, nonenhancement of local extrema generally does not imply
that their number is nonincreasing, cf. 1.2.5 and [342].
66 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
We know that the diﬀusion tensor D := D(J
ρ
(∇u
σ
(ξ, θ))) is positive deﬁnite.
Hence, there exists an orthogonal matrix S ∈ IR
2×2
such that
S
T
DS = diag(λ
1
, λ
2
) =: Λ
with λ
1
, λ
2
being the positive eigenvalues of D.
Now, let us assume that (ξ, θ) is a local maximum where H := Hess(u(ξ, θ))
and, thus, B := (b
ij
) := S
T
HS are negative deﬁnite. Then we have
b
ii
< 0 (i = 1, 2),
and by the invariance of the trace with respect to orthogonal transformations it
follows that
∂
t
u(ξ, θ) = trace (DH)
= trace (S
T
DS S
T
HS)
= trace (ΛB)
=
2
¸
i=1
λ
i
b
ii
< 0.
If ξ is a local minimum of u(x, θ), one proceeds in the same way utilizing the
positive deﬁniteness of the Hessian. 2
Nonenhancement of local extrema distinguishes anisotropic diﬀusion from clas
sical contrast enhancing methods such as highfrequency emphasis [163, pp. 182–
183], which do violate this principle. Although possibly behaving like backward
diﬀusion across edges, nonlinear diﬀusion is always in the forward region at ex
trema. This ensures its stability.
It should be noted that nonenhancement of local extrema is just one possibil
ity to end up with Koenderink’s causality requirement. Another way to establish
causality is via the extremum principle (2.15) following Hummel’s reasoning; see
[189] for more details.
Lyapunov functionals and behaviour for t →∞
Since scalespaces are intended to subsequently simplify an image, it is desirable
that, for t → ∞, we obtain the simplest possible image representation, namely a
constant image with the same average grey value as the original one. The following
theorem states that anisotropic diﬀusion ﬁltering always leads to a constant steady
state. This is due to the class of Lyapunov functionals associated with the diﬀusion
process.
2.4. SCALESPACE PROPERTIES 67
Theorem 3 (Lyapunov functionals and behaviour for t →∞).
Suppose that u is the solution of (P
c
) and let a, b, µ and M be deﬁned as in (P
c
),
(2.22) and (2.24), respectively. Then the following properties are valid:
(a) (Lyapunov functionals)
For all r ∈ C
2
[a, b] with r
′′
≥ 0 on [a, b], the function
V (t) := Φ(u(t)) :=
Ω
r(u(x, t)) dx (2.33)
is a Lyapunov functional:
(i) Φ(u(t)) ≥ Φ(Mf) for all t ≥ 0.
(ii) V ∈ C[0, ∞) ∩ C
1
(0, ∞) and V
′
(t) ≤ 0 for all t > 0.
Moreover, if r
′′
>0 on [a, b], then V (t) = Φ(u(t)) is a strict Lyapunov func
tional:
(iii) Φ(u(t)) = Φ(Mf) ⇐⇒
u(t) = Mf on
¯
Ω (if t > 0)
u(t) = Mf a.e. on Ω (if t = 0)
(iv) If t > 0, then V
′
(t) = 0 if and only if u(t) = Mf on
¯
Ω.
(v) V (0) = V (T) for T > 0 ⇐⇒
f = Mf a.e. on Ω and
u(t) = Mf on
¯
Ω (0, T]
(b) (Convergence)
(i) lim
t→∞
u(t) −Mf
L
p
(Ω)
= 0 for p ∈ [1, ∞).
(ii) In the 1D case, the convergence lim
t→∞
u(x, t) = µ is uniform on
¯
Ω.
Proof:
(a) (i) Since r ∈ C
2
[a, b] with r
′′
≥ 0 on [a, b], we know that r is convex on
[a, b]. Using the average grey level invariance and Jensen’s inequality we
obtain, for all t ≥ 0,
Φ(Mf) =
Ω
r
¸
1
[Ω[
Ω
u(x, t) dx
¸
dy
≤
Ω
¸
1
[Ω[
Ω
r(u(x, t)) dx
¸
dy
=
Ω
r(u(x, t)) dx
= Φ(u(t)). (2.34)
68 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
(ii) Let us start by proving the continuity of V (t) in 0. Thanks to the
maximum–minimum principle, we may choose a constant
L := max
s∈[a,b]
[r
′
(s)[
such that for all t > 0, the Lipschitz condition
[r(u(x, t))−r(f(x))[ ≤ L [u(x, t)−f(x)[
is veriﬁed a.e. on Ω. From this and the Cauchy–Schwarz inequality, we
get
[V (t)−V (0)[ ≤ [Ω[
1/2
r(u(t))−r(f)
L
2
(Ω)
≤ [Ω[
1/2
L u(t)−f
L
2
(Ω)
,
and by virtue of u ∈ C([0, T]; L
2
(Ω)), the limit t → 0
+
gives the an
nounced continuity in 0.
By Theorem 1 and the boundedness of r
′
on [a, b], we know that V is
diﬀerentiable for t > 0 and V
′
(t) =
Ω
r
′
(u) u
t
dx. Thus, the divergence
theorem yields
V
′
(t) =
Ω
r
′
(u) div (D(J
ρ
(∇u
σ
))∇u) dx
=
Γ
r
′
(u) 'D(J
ρ
(∇u
σ
))∇u, n`
. .. .
=0
dS
−
Ω
r
′′
(u)
. .. .
≥0
'∇u, D(J
ρ
(∇u
σ
))∇u`
. .. .
≥0
dx
≤ 0.
(iii) Let Φ(u(t)) = Φ(Mf).
If t > 0, then u(t) is continuous in
¯
Ω. Let us now show that equality in
the estimate (2.34) implies that u(t) = const. on
¯
Ω. To this end, assume
that u is not constant on
¯
Ω. Then, by the continuity of u, there exists
a partition Ω = Ω
1
∪ Ω
2
with [Ω
1
[, [Ω
2
[ ∈ (0, [Ω[) and
α :=
1
[Ω
1
[
Ω
1
u dx =
1
[Ω
2
[
Ω
2
u dx =: β.
From r
′′
>0 on [a, b] it follows that r is strictly convex on [a, b] and
r
¸
1
[Ω[
Ω
u dx
¸
= r
[Ω
1
[
[Ω[
α +
[Ω
2
[
[Ω[
β
2.4. SCALESPACE PROPERTIES 69
<
[Ω
1
[
[Ω[
r(α) +
[Ω
1
[
[Ω[
r(β)
≤
1
[Ω[
Ω
1
r(u) dx +
1
[Ω[
Ω
2
r(u) dx
=
1
[Ω[
Ω
r(u) dx.
If we utilize this result in the estimate (2.34) we observe that, for t >0,
Φ(u(t)) = Φ(Mf) implies that u(t) = const. on
¯
Ω. Thanks to the
average grey value invariance we ﬁnally obtain u(t) = Mf on
¯
Ω.
So let us turn to the case t = 0. From (i) and (ii), we conclude that
Φ(u(θ)) = Φ(Mf) for all θ > 0. Thus, we have u(θ) = Mf for all θ > 0.
For θ > 0, the Cauchy–Schwarz inequality gives
Ω
[u(x, θ) −µ[ dx ≤ [Ω[
1/2
u(θ) −Mf
L
2
(Ω)
= 0.
Since u ∈ C([0, T]; L
2
(Ω)), the limit θ → 0
+
ﬁnally yields u(0) = Mf
a.e. on Ω.
Conversely, it is obvious that u(t) = Mf (a.e.) on Ω implies Φ(u(t)) =
Φ(Mf).
(iv) Let t > 0 and V
′
(t) = 0. Then from
0 = V
′
(t) = −
Ω
r
′′
(u(x, t))
. .. .
>0
'∇u(x, t), D(J
ρ
(∇u
σ
(x, t)))∇u(x, t)` dx
and the smoothness of u we obtain
'∇u, D(J
ρ
(∇u
σ
))∇u` = 0 on
¯
Ω.
By the uniform boundedness of D, there exists some constant ν > 0,
such that
ν [∇u[
2
≤ '∇u, D(J
ρ
(∇u
σ
))∇u` on
¯
Ω (0, ∞).
Thus, we have ∇u(x, t) = 0 a.e. on Ω. Due to the continuity of ∇u,
this yields u(x, t) = const. for all x ∈ Ω, and the average grey level
invariance ﬁnally gives u(x, t) = µ on Ω.
Conversely, let u(x, t) = µ on Ω. Then,
V
′
(t) = −
Ω
r
′′
(u) '∇u, D(J
ρ
(∇u
σ
))∇u` dx = 0.
70 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
(v) Suppose that V (T) = V (0). Since V is decreasing, we have
V (t) = const. on [0, T].
Let ǫ > 0. Then for any t ∈ [ǫ, T], we have V
′
(t) = 0, and part (iv)
implies that u(t) = Mf on Ω. Now, the Cauchy–Schwarz inequality
gives
Ω
[f −Mf[ dx ≤ [Ω[
1/2
f −u(t)
L
2
(Ω)
.
As u ∈ C([0, T]; L
2
(Ω)), the limit t →0
+
yields f = Mf a.e. on Ω.
Conversely, if u(t) = Mf (a.e.) on Ω holds for all t ∈ [0, T], it is evident
that V (0) = V (T).
(b) (i) By the grey level shift invariance we know that v := u−Mf satisﬁes the
diﬀusion equation as well. We multiply this equation by v, integrate,
and use the divergence theorem to obtain
Ω
vv
t
dx = −
Ω
'∇v, D(J
ρ
(∇v
σ
))∇v` dx.
Since ∇v
σ
is bounded, we ﬁnd some ν > 0 such that
1
2
d
dt
(v
2
L
2
(Ω)
) ≤ −ν ∇v
2
L
2
(Ω)
.
For t > 0, there exists some x
0
with v(x
0
) = 0. Therefore, Poincar´e’s
inequality (cf. [9, p. 122]) may be applied giving
v
2
L
2
(Ω)
≤ C
0
∇v
2
L
2
(Ω)
with some constant C
0
= C
0
(Ω) > 0. This yields
d
dt
v
2
L
2
(Ω)
≤ −2ν C
0
v
2
L
2
(Ω)
and hence the exponential decay of v
L
2
(Ω)
to 0.
By the maximum principle, we know that v(t)
L
∞
(Ω)
is bounded by
f −Mf
L
∞
(Ω)
. Thus, for q ∈ IN, q ≥ 2, we get
v(t)
q
L
q
(Ω)
≤ f −Mf
q−2
L
∞
(Ω)
v(t)
2
L
2
(Ω)
→ 0,
and H¨older’s inequality gives, for 1 ≤ p < q < ∞,
v(t)
L
p
(Ω)
≤ [Ω[
(1/p)−(1/q)
v(t)
L
q
(Ω)
→ 0.
This proves the assertion.
2.4. SCALESPACE PROPERTIES 71
(ii) To prove uniform convergence in the onedimensional setting, we can
generalize and adapt methods from [202] to our case.
Let Ω = (0, a). From part (a) we know that V (t) :=
a
0
u
2
(x, t) dx is
nonincreasing and bounded from below. Thus, the sequence (V (i))
i∈IN
converges.
Since V ∈ C[0, ∞) ∩ C
1
(0, ∞) the mean value theorem implies
∃ t
i
∈ (i, i+1) : V
′
(t
i
) = V (i + 1) −V (i).
Thus, (t
i
)
i∈IN
→ ∞ and from the convergence of (V (i))
i∈IN
it follows
that
V
′
(t
i
) →0. (2.35)
Thanks to the uniform positive deﬁniteness of D there exists some ν > 0
such that, for t > 0,
V
′
(t) = −2
a
0
u
2
x
D(J
ρ
(∂
x
u
σ
)) dx
≤ −2ν
a
0
u
2
x
dx
≤ 0. (2.36)
Equations (2.35) and (2.36) yield
u
x
(t
i
)
L
2
(Ω)
→0.
Hence, u(t
i
) is a bounded sequence in H
1
(0, a). By virtue of the Rellich–
Kondrachov theorem [7, p. 144] we know that the embedding from
H
1
(0, a) into C
0,α
[0, a], the space of H¨oldercontinuous functions on [0, a]
[7, pp. 9–12], is compact for α∈(0,
1
2
). Therefore, there exists a subse
quence (t
i
j
) →∞ and some ¯ u with
u(t
i
j
) → ¯ u in C
0,α
[0, a].
This also gives u(t
i
j
) → ¯ u in L
2
(0, a). Since we already know from (b)(i)
that u(t
i
j
) →Mf in L
2
(0, a), it follows that ¯ u = Mf. Hence,
lim
j→∞
u(t
i
j
) −Mf
L
∞
(Ω)
= 0. (2.37)
Part (a) tells us that u(t)−Mf
p
L
p
(Ω)
is a Lyapunov function for p ≥ 2.
Thus,
u(t)−Mf
L
∞
(Ω)
= lim
p→∞
u(t)−Mf
L
p
(Ω)
72 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
is also nonincreasing. Therefore, lim
t→∞
u(t)−Mf
L
∞
(Ω)
exists and from
(2.37) we conclude that
lim
t→∞
u(t)−Mf
L
∞
(Ω)
= 0.
The smoothness of u establishes ﬁnally that the convergence
lim
t→∞
u(x, t) = µ
is uniform on
¯
Ω. 2
Since the class (P
c
) does not forbid contrast enhancement it admits processes
where forward diﬀusion has to compete with backward diﬀusion. Theorem 3 is
of importance as it states that the regularization by convolving with K
σ
tames
the backward diﬀusion in such a way that forward diﬀusion wins in the long run.
Moreover, the competition evolves in a certain direction all the time: although
backward diﬀusion may be locally superior, the global result – denoted by the
Lyapunov functional – becomes permanently better for forward diﬀusion.
Let us have a closer look at what might be the meaning of this global result in
the context of image processing. Considering the Lyapunov functions associated
with r(s) := [s[
p
, r(s) := (s−µ)
2n
and r(s) := s ln s, respectively, the preceding
theorem gives the following corollary.
Corollary 1 (Special Lyapunov functionals).
Let u be the solution of (P
c
) and a and µ be deﬁned as in (P
c
) and (2.22). Then
the following functions are decreasing for t ∈ [0, ∞):
(a) u(t)
L
p
(Ω)
for all p ≥ 2.
(b) M
2n
[u(t)] :=
1
[Ω[
Ω
(u(x, t) −µ)
2n
dx for all n ∈ IN.
(c) H[u(t)] :=
Ω
u(x, t) ln(u(x, t)) dx, if a > 0.
Corollary 1 oﬀers multiple possibilities of how to interpret nonlinear anisotropic
diﬀusion ﬁltering as a smoothing transformation.
As a special case of (a) it follows that the energy u(t)
2
L
2
(Ω)
is reduced by
diﬀusion.
Part (b) gives a probabilistic interpretation of anisotropic diﬀusion ﬁltering.
Consider the intensity in an image f as a random variable Z
f
with distribution
2.4. SCALESPACE PROPERTIES 73
F
f
(z), i.e. F
f
(z) is the probability that an arbitrary grey value Z
f
of f does not
exceed z. By the average grey level invariance, µ is equal to the expected value
EZ
u(t)
:=
IR
z dF
u(t)
(z), (2.38)
and it follows that M
2n
[u(t)] is just the even central moment
IR
z−EZ
u(t)
2n
dF
u(t)
(z). (2.39)
The second central moment (the variance) characterizes the spread of the intensity
about its mean. It is a common tool for constructing measures for the relative
smoothness of the intensity distribution. The fourth moment is frequently used to
describe the relative ﬂatness of the grey value distribution. Higher moments are
more diﬃcult to interpret, although they do provide important information for
tasks like texture discrimination [163, pp. 414–415]. All decreasing even moments
demonstrate that the image becomes smoother during diﬀusion ﬁltering. Hence,
local eﬀects such as edge enhancement, which object to increase central moments,
are overcompensated by smoothing in other areas.
If we choose another probabilistic model of images, then part (c) characterizes
the informationtheoretical side of our scalespace. Provided the initial image f is
strictly positive on Ω, we may regard it also as a twodimensional density.
4
Then,
S[u(t)] := −
Ω
u(x, t) ln(u(x, t)) dx (2.40)
is called the entropy of u(t), a measure of uncertainty and missing information [63].
Since anisotropic diﬀusion ﬁlters increase the entropy the corresponding scalespace
embeds the genuine image f into a family of subsequently likelier versions of it
which contain less information. Moreover, for t →∞, the process reaches the state
with the lowest possible information, namely a constant image. This information
reducing property indicates that anisotropic diﬀusion might be generally useful in
the context of image compression. In particular, it helps to explain the success of
nonlinear diﬀusion ﬁltering as a preprocessing step for subsampling as observed in
[144]. The interpretation of the entropy in terms of Lyapunov functionals carries
also over to generalized entropies; see [390] for more details.
From all the previous considerations, we recognize that, in spite of possible
contrastenhancing properties, anisotropic diﬀusion does really simplify the ori
ginal image in a steady way.
Let us ﬁnally point out another interpretation of the Lyapunov functionals. In
a classic scalespace representation, the time t plays the role of the scale para
meter. By increasing t, one transforms the image from a local to a more global
4
Without loss of generality we omit the normalization.
74 CHAPTER 2. CONTINUOUS DIFFUSION FILTERING
representation. We have seen in Chapter 1 that, for linear diﬀusion scalespaces
and morphological scalespaces, it is possible to associate with the evolution time
a corresponding spatial scale.
In the nonlinear diﬀusion case, however, the situation is more complicated.
Since the smoothing is nonuniform, one can only deﬁne an average measure for
the globality of the representation. This can be achieved by taking some Lyapunov
function Φ(u(t)) and investigating the expression
Ψ(u(t)) :=
Φ(f) −Φ(u(t))
Φ(f) −Φ(Mf)
. (2.41)
We observe that Ψ(t) increases from 0 to 1. It gives the average globality of u(t) and
its value can be used to measure the distance of u(t) from the initial state f and the
ﬁnal state Mf. Prescribing a certain value for Ψ provides us with an aposteriori
criterion for the stopping time of the nonlinear diﬀusion process. Experiments in
this direction can be found in [431, 308].
Chapter 3
Semidiscrete diﬀusion ﬁltering
The goal of this chapter is to study a semidiscrete framework for diﬀusion scale
spaces where the image is sampled on a ﬁnite grid and the scale parameter is
continuous. This leads to a system of nonlinear ordinary diﬀerential equations
(ODEs). We shall investigate conditions under which one can establish similar
properties as in the continuous setting concerning wellposedness, extremum prin
ciples, average grey level invariance, Lyapunov functions, and convergence to a
constant steadystate. Afterwards we shall discuss whether it is possible to obtain
such ﬁlters from spatial discretizations of the continuous models that have been
investigated in Chapter 2. We will see that there exists a ﬁnite stencil on which
a diﬀerence approximation of the spatial derivatives are in accordance with the
semidiscrete scalespace framework.
3.1 The general model
A discrete image can be regarded as a vector f ∈IR
N
, N ≥2, whose components
f
j
, j = 1,...,N represent the grey values at each pixel. We denote the index set
¦1, ..., N¦ by J. In order to specify the requirements for our semidiscrete ﬁlter
class we ﬁrst recall a useful deﬁnition of irreducible matrices [407, pp. 18–20].
Deﬁnition 1 (Irreducibility). A matrix A = (a
ij
) ∈ IR
N×N
is called irreducible
if for any i, j ∈ J there exist k
0
,...,k
r
∈ J with k
0
= i and k
r
= j such that
a
kpk
p+1
= 0 for p = 0,...,r−1.
The semidiscrete problem class (P
s
) we are concerned with is deﬁned in the
following way:
75
76 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
Let f ∈ IR
N
. Find a function u ∈ C
1
([0, ∞), IR
N
) which satisﬁes an
initial value problem of type
du
dt
= A(u) u,
u(0) = f,
where A = (a
ij
) has the following properties:
(S1) Lipschitzcontinuity of A ∈ C(IR
N
, IR
N×N
) for every bounded
subset of IR
N
,
(S2) symmetry: a
ij
(u) = a
ji
(u) ∀ i, j ∈ J, ∀ u ∈ IR
N
,
(S3) vanishing row sums:
¸
j∈J
a
ij
(u) = 0 ∀ i ∈ J, ∀ u ∈ IR
N
,
(S4) nonnegative oﬀdiagonals: a
ij
(u) ≥ 0 ∀ i = j, ∀ u ∈ IR
N
,
(S5) irreducibility for all u ∈ IR
N
.
(P
s
)
Not all of these requirements are necessary for every theoretical result below.
(S1) is needed for wellposedness, the proof of a maximum–minimum principle
involves (S3) and (S4), while average grey value invariance uses (S2) and (S3).
The existence of Lyapunov functions can be established by means of (S2)–(S4),
and strict Lyapunov functions and the convergence to a constant steadystate
require (S5) in addition to (S2)–(S4).
This indicates that these properties reveal some interesting parallels to the
continuous setting from Chapter 2: In both cases we need smoothness assumptions
to ensure wellposedness; (S2) and (S3) correspond to the speciﬁc structure of the
divergence expression with a symmetric diﬀusion tensor D, while (S4) and (S5)
play a similar role as the nonnegativity of the eigenvalues of D and its uniform
positive deﬁniteness, respectively.
3.2 Theoretical results
Before we can establish scalespace results, it is of importance to ensure the ex
istence of a unique solution. This is done in the theorem below which also states
the continuous dependence of the solution and a maximum–minimum principle.
3.2. THEORETICAL RESULTS 77
Theorem 4 (Wellposedness, extremum principle).
For every T > 0 the problem (P
s
) has a unique solution u(t) ∈ C
1
([0, T], IR
N
).
This solution depends continuously on the initial value and the righthand side of
the ODE system, and it satisﬁes the extremum principle
a ≤ u
i
(t) ≤ b ∀ i ∈ J, ∀ t ∈ [0, T], (3.1)
where
a := min
j∈J
f
j
, (3.2)
b := max
j∈J
f
j
. (3.3)
Proof:
(a) Local existence and uniqueness
Local existence and uniqueness are proved by showing that our problem
satisﬁes the requirements of the Picard–Lindel¨of theorem [432, p. 59].
Let t
0
:= 0 and β > 0. Evidently, φ(t, u) := ψ(u) := A(u) u is continuous on
B
0
:= [0, T]
u ∈ IR
N
u
∞
≤ f
∞
+β
¸
,
since it is a composition of continuous functions. Moreover, by the compact
ness of B
0
there exists some c > 0 with
φ(t, u)
∞
≤ c ∀ (t, u) ∈ B
0
.
In order to prove existence and uniqueness of a solution of (P
s
) in
R
0
:=
(t, u)
t ∈ [t
0
, t
0
+min(
β
c
, T)], u−f
∞
≤ β
¸
⊂ B
0
we have to show that φ(t, u) satisﬁes a global Lipschitz condition on R
0
with
respect to u. However, this follows directly from the fact that A is Lipschitz
continuous on ¦u ∈ IR
N
[ u−f
∞
≤ β¦.
(b) Maximum–minimum principle
We prove only the maximum principle, since the proof for the minimum
principle is analogous.
Assume that the problem (P
s
) has a unique solution on [0, θ]. First we show
that the derivative of the largest component of u(t) is nonpositive for every
78 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
t ∈ [0, θ]. Let u
k
(ϑ) := max
j∈J
u
j
(ϑ) for some arbitrary ϑ ∈ [0, θ]. If we keep
this k ﬁxed we obtain, for t = ϑ,
du
k
dt
=
¸
j∈J
a
kj
(u) u
j
= a
kk
(u) u
k
+
¸
j∈J\{k}
a
kj
(u)
. .. .
≥0
u
j
....
≤u
k
≤ u
k
¸
j∈J
a
kj
(u)
(S4)
= 0. (3.4)
Let us now prove that this implies a maximum principle (cf. [201]).
Let ε > 0 and set
u
ε
(t) := u(t) −
¸
¸
¸
εt
.
.
.
εt
¸
.
Moreover, let P := ¦p∈J [ u
εp
(0) = max
j∈J
u
εj
(0)¦. Then, by (3.4),
du
εp
dt
(0) =
du
p
dt
(0)
. .. .
≤0
−ε < 0 ∀ p ∈ P. (3.5)
By means of
max
i∈J\P
u
εi
(0) < max
j∈J
u
εj
(0),
and the continuity of u there exists some t
1
∈ (0, θ) such that
max
i∈J\P
u
εi
(t) < max
j∈J
u
εj
(0) ∀ t ∈ [0, t
1
). (3.6)
Next, let us consider some p ∈ P. Due to (3.5) and the smoothness of u we
may ﬁnd a ϑ
p
∈ (0, θ) with
du
εp
dt
(t) < 0 ∀ t ∈ [0, ϑ
p
).
Thus, we have
u
εp
(t) < u
εp
(0) ∀ t ∈ (0, ϑ
p
)
and, for t
2
:= min
p∈P
ϑ
p
, it follows that
max
p∈P
u
εp
(t) < max
j∈J
u
εj
(0) ∀ t ∈ (0, t
2
). (3.7)
3.2. THEORETICAL RESULTS 79
Hence, for t
0
:= min(t
1
, t
2
), the estimates (3.6) and (3.7) give
max
j∈J
u
εj
(t) < max
j∈J
u
εj
(0) ∀ t ∈ (0, t
0
). (3.8)
Now we prove that this estimate can be extended to the case t ∈ (0, θ). To
this end, assume the opposite is true. Then, by virtue of the intermediate
value theorem, there exists some t
3
which is the smallest time in (0, θ) such
that
max
j∈J
u
εj
(t
3
) = max
j∈J
u
εj
(0).
Let u
εk
:= max
j∈J
u
εj
(t
3
). Then the minimality of t
3
yields
u
εk
(t) < u
εk
(t
3
) ∀ t ∈ (0, t
3
), (3.9)
and inequality (3.4) gives
du
εk
dt
(t
3
) =
du
k
dt
(t
3
)
. .. .
≤0
−ε < 0.
Due to the continuity of
du
dt
there exists some t
4
∈ (0, t
3
) with
du
εk
dt
(t) < 0 ∀ t ∈ (t
4
, t
3
]. (3.10)
The mean value theorem, however, implies that we ﬁnd a t
5
∈ (t
4
, t
3
) with
du
εk
dt
(t
5
) =
u
εk
(t
3
) −u
εk
(t
4
)
t
3
−t
4
(3.9)
> 0,
which contradicts (3.10). Hence, (3.8) must be valid on the entire interval
(0, θ).
Together with u = lim
ε→0
u
ε
and the continuity of u this yields the announced
maximum principle
max
j∈J
u
j
(t) ≤ max
j∈J
u
j
(0) ∀ t ∈ [0, θ].
(c) Global existence and uniqueness
Global existence and uniqueness follow from local existence and uniqueness
when being combined with the extremum principle.
Using the notations and results from (a), we know that the problem (P
s
) has
a unique solution u(t) for t ∈ [t
0
, t
0
+ min(
β
c
, T)].
80 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
Now let t
1
:= t
0
+ min(
β
c
, T), g := u(t
1
), and consider the problem
du
dt
= A(u) u,
u(t
1
) = g.
Clearly, φ(t, u) = A(u)u is continuous on
B
1
:= [0, T]
u ∈ IR
N
u
∞
≤ g
∞
+β
¸
,
and by the extremum principle we know that B
1
⊂ B
0
. Hence,
φ(t, u)
∞
≤ c ∀ (t, u) ∈ B
1
.
with the same c as in (a). Using the same considerations as in (a) one shows
that φ is Lipschitzcontinuous on
R
1
:=
(t, u)
t ∈ [t
1
, t
1
+min(
β
c
, T)], u−g
∞
≤ β
¸
.
Hence, the considered problem has a unique solution on [t
1
, t
1
+ min(
β
c
, T)].
Therefore, (P
s
) reveals a unique solution on [0, min(
2β
c
, T)], and, by iterating
this reasoning, the existence of a unique solution can be extended to the
entire interval [0, T]. As a consequence, the extremum principle is valid on
[0, T] as well.
(d) Continuous dependence
Let u(t) be the solution of
du
dt
= φ(t, u),
u(0) = f
for t ∈[0, T] and φ(u, t) =ψ(u) =A(u)u. In order to show that u(t) depends
continuously on the initial data and the righthand side of the ODE system,
it is suﬃcient to prove that φ(t, u) is continuous, and that there exists some
α > 0 such that φ(t, u) satisﬁes a global Lipschitz condition on
S
α
:=
(t, v)
t ∈ [0, T], v−u
∞
≤ α
¸
.
with respect to its second argument. In this case the results in [412, p. 93]
ensure that for every ε > 0 there exists a δ > 0 such that the solution ˜ u of
the perturbed problem
d˜ u
dt
=
˜
φ(t, ˜ u),
˜ u(0) =
˜
f
3.3. SCALESPACE PROPERTIES 81
with continuous
˜
φ and

˜
f −f
∞
< δ,

˜
φ(t, v) −φ(t, v)
∞
< δ for v −u
∞
< α
exists in [0, T] and satisﬁes the inequality
˜ u(t) −u(t)
∞
< ε.
Similar to the the local existence and uniqueness proof, the global Lipschitz
condition on S
α
follows direcly from the fact that A is Lipschitzcontinuous
on ¦v ∈ IR
N
[ v−u
∞
≤ α¦. 2
3.3 Scalespace properties
It is evident that properties such as grey level shift invariance or reverse contrast
invariance are automatically satisﬁed by every consistent semidiscrete approxima
tion of the continuous ﬁlter class (P
c
). On the other hand, translation invariance
only makes sense for translations in grid direction with multiples of the grid size,
and isometry invariance is satisﬁed for consistent schemes up to an discretization
error. So let us focus on average grey level invariance now.
Proposition 2 (Conservation of average grey value).
The average grey level
µ :=
1
N
¸
j∈J
f
j
(3.11)
is not aﬀected by the semidiscrete diﬀusion ﬁlter:
1
N
¸
j∈J
u
j
(t) = µ ∀ t ≥ 0. (3.12)
Proof:
By virtue of (S2) and (S3) we have
¸
j∈J
a
jk
(u) = 0 for all k ∈ J. Thus, for t ≥ 0,
¸
j∈J
du
j
dt
=
¸
j∈J
¸
k∈J
a
jk
(u) u
k
=
¸
k∈J
¸
j∈J
a
jk
(u)
u
k
= 0,
which shows that
¸
j∈J
u
j
(t) is constant on [0, ∞) and concludes the proof. 2
This property is in complete accordance with the result for the continuous ﬁlter
class.
82 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
Similar to the continuous setting, it is possible to ﬁnd a large class of Lyapunov
functions which establish smoothing scalespace properties and ensure that the
image tends to a constant steadystate with the same average grey level as the
initial image.
Theorem 5 (Lyapunov functions and behaviour for t →∞).
Let u(t) be the solution of (P
s
), let a, b, and µ be deﬁned as in (3.2), (3.3), and
(3.11), respectively, and let c := (µ, µ, ..., µ)
⊤
∈ IR
N
.
Then the following properties are valid:
(a) (Lyapunov functions)
For all r ∈ C
1
[a, b] with increasing r
′
on [a, b], the function
V (t) := Φ(u(t)) :=
¸
i∈J
r(u
i
(t))
is a Lyapunov function:
(i) Φ(u) ≥ Φ(c) for all t ≥ 0.
(ii) V ∈ C
1
[0, ∞) and V
′
(t) ≤ 0 for all t ≥ 0.
Moreover, if r
′
is strictly increasing on [a, b], then V (t) = Φ(u(t)) is a strict
Lyapunov function:
(iii) Φ(u) = Φ(c) ⇐⇒ u = c
(iv) V
′
(t) = 0 ⇐⇒ u = c
(b) (Convergence)
lim
t→∞
u(t) = c.
Proof:
(a) (i) Since r
′
is increasing on [a, b] we know that r is convex on [a, b]. Average
grey level invariance and this convexity yield, for all t ≥ 0,
Φ(c) =
N
¸
i=1
r
¸
N
¸
j=1
1
N
u
j
¸
≤
N
¸
i=1
¸
1
N
N
¸
j=1
r(u
j
)
¸
=
N
¸
j=1
r(u
j
)
= Φ(u). (3.13)
3.3. SCALESPACE PROPERTIES 83
(ii) Since u ∈ C
1
([0, ∞), IR
N
) and r ∈ C
1
[a, b], it follows that V ∈ C
1
[0, ∞).
Using the prerequisites (S2) and (S3) we get
V
′
(t) =
N
¸
i=1
du
i
dt
r
′
(u
i
)
(S3)
=
N
¸
i=1
N
¸
j=1
a
ij
(u) (u
j
−u
i
) r
′
(u
i
)
=
N
¸
i=1
¸
N
¸
j=i+1
+
i−1
¸
j=1
¸
a
ij
(u) (u
j
−u
i
) r
′
(u
i
)
=
N
¸
i=1
N−i
¸
k=1
a
i,i+k
(u) (u
i+k
−u
i
) r
′
(u
i
)
+
N
¸
i=1
N−i
¸
k=1
a
i+k,i
(u) (u
i
−u
i+k
) r
′
(u
i+k
)
(S2)
=
N
¸
i=1
N−i
¸
k=1
a
i,i+k
(u) (u
i+k
−u
i
)
r
′
(u
i
)−r
′
(u
i+k
)
. (3.14)
Since r
′
is increasing, we always have
(u
i+k
−u
i
)
r
′
(u
i
) −r
′
(u
i+k
)
≤ 0.
With this and (S4), equation (3.14) implies that V
′
(t) ≤ 0 for t ≥ 0.
(iii) Let us ﬁrst prove that equality in the estimate (3.13) implies that all
components of u are equal.
To this end, suppose that u
i
0
:= min
i
u
i
< max
j
u
j
=: u
j
0
and let
η :=
N
¸
j=1
j=j
0
1
N
u
j
1 −
1
N
.
Then, η <u
j
0
. Since r
′
is strictly increasing on [a, b], we know that r is
strictly convex. Hence, we get
r
N
¸
i=1
1
N
u
j
= r
1
N
u
j
0
+
1−
1
N
η
<
1
N
r(u
j
0
) +
1−
1
N
r(η)
≤
1
N
r(u
j
0
) +
N
¸
j=1
j=j
0
1
N
r(u
j
)
=
N
¸
j=1
1
N
r(u
j
).
84 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
This shows that equality in (3.13) implies that u
1
= ... = u
N
. By virtue
of the grey level shift invariance we conclude that u = c.
Conversely, it is trivial that Φ(u) = Φ(c) for u = c.
(iv) Let V
′
(t) = 0. From (3.14) we have
0 = V
′
(t) =
N
¸
i=1
N−i
¸
k=1
a
i,i+k
(u) (u
i+k
−u
i
)
r
′
(u
i
)−r
′
(u
i+k
)
. .. .
≤0
,
and by virtue of the symmetry of A(u) it follows that
a
ij
(u) (u
j
−u
i
)
r
′
(u
i
)−r
′
(u
j
)
= 0 ∀ i, j ∈ J. (3.15)
Now consider two arbitrary i
0
, j
0
∈ J. The irreducibility of A(u) implies
that there exist k
0
,...,k
r
∈ J with k
0
= i
0
, k
r
= j
0
, and
a
kpk
p+1
(u) = 0, p = 0, ..., r −1.
As r
′
is strictly increasing we have, for p = 0,...,r −1,
(u
kp
−u
k
p+1
)
r
′
(u
k
p+1
)−r
′
(u
kp
)
= 0 ⇐⇒ u
kp
= u
k
p+1
.
From this and (3.15) we get
u
i
0
= u
k
0
= u
k
1
= ... = u
kr
= u
j
0
.
Since i
0
and j
0
are arbitrary, we obtain u
i
= const. for all i ∈ J,
and the average grey level invariance gives u = c. This proves the ﬁrst
implication.
Conversely, let u
i
= const. for all i ∈ J. Then from the representation
(3.14) we immediately conclude that V
′
(t) = 0.
(b) The convergence proof is based on classical Lyapunov reasonings, see e.g.
[180] for an introduction to these techniques.
Consider the Lyapunov function V (t) := Φ(u(t)) := [u(t)−c[
2
, which results
from the choice r(s) := (s−µ)
2
. Since V (t) is decreasing and bounded from
below by 0, we know that lim
t→∞
V (t) =: η exists and η ≥ 0.
Now assume that η > 0.
Since [u(t)−c[ is bounded from above by α := [f −c[ we have
[u(t)−c[ ≤ α ∀ t ≥ 0. (3.16)
By virtue of Φ(x) = [x −c[
2
we know that, for β ∈ (0,
√
η),
Φ(x) < η ∀ x ∈ IR
N
, [x−c[ < β.
3.3. SCALESPACE PROPERTIES 85
Let w.l.o.g. β < α. Since Φ(u(t)) ≥ η we conclude that
[u(t)−c[ ≥ β ∀ t ≥ 0. (3.17)
So from (3.16) and (3.17) we have
u(t) ∈ ¦x ∈ IR
N
[ β ≤ [x−c[ ≤ α¦ =: S ∀ t ≥ 0.
By (a)(ii),(iv), the compactness of S, and β > 0 there exists some M > 0
such that
V
′
(t) ≤ −M ∀ t ≥ 0.
Therefore, it follows
V (t) = V (0) +
t
0
V
′
(θ) dθ ≤ V (0) −tM
which implies lim
t→∞
V (t) = −∞ and, thus, contradicts (a)(i).
Hence the assumption η > ρ is wrong and we must have η = ρ.
According to (a)(iii) this yields lim
t→∞
u(t) = c. 2
As in the continuous case, we can consider the Lyapunov functions associated
with r(s) := [s[
p
, r(s) := (s−µ)
2n
and r(s) := s ln s, respectively, and obtain the
following corollary.
Corollary 2 (Special Lyapunov functions).
Let u be the solution of (P
s
) and a and µ be deﬁned as in (3.2) and (3.11). Then
the following functions are decreasing for t ∈ [0, ∞):
(a) u(t)
p
for all p ≥ 2.
(b) M
2n
[u(t)] :=
1
N
N
¸
j=1
(u
j
(t) −µ)
2n
for all n ∈ IN.
(c) H[u(t)] :=
N
¸
j=1
u
j
(t) ln(u
j
(t)), if a > 0.
Since all pnorms (p ≥ 2) and all central moments are decreasing, while the
discrete entropy
S[u(t)] := −
N
¸
j=1
u
j
(t) ln(u
j
(t)) (3.18)
is increasing with respect to t, we observe that the semidiscrete setting reveals
smoothing scalespace properties which are closely related to the continuous case.
86 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
3.4 Relation to continuous models
In this section we investigate whether it is possible to use spatial discretizations of
the continuous ﬁlter class (P
c
) in order to construct semidiscrete diﬀusion models
satisfying (S1)–(S5). First we shall verify that this is easily done for isotropic
models. In the anisotropic case, however, the mixed derivative terms make it more
diﬃcult to ensure nonnegative oﬀdiagonal elements. A constructive existence proof
is presented showing that for a suﬃciently large stencil it is always possible to ﬁnd
such a nonnegative discretization. This concept is illustrated by investigating the
situation on a (33)stencil in detail.
3.4.1 Isotropic case
Let the rectangle Ω = (0, a
1
) (0, a
2
) be discretized by a grid of N = n
1
n
2
pixels
such that a pixel (i, j) with 1≤i ≤n
1
and 1≤j ≤n
2
represents the location (x
i
, y
j
)
where
x
i
:= (i −
1
2
) h
1
, (3.19)
y
j
:= (j −
1
2
) h
2
, (3.20)
and the grid sizes h
1
, h
2
are given by h
1
:= a
1
/n
1
and h
2
:= a
2
/n
2
, respectively.
These pixels can be numbered by means of an arbitrary bijection
p : ¦1, ..., n
1
¦ ¦1, ..., n
2
¦ → ¦1, ..., N¦. (3.21)
Thus, pixel (i, j) is represented by a single index p(i, j).
Let us now verify that a standard FD space discretization of an isotropic variant
of (P
c
) leads to a semidiscrete ﬁlter satisfying the requirements (S1)–(S5). To
this end, we may replace the diﬀusion tensor D(J
ρ
(∇u
σ
)) by some scalarvalued
function g(J
ρ
(∇u
σ
)). The structure tensor requires the calculations of convolutions
with ∇K
σ
and K
ρ
, respectively. In the spatially discrete case this comes down to
speciﬁc vector–matrix multiplications. For this reason, we may approximate the
structure tensor by some matrix H(u) = (h
ij
(u)) where H ∈ C
∞
(IR
N
, IR
2×2
).
Next, consider some pixel k = p(i, j). Then a consistent spatial discretization of
the isotropic diﬀusion equation with homogeneous Neumann boundary conditions
can be written as
du
k
dt
=
2
¸
n=1
¸
l∈Nn(k)
g
l
+g
k
2h
2
l
(u
l
−u
k
), (3.22)
where ^
n
(k) consists of the one or two neighbours of pixel k along the nth coor
dinate axis (boundary pixels have only one neighbour) and g
k
:= g ((H(u))
k
).
In vector–matrix notation (3.22) becomes
du
dt
= A(u) u, (3.23)
3.4. RELATION TO CONTINUOUS MODELS 87
where the matrix A(u) = (a
kl
(u))
kl
is given by
a
kl
:=
g
k
+g
l
2h
2
n
(l ∈ ^
n
(k)),
−
2
¸
n=1
¸
l∈Nn(k)
g
k
+g
l
2h
2
n
(l = k),
0 (else).
(3.24)
Let us now verify that (S1)–(S5) are fulﬁlled.
Since H ∈ C
∞
(IR
N
, IR
2×2
) and g ∈ C
∞
(IR
2×2
), we have A ∈ C
∞
(IR
N
, IR
N×N
).
This proves (S1).
The symmetry of A follows directly from (3.24) and the symmetry of the neigh
bourhood relation:
l ∈ ^
n
(k) ⇐⇒ k ∈ ^
n
(l).
By the construction of A it is also evident that all row sums vanish, i.e. (S3)
is satisﬁed. Moreover, since g is positive, it follows that a
kl
≥ 0 for all k = l and,
thus, (S4) holds.
In order to show that A is irreducible, let us consider two arbitrary pixels k and
l. Then we have to ﬁnd k
0
,...,k
r
∈ J with k
0
=k and k
r
=l such that a
kqk
q+1
= 0
for q = 0,...,r−1. If k = l, we already know from (3.24) that a
kk
< 0. In this case
we have the trivial path k = k
0
= k
r
= l. For k = l, we may choose any arbitrary
path k
0
,...,k
r
, such that k
q
and k
q+1
are neighbours for q = 0,...,r−1. Then,
a
kqk
q+1
=
g
kq
+g
k
q+1
2h
2
n
> 0
for some n ∈ ¦1, 2¦. This proves (S5).
Remarks:
(a) We observe that (S1)–(S5) are properties which are valid for all arbitrary
pixel numberings.
(b) The ﬁlter class (P
c
) is not the only family which leads to semidiscrete ﬁl
ters satisfying (S1)–(S5). Interestingly, a semidiscrete version of the Perona–
Malik ﬁlter – which is to a certain degree illposed in the continuous setting
(cf. 1.3.1) – also satisﬁes (S1)–(S5) and, thus, reveals all the discussed well
posedness and scalespace properties [425]. This is due to the fact that the
extremum principle limits the modulus of discrete gradient approximations.
Hence, the spatial discretization implicitly causes a regularization. These
results are also in accordance with a recent paper by Pollak et al. [332].
They study an image evolution under an ODE system with a discontinuous
right hand side, which has some interesting relations to the limit case of a
semidiscrete Perona–Malik model. They also report stable behaviour of their
process.
88 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
3.4.2 Anisotropic case
If one wishes to transfer the results from the isotropic case to the general aniso
tropic setting the main diﬃculty arises from the fact that, due to the mixed deriva
tive expressions, it is not obvious how to ensure (S4), the nonnegativity of all oﬀ
diagonal elements of A(u). The theorem below states that this is always possible
for a suﬃciently large stencil.
Theorem 6 (Existence of a nonnegative discretization).
Let D ∈ IR
2×2
be symmetric positive deﬁnite with a spectral condition number
κ. Then there exists some m(κ) ∈IN such that div (D∇u) reveals a secondorder
nonnegative FD discretization on a (2m+1)(2m+1)stencil.
Proof:
Let us consider some m ∈ IN and the corresponding (2m+1)(2m+1)stencil. The
“boundary pixels” of this stencil deﬁne 4m principal orientations β
i
∈ (−
π
2
,
π
2
],
i = −2m+1, ..., 2m according to
β
i
:=
arctan
ih
2
mh
1
([i[ ≤ m),
arccot
(2m−i)h
1
mh
2
(m < i ≤ 2m),
arccot
(i−2m)h
1
mh
2
(−2m+ 1 ≤ i < −m).
Now let J
m
:= ¦1, ..., 2m−1¦ and deﬁne a partition of (−
π
2
,
π
2
] into 4m−2 subin
tervals I
i
, [i[ ∈J
m
:
(−
π
2
,
π
2
] =
−1
¸
i=−2m+1
(θ
i
, θ
i+1
]
. .. .
=:I
i
∪
2m−1
¸
i=1
(θ
i−1
, θ
i
]
. .. .
=:I
i
,
where
θ
i
:=
0 (i = 0),
1
2
arctan
2
cot β
i
−tan β
i+1
(i ∈ ¦1, ..., 2m−2¦, β
i
+β
i+1
<
π
2
),
π
4
(i ∈ ¦1, ..., 2m−2¦, β
i
+β
i+1
=
π
2
),
π
2
+
1
2
arctan
2
cot β
i
−tan β
i+1
(i ∈ ¦1, ..., 2m−2¦, β
i
+β
i+1
>
π
2
),
π
2
(i = 2m−1),
and
θ
i
:= −θ
−i
(i ∈ ¦−2m+1, ..., −1¦).
It is not hard to verify that β
i
∈I
i
for [i[ ∈J
m
.
Let λ
1
≥ λ
2
> 0 be the eigenvalues of D with corresponding eigenvectors
(cos ψ, sin ψ)
⊤
and (−sin ψ, cos ψ)
⊤
, where ψ ∈ (−
π
2
,
π
2
]. Now we show that for
a suitable m there exists a stencil direction β
k
, [k[ ∈ J
m
such that the splitting
div (D∇u) = ∂
e
β
0
α
0
∂
e
β
0
u
+∂
e
β
k
α
k
∂
e
β
k
u
+∂
e
β
2m
α
2m
∂
e
β
2m
u
(3.25)
3.4. RELATION TO CONTINUOUS MODELS 89
with e
β
i
:= (cos β
i
, sin β
i
)
⊤
reveals nonnegative “directional diﬀusivities” α
0
, α
k
,
α
2m
along the stencil orientations β
0
, β
k
, β
2m
. This can be done by proving the
following properties:
(a) Let ψ ∈ I
k
and D =
a
b
b
c
. Then a nonnegative splitting of type (3.25) is
possible if
min
a −b cot β
k
, c −b tan β
k
≥ 0. (3.26)
(b) Inequality (3.26) is satisﬁed for
λ
1
λ
2
≤ min
cot(ρ
k
−β
k
) tanρ
k
, cot(β
k
−η
k
) cot η
k
=: κ
k,m
(3.27)
with
ρ
k
:=
θ
k
([k[ ∈ ¦1, ..., 2m−2¦),
1
2
(θ
k
+β
k
) ([k[ = 2m−1),
η
k
:=
1
2
β
k
([k[ = 1),
θ
k−1
([k[ ∈ ¦2, ..., 2m−1¦).
(c) lim
m→∞
min
i∈Jm
κ
i,m
= ∞.
Once these assertions are proved a nonnegative secondorder discretization of (3.25)
arises in a natural way, as we shall see at the end of this chapter. So let us now
verify (a)–(c).
(a) In order to use subsequent indices, let ϕ
0
:= 0, ϕ
1
:= β
k
where ψ ∈ I
k
, and
ϕ
2
:=
π
2
. Furthermore, let γ
0
:= α
0
, γ
1
:= α
k
, and γ
2
:= α
2m
. Then (3.25)
requires that
div
a b
b c
∇u
=
2
¸
i=0
∂
∂e
ϕ
i
γ
i
∂u
∂e
ϕ
i
=
∂
∂x
2
¸
i=0
cos ϕ
i
γ
i
(u
x
cos ϕ
i
+u
y
sin ϕ
i
)
+
∂
∂y
2
¸
i=0
sin ϕ
i
γ
i
(u
x
cos ϕ
i
+u
y
sin ϕ
i
)
= div
¸
¸
¸
¸
¸
¸
2
¸
i=0
γ
i
cos
2
ϕ
i
2
¸
i=0
γ
i
sin ϕ
i
cos ϕ
i
2
¸
i=0
γ
i
sin ϕ
i
cos ϕ
i
2
¸
i=0
γ
i
sin
2
ϕ
i
¸
∇u
¸
.
By comparing the coeﬃcients and using the deﬁnition of ϕ
0
, ϕ
1
and ϕ
2
we
obtain the linear system
¸
¸
1 cos
2
β
k
0
0 sin β
k
cos β
k
0
0 sin
2
β
k
1
¸
¸
¸
γ
0
γ
1
γ
2
¸
=
¸
¸
a
b
c
¸
90 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
which has the unique solution
γ
0
= a −b cot β
k
, (3.28)
γ
1
=
b
sin β
k
cos β
k
, (3.29)
γ
2
= c −b tanβ
k
. (3.30)
From the structure of the eigenvalues and eigenvectors of D it is easily seen
that
b = (λ
1
−λ
2
) sin ψ cos ψ.
Now, λ
1
−λ
2
≥ 0, and since ψ, β
k
∈ I
k
we conclude that ψ and β
k
belong to
the same quadrant. Thus, γ
1
is always nonnegative. In order to satisfy the
nonnegativity of γ
0
and γ
2
we need that
min
a −b cot β
k
, c −b tan β
k
≥ 0.
(b) Let
λ
1
λ
2
≤ κ
k,m
and consider the case 0 < β
k
<
π
2
. By deﬁning
B(ϕ) := cos
2
ϕ −sin ϕcos ϕcot β
k
,
C(ϕ) := sin
2
ϕ + sin ϕcos ϕcot β
k
we get
λ
1
λ
2
≤ cot(ρ
k
−β
k
) tanρ
k
= −
C(ρ
k
)
B(ρ
k
)
= min
ϕ∈(β
k
,θ
k
)
−
C(ϕ)
B(ϕ)
.
Since B(ϕ) < 0 on (β
k
,
π
2
) we have
λ
1
B(ϕ) +λ
2
C(ϕ) ≥ 0 ∀ ϕ ∈ (β
k
, θ
k
). (3.31)
Because of
B(ϕ) ≥ 0 ∀ ϕ ∈ [−
π
2
, β
k
],
C(ϕ) ≥ 0 ∀ ϕ ∈ [0,
π
2
],
and the continuity of B(ϕ) and C(ϕ) we may extend (3.31) to the entire
interval I
k
= (θ
k−1
, θ
k
]. In particular, since ψ ∈ I
k
, we have
0 ≤ λ
1
B(ψ) +λ
2
C(ψ)
= (λ
1
cos
2
ψ +λ
2
sin
2
ψ) −(λ
1
−λ
2
) sin ψ cos ψ cot β
k
.
By the representation
a b
b c
=
cos ψ −sin ψ
sin ψ cos ψ
λ
1
0
0 λ
2
cos ψ sin ψ
−sin ψ cos ψ
=
λ
1
cos
2
ψ +λ
2
sin
2
ψ (λ
1
−λ
2
) sin ψ cos ψ
(λ
1
−λ
2
) sin ψ cos ψ λ
1
sin
2
ψ +λ
2
cos
2
ψ
3.4. RELATION TO CONTINUOUS MODELS 91
we recognize that this is just the desired condition
a −b cot β
k
≥ 0. (3.32)
For the case −
π
2
< β
k
< 0 a similar reasoning can be applied leading also to
(3.32).
In an analogous way one veriﬁes that
λ
1
λ
2
≤ cot(β
k
−η
k
) cot η
k
=⇒ c −b tan β
k
≥ 0.
(c) Let us ﬁrst consider the case 1 ≤ i ≤ 2m−2. Then, ρ
i
= θ
i
, and the deﬁnition
of θ
i
implies that
cot ρ
i
−tan ρ
i
= cot β
i
−tanβ
i+1
.
Solving for cot ρ
i
and tan ρ
i
, respectively, yields
cot ρ
i
=
1
2
cot β
i
−tan β
i+1
+
(cot β
i
−tanβ
i+1
)
2
+ 4
,
tan ρ
i
=
1
2
−cot β
i
+ tanβ
i+1
+
(cot β
i
−tan β
i+1
)
2
+ 4
.
By means of these results we obtain
cot(ρ
i
−β
i
) tan ρ
i
=
cot β
i
+ tanρ
i
cot β
i
−cot ρ
i
= 1 +
2
(cot β
i
+tan β
i+1
)
2
(cot β
i
−tan β
i+1
)
2
+4
−1
.
Let us now assume that 1 ≤ i ≤ m−1. Then we have
tanβ
i+1
=
(i + 1) h
2
mh
1
,
cot β
i
=
mh
1
ih
2
.
This gives
(cot β
i
+ tanβ
i+1
)
2
(cot β
i
−tan β
i+1
)
2
+ 4
=
1
1 −
4m
2
i
m
2
h
1
h
2
+i(i+1)
h
2
h
1
2
=:
1
1 −g
m
(i)
=: f
m
(i).
For m >
1
2
h
2
h
1
the function g
m
(x) is bounded and attains its global maximum
in
x
m
:= −
1
6
+
1
6
1 + 12 m
2
h
2
1
h
2
2
.
92 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
Thus, for 1 ≤ i ≤ m−1,
g
m
(i) ≤ g
m
(x
m
) → 0
+
for m →∞,
which yields
f
m
(i) ≤
1
1 −g
m
(x
m
)
→ 1
+
for m →∞.
This gives
min
1≤i≤m−1
cot(ρ
i
−β
i
) tan ρ
i
≥ 1 +
2
f
m
(x
m
) −1
→ ∞ for m →∞.
For m ≤ i ≤ 2m−2 similar calculations show that by means of
tan β
i+1
=
mh
2
(2m−i−1) h
1
,
cot β
i
=
(2m−i) h
1
mh
2
one obtains
min
m≤i≤2m−2
cot(ρ
i
−β
i
) tan ρ
i
→ ∞ for m →∞.
For i = 2m−1 we have
cot
ρ
2m−1
−β
2m−1
tanρ
2m−1
= cot
π
4
−
β
2m−1
2
tan
π
4
+
β
2m−1
2
= tan
2
π
4
+
β
2m−1
2
→ ∞ for m →∞.
It is not hard to verify that for −2m + 1 ≤ i ≤ −1 the preceding results
carry over. Hence,
lim
m→∞
min
i∈Jm
cot(ρ
i
−β
i
) tanρ
i
= ∞. (3.33)
Now, in a similar way as above, one establishes that
lim
m→∞
min
i∈Jm
cot(β
i
−η
i
) cot η
i
= ∞. (3.34)
From (3.33) and (3.34) we ﬁnally end up with the assertion
lim
m→∞
min
i∈Jm
κ
i,m
= ∞.
2
3.4. RELATION TO CONTINUOUS MODELS 93
Remarks:
(a) We observe that the preceding existence proof is constructive. Moreover,
only three directions are suﬃcient to guarantee a nonnegative directional
splitting. Thus, unless m is very small, most of the stencil coeﬃcients can be
set to zero.
(b) Especially for large m, a (2m+1)(2m+1)stencil reveals much more directions
than those 4m that are induced by the 8m “boundary pixels”. Therefore,
even if we use only 3 directions, we may expect to ﬁnd stricter estimates
than those given in the proof. These estimates might be improved further by
admitting more than 3 directions.
(c) For a speciﬁed diﬀusion tensor function D it is possible to give apriori es
timates for the required stencil size: using the extremum principle it is not
hard to show that
[∇u
σ
(x, t)[ = [(∇K
σ
∗u)(x, t)[ ≤
4 f
L
∞
(Ω)
√
2π σ
on
¯
Ω (0, ∞),
where the notations from Chapter 2 have been used. Thanks to the uniform
positive deﬁniteness of D there exists an upper limit for the spectral condition
number of D. This condition limit can be used to ﬁx a suitable stencil size.
(d) The existence of a nonnegative directional splitting distinguishes the ﬁlter
class (P
c
) from morphological anisotropic equations such as mean curvature
motion. In this case it has been proved that it is impossible to ﬁnd a nonneg
ative directional splitting on a ﬁnite stencil [13]. As a remedy, Crandall and
Lions [104] propose to study a convergent sequence of regularizations which
can be approximated on a ﬁnite stencil.
Let us now illustrate the ideas in the proof of Theorem 6 by applying them
to a practical example: We want to ﬁnd a nonnegative spatial discretization of
div (D∇u) on a (33)stencil, where
D =
a
b
b
c
and a, b and c may be functions of J
ρ
(∇u
σ
).
Since m = 1 we have a partition of (−
π
2
,
π
2
] into 4m−2 = 2 subintervals:
(−
π
2
,
π
2
] = (−
π
2
, 0] ∪ (0,
π
2
] =: I
−1
∪ I
1
.
94 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
I
−1
and I
1
belong to the grid angles
β
−1
= arctan
−
h
2
h
1
,
β
1
= arctan
h
2
h
1
=: β.
First we focus on the case ψ ∈ I
1
where (cos ψ, sin ψ) denotes the eigenvector to
the larger eigenvalue λ
1
of D. With the notations from the proof of Theorem 6 we
obtain
θ
1
=
π
2
,
ρ
1
=
θ
1
+β
1
2
=
π
4
+
β
2
,
η
1
=
β
2
.
Therefore, we get
cot(ρ
1
−β
1
) tan ρ
1
= cot
π
4
−
β
2
tan
π
4
+
β
2
=
1 + sin β
1 −sin β
,
cot(β
1
−η
1
) cot η
1
= cot
2
β
2
=
1 + cos β
1 −cos β
,
which restricts the upper condition number for a nonnegative discretization with
ψ ∈ I
1
to
κ
1,1
:= min
1 + sin β
1 −sin β
,
1 + cos β
1 −cos β
. (3.35)
Thanks to the symmetry we obtain the same condition restriction for ψ ∈ I
−1
.
These bounds on the condition number attain their maximal value for h
1
= h
2
. In
this case β =
π
4
gives
κ
1,1
= κ
−1,1
=
1 +
1
2
√
2
1 −
1
2
√
2
= 3 + 2
√
2 ≈ 5.8284. (3.36)
By virtue of (3.28)–(3.30) we obtain as expressions for the directional diﬀusivities
α
−1
=
[b[ −b
h
2
1
+h
2
2
2h
1
h
2
,
α
0
= a −[b[
h
1
h
2
,
α
1
=
[b[ +b
h
2
1
+h
2
2
2h
1
h
2
,
α
2
= c −[b[
h
2
h
1
.
3.4. RELATION TO CONTINUOUS MODELS 95
This induces in a natural way the following secondorder discretization for div (D∇u):
b
i−1,j+1
−b
i−1,j+1
4h
1
h
2
+
b
i,j
−b
i,j
4h
1
h
2
c
i,j+1
+c
i,j
2h
2
2
−
b
i,j+1
+b
i,j

2h
1
h
2
b
i+1,j+1
+b
i+1,j+1
4h
1
h
2
+
b
i,j
+b
i,j
4h
1
h
2
a
i−1,j
+a
i,j
2h
2
1
−
b
i−1,j
+b
i,j

2h
1
h
2
−
a
i−1,j
+2a
i,j
+a
i+1,j
2h
2
1
−
b
i−1,j+1
−b
i−1,j+1
+b
i+1,j+1
+b
i+1,j+1
4h
1
h
2
−
b
i−1,j−1
+b
i−1,j−1
+b
i+1,j−1
−b
i+1,j−1
4h
1
h
2
+
b
i−1,j
+b
i+1,j
+b
i,j−1
+b
i,j+1
+2b
i,j

2h
1
h
2
−
c
i,j−1
+2c
i,j
+c
i,j+1
2h
2
2
a
i+1,j
+a
i,j
2h
2
1
−
b
i+1,j
+b
i,j

2h
1
h
2
b
i−1,j−1
+b
i−1,j−1
4h
1
h
2
+
b
i,j
+b
i,j
4h
1
h
2
c
i,j−1
+c
i,j
2h
2
2
−
b
i,j−1
+b
i,j

2h
1
h
2
b
i+1,j−1
−b
i+1,j−1
4h
1
h
2
+
b
i,j
−b
i,j
4h
1
h
2
All nonvanishing entries of the pth row of A(u) are represented in this stencil,
where p(i, j) is the index of some inner pixel (i, j). Thus, for instance, the upper
left stencil entry gives the element (p(i, j), p(i−1, j+1)) of A(u). The other nota
tions should be clear from the context as well, e.g. b
i,j
denotes a ﬁnite diﬀerence
approximation of b(J
ρ
(∇u
σ
)) at some grid point (x
i
, y
j
).
The problem of ﬁnding nonnegative diﬀerence approximations to elliptic expres
sions with mixed derivatives has a long history; see e.g. [294, 120, 170]. Usually it
is studied for the expression
a(x, y) ∂
xx
u + 2b(x, y) ∂
xy
u +c(x, y) ∂
yy
u.
The approach presented here extends these results to
∂
x
a(x, y) ∂
x
u
+∂
x
b(x, y) ∂
y
u
+∂
y
b(x, y) ∂
x
u
+∂
y
c(x, y) ∂
y
u
and establishes the relation between the condition number of
a
b
b
c
and the non
negativity of the diﬀerence operator. Recently Kocan described an interesting al
ternative to obtain upper bounds for the stencil size as a function of the condition
number [238]. His derivation is based on the diophantine problem of approximating
irrationals by rational numbers.
96 CHAPTER 3. SEMIDISCRETE DIFFUSION FILTERING
Chapter 4
Discrete diﬀusion ﬁltering
This chapter presents a discrete class of diﬀusion processes for which one can estab
lish similar properties as in the semidiscrete case concerning existence, uniqueness,
continuous dependence of the solution on the initial image, maximumminimum
principle, average grey level invariance, Lyapunov sequences and convergence to a
constant steadystate. We shall see that this class comprises αsemiimplicit dis
cretizations of the semidiscrete ﬁlter class (P
s
) as well as certain variants of them
which are based on an additive operator splitting.
4.1 The general model
As in Chapter 3 we regard a discrete image as a vector f ∈IR
N
, N≥2, and denote
the index set ¦1, ..., N¦ by J. We consider the following discrete ﬁlter class (P
d
):
Let f ∈IR
N
. Calculate a sequence (u
(k)
)
k∈IN
0
of processed versions
of f by means of
u
(0)
= f,
u
(k+1)
= Q(u
(k)
) u
(k)
, ∀ k ∈ IN
0
,
where Q = (q
ij
) has the following properties:
(D1) continuity in its argument: Q ∈ C(IR
N
, IR
N×N
),
(D2) symmetry: q
ij
(v) = q
ji
(v) ∀ i, j ∈ J, ∀ v ∈ IR
N
,
(D3) unit row sum:
¸
j∈J
q
ij
(v) = 1 ∀ i ∈ J, ∀ v ∈ IR
N
,
(D4) nonnegativity: q
ij
(v) ≥ 0 ∀ i, j ∈ J, ∀ v ∈ IR
N
,
(D5) irreducibility for all v ∈ IR
N
,
(D6) positive diagonal: q
ii
(v) > 0 ∀ i ∈ J, ∀ v ∈ IR
N
.
(P
d
)
97
98 CHAPTER 4. DISCRETE DIFFUSION FILTERING
Remarks:
(a) Although the basic idea behind scalespaces is to have a continuous scale
parameter, it is evident that fully discrete results are of importance since, in
practice, scalespace evolutions are evaluated exclusively at a ﬁnite number
of scales.
(b) The requirements (D1)–(D5) have a similar meaning as their semidiscrete
counterparts (S1)–(S5). Indeed, (D1) immediately gives wellposedness re
sults, while the proof of the extremum principle requires (D3) and (D4),
and average grey value invariance is based on (D2) and (D3). The existence
of Lyapunov sequences is a consequence of (D2)–(D4), strict Lyapunov se
quences need (D5) and (D6) in addition to (D2)–(D4), and the convergence
to a constant steadystate utilizes (D2)–(D5).
(c) Nonnegative matrices Q = (q
ij
) ∈IR
N×N
satisfying
¸
j∈J
q
ij
=1 for all i ∈J
are also called stochastic matrices. Moreover, if Q is stochastic and
¸
i∈J
q
ij
=
1 for all j ∈J, then Q is doubly stochastic. This indicates that our discrete
diﬀusion processes are related to the theory of Markov chains [370, 223].
4.2 Theoretical results
It is obvious that for a ﬁxed ﬁlter belonging to the class (P
d
) every initial image
f ∈ IR
N
generates a unique sequence (u
(k)
)
k∈IN
0
. Moreover, by means of (D1) we
know that, for every ﬁnite k, u
(k)
depends continuously on f. Therefore, let us now
prove a maximum–minimum principle.
Proposition 3 (Extremum principle).
Let f ∈IR
N
and let (u
(k)
)
k∈IN
0
be the sequence of ﬁltered images according to (P
d
).
Then,
a ≤ u
(k)
i
≤ b ∀ i ∈ J, ∀ k ∈ IN
0
, (4.1)
where
a := min
j∈J
f
j
, (4.2)
b := max
j∈J
f
j
. (4.3)
Proof:
The maximum–minimum principle follows directly from the fact that, for all i ∈J
and k∈IN
0
, the following inequalities hold:
(i) u
(k+1)
i
=
¸
j∈J
q
ij
(u
(k)
)u
(k)
j
(D4)
≤ max
m∈J
u
(k)
m
¸
j∈J
q
ij
(u
(k)
)
(D3)
= max
m∈J
u
(k)
m
.
4.3. SCALESPACE PROPERTIES 99
(ii) u
(k+1)
i
=
¸
j∈J
q
ij
(u
(k)
)u
(k)
j
(D4)
≥ min
m∈J
u
(k)
m
¸
j∈J
q
ij
(u
(k)
)
(D3)
= min
m∈J
u
(k)
m
.
2
4.3 Scalespace properties
All statements from Chapter 3 with respect to invariances are valid in the discrete
framework as well. Below we focus on proving average grey level invariance.
Proposition 4 (Conservation of average grey value).
The average grey level
µ :=
1
N
¸
j∈J
f
j
(4.4)
is not aﬀected by the discrete diﬀusion ﬁlter:
1
N
¸
j∈J
u
(k)
j
= µ ∀ k ∈ IN
0
. (4.5)
Proof:
By virtue of (D2) and (D3) we have
¸
i∈J
q
ij
(u
(k)
)=1 for all j ∈J and k∈IN
0
. This
socalled redistribution property [164] ensures that, for all k ∈ IN
0
,
¸
i∈J
u
(k+1)
i
=
¸
i∈J
¸
j∈J
q
ij
(u
(k)
)u
(k)
j
=
¸
j∈J
¸
i∈J
q
ij
(u
(k)
)
u
(k)
j
=
¸
j∈J
u
(k)
j
,
which proves the proposition. 2
As one might expect, the class (P
d
) allows an interpretation as a transformation
which is smoothing in terms of Lyapunov sequences. These functions ensure that
u
(k)
converges to a constant image as k → ∞. However, we need less regularity
than in the semidiscrete case: The convex function r, which generates the Lyapunov
sequences, needs only to be continuous, but no more diﬀerentiable.
Theorem 7 (Lyapunov sequences and behaviour for k →∞).
Assume that (u
(k)
)
k∈IN
0
satisﬁes the requirements of (P
d
), let a, b, and µ be deﬁned
as in (4.2), (4.3), and (4.4), respectively, and let c := (µ, µ, ..., µ)
⊤
∈ IR
N
.
Then the following properties are fulﬁlled:
(a) (Lyapunov sequences)
For all convex r ∈ C[a, b] the sequence
V
(k)
:= Φ(u
(k)
) :=
¸
i∈J
r(u
(k)
i
), k ∈ IN
0
is a Lyapunov sequence:
100 CHAPTER 4. DISCRETE DIFFUSION FILTERING
(i) Φ(u
(k)
) ≥ Φ(c) ∀ k ∈ IN
0
(ii) V
(k+1)
−V
(k)
≤ 0 ∀ k ∈ IN
0
Moreover, if r is strictly convex, then V
(k)
= Φ(u
(k)
) is a strict Lyapunov
sequence:
(iii) Φ(u
(k)
) = Φ(c) ⇐⇒ u
(k)
= c
(iv) V
(k+1)
−V
(k)
= 0 ⇐⇒ u
(k)
= c
(b) (Convergence)
lim
k→∞
u
(k)
= c.
Proof:
(a) (i) Average grey level invariance and the convexity of r give
Φ(c) =
N
¸
i=1
r
¸
N
¸
j=1
1
N
u
(k)
j
¸
≤
N
¸
i=1
¸
1
N
N
¸
j=1
r(u
(k)
j
)
¸
=
N
¸
j=1
r(u
(k)
j
)
= Φ(u
(k)
). (4.6)
(ii) For i, j ∈ J we deﬁne
a
ij
(u
(k)
) :=
q
ij
(u
(k)
) −1 (i = j)
q
ij
(u
(k)
) (i = j).
(4.7)
Using the convexity of r, the preceding deﬁnition, and the prerequisites
(D2) and (D3) we obtain
V
(k+1)
−V
(k)
=
N
¸
i=1
¸
r
¸
N
¸
j=1
q
ij
(u
(k)
) u
(k)
j
¸
−r(u
(k)
i
)
¸
conv.
≤
N
¸
i=1
¸
N
¸
j=1
q
ij
(u
(k)
) r(u
(k)
j
) −r(u
(k)
i
)
¸
(4.7)
=
N
¸
i=1
N
¸
j=1
a
ij
(u
(k)
) r(u
(k)
j
)
(D3)
=
N
¸
i=1
N
¸
j=1
a
ij
(u
(k)
)
r(u
(k)
j
) −r(u
(k)
i
)
4.3. SCALESPACE PROPERTIES 101
=
N
¸
i=1
N−i
¸
m=1
a
i+m,i
(u
(k)
)
r(u
(k)
i
) −r(u
(k)
i+m
)
+
N
¸
i=1
N−i
¸
m=1
a
i,i+m
(u
(k)
)
r(u
(k)
i+m
) −r(u
(k)
i
)
(D2)
= 0. (4.8)
(iii) This part of the proof can be shown in exactly the same manner as in
the semidiscrete case (Chapter 3, Theorem 5): Equality in the estimate
(4.6) holds due to the strict convexity of r if and only if u
(k)
= c.
(iv) In order to verify the ﬁrst implication, let us start with a proof that
V
(k+1)
= V
(k)
implies u
(k)
1
= ... = u
(k)
N
. To this end, assume that u
(k)
is
not constant:
u
(k)
i
0
:= min
i∈J
u
(k)
i
< max
j∈J
u
(k)
j
=: u
(k)
j
0
.
Then, by the irreducibility of Q(u
(k)
), we ﬁnd l
0
, ..., l
r
∈ J with l
0
= i
0
,
l
r
= j
0
and q
lpl
p+1
= 0 for p = 0, ..., r −1. Hence, there exists some
p
0
∈ ¦0, ..., r −1¦ such that n := l
p
0
, m := l
p
0
+1
, q
nm
(u
(k)
) = 0, and
u
(k)
m
= u
(k)
n
. Moreover, the nonnegativity of Q(u
(k)
) gives q
nm
(u
(k)
) > 0,
and by (D6) we have q
nn
(u
(k)
) > 0. Together with the strict convexity
of r these properties lead to
r
¸
N
¸
j=1
q
nj
(u
(k)
) u
(k)
j
¸
= r
¸
¸
N
¸
j=1
j=n,m
q
nj
(u
(k)
) u
(k)
j
+q
nn
(u
(k)
) u
(k)
n
+q
nm
(u
(k)
) u
(k)
m
¸
<
N
¸
j=1
j=n,m
q
nj
(u
(k)
) r(u
(k)
j
) + q
nn
(u
(k)
) r(u
(k)
n
) +q
nm
(u
(k)
) r(u
(k)
m
)
=
N
¸
j=1
q
nj
(u
(k)
) r(u
(k)
j
).
If we combine this with the results in (4.8), we obtain
V
(k+1)
−V
(k)
=
N
¸
i=1
i=n
¸
r
¸
N
¸
j=1
q
ij
(u
(k)
) u
(k)
j
¸
−r(u
(k)
i
)
¸
+ r
¸
N
¸
j=1
q
nj
(u
(k)
) u
(k)
j
¸
−r(u
(k)
n
)
102 CHAPTER 4. DISCRETE DIFFUSION FILTERING
<
N
¸
i=1
¸
N
¸
j=1
q
ij
(u
(k)
) r(u
(k)
j
) −r(u
(k)
i
)
¸
(4.8)
= 0.
This establishes that V
(k+1)
= V
(k)
implies u
(k)
1
= ... = u
(k)
N
. Then, by
virtue of the grey value invariance, we conclude that u
(k)
= c.
Conversely, let u
(k)
= c. By means of prerequisite (D3) we obtain
V
(k+1)
−V
(k)
=
N
¸
i=1
r
¸
N
¸
j=1
q
ij
(u
(k)
)µ
¸
−
N
¸
i=1
r(µ) = 0.
(c) In order to prove convergence to a constant steadystate, we can argue exactly
in the same way as in the semidiscrete case if we replace Lyapunov functions
by Lyapunov sequences and integrals by sums. See Chapter 3, Theorem 5 for
more details. 2
In analogy to the semidiscrete case the preceding theorem comprises many
Lyapunov functions which demonstrate the informationreducing qualities of our
ﬁlter class. Choosing the convex functions r(s) := [s[
p
, r(s) := (s −µ)
2n
and
r(s) := s ln s, we immediately obtain the following corollary.
Corollary 3 (Special Lyapunov sequences).
Let (u
(k)
)
k∈IN
0
be a diﬀusion sequence according to (P
d
), and let a and µ be deﬁned
as in (4.2) and (4.4). Then the following functions are decreasing in k:
(a) u
(k)

p
for all p ≥ 1.
(b) M
2n
[u
(k)
] :=
1
N
N
¸
j=1
(u
(k)
j
−µ)
2n
for all n ∈ IN.
(c) H[u
(k)
] :=
N
¸
j=1
u
(k)
j
ln(u
(k)
j
), if a > 0.
An interpretation of these results in terms of decreasing energy, decreasing
central moments and increasing entropy is evident.
4.4 Relation to semidiscrete models
4.4.1 Semiimplicit schemes
Let us now investigate in which sense our discrete ﬁlter class covers in a natural
way time discretizations of semidiscrete ﬁlters. To this end, we regard u
(k)
as an
4.4. RELATION TO SEMIDISCRETE MODELS 103
approximation of the solution u of (P
s
) at time t = kτ, where τ denotes the time
step size. We consider a ﬁnite diﬀerence scheme with two time levels where the
operator A – which depends nonlinearly on u – is evaluated in an explicit way,
while the linear remainder is discretized in an αimplicit manner. Such schemes
are called αsemiimplicit. They reveal the advantage that the linear implicit part
ensures good stability properties, while the explicit evaluation of the nonlinear
terms avoids the necessity to solve nonlinear systems of equations. The theorem
below states that this class of schemes is covered by the discrete framework, for
which we have established scalespace results.
Theorem 8 (Scalespace interpretation for αsemiimplicit schemes).
Let α ∈ [0, 1], τ > 0, and let A = (a
ij
) : IR
N
→ IR
N×N
satisfy the requirements
(S1)–(S5) of section 3.1. Then the αsemiimplicit scheme
u
(k+1)
−u
(k)
τ
= A(u
(k)
)
αu
(k+1)
+ (1−α)u
(k)
(4.9)
fulﬁls the prerequisites (D1)–(D6) for discrete diﬀusion models provided that
τ ≤
1
(1−α) max
i∈J
[a
ii
(u
(k)
)[
(4.10)
for α ∈ (0, 1). In the explicit case (α=0) the properties (D1)–(D6) hold for
τ <
1
max
i∈J
[a
ii
(u
(k)
)[
, (4.11)
and the semiimplicit case (α=1) satisﬁes (D1)–(D6) unconditionally.
Proof: Let
B(u
(k)
) := (b
ij
(u
(k)
)) := I −ατA(u
(k)
),
C(u
(k)
) := (c
ij
(u
(k)
)) := I + (1−α)τA(u
(k)
),
where I ∈IR
N
denotes the unit matrix. Since (4.9) can be written as
B(u
(k)
) u
(k+1)
= C(u
(k)
) u
(k)
we ﬁrst have to show that B(u
(k)
) is invertible for all u
(k)
∈ IR
N
. Henceforth, the
argument u
(k)
is suppressed frequently since the considerations below are valid for
all u
(k)
∈ IR
N
.
If α=0, then B=I and hence invertible. Now assume that α > 0. Then B is
strictly diagonally dominant, since
b
ii
= 1 −ατa
ii
(S3)
= 1 +ατ
¸
j∈J
j=i
a
ij
> ατ
¸
j∈J
j=i
a
ij
(S4)
=
¸
j∈J
j=i
[b
ij
[ ∀ i ∈ J.
104 CHAPTER 4. DISCRETE DIFFUSION FILTERING
This also shows that b
ii
> 0 for all i ∈ J, and by the structure of the oﬀdiagonal
elements of B we observe that the irreducibility of A implies the irreducibility of
B. Thanks to the fact that B is irreducibly diagonally dominant, b
ij
≤ 0 for all
i = j, and b
ii
> 0 for all i ∈ J, we know from [407, p. 85] that B
−1
=: H =: (h
ij
)
exists and h
ij
> 0 for all i, j ∈ J. Thus, Q := (q
ij
) := B
−1
C exists and by (S1) it
follows that Q ∈ C(IR
N
, IR
N×N
). This proves (D1).
The requirement (D2) is not hard to satisfy: Since B
−1
and C are symmetric
and reveal the same set of eigenvectors – namely those of A – it follows that
Q = B
−1
C is symmetric as well.
Let us now verify (D3). By means of (S3) we obtain
¸
j∈J
b
ij
= 1 =
¸
j∈J
c
ij
∀ i ∈ J. (4.12)
Let v := (1, ..., 1)
⊤
∈ IR
N
. Then (4.12) is equivalent to
Bv = v = Cv, (4.13)
and the invertibility of B gives
v = B
−1
v = Hv. (4.14)
Therefore, from
Qv = HCv
(4.13)
= Hv
(4.14)
= v
we conclude that
¸
j∈J
q
ij
= 1 for all i ∈ J. This proves (D3).
In order to show that (D4) is fulﬁlled, we ﬁrst check the nonnegativity of C.
For i = j we have
c
ij
= (1−α)τa
ij
(S4)
≥ 0.
The diagonal entries yield
c
ii
= 1 + (1−α)τa
ii
.
If α = 1 we have c
ii
= 1 for all i ∈ J. For 0 ≤ α < 1, however, nonnegativity of C
is not automatically guaranteed: Using (S3)–(S5) we obtain
a
ii
(S3)
= −
¸
j∈J
j=i
a
ij
(S4),(S5)
< 0 ∀ i ∈ J. (4.15)
Hence, C(u
(k)
) is nonnegative if
τ ≤
1
(1−α) max
i∈J
[a
ii
(u
(k)
)[
=: τ
α
(u
(k)
).
4.4. RELATION TO SEMIDISCRETE MODELS 105
Since H is nonnegative, we know that the nonnegativity of C implies the nonneg
ativity of Q=HC.
Now we want to prove (D5). If α = 1, then C = I, and by the positivity of H
we have q
ij
> 0 for all i, j ∈ J. Thus, Q is irreducible.
Next let us consider the case 0 <α <1 and τ ≤τ
α
(u
(k)
). Then we know that
C is nonnegative. Using this information, the positivity of H, the symmetry of C,
and (4.12) we obtain
q
ij
=
¸
k∈J
h
ik
c
kj
≥ min
k∈J
h
ik
. .. .
>0
¸
k∈J
c
kj
. .. .
=1
> 0 ∀ i, j ∈ J,
which establishes the irreducibility of Q.
Finally, for α = 0, we have Q = C. For i, j ∈ J with i = j we know that
a
ij
(u
(k)
) > 0 implies c
ij
(u
(k)
) > 0. Now for
τ <
1
max
i∈J
[a
ii
(u
(k)
)[
it follows that c
ii
(u
(k)
) > 0 for all i ∈ J and, thus, the irreducibility of A(u
(k)
)
carries over to Q(u
(k)
).
In all the abovementioned cases the time step size restrictions for ensuring irre
ducibility imply that all diagonal elements of Q(u
(k)
) are positive. This establishes
(D6). 2
Remarks:
(a) We have seen that the discrete ﬁlter class (P
d
) – although at ﬁrst glance
looking like a pure explicit discretization – covers the αsemiimplicit case as
well. Explicit twolevel schemes are comprised by the choice α=0. Equation
(4.11) shows that they reveal the most prohibitive time step size restrictions.
(b) The conditions (4.10) and (4.11) can be satisﬁed by means of an apriori
estimate. Since the semiimplicit scheme fulﬁls (D1)–(D6) we know by The
orem 3 that the solution obeys an extremum principle. This means that u
(k)
belongs to the compact set ¦v ∈ IR
N
v
∞
≤ f
∞
¦ for all k ∈ IN
0
. By
A∈C(IR
N
, IR
N×N
) it follows that
K
f
:= max
[a
ii
(v)[
i ∈ J, v ∈ IR
N
, v
∞
≤ f
∞
¸
exists, and (4.15) shows that K
f
> 0. Thus, choosing
τ ≤
1
(1−α) K
f
106 CHAPTER 4. DISCRETE DIFFUSION FILTERING
ensures that (4.10) is always satisﬁed, and
τ <
1
K
f
guarantees that (4.11) holds.
(c) If α>0, a large linear system of equations has to be solved. Its system matrix
is symmetric, diagonally dominant, and positive deﬁnite. Usually, it is also
sparse: For instance, if it results from a ﬁnite diﬀerence discretization on a
(2p+1)(2p+1)stencil it contains at most 4p
2
+4p+1 nonvanishing entries
per row. One should not expect, however, that in the ith row these entries
can be found within the positions [i, i−2p
2
−2p] to [i, i+2p
2
+2p]. In general,
the matrix reveals a much larger bandwidth.
Applying standard direct algorithms such as Gaussian elimination would
destroy the zeros within the band and would lead to an immense storage and
computation eﬀort. Modiﬁcations in order to reduce these problems [122] are
quite diﬃcult to implement.
Iterative algorithms appear to be better suited. Classical methods such as
Gauß–Seidel or SOR [447] are easy to code, they do not need additional
storage, and their convergence can be guaranteed for the special structure of
A. Unfortunately, they converge rather slowly. Faster iterative methods such
as preconditioned conjugate gradient algorithms [348] need signiﬁcantly more
storage, which can become prohibitive for large images. A typical problem of
iterative methods is that their convergence slows down for larger τ, since this
increases the condition number of the system matrix. Multigrid methods [59,
179] are one possibility to circumvent these diﬃculties; another possibility
will be studied in Section 4.4.2.
(d) For α = 1 we obtain semiimplicit schemes which do not suﬀer from time
step size restrictions. In spite of the fact that the nonlinearity is discretized
in an explicit way they are absolutely stable in the maximum norm, and
they inherit the scalespace properties from the semidiscrete setting regard
less of the step size. Compared to explicit schemes, this advantage usually
overcompensates for the additional eﬀort of resolving a linear system.
(e) By the explicit discretization of the nonlinear operator A it follows that all
schemes in the preceding theorem are of ﬁrst order in time. This should not
give rise to concern, since in image processing one is in general more inter
ested in maintaining qualitative properties such as maximum principles or
invariances rather than having an accurate approximation of the continu
ous equation. However, if one insists in secondorder schemes, one may for
4.4. RELATION TO SEMIDISCRETE MODELS 107
instance use the predictor–corrector approach by Douglas and Jones [121]:
u
(k+1/2)
−u
(k)
τ/2
= A(u
(k)
) u
(k+1/2)
,
u
(k+1)
−u
(k)
τ
= A(u
(k+1/2)
)
1
2
u
(k+1)
+
1
2
u
(k)
.
This scheme satisﬁes the properties (D1)–(D6) for τ ≤ 2/K
f
.
(f) The assumptions (S1)–(S5) are suﬃcient conditions for the αsemiimplicit
scheme to fulﬁl (D1)–(D6), but they are not necessary. Nonnegativity of
Q(u
(k)
) may also be achieved using spatial discretizations where A(u
(k)
) has
negative oﬀdiagonal elements (see [55] for examples).
4.4.2 AOS schemes
We have seen that, for α > 0, the preceding αsemiimplicit schemes require to
solve a linear system with the system matrix (I − ατA(u
(k)
)). Since this can be
numerically expensive, it would be nice to have an eﬃcient alternative. Suppose
we know a splitting
A(u
(k)
) =
m
¸
l=1
A
l
(u
(k)
), (4.16)
such that the m linear systems with system matrices (I −mατA
l
(u
(k)
)), l = 1,...,m
can be solved more eﬃciently. Then it is advantageous to study instead of the α
semiimplicit scheme
u
(k+1)
=
I −ατ
m
¸
l=1
A
l
(u
(k)
)
−1
I + (1−α)τ
m
¸
l=1
A
l
(u
(k)
)
u
(k)
(4.17)
its additive operator splitting (AOS) variant [424]
u
(k+1)
=
1
m
m
¸
l=1
I −αmτA
l
(u
(k)
)
−1
I + (1−α)mτA
l
(u
(k)
)
u
(k)
. (4.18)
By means of a Taylor expansion one can verify that, although both schemes are not
identical, they have the same approximation order in space and time. Hence, from
a numerical viewpoint, they are both consistent approximations to the semidiscrete
ODE system from (P
s
).
The following theorem clariﬁes the conditions under which AOS schemes create
discrete scalespaces.
108 CHAPTER 4. DISCRETE DIFFUSION FILTERING
Theorem 9 (Scalespace interpretation for AOS schemes).
Let α ∈ (0, 1], τ >0, and let A
l
= (a
ijl
)
ij
: IR
N
→ IR
N×N
, l = 1,...,m satisfy the
requirements (S1)–(S4) of section 3.1. Moreover, assume that A(u) =
¸
m
l=1
A
l
(u)
is irreducible for all u ∈ IR
N
, and that for each A
l
there exists a permutation
matrix P
l
∈ IR
N×N
such that P
l
A
l
P
T
l
is block diagonal and irreducible within each
block. Then the following holds:
For α ∈ (0, 1), the AOS scheme (4.18) fulﬁls the prerequisites (D1)–(D6) of
discrete diﬀusion scalespaces provided that
τ <
1
(1−α) m max
i,l
[a
iil
(u
(k)
)[
. (4.19)
In the semiimplicit case (α = 1), the properties (D1)–(D6) are unconditionally
satisﬁed.
Proof: The reasoning is similar to the proof of Theorem 8. Let
B
l
(u
(k)
) := (b
ijl
(u
(k)
))
ij
:= I −αmτA
l
(u
(k)
),
C
l
(u
(k)
) := (c
ijl
(u
(k)
))
ij
:= I + (1−α)mτA
l
(u
(k)
).
B
l
is invertible because of its strict diagonal dominance:
b
iil
(S3)
= 1 +αmτ
¸
j∈J
j=i
a
ijl
> αmτ
¸
j∈J
j=i
a
ijl
(S4)
=
¸
j∈J
j=i
[b
ijl
[ ∀ i ∈ J.
Since A
l
(u) is continuous in u by virtue of (S1), it follows that
Q(u) :=
1
m
m
¸
l=1
B
−1
l
(u) C
l
(u)
is also continuous in u. This proves (D1).
The symmetry property (D2) of Q results directly from the fact that B
−1
l
and
C
l
are symmetric and share their eigenvectors with those of A
l
.
In the same way as in the proof of Theorem 8 one shows that B
−1
l
C
l
has only
unit row sums for all l. Thus, the row sums of Q are 1 as well, and (D3) is satisﬁed.
To verify (D4), we utilize that B
l
is strictly diagonally dominant, b
iil
> 0 for
all i, and b
ijl
≤ 0 for i = j. Under these circumstances it follows from [284, p. 192]
that H
l
:= B
−1
l
is nonnegative in all components. Thus, a suﬃcient condition for
proving (D4) is to ensure that C
l
is nonnegative for all l. For i = j we have
c
ijl
= (1−α)mτa
ijl
(S4)
≥ 0.
4.4. RELATION TO SEMIDISCRETE MODELS 109
The diagonal entries yield
c
iil
= 1 + (1−α)mτa
iil
.
If α = 1 we have c
ii
= 1 for all i ∈ J. For 0 < α < 1, however, nonnegativity of
C is not automatically guaranteed: Since A
l
satisﬁes (S3) and (S4), we know that
a
iil
≤ 0, for all i. Moreover, by (4.15) it follows that for every i there exists an l
with a
iil
< 0. Thus, requiring
τ <
1
(1−α) m max
i,l
[a
iil
(u
(k)
)[
=: τ
α
(u
(k)
)
guarantees that
c
iil
> 0 ∀ i ∈ J, ∀ l = 1, ..., m. (4.20)
Next we prove (D5), the irreducibility of Q. Suppose that a
i
0
j
0
= 0 for some
i
0
, j
0
∈ J. Then there exists an l
0
∈ ¦1, ..., m¦ such that a
i
0
j
0
l
0
= 0. Denoting B
−1
l
by H
l
= (h
ijl
)
ij
, we show now that a
i
0
j
0
l
0
= 0 implies h
i
0
j
0
l
0
> 0.
This can be seen as follows: There exist permutation matrices P
l
, l = 1, ..., m
such that P
l
B
l
P
T
l
is block diagonal. Each block is irreducible and strictly diagonally
dominant with a positive diagonal and nonpositive oﬀdiagonals. Thus, a theorem
by Varga [407, p. 85] ensures that the inverse of each block contains only positive
elements. As a consequence, a
i
0
j
0
l
0
= 0 implies h
i
0
j
0
l
0
> 0.
Together with (4.20) this yields
q
i
0
j
0
=
1
m
¸
¸
¸
(l,n)=(l
0
,j
0
)
h
i
0
nl
. .. .
≥0
c
nj
0
l
. .. .
≥0
+ h
i
0
j
0
l
0
. .. .
>0
c
j
0
j
0
l
0
. .. .
>0
¸
> 0.
Recapitulating, this means that, for τ < τ
α
(u
(k)
),
a
i
0
j
0
= 0 =⇒ q
i
0
j
0
> 0. (4.21)
Thus, the irreducibility of A carries over to Q, and (D5) is proved.
Moreover, (4.21) also proves (D6): By virtue of (4.15) we have a
ii
< 0 for all
i ∈ J. Therefore, Q must have a positive diagonal. 2
Remarks:
(a) In analogy to the unsplit αsemiimplicit schemes, the case α = 1 is especially
interesting, because no time step size restriction occurs. Again, it is also
possible to construct a predictor–corrector scheme of Douglas–Jones type
[121] within the AOS framework:
u
(k+1/2)
=
1
m
m
¸
l=1
I −
1
2
mτA
l
(u
(k)
)
−1
u
(k)
,
u
(k+1)
=
1
m
m
¸
l=1
I −
1
2
mτA
l
(u
(k+1/2)
)
−1
I +
1
2
mτA
l
(u
(k+1/2)
)
u
(k)
.
110 CHAPTER 4. DISCRETE DIFFUSION FILTERING
It satisﬁes (D1)–(D6) for τ < 2/K
f
, where K
f
is determined by the apriori
estimate
K
f
:= m max
[a
iil
(v)[
i ∈ J, l ∈ ¦1, ..., m¦, v ∈ IR
N
, v
∞
≤ f
∞
¸
.
However, the AOS–Douglas–Jones scheme is only ﬁrst order accurate in time.
(b) The fact that AOS schemes use an additive splitting ensures that all coordi
nate axes are treated in exactly the same manner. This is in contrast to the
various conventional splitting techniques from the literature [120, 277, 286,
354, 442]. They are multiplicative splittings. A typical representative is
u
(k+1)
=
m
¸
l=1
I −τA
l
(u
(k)
)
−1
u
(k)
.
Since in the general nonlinear case the split operators A
l
, l = 1,..., m do not
commute, the result of multiplicative splittings will depend on the order of
the operators. In practice, this means that these schemes produce diﬀerent
results if the image is rotated by 90 degrees. Moreover, most multiplicative
splittings lead to a nonsymmetric matrix Q(u
(k)
). This violates requirement
(D2) for discrete scalespaces.
(c) The result u
(k+1)
of an AOS scheme can be regarded as the average of m
ﬁlters of type
v
(k+1)
l
:=
I −αmτA
l
(u
(k)
)
−1
I +(1−α)mτA
l
(u
(k)
)
u
(k)
(l = 1, ..., m).
Since v
(k+1)
l
, l = 1,...,m can be calculated independently from each other, it
is possible to distribute their computation to diﬀerent processors of a parallel
machine.
(d) AOS schemes with α =1 have been presented in [424, 430] as eﬃcient dis
cretizations of the isotropic nonlinear diﬀusion ﬁlter of Catt´e et al. [81]. In
Section 1.3.2 we have seen that this ﬁlter is based on the PDE
∂
t
u = div (g([∇u
σ
[
2
) ∇u).
In this case, a natural operator splitting A =
¸
m
l=1
A
l
results from a decom
position of the divergence expression into onedimensional terms of type
∂
x
l
(g([∇u
σ
[
2
) ∂
x
l
u) (l = 1, ..., m).
This separation is very eﬃcient: There exist permutation matrices P
l
(pixel
orderings) such that P
l
A
l
P
T
l
is block diagonal and each block is diagonally
dominant and tridiagonal. Hence, the corresponding linear systems can be
4.4. RELATION TO SEMIDISCRETE MODELS 111
solved in linear eﬀort by means of a simple Gaussian algorithm. The resulting
forward substitution and backward elimination can be regarded as a recursive
ﬁlter.
A parallel implementation assigning these tridiagonal subsystems to diﬀerent
processors is described in [431]. The denoising of a medical 3D ultrasound
data set with 138 208 138 voxels on an SGI Power Challenge XL with
eight 195 MHz R10000 processors was possible in less than 1 minute.
(e) The idea to base AOS schemes on decompositions into onedimensional op
erators can also be generalized to anisotropic diﬀusion ﬁlters: Consider for
instance the discretization on a (3 3)stencil at the end of Section 3.4.2. If
it fulﬁls (S1)–(S5), then a splitting of such a 2D ﬁlter into 4 onedimensional
diﬀusion processes acting along the 4 stencil directions satisﬁes all prerequi
sites of Theorem 9.
112 CHAPTER 4. DISCRETE DIFFUSION FILTERING
Chapter 5
Examples and applications
The scalespace theory from Chapters 2–4 also covers methods such as linear or
nonlinear isotropic diﬀusion ﬁltering, for which many interesting applications have
already been mentioned in Chapter 1. Therefore, the goal of the present chapter is
to show that a generalization to anisotropic models with diﬀusion tensors depend
ing on the structure tensor oﬀers novel properties and application ﬁelds. Thus, we
focus mainly on these anisotropic techniques and juxtapose the results to other
methods. In order to demonstrate the ﬂexibility of anisotropic diﬀusion ﬁltering,
we shall pursue two diﬀerent objectives:
• smoothing with simultaneous edgeenhancement,
• smoothing with enhancement of coherent ﬂowlike textures.
All calculations for diﬀusion ﬁltering are performed using semiimplicit FD
schemes with time steps ∆t ∈ [2, 5]. In order to compare anisotropic diﬀusion to
other methods, morphological scalespaces and modiﬁcations of them have been
discretized as well. For MCM and AMSS this is achieved by means of explicit FD
schemes (cf. 1.6.4) with ∆t := 0.1 and ∆t := 0.01, respectively. Dilation with a ﬂat
structuring element is approximated by an OsherSethian scheme of type (1.88)
with ∆t := 0.5, and dilation with a quadratic structuring function is performed
in a noniterative way using van den Boomgaard’s algorithm from [51]. On an
HP 9000/889 workstation it takes less than 0.3 CPU seconds to calculate one
nonlinear diﬀusion step for a 256 256 image, and MCM, AMSS or dilation with
a disc require approximately 0.06 seconds per iteration. Typical dilations with a
quadratic structuring function take less than 0.3 seconds.
113
114 CHAPTER 5. EXAMPLES AND APPLICATIONS
5.1 Edgeenhancing diﬀusion
5.1.1 Filter design
In accordance with the notations in 2.2, let µ
1
, µ
2
with µ
1
≥ µ
2
be the eigenvalues
of the structure tensor J
ρ
, and v
1
, v
2
the corresponding orthonormal eigenvectors.
Since the diﬀusion tensor should reﬂect the local image structure it ought to be
chosen in such a way that it reveals the same set of eigenvectors v
1
, v
2
as J
ρ
. The
choice of the corresponding eigenvalues λ
1
, λ
2
depends on the desired goal of the
ﬁlter.
If one wants to smooth preferably within each region and aims to inhibit dif
fusion across edges, then one can reduce the diﬀusivity λ
1
perpendicular to edges
the more the higher the contrast µ
1
is, see 1.3.3 and [415]. This behaviour may be
accomplished by the following choice (m ∈ IN, C
m
> 0, λ > 0):
λ
1
(µ
1
) := g(µ
1
), (5.1)
λ
2
:= 1 (5.2)
with
g(s) :=
1 (s ≤ 0)
1 −exp
−Cm
(s/λ)
m
(s > 0).
(5.3)
This exponentially decreasing function is chosen in order to fulﬁl the smoothness
requirement stated in (P
c
), cf. 2.3. Since ∇u
σ
remains bounded on Ω[0, ∞) and
µ
1
= [∇u
σ
[
2
, we know that the uniform positive deﬁniteness of D is automatically
satisﬁed by this ﬁlter.
The constant C
m
is calculated in such a way that the ﬂux Φ(s) = sg(s) is
increasing for s ∈ [0, λ] and decreasing for s ∈ (λ, ∞). Thus, the preceding ﬁlter
strategy can be regarded as an anisotropic regularization of the Perona–Malik
model.
The choice m := 4 (which implies C
4
= 3.31488) gives visually good results
and is used exclusively in the examples below. Since in this section we are only
interested in edgeenhancing diﬀusion we may set the integration scale ρ of the
structure tensor equal to 0. Applications which require nonvanishing integration
scales shall be studied in section 5.2.
5.1.2 Applications
Figure 5.1 illustrates that anisotropic diﬀusion ﬁltering is still capable of possessing
the contrastenhancing properties of the Perona–Malik ﬁlter (provided the regu
larization parameter σ is not too large). It depicts the temporal evolution of a
5.1. EDGEENHANCING DIFFUSION 115
Figure 5.1: Anisotropic diﬀusion ﬁltering of a Gaussiantype function,
Ω = (0, 256)
2
, λ = 3.6, σ = 2. From top left to bottom right: t = 0,
125, 625, 3125, 15625, 78125.
116 CHAPTER 5. EXAMPLES AND APPLICATIONS
Gaussianlike function and its isolines.
1
It can be observed that two regions with
almost constant grey value evolve which are separated by a fairly steep edge. Edge
enhancement is caused by the fact that, due to the rapidly decreasing diﬀusivity,
smoothing within each region is strongly preferred to diﬀusion between the two
adjacent regions. The edge location remains stable over a very long time interval.
This indicates that, in practice, the determination of a suitable stopping time is
not a critical problem. After the process of contrast enhancement is concluded, the
steepness of edges decreases very slowly until the gradient reaches a value where no
backward diﬀusion is possible anymore. Then the image converges quickly towards
a constant image.
Let us now compare the denoising properties of diﬀerent diﬀusion ﬁlters. Figure
5.2(a) consists of a triangle and a rectangle with 70 % of all pixels being completely
degraded by noise. This image is taken from the software package MegaWave.
Test images of this type have been used to study the behaviour of ﬁlters such
as [13, 15, 16, 99, 102]. In Fig. 5.2(b) we observe that linear diﬀusion ﬁltering
is capable of removing all noise, but we have to pay a price: the image becomes
completely blurred. Besides the fact that edges get smoothed so that they are
harder to identify, the correspondence problem appears: edges become dislocated.
Thus, once they are identiﬁed at a coarse scale, they have to be traced back in
order to ﬁnd their true location, a theoretically and practically rather diﬃcult
problem.
Fig. 5.2(c) shows the eﬀect when applying the isotropic nonlinear diﬀusion
equation [81]
∂
t
u = div (g([∇u
σ
[
2
)∇u) (5.4)
with g as in (5.3). Since edges are hardly aﬀected by this process, nonlinear
isotropic diﬀusion does not lead to correspondence problems which are charac
teristic for linear ﬁltering. On the other hand, the drastically reduced diﬀusivity
at edges is also responsible for the drawback that noise at edges is preserved.
Figure 5.2(d) demonstrates that nonlinear anisotropic ﬁltering shares the ad
vantages of both methods. It combines the good noise eliminating properties of
linear diﬀusion with the stable edge structure of nonlinear isotropic ﬁltering. Due
to the permitted smoothing along edges, however, corners get more rounded than
in the nonlinear isotropic case.
The scalespace behaviour of diﬀerent PDEbased methods is juxtaposed in
Figures 5.3–5.6, where an MRI slice of a human head is processed [414, 423].
1
Except for Figs. 5.1, 5.3–5.6, where contrast enhancement is to be demonstrated, all images
in the present work are depicted in such a way that the lowest value is black and the highest one
appears white. They reveal a range within the interval [0, 255] and all pixels have unit length in
both directions.
5.1. EDGEENHANCING DIFFUSION 117
Figure 5.2: Restoration properties of diﬀusion ﬁlters. (a) Top Left:
Test image, Ω = (0, 128)
2
. (b) Top Right: Linear diﬀusion, t = 80.
(c) Bottom Left: Nonlinear isotropic diﬀusion, λ = 3.5, σ = 3,
t = 80. (d) Bottom Right: Nonlinear anisotropic diﬀusion, λ = 3.5,
σ = 3, t = 80.
Again we observe that linear diﬀusion (Fig. 5.3(a)) does not only blur all struc
tures in an equal amount but also dislocates them more and more with increasing
scale.
A ﬁrst step to reduce these problems is to adapt the diﬀusivity to the gradient
of the initial image f [147]. Fig. 5.3(b) shows the evolution under
∂
t
u = div (g([∇f[
2
) ∇u), (5.5)
118 CHAPTER 5. EXAMPLES AND APPLICATIONS
where a diﬀusivity of type [88]
g([∇f[
2
) :=
1
1 +[∇f[
2
/λ
2
(λ > 0). (5.6)
is used. Compared with homogeneous linear diﬀusion, edges remain better localized
and their blurring is reduced. On the other hand, for large t the ﬁltered image
reveals some artifacts which reﬂect the diﬀerential structure of the initial image.
A natural idea to reduce the artifacts of inhomogeneous linear diﬀusion ﬁltering
would be to introduce a feedback in the process by adapting the diﬀusivity g to
the gradient of the actual image u(x, t) instead of the original image f(x). This
leads to the nonlinear diﬀusion equation [326]
∂
t
u = div (g([∇u[
2
) ∇u). (5.7)
Figure 5.3(c) shows how such a nonlinear feedback is useful to increase the edge
localization in a signiﬁcant way: Structures remain welllocalized as long as they
can be recognized. Also blurring at edges is reduced very much. The absolute
contrast at edges, however, becomes smaller.
The latter problem can be avoided using a diﬀusivity which decreases faster
than (5.6) and leads to a nonmonotone ﬂux function. This is illustrated in Figure
5.4(a) where the regularized isotropic nonlinear diﬀusion ﬁlter (5.4) with the diﬀu
sivity (5.3) is applied. At the chin we observe that this equation is indeed capable
of enhancing edges. All structures are extremely welllocalized and the results are
segmentationlike. On the other hand, also small structures exist over a long range
of scales if they diﬀer from their vicinity by a suﬃciently large contrast. One can
try to make this ﬁlter faster and more insensitive to smallsize structures by in
creasing the regularizing Gaussian kernel size σ (cf. Fig. 5.4(b)), but this also leads
to stronger blurring of large structures, and it is no longer possible to enhance the
contour of the entire head.
Anisotropic nonlinear diﬀusion (Fig. 5.4(c)) permits diﬀusion along edges and
inhibits smoothing across them. As in Figure 5.2(d), this causes a stronger round
ing of structures, which can be seen at the nose. A positive consequence of this
slight shrinking eﬀect is the fact that small or elongated and thin structures are
better eliminated than in the isotropic case. Thus, we recognize that most of the
depicted “segments” coincide with semantically correct objects that one would ex
pect at these scales. Finally the image turns into a silhouette of the head, before
it converges to a constant image.
The tendency to produce piecewise almost constant regions indicates that dif
fusion scalespaces with nonmonotone ﬂux are ideal preprocessing tools for seg
mentation. Unlike diﬀusion–reaction models aiming to yield one segmentationlike
result for t → ∞ (cf. 1.4), the temporal evolution of these models generates a
5.1. EDGEENHANCING DIFFUSION 119
complete hierarchical family of segmentationlike images. The contrastenhancing
quality distinguishes nonlinear diﬀusion ﬁlters from most other scalespaces. It
should be noted that contrast enhancement is a local phenomenon which cannot
be replaced by simple global rescalings of the grey value range. Therefore, it is
generally not possible to obtain similar segmentationlike results by just rescaling
the grey values from a scalespace which is only contrastreducing.
The contrast and noise parameters λ and σ give the user the liberty to adapt
nonlinear diﬀusion scalespaces to the desired purpose in order to reward inter
esting features with a longer lifetime. Suitable values for them should result in a
natural way from the speciﬁc problem. In this sense, the time t is rather a para
meter of importance, with respect to the speciﬁed task, than a descriptor of spatial
scale. The traditional opinion that the evolution parameter t of scalespaces should
be related to the spatial scale reﬂects the assumption that a scalespace analysis
should be uncommitted. Nonlinear diﬀusion ﬁltering renounces this requirement by
allowing to incorporate apriori information (e.g. about the contrast of semantically
important structures) into the evolution process. The basic idea of scalespaces,
however, is maintained: to provide a family of subsequently simpliﬁed versions of
the original image, which gives a hierarchy of structures and allows to extract the
relevant information from a certain scale.
Besides these speciﬁc features of nonlinear diﬀusion scalespaces it should be
mentioned that, due to the homogeneous Neumann boundary condition and the
divergence form, both linear and nonlinear diﬀusion ﬁlters preserve the average
grey level of the image.
This is not true for the morphological ﬁlters and their modiﬁcations which are
depicted in Fig. 5.5 and 5.6.
Figure 5.5(a) and (b) show the result under continuousscale dilation with a ﬂat
discshaped structuring element and a quadratic structuring function, respectively.
From Section 1.5.3 and 1.5.6 we know that their evolution equations are given by
∂
t
u = [∇u[, (5.8)
for the disc, and by
∂
t
u = [∇u[
2
(5.9)
for the quadratic structuring function. In both cases the number of local maxima
is decreasing, and maxima keep their location in scalespace until they disappear
[206, 207]. The fact that the maximum with the largest grey value will dominate
at the end shows that these processes can be sensitive to noise (maxima might
be caused by noise), and that they usually do not preserve the average grey level.
It is not very diﬃcult to guess the shape of the structuring function from the
scalespace evolution.
120 CHAPTER 5. EXAMPLES AND APPLICATIONS
Figure 5.3: Evolution of an MRI slice under diﬀerent PDEs. Top: Original im
age, Ω = (0, 236)
2
. (a) Left Column: Linear diﬀusion, top to bottom: t = 0,
12.5, 50, 200. (b) Middle Column: Inhomogeneous linear diﬀusion (λ = 8),
t = 0, 70, 200, 600. (c) Right Column: Nonlinear isotropic diﬀusion with the
Charbonnier diﬀusivity (λ = 3), t = 0, 70, 150, 400.
5.1. EDGEENHANCING DIFFUSION 121
Figure 5.4: Evolution of an MRI slice under diﬀerent PDEs. Top: Original image,
Ω = (0, 236)
2
. (a) Left Column: Isotropic nonlinear diﬀusion (λ = 3, σ =
1), t = 0, 25000, 500000, 7000000. (b) Middle Column: Isotropic nonlinear
diﬀusion (λ = 3, σ = 4), t = 0, 40, 400, 1500. (c) Right Column: Edge
enhancing anisotropic diﬀusion (λ = 3, σ = 1), t = 0, 250, 875, 3000.
122 CHAPTER 5. EXAMPLES AND APPLICATIONS
Figure 5.5: Evolution of an MRI slice under diﬀerent PDEs. Top: Original image,
Ω = (0, 236)
2
. (a) Left Column: Dilation with a disc, t = 0, 4, 10, 20. (b)
Middle Column: Dilation with a quadratic structuring function, t = 0, 0.25,
1, 4. (c) Right Column: Mean curvature motion, t = 0, 70, 275, 1275.
5.1. EDGEENHANCING DIFFUSION 123
Figure 5.6: Evolution of an MRI slice under diﬀerent PDEs. Top: Original image,
Ω = (0, 236)
2
. (a) Left Column: Aﬃne morphological scalespace, t = 0, 20,
50, 140. (b) Middle Column: Modiﬁed mean curvature motion (λ = 3, σ = 1),
t = 0, 100, 350, 1500. (c) Right Column: Selfsnakes (λ = 3, σ = 1), t = 0,
600, 5000, 40000.
124 CHAPTER 5. EXAMPLES AND APPLICATIONS
A completely diﬀerent morphological evolution is given by the mean curvature
motion (1.90) depicted in Fig. 5.5(c). Since MCM shrinks level lines with a ve
locity that is proportional to their curvature, lowcurved object boundaries are
less aﬀected by this process, while highcurved structures (e.g. the nose) exhibit
roundings at an earlier stage. This also explains its excellent noise elimination
qualities. After some time, however, the head looks almost like a ball. This is in
accordance with the theory which predicts convergence of all closed level lines to
circular points.
A similar behaviour can be observed for the aﬃne invariant morphological scale
space (1.108) shown in Fig. 5.6(a). Since it takes the time T =
3
4
s
4
3
to remove
all isolines within a circle of radius s – in contrast to T =
1
2
s
2
for MCM – we
see that, for a comparable elimination of small structure, the shrinking eﬀect of
large structures is stronger for AMSS than for MCM. Thus, the correspondence
problem is more severe than for MCM. Nevertheless, the advantage of having aﬃne
invariance may counterbalance the correspondence problem in certain applications.
Since the AMSS involves no additional parameters and oﬀers more invariances
than other scalespaces, it is ideal for uncommitted image analysis and shape
recognition. Both MCM and AMSS give signiﬁcantly sharper edges than linear
diﬀusion ﬁltering, but they are not designed to act contrastenhancing.
One possibility to reduce the correspondence problem of morphological scale
spaces is to attenuate the curve evolution at highcontrast edges. This is at the
expense of withdrawing morphology in terms of invariance under monotone grey
scale transformations.
One possibility is to use the damping function outside the divergence expression.
Processes of this type are studied in [13, 364, 365, 304]. As a simple prototype for
this idea, let us investigate the modiﬁed MCM
∂
t
u = g([∇u
σ
[
2
) [∇u[ div
∇u
[∇u[
(5.10)
with g([∇u
σ
[
2
) as in (1.32). The corresponding evolution is depicted in Fig. 5.6(b).
We observe that structures remain much better localized than in the original MCM.
On the other hand, the experiments give evidence that this process is probably
not contrastenhancing, see e.g. the chin. As a consequence, the results appear less
segmentationlike than those for nonlinear diﬀusion ﬁltering.
Using g([∇u
σ
[
2
) inside the divergence expression leads to
∂
t
u = [∇u[ div
g([∇u
σ
[
2
)
∇u
[∇u[
. (5.11)
In Section 1.6.6 we have seen that processes of this type are called selfsnakes
[357]. Since they diﬀer from isotropic nonlinear diﬀusion ﬁlters by the [∇u[ terms
inside and outside the divergence expression, they will not preserve the average
5.1. EDGEENHANCING DIFFUSION 125
Figure 5.7: Preprocessing of a fabric image. (a) Left: Fabric, Ω =
(0, 257)
2
. (b) Right: Anisotropic diﬀusion, λ = 4, σ = 2, t = 240.
grey value. The evolution in Fig. 5.6(c) indicates that g([∇u
σ
[
2
) gives similar edge
enhancing eﬀects as in a nonlinear diﬀusion ﬁlter, but one can observe a stronger
tendency to create circular structures. This behaviour which resembles MCM is
not surprising if one compares (1.120) with (1.121).
Let us now study two applications of nonlinear diﬀusion ﬁltering in computer
aided quality control (CAQ): the grading of fabrics and wood surfaces (see also
[413]).
The quality of a fabric is determined by two criteria, namely clouds and stripes.
Clouds result from isotropic inhomogeneities of the density distribution, whereas
stripes are an anisotropic phenomenon caused by adjacent ﬁbres pointing in the
same direction. Anisotropic diﬀusion ﬁlters are capable of visualizing both quality
relevant features simultaneously (Fig. 5.7). For a suitable parameter choice, they
perform isotropic smoothing at clouds and diﬀuse in an anisotropic way along ﬁ
bres in order to enhance them. However, if one wants to visualize both features
separately, one can use a fast pyramid algorithm based on linear diﬀusion ﬁlter
ing for the clouds [417], whereas stripes can be enhanced by a special nonlinear
diﬀusion ﬁlter which is designed for closing interrupted lines and which shall be
discussed in Section 5.2.
For furniture production it is of importance to classify the quality of wood
surfaces. If one aims to automize this evaluation, one has to process the image in
such a way that quality relevant features become better visible und unimportant
structures disappear. Fig. 5.8(a) depicts a wood surface possessing one defect. To
visualize this defect, equation (5.4) can be applied with good success (Fig. 5.8(b)).
126 CHAPTER 5. EXAMPLES AND APPLICATIONS
Figure 5.8: Defect detection in wood. (a) Left: Wood surface, Ω =
(0, 256)
2
. (b) Right: Isotropic nonlinear diﬀusion, λ = 4, σ = 2,
t = 2000.
In [413] it is demonstrated how a modiﬁed anisotropic diﬀusion process yields even
more accurate results with less roundings at the corners.
Fig. 5.9(a) gives an example for possible medical applications of nonlinear dif
fusion ﬁltering as a preprocessing tool for segmentation (see also [415] for another
example). It depicts an MRI slice of the human head. For detecting Alzheimer’s
disease one is interested in determining the ratio between the ventricle areas, which
are given by the two white longitudinal objects in the centre, and the entire head
area.
In order to make the diagnosis more objective and reliable, it is intended to
automize this feature extraction step by a segmentation algorithm. Figure 5.9(c)
shows a segmentation according to the following simpliﬁcation of the Mumford–
Shah functional (1.58):
E
f
(u, K) =
Ω
(u−f)
2
dx + α[K[. (5.12)
It has been obtained by a MegaWave programme using a hierarchical region grow
ing algorithm due to Koepﬂer et al. [239]. As is seen in Fig. 5.9(d), one gets a better
segmentation when processing the original image slightly by means of nonlinear
diﬀusion ﬁltering (Fig. 5.9(b)) prior to segmenting it.
5.2. COHERENCEENHANCING DIFFUSION 127
Figure 5.9: Preprocessing of an MRI slice. (a) Top Left: Head, Ω =
(0, 256)
2
. (b) Top Right: Diﬀusionﬁltered, λ = 5, σ = 0.1, t =
2.5. (c) Bottom Left: Segmented original image, α = 8192. (d)
Bottom Right: Segmented ﬁltered image, α = 8192.
5.2 Coherenceenhancing diﬀusion
5.2.1 Filter design
In this section we shall investigate how the structure tensor information can be
used to design anisotropic diﬀusion scalespaces which enhance the coherence of
ﬂowlike textures [418]. This requires a nonvanishing integration scale ρ.
Let again µ
1
, µ
2
with µ
1
≥ µ
2
be the eigenvalues of J
ρ
, and v
1
, v
2
the cor
responding orthonormal eigenvectors. As in 5.1 the diﬀusion tensor D(J
ρ
(∇u
σ
))
ought to possess the same set of eigenvectors as J
ρ
(∇u
σ
).
If one wants to enhance coherent structures, one should smooth preferably
128 CHAPTER 5. EXAMPLES AND APPLICATIONS
Figure 5.10: Local orientation in a ﬁngerprint image. (a) Top Left:
Original ﬁngerprint, Ω = (0, 256)
2
. (b) Top Right: Orientation
of smoothed gradient, σ = 0.5. (c) Bottom Left: Orientation of
smoothed gradient, σ = 5. (d) Bottom Right: Structure tensor
orientation, σ = 0.5, ρ = 4.
along the coherence direction v
2
with a diﬀusivity λ
2
which increases with respect
to the coherence (µ
1
−µ
2
)
2
. This may be achieved by the following choice for the
eigenvalues of the diﬀusion tensor (C > 0, m ∈ IN):
λ
1
:= α,
λ
2
:=
α if µ
1
=µ
2
,
α + (1−α) exp
−C
(µ
1
−µ
2
)
2m
else,
where the exponential function was chosen to ensure the smoothness of D and the
5.2. COHERENCEENHANCING DIFFUSION 129
Figure 5.11: Anisotropic equations applied to the ﬁngerprint image.
(a) Left: Meancurvature motion, t = 5. (b) Right: Coherence
enhancing anisotropic diﬀusion, σ = 0.5, ρ = 4, t = 20.
small positive parameter α∈(0, 1) keeps D(J
ρ
(∇u
σ
)) uniformly positive deﬁnite.
2
All examples below are calculated using C := 1, m := 1, and α := 0.001.
5.2.2 Applications
Figure 5.10 illustrates the advantages of local orientation analysis by means of
the structure tensor. In order to detect the local orientation of the ﬁngerprint
depicted in Fig. 5.10(a), the gradient orientation of a slightly smoothed image has
been calculated (Fig. 5.10(b)). Horizontally oriented structures appear black, while
vertical structures are represented in white. We observe very high ﬂuctuations in
the local orientation. When applying a larger smoothing kernel it is clear that
adjacent gradients having the same orientation but opposite direction cancel out.
Therefore, the results in (c) are much worse than in (b). The structure tensor,
however, averages the gradient orientation instead of its direction. This is the
reason for the reliable estimates of local orientation that can be obtained with this
method (Fig. 5.10(d)).
To illustrate how the result of anisotropic PDE methods depends on the direc
tion in which they smooth, let us recall the example of mean curvature motion (cf.
1.6.1):
∂
t
u = u
ξξ
= [∇u[ curv(u) (5.13)
with ξ being the direction perpendicular to ∇u. Since MCM smoothes by prop
agating level lines in inner normal direction we recognize that its smoothing di
2
Evidently, ﬁlters of this type are not regularizations of the Perona–Malik process: the limit
σ →0, ρ →0 leads to a linear diﬀusion process with constant diﬀusivity α.
130 CHAPTER 5. EXAMPLES AND APPLICATIONS
Figure 5.12: Scalespace behaviour of coherenceenhancing diﬀusion
(σ = 0.5, ρ = 2). (a) Top Left: Original fabric image, Ω = (0, 257)
2
.
(b) Top Right: t = 20. (c) Bottom Left: t = 120. (d) Bottom
Right: t = 640.
rection depends exclusively on ∇u. Thus, although this method is in a complete
anisotropic spirit, we should not expect it to be capable of closing interrupted
linelike structures. The results in Fig. 5.11(a) conﬁrm this impression.
The proposed anisotropic diﬀusion ﬁlter, however, biases the diﬀusive ﬂux to
wards the coherence orientation v
2
and is therefore wellsuited for closing inter
rupted lines in coherent ﬂowlike textures, see Fig. 5.11(b). Due to its reduced
diﬀusivity at noncoherent structures, the locations of the semantically important
singularities in the ﬁngerprint remain the same. This is an important prerequisite
that any image processing method has to satisfy if it is to be applied to ﬁngerprint
analysis.
5.2. COHERENCEENHANCING DIFFUSION 131
Figure 5.13: (a) Top: High resolution slipring CT scan of a femural bone,
Ω = (0, 300) (0, 186). (b) Bottom Left: Filtered by coherenceenhancing
anisotropic diﬀusion, σ = 0.5, ρ = 6, t = 16. (c) Bottom Right: Dito with
t =128.
Figure 5.12 depicts the scalespace behaviour of coherenceenhancing aniso
tropic diﬀusion applied to the fabric image from Fig. 5.7. The temporal behaviour
of this diﬀusion ﬁlter seems to be appropriate for visualizing coherent ﬁbre agglom
erations (stripes) at diﬀerent scales, a diﬃcult problem for the automatic grading
of nonwovens [299].
Figure 5.13 illustrates the potential of CED for medical applications. It depicts
a human bone. Its internal structure has a distinctive texture through the presence
of tiny elongated bony structural elements, the trabeculae. There is evidence that
the trabecular formation is for a great deal determined by the external load. For
this reason the trabecular structure constitutes an important clinical parameter in
orthopedics. Examples are the control of recovery after surgical procedures, such
as the placement or removal of metal implants, quantifying the rate of progression
of rheumatism and osteoporosis, the determination of leftright deviations of sym
metry in the load or establishing optimal load corrections by physiotherapy. From
Figure 5.13(b),(c) we observe that CED is capable of enhancing the trabecular
structures in order to ease their subsequent orientation analysis.
132 CHAPTER 5. EXAMPLES AND APPLICATIONS
Figure 5.14: Image restoration using coherenceenhancing anisotropic
diﬀusion. (a) Left: “Selfportrait” by van Gogh (SaintR´emy, 1889;
Paris, Muse´e d’Orsay), Ω = (0, 215) (0, 275). (b) Right: Filtered,
σ = 0.5, ρ = 4, t = 6.
Let us now investigate the impact of coherenceenhancing diﬀusion on images,
which are not typical texture images, but still reveal a ﬂowlike character. To this
end, we shall process impressionistic paintings by Vincent van Gogh.
Fig. 5.14 shows the restoration properties of coherenceenhancing anisotropic
diﬀusion when being applied to a selfportrait of the artist [161]. We observe that
the diﬀusion ﬁlter can close interrupted lines and enhance the ﬂowlike character
which is typical for van Gogh paintings.
The next painting we are concerned with is called “Road with Cypress and
Star” [162, 429]. It is depicted in Fig. 5.15. In order to demonstrate the inﬂuence
of the integration scale ρ, all ﬁlter parameters are ﬁxed except for ρ. Fig. 5.15(b)
shows that a value for ρ which is too small does not lead to the visually dominant
coherence orientation and creates structures with a lot of undesired ﬂuctuations.
Increasing the value for ρ improves the image signiﬁcantly (Fig. 5.15(c)). Interest
ingly, a further increasing of ρ does hardly alter this result (Fig. 5.15(d)), which
indicates that this van Gogh painting possesses a uniform “texture scale” reﬂecting
the characteristic painting style of the artist.
In a last example the temporal evolution of ﬂowlike images is illustrated by
virtue of the “Starry Night” painting in Fig. 5.16 [160, 419]. Due to the established
5.2. COHERENCEENHANCING DIFFUSION 133
Figure 5.15: Impact of the integration scale on coherenceenhancing
anisotropic diﬀusion (σ = 0.5, t = 8). (a) Top Left: “Road with
Cypress and Star” by van Gogh (AuverssurOise, 1890; Otterlo, Ri
jksmuseum Kr¨oller–M¨ uller), Ω = (0, 203) (0, 290). (b) Top Right:
Filtered with ρ = 1. (c) Bottom Left: ρ = 4. (d) Bottom Right:
ρ = 6.
134 CHAPTER 5. EXAMPLES AND APPLICATIONS
Figure 5.16: Scalespace properties of coherenceenhancing anisotropic diﬀusion
(σ = 0.5, ρ = 4). (a) Top Left: “Starry Night” by van Gogh (SaintR´emy,
1889; New York, The Museum of Modern Art), Ω = (0, 255) (0, 199). (b) Top
Right: t = 8. (c) Bottom Left: t = 64. (d) Bottom Right: t = 512.
scalespace properties, the image becomes gradually simpler in many aspects, be
fore it ﬁnally will tend to its simplest representation, a constant image with the
same average grey value as the original one. The ﬂowlike character, however, is
maintained for a very long time.
3
3
Results for AMSS ﬁltering of this image can be found in [305].
Chapter 6
Conclusions and perspectives
While Chapter 1 has given a general overview of PDEbased smoothing and restora
tion methods, the goal of Chapters 2–5 has been to present a scalespace framework
for nonlinear diﬀusion ﬁltering which does not require any monotony assump
tion (comparison principle). We have seen that, besides the fact that many global
smoothing scalespace properties are maintained, new possibilities with respect to
image restoration appear.
Rather than deducing a unique equation from ﬁrst principles we have ana
lysed wellposedness and scalespace properties of a general family of regularized
anisotropic diﬀusion ﬁlters. Existence and uniqueness results, continuous depen
dence of the solution on the initial image, maximum–minimum principles, invari
ances, Lyapunov functionals, and convergence to a constant steadystate have been
established.
The large class of Lyapunov functionals permits to regard these ﬁlters in many
ways as simplifying, informationreducing transformations. These global smooth
ing properties do not contradict seemingly opposite local eﬀects such as edge en
hancement. For this reason it is possible to design scalespaces with restoration
properties giving segmentationlike results.
Prerequisites have been stated under which one can prove wellposedness and
scalespace results in the continuous, semidiscrete and discrete setting. Each of
these frameworks is selfcontained and does not require the others. On the other
hand, the prerequisites in all three settings reveal many similarities and, as a
consequence, representatives of the semidiscrete class can be obtained by suitable
spatial discretizations of the continuous class, while representatives of the discrete
class may arise from time discretizations of semidiscrete ﬁlters.
The degree of freedom within the proposed class of ﬁlters can be used to tailor
the ﬁlters towards speciﬁc restoration tasks. Therefore, these scalespaces do not
need to be uncommitted; they give the user the liberty to incorporate apriori
knowledge, for instance concerning size and contrast of especially interesting fea
tures.
135
136 CHAPTER 6. CONCLUSIONS AND PERSPECTIVES
The analysed class comprises linear diﬀusion ﬁltering and the nonlinear iso
tropic model of Catt´e et al. [81] and Whitaker and Pizer [438], but also novel ap
proaches have been proposed: The use of diﬀusion tensors instead of scalarvalued
diﬀusivities puts us in a position to design real anisotropic diﬀusion processes
which may reveal advantages at noisy edges. Last but not least, the fact that these
ﬁlters are steered by the structure tensor instead of the regularized gradient allows
to adapt them to more sophisticated tasks such as the enhancement of coherent
ﬂowlike structures.
In view of these results, anisotropic diﬀusion deserves to be regarded as much
more than an adhoc strategy for transforming a degraded image into a more
pleasant looking one. It is a ﬂexible and mathematically sound class of methods
which ties the advantages of two worlds: scalespace analysis and image restoration.
It is clear, however, that nonlinear diﬀusion ﬁltering is a young ﬁeld which has
certainly not reached its ﬁnal state yet. Thus, we can expect a lot of new results
in the near future. Some of its future developments, however, are likely to consist
of straightforward extensions of topics presented in this text:
• While the theory and the examples in the present book focus on 2D grey
scale images, it is evident that most of its results can easily be generalized to
higher dimensions and vectorvalued images. The need for such extensions
grows with the rapid progress in the development of faster computers, the
general availability of aﬀordable colour scanners and printers, and the wish
to integrate information from diﬀerent channels. Some of the references in
Chapter 1 point in directions how this can be accomplished.
• The various possibilities to include semilocal or global information constitute
another future perspective. This could lead to speciﬁcally tuned ﬁlters for
topics such as perceptual grouping. Coherenceenhancing anisotropic diﬀu
sion is only a ﬁrst step in this direction. New ﬁlter models might arise using
other structure descriptors than the (regularized) gradient or the structure
tensor. Interesting candidates could be wavelets, Gabor ﬁlters, or steerable
ﬁlters.
• In contrast to the linear diﬀusion case, the relation between structures at
diﬀerent scales has rarely been exploited in the nonlinear context. Although
this problem is less severe, since the avoidance of correspondence problems
was one of the key motivations to study nonlinear scalespaces, it would cer
tainly be useful to better understand the deep structure in nonlinear diﬀusion
processes. The scalespace stack of these ﬁlters appears to be wellsuited to
extract semantically important information with respect to a speciﬁed task.
This ﬁeld oﬀers a lot of challenging mathematical questions.
137
• Most people working in computer vision do not have a speciﬁc knowledge
on numerical methods for PDEs. As a consequence, the most widelyused
numerical methods for nonlinear diﬀusion ﬁltering are still the simple, but
ineﬃcient explicit (Euler forward) schemes. Novel, more timecritical appli
cation areas could be explored by applying implicit schemes, splitting and
multigrid techniques, or grid adaptation strategies. In this context it would be
helpful to have software packages, where diﬀerent nonlinear diﬀusion ﬁlters
are implemented in an eﬃcient way, and which are easy to use for everyone.
• The price one has to pay for the ﬂexibility of nonlinear diﬀusion ﬁltering is
the speciﬁcation of some parameters. Since these parameters have a rather
natural meaning, this is not a very diﬃcult problem for someone with ex
perience in computer vision. Somebody with another primary interest, for
instance a physician who wants to denoise ultrasound images, may be fright
ened by this perspective. Thus, more research on ﬁnding some guidelines
for automatic parameter determination for a task at hand would encourage
also people without a speciﬁc image processing background to apply nonlin
ear diﬀusion ﬁlters. Several useful suggestions for parameter adaptation can
already be found in [328, 36, 431, 270, 444].
• There are not yet many studies which explore the potential of nonlinear dif
fusion ﬁltering when being combined with other image processing techniques.
Especially combinations with concepts such as data compression, segmenta
tion algorithms, tomographic reconstruction techniques, or neural networks
for learning apriori information might lead to novel application areas for
these techniques.
Thus, there still remains a lot of work to be done. It would be nice if this book has
inspired its readers to contribute to the solution of some of the remaining open
problems.
138 CHAPTER 6. CONCLUSIONS AND PERSPECTIVES
Bibliography
[1] S. Abbasi, F. Mokhtarian, J. Kittler, Reliable classiﬁcation of chrysanthemum leaves
through curvature scale space, B. ter Haar Romeny, L. Florack, J. Koenderink, M.
Viergever (Eds.), Scalespace theory in computer vision, Lecture Notes in Computer Sci
ence, Vol. 1252, Springer, Berlin, 284–295, 1997.
[2] R. Acar, C.R. Vogel Analysis of bounded variation penalty methods for illposed problems,
Inverse Problems, Vol. 10, 1217–1229, 1994.
[3] A. Ackah–Miezan, A. Gagalowicz, Discrete models for energyminimizing segmentation,
Proc. Fourth Int. Conf. on Computer Vision (ICCV ’93, Berlin, May 11–14, 1993), IEEE
Computer Society Press, Los Alamitos, 200–207, 1993.
[4] S.T. Acton, Edge enhancement of infrared imagery by way of the anisotropic diﬀusion
pyramid, Proc. IEEE Int. Conf. Image Processing (ICIP–96, Lausanne, Sept. 16–19, 1996),
Vol. 1, 865–868, 1996.
[5] S.T. Acton, A.C. Bovik, M.M. Crawford, Anisotropic diﬀusion pyramids for image seg
mentation, Proc. IEEE Int. Conf. Image Processing (ICIP–94, Austin, Nov. 13–16, 1994),
Vol. 3, IEEE Computer Society Press, Los Alamitos, 478–482, 1994.
[6] S.T. Acton, M.M. Crawford, A mean ﬁeld solution to anisotropic edge detection of remotely
sensed data, Proc. 12th Int. Geoscience and Remote Sensing Symposium (IGARSS ’92,
Houston, May 26–29, 1992), Vol. 2, 845–847, 1992.
[7] R.A. Adams, Sobolev spaces, Academic Press, New York, 1975.
[8] D. Adalsteinsson, J.A. Sethian, A fast level set method for propagating interfaces, J. Com
put. Phys., Vol. 118, 269–277, 1995.
[9] H.W. Alt, Lineare Funktionalanalysis, Springer, Berlin, 1992.
[10] L. Alvarez, Images and PDE’s, M.O. Berger, R. Deriche, I. Herlin, J. Jaﬀr´e, J.M. Morel
(Eds.), ICAOS ’96: Images, wavelets and PDEs, Lecture Notes in Control and Information
Sciences, Vol. 219, Springer, London, 3–14, 1996.
[11] L. Alvarez, J. Esclarin, Image quantization using reaction–diﬀusion equations, SIAM J.
Appl. Math., Vol. 57, 153–175, 1997.
[12] L. Alvarez, F. Guichard, P.L. Lions, J.M. Morel, Axioms and fundamental equations in
image processing, Arch. Rational Mech. Anal., Vol. 123, 199–257, 1993.
[13] L. Alvarez, P.L. Lions, J.M. Morel, Image selective smoothing and edge detection by
nonlinear diﬀusion. II, SIAM J. Numer. Anal., Vol. 29, 845–866, 1992.
[14] L. Alvarez, L. Mazorra, Signal and image restoration using shock ﬁlters and anisotropic
diﬀusion, SIAM J. Numer. Anal., Vol. 31, 590–605, 1994.
139
140 BIBLIOGRAPHY
[15] L. Alvarez, F. Morales, Aﬃne morphological multiscale analysis of corners and multiple
junctions, Report No. 9402, Departamento de Inform´atica y Sistemas, Universidad de Las
Palmas, Campus Universitario de Taﬁra, 35017 Las Palmas, Spain, 1994; to appear in Int.
J. Comput. Vision.
[16] L. Alvarez, J.M. Morel, Formalization and computational aspects of image analysis, Acta
Numerica, 1–59, 1994.
[17] L. Ambrosio, A compactness theorem for a special class of functions of bounded variations,
Boll. Un. Math. Ital., Vol. 3B, 857–881, 1990.
[18] L. Ambrosio, N. Fusco, D. Pallara, Partial regularity of free discontinuity sets II, Preprint
CVGMT96040106, Math. Dept., Scuola Normale Superiore, Piazza dei Cavalieri, 56126
Pisa, Italy, 1995; submitted to Annali della Scuola Normale Superiore di Pisa.
[19] L. Ambrosio, N. Fusco, D. Pallara, Higher regularity of solutions of free discontinuity
problems, Preprint CVGMT96040107, Math. Dept., Scuola Normale Superiore, Piazza
dei Cavalieri, 56126 Pisa, Italy, 1995.
[20] L. Ambrosio, D. Pallara, Partial regularity of free discontinuity sets I, Preprint CVGMT
96040105, Math. Dept., Scuola Normale Superiore, Piazza dei Cavalieri, 56126 Pisa, Italy,
1995; submitted to Annali della Scuola Normale Superiore di Pisa.
[21] L. Ambrosio, V. Tortorelli, On the approximation of functionals depending on jumps by
elliptic functionals, Comm. Pure Appl. Math., Vol. 43, 999–1036, 1990.
[22] L. Ambrosio, V. Tortorelli, Approximation of functionals depending on jumps by elliptic
functionals via Γconvergence, Boll. Un. Math. Ital., Vol. 7, 105–123, 1992.
[23] S. Angenent, Parabolic equations on curves and surfaces. Part I. Curves with pintegrable
curvature, Annals of Mathematics, Vol. 132, 451–483, 1990.
[24] S. Angenent, Parabolic equations on curves and surfaces. Part II. Intersections, blowup,
and generalized solutions, Annals of Mathematics, Vol. 133, 171–215, 1991.
[25] A.B. Arehart, L. Vincent, B.B. Kimia, Mathematical morphology: The Hamilton–Jacobi
connection, Proc. Fourth Int. Conf. on Computer Vision (ICCV ’93, Berlin, May 11–14,
1993), IEEE Computer Society Press, Los Alamitos, 215–219, 1993.
[26] S.R. Arridge, A. Simmons, Multispectral probabilistic diﬀusion using Bayesian classiﬁca
tion, B. ter Haar Romeny, L. Florack, J. Koenderink, M. Viergever (Eds.), Scalespace
theory in computer vision, Lecture Notes in Computer Science, Vol. 1252, Springer, Berlin,
224–235, 1997.
[27] K.
˚
Astr¨om, A. Heyden, Stochastic analysis of image acquisition and scalespace smoothing,
J. Sporring, M. Nielsen, L. Florack, P. Johansen (Eds.), Gaussian scalespace theory,
Kluwer, Dordrecht, 129–136, 1997.
[28] I. Bajla, I. Holl¨ander, Geometrydriven diﬀusion smoothing of the MR brain images using
a novel variable conductance, Proc. 18th Annual Int. Conf. of the IEEE Engineering in
Medicine and Biology Soc. (EMB ’96, Amsterdam, Oct. 31 – Nov. 3, 1996), CDROM
document N21.
[29] I. Bajla, M. Maruˇsiak, M.
ˇ
Sr´amek, Anisotropic ﬁltering of MRI data based upon image
gradient histogram, D. Chetverikov, W.G. Kropatsch (Eds.), Computer analysis of images
and patterns, Lecture Notes in Computer Science, Vol. 719, Springer, Berlin, 90–97, 1993.
[30] J. Babaud, A.P. Witkin, M. Baudin, R.O. Duda, Uniqueness of the Gaussian kernel for
scale space ﬁltering, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 8, 26–33, 1986.
BIBLIOGRAPHY 141
[31] C. Ballester Nicolau, An aﬃne invariant model for image segmentation: Mathematical
analysis and applications, Ph.D. thesis, University of Illes Baleares, Palma de Mallorca,
Spain, 1995.
[32] C. Ballester, V. Caselles, M. Gonz´alez, Aﬃne invariant segmentation by variational
method, SIAM J. Appl. Math., Vol. 56, 294–325, 1996.
[33] C. Ballester, M. Gonz´alez, Texture segmentation by variational methods, M.O. Berger, R.
Deriche, I. Herlin, J. Jaﬀr´e, J.M. Morel (Eds.), ICAOS ’96: Images, wavelets and PDEs,
Lecture Notes in Control and Information Sciences, Vol. 219, Springer, London, 187–193,
1996.
[34] E. B¨ansch, K. Mikula, A coarsening ﬁnite element strategy in image selective smoothing,
Preprint No. 18/1996, Faculty of Mathematics, University of Freiburg, Hebelstr. 27, 79104
Freiburg, Germany, 1996; to appear in Computation and Visualization in Science.
[35] G.I. Barenblatt, M. Bertsch, R. Dal Passo, M. Ughi, A degenerate pseudoparabolic reg
ularization of a nonlinear forward–backward heat equation arising in the theory of heat
and mass exchange in stably stratiﬁed turbulent shear ﬂow, SIAM J. Math. Anal., Vol. 24,
1414–1439, 1993.
[36] B. Benhamouda, Parameter adaptation for nonlinear diﬀusion in image processing, mas
ter thesis, Dept. of Mathematics, University of Kaiserslautern, P.O. Box 3049, 67653
Kaiserslautern, Germany, 1994.
[37] P.M. van den Berg, R.E. Kleinman, A total variation enhanced modiﬁed gradient algorithm
for proﬁle reconstruction, Physics Express Letters, Inverse Problems, Vol. 11, pp. L5–L10,
1995.
[38] F. Bergholm, Edge focusing, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 9, 726–741,
1987.
[39] J. Big¨ un, G.H. Granlund, Optimal orientation detection of linear symmetry, Proc. First
Int. Conf. on Computer Vision (ICCV ’87, London, June 8–11, 1987), IEEE Computer
Society Press, Washington, 433–438, 1987.
[40] J. Big¨ un, G.H. Granlund, J. Wiklund, Multidimensional orientation estimation with ap
plications to texture analysis and optical ﬂow, IEEE Trans. Pattern Anal. Mach. Intell.,
Vol. 13, 775–790, 1991.
[41] M.J. Black, G. Sapiro, D. Marimont, D. Heeger, Robust anisotropic diﬀusion: connections
between robust statistics, line processing, and anisotropic diﬀusion, B. ter Haar Romeny,
L. Florack, J. Koenderink, M. Viergever (Eds.), Scalespace theory in computer vision,
Lecture Notes in Computer Science, Vol. 1252, Springer, Berlin, 323–326, 1997.
[42] A. Blake, A. Zisserman, Visual reconstruction, MIT Press, Cambridge (Mass.), 1987.
[43] L. Blanc–F´eraud, P. Charbonnier, G. Aubert, M. Barlaud, Nonlinear image processing:
modelling and fast algorithm for regularization with edge detection, Proc. IEEE Int. Conf.
Image Processing (ICIP–95, Washington, Oct. 22–25, 1995), Vol. 1, 474–477, 1995.
[44] W. Blaschke, Vorlesungen ¨ uber Diﬀerentialgeometrie II, Springer, Berlin, 1923.
[45] P. Blomgren, T.F. Chan, Color TV: total variation methods for restoration of vector valued
images, CAM Report 965, Dept. of Mathematics, Univ. of California, Los Angeles, CA
90024, U.S.A., 1996; to appear in IEEE Trans. Image Proc.
142 BIBLIOGRAPHY
[46] P. Blomgren, T.F. Chan, P. Mulet, Extensions to total variation denoising, F.T. Luk
(Ed.), Advanced signal processing: algorithms, architectures, and implementations VII,
SPIE Vol. 3162, 1997, to appear.
[47] H. Blum, A transformation for extracting new descriptors of shape, W. Wathen–Dunn
(Ed.), Models for the perception of speech and visual form, M.I.T. Press, Cambridge,
362–380, 1967.
[48] A. Bonnet, Charact´erisation des minima globaux de la fonctionelle du Mumford–Shah en
segmentation d’images, C. R. Acad. Sci. Paris, t. 321, S´erie I, 1121–1126, 1995.
[49] A. Bonnet, Sur la r´egularit´e des bords des minima de la fonctionelle de Mumford–Shah,
C. R. Acad. Sci. Paris, t. 321, S´erie I, 1275–1279, 1995.
[50] R. van den Boomgaard, Mathematical morphology: Extensions towards computer vision,
Ph.D. thesis, University of Amsterdam, The Netherlands, 1992.
[51] R. van den Boomgaard, The morphological equivalent of the Gauss convolution, Nieuw
Archief Voor Wiskunde, Vierde Serie, Deel 10, 219–236, 1992.
[52] R. van den Boomgaard, Morphology, oral presentation, Ph.D. School on Classical Scale
Space Theory, Copenhagen, May 10–13, 1996.
[53] R. van den Boomgaard, A. Smeulders, The morphological structure of images: The dif
ferential equations of morphological scalespace, IEEE Trans. Pattern Anal. Mach. Intell.,
Vol. 16, 1101–1113, 1994.
[54] K.A. Brakke, The motion of a surface by its mean curvature, Princeton University Press,
Princeton, 1978.
[55] J.H. Bramble, B.E. Hubbard, New monotone type approximations for elliptic problems,
Math. Comp., Vol. 18, 349–367, 1964.
[56] F. Brauer, J.A. Nohel, Ordinary diﬀerential equations: A ﬁrst course, Benjamin, New
York, 1967.
[57] H. Brezis, Op´erateurs maximaux monotones et semigroupes de contractions dans les es
paces de Hilbert, North Holland, Amsterdam, 1973.
[58] H. Brezis, Analyse fonctionelle, Masson, Paris, 1992.
[59] W.L. Briggs, A multigrid tutorial, SIAM, Philadelphia, 1987.
[60] R.W. Brockett, P. Maragos, Evolution equations for continuousscale morphological ﬁlter
ing, IEEE Trans. Signal Processing, Vol. 42, 3377–3386, 1994.
[61] A.M. Bruckstein, G. Sapiro, D. Shaked, Evolution of planar polygons, Int. J. Pattern
Recogn. Artif. Intell., Vol. 9, 991–1014, 1995.
[62] A.M. Bruckstein, D. Shaked, On projective invariant smoothing and evolution of planar
curves and polygons, J. Math. Imag. Vision, Vol. 7, 225–240, 1997.
[63] B. Buck, V. Macaulay (Eds.), Maximum entropy in action, Clarendon, Oxford, 1991.
[64] P.J. Burt, Fast ﬁlter transforms for image processing, Computer Graphics and Image
Processing, Vol. 16, 20–51, 1981.
[65] M. Burton, S. Acton, Target tracking using the anisotropic diﬀusion pyramid, Proc. IEEE
Southwest Symposium on Image Analysis and Interpetation (SSIAI ’96, San Antonio,
April 8–9, 1996).
BIBLIOGRAPHY 143
[66] M.A. Butt, P. Maragos, Comparison of multiscale morphology approaches: PDE imple
mented via curve evolution versus Chamfer distance transform, P. Maragos, R.W. Schafer,
M.A. Butt (Eds.), Mathematical morphology and its applications to image and signal pro
cessing, Kluwer, Boston, 31–40, 1996.
[67] L.D. Cai, Some notes on repeated averaging smoothing, J. Kittler (Ed.), Pattern Recog
nition, Lecture Notes in Computer Science, Vol. 301, Springer, Berlin, 597–605, 1988.
[68] L.D. Cai, A “small leakage” model for diﬀusion smoothing of image data, Proc. 11th
Int. Joint Conf. on Artiﬁcial Intelligence (IJCAI ’89, Detroit, Aug. 20–25, 1989), Vol. 2,
1585–1590, 1989.
[69] J. Canny, A computational approach to edge detection, IEEE Trans. Pattern Anal. Mach.
Intell., Vol. 8, 679–698, 1986.
[70] V. Cantoni, M. Ferretti, Pyramidal architectures for computer vision, Plenum Press, New
York, 1994.
[71] M.P. do Carmo, Diﬀerentialgeometrie von Kurven und Fl¨ achen, Vieweg, Braunschweig,
1983.
[72] R. Carmona, S. Zhong, Adaptive image smoothing respecting feature directions, Preprint,
Dept. of Civil Engineering, Princeton University, Princeton, NJ 08544–5263, U.S.A., 1996;
submitted to IEEE Trans. Image Proc.
[73] V. Caselles, F. Catt´e, T. Coll, F. Dibos, A geometric model for active contours in image
processing, Numer. Math., Vol. 66, 1–31, 1993.
[74] V. Caselles, B. Coll, Snakes in movement, SIAM J. Numer. Anal., Vol. 33, 2445–2456,
1996.
[75] V. Caselles, B. Coll, J.M. Morel, Scalespace versus topographic map for natural images,
B. ter Haar Romeny, L. Florack, J. Koenderink, M. Viergever (Eds.), Scalespace theory in
computer vision, Lecture Notes in Computer Science, Vol. 1252, Springer, Berlin, 29–49,
1997.
[76] V. Caselles, R. Kimmel, G. Sapiro, Geodesic active contours, Int. J. Comput. Vision., Vol.
22, 61–79, 1997.
[77] V. Caselles, R. Kimmel, G. Sapiro, C. Sbert, Minimal surfaces: a geometric three dimen
sional segmentation approach, Numer. Math., Vol. 77, 423–451, 1997.
[78] V. Caselles, C. Sbert, What is the best causal scalespace for 3D images?, SIAM J. Appl.
Math., Vol. 56, 1199–1246, 1996.
[79] F. Catt´e, Convergence of iterated aﬃne and morphological ﬁlters by nonlinear semigroup
theory, M.O. Berger, R. Deriche, I. Herlin, J. Jaﬀr´e, J.M. Morel (Eds.), ICAOS ’96:
Images, wavelets and PDEs, Lecture Notes in Control and Information Sciences, Vol. 219,
Springer, London, 125–133, 1996.
[80] F. Catt´e, F. Dibos, G. Koepﬂer, A morphological scheme for mean curvature motion and
applications to anisotropic diﬀusion and motion of level sets, SIAM J. Numer. Anal., Vol.
32, 1895–1909, 1995.
[81] F. Catt´e, P.L. Lions, J.M. Morel, T. Coll, Image selective smoothing and edge detection
by nonlinear diﬀusion, SIAM J. Numer. Anal., Vol. 29, 182–193, 1992.
[82] A. Chambolle, Partial diﬀerential equations and image processing, Proc. IEEE Int. Conf.
Image Processing (ICIP–94, Austin, Nov. 13–16, 1994), Vol. 1, IEEE Computer Society
Press, Los Alamitos, 16–20, 1994.
144 BIBLIOGRAPHY
[83] A. Chambolle, P.L. Lions, Image recovery via total variation minimization and related
problems, Numer. Math., Vol. 76, 167–188, 1997.
[84] R.H. Chan, T.F. Chan, C.K. Wong, Cosine transform based preconditioners for total varia
tion minimization problems in image processing, S. Margenov, P. Vassilevski (Eds.), Proc.
Second IMACS Symposium on Iterative Methods, IMACS Press, 311–329, 1995.
[85] T.F. Chan, G.H. Golub, P. Mulet, A nonlinear primal–dual method for totalvariation
based image restoration, M.O. Berger, R. Deriche, I. Herlin, J. Jaﬀr´e, J.M. Morel (Eds.),
ICAOS ’96: Images, wavelets and PDEs, Lecture Notes in Control and Information Sci
ences, Vol. 219, Springer, London, 241–252, 1996.
[86] T.F. Chan, C.K. Wong, Total variation blind deconvolution, CAM Report 9645, Dept.
of Mathematics, Univ. of California, Los Angeles, CA 90024, U.S.A., 1996; to appear in
IEEE Trans. Image Proc.
[87] T.F. Chan, H.M. Zhou, R.H. Chan, Continuation method for total variation minimizing
problems in image processing, F.T. Luk (Ed.), Advanced signal processing algorithms,
SPIE Vol. 2563, 314–325, 1995.
[88] P. Charbonnier, L. Blanc–F´eraud, G. Aubert, M. Barlaud, Two deterministic half
quadratic regularization algorithms for computed imaging, Proc. IEEE Int. Conf. Image
Processing (ICIP–94, Austin, Nov. 13–16, 1994), Vol. 2, IEEE Computer Society Press,
Los Alamitos, 168–172, 1994.
[89] Y. Chen, B.C. Vemuri, L. Wang, Image denoising and segmentation via nonlinear dif
fusion, Report UFCISETR97015, CISE Dept., University of Florida, Gainesville, FL
32611, U.S.A., 1997; submitted to SIAM J. Appl. Math.
[90] Y.G. Chen, Y. Giga, S. Goto, Uniqueness and existence of viscosity solutions of general
ized mean curvature ﬂow equations, J. Diﬀerential Geometry, Vol. 33, 749–786, 1991.
[91] D.L. Chopp, Computing minimal surfaces via level set curvature ﬂow, Ph.D. thesis LBL
30685, Physics Division, Lawrence Berkeley Laboratory, University of California, Berkeley,
CA 94720, U.S.A., 1991 (see also J. Comput. Phys., Vol. 106, 77–91, 1993).
[92] C.K. Chui, Wavelets: A tutorial in theory and applications, Academic Press, New York,
1992.
[93] L.D. Cohen, On active contour models and balloons, CVGIP: Image Understanding, Vol.
53, 211–218, 1991.
[94] M.A. Cohen, S. Grossberg, Neural dynamics of brightness perception: Features, bound
aries, diﬀusion and resonance, Perception and Psychophysics, Vol. 36, 428–456, 1984.
[95] T. Cohignac, F. Eve, F. Guichard, C. Lopez, J.M. Morel, Numerical analysis of the
fundamental equation of image processing, Preprint No. 9254, CEREMADE, Universit´e
Paris IX – Dauphine, Place du Mar´echal de Lattre de Tassigny, 75775 Paris Cedex 16,
France, 1992.
[96] T. Cohignac, C. Lopez, J.M. Morel, Integral and local aﬃne invariant parameter and
application to shape recognition, Proc. 12th Int. Conf. Pattern Recognition (ICPR 12,
Jerusalem, Oct. 9–13, 1994), Vol. A, IEEE Computer Society Press, Los Alamitos, 164–
168, 1994.
[97] T. Cohignac, J.M. Morel, Scalespace and aﬃne invariant recognition of occluded shapes,
L.I. Rudin, S.K. Bramble (Eds.), Investigative and trial image processing, SPIE Vol. 2567,
214–222, 1995.
BIBLIOGRAPHY 145
[98] G. Cong, S.D. Ma, Nonlinear diﬀusion for early vision, Proc. 13th Int. Conf. Pattern
Recognition (ICPR 13, Vienna, Aug. 25–30, 1996), Vol. A, 403–406, 1996.
[99] G.H. Cottet, Diﬀusion approximation on neural networks and applications for image
processing, F. Hodnett (Ed.), Proc. Sixth European Conf. on Mathematics in Industry,
Teubner, Stuttgart, 3–9, 1992.
[100] G.H. Cottet, Neural networks: Continuous approach and applications to image processing,
J. Biological Systems, Vol. 3, 1131–1139, 1995.
[101] G.H. Cottet, M. El Ayyadi, Nonlinear PDE operators with memory terms for image
processing, Proc. IEEE Int. Conf. Image Processing (ICIP–96, Lausanne, Sept. 16–19,
1996), Vol. 1, 481–483, 1996.
[102] G.H. Cottet, L. Germain, Image processing through reaction combined with nonlinear
diﬀusion, Math. Comp., Vol. 61, 659–673, 1993.
[103] M.G. Crandall, I. Ishii, P.L. Lions, User’s guide to viscosity solutions of second order
partial diﬀerential equations, Bull. Am. Math. Soc., Vol. 27, 1–67, 1992.
[104] M.G. Crandall, P.L. Lions, Convergent diﬀerence schemes for nonlinear parabolic equa
tions and mean curvature motion, Numer. Math., Vol. 75, 17–41, 1996.
[105] J. Damon, Local Morse theory for solutions to the heat equation and Gaussian blurring,
J. Diﬀerential Equations, Vol. 115, 368–401, 1995.
[106] I. Daubechies, Ten lectures on wavelets, SIAM, Philadelphia, 1992.
[107] G. David, C
1
arcs for minimizers of the Mumford–Shah functional, SIAM J. Appl. Math.,
Vol. 56, 783–888, 1996.
[108] E. De Giorgi, M. Carriero, A. Leaci, Existence theorem for a minimum problem with a
free discontinuity set, Arch. Rat. Mech. Anal., Vol. 108, 195–218, 1989.
[109] R. Deriche, Recursively implementing the Gaussian and its derivatives, V. Srinivasan, Ong
S.H., Ang Y.H. (Eds.), Proc. Second Int. Singapore Conf. on Image Proc. (Singapore, Sept.
7–11, 1992), 263–267, 1992.
[110] R. Deriche, O. Faugeras, Les EDP en traitement des images et vision par ordinateur,
Traitement du Signal, Vol. 13, No. 6, 1997.
[111] R. Deriche, P. Kornprobst, G. Aubert, Opticalﬂow estimation while preserving its dis
continuities: A variational approach, Proc. Second Asian Conf. Computer Vision (ACCV
’95, Singapore, December 5–8, 1995), Vol. 2, 290–295, 1995.
[112] F. Dibos, Analyse multi´echelle invariante par transformations projectives, C. R. Acad.
Sci. Paris, t. 320, S´erie I, 917–921, 1995.
[113] F. Dibos, Projective invariant multiscale analysis, Proc. IEEE Int. Conf. Image Processing
(ICIP–96, Lausanne, Sept. 16–19, 1996), Vol. 1, 485–488, 1996.
[114] S. Di Zenzo, A note on the gradient of a multiimage, Computer Vision, Graphics, and
Image Processing, Vol. 33, 116–125, 1986.
[115] D.C. Dobson, F. Santosa, An image enhancement technique for electrical impedance to
mography, Inverse Problems, Vol. 10, 317–334, 1994.
[116] D.C. Dobson, F. Santosa, Recovery of blocky images from noisy and blurred data, SIAM
J. Appl. Math., Vol. 56, 1181–1198, 1996.
146 BIBLIOGRAPHY
[117] D. Dobson, O. Scherzer, Analysis of regularized total variation penalty methods for de
noising, Inverse Problems, Vol. 12, 601–617, 1996.
[118] D.C. Dobson, C.R. Vogel, Convergence of an iterative method for total variation denois
ing, Preprint, Dept. of Mathematical Sciences, Montana State University, Bozeman, MT
597170240, U.S.A., 1995; SIAM J. Numer. Anal., to appear.
[119] L. Dorst, R. van den Boomgaard, Morphological signal processing and the slope transform,
Signal Processing, Vol. 38, 79–98, 1994.
[120] J. Douglas, J.E. Gunn, A general formulation of alternating direction methods. Part I.
Parabolic and hyperbolic problems, Numer. Math., Vol. 6, 428–453, 1964.
[121] J. Douglas, B.F. Jones, On predictor–corrector methods for nonlinear parabolic diﬀerential
equations, J. Soc. Ind. Appl. Math., Vol. 11, 195–204, 1963.
[122] I.S. Duﬀ, A.M. Erisman, J.K. Reid, Direct methods for sparse systems, Clarendon, Oxford,
1986.
[123] S.K. Dzhu Magazieva, Numerical study of a partial diﬀerential equation, U.S.S.R. Comput.
Maths. Math. Phys., Vol. 23, No. 4, 45–49, 1983.
[124] D. Eberly, A diﬀerential geometric approach to anisotropic diﬀusion, B.M. ter Haar
Romeny (Ed.), Geometrydriven diﬀusion in computer vision, Kluwer, Dordrecht, 371–
391, 1994.
[125] S. Eidelman, W. Grossmann, A. Friedman, Nonlinear signal processing using integration
of ﬂuid dynamics equations, A.G. Tescher (Ed.): Applications of digital image processing
XIV, SPIE Vol. 1567, 439–450, 1991.
[126] A.I. ElFallah, G.E. Ford, Nonlinear adaptive image ﬁltering based on inhomogeneous
diﬀusion and diﬀerential geometry, S.A. Rajala, R.L. Stevenson (Eds.), Image and video
processing II, SPIE Vol. 2182, 49–63, 1994.
[127] A.I. ElFallah, G.E. Ford, The evolution of mean curvature in image ﬁltering, Proc. IEEE
Int. Conf. Image Processing (ICIP–94, Austin, Nov. 13–16, 1994), Vol. 1, IEEE Computer
Society Press, Los Alamitos, 298–302, 1994.
[128] H.W. Engl, M. Hanke, A. Neubauer, Regularization of inverse problems, Kluwer, Dor
drecht, 1996.
[129] L.C. Evans, J. Spruck, Motion of level sets by mean curvature I., J. Diﬀerential Geometry,
Vol. 33, 635–681, 1991.
[130] O. Faugeras, Sur l’´evolution de courbes simples du plan projectif r´eel, C. R. Acad. Sci.
Paris, t. 317, S´erie I, 565–570, 1993.
[131] O. Faugeras, Cartan’s moving frame method and its application to the geometry and evolu
tion of curves in the Euclidean, aﬃne and projective planes, J.L. Mundy, A. Zisserman, D.
Forsyth (Eds.), Applications of invariance in computer vision, Lecture Notes in Computer
Science, Vol. 825, Springer, Berlin, 11–46, 1994.
[132] O. Faugeras, R. Keriven, Scalespaces and aﬃne curvature, W. Chenke, R. Mohr (Eds.),
Proc. Workshop on Geometrical Modelling and Invariants for Computer Vision (Xi’an,
April 27–30, 1995), Xidian University Press, Xi’an, 1995.
[133] O. Faugeras, R. Keriven, On projective plane curve evolution, M.O. Berger, R. Deriche,
I. Herlin, J. Jaﬀr´e, J.M. Morel (Eds.), ICAOS ’96: Images, wavelets and PDEs, Lecture
Notes in Control and Information Sciences, Vol. 219, Springer, London, 66–73, 1996.
BIBLIOGRAPHY 147
[134] O. Faugeras, R. Keriven, Level set methods and the stereo problem, B. ter Haar Romeny,
L. Florack, J. Koenderink, M. Viergever (Eds.), Scalespace theory in computer vision,
Lecture Notes in Computer Science, Vol. 1252, Springer, Berlin, 272–283, 1997.
[135] S. Fejes, A. Rosenfeld, Discrete active models and applications, Pattern Recognition, Vol.
30, 817–835, 1997.
[136] P. Fiddelaers, E.J. Pauwels, L.J. Van Gool, Geometrydriven curve evolution, J.O. Ek
lundh (Ed.), Computer Vision – ECCV ’94, Lecture Notes in Computer Science, Vol. 800,
Springer, Berlin, 427–432, 1994.
[137] B. Fischl, E.L. Schwartz, Adaptive nonlinear ﬁltering for nonlinear diﬀusion approxima
tion in image processing, Proc. 13th Int. Conf. Pattern Recognition (ICPR 13, Vienna,
Aug. 25–30, 1996), Vol. D, 276–280, 1996.
[138] L.M.J. Florack, Data, models, and images, Proc. 1996 IEEE Int. Conf. Image Processing
(ICIP–96, Lausanne, Sept. 16–19, 1996), Vol. I, 469–472, 1996.
[139] L.M.J. Florack, Image structure, Kluwer, Dordrecht, 1997.
[140] L.M.J. Florack, B.M. ter Haar Romeny, J.J. Koenderink, M.A. Viergever, Scale and the
diﬀerential structure of images, Image Vision Comput., Vol. 10, 376–388, 1992.
[141] L.M.J. Florack, B.M. ter Haar Romeny, J.J. Koenderink, M.A. Viergever, The Gaussian
scalespace paradigm and the multiscale local jet, Int. J. Comput. Vision, Vol. 18, 61–75,
1996.
[142] L. Florack, R. Maas, Pseudolinear scalespace theory: towards the integration of linear
and morphological scalespace paradigms, Technical Report UUCS199725, Dept. of Com
puter Science, Padualaan 14, 3584 CH Utrecht, The Netherlands, 1997.
[143] L.M.J. Florack, A.H. Salden, B.M. ter Haar Romeny, J.J. Koenderink, M.A. Viergever,
Nonlinear scalespace, Image Vision Comput., Vol. 13, 279–294, 1995.
[144] G.E. Ford, R.R. Estes, H. Chen, Scalespace analysis for image sampling and interpola
tion, Proc. IEEE Int. Conf. Acoustics, Speech and Signal Processing (ICASSP–92, San
Francisco, March 23–26, 1992), Vol. 3, 165–168, 1992.
[145] M.A. F¨orstner, E. G¨ ulch, A fast operator for detection and precise location of distinct
points, corners and centres of circular features, Proc. ISPRS Intercommission Conf. on
Fast Processing of Photogrammetric Data (Interlaken, June 2–4, 1987), 281–305, 1987.
[146] T. Franke, H. Neumann, R. Seydel, Anisotropic diﬀusion based on mean curvature motion:
A computational study, B. J¨ ahne, P. Geißler, H. Haußecker, F. Hering (Eds.), Mustererken
nung 1996, Springer, Berlin, 47–54, 1996.
[147] D.S. Fritsch, A medial description of greyscale image structure by gradientlimited dif
fusion, R.A. Robb (Ed.), Visualization in biomedical computing ’92, SPIE Vol. 1808,
105–117, 1992.
[148] J. Fr¨ ohlich, J. Weickert, Image processing using a wavelet algorithm for nonlinear diﬀu
sion, Report No. 104, Laboratory of Technomathematics, University of Kaiserslautern,
P.O. Box 3049, 67653 Kaiserslautern, Germany, 1994.
[149] D. Gabor, Information theory in electron microscopy, Laboratory Investigation, Vol. 14,
801–807, 1965.
[150] M. Gage, On areapreserving evolution equation for plane curves, Contemp. Math., Vol.
23, 51–62, 1986.
148 BIBLIOGRAPHY
[151] M. Gage, R.S. Hamilton, The heat equation shrinking convex plane curves, J. Diﬀerential
Geometry, Vol. 23, 69–96, 1986.
[152] D. Geiger, A. Yuille, A common framework for image segmentation, Int. J. Comput.
Vision, Vol. 6, 227–243, 1991.
[153] D. Geman, G. Reynolds, Constrained restoration and the recovery of discontinuities, IEEE
Trans. Pattern Anal. Mach. Intell., Vol. 14, 367–383, 1992.
[154] S. Geman, D. Geman, Stochastic relaxation, Gibbs distributions, and the Bayesian restora
tion of images, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 6, 721–741, 1984.
[155] G. Gerig, O. K¨ ubler, R. Kikinis, F.A. Jolesz, Nonlinear anisotropic ﬁltering of MRI data,
IEEE Trans. Medical Imaging, Vol. 11, 221–232, 1992.
[156] G. Gerig, G. Sz´ekely, Finding line structures by multivalued nonlinear diﬀusion of feature
maps, Report, Communication Technology Laboratory, Image Science Division, ETH–
Zentrum, 8092 Z¨ urich, Switzerland, 1993.
[157] G. Gerig, G. Sz´ekely, G. Israel, M. Berger, Detection and characterization of unsharp blobs
by curve evolution, Y. Bizais, C. Barillot, R. Di Paola (Eds.), Information Processing in
Medical Imaging, Kluwer, Dordrecht, 165–176, 1995.
[158] T. Gijbels, P. Six, L. Van Gool, F. Catthoor, H. De Man, A. Oosterlinck, A VLSI
architecture for parallel nonlinear diﬀusion with applications in vision, Proc. IEEE Work
shop on VLSI Signal Processing, Vol. 7, 398–407, 1994.
[159] D. Giuiliani, Directed diﬀusion equations and edge detection, master thesis, Dept. of
Mathematics, University of Kaiserslautern, P.O. Box 3049, 67653 Kaiserslautern, Ger
many, 1991.
[160] V. van Gogh, Starry night, SaintR´emy, 1889. New York, The Museum of Modern Art.
[161] V. van Gogh, Selfportrait, SaintR´emy, 1889. Paris, Muse´e d’Orsay.
[162] V. van Gogh, Road with cypress and star, AuverssurOise, 1890. Otterlo, Rijksmuseum
Kr¨ oller–M¨ uller.
[163] R.C. Gonzalez, P. Wintz, Digital image processing, Addison–Wesley, Reading, 1987.
[164] R. Gorenﬂo, Conservative diﬀerence schemes for diﬀusion problems, Preprint No. 39, FB
Mathematik, FU Berlin, Arnimallee 2–6, W1000 Berlin 33, Germany, 1977.
[165] A. Goshtasby, On edge focusing, Image Vision Comput., Vol. 12, 247–256, 1994.
[166] A.R. Gourlay, Implicit convolution, Image Vision Comput., Vol. 3, 15–23, 1985.
[167] R. Graham, Snow removal: A noisestripping process for TV signals, IRE Trans. Infor
mation Theory, Vol. 8, 129–144, 1962.
[168] G.H. Granlund, H. Knutsson, Signal processing for computer vision, Kluwer, Dordrecht,
1995.
[169] M. Grayson, The heat equation shrinks embedded plane curves to round points, J. Diﬀer
ential Geometry, Vol. 26, 285–314, 1987.
[170] D. Greenspan, P.C. Jain, On nonnegative diﬀerence analogues of elliptic diﬀerential equa
tions, J. Franklin Inst., Vol. 279, 360–365, 1965.
[171] L.D. Griﬃn, Scaleimprecision space, Image Vision Comput., Vol. 15, 369–398, 1997.
[172] L.D. Griﬃn, A.C.F. Colchester, G.P. Robinson, Scale and segmentation of greylevel im
ages using maximum gradient paths, Image Vision Comput., Vol. 10, 389–402, 1992.
BIBLIOGRAPHY 149
[173] F. Guichard, J.M. Morel, Partial diﬀerential equations and image iterative ﬁltering,
Preprint No. 9535, CEREMADE, Universit´e Paris IX – Dauphine, Place du Mar´echal
de Lattre de Tassigny, 75775 Paris Cedex 16, France, 1995; to appear in I.S. Duﬀ, G.A.
Watson (Eds.), The state of the art in numerical analysis, Oxford University Press.
[174] B.M. ter Haar Romeny (Ed.), Geometrydriven diﬀusion in computer vision, Kluwer,
Dordrecht, 1994.
[175] B.M. ter Haar Romeny, Applications of scalespace theory, J. Sporring, M. Nielsen, L.
Florack, P. Johansen (Eds.), Gaussian scalespace theory, Kluwer, Dordrecht, 3–19, 1997.
[176] B.M. ter Haar Romeny, Frontend vision and multiscale image analysis: Introduction to
scalespace theory, Kluwer, Dordrecht, in preparation.
[177] B.M. ter Haar Romeny, L.M.J. Florack, M. de Swart, J. Wilting, M.A. Viergever, Deblur
ring Gaussian blur, F.L. Bookstein, J.S. Duncan, N. Lange, D.C. Wilson (Eds.), Mathe
matical methods in medical imaging III, SPIE Vol. 2299, 139–148, 1994.
[178] B.M. ter Haar Romeny, W.J. Niessen, J. Wilting, L.M.J. Florack, Diﬀerential structure
of images: accuracy of representation, Proc. IEEE Int. Conf. Image Processing (ICIP–94,
Austin, Nov. 13–16, 1994), Vol. 1, IEEE Computer Society Press, Los Alamitos, 21–25,
1994.
[179] W. Hackbusch, Multigrid methods and applications, Springer, New York, 1985.
[180] W. Hahn, Theorie und Anwendung der direkten Methode von Ljapunov, Springer, Berlin,
1959.
[181] R.M. Haralick, S.R. Sternberg, X. Zhuang, Image analysis using mathematical morphology,
IEEE Trans. Pattern Anal. Mach. Intell., Vol. 9, 532–550, 1987.
[182] C.G. Harris, M. Stephens, A combined corner and edge detector, Proc. Fouth Alvey Vision
Conf. (AVC 88, Manchester, Aug. 31 – Sept. 2, 1988), 147–152, 1988.
[183] A. Harten, B. Enquist, S. Osher, S. Chakravarthy, Uniformly high order accurrate essen
tially nonoscillatory schemes, III, J. Comput. Phys., Vol. 71, 231–303, 1987.
[184] H.J.A.M. Heijmans, Mathematical morphology: A modern approach in image processing
based on algebra and geometry, SIAM Review, Vol. 37, 1–36, 1995.
[185] G. Hellwig, Partial diﬀerential equations, Teubner, Stuttgart, 1977.
[186] G. Hewer, C. Kenney, B.S. Manjunath, Variational image segmentation using boundary
functions, Report No. 9607, ECE Dept., University of California, Santa Barbara, CA
93106, U.S.A., 1996; submitted to IEEE Trans. Image Proc.
[187] K. H¨ ollig, Existence of inﬁnitely many solutions for a forward–backward heat equation,
Trans. Amer. Math. Soc., Vol. 278, 299–316, 1983.
[188] G. Huisken, Flow by mean curvature of convex surfaces into spheres, J. Diﬀerential Geo
metry, Vol. 20, 237–266, 1984.
[189] R.A. Hummel, Representations based on zerocrossings in scale space, Proc. IEEE Comp.
Soc. Conf. Computer Vision and Pattern Recognition (CVPR ’86, Miami Beach, June
22–26, 1986), IEEE Computer Society Press, Washington, 204–209, 1986.
[190] R. Hummel, R. Moniot, Reconstructions from zerocrossings in scale space, IEEE Trans.
Acoustics, Speech and Signal Processing, Vol. 37, 2111–2130, 1989.
[191] T. Iijima, Basic theory of pattern observation, Papers of Technical Group on Automata
and Automatic Control, IECE, Japan, Dec. 1959 (in Japanese).
150 BIBLIOGRAPHY
[192] T. Iijima, Basic theory on normalization of a pattern (for the case of a typical one
dimensional pattern), Bulletin of the Electrotechnical Laboratory, Vol. 26, 368–388, 1962
(in Japanese).
[193] T. Iijima, Observation theory of twodimensional visual patterns, Papers of Technical
Group on Automata and Automatic Control, IECE, Japan, Oct. 1962 (in Japanese).
[194] T. Iijima, Basic theory on normalization of twodimensional visual pattern, Studies on
Information and Control, Pattern Recognition Issue, IECE, Japan, No. 1, 15–22, 1963 (in
Japanese).
[195] T. Iijima, Theory of pattern recognition, Electronics and Communications in Japan, 123–
134, Nov. 1963 (in English).
[196] T. Iijima, Basic equation of ﬁgure and observational transformation, Systems, Computers,
Controls, Vol. 2, No. 4, 70–77, 1971 (in English).
[197] T. Iijima, Pattern recognition, Coronasha, 1973 (in Japanese).
[198] T. Iijima, Theory of pattern recognition, Series of Basic Information Technology, Vol. 6,
Morishita Publishing, 1989 (in Japanese).
[199] T. Iijima, H. Genchi, K. Mori, A theoretical study of pattern identiﬁcation by matching
method, Proc. First USA–Japan Computer Conference (Tokyo, Oct. 3–5, 1972), 42–48,
1972 (in English).
[200] T. Iijima, H. Genchi, K. Mori, A theory of character recognition by pattern matching
method, Proc. First Int. Joint Conf. Pattern Recognition (IJCPR 1, Washington, Oct.
1973), 50–56, 1973 (in English).
[201] R. Illner, Global existence for twovelocity models of the Boltzmann equation, Math. Meth.
Appl. Sci., Vol. 1, 187–193, 1979.
[202] R. Illner, H. Neunzert, Relative entropy maximization and directed diﬀusion equations,
Math. Meth. Appl. Sci., Vol. 16, 545–554, 1993.
[203] R. Illner, J. Tie, On directed diﬀusion with measurable background, Math. Meth. Appl.
Sci., Vol. 16, 681–690, 1993.
[204] K. Ito, K. Kunisch, Augmented Lagrangian methods for nonsmooth, convex optimization
in Hilbert spaces, Preprint No. 409/1994, Fachbereich Mathematik (3), Technische Uni
versit¨ at Berlin, Straße des 17. Juni 136, 10623 Berlin, Germany, 1994.
[205] K. Ito, K. Kunisch, An active set strategy for image restoration based on the augmented
Lagrangian formulation, Preprint No. 410/1994, Fachbereich Mathematik (3), Technische
Universit¨ at Berlin, Straße des 17. Juni 136, 10623 Berlin, Germany, 1994.
[206] P.T. Jackway, Morphological scalespace, Proc. 11th Int. Conf. Pattern Recognition (ICPR
11, The Hague, Aug. 30 – Sept. 3, 1992), Vol. C, 252–255, 1992.
[207] P.T. Jackway, M. Deriche, Scalespace properties of the multiscale morphological dilation–
erosion, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 18, 38–51, 1996.
[208] M. J¨ agersand, Saliency maps and attention selection in scale and spatial coordinates: an
information theoretic approach, Proc. Fifth Int. Conf. on Computer Vision (ICCV ’95,
Cambridge, June 20–23, 1995), IEEE Computer Society Press, Los Alamitos, 195–202,
1995.
[209] B. J¨ ahne, Spatiotemporal image processing, Lecture Notes in Computer Science, Vol. 751,
Springer, Berlin, 1993.
BIBLIOGRAPHY 151
[210] P. Johansen, On the classiﬁcation of toppoints in scalespace, J. Math. Imag. Vision, Vol.
4, 57–67, 1994.
[211] P. Johansen (Ed.), Proceedings of the workshop on Gaussian scalespace theory, Report
DIKU 96/19, Dept. of Computer Science, University of Copenhagen, Universitetsparken
1, 2100 Copenhagen, Denmark, 1996.
[212] P. Johansen, S. Skelboe, K. Grue, J.D. Andersen, Representing signals by their toppoints
in scale space, Proc. Eighth Int. Conf. Pattern Recognition (ICPR 8, Paris, Oct. 27–31,
1986), 215–217, 1986.
[213] F. John, Partial diﬀerential equations, Springer, New York, 1978.
[214] J.M. Jolion, A. Rosenfeld, A pyramid framework for early vision, Kluwer, Dordrecht,
1994.
[215] D.G. Jones, J. Malik, A computational framework for determining stereo correspondence
from a set of linear spatial ﬁlters, G. Sandini (Ed.), Computer vision – ECCV ’92, Lecture
Notes in Computer Science, Vol. 588, Springer, Berlin, 395–410, 1992.
[216] J. Kaˇcur, K. Mikula, Solution of nonlinear diﬀusion appearing in image smoothing and
edge detection, Appl. Num. Math., Vol. 17, 47–59, 1995.
[217] J. Kaˇcur, K. Mikula, Slowed anisotropic diﬀusion, B. ter Haar Romeny, L. Florack, J.
Koenderink, M. Viergever (Eds.), Scalespace theory in computer vision, Lecture Notes
in Computer Science, Vol. 1252, Springer, Berlin, 357–360, 1997.
[218] H.W. Kaﬁtz, Morphologische Diﬀerentialgleichungen in der Bildverarbeitung, diploma
thesis, Dept. of Mathematics, University of Kaiserslautern, P.O. Box 3049, 67653 Kaisers
lautern, Germany, 1996.
[219] G. Kanizsa, Organization in vision, Praeger, New York, 1979.
[220] M. Kass, A. Witkin, Analyzing oriented patterns, Computer Vision, Graphics, and Image
Processing, Vol. 37, 362–385, 1987.
[221] M. Kass, A. Witkin, D. Terzopoulos, Snakes: Active contour models, Int. J. Comput.
Vision, Vol. 1, 321–331, 1988.
[222] B. Kawohl, N. Kutev, Maximum and comparison principles for anisotropic diﬀusion,
Preprint, Mathematical Institute, University of Cologne, 50923 Cologne, Germany, 1997.
[223] J.G. Kemeny, J.L. Snell, Finite Markov chains, D. Van Nostrand, Princeton, 1960.
[224] S. Kichenassamy, Nonlinear diﬀusions and hyperbolic smoothing for edge enhancement,
M.O. Berger, R. Deriche, I. Herlin, J. Jaﬀr´e, J.M. Morel (Eds.), ICAOS ’96: Im
ages, Wavelets and PDEs, Lecture Notes in Control and Information Sciences, Vol. 219,
Springer, London, 119–124, 1996.
[225] S. Kichenassamy, The Perona–Malik paradox, SIAM J. Appl. Math., to appear.
[226] S. Kichenassamy, A. Kumar, P. Olver, A. Tannenbaum, A. Yezzi, Conformal curvature
ﬂows: from phase transitions to active vision, Arch. Rat. Mech. Anal., Vol. 134, 275–301,
1996.
[227] B.B. Kimia, K. Siddiqi, Geometric heat equation and nonlinear diﬀusion of shapes and
images, Computer Vision and Image Understanding, Vol. 64, 305–322, 1996.
[228] B.B. Kimia, A. Tannenbaum, S.W. Zucker, Toward a computational theory of shape: An
overview, O. Faugeras (Ed.), Computer Vision – ECCV ’90, Lecture Notes in Computer
Science, Vol. 427, Springer, Berlin, 402–407, 1990.
152 BIBLIOGRAPHY
[229] B.B. Kimia, A. Tannenbaum, S.W. Zucker, On the evolution of curves via a function of
curvature. I. The classical case, J. Math. Anal. Appl., Vol. 163, 438–458, 1992.
[230] B. Kimia, A. Tannenbaum, S.Zucker, Shapes, shocks, and deformations I: The components
of twodimensional shape and the reaction–diﬀusion space, Int. J. Comput. Vision, Vol.
15, 189–224, 1995.
[231] R. Kimmel, Aﬃne diﬀerential signatures for gray level images of planar shape, Proc. 13th
Int. Conf. Pattern Recognition (ICPR 13, Vienna, Aug. 25–30, 1996), Vol. A, 45–49, 1996.
[232] R. Kimmel, Intrinsic scalespace for images on surfaces: The geodesic curvature ﬂow, B.
ter Haar Romeny, L. Florack, J. Koenderink, M. Viergever (Eds.), Scalespace theory in
computer vision, Lecture Notes in Computer Science, Vol. 1252, Springer, Berlin, 212–223,
1997.
[233] R. Kimmel, N. Kiryati, A.M. Bruckstein, Analyzing and synthesizing images by evolving
curves with the OsherSethian method, Int. J. Comput. Vision, Vol. 24, 37–56, 1997.
[234] R. Kimmel, N. Sochen, R. Malladi, On the geometry of texture, Preprint LBNL39640,
Lawrence Berkeley National Laboratory, University of California, Berkeley, CA 94720,
U.S.A., 1996.
[235] R. Kimmel, N. Sochen, R. Malladi, Images as embedding maps and minimal surfaces:
Movies, color, and volumetric medical images, Proc. IEEE Comp. Soc. Conf. Computer
Vision and Pattern Recognition (CVPR ’97, San Juan, June 17–19, 1997), IEEE Computer
Society Press, Los Alamitos, 350–355, 1997.
[236] G. Kin, M. Sato, Scale space ﬁltering on spherical pattern, Proc. 11th Int. Conf. Pattern
Recognition (ICPR 11, The Hague, Aug. 30 – Sept. 3, 1992), Vol. C, 638–641, 1992.
[237] H.E. Knutsson, R. Wilson, G.H. Granlund, Anisotropic nonstationary image estimation
and its applications: Part I – Restoration of noisy images, IEEE Trans. Comm., Vol. 31,
388–397, 1983.
[238] M. Kocan, Approximation of viscosity solutions of elliptic partial diﬀerential equations on
minimal grids, Numer. Math., Vol. 72, 73–92, 1996.
[239] G. Koepﬂer, C. Lopez, J.M. Morel, A multiscale algorithm for image segmentation by
variational method, SIAM J. Numer. Anal., Vol. 31, 282–299, 1994.
[240] J.J. Koenderink, The structure of images, Biol. Cybern., Vol. 50, 363–370, 1984.
[241] J.J. Koenderink, Scaletime, Biol. Cybern., Vol. 58, 159–162, 1988.
[242] J.J. Koenderink, A.J. van Doorn, Representation of local geometry in the visual system,
Biol. Cybern., Vol. 55, 367–375, 1987.
[243] H.P. Kramer, J.B. Bruckner, Iterations of a nonlinear transformation for enhancement
of digital images, Pattern Recognition, Vol. 7, 53–58, 1975.
[244] K. Krissian, G. Malandain, N. Ayache, Directional anisotropic diﬀusion applied to seg
mentation of vessels in 3D images, B. ter Haar Romeny, L. Florack, J. Koenderink, M.
Viergever (Eds.), Scalespace theory in computer vision, Lecture Notes in Computer Sci
ence, Vol. 1252, Springer, Berlin, 345–348, 1997.
[245] A. Kumar, A.R. Tannenbaum, G.J. Balas, Optic ﬂow: a curve evolution approach, IEEE
Trans. Image Proc., Vol. 5, 598–610, 1996.
[246] O.A. Ladyˇzenskaja, V.A. Solonnikov, N.N. Ural’ceva, Linear and quasilinear equations of
parabolic type, American Mathematical Society, Providence, 1968.
BIBLIOGRAPHY 153
[247] D.S. Lalush, The relationship between diﬀusion ﬁltering and Gibbs smoothing, personal
note, Dept. of Biomedical Engineering, University of North Carolina, Chapel Hill, NC
27599, U.S.A., 1994.
[248] C. Lamberti, M. Sitta, F. Sgallari, Improvements to the anisotropic diﬀusion model for
2D echo image processing, Proc. Annual Int. Conf. of the IEEE Engineering in Medicine
and Biology Society, Vol. 14, 1872–1873, 1992.
[249] A. Leaci, S. Solimini, Variational problems with a free discontinuity set, B.M. ter Haar
Romeny (Ed.), Geometrydriven diﬀusion in computer vision, Kluwer, Dordrecht, 147–
154, 1994.
[250] A. Lev, S. Zucker, A. Rosenfeld, Iterative enhancement of noisy images, IEEE Trans.
Systems, Man and Cybernetics, Vol. 7, 435–442, 1977.
[251] S.Z. Li, On discontinuityadaptive smoothness priors in computer vision, IEEE Trans.
Pattern Anal. Mach. Intell., Vol. 17, 576–586, 1995.
[252] X. Li, T. Chen, Nonlinear diﬀusion with multiple edginess thresholds, Pattern Recognition,
Vol. 27, 1029–1037, 1994.
[253] Y. Li, F. Santosa, A computational algorithm for minimizing total variation in image
enhancement, IEEE Trans. Image Proc., Vol. 5, 987–995, 1996.
[254] L.M. Lifshitz, S.M. Pizer, A multiresolution hierarchical approach to image segmentation
based on intensity extrema, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 12, 529–540,
1990.
[255] T. Lindeberg, Scalespace for discrete signals, IEEE Trans. Pattern Anal. Mach. Intell.,
Vol. 12, 234–254, 1990.
[256] T. Lindeberg, Scalespace theory in computer vision, Kluwer, Boston, 1994.
[257] T. Lindeberg, On the axiomatic foundations of linear scalespace, J. Sporring, M. Nielsen,
L. Florack, P. Johansen (Eds.), Gaussian scalespace theory, Kluwer, Dordrecht, 75–97,
1997.
[258] T. Lindeberg, D. Fagerstr¨ om, Scalespace with causal time direction, B. Buxton, R. Cipolla
(Eds.), Computer vision – ECCV ’96, Volume I, Lecture Notes in Computer Science, Vol.
1064, Springer, Berlin, 229–240, 1996.
[259] T. Lindeberg, J. G˚arding, Shapeadapted smoothing in estimation of 3D depth cues from
aﬃne distortions of local 2D brightness structure, Image Vision Comput., Vol. 15, 415–
434, 1997.
[260] M. Lindenbaum, M. Fischer, A. Bruckstein, On Gabor’s contribution to image enhance
ment, Pattern Recognition, Vol. 27, 1–8, 1994.
[261] P.L. Lions, Axiomatic derivation of image processing models, Math. Mod. Meth. Appl.
Sci., Vol. 4, 467–475, 1994.
[262] P.L. Lions, S. Osher, L.I. Rudin, Denoising and deblurring images using constrained
nonlinear partial diﬀerential equations, Preprint, Cognitech, Inc., 2800 – 28 St., Suite 101,
Santa Monica, CA 90405, U.S.A., 1993.
[263] P. Lobel, L. BlancF´eraud, C. Pichot, M. Barlaud, A new regularization scheme for inverse
scattering, Inverse Problems, Vol. 13, 403–410, 1997.
[264] M.H. Loew, J. Rosenman, J. Chen, A clinical tool for enhancement of portal images, M.H.
Loew (Ed.), Image processing, SPIE Vol. 2167, 543–550, 1994.
154 BIBLIOGRAPHY
[265] C. Lopez, J.M. Morel, Axiomatization of shape analysis and applications to texture hyper
discrimination, Proc. IEEE Comp. Soc. Conf. Computer Vision and Pattern Recognition
(CVPR ’93, New York, June 15–17, 1993), IEEE Computer Society Press, Los Alamitos,
646–647, 1993.
[266] A.K. Louis, Inverse und schlecht gestellte Probleme, Teubner, Stuttgart, 1989.
[267] L. Lucido, R. Deriche, L. Alvarez, V. Rigaud, Sur quelques sch´emas num´eriques de
r´esolution d’´equations aux d´eriv´ees partielles pour le traitement d’images, Report No.
3192, ROBOTVIS, INRIA, 2004 route des Lucioles, BP 93, 06902 Sophia–Antipolis Cedex,
France, 1997.
[268] L. Lucido, J. Opderbecke, V. Rigaud, R. Deriche, Z. Zhang, An integrated multiscale
approach for terrain referenced underwater navigation, Proc. IEEE Int. Conf. Image Pro
cessing (ICIP–96, Lausanne, Sept. 16–19, 1996), Vol. 2, 633–636, 1996.
[269] A. Lukschin, H. Neunzert, J. Struckmeier, Interim report of the project DPH 6473/91 –
Coupling of NavierStokes and Boltzmann regions, Internal Report, Laboratory of Tech
nomathematics, Kaiserslautern, 1993.
[270] D.S. Luo, M.A. King, S. Glick, Local geometry variable conductance diﬀusion for post
reconstruction ﬁltering, IEEE Trans. Nuclear Sci., Vol. 41, 2800–2806, 1994.
[271] R. Malladi, R. Kimmel, D. Adalsteinsson, G. Sapiro, V. Caselles, J.A. Sethian, A geometric
approach to segmentation and analysis of 3D medical images, Proc. IEEE Workshop on
Mathematical Models in Biomedical Image Analysis (MMBIA ’96, San Francisco, June
21–22, 1996), IEEE Computer Society Press, Los Alamitos, 244–252, 1996.
[272] R. Malladi, J.A. Sethian, Image processing: ﬂows under min/max curvature and mean
curvature, Graphical Models and Image Processing, Vol. 58, 127–141, 1996.
[273] R. Malladi, J.A. Sethian, B.C. Vemuri, Shape modeling with front propagation: A level set
approach, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 17, 158–175, 1995.
[274] P. Maragos, A representation theory for morphological image and signal processing, IEEE
Trans. Pattern Anal. Mach. Intell., Vol. 11, 586–599, 1989.
[275] P. Maragos, Morphological systems: Slope transforms and max–min diﬀerence and diﬀer
ential equations, Signal Processing, Vol. 38, 57–77, 1994.
[276] P. Maragos, Diﬀerential morphology and image processing, IEEE Trans. Image Proc., Vol.
5, 922–937, 1996.
[277] G.I. Marchuk, Splitting and alternating direction methods, P.G. Ciarlet, J.L. Lions (Eds.),
Handbook of numerical analysis, Vol. I, 197–462, 1990.
[278] D. Marr, E. Hildreth, Theory of edge detection, Proc. R. Soc. London Ser. B, Vol. 207,
187–217, 1980.
[279] J.L. Marroquin, Deterministic interactive particle models for image processing and com
puter graphics, CVGIP: Graphical Models and Image Processing, Vol. 55, 408–417, 1993.
[280] G. Matheron, Random sets and integral geometry, Wiley, New York, 1975.
[281] T. McInerney, D. Terzopoulos, Topologically adaptable snakes, Proc. Fifth Int. Conf. on
Computer Vision (ICCV ’95, Cambridge, June 20–23, 1995), IEEE Computer Society
Press, Los Alamitos, 840–845, 1995.
[282] T. McInerney, D. Terzopoulos, Deformable models in medical image analysis: A survey,
Medical Image Analysis, Vol. 1, 91–108, 1996.
BIBLIOGRAPHY 155
[283] T. McInerney, D. Terzopoulos, Medical image segmentation using topologically adaptable
surfaces, J. Troccaz, E. Grimson, R. M¨ osges (Eds.), CVRMed–MRCAS ’97, Lecture Notes
in Computer Science, Vol. 1205, Springer, Berlin, 23–32, 1997.
[284] T. Meis, U. Marcowitz, Numerische Behandlung partieller Diﬀerentialgleichungen,
Springer, Berlin, 1978.
[285] B. Merriman, J. Bence, S. Osher, Motion of multiple junctions: a level set approach, J.
Comput. Phys., Vol. 112, 334–363, 1994.
[286] A.R. Mitchell, D.F. Griﬃths, The ﬁnite diﬀerence method in partial diﬀerential equations,
Wiley, Chichester, 1980.
[287] L. Moisan, Analyse multi´echelle de ﬁlms pour la reconstruction du relief, C. R. Acad. Sci.
Paris, t. 320, S´erie I, 279–284, 1995.
[288] L. Moisan, Perspective invariant movie analysis for depth recovery, L.I. Rudin, S.K. Bram
ble (Eds.), Investigative and trial image processing, SPIE Vol. 2567, 84–94, 1995.
[289] L. Moisan, Aﬃne plane curve evolution: a fully consistent scheme, Preprint No. 9628,
CEREMADE, Universit´e Paris IX – Dauphine, Place du Mar´echal de Lattre de Tassigny,
75775 Paris Cedex 16, France, 1996.
[290] F. Mokhtarian, Multiscale contour segmentation, B. ter Haar Romeny, L. Florack, J.
Koenderink, M. Viergever (Eds.), Scalespace theory in computer vision, Lecture Notes
in Computer Science, Vol. 1252, Springer, Berlin, 296–307, 1997.
[291] F. Mokhtarian, A.K. Mackworth, A theory of multiscale, curvaturebased shape represen
tation for planar curves, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 14, 789–805, 1992.
[292] J.M. Morel, S. Solimini, Variational methods in image segmentation, Birkh¨auser, Boston,
1994.
[293] K.W. Morton, D.F. Mayers, Numerical solution of partial diﬀerential equations, Cam
bridge University Press, Cambridge, 1994.
[294] T.S. Motzkin, W. Wasow, On the approximation of linear elliptic diﬀerential equations
by diﬀerence equations with positive coeﬃcients, J. Mathematics and Physics, Vol. 31,
253–259, 1952.
[295] D. Mumford, J. Shah, Boundary detection by minimizing functionals, I, Proc. IEEE Comp.
Soc. Conf. Computer Vision and Pattern Recognition (CVPR ’85, San Francisco, June
19–23, 1985), IEEE Computer Society Press, Washington, 22–26, 1985.
[296] D. Mumford, J. Shah, Optimal approximations by piecewise smooth functions and associ
ated variational problems, Comm. Pure Appl. Math., Vol. 42, 577–685, 1989.
[297] M. Nagao, T. Matsuyama, Edge preserving smoothing, Computer Graphics and Image
Processing, Vol. 9, 394–407, 1979.
[298] P. Neskovic, B.B. Kimia, Threedimensional shape representation from curvature depen
dent surface evolution, Proc. IEEE Int. Conf. Image Processing (ICIP–94, Austin, Nov.
13–16, 1994), Vol. 1, IEEE Computer Society Press, Los Alamitos, 6–10, 1994.
[299] H. Neunzert, B. Claus, K. Rjasanowa, R. R¨ osch, J. Weickert, Mathematische Werkzeuge in
der Bildverarbeitung zur Qualit¨ atsbeurteilung von Oberﬂ¨ achen, K.H. Hoﬀmann, W. J¨ ager,
T. Lohmann, H. Schunck (Eds.), Mathematik – Schl¨ usseltechnologie f¨ ur die Zukunft,
Springer, Berlin, 449–462, 1997.
156 BIBLIOGRAPHY
[300] T.G. Newman, H. Dirilten, A nonlinear transformation for digital picture processing, IEEE
Trans. Computers, Vol. 22, 869–873, 1973.
[301] E. Nicolet, M. Sp¨ uhler, Charakterisierung von unscharfen blobf¨ ormigen 3D Strukturen,
diploma thesis, Communication Technology Laboratory, Image Science Division, ETH–
Zentrum, 8092 Z¨ urich, Switzerland, 1996.
[302] M. Nielsen, From paradigm to algorithms in computer vision, Ph.D. thesis, Report DIKU
95/8, Dept. of Computer Science, University of Copenhagen, Universitetsparken 1, 2100
Copenhagen, Denmark, 1995.
[303] M. Nielsen, L. Florack, R. Deriche, Regularization, scalespace and edge detection ﬁlters,
J. Math. Imag. Vision, Vol. 7, 291–307, 1997.
[304] W.J. Niessen, B.M. ter Haar Romeny, L.M.J. Florack, M.A. Viergever, A general frame
work for geometrydriven evolution equations, Int. J. Comput. Vision, Vol. 21, 187–205,
1997.
[305] W.J. Niessen, B.M. ter Haar Romeny, M.A. Viergever, Numerical analysis of geometry
driven diﬀusion equations, B.M. ter Haar Romeny (Ed.), Geometrydriven diﬀusion in
computer vision, Kluwer, Dordrecht, 392–410, 1994.
[306] W.J. Niessen, R. Maas, Multiscale optic ﬂow and stereo, J. Sporring, M. Nielsen, L. Flo
rack, P. Johansen (Eds.), Gaussian scalespace theory, Kluwer, Dordrecht, 31–42, 1997.
[307] W.J. Niessen, K.L. Vincken, J. Weickert, M.A. Viergever, Nonlinear multiscale repre
sentations for image segmentation, Computer Vision and Image Understanding, Vol. 66,
233–245, 1997.
[308] W.J. Niessen, K.L. Vincken, J. Weickert, M.A. Viergever, Threedimensional MR brain
segmentation, Proc. Sixth Int. Conf. on Computer Vision (ICCV ’98, Bombay, Jan. 4–7,
1998), in press.
[309] M. Nitzberg, D. Mumford, T. Shiota, Filtering, segmentation and depth, Lecture Notes in
Computer Science, Vol. 662, Springer, Berlin, 1993.
[310] M. Nitzberg, T. Shiota, Nonlinear image ﬁltering with edge and corner enhancement,
IEEE Trans. Pattern Anal. Mach. Intell., Vol. 14, 826–833, 1992.
[311] N. Nordstr¨ om, Biased anisotropic diﬀusion  a uniﬁed regularization and diﬀusion ap
proach to edge detection, Image Vision Comput., Vol. 8, 318–327, 1990.
[312] E. Norman, A discrete analogue of the Weierstrass transform, Proc. Amer. Math. Soc.,
Vol. 11, 596–604, 1960.
[313] O.F. Olsen, Multiscale watershed segmentation, J. Sporring, M. Nielsen, L. Florack, P.
Johansen (Eds.), Gaussian scalespace theory, Kluwer, Dordrecht, 191–200, 1997.
[314] P.J. Olver, G. Sapiro, A. Tannenbaum, Classiﬁcation and uniqueness of invariant geomet
ric ﬂows, C. R. Acad. Sci. Paris, t. 319, S´erie I, 339–344, 1994.
[315] P.J. Olver, G. Sapiro, A. Tannenbaum, Aﬃne invariant detection: edges, active contours,
and segments, Proc. IEEE Comp. Soc. Conf. Computer Vision and Pattern Recognition
(CVPR ’96, San Francisco, June 18–20, 1996), IEEE Computer Society Press, Los Alami
tos, 520–525, 1996.
[316] P.J. Olver, G. Sapiro, A. Tannenbaum, Invariant geometric evolutions of surfaces and
volumetric smoothing, SIAM J. Appl. Math., Vol. 57, 176–194, 1997.
BIBLIOGRAPHY 157
[317] S. Osher, L. Rudin, Featureoriented image enhancement using shock ﬁlters, SIAM J.
Numer. Anal., Vol. 27, 919–949, 1990.
[318] S. Osher, L. Rudin, Shocks and other nonlinear ﬁltering applied to image processing, A.G.
Tescher (Ed.): Applications of digital image processing XIV, SPIE Vol. 1567, 414–431,
1991.
[319] S. Osher, J.A. Sethian, Fronts propagating with curvaturedependent speed: Algorithms
based on Hamilton–Jacobi formulations, J. Comput. Phys., Vol. 79, 12–49, 1988.
[320] N. Otsu, Mathematical studies on feature extraction in pattern recognition, Ph.D. thesis,
Report No. 818, Electrotechnical Laboratory, 114, Umezono, Sakuramura, Niiharigun,
Ibaraki, Japan, 1981 (in Japanese).
[321] K. Ottenberg, Modelbased extraction of geometric structure from digital images, Ph.D.
thesis, Utrecht University, The Netherlands, 1993.
[322] N. Paragios, R. Deriche, A PDEbased levelset approach for detection and tracking of
moving objects, Proc. Sixth Int. Conf. on Computer Vision (ICCV ’98, Bombay, Jan. 4–7,
1998), in press.
[323] E.J. Pauwels, P. Fiddelaers, L.J. Van Gool, Enhancement of planar shape through op
timization of functionals for curves, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 17,
1101–1105, 1995.
[324] E.J. Pauwels, L.J. Van Gool, P. Fiddelaers, T. Moons, An extended class of scaleinvariant
and recursive scale space ﬁlters, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 17, 691–701,
1995.
[325] E. Payot, R. Guillemaud, Y. Trousset, F. Preteux, An adaptive and constrained model for
3D Xray vascular reconstruction, P. Grangeat, J.L. Amans (Eds.), Threedimensional
image reconstruction in radiation and nuclear medicine, Kluwer, Dordrecht, 47–57, 1996.
[326] P. Perona, J. Malik, Scale space and edge detection using anisotropic diﬀusion, Proc. IEEE
Comp. Soc. Workshop on Computer Vision (Miami Beach, Nov. 30 – Dec. 2, 1987), IEEE
Computer Society Press, Washington, 16–22, 1987.
[327] P. Perona, J. Malik, A network for multiscale image segmentation, Proc. IEEE Int. Symp.
Circuits and Systems (ISCAS–88, Espoo, June 7–9, 1988), 2565–2568, 1988.
[328] P. Perona, J. Malik, Scale space and edge detection using anisotropic diﬀusion, IEEE
Trans. Pattern Anal. Mach. Intell., Vol. 12, 629–639, 1990.
[329] P. Perona, T. Shiota, J. Malik, Anisotropic diﬀusion, B.M. ter Haar Romeny (Ed.),
Geometrydriven diﬀusion in computer vision, Kluwer, Dordrecht, 72–92, 1994.
[330] P. Perona, M. Tartagni, Diﬀusion networks for onchip image contrast normalization,
Proc. IEEE Int. Conf. Image Processing (ICIP–94, Austin, Nov. 13–16, 1994), Vol. 1,
IEEE Computer Society Press, Los Alamitos, 1–5, 1994.
[331] I.G. Petrowski, Vorlesungen ¨ uber partielle Diﬀerentialgleichungen, Teubner, Leipzig, 1955.
[332] I. Pollak, A.S. Willsky, H. Krim, Scale space analysis by stabilized inverse diﬀusion equa
tions, B. ter Haar Romeny, L. Florack, J. Koenderink, M. Viergever (Eds.), Scalespace
theory in computer vision, Lecture Notes in Computer Science, Vol. 1252, Springer, Berlin,
200–211, 1997.
[333] E.S. Posmentier, The generation of salinity ﬁnestructure by vertical diﬀusion, J. Phys.
Oceanogr., Vol. 7, 298–300, 1977.
158 BIBLIOGRAPHY
[334] C.B. Price, P. Wambacq, A. Oosterlinck, Applications of reaction–diﬀusion equations to
image processing, IEE Third Int. Conf. Image Processing and its Applications (Warwick,
18–20 July 1989), IEE, London, 49–53, 1989.
[335] C.B. Price, P. Wambacq, A. Oosterlinck, Computing with reaction–diﬀusion systems: ap
plications in image processing, D. Roose et al. (Eds.), Continuation and bifurcations:
Numerical techniques and applications, Kluwer, 379–387, 1990.
[336] M. Proesmans, E.J. Pauwels, L.J. Van Gool, Coupled geometrydriven diﬀusion equations
for lowlevel vision, B.M. ter Haar Romeny (Ed.), Geometrydriven diﬀusion in computer
vision, Kluwer, Dordrecht, 191–228, 1994.
[337] M. Proesmans, E.J. Pauwels, L.J. Van Gool, T. Moons, A. Oosterlinck, Image enhance
ment using nonlinear diﬀusion, Proc. IEEE Comp. Soc. Conf. Computer Vision and
Pattern Recognition (CVPR ’93, New York, June 15–17, 1993), IEEE Computer Society
Press, Los Alamitos, 680–681, 1993.
[338] M. Proesmans, L. Van Gool, E. Pauwels, A. Oosterlinck, Determination of optical ﬂow
and its discontinuities using nonlinear diﬀusion, J.O. Eklundh (Ed.), Computer Vision –
ECCV ’94, Lecture Notes in Computer Science, Vol. 801, Springer, Berlin, 295–304, 1994.
[339] I. Rambaux, P. Gar¸ con, Nonlinear anisotropic diﬀusion ﬁltering of 3D images, project
work, D´epartement G´enie Math´ematique, INSA de Rouen and Laboratory of Techno
mathematics, University of Kaiserslautern, 1994.
[340] A.R. Rao, B.G. Schunck, Computing oriented texture ﬁelds, CVGIP: Graphical Models
and Image Processing, Vol. 53, 157–185, 1991.
[341] T. Richardson, S. Mitter, Approximation, computation, and distortion in the variational
formulation, B.M. ter Haar Romeny (Ed.), Geometrydriven diﬀusion in computer vision,
Kluwer, Dordrecht, 169–190, 1994.
[342] J.H. Rieger, Generic evolution of edges on families of diﬀused greyvalue surfaces, J. Math.
Imag. Vision, Vol. 5, 207–217, 1995.
[343] L. Robert, R. Deriche, Dense depth map reconstruction: A minimization and regularization
approach which preserves discontinuities, B. Buxton, R. Cipolla (Eds.), Computer vision
– ECCV ’96, Volume I, Lecture Notes in Computer Science, Vol. 1064, Springer, Berlin,
439–451, 1996.
[344] L.I. Rudin, S. Osher, Total variation based image restoration with free local constraints,
Proc. IEEE Int. Conf. Image Processing (ICIP–94, Austin, Nov. 13–16, 1994), Vol. 1,
IEEE Computer Society Press, Los Alamitos, 31–35, 1994.
[345] L.I. Rudin, S. Osher, E. Fatemi, Nonlinear total variation based noise removal algorithms,
Physica D, Vol. 60, 259–268, 1992.
[346] L.I. Rudin, S. Osher, C. Fu, Total variation based restoration of noisy, blurred images,
Preprint, Cognitech, Inc., 2800 – 28 St., Suite 101, Santa Monica, CA 90405, U.S.A., 1992.
[347] S.J. Ruuth, An algorithm for generating motion by mean curvature, M.O. Berger, R.
Deriche, I. Herlin, J. Jaﬀr´e, J.M. Morel (Eds.), ICAOS ’96: Images, wavelets and PDEs,
Lecture Notes in Control and Information Sciences, Vol. 219, Springer, London, 82–91,
1996.
[348] Y. Saad, Iterative methods for sparse linear systems, PWS Publishing, Boston, 1996.
[349] P. SaintMarc, J.S. Chen, G. Medioni, Adaptive smoothing: a general tool for early vision,
IEEE Trans. Pattern Anal. Mach. Intell., Vol. 13, 514–529, 1990.
BIBLIOGRAPHY 159
[350] A.H. Salden, The nonlinear rescaling process, Report 3DCV 9228, 3D Computer Vision,
AZU, Room E.02.222, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands, 1992.
[351] A.H. Salden, Dynamic scalespace paradigms, Ph.D. thesis, Utrecht University, The
Netherlands, 1996.
[352] A.H. Salden, B.M. ter Haar Romeny, M.A. Viergever, Image structure generating nor
malised geometric scale spaces, Proc. Int. Conf. Volume Image Processing (VIP ’93,
Utrecht, June 2–4, 1993), 141–143, 1993.
[353] A.H. Salden, B.M. ter Haar Romeny, M.A. Viergever, Linear scalespace theory from
physical principles, J. Math. Imag. Vision, to appear.
[354] A.A. Samarskij, Theorie der Diﬀerenzenverfahren, Geest & Portig, Leipzig, 1984.
[355] G.I. S´ anchez–Ortiz, D. Rueckert, P. Burger, Knowledgebased anisotropic diﬀusion of
vectorvalued 4dimensional cardiac MR images, R.B. Fisher, E. Trucco (Eds.), Proc.
Seventh British Machine Vision Conference (BMVC 1996, Edinburgh, Sept. 9–12, 1996),
BMVA Press.
[356] G. Sapiro, From active contours to anisotropic diﬀusion: connections between basic PDEs
in image processing, Proc. IEEE Int. Conf. Image Processing (ICIP–96, Lausanne, Sept.
16–19, 1996), Vol. 1, 477–480, 1996.
[357] G. Sapiro, Vector (self ) snakes: a geometric framework for color, texture and multiscale
image segmentation, Proc. IEEE Int. Conf. Image Processing (ICIP–96, Lausanne, Sept.
16–19, 1996), Vol. 1, 817–820, 1996.
[358] G. Sapiro, V. Caselles, Histogram modiﬁcation via diﬀerential equations, J. Diﬀerential
Equations, Vol 135, 238–268, 1997.
[359] G. Sapiro, A. Cohen, A.M. Bruckstein, A subdivision scheme for continuousscale B
splines and aﬃneinvariant progressive smoothing, J. Math. Imag. Vision, Vol. 7, 23–40,
1997.
[360] G. Sapiro, R. Kimmel, D. Shaked, B.B. Kimia, A.M. Bruckstein, Implementing continuous
scale morphology via curve evolution, Pattern Recognition, Vol. 26, 1363–1372, 1993.
[361] G. Sapiro, D.L. Ringach, Anisotropic diﬀusion of multivalued images with applications to
color ﬁltering, IEEE Trans. Image Proc., Vol. 5, 1582–1586, 1996.
[362] G. Sapiro, A. Tannenbaum, Aﬃne invariant scalespace, Int. J. Comput. Vision, Vol. 11,
25–44, 1993.
[363] G. Sapiro, A. Tannenbaum, On aﬃne plane curve evolution, J. Funct. Anal., Vol. 119,
79–120, 1994.
[364] G. Sapiro, A. Tannenbaum, Edge preserving geometric smoothing of MRI data, Technical
Report, Dept. of Electical Engineering, University of Minnesota, Minneapolis, MN 55455,
U.S.A., 1994; submitted to IEEE Trans. Medical Imaging.
[365] G. Sapiro, A. Tannenbaum, Y.L. You, M. Kaveh, Experiments on geometric image en
hancement, Proc. IEEE Int. Conf. Image Processing (ICIP–94, Austin, Nov. 13–16, 1994),
Vol. 2, IEEE Computer Society Press, Los Alamitos, 472–476, 1994.
[366] G. Sapiro, A. Tannenbaum, Area and length preserving geometric invariant scalespaces,
IEEE Trans. Pattern Anal. Mach. Intell., Vol. 17, 67–72, 1995.
[367] C. Schn¨ orr, Unique reconstruction of piecewise smooth images by minimizing strictly con
vex nonquadratic functionals, J. Math. Imag. Vision, Vol. 4, 189–198, 1994.
160 BIBLIOGRAPHY
[368] C. Schn¨ orr, Segmentation of visual motion by minimizing convex nonquadratic function
als, Proc. 12th Int. Conf. Pattern Recognition (ICPR 12, Jerusalem, Oct. 9–13, 1994),
Vol. A, IEEE Computer Society Press, Los Alamitos, 661–663, 1994.
[369] C. Schn¨ orr, Convex variational segmentation of multichannel images, M.O. Berger, R.
Deriche, I. Herlin, J. Jaﬀr´e, J.M. Morel (Eds.), ICAOS ’96: Images, wavelets and PDEs,
Lecture Notes in Control and Information Sciences, Vol. 219, Springer, London, 201–207,
1996.
[370] E. Seneta, Nonnegative matrices and Markov chains, Springer, New York, 1981.
[371] J. Serra, Image analysis and mathematical morphology, Academic Press, London, 1982.
[372] J.A. Sethian, Level set methods, Cambridge University Press, Cambridge, 1996.
[373] J.A. Sethian, A fast marching level set method for monotonically advancing fronts, Proc.
Nat. Acad. Sci., Vol. 93, 1591–1595, 1996.
[374] J.A. Sethian, A review of the theory, algorithms, and applications of level set methods for
propagating interfaces, Acta Numerica, 309–395, 1996.
[375] J. Shah, Segmentation by nonlinear diﬀusion, Proc. IEEE Comp. Soc. Conf. on Computer
Vision and Pattern Recognition (CVPR ’91, Maui, June 3–6, 1992), IEEE Computer
Society Press, Los Alamitos, 202–207, 1991.
[376] J. Shah, Segmentation by nonlinear diﬀusion, II, Proc. IEEE Comp. Soc. Conf. Com
puter Vision and Pattern Recognition (CVPR ’92, Champaign, June 15–18, 1992), IEEE
Computer Society Press, Los Alamitos, 644–647, 1992.
[377] J. Shah, Properties of energyminimizing segmentations, SIAM J. Control and Optimiza
tion, Vol. 30, 99–111, 1992.
[378] J. Shah, A nonlinear diﬀusion model for discontinuous disparity and halfocclusions in
stereo, Proc. IEEE Comp. Soc. Conf. Computer Vision and Pattern Recognition (CVPR
’93, New York, June 15–17, 1993), IEEE Computer Society Press, Los Alamitos, 34–40,
1993.
[379] J. Shah, Piecewise smooth approximations of functions, Calc. Var., Vol. 2, 315–328, 1994.
[380] J. Shah, A common framework for curve evolution, segmentation and anisotropic diﬀusion,
Proc. IEEE Comp. Soc. Conf. Computer Vision and Pattern Recognition (CVPR ’96, San
Francisco, June 18–20, 1996), IEEE Computer Society Press, Los Alamitos, 136–142, 1996.
[381] J. Shen, S. Castan, An optimal linear operator for step edge detection, Graphical Models
and Image Processing, Vol. 54, 112–133, 1992.
[382] A. Sherstinsky, R.W. Picard, Mlattice: From morphogenesis to image processing, IEEE
Trans. Image Proc., Vol. 5, 1137–1150, 1996.
[383] K. Siddiqi, B.B. Kimia, Parts of visual form: computational aspects, IEEE Trans. Pattern
Anal. Mach. Intell., Vol. 17, 239–251, 1995.
[384] K. Siddiqi, B.B. Kimia, A shock grammar for recognition, Proc. IEEE Comp. Soc. Conf.
Computer Vision and Pattern Recognition (CVPR ’96, San Francisco, June 18–20, 1996),
IEEE Computer Society Press, Los Alamitos, 507–513, 1996.
[385] K. Siddiqi, B.B. Kimia, C.W. Shu, Geometric shockcapturing ENO schemes for subpixel
interpolation, computation and curve evolution, Graphical Models and Image Processing,
in press.
BIBLIOGRAPHY 161
[386] J. Sijbers, P. Scheunders, M. Verhoye, A. Van der Linden, D. Van Dyck, E. Raman,
Watershedbased segmentation of 3D MR data for volume quantization, Magnetic Reso
nance Imaging, Vol. 15, 679–688, 1997.
[387] W. Snyder, Y.S. Han, G. Bilbro, R. Whitaker, S. Pizer, Image relaxation: Restoration
and feature extraction, IEEE Trans. Pattern Anal. Mach. Intell., Vol. 17, 620–624, 1995.
[388] J. Sporring, The entropy of scalespace, Proc. 13th Int. Conf. Pattern Recognition (ICPR
13, Vienna, Aug. 25–30, 1996), Vol. A, 900–904, 1996.
[389] J. Sporring, M. Nielsen, L. Florack, P. Johansen (Eds.), Gaussian scalespace theory,
Kluwer, Dordrecht, 1997.
[390] J. Sporring, J. Weickert, On generalized entropies and scalespace, B. ter Haar Romeny,
L. Florack, J. Koenderink, M. Viergever (Eds.), Scalespace theory in computer vision,
Lecture Notes in Computer Science, Vol. 1252, Springer, Berlin, 53–64, 1997.
[391] R. Sprengel, Entwurf und Analyse nichtlinearer Diﬀusionsverfahren f¨ ur die Bildverar
beitung, Ph.D. thesis, Dept. of Computer Science, University of Hamburg, Germany, 1995;
DISKI 123, Inﬁx Verlag, Sankt Augustin, 1996.
[392] J.L. Stansﬁeld, Conclusions from the commodity expert project, A.I. Memo 671, Artiﬁcial
Intelligence Lab., Massachusetts Inst. of Technology, Cambridge, MA, U.S.A., 1980.
[393] E. Steen, B. Olstad, Scalespace and boundary detection in ultrasonic imaging using non
linear signaladaptive anisotropic diﬀusion, M.H. Loew (Ed.), Image processing, SPIE Vol.
2167, 116–127, 1994.
[394] A. Steiner, R. Kimmel, A.M. Bruckstein, Planar shape enhancement and exaggeration,
Proc. 13th Int. Conf. Pattern Recognition (ICPR 13, Vienna, Aug. 25–30, 1996), Vol. A,
523–527, 1996.
[395] P. Stoll, C.W. Shu, B.B. Kimia, Shockcapturing numerical methods for viscosity solutions
of certain PDEs in computer vision: The Godunov, Osher–Sethian and ENO schemes,
Technical Report LEMS132, Division of Engineering, Brown University, Providence, R.I.
02912, U.S.A., 1994.
[396] D.M. Strong, T.F. Chan, Relation of regularization parameter and scale in total variation
based image denoising, CAM Report 967, Dept. of Mathematics, Univ. of California, Los
Angeles, CA 90024, U.S.A., 1996.
[397] D.M. Strong, T.F. Chan, Spatially and scale adaptive total variation based regularization
and anisotropic diﬀusion in image processing, CAM Report 9646, Dept. of Mathematics,
Univ. of California, Los Angeles, CA 90024, U.S.A., 1996.
[398] S. Teboul, L. Blanc–F´eraud, G. Aubert, M. Barlaud, Variational approach for edge
preserving regularization using coupled PDEs, IEEE Trans. Image Proc., March 1998,
in press.
[399] H. Tek, B.B. Kimia, Image segmentation by reaction–diﬀusion bubbles, Proc. Fifth Int.
Conf. on Computer Vision (ICCV ’95, Cambridge, June 20–23, 1995), IEEE Computer
Society Press, Los Alamitos, 156–162, 1995.
[400] H. Tek, B.B. Kimia, Volumetric segmentation of medical images by threedimensional
bubbles, Computer Vision and Image Understanding, Vol. 65, 246–258, 1997.
[401] D. Terzopoulos, A. Witkin, M. Kass, Constraints on deformable models: Recovering 3D
shape and nonrigid motion, Artif. Intell., Vol. 36, 91–123, 1988.
162 BIBLIOGRAPHY
[402] A. Tikhonov, V. Arsenin, Solutions of illposed problems, Wiley, New York, 1977.
[403] F. Torkamani–Azar, K.E. Tait, Image recovery using the anisotropic diﬀusion equation,
IEEE Trans. Image Proc., Vol. 5, 1573–1578, 1996.
[404] V. Torre, T.A. Poggio, On edge detection, IEEE Trans. Pattern Anal. Mach. Intell., Vol.
8, 148–163, 1986.
[405] A.M. Turing, The chemical basis of morphogenesis, Philosoph. Trans. Royal Soc. London
B, Vol. 237, 37–72, 1952.
[406] G. Turk, Generating textures on arbitrary surfaces using reaction–diﬀusion, Computer
Graphics, Vol. 25, No. 4, 289–298, 1991.
[407] R.A. Varga, Matrix iterative analysis, Prentice Hall, Englewood Cliﬀs, 1962.
[408] K.L. Vincken, A.S.E. Koster, M.A. Viergever, Probabilistic multiscale image segmentation,
IEEE Trans. Pattern Anal. Mach. Intell., Vol. 19, 109–120, 1997.
[409] C.R. Vogel, M.E. Oman, Iterative methods for total variation denoising, SIAM J. Sci.
Comput., Vol. 17, 227–238, 1996.
[410] C.R. Vogel, M.E. Oman, Fast numerical methods for total variation minimization in image
reconstruction, F.T. Luk (Ed.), Advanced signal processing algorithms, SPIE Vol. 2563,
359–367, 1995.
[411] W. Walter, Diﬀerential and integral inequalities, Springer, Berlin, 1970.
[412] W. Walter, Gew¨ ohnliche Diﬀerentialgleichungen, Springer, Berlin, 1976.
[413] J. Weickert, Anisotropic diﬀusion ﬁlters for image processing based quality control, A.
Fasano, M. Primicerio (Eds.), Proc. Seventh European Conf. on Mathematics in Industry,
Teubner, Stuttgart, 355–362, 1994.
[414] J. Weickert, Scalespace properties of nonlinear diﬀusion ﬁltering with a diﬀusion tensor,
Report No. 110, Laboratory of Technomathematics, University of Kaiserslautern, P.O.
Box 3049, 67653 Kaiserslautern, Germany, 1994.
[415] J. Weickert, Theoretical foundations of anisotropic diﬀusion in image processing, Com
puting, Suppl. 11, 221–236, 1996.
[416] J. Weickert, Anisotropic diﬀusion in image processing, Ph.D. thesis, Dept. of Mathematics,
University of Kaiserslautern, P.O. Box 3049, 67653 Kaiserslautern, Germany, 1996.
[417] J. Weickert, A model for the cloudiness of fabrics, H. Neunzert (Ed.), Progress in industrial
mathematics at ECMI 94, Wiley–Teubner, Chichester, 258–265, 1996.
[418] J. Weickert, Multiscale texture enhancement, V. Hlav´aˇc, R.
ˇ
S´ ara (Eds.), Computer analysis
of images and patterns, Lecture Notes in Computer Science, Vol. 970, Springer, Berlin,
230–237, 1995.
[419] J. Weickert, Foundations and applications of nonlinear anisotropic diﬀusion ﬁltering, Z.
Angew. Math. Mech., Vol. 76, Suppl. 1, 283–286, 1996.
[420] J. Weickert, Nonlinear diﬀusion scalespaces: From the continuous to the discrete setting,
M.O. Berger, R. Deriche, I. Herlin, J. Jaﬀr´e, J.M. Morel (Eds.), ICAOS ’96: Images,
wavelets and PDEs, Lecture Notes in Control and Information Sciences, Vol. 219, Springer,
London, 111–118, 1996.
[421] J. Weickert, Nonlinear scalespaces, J. Sporring, M. Nielsen, L. Florack, P. Johansen
(Eds.), Gaussian scalespace theory, Kluwer, Dordrecht, 221–234, 1997.
BIBLIOGRAPHY 163
[422] J. Weickert, Coherenceenhancing diﬀusion of colour images, A. Sanfeliu, J.J. Villanueva,
J. Vitri` a (Eds.), Pattern Recognition and Image Analysis (VII NSPRIA, Barcelona, April
21–25, 1997), Vol. 1, 239–244, 1997.
[423] J. Weickert, A review of nonlinear diﬀusion ﬁltering, B. ter Haar Romeny, L. Florack, J.
Koenderink, M. Viergever (Eds.), Scalespace theory in computer vision, Lecture Notes
in Computer Science, Vol. 1252, Springer, Berlin, 3–28, 1997.
[424] J. Weickert, Recursive separable schemes for nonlinear diﬀusion ﬁlters, B. ter Haar
Romeny, L. Florack, J. Koenderink, M. Viergever (Eds.), Scalespace theory in computer
vision, Lecture Notes in Computer Science, Vol. 1252, Springer, Berlin, 260–271, 1997.
[425] J. Weickert, B. Benhamouda, A semidiscrete nonlinear scalespace theory and its relation
to the Perona–Malik paradox, F. Solina, W.G. Kropatsch, R. Klette, R. Bajcsy (Eds.),
Theoretical foundations of computer vision (TFCV ’96, Dagstuhl, March 18–22, 1996),
Springer, Wien, in press.
[426] J. Weickert, S. Ishikawa, A. Imiya, On the history of Gaussian scalespace axiomatics, J.
Sporring, M. Nielsen, L. Florack, P. Johansen (Eds.), Gaussian scalespace theory, Kluwer,
Dordrecht, 45–59, 1997.
[427] J. Weickert, S. Ishikawa, A. Imiya, Scalespace has been discovered in Japan, Report DIKU
97/18, Dept. of Computer Science, University of Copenhagen, Universitetsparken 1, 2100
Copenhagen, Denmark, 1997, submitted.
[428] J. Weickert, B.M. ter Haar Romeny, A. Lopez, W.J. van Enk, Orientation analysis by
coherenceenhancing diﬀusion, Proc. Symp. Real World Computing (RWC ’97, Tokyo,
Jan. 29–31, 1997), 96–103, 1997.
[429] J. Weickert, B.M. ter Haar Romeny, M.A. Viergever, Conservative image transformations
with restoration and scalespace properties, Proc. IEEE Int. Conf. Image Processing (ICIP–
96, Lausanne, Sept. 16–19, 1996), Vol. 1, 465–468, 1996.
[430] J. Weickert, B.M. ter Haar Romeny, M.A. Viergever, Eﬃcient and reliable schemes for
nonlinear diﬀusion ﬁltering, IEEE Trans. Image Proc., March 1998, to appear.
[431] J. Weickert, K.J. Zuiderveld, B.M. ter Haar Romeny, W.J. Niessen, Parallel implementa
tions of AOS schemes: A fast way of nonlinear diﬀusion ﬁltering, Proc. 1997 IEEE Inter
national Conference on Image Processing (ICIP–97, Santa Barbara, Oct. 26–29, 1997), to
appear.
[432] H. Werner, H. Arndt, Gew¨ ohnliche Diﬀerentialgleichungen, Springer, Berlin, 1986.
[433] R.T. Whitaker, Geometry limited diﬀusion in the characterization of geometric patches in
images, CVGIP: Image Understanding, Vol. 57, 111–120, 1993.
[434] R.T. Whitaker, Characterizing ﬁrst and secondorder patches using geometrylimited dif
fusion, H.H. Barrett, A.F. Gmitro (Eds.), Information processing in medical imaging,
Lecture Notes in Computer Science, Vol. 687, Springer, Berlin, 149–167, 1993.
[435] R.T. Whitaker, Algorithms for implicit deformable models, Proc. Fifth Int. Conf. on Com
puter Vision (ICCV ’95, Cambridge, June 20–23, 1995), IEEE Computer Society Press,
Los Alamitos, 822–827, 1995.
[436] R.T. Whitaker, D.T. Chen, Embedded active surfaces for volume visualization, M.H. Loew
(Ed.), Image processing, SPIE Vol. 2167, 340–352, 1994.
[437] R. Whitaker, G. Gerig, Vectorvalued diﬀusion, B.M. ter Haar Romeny (Ed.), Geometry
driven diﬀusion in computer vision, Kluwer, Dordrecht, 93–134, 1994.
164 BIBLIOGRAPHY
[438] R.T. Whitaker, S.M. Pizer, A multiscale approach to nonuniform diﬀusion, CVGIP: Im
age Understanding, Vol. 57, 99–110, 1993.
[439] A.P. Witkin, Scalespace ﬁltering, Proc. Eighth Int. Joint Conf. on Artiﬁcial Intelligence
(IJCAI ’83, Karlsruhe, Aug. 8–12, 1983), Vol. 2, 1019–1022, 1983.
[440] A. Witkin, M. Kass, Reaction–diﬀusion textures, Computer Graphics, Vol. 25, No. 4,
299–308, 1991.
[441] A. Witkin, D. Terzopoulos, M. Kass, Signal matching through scalespace, Int. J. Comput.
Vision, Vol. 1, 134–144, 1987.
[442] N.N. Yanenko, The method of fractional steps: the solution of problems of mathematical
physics in several variables, Springer, New York, 1971.
[443] G.Z. Yang, P. Burger, D.N. Firmin, S.R. Underwood, Structure adaptive anisotropic ﬁl
tering, Image Vision Comput., Vol. 14, 135–145, 1996.
[444] T.S. Yoo, J.M. Coggins, Using statistical pattern recognition techniques to control vari
able conductance diﬀusion, H.H. Barrett, A.F. Gmitro (Eds.), Information processing in
medical imaging, Lecture Notes in Computer Science, Vol. 687, Springer, Berlin, 459–471,
1993.
[445] Y.L. You, M. Kaveh, Anisotropic blind image restoration, Proc. IEEE Int. Conf. Image
Processing (ICIP–96, Lausanne, Sept. 16–19, 1996), Vol. 2, 461–464, 1996.
[446] Y.L. You, W. Xu, A. Tannenbaum, M. Kaveh, Behavioral analysis of anisotropic diﬀusion
in image processing, IEEE Trans. Image Proc., Vol. 5, 1539–1553, 1996.
[447] D.M. Young, Iterative solution of large linear systems, Academic Press, New York, 1971.
[448] I.T. Young, L.J. van Vliet, Recursive implementation of the Gaussian ﬁlter, Signal Pro
cessing, Vol. 44, 139–151, 1995.
[449] R.A. Young, The Gaussian derivative model for spatial vision: I. Retinal mechanisms,
Spatial Vision, Vol. 2, 273–293, 1987.
[450] A.L. Yuille, T.A. Poggio, Scaling theorems for zero crossings, IEEE Trans. Pattern Anal.
Mach. Intell., Vol. 8, 15–25, 1986.
[451] D. Zhao, B. Li, A new implementation of discrete multiscale ﬁltering, Proc. IEEE Int.
Conf. Image Processing (ICIP–96, Lausanne, Sept. 16–19, 1996), Vol. 1, 383–386, 1996.
[452] S.C. Zhu, A.L. Yuille, Region competition: unifying snakes, region growing, en
ergy/Bayes/MDL for multiband image segmentation, IEEE Trans. Pattern Anal. Mach.
Intell., Vol. 18, 884–900, 1996.
[453] S.W. Zucker, A. Dobbins, L. Iverson, B. Kimia, A. Tannenbaum, From curve detection
to shape description: an outline, A. Basu, X. Li (Eds.), Computer vision: systems, theory
and applications, World Scientiﬁc, Singapore, 25–39, 1993.
Index
apriori estimates, 93, 105, 110
apriori knowledge, 12–14, 119, 135
active
blobs, 46
contour models, 44–49
surfaces, 46
adaptive mesh coarsening, 25
adaptive smoothing, 14
additive operator splitting, 26, 107–111
aﬃne
arclength, see arclength
Gaussian scalespace, 13
invariance, see invariance
invariant geodesic snakes, 48
invariant gradient, 45
invariant heat ﬂow, 40
invariant texture segmentation, 29
morphological scalespace, 40–45, 113,
124
Mumford–Shah functional, 29
perimeter, see perimeter
shortening ﬂow, see shortening ﬂow
transformation, see transformation
AMSS, see aﬃne morphological scalespace
anisotropic diﬀusion, see diﬀusion
AOS, see additive operator splitting
applications, 11, 13, 26, 28, 30, 37, 44, 52,
114–126, 129–134
arclength
aﬃne, 40
Euclidean, 38
areapreserving ﬂows, 42
average globality, 74
average grey level invariance, see invariance
axiomatics, see scalespace
balloon force, 47, 48
basic equation of ﬁgure, 13
bias term, 28, 49
blind image restoration, 27, 52
blob detection, 5, 45
BUC, 34, 61
CAD, 37
Canny edge detector, 5, 15
causality, 5, 34, 65
CED, see diﬀusion, coherenceenhancing
central moments, 73, 85, 102
circular point, 39
closing, 32
coherence, 57, 128
coherence direction, 57, 128
coherenceenhancing diﬀusion, see diﬀusion
colour images, see vectorvalued images
commuting operators, 62
comparison principle, 20, 61, 135
computer aided quality control, 26, 125–126,
130
condition number, 88, 95, 106
conjugate gradient methods, 106
continuity, 97
continuity equation, 2
continuous dependence, see wellposedness
continuous diﬀusion, see diﬀusion
contrast enhancement, 15, 40, 53, 61, 62, 72,
113–126
contrast parameter, 16
convergence, 14, 19, 21, 22, 26, 28, 30, 50,
52, 67, 76, 82, 98, 99, 106, 116, 118,
124, 134
convolution theorem, 3, 10
corners, 24, 28, 29, 45, 56, 57, 116
correspondence problem, 12, 19, 53, 116, 124
CPU times, 113
crystal growth, 38
curvature
of a curve, 33
of an image, 33
curve evolution, 33, 38–43, 46, 47
curve evolution schemes, 36, 43
deblurring, 5, 50, 52
165
166 INDEX
deep structure, 11, 136
defect detection, see computer aided quality
control
defocusing, 7
deviation cost, 27, 29
diﬀerencesofGaussians, 5
diﬀerential inequalities, 17
diﬀerential invariants, 5
diﬀusion
anisotropic, 2, 13, 22, 24, 55–74, 87–95,
111, 113–137
backward, 5, 50, 116
coherenceenhancing, 24, 127–134
continuous, 55–74
directed, 14
discrete, 97–111
edgeenhancing, 23, 113–126
equation, 2
forward–backward, 15, 45, 66, 72
homogeneous, 2
inhomogeneous, 2, 117
isotropic, 2–22, 86–87, 110, 116
linear, 2–14, 36, 116, 125
nonlinear, 2, 14–27, 52, 55–137
on a sphere, 12
physical background, 2
semidiscrete, 75–95
tensor, 2, 13, 22, 58, 62, 88, 95, 114, 127,
135
diﬀusion–reaction, 27–30, 118
single equations, 27
systems, 28
diﬀusivity, 2, 15, 16, 47, 48, 116
digitally scalable, 36
dilation, 32, 35, 47, 63, 113, 119
direction, 56
directional diﬀusivities, 89, 94
directional splitting, 88–95, 107–111
discrete analogue
AMSS, MCM, 43
Gaussian, 10
linear diﬀusion, 10
nonlinear diﬀusion, 25, 97–111
discrete diﬀusion, see diﬀusion
dissipative eﬀects, 37, 43, 44
distance transformation, 37, 47
DoGs, see diﬀerencesofGaussians
doubly stochastic matrix, 98
edge cost, 27, 29
edge detection, 5, 15, 24, 28, 30, 50, 53, 55,
57
edge enhancement, see contrast enhancement
edgeenhancing diﬀusion, see diﬀusion
elliptic point, 41
energy, 72, 85, 102
energy functional
explicit snakes, 46
for curves, 28, 29
geodesic snakes, 48
Mumford–Shah, 28, 44, 126
Nordstr¨ om, 27, 30, 51
Perona–Malik, 18
Polyakov, 42
TVminimization, 51
ENO schemes, 37
entropy, 11, 73, 85, 102
generalized, 11, 73
entropy scalespace, see scalespace, reaction–
diﬀusion
erosion, 32, 47, 63
Euclidean
arclength, see arclength
perimeter, see perimeter
shortening ﬂow, see shortening ﬂow
transformation, see transformation
Eulerian formulation, 33
existence, see wellposedness
expected value, 73
explicit schemes, 11, 19, 25, 43, 53, 103, 105,
106, 113, 137
external energy, 46
extremum principle, 4, 6, 18–20, 30, 34, 39,
41, 44, 50, 58, 62, 66, 76, 77, 87, 93,
98, 105
fabrics, 125, 130
FD, see ﬁnite diﬀerences
feature vectors, 24
Fick’s law, 2
ﬁlter design, 114, 127
ﬁngerprint enhancement, 26, 129
ﬁnite diﬀerences, 11, 25, 30, 43, 46, 50, 51,
85–95, 102–111, 113
ﬁnite elements, 25, 28
ﬂame propagation, 38
ﬂowline, 16
ﬂux, 2, 15, 22, 114
focusofattention, 11
forensic applications, 52
INDEX 167
Fourier transformation, 3, 10, 35, 44
free discontinuity problems, 29
fundamental equation in image processing,
41, 42
Gabor ﬁlters, 45, 136
Gabor’s restoration method, 45
gas dynamics, 49
gauge coordinates, 16
Gaussian
derivatives, 4, 43, 44
elimination, 106
kernel, 3, 35
pyramid, 11
scalespace, see scalespace
smoothing, 2–14, 56
Gauß–Seidel method, 44, 106
generalized entropy, see entropy
generalized ﬁgure, 13
geodesic, 48
geodesic curvature ﬂow, 42
geometric heat equation, 38, 39
grassﬁre ﬂow, 33
grey level shift invariance, see invariance
greyscale invariance, see invariance
halftoning, 30, 37
hardware realizations, 11, 13, 25, 33
heat equation, 2
histogramme enhancement, 28, 45
Huygens principle, 33
hyperstack, 63
hysteresis thresholding, 5, 15
Iijima’s axiomatic, 7
illposedness, 4, 17, 28–30, 49, 53
image
colour, see vectorvalued images
discrete, 75
greyscale, 3, 55
higher dimensional, see threedimensional
images
vectorvalued, see vectorvalued images
image compression, 73, 137
image enhancement, 14–30, 45, 48–53, 55–
74, 113–134
image restoration, see image enhancement
image sequences, 11, 24, 30, 47, 48, 56, 62
immediate localization, 19
implicit schemes, 11, 44, 137
information theory, 73
integral geometry, 10
integration scale, 57, 114, 127, 132
interest operator, see structure tensor
internal energy, 46
intrinsic heat ﬂows, 39, 40, 42
invariance
aﬃne, 42–45
average grey level, 13, 63, 76, 81, 98, 99,
119
grey level shift, 62, 81
greyscale, 31, 43, 45, 124
isometry, 64, 81
morphological, see invariance, greyscale
reverse contrast, 63, 81
rotational, 13, 21, 35, 43
scale, 8
translation, 8, 64, 81
irreducibility, 75, 76, 97
isometry invariance, see invariance
isophote, 16
Japanese scalespace research, 7, 13
jet space, 24
junctions, 5, 29, 42, 56
Kramer–Bruckner ﬁlter, 50
LaplacianofGaussian, 5
lengthpreserving ﬂows, 42
level set, 31, 44
level set methods, 36
line detection, 26
linear diﬀusion, see diﬀusion
linear systems, 106, 107, 110
linearity, 8
Lipschitzcontinuity, 76
local extrema
creation, 12, 43, 65
noncreation, 34, 39, 119
nonenhancement, 65
local scale, 57
LoG, see LaplacianofGaussian
Lyapunov
functionals, 66–74, 135
functions, 76, 82–85
sequences, 98–102
Markov chains, 98
Markov processes, 11
168 INDEX
Markov random ﬁelds, 27, 29
Marr–Hildreth operator, 5
maximal monotone operator, 17
maximum–minimum principle, see extremum
principle
MCM, see mean curvature motion
mean curvature motion, 37–40, 42–45, 47,
50, 51, 93, 113, 119, 125, 129
backward, 45
mean ﬁeld annealing, 27
median ﬁltering, 38
medical imaging, 27, 28, 45, 46, 52, 62, 63,
111, 116–126, 131
MegaWave, 44, 116, 126
min–max ﬂow, 45
minimal surfaces, 29, 38
monotonicity preservation, 19
morphological invariance, see invariance, grey
scale
morphology, 30–49, 119–124
binary, 31
classical, 30–37
continuousscale, 32
curvaturebased, 37–49
greyscale, 31
multigrid, 11, 25, 106, 137
multiplicative operator splittings, 110
Mumford–Shah functional, see energy func
tional
neural networks, 25, 137
noise elimination, 12, 21, 23, 45, 116, 124
noise scale, 21, 57
nonmaxima suppression, 5, 15
noncreation of local extrema, see local ex
trema
nonenhancement of local extrema, see local
extrema
nonlinear diﬀusion, see diﬀusion
nonnegativity, 76, 88, 97, 107
Nordstr¨ om functional, see energy function
als
numerical aspects, 10, 25, 36, 43, 51, 85–95,
102–111, 113
oceanography, 18
ODE, see ordinary diﬀerential equations
opening, 32
optic ﬂow, 11, 28, 52
optical character recognition, 13
optimization
convex, 28
nonconvex, 53
ordinary diﬀerential equations, 75–95
orientation, 56, 129
oscillations, 10, 19, 21, 44, 50
Osher–Sethian schemes, 36, 113
oversegmentation, 29
parallel computing, 25, 26, 30, 110, 111
parameter determination, 17, 46, 114, 119,
129, 137
partial diﬀerential equations, 1–53
particle methods, 27
path planning, 37
PDE, see partial diﬀerential equations
perceptual grouping, 136
perimeter
aﬃne, 41
Euclidean, 39
Perona–Malik ﬁlter, 14–20, 28, 30, 45, 49,
114
basic idea, 15
edge enhancement, 15
illposedness, 17
regularizations, 20–24
scalespace properties, 19
semidiscrete, 87
piecewise smoothing, 19
pixel numbering, 87
positivity, 97
positivity preservation, 8
postprocessing, 26
prairie ﬂow, 33
predictor–corrector schemes, 107, 109
preprocessing, 118, 126
pseudospectral methods, 25
pyramids, 11, 25, 125
quantization, 45
random variable, 72
reaction–diﬀusion bubbles, 48
reaction–diﬀusion scalespace, see scalespace
recursive ﬁlters, 10, 11, 110, 111
recursivity, 6
redistribution property, 99
region growing, 29, 126
regularity, 58, 61
INDEX 169
regularization, 4, 10, 49, 50, 52, 62, 72, 93,
114
remote sensing, 27
rescaling, 20
reverse contrast invariance, see invariance
robust statistics, 27
row sum, 76, 97
scale invariance, see invariance
scale selection, 11
scale–imprecision space, 6
scalespace
aﬃne Gaussian, 13
aﬃne morphological, 41
anisotropic diﬀusion, 62–74
architectural principles, 6
axioms, 6, 34, 42, 53
dilation–erosion, 34
discrete nonlinear diﬀusion, 25, 99–111
for curves, 39, 42
for image sequences, 11, 43
for shapes, 42
Gaussian, 7
general concept, 6
invariance principles, 7
linear diﬀusion, 7, 116
linearity principle, 7
mean curvature, 39
morphological equivalent of Gaussian scale
space, 35
nonlinear diﬀusion, 62–74, 118
projective, 42
reaction–diﬀusion, 45
semidiscrete linear, 10
semidiscrete nonlinear diﬀusion, 25, 81–
95
smoothing principles, 6, 62
scatter matrix, see structure tensor
second moment matrix, see structure tensor
segmentation, 11, 21, 27, 44, 53, 63, 118, 126,
137
selfsnakes, see snakes
semiimplicit schemes, 25, 44, 50, 102–111,
113
semidiscrete, 10
semidiscrete analogue
linear diﬀusion, 10
nonlinear diﬀusion, 25, 75–95
semidiscrete diﬀusion, see diﬀusion
semigroup property, 6, 8, 34
separability, 10, 35, 36, 110
settheoretic schemes, 36, 43
shape, 31
cues, 56
exaggeration, 45
oﬀset, 37
recognition, 45, 124
segmentation, 45
shape inclusion principle, 39
shapeadapted Gaussian smoothing, 13, 22
shapefromshading, 37
shock
capturing schemes, 37, 49
creation, 34, 50
ﬁlters, 50
shortening ﬂow
aﬃne, 41
Euclidean, 38
silhouette, 31
simulated annealing, 29
skeletonization, 37
slope transform, 35
smoothness, 58, 76, 97, 99, 128
snakes, 46
explicit, 46
geodesic, 48
implicit, 47
selfsnakes, 48, 124
software, 27, 44, 116, 137
SOR, 106
stability, 25, 43, 44, 106, 109
stabilizing cost, 27, 29
staircasing eﬀect, 18, 29, 52
Stampacchia’s truncation method, 59
steadystate, see convergence
steerable ﬁlters, 136
stencil size, 88–95
stereo, 11, 28, 30, 48, 62
stochastic matrix, 98
stopping time, 4, 28, 39, 41, 47, 74, 116
string theory, 42
structure tensor, 56, 86, 114, 136
structuring element, 32
structuring function, 35
subpixel accuracy, 36
subsampling, 26, 73
symmetry, 58, 76, 97
target tracking, 27
terrain matching, 45
170 INDEX
texture
analysis, 56
discrimination, 45, 73
enhancement, 26, 45, 127
generation, 30
segmentation, 24, 27, 29
threedimensional images, 24, 43, 46–48, 136
topological changes, 43, 46, 47
toppoint, 12
total variation, 49
minimizing methods, 50
preserving methods, 49
transformation
aﬃne, 40
Euclidean, 38
projective, 42
translation invariance, see invariance
Turing’s pattern formation model, 30
TV, see total variation
uniform positive deﬁniteness, 58, 62, 76, 93,
129
uniqueness, see wellposedness
van Gogh, 131
variance, 73
vectorvalued images, 24, 28, 30, 42, 43, 49,
52, 136
viscosity solution, 17, 34, 41, 45, 47–49, 61
visual system, 5, 15, 33
wavelets, 11, 25, 44, 136
weak membrane model, 29
wellposedness, 4, 20, 23, 28, 29, 34, 39, 41,
47–49, 52, 53, 58, 62, 76, 77, 98, 135
wood surfaces, 125
zerocrossings, 5, 34, 50
Anisotropic Diﬀusion
in Image Processing
Joachim Weickert
University of Copenhagen
Many recent techniques for digital image enhancement and multiscale image rep
resentations are based on nonlinear partial diﬀerential equations.
This book gives an introduction to the main ideas behind these methods, and
it describes in a systematic way their theoretical foundations, numerical aspects,
and applications. A large number of references enables the reader to acquire an
uptodate overview of the original literature.
The central emphasis is on anisotropic nonlinear diﬀusion ﬁlters. Their ﬂexibil
ity allows to combine smoothing properties with image enhancement qualities. A
general framework is explored covering wellposedness and scalespace results not
only for the continuous, but also for the algorithmically important semidiscrete and
fully discrete settings. The presented examples range from applications in medical
image analysis to problems in computer aided quality control.
B.G. Teubner Stuttgart
Anisotropic Diﬀusion in Image Processing
Joachim Weickert
Department of Computer Science University of Copenhagen Copenhagen, Denmark
B.G. Teubner Stuttgart 1998
1996 Ph.Sc. in physics and industrial mathematics. semidiscrete. Joachim Weickert Born in 1965 in Ludwigshafen/Germany. All rights reserved. .Dr. 1987 B. The goal was to visualize the quality relevant adjacent ﬁbre structures. The copyright has been returned to the author in 2008. G. physics and computer science at the University of Kaiserslautern. In the current version a few typos and other errors have been corrected. rer. or translated into a machine language without the written permission of the author. This book had been published by B. Copenhagen University. c Copyright 2008 by Joachim Weickert. Since then visiting assistant research professor at the Department of Computer Science. Teubner (Stuttgart) in 1998 and went out of print in 2001. No part of this book may be reproduced by any means. 1991 M.D. socalled stripes. or transmitted. Studies in mathematics. The displayed equations describe the basic structure of nonlinear diﬀusion ﬁltering in the continuous.2 for more details). and fully discrete setting.Sc. Postdoctoral researcher at the Image Sciences Institute at Utrecht University from 2/96 to 3/97. The cover image shows a thresholded nonwoven fabric image which was processed by applying a coherenceenhancing anisotropic diﬀusion ﬁlter (see Section 5. nat. Their theoretical foundations are treated in Chapters 2–4. in industrial mathematics. in mathematics.
To my parents Gerda and Norbert .
.
This includes many wellknown techniques such as Gaussian convolution. Thus. they oﬀer several advantages: • Deep mathematical results with respect to wellposedness are available. One of the two goals of this book is to give an overview of the stateoftheart of PDEbased methods for image enhancement and smoothing. • They allow a reinterpretation of several classical methods under a novel unifying framework. Emphasis is put on a v . dilation or erosion. Hundreds of publications have appeared in the last decade. In image processing and computer vision. PDEbased methods are one of the mathematically bestfounded techniques in image processing. or describe novel ways of shape simpliﬁcation. and the restored image is given by the steadystate of this process. Typical PDE techniques for image smoothing regard the original image as initial state of a parabolic (diﬀusionlike) process. since PDEs have proved their usefulness in areas such as physics and engineering sciences for a very long time. their approximations aim to be independent of the underlying grid and may reveal good rotational invariance. • The PDE formulation is genuinely continuous. Many evolution equations for restoring images can be derived as gradient descent methods for minimizing a suitable energy functional.Preface Partial diﬀerential equations (PDEs) have led to an entire new ﬁeld in image processing and computer vision. median ﬁltering. more global representations of it. and PDEbased methods have played a central role at several conferences and workshops. The success of these techniques is not really surprising. an embedding of the original image into a family of subsequently simpler. structure preserving ﬁltering. The whole evolution can be regarded as a socalled scalespace. one can use a scalespace representation for extracting semantically important information. • This understanding has also led to the discovery of new methods. such that stable algorithms can be found. and extract ﬁltered versions from its temporal evolution. PDEbased image processing techniques are mainly used for smoothing and restoration purposes. and enhancement of coherent linelike structures. They can oﬀer more invariances than classical techniques. Since this introduces a hierarchy into the image structures.
numerical approximations. Thus. Mathematically. Image processing and computer vision are interdisciplinary areas. generalizations and applications. These ﬁlters are diﬃcult to analyse mathematically. and invariances. As a consequence. In order to establish wellposedness and scalespace properties for this class.vi PREFACE uniﬁed description of the underlying ideas. uniqueness. The second goal of this book is to present an indepth treatment of an interesting class of parabolic equations which may bridge the gap between scalespace and restoration ideas: nonlinear diﬀusion ﬁlters. which reveal similar properties as their continuous diﬀusion counterparts. we shall investigate existence. the diﬀusive ﬂux does not have to be parallel to the grey value gradient: the ﬁlters may become anisotropic. Many references are given which point the reader to useful original literature for a task at hand. Lyapunov functionals. as they may act locally like a backward diﬀusion process. Anisotropic diﬀusion ﬁlters can outperform isotropic ones with respect to certain applications such as denoising of highly degraded edges or enhancing coherent ﬂowlike images by closing interrupted onedimensional structures. In order to smooth an image and to simultaneously enhance important features such as edges. but it also allows processes which replace the scalar diﬀusivity by a diﬀusion tensor. where researchers. and to include various aspects . Although being concise. maximum–minimum principles. The proofs present mathematical results from the nonlinear analysis of partial diﬀerential equations. but also historical remarks and pointers to open questions can be found. stability. It leads to a semidiscrete and fully discrete scalespace theory for nonlinear diﬀusion processes. Organization of the book. We shall develop results in this direction by investigating a general class of nonlinear diﬀusion processes. I have tried to keep this book as selfcontained as possible. the Mumford–Shah functional for image segmentation. nonlinear diﬀusion ﬁlters are frequently applied with very impressive results. it is necessary to know if the results for the continuous framework carry over to the practically relevant discrete setting. continuousscale morphology. A general characterization of semidiscrete and fully discrete ﬁlters. theoretical results. Since digital images are always sampled on a pixel grid. These questions are an important topic of the present book as well. On the other hand. practitioners and students may have a very diﬀerent scientiﬁc background and diﬀering intentions. this part covers a broad spectrum: it includes for instance an early Japanese scalespace axiomatic. this comes down to the study of nonlinear systems of ordinary diﬀerential equations and the theory of nonnegative matrices. so there appears the need for a theoretical foundation. active contour models and shock ﬁlters. This gives rise to wellposedness questions. is presented. This class comprises linear diﬀusion ﬁlters as well as spatial regularizations of the Perona–Malik process. they apply a diﬀusion process whose diﬀusivity is steered by derivatives of the evolving image. Methods of this type have been proposed for the ﬁrst time by Perona and Malik in 1987 [326].
secondmoment matrix). coupled diﬀusion–reaction methods. 286]. The discussed methods include linear and nonlinear diﬀusion ﬁltering. Lyapunov functionals. partial diﬀerential equations [185] and their numerical aspects [293. examples and applications of anisotropic diﬀusion ﬁltering. applications and generalizations. ordinary diﬀerential equations [56. The subsequent three chapters explore a theoretical framework for anisotropic diﬀusion ﬁltering. and total variation methods. We shall investigate under which conditions it is possible to get consistent approximations of the continuous anisotropic ﬁlter class which satisfy the abovementioned requirements. functional analysis [9. 58. for instance wellposedness results. Their qualities are illustrated using images arising from computer aided quality control and medical applications. 412]. though. Chapter 3 establishes conditions under which comparable wellposedness and scalespace results can be proved for the semidiscrete framework. extremum principles. Chapter 5 is devoted to practical topics such as ﬁlter design. and stability and an extremum principle are proved. Speciﬁc models are proposed which are tailored towards smoothing with edge enhancement and multiscale enhancement of coherent structures. Existence and uniqueness are discussed. This case takes into account the spatial discretization which is characteristic for digital images. 7]. We shall see that many numerical schemes share typical features with their semidiscrete counterparts. The prerequisites are kept to a minimum and can be found in standard textbooks on image processing [163]. It leads to nonlinear systems of ordinary diﬀerential equations. but also ﬁngerprint im . matrix analysis [407]. This corresponds to the fully discrete case which is treated in Chapter 4. Scalespace properties are investigated with respect to invariances and informationreducing qualities resulting from associated Lyapunov functionals. numerical methods. Euclidean and aﬃne invariant curve evolutions. This chapter also shows how one can design eﬃcient numerical methods which are in accordance with the fully discrete scalespace framework and which are based on an additive operator splitting (AOS). Chapter 2 presents a general model for the continuous setting where the diﬀusion tensor depends on the structure tensor (interest operator. a generalization of the Gaussiansmoothed gradient allowing a more sophisticated description of local image structure.PREFACE vii such that it should contain interesting material for many readers. In practice. and convergence to a constant steadystate. The book is organized as follows: Chapter 1 surveys the fundamental ideas behind PDEbased smoothing and restoration methods. PDE analogues of classical morphological processes. The investigated discrete ﬁlter class comes down to solving linear systems of equations which may arise from semiimplicit time discretizations of the semidiscrete ﬁlters. scalespaces can only be calculated for a ﬁnite number of scales. but it keeps the scalespace idea of using a continuous scale parameter. This general overview sketches their theoretical properties.
Copenhagen University). the Danish Research u Council. The Netherlands. and he provided the possibility to carry out this work at his laboratory. University Paris IX) invited me to the CEREMADE. The results are juxtaposed to related methods from Chapter 1. thesis. thesis [416]. which was written at the Department of Mathematics at the University of Kaiserslautern. Helmut Neunzert. During that time I also met Atsushi Imiya (Chiba University. October 1997 Joachim Weickert . and the EU–TMR Research Network VIRGO. Acknowledgments. In writing this book I have been helped and inﬂuenced by many people. Germany. Martin Reißel undertook the hard job of checking the whole manuscript for its mathematical correctness. This work has been funded by Stiftung Volkswagenwerk. October 1994) to an international audience. Stiftung Rheinland– Pfalz f¨r Innovation. drew my interest to diﬀusion processes in image processing. The proofreading of this book was done by Martin Reißel and Andrea Bechtold (Kaiserslautern). The present book is an extended and revised version of my Ph. In the latter ﬁeld I share many common interests with Jon Sporring. After the defence of my thesis in Kaiserslautern. I also thank him and the other editors of the ECMI Series as well as Teubner Verlag for their interest in publishing this work. Copenhagen. one of the birthplaces of many important ideas in this ﬁeld. The discussions and collaborations with the members of this group increased my interest in scalespace related deep structure analysis and information theory. Chapter 6 concludes the book by giving a summary and discussing possible future perspectives for nonlinear diﬀusion ﬁltering. Finally. He also gave me the honour to present my results as an invited speaker at the EMS Conference Multiscale Analysis in Image Processing (Lunteren. the Real World Computing Partnership. In the meantime I am with the computer vision group of Peter Johansen and Jens Arnspang (DIKU. Japan) at a workshop in Dagstuhl (Germany).D. Also Robert Maas (Utrecht University Hospital) contributed several useful hints. and it is a pleasure to take this opportunity to express my gratitude to them. and Andrea Bechtold was a great help in all kinds of diﬃculties with the English language.viii PREFACE ages and impressionistic paintings shall be processed. I joined the TGV (“tools for geometry in vision”) group at Utrecht University Hospital for 14 months.D. and he acted as a referee for the Ph. head of the Laboratory of Technomathematics. He introduced me into the fascinating world of early Japanese scalespace research conducted by Taizo Iijima decades before scalespace became popular in America and Europe. In this young and dynamic group I had the possibility to learn a lot about medical image analysis. and to experience Bart ter Haar Romeny’s enthusiasm for scalespace. Pierre–Louis Lions (CEREMADE.
equations . . . . . .6 Curvaturebased morphological processes . . . . . . . . . . . .6. . . .2. 1. . . . . . . . . . . . . .4 Methods of diﬀusion–reaction type . . . . 1. . . . . .6 Applications . . . . . . . . . . . . . . .6 Generalizations . . . .4 Generalizations . . . . .4 Applications . . . . . . . . . 1. . . . . . . . . . . 1. . . . . . . . . . . . 1. . . . . . . .3. . . . . .5. . 1. . . . . . .6. . . . . . . . . ix . . 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. . . . . . . 1. . . . . . . . . . . . 1. . . . . . . . . . . . . . 1. . . 1. . . . . .5. . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . .2 Regularized nonlinear models . . . .5 Classic morphological processes . . . 1. .5. . . . . 1 2 3 3 6 10 11 12 13 14 15 20 22 24 25 26 27 27 29 31 31 32 32 34 34 35 36 37 37 37 40 42 43 . . . . . . . . . . . . . . . . .1 Single diﬀusion–reaction equations . . . . . . . . . . . . . . . . . .1 Physical background of diﬀusion processes . . . . . . . . . .2 Linear diﬀusion ﬁltering . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . .2. . .7 Numerical aspects . . . . . .5 Scalespace properties . . . . . 1. . . . . . . . . . . . .2 Aﬃne invariant ﬁltering . . . . 1. . . . . . . . . . . 1. . . . . . . . . . . . . 1. . . . . .4 Theoretical results . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . 1. . . . .4 Numerical aspects . .2.Contents 1 Image smoothing and restoration by PDEs 1. .5 Numerical aspects . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . .3 Nonlinear diﬀusion ﬁltering . . . .2 Basic operations . . . . . . . . . . 1. .8 Applications . .3 Anisotropic nonlinear models . . . . . . . .2. . . . .3 Numerical aspects .6. .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. . .1 Meancurvature ﬁltering . . . . 1. . . 1.3 Generalizations . . .2 Coupled systems of diﬀusion–reaction 1. . . . . . . . . . . . . . . . . .1 Binary and greyscale morphology . . . .6 Generalizations . 1. . . . . .5. . . . . . . . . . . . . . .3. . . . . .3.3 Continuousscale morphology . . . . . . . . . . . . . . . . . 1. .1 Relations to Gaussian smoothing . . . . 1.5. . 1.5. . .6. .4. . . 1. . . 1. . . . . . . . .2 Scalespace properties . . . . 1. . . . . . 1. . .2. . . . . . .5. . . . . . . . . . .5 Limitations . . . . . . . . . . . . . . . . . .1 The Perona–Malik model . . .
. . . . . . . . . . . . . . . . . . . .2 The structure tensor . . 3. . . . . . . . . . . . . . . . . . . . . . . 1. .2 Informationreducing properties 3 Semidiscrete diﬀusion ﬁltering 3. . . . . .x 1. . . . . 3. . . . . . . . . . 5 Examples and applications 5. . . . . . . . . . . . . . . . . .2 Applications . . . . . . . .3 Theoretical results . . . . 3. . . . .7. . .2 Theoretical results . . .4 Scalespace properties . . . . .6 Active contour models . . . . . . . . . . . .1 Isotropic case . . . . . . . . . . . . . . . . . .1. . . . Conclusions and further scope of the book . . . . .2 TVminimizing methods . . . . 5. . . . . .6. . . . . . . . . . . . 3. . .2 AOS schemes . . .1 Edgeenhancing diﬀusion . . . . . . . . . . . . . . . . . . .1 Filter design . . . . . . . . . . . . . . . . . . . . . . 6 Conclusions and perspectives Bibliography Index .4. . .3 Scalespace properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1. . . . . . . . . . . . . . . 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Invariances . . . . . . . . . . . . . . . . . . . 4. . . . .4 Relation to continuous models 3. . .1 TVpreserving methods . . .7. . . . .3 Scalespace properties . . . . 5. . . . . . . . . . . .8 2 Continuous diﬀusion ﬁltering 2. . . . . . . .1 Semiimplicit schemes . . . . . . . . 45 46 49 50 50 53 55 55 56 57 62 62 65 75 75 76 81 86 86 88 97 97 98 99 102 102 107 113 114 114 114 127 127 129 135 139 165 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Total variation methods . . . . . . . . . . . . . . . . 2. 4. . .2 Anisotropic case . . .4. . . . . . . . .2. . 1. .4. . . . . .2 Theoretical results . . .2 Applications . . . . . . . . . . . . . . . .2. . . . . . . . . 4 Discrete diﬀusion ﬁltering 4. . . . . . . . . . . .5 Applications . . . . . . . . . . . . . . . . . . . . . 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Relation to semidiscrete models 4. .1 Filter design . . 1. . . . . . . . . . . . . . 5. .7 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Basic ﬁlter structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CONTENTS . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . .4. . . . . .1 The general model . . . . . 5. . . . . 2. . .2 Coherenceenhancing diﬀusion 5. . . . . .6. . 2. . . . . . . . . . . . . . . . . . . . . . .1 The general model . . . . 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
the study of salinity proﬁles in oceanography. they are mentioned wherever they appear. a topic which looks at ﬁrst glance fairly diﬀerent to the diﬀusion approach. The outline of this chapter is as follows: We start with reviewing the physical ideas behind diﬀusion processes. in order to allow the interested reader to pursue these ideas. generalizations. namely the smoothing and restoration of images. This chapter reviews their main application. and applications are discussed. The goal is to make the reader sensitive to the similarities. This helps us to better understand the next sections which are concerned with the properties of linear and nonlinear diﬀusion ﬁlters in image processing. Also many historical notes are added. it reveals some interesting relations when it is interpreted within a PDE framework. numerical aspects. After that we investigate morphological ﬁlters. Nevertheless. The subsequent study of image enhancement methods of diﬀusion–reaction type relates diﬀusion ﬁlters to variational image restoration techniques. where proofs and full mathematical details can be found. crack propagation. Many of these ideas are borrowed from physical phenomena such as wave propagation or transport of heat and mass. and to point out the main results and open problems in this rapidly evolving area. grassﬁre ﬂow.Chapter 1 Image smoothing and restoration by PDEs PDEbased methods appear in a large variety of image processing and computer vision areas ranging from shapefromshading and histogramme modiﬁcation to optic ﬂow and stereo vision. also gas dynamics. Nevertheless. For each class of methods the basic ideas are explained and their theoretical background. This becomes 1 . Although a detailed discussion of these connections would be far beyond the scope of this work. advantages and shortcomings of these techniques. diﬀerences. or mechanisms of the retina and the brain are closely related to some of these approaches. It is written in an informal style and refers to a large amount of original literature.
Diﬀusion which does not depend on the evolving image is called linear. even if it uses a scalarvalued diﬀusivity instead of a diﬀusion tensor.2) where t denotes the time. Finally we shall discuss total variation image restoration techniques which permit discontinuous solutions. The observation that diﬀusion does only transport mass without destroying it or creating new mass is expressed by the continuity equation ∂t u = −div j (1. . Then we may replace the diﬀusion tensor by a positive scalarvalued diﬀusivity g. PARTIAL DIFFERENTIAL EQUATIONS especially evident when considering curvaturebased morphological PDEs. If the diﬀusion tensor is constant over the whole image domain. In image processing we may identify the concentration with the grey value at a certain location.3) This equation appears in many physical transport processes. and a spacedependent ﬁltering is called inhomogeneous. In the general anisotropic case. In the context of heat transfer it is called heat equation. a positive deﬁnite symmetric matrix.2 CHAPTER 1. (1. and inhomogeneous blurring is called anisotropic. 1. The case where j and ∇u are parallel is called isotropic. If we plug in Fick’s law into the continuity equation we end up with the diﬀusion equation ∂t u = div (D · ∇u).1 Physical background of diﬀusion processes Most people have an intuitive impression of diﬀusion as a physical process that equilibrates concentration diﬀerences without creating or destroying mass. Such a feedback leads to nonlinear diﬀusion ﬁlters. The last section summarizes the advantages and shortcomings of the main methods and gives an outline of the questions we are concerned with in the subsequent chapters.1) This equation states that a concentration gradient ∇u causes a ﬂux j which aims to compensate for this gradient. The equilibration property is expressed by Fick’s law: j = −D · ∇u. The relation between ∇u and j is described by the diﬀusion tensor D. This physical observation can be easily cast in a mathematical formulation. one speaks of homogeneous diﬀusion. Sometimes the computer vision literature deviates from the preceding notations: It can happen that homogeneous ﬁltering is named isotropic. (1. j and ∇u are not parallel. Often the diﬀusion tensor is a function of the diﬀerential structure of the evolving image itself.
6) we obtain by the convolution theorem that (F (Kσ ∗f )) (ω) = (F Kσ )(ω) · (F f )(ω). let us investigate the behaviour in the frequency domain.2 LINEAR DIFFUSION FILTERING 3 1.2. even if f is only absolutely integrable. We shall focus on the relation between linear diﬀusion ﬁltering and the convolution with a Gaussian.1 Relations to Gaussian smoothing Gaussian smoothing Let a greyscale image f be represented by a realvalued mapping f ∈ L1 (IR2 ). Afterwards a survey on discrete aspects is given and applications and limitations of the linear diﬀusion paradigm are discussed.5) 2πσ 2 2σ There are several reasons for the excellent smoothing properties of this method: First we observe that since Kσ ∈ C ∞ (IR2 ) we get Kσ ∗f ∈ C ∞ (IR2 ). Since the Fourier transform of a Gaussian is again Gaussianshaped. (F Kσ )(ω) = exp − ω2 . x ) dx (1. A widelyused way to smooth f is by calculating the convolution (Kσ ∗f )(x) := Kσ (x−y) f (y) dy IR 2 (1.4) where Kσ denotes the twodimensional Gaussian of width (standard deviation) σ >0 : 1 x2 Kσ (x) := · exp − 2 . Next.8) (1.7) . The section is concluded by sketching two linear generalizations which can incorporate apriori knowledge: aﬃne Gaussian scalespace and directed diﬀusion processes.2 Linear diﬀusion ﬁltering The simplest and best investigated PDE method for smoothing images is to apply a linear diﬀusion process. When deﬁning the Fourier transformation F by (F f )(ω) := IR2 f (x) exp(−i ω.1. and review the fundamental properties of the Gaussian scalespace induced by linear diﬀusion ﬁltering. analyse its smoothing properties for the image as well as its derivatives. 1. (1. 2/σ 2 (1.
(1. u(x. t) = f (x) (K√2t ∗ f )(x) (t = 0) (t > 0).12) and it (1. (1. it is called illposed.g. Figure 5. pp. smoothing structures of order σ requires to stop the diﬀusion process at time 1 (1.10) This solution is unique. diﬀerentiation is illposed1 . [331.13) It depends continuously on the initial image f with respect to fulﬁls the maximum–minimum principle inf f ≤ u(x. PARTIAL DIFFERENTIAL EQUATIONS we observe that (1. if it has a unique solution which depends continuously on the input data and parameters. 43–56]) that for any bounded f ∈ C(IR2 ) the linear diﬀusion process ∂t u = ∆u. However. L∞ (IR2 ) . 1 .4) is a lowpass ﬁlter that attenuates high frequencies in a monotone way. A thorough A problem is called wellposed. provided we restrict ourselves to functions satisfying u(x. .9) (1. If one of these conditions is violated. a > 0).14) T = 2 σ2 . the need for regularization methods arises.2 (b) and 5.4 CHAPTER 1. that is to say.3 (c) in Chapter 5 illustrate the eﬀect of linear diﬀusion ﬁltering. Equivalence to linear diﬀusion ﬁltering It is a classical result (cf. Interestingly. Hence. we need information about its derivatives.11) (1. the smoothing behaviour can also be understood in the context of a PDE interpretation. Hence. pp. as small perturbations in the original image can lead to arbitrarily large ﬂuctuations in the derivatives. t) ≤ sup f 2 IR IR2 on IR2 × [0. ∞).11) we observe that the time t is related to the spatial width σ = 2t of the Gaussian. √ From (1. 0) = f (x) possesses the solution u(x. t) ≤ M · exp (ax2 ) (M. e. Gaussian derivatives In order to understand the structure of an image we have to analyse grey value ﬂuctuations within a neighbourhood of each image point. 267–271] and [185.
no new zerocrossings are created which cannot be traced back to ﬁner scales [439]. for instance ∇Kσ ∗u or ∆Kσ ∗u. increasing the order of applied Gaussian derivatives or reducing the kernel size will ﬁnally deteriorate the results of deblurring. 177].2 LINEAR DIFFUSION FILTERING 5 treatment of this mathematical theory can be found in the books of Tikhonov and Arsenin [402]. To illustrate this. we focus on two applications for detecting edges. [128]. which uses the LaplacianofGaussian (LoG) ∆Kσ as convolution kernel. ridges. This method is often acknowledged to be the best linear edge detector. expressions that are invariant under transformations such as rotations. It is Of course.1. Another important edge detector is the Marr–Hildreth operator [278]. see [256] for an overview. After applying sophisticated thinning and linking mechanisms (nonmaxima suppression and hysteresis thresholding). Moreover. and Koenderink and van Doorn suggested the set of Gaussian derivatives as a general model for the visual system [242]. Replacing derivatives by these Gaussian derivatives has a strong regularizing eﬀect. we observe that all derivatives undergo the same Gaussian smoothing process as the image itself and this process is equivalent to convolving the image with derivatives of a Gaussian. It is based on calculating the ﬁrst derivatives of the Gaussiansmoothed image. If one investigates the temporal evolution of the zerocrossings of an image ﬁltered by linear diﬀusion. edges are identiﬁed as locations where the gradient magnitude has a maximum. and it has become a standard in edge detection. One possibility to regularize is to convolve the image with a Gaussian prior to diﬀerentiation [404]. see [278] and the references therein. solutions of the regularization can only approximate the solution of the original problem (if it exists). By the equality n m n m n m ∂x1 ∂x2 (Kσ ∗f ) = Kσ ∗ (∂x1 ∂x2 f ) = (∂x1 ∂x2 Kσ ) ∗ f (1. and blobs. junctions. Louis [266] and Engl et al. Gaussian derivatives can be combined to socalled diﬀerential invariants. A frequently used method is the Canny edge detector [69]. This property has been used to stabilize illposed problems like deblurring images by solving the heat equation backwards in time2 [141. There are indications that LoGs and especially their approximation by diﬀerencesofGaussians (DoGs) play an important role in the visual system of mammals. Diﬀerential invariants are useful for the detection of features such as edges.15) for suﬃciently smooth f . one observes an interesting phenomenon: When increasing the smoothing scale σ. Young developed this theory further by presenting evidence that the receptive ﬁelds in primate eyes are shaped like the sum of a Gaussian and its Laplacian [449]. Edges of f are identiﬁed as zerocrossings of ∆Kσ ∗f . 2 . This evolution property is called causality [240]. In practice. This needs no further postprocessing and always gives closed contours.
Attempts to reconstruct the original image from the temporal evolution of the zerocrossings of the Laplacian have been carried out by Hummel and Moniot [190]. however. In the western world the evolution property of the zerocrossings was the key investigation which has inspired Witkin to the socalled scalespace concept [439]. Smoothing properties and information reduction arise from the wish that the transformation should not create artifacts when passing from ﬁne to coarse representation. 450. smoothing (informationreducing) or invariance requirements [12]. and the ﬁltering may be split into a sequence of ﬁlter banks: T0 f = f.6 CHAPTER 1.16) (1. 189. 1. Most of these properties can be classiﬁed as architectural. A scalespace is an image representation at a continuum of scales. This simpliﬁcation property is speciﬁed by numerous authors in diﬀerent ways. 257]. Moreover. They concluded. that this is practically unstable unless very much additional information is provided. Tt+s f = Tt (Ts f ) ∀ s. t ≥ 0. i.e. This shall be discussed next. provided that it fulﬁls certain requirements3 . (1. 3 .2 Scalespace properties The general scalespace concept It is a wellknown fact that images usually contain structures at a large variety of scales. In those cases where it is not clear in advance which is the right scale for the depicted information it is desirable to have an image representation at multiple scales. we should not have additional structures which are caused by the ﬁltering method itself and not by underlying structures at ﬁner scales. Other architectural principles comprise for instance regularity properties of Tt and local behaviour as t tends to 0. An important architectural assumption is recursivity. using concepts such as no creation of new level curves (causality) [240.2. at a coarse scale. 255].17) This property is very often referred to as the semigroup property. PARTIAL DIFFERENTIAL EQUATIONS closely connected to the maximum–minimum principle of certain parabolic operators [189]. nonenhancement of local extrema [30. for t = 0. the scalespace representation gives the original image f . embedding the image f into a family {Tt f t ≥ 0} of gradually simpliﬁed versions of it. by comparing the structures at diﬀerent scales. one obtains a hierarchy of image structures which eases a subsequent image interpretation. decreasing number Recently it has also been proposed to extend the scalespace concept to scale–imprecision space by taking into account the imprecision of the measurement device [171]. Thus.
This class of blurring transformations is called boke (defocusing). 320]. Ishikawa and Imiya [426. comparison principle [12]. In [192] Iijima considers an observation transformation Φ which depends on a scale parameter σ and which transforms the original image f (x) into a blurred version4 Φ[f (x′ ). He assumes that it has the structure Φ[f (x ). or – equivalently – by linear diﬀusion ﬁltering according to (1.2 LINEAR DIFFUSION FILTERING 7 of local extrema [255]. b ∈ IR. Often these requirements are supplemented with an additional assumption which is equivalent to the superposition principle. Thus. 303]. Guichard.10). which is obtained via convolution with Gaussians of increasing variance. see [426] for more details. maximum–minimum principle [189. σ]. 193. 138]. x. σ] = 4 ′ ∞ −∞ φ{f (x′ ). and Lyapunov functionals [415. 16]. Lions and Morel [12] shows that every scalespace fulﬁlling some fairly natural architectural. maximum loss of ﬁgure impression [196]. 429]. translations and rotations or even more complicated transformations such as aﬃne mappings. (1. in order to analyse only the depicted object [196. ∀ a. x. positivity [324. x. 427]. Especially in the linear setting. namely linearity: Tt (af + bg) = a Tt f + b Tt g ∀ t ≥ 0. 328]. imposing linearity restricts the scalespace idea to essentially one representative. PDEs are the suitable framework for scalespaces. informationreducing and invariance axioms is governed by a PDE with the original image as initial condition.1. This makes the requirement plausible that the scalespace analysis should be invariant to as many of these transformations as possible. Usually a 1983 paper by Witkin [439] or a 1980 report by Stansﬁeld [392] are regarded as the ﬁrst references to the linear scalespace idea. preservation of positivity [191. σ} dx′ . Two images of this class diﬀer e. (1.9). (1.g. shows that scalespace is more than 20 years older: An axiomatic derivation of 1D Gaussian scalespace has already been presented by Taizo Iijima in a technical paper from 1959 [191] followed by a journal version in 1962 [192]. Tikhonov regularization [302. We may regard an image as a representative of an equivalence class containing all images that depict the same object. Gaussian scalespace The historically ﬁrst and best investigated scalespace is the Gaussian scalespace. however. x′ .19) The variable x′ serves as a dummy variable. Recent work by Weickert. many of these properties are equivalent or closely related.18) As we shall see below. Both papers are written in Japanese. The pioneering work of Alvarez. . by greylevel shifts.
8
CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
and that it should satisfy ﬁve conditions: (I) Linearity (with respect to multiplications): If the intensity of a pattern becomes A times its original intensity, then the same should happen to the observed pattern: Φ[Af (x′ ), x, σ] = A Φ[f (x′ ), x, σ]. (1.20)
(II) Translation invariance: Filtering a translated image is the same as translating the ﬁltered image: Φ[f (x′ −a), x, σ] = Φ[f (x′ ), x−a, σ]. (1.21)
(III) Scale invariance: If a pattern is spatially enlarged by some factor λ, then there exists a σ ′ = σ ′ (σ, λ) such that Φ[f (x′ /λ), x, σ] = Φ[f (x′ ), x/λ, σ ′ ]. (1.22)
(IV) (Generalized) semigroup property: If f is observed under a parameter σ1 and this observation is observed under a parameter σ2 , then this is equivalent to observing f under a suitable parameter σ3 = σ3 (σ1 , σ2 ): Φ Φ[f (x′′ ), x′ , σ1 ], x, σ2 = Φ[f (x′′ ), x, σ3 ]. (1.23)
(V) Preservation of positivity: If the original image is positive, then the observed image is positive as well: Φ[f (x′ ), x, σ] > 0 ∀ f (x′ ) > 0, ∀ σ > 0. (1.24)
Under these requirements Iijima derives in a very systematic way that 1 Φ[f (x ), x, σ] = √ 2 πσ
′ ∞ −∞
f (x′ ) exp
−(x − x′ )2 4σ 2
dx′ .
(1.25)
Thus, Φ[f (x′ ), x, σ] is just the convolution between f and a Gaussian with standard √ deviation σ 2. This has been the starting point of an entire world of linear scalespace research in Japan, which is basically unknown in the western world. Japanese scalespace theory was wellembedded in a general framework for pattern recognition, feature extraction and object classiﬁcation [195, 197, 200, 320], and many results have
1.2 LINEAR DIFFUSION FILTERING
9
been established earlier than in the western world. Apart from their historical merits, these Japanese results reveal many interesting qualities which should induce everyone who is interested in scalespace theory to have a closer look at them. More details can be found in [426, 427] as well as in some English scalespace papers by Iijima such as [195, 197]. In particular, the latter ones show that there is no justiﬁcation to deny Iijima’s pioneering role in linear scalespace theory because of language reasons. Table 1.1: Overview of continuous Gaussian scalespace axiomatics (I1 = Iijima [191, 192], I2 = Iijima [193, 194], I3 = Iijima [196], O = Otsu [320], K = Koenderink [240], Y = Yuille/Poggio [450], B = Babaud et al. [30], L1 = Lindeberg [255], F1 = Florack et al. [140], A = Alvarez et al. [12], P = Pauwels et al. [324], N = Nielsen et al. [303], L2 = Lindeberg [257], F2 = Florack [138]).
convolution kernel semigroup property locality regularity inﬁnetes. generator max. loss principle causality nonnegativity Tikhonov regulariz. aver. grey level invar. ﬂat kernel for t → ∞ isometry invariance homogen. & isotropy separability scale invariance valid for dimension I1 • • I2 • • I3 O • K Y • • • • • • • • 1 • 2 • • • • 2 2 • 1,2 • • B • • • • • • 1 L1 • • • • • • F1 • • • A • • • • P • • • • • • • • 1,2 N • L2 • • • • • • • N • F2 • • •
• • • • • >1 • •
•
• • • 1,2
• • N
1
N
N
Table 1.1 presents an overview of the current Japanese and western Gaussian scalespace axiomatics (see [426, 427] for detailed explanations). All of these axiomatics use explicitly or implicitly5 a linearity assumption. We observe that – despite the fact that many axiomatics reveal similar ﬁrst principles – not two of them are identical. Each of the 14 axiomatics conﬁrms and enhances the evidence that the others give: that Gaussian scalespace is unique within a linear framework. A detailed treatment of Gaussian scalespace theory can be found in two Japanese monographs by Iijima [197, 198], as well as in English books by Lindeberg [256], Florack [139], and ter Haar Romeny [176]. A collection edited by
Often it is assumed that the ﬁlter is a convolution integral. This is equivalent to linearity and translation invariance.
5
10
CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
Sporring, Nielsen, Florack and Johansen [389] gives an excellent overview of the various aspects of this theory, and additional material is presented in [211]. Many relations between Gaussian scalespace and regularization theory have been elaborated by Nielsen [302], and readers who wish to analyse linear and nonlinear scalespace concepts in terms of diﬀerential and integral geometry can ﬁnd a lot of material in the thesis of Salden [351].
1.2.3
Numerical aspects
The preceding theory is entirely continuous. However, in practical problems, the image is sampled at the nodes (pixels) of a ﬁxed equidistant grid. Thus, the diﬀusion ﬁlter has to be discretized. By virtue of the equivalence of solving the linear diﬀusion equation and convolving with a Gaussian, we can either approximate the convolution process or the diﬀusion equation. When restricting the image to a ﬁnite domain and applying the Fast Fourier Transformation (FFT), convolution in the spatial domain can be reduced to multiplication in the frequency domain, cf. (1.7). This proceeding requires a ﬁxed computational eﬀort of order N log N, which depends only on the pixel number N, but not on the kernel size σ. For large kernels this is faster than most spatial techniques. Especially for small kernels, however, aliasing eﬀects in the Fourier domain may create oscillations and over and undershoots [178]. One eﬃcient possibility to approximate Gaussian convolution in the spatial domain consists of applying recursive ﬁlters [109, 448]. More frequently the Gaussian kernel is just sampled and truncated at some multiple of its standard deviation σ. Factorizing a higherdimensional Gaussian into onedimensional Gaussians reduces the computational eﬀort to O(Nσ). Convolution with a truncated Gaussian, however, reveals the drawback that it does not preserve the semigroup property of the continuous Gaussian scalespace [255]. Lindeberg [255] has established a linear scalespace theory for the semidiscrete6 case. His results are in accordance with those of Norman [312], who proposed in 1960 that the discrete analogue of the Gaussian kernel should be given in terms of modiﬁed Bessel functions of integer order. Since this scalespace family arises naturally from a semidiscretized version of the diﬀusion equation, it has been argued that approximating the diﬀusion equation should be preferred to discretizing the convolution integral [255]. Recently, interesting semidiscrete and fully discrete linear scalespace formulations have been established utilizing stochastic principles: ˚str¨m and Heyden [27] A o study a framework based on stationary random ﬁelds, while the theory by Salden et al. [353] exploits the relations between diﬀusion and Markov processes.
6
By semidiscrete we mean discrete in space and continuous in time throughout this work.
258. • for multiscale segmentation of images [172. 1. It can be found in almost every standard textbook in these ﬁelds. while scalespace analysis intends to extract semantical information by tracing signals through a continuum of scales. 256. A very eﬃcient approximation of the Gaussian scalespace results from applying multigrid ideas. 8 Historically. 67.2.1. Apart from some implicit approaches [166. however. linear diﬀusion ﬁltering has been applied in numerous ﬁelds of image processing and computer vision. pyramid decompositions have become very popular and have been integrated into commercially available hardware [70. Less frequent are applications which exploit the evolution of an image under Gaussian scalespace. By subsequently smoothing the image with an explicit scheme for the diﬀusion equation and restricting the result to a coarser grid.4 Applications Due to its equivalence to convolution with a Gaussian. The Gaussian pyramid [64] has the computational complexity O(N) and gives a multilevel representation at ﬁnitely many scales of diﬀerent resolution. 408]. Several scalespace applications are summarized in a survey paper by ter Haar Romeny [175]. 106] are just designed to have few or no redundancies. focusofattention). 306. 254. 241. ﬁnite diﬀerence (FD) schemes dominate the ﬁeld. 259. this is incorrect: Iijima’s scalespace work [191] is much older than multigrid ideas in image processing. 313. 388] or generalized entropies [390] over scales. The idea is to identify segments at coarse scales and to link backwards to the original image in order to improve the localization. Of course. Pyramids are not invariant under translations. and sometimes it is argued that they are undersampled and that the pyramid levels should be closer7 . Due to their simplicity and eﬃciency. 10. multiresolution techniques such as pyramids or discrete wavelet transforms [92. for instance • for ﬁnding the most relevant scales (scale selection.2 LINEAR DIFFUSION FILTERING 11 Among the numerous numerical possibilities to approximate the linear diﬀusion equation. one obtains a simpliﬁed image representation at the next coarser grid. 7 . This deep structure analysis [240] provides useful information for extracting semantic information from an image. In recent years also applications of Gaussian scalespace to stereo. This may be done by searching for extrema of (nonlinear) combinations of normalized Gaussian derivatives [256] or by analysing information theoretic measures such as the entropy [208. 451]. explicit schemes are mainly used. 441]. 302. 214]. 215. 68] allowing realizations as a recursive ﬁlter [14. These are the reasons why some people regard pyramids rather as predecessors of the scalespace idea than as a numerical approximation8 . optic ﬂow and image sequences have become an active research ﬁeld [139.
PARTIAL DIFFERENTIAL EQUATIONS Interesting results arise when one studies linear scalespace on a sphere [236. In practice. Part III of the book edited by Sporring et al. Due to the uniqueness of Gaussian scalespace within a linear framework we know that any modiﬁcation in order to overcome the problems (a)–(c) will either . and it is generally not true that the number of local extrema is nonincreasing. for instance by analysing its toppoints [210]. but happens generically. Witkin [439]. Since Gaussian smoothing is designed to be completely uncommitted. [389] and the references therein give an overview of the stateoftheart in deep structure analysis. There is some evidence that these points. it cannot take into account any apriori information on structures which are worth being preserved (or even enhanced). Regarding (b) and (c).5 Limitations In spite of several properties that make linear diﬀusion ﬁltering unique and easy to handle. thus. (b) Linear diﬀusion ﬁltering dislocates edges when moving from ﬁner to coarser scales. This and other results [12. (c) Some smoothing properties of Gaussian scalespace do not carry over from the 1D case to higher dimensions: A closed zerocrossing contour can split into two as the scale increases [450]. 353]: while the diﬀusion equation remains the correct concept. 255] for illustrative counterexamples. It turned out that the pairwise creation of an extremum and a saddle point is not an exception. see e. carry essential image information [212]. A deep mathematical analysis of such phenomena has been carried out by Damon [105] and Rieger [342]. much eﬀorts have been spent in order to understand the deep structure in Gaussian scalespace. where the gradient vanishes and the Hessian does not have full rank.2. relating dislocated information obtained at diﬀerent scales is diﬃcult and bifurcations may give rise to instabilities. see [254.g. 255] indicate that the PDE formulation of linear scalespace in terms of a diﬀusion equation is more natural and has a larger generalization potential than convolution with Gaussians.12 CHAPTER 1. but also blurs important features such as edges and. Gaussian kernels are of no use anymore: appropriate kernels have to be expressed in terms of Legendre functions [236]. So structures which are identiﬁed at a coarse scale do not give the right location and have to be traced back to the original image [439. These coarsetoﬁne tracking diﬃculties are generally denoted as the correspondence problem. it reveals some drawbacks as well: (a) An obvious disadvantage of Gaussian smoothing is the fact that it does not only smooth noise. 165]. makes them harder to identify. 1. 38.
2) of f . t > 0. let us ﬁrst investigate two linear modiﬁcations which have been introduced in order to address the problems (a) and (b) from the previous section. In 1992 Nitzberg and Shiota [310] proposed to adapt the Gaussian kernel shape to the structure of the original image.1. (1. If one wants to relate 9 Isotropic Gaussian scalespace can be recovered using the unit matrix for D. 218].27) the basic equation of ﬁgure [196].27). In practice this can be experienced by the fact that shapeadaptation of Gaussian smoothing does not preserve the average grey level. Section 2. t) := IR 2 −1 (x−y)⊤ Dt (x−y) exp − 4 4π det(Dt ) 1 f (y) dy (1. while the divergence formulation ensures that this is still possible for nonuniform diﬀusion ﬁltering. We shall see that appropriate methods to avoid the shortcomings (a) and (b) are nonlinear diﬀusion processes. For a ﬁxed matrix D. 200]. and D ∈ IR2×2 is symmetric positive deﬁnite9 . see Section 1. In 1971 this concept was realized in hardware in the optical character reader ASPET/71 [199. By chosing D in (1. 1.26) as a function of the structure tensor (cf. .1. He named u(x. calculating the convolution integral (1. Aﬃne Gaussian scalespace A straightforward generalization of Gaussian scalespace results from renouncing invariance under rotations.27) (1.26) is equivalent to solving a linear anisotropic diﬀusion problem with D as diﬀusion tensor: ∂t u = div (D ∇u). they combined nonlinear shape adaptation with linear smoothing. 0) = f (x). In [427] it is shown that aﬃne Gaussian scalespace has been axiomatically derived by Iijima in 1962 [193. 207.6 Generalizations Before we turn our attention to nonlinear processes. It should be noted that shapeadapted Gaussian smoothing with a spatially varying D is no longer equivalent to a diﬀusion process of type (1. Later on similar ideas have been developed in [259. while (c) requires morphological equations [206. 443]. 194].2 LINEAR DIFFUSION FILTERING 13 renounce linearity or some scalespace properties.26) where Dt := tD. The scalespace part has been regarded as the reason for its reliability and robustness. t) the generalized ﬁgure of f . This leads to the aﬃne Gaussian scalespace u(x. and (1.28) u(x. Also in this case the diﬀusion equation seems to be more general.2.
29) (1. especially its illposedness aspects. This gives rise to study regularizations. one has to carry out sophisticated scaling limits [310]. Such concepts will be investigated in Section 1. by prescribing that features within a certain contrast and scale range are considered to be semantically important and processed diﬀerently.30) u(x. . PARTIAL DIFFERENTIAL EQUATIONS shapeadapted Gaussian smoothing to a PDE.14 CHAPTER 1.3. Although this concept is wellknown in the image processing community (see [349] and the references therein for an overview). Such a directed diﬀusion process requires to specify an entire image as background information in advance. (1. Such demands can be satisﬁed by nonlinear diﬀusion ﬁlters. These techniques can be extended to anisotropic processes which make use of an adapted diﬀusion tensor instead of a scalar diﬀusivity. Numerical experiments have been carried out by Giuliani [159]. 1. and an analysis in terms of nonsmooth b and weak solutions is due to Illner and Tie [203]. in many applications it would be desirable to include a priori knowledge in a less speciﬁc way. a corresponding PDE formulation was ﬁrst given by Perona and Malik [326] in 1987. Provided we are given some background information in form of a smooth image b. We shall discuss this model in detail. e. they show that under some technical requirements and suitable boundary conditions the classical solution u of ∂t u = b ∆u − u ∆b. 0) = f (x) converges to b along a path where the relative entropy with respect to b increases in a monotone way.3 Nonlinear diﬀusion ﬁltering Adaptive smoothing methods are based on the idea of applying a process which itself depends on local properties of the image. Noniterative shapeadapted Gaussian smoothing diﬀers from nonlinear anisotropic diﬀusion ﬁltering by the fact that the latter one introduces a feedback into the process: it adapts the diﬀusion tensor in (1.27) to the diﬀerential structure of the ﬁltered image instead of the original image.g.3 and in the remaining chapters of this book. Directed diﬀusion Another method for incorporating apriori knowledge into a linear diﬀusion process is suggested by Illner and Neunzert [202].
1.3 NONLINEAR DIFFUSION FILTERING
15
1.3.1
The Perona–Malik model
Basic idea Perona and Malik propose a nonlinear diﬀusion method for avoiding the blurring and localization problems of linear diﬀusion ﬁltering [326, 328]. They apply an inhomogeneous process that reduces the diﬀusivity at those locations which have a larger likelihood to be edges. This likelihood is measured by ∇u2. The Perona– Malik ﬁlter is based on the equation ∂t u = div (g(∇u2) ∇u). and it uses diﬀusivities such as g(s2) = 1 1 + s2 /λ2 (λ > 0). (1.32) (1.31)
Although Perona and Malik name their ﬁlter anisotropic, it should be noted that – in our terminology – it would be regarded as an isotropic model, since it utilizes a scalarvalued diﬀusivity and not a diﬀusion tensor. Interestingly, there exists a relation between (1.31) and the neural dynamics of brightness perception: In 1984 Cohen and Grossberg [94] proposed a model of the primary visual cortex with similar inhibition eﬀects as in the Perona–Malik model. The experiments of Perona and Malik were visually very impressive: edges remained stable over a very long time. It was demonstrated [328] that edge detection based on this process clearly outperforms the linear Canny edge detector, even without applying nonmaxima suppression and hysteresis thresholding. This is due to the fact that diﬀusion and edge detection interact in one single process instead of being treated as two independent processes which are to be applied subsequently. Moreover, there is another reason for the impressive behaviour at edges, which we shall discuss next. Edge enhancement To study the behaviour of the Perona–Malik ﬁlter at edges, let us for a moment restrict ourselves to the onedimensional case. This simpliﬁes the notation and illustrates the main behaviour since near a straight edge a twodimensional image approximates a function of one variable. For the diﬀusivity (1.32) it follows that the ﬂux function Φ(s) := sg(s2 ) satisﬁes ′ Φ (s) ≥ 0 for s ≤ λ, and Φ′ (s) < 0 for s > λ, see Figure 1.1. Since (1.31) can be rewritten as ∂t u = Φ′ (ux )uxx , (1.33) we observe that – in spite of its nonnegative diﬀusivity – the Perona–Malik model is of forward parabolic type for ux  ≤ λ, and of backward parabolic type for ux  > λ.
16
1
CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
0.6 diffusivity flux function
0 0 lambda
0 0 lambda
Figure 1.1: (a) Left: Diﬀusivity g(s2 ) = 1+s1 /λ2 . (b) Right: Flux function 2 Φ(s) = 1+ss /λ2 . 2
Hence, λ plays the role of a contrast parameter separating forward (low contrast) from backward (high contrast) diﬀusion areas. It is not hard to verify that the Perona–Malik ﬁlter increases the slope at inﬂection points of edges within a backward area: If there exists a suﬃciently smooth solution u it satisﬁes ∂t (u2 ) = 2ux ∂x (ut ) = 2Φ′′ (ux )ux u2 + 2Φ′ (ux )ux uxxx. x xx (1.34)
A location x0 where u2 is maximal at some time t is characterized by ux uxx = 0 x and ux uxxx ≤ 0. Therefore,
2 (∂t (ux )) (x0 , t) ≥ 0 for ux (x0 , t) > λ
(1.35)
with strict inequality for ux uxxx < 0. In the twodimensional case, (1.33) is replaced by [12] ∂t u = Φ′ (∇u)uηη + g(∇u2)uξξ (1.36)
where the gauge coordinates ξ and η denote the directions perpendicular and parallel to ∇u, respectively. Hence, we have forward diﬀusion along isophotes (i.e. lines of constant grey value) combined with forward–backward diﬀusion along ﬂowlines (lines of maximal grey value variation). We observe that the forward–backward diﬀusion behaviour is not only restricted to the special diﬀusivity (1.32), it appears for all diﬀusivities g(s2 ) whose rapid decay causes nonmonotone ﬂux functions Φ(s) = sg(s2 ). Overviews of several common diﬀusivities for the Perona–Malik model can be found in [43, 343], and a family of diﬀusivities with diﬀerent decay rates is investigated in [36]. Rapidly decreasing diﬀusivities are explicitly intended in the Perona–Malik method as they
1.3 NONLINEAR DIFFUSION FILTERING
17
give the desirable result of blurring small ﬂuctuations and sharpening edges. Therefore, they are the main reason for the visually impressive results of this restoration technique. It is evident that the “optimal” value for the contrast parameter λ has to depend on the problem. Several proposals have been made to facilitate such a choice in practice, for instance adapting it to a speciﬁed quantile in the cumulative gradient histogramme [328], using statistical properties of a training set of regions which are considered as ﬂat [444], or estimating it by means of the local image geometry [270]. Illposedness Unfortunately, forward–backward equations of Perona–Malik type cause some theoretical problems. Although there is no general theory for nonlinear parabolic processes, there exist certain frameworks which allow to establish wellposedness results for a large class of equations. Let us recall three examples: • Let S(N) denote the set of symmetric N × N matrices and Hess(u) the Hessian of u. Classical diﬀerential inequality techniques [411] based on the Nagumo–Westphal lemma require that the underlying nonlinear evolution equation ∂t u = F (t, x, u, ∇u, Hess(u)) (1.37) satisﬁes the monotony property F (t, x, r, p, Y ) ≥ F (t, x, r, p, X) for all X, Y ∈ S(2) where Y −X is positive semideﬁnite. • The same requirement is needed for applying the theory of viscosity solutions. A detailed introduction into this framework can be found in a paper by Crandall, Ishii and Lions [103]. • Let H be a Hilbert space with scalar product (., .) and A : H → H . In order to apply the concept of maximal monotone operators [57] to the problem du + Au = 0, dt u(0) = f one has to ensure that A is monotone, i.e. (Au−Av, u−v) ≥ 0 ∀ u, v ∈ H. (1.41) (1.39) (1.40) (1.38)
Such a model is used to explain the evolution of stepwise constant temperature or salinity proﬁles in the ocean. but one should neither expect uniqueness nor stability [329].18 CHAPTER 1. [446] give evidence that the Perona–Malik process is unstable with respect to perturbations of the initial image. It has already been observed by Posmentier in 1977 [333]. the mainly observable instability is the socalled staircasing eﬀect. In the meantime several theoretical results are available which provide some insights into the actual degree of illposedness of the Perona–Malik ﬁlter. If they exist. Kawohl and Kutev showed that these solutions are unique and satisfy a maximumminimum principle. . The existence of local C 1 solutions remained unproven. Kawohl and Kutev [222] proved that the Perona–Malik process does not have global (weak) C 1 solutions for intial data that involve backward diﬀusion. Each of these minima corresponds to a piecewise constant image. Therefore. 225] proposed a notion of generalized solutions. it was possible to establish a comparison principle. Although this process was diﬀerent from the Perona–Malik process. PARTIAL DIFFERENTIAL EQUATIONS We observe that the nonmonotone ﬂux function of the Perona–Malik process implies that neither (1. one was warned what can happen. In 1994 the general conjecture was that the Perona–Malik ﬁlter might have weak solutions. They showed that the energy functional leading to the Perona–Malik process as steepest descent method has an inﬁnite number of global minima which are dense in the image space. however. under special assumptions on the initial data. see [35] and the references therein. Related equations also play a role in population dynamics and viscoelasiticity. forward–backward diﬀusion equations of Perona–Malik type are not as unnatural as they look at ﬁrst glance: besides their importance in computer vision they have been proposed as a mathematical model for heat and mass transfer in a stably stratiﬁed turbulent shear ﬂow. Numerically. Interestingly. He used an equation of Perona–Malik type for numerical simulations of the salinity proﬁles in oceans. Moreover. Starting from a smoothly increasing initial distribution he reported the creation of perturbations which led to a stepwise constant proﬁle after some time. One reason why people became pessimistic about the wellposedness of the Perona–Malik equation was a result by H¨llig [187]. where a sigmoid edge evolves into piecewise linear segments which are separated by jumps. and he showed that an analysis of their moving and merging gives similar eﬀects to those one can observe in practice. none of these frameworks is applicable to ensure wellposedness results. Results of You et al. Kichenassamy [224. and slightly diﬀerent initial images may end up in diﬀerent minima for t → ∞. which are piecewise linear and contain jumps.38) is satisﬁed nor A deﬁned by Au := −div (g(∇u2) ∇u) is monotone. He constructed a forward– o backward diﬀusion process which can have inﬁnitely many solutions.
provided one knows that there exists a suﬃciently smooth solution. ∞). Of course.1. There exists also a discrete explanation why staircasing is essentially the only observable instability: In 1D. 225. Nevertheless. and that intraregion smoothing should be preferred to interregion smoothing (piecewise smoothing). Further contributions to the explanation and avoidance of staircasing can be found in [4. Scalespace interpretation Perona and Malik renounced the assumption of Koenderink’s linear scalespace axiomatic [240] that the smoothing should treat all spatial points and scale levels equally. also other attempts to apply scalespace frameworks to the Perona– Malik process have not been more successful yet: • Salden [350]. This has been shown by Dzu Magaziewa [123] in the semidiscrete case and by Benhamouda [36. a natural way would be to prove a maximum–minimum principle. Its global solution satisﬁes a maximum–minimum principle and converges to a constant steadystate. 425] in the fully discrete case with an explicit time discretization.3 NONLINEAR DIFFUSION FILTERING 19 In image processing. Finer discretizations are less regularizing and lead to more “stairs”. since in scalespace theory one is interested in having an extremum principle for the entire time interval [0. as they guarantee that structures can be detected easily and correspondence problems can be neglected. this is only partly satisfying. Since the existence question used to be the bottleneck in the past. they required that region boundaries should be sharp and should coincide with the semantically meaningful boundaries at each resolution level (immediate localization). All results point in the same direction: the number of created plateaus depends strongly on the regularizing eﬀect of the discretization. the ﬁrst proof is due to Kawohl and Kutev who established an extremum principle for their local weak C 1 solution to the Perona–Malik ﬁlter [222]. Weickert and Benhamouda [425] showed that the regularizing eﬀect of a standard ﬁnite diﬀerence discretization is suﬃcient to turn the Perona–Malik ﬁlter into a wellposed initial value problem for a nonlinear system of ordinary diﬀerential equations. numerical studies of the staircasing eﬀect have been carried out by Nitzberg and Shiota [310]. Experiments demonstrated that the Perona–Malik ﬁlter satisﬁes these requirements fairly well [328]. standard FD discretizations are monotonicity preserving. Fr¨hlich and Weickert [148]. Instead of this. 438]. 36. and Benhamouda o [36]. In order to establish a smoothing scalespace property for this nonlinear diffusion process. 98. which guarantees that no additional oscillations occur during the evolution. Florack [143] and Eberly [124] proposed to carry over the linear scalespace theory to the nonlinear case by considering nonlinear diﬀusion . These properties are of signiﬁcant practical interest. The theoretical framework for this analysis will be presented in Chapter 3.
but not for all t > 0.43) ∂t u = div (g(∇uσ 2 ) ∇u). A mathematically sound formulation of this idea is given by Catt´. a combination of both). Lions and Morel [12] have developed a nonlinear scalespace axiomatic which comprises the linear scalespace theory as well as nonlinear morphological processes (which we will discuss in 1. because its local weak solution satisﬁes a comparison principle only for some ﬁnite time. e 2 Morel and Coll [81]. PARTIAL DIFFERENTIAL EQUATIONS processes which result from special rescalings of the linear one. the Perona–Malik model does not ﬁt into this framework. Below we shall discuss three examples which illustrate the variety of possibilities and their tailoring towards a speciﬁc task. However.6). In [81] existence.3. see [222]. one should adjust the regularization to the desired goal of the forward– backward heat equation [35]. the Perona–Malik ﬁlter turned out not to belong to this class [143]. Hence. Unfortunately.5 and 1. By replacing the diﬀusivity g(∇u ) of the Perona– Malik model by a Gaussiansmoothed version g(∇uσ 2 ) with uσ := Kσ ∗ u they end up with (1. it has been suggested to introduce the regularization directly into the continuous equation in order to become more independent of the numerical implementation [81. 261].2 Regularized nonlinear models It has already been mentioned that numerical schemes may provide implicit regularizations which stabilize the Perona–Malik process [425]. Guichard.42) This property is closely related to a maximum–minimum principle and to L∞ stability of the solution [12. This process has been analysed and modiﬁed in many ways: Whitaker and Pizer [438] have suggested that the regularization parameter σ should be . One can apply spatial or temporal regularization (and of course. (1. • Alvarez. uniqueness and regularity of a solution for σ > 0 have been established. Their smoothing axiom is a monotony assumption (comparison principle) requiring that the scalespace is orderpreserving: f ≤ g =⇒ Tt f ≤ Tt g ∀ t ≥ 0. 310]. 1. Since the dynamics of the solution may critically depend on the sort of regularization.20 CHAPTER 1. (a) The ﬁrst spatial regularization attempt is probably due to Posmentier who observed numerically the stabilizing eﬀect of averaging the gradient within the diﬀusivity [333]. Lions.
In this case. A detailed study of the inﬂuence of the parameters in a regularized Perona–Malik model has been carried out by Benhamouda [36]. spatial regularizations oﬀer the advantage that they make the ﬁlter insensitive to noise at scales smaller than σ. [35] regularized the onedimensional forward–backward heat equation by considering the thirdorder equation ∂t u = ∂x (Φ(ux )) + τ ∂xt (Ψ(ux )) (1.45) (∂x u −v) 2 leads to a wellposed ﬁlter (cf. Spatial regularizations of the Perona–Malik process leading to anisotropic diﬀusion equations have been proposed by Weickert [413. (b) P. Examples for spatially regularized nonlinear diﬀusion ﬁltering can be found in Figure 5.3. From a practical point of view. These equations arise as a special case of the spatiotemporal regularizations of Nitzberg and Shiota [310] when neglecting any spatial regularization. Therefore. Torkamani–Azar and Tait [403] suggest to replace the Gaussian convolution by the exponential ﬁlter of Shen and Castan10 [381]. Mumford conjectures that this model gives piecewise constant steadystates. it exhibits besides the contrast parameter λ an additional noise scale σ. Lions proved in a private communication to Mumford that the onedimensional process ∂t u = ∂x (g(v) ∂x u).3 NONLINEAR DIFFUSION FILTERING 21 a decreasing function in t. when regarding (1.(b).4 (a). In Chapter 2 we shall see that spatial regularizations lead to wellposed scalespaces with a large class of Lyapunov functionals which guarantee that the solution converges to a constant steadystate. ∂t v = 1 τ (1.46) 10 This renounces invariance under rotation.3. the steadystate solution would solve a segmentation problem. This avoids a shortcoming of the genuine Perona–Malik process which misinterprets strong oscillations due to noise as edges which should be preserved or even enhanced.2 (c) and 5. .1.44) (1. Barenblatt et al. Kaˇur and Mikula [217] have investigated a modiﬁcac tion which allows to diﬀuse diﬀerently in diﬀerent grey value ranges. 415] and will be described in 1. (c) In the context of shear ﬂows.43) as an image restoration equation. [329]). and Li and Chen [252] have proposed to subsequently decrease the contrast parameter λ. We observe that v is intended as a timedelay regularization of ∂x u2 where the parameter τ > 0 determines the delay.L.
Below we study two representatives of anisotropic diﬀusion processes.2. [443]. We observe that spatial regularizations are closer to scalespace ideas while temporal regularization are more related to image restoration and segmentation. 1. Nevertheless. 11 An exception is the timedelay regularization of Cottet and ElAyyadi [100. These requirements cannot be satisﬁed by a scalar diﬀusivity anymore.3 Anisotropic nonlinear models All nonlinear diﬀusion ﬁlters that we have investigated so far utilize a scalarvalued diﬀusivity g which is adapted to the underlying image structure. Lev. whereas the second one is welladapted to the processing of onedimensional features. A general theoretical framework for spatially regularized anisotropic diﬀusion ﬁlters will be presented in the remaining chapters of this book. They also showed that smooth solutions may become discontinuous within ﬁnite time. since they may lead to nontrivial steadystates. before they ﬁnally converge to a piecewise constant steadystate. These ideas have been further developed by Nitzberg and Shiota [310]. and Yang et al. and Nagao and Matsuyama [297]. First anisotropic ideas in image processing date back to Graham [167] in 1962. a diﬀusion tensor leading to anisotropic diﬀusion ﬁlters has to be introduced. in certain applications it would be desirable to bias the ﬂux towards the orientation of interesting features. . For the corresponding initial boundary value problem with homogeneous Neumann boundary conditions they proved the existence of a unique generalized solution. Zucker and Rosenfeld [250]. Appropriately chosen regularizations create the desired ﬁlter features. PARTIAL DIFFERENTIAL EQUATIONS where Ψ is strictly increasing and uniformly bounded in IR. Therefore. Anisotropic diﬀusion ﬁlters usually apply spatial regularization strategies11 . 101]. Their suggestion to use shapeadapted Gaussian masks has been discussed in Section 1. they are isotropic and the ﬂux j = −g∇u is always parallel to ∇u. They used convolution masks that depended on the underlying image structure. These examples demonstrate that regularization is much more than stabilizing an illposed process: Regularization is modeling. Related statistical approaches were proposed by Knutsson. The ﬁrst one oﬀers advantages at noisy edges. and Φ′ (s) = O(Ψ′(s)) as s → ±∞. respectively. This regularization was physically motivated by introducing a relaxation time τ into the diﬀusivity.6. followed by Newman and Dirilten [300]. Wilson and Granlund [237]. Lindeberg and G˚ arding [259]. They are called edgeenhancing anisotropic diﬀusion and coherenceenhancing anisotropic diﬀusion.22 CHAPTER 1.3.
43) behaves almost like the linear diﬀusion ﬁlter (1. a desirable method should prefer diﬀusion along edges to diﬀusion perpendicular to them.3 NONLINEAR DIFFUSION FILTERING 23 (a) Anisotropic regularization of the Perona–Malik process In the interior of a segment the nonlinear isotropic diﬀusion equation (1. noise at edges cannot be eliminated successfully by this process. we observe that ∇u becomes an eigenvector of D with corresponding eigenvalue 0.51) This is a process diﬀusing solely perpendicular to ∇uσ . λ2 (∇uσ ) := 1. Weickert [415] proposed to choose the corresponding eigenvalues λ1 and λ2 as λ1 (∇uσ ) := g(∇uσ 2 ). we end up with the isotropic Perona–Malik process. (1.9). ∇u does not coincide with one of the eigenvectors of D as long as σ > 0. the process stops completely. we construct the orthonormal system of eigenvectors v1 . Therefore. Another anisotropic model which can be regarded as a regularization of an isotropic nonlinear diﬀusion ﬁlter has been described in [413]. In general. but at edges diﬀusion is inhibited. (1. (b) Anisotropic models for smoothing onedimensional objects A second motivation for introducing anisotropy into diﬀusion processes arises from the wish to process onedimensional features such as linelike structures.48) (1. To this end. Anisotropic models do not only take into account the modulus of the edge detector ∇uσ . Cottet and Germain [102] constructed a diﬀusion tensor with eigenvectors as in (1.1 presents several examples where this process is applied to test images. If we let the regularization parameter σ tend to 0. λ2 (∇uσ ) := η∇uσ  1 + (∇uσ /σ)2 2 (1. To overcome this problem. it is not intended as an anisotropic regularization of the Perona–Malik equation. v2 of the diﬀusion tensor D such that they reﬂect the estimated edge structure: v1 ∇uσ . To this end. Hence. v2 ⊥ ∇uσ . In this sense.47) In order to prefer smoothing along the edge to smoothing across it.1. this model behaves really anisotropic. but also its direction. Therefore. (1. For σ → 0.50) (η > 0).49) Section 5.47) and corresponding eigenvalues λ1 (∇uσ ) := 0. Wellposedness .
Vectorvalued images can arise either from devices measuring multiple physical properties or from a feature analysis of one single image. 418]. 1.g. This leads to coherenceenhancing anisotropic diﬀusion [418]. whereas representatives of the second class are given by statistical moments or the jet space induced by the image itself and its partial derivatives up to a given order. As a remedy. . The theoretical analysis in the present work shall comprise the second possibility.3.2). More sophisticated structure descriptors. For enhancing parallel linelike structures. the structure tensor (cf.2. This may be useful when considering e. Since the Cottet–Germain model diﬀuses only in one direction. which shall be discussed in Section 5. It has also been proposed to replace ∇uσ by a Bayesian classiﬁcation result in feature space [26]. It is easily seen that many of the previous results can be generalized to higher dimensions. one can steer the smoothing process by more advanced structure descriptors such as higherorder derivatives [127] or tensorvalued expressions of ﬁrstorder derivatives [414. Section 2. multispectral Landsat exposures and multispin echo MR images. it is clear that its result depends very much on the smoothing direction. [355] describe nonlinear diﬀusion ﬁltering of 3D a image sequences by treating them as 4D data sets. The edge detector ∇uσ enables us to adapt the diﬀusion to magnitude and direction of edges.24 CHAPTER 1. where also many examples can be found. one can improve this model when replacing ∇u⊥ σ by a more robust descriptor of local orientation. PARTIAL DIFFERENTIAL EQUATIONS results for the Cottet–Germain ﬁlter comprise an existence proof for weak solutions.4 Generalizations Higher dimensions. Vectorvalued models. A generalization c of coherenceenhancing anisotropic diﬀusion to higher dimensions is proposed in [428]. Feature vectors play an important role for tasks like texture segmentation. or when applying diﬀusion ﬁlters to the postprocessing of ﬂuctuating higherdimensional numerical data. but it can neither distinguish between edges and corners nor does it give a reliable measure of local orientation. medical image sequences from computerized tomography (CT) or magnetic resonance imaging (MRI). The ﬁrst threedimensional nonlinear diﬀusion ﬁlters have been investigated by Gerig et al. Examples for the ﬁrst category are colour images. [155] in the isotropic case and by Rambaux and Gar¸on [339] in the anisotropic case. and S´nchez–Ortiz et al.
Extensions to anisotropic vectorvalued models with a common tensorvalued structure descriptor for all channels have been investigated by Weickert [422]. In order to avoid this. see also [349] for related ideas. It turned out that all results became fairly similar.3. Semiimplicit schemes – which approximate the diﬀusivity or the diﬀusion tensor in an explicit way and the rest implicitly – are considered in [81].3 NONLINEAR DIFFUSION FILTERING 25 The simplest idea how to apply diﬀusion ﬁltering to multichannel images would be to diﬀuse all channels separately and independently from each other.5 Numerical aspects For nonlinear diﬀusion ﬁltering numerous numerical methods have been applied: Finite element techniques are described in [367. Due to their local behaviour. Nevertheless. 4]. A fast multigrid technique using a pyramid algorithm for the Perona–Malik ﬁlter has been studied by Acton et al. [158]. [155. 1. when the regularization parameter σ was suﬃciently large.1. 421]. Neural network approximations to nonlinear diﬀusion ﬁlters are investigated by Cottet [100. 216]. Explicit schemes are the most simple to code and. Most implementations of nonlinear diﬀusion ﬁlters are based on ﬁnite diﬀerence methods. one should use a common diﬀusivity which combines information from all channels. They possess much better stability properties. they are wellsuited for parallel architectures. This leads to the undesirable eﬀect that edges may be formed at diﬀerent locations for each channel. B¨nsch and a Mikula reported a signiﬁcant speedup by supplementing them with an adaptive mesh coarsening [34]. therefore. both for the semidiscrete . 156] and Whitaker [433. 391. While the preceding techniques are focusing on approximating a continuous equation. Such isotropic vectorvalued diﬀusion models were studied by Gerig et al. Since the computational eﬀort is of a comparable order of magnitude. 434] in the context of medical imagery. a pseudospectral method and a ﬁnitediﬀerence scheme. they suﬀer from the fact that fairly small time step sizes are needed in order to ensure stability. since they are easy to handle and the pixel structure of digital images already provides a natural discretization on a ﬁxed rectangular grid. [5. Perona and Malik [327] propose hardware realizations by means of analogue VLSI networks with nonlinear resistors. it seems to be a matter of taste which scheme is preferred. 99] and Fischl and Schwartz [137]. it is often desirable to have a genuinely discrete theory which guarantees that an algorithm exactly reveals the same qualitative properties as its continuous counterpart. Such a framework is presented in [420. they are used almost exclusively. In [148] three schemes for a spatially regularized 1D Perona–Malik ﬁlter are compared: a wavelet method of Petrov–Galerkin type. 34. A very detailed VLSI proposal has been developed by Gijbels et al.
c. it turns out that suitable explicit and semiimplicit ﬁnite diﬀerence discretizations of many discussed models create discrete scalespaces. semidiscrete and fully discrete nonlinear diﬀusion scalespace. and a speedup by another order of magnitude can be achieved by a parallel implementation [431]. A detailed treatment of this theory can be found in Chapter 3 and 4. 418]. We observe that the requirements belong to ﬁve categories: smoothness. These criteria are easy to check for many discretizations. and for enhancing textures such as ﬁngerprints [418]. requirement continuous semidiscrete ut = div (D∇u) du = A(u)u dt u(t = 0) = f u(0) = f D∇u. they are about 10 times more eﬃcient than the widely used explicit schemes. are 0 positive nonnegative semideﬁnite oﬀdiagonals uniformly pos. They have proved to be useful for improving subsampling [144] and line detection [156. symmetry. nonnegativity and connectivity requirements. average grey value invariance. which are based on an additive operator splitting (AOS) [424. 415]. and convergence to a constant steadystate [423].26 CHAPTER 1. n = 0 D ∈ C∞ A Lipschitzcontinuous D symmetric A symmetric div form. pos. column sums reﬂective b. diagonal smoothness symmetry conservation nonnegativity connectivity and for the fully discrete case. PARTIAL DIFFERENTIAL EQUATIONS Table 1. Table 1. 418]. conservation.2: Requirements for continuous. for blind image restoration [445]. 430]. irreducible deﬁnite discrete u0 = f uk+1 = Q(uk )uk Q continuous Q symmetric column sums are 1 nonnegative elements irreducible. The discrete nonlinear scalespace concept has also led to the development of fast novel schemes. respectively. for visualizing qualityrelevant features in computer aided quality control [299.3. for scalespace based segmentation algorithms .2 gives an overview of the requirements which are needed in order to prove wellposedness. 1. 413. A general framework for AOS schemes will be presented in Section 4.4. causality in terms of an extremum principle and Lyapunov functionals.2.6 Applications Nonlinear diﬀusion ﬁlters have been applied for postprocessing ﬂuctuating data [269. In particular. Under typical accuracy requirements.
1.4 METHODS OF DIFFUSION–REACTION TYPE
27
[307, 308], for segmentation of textures [433, 437] and remotely sensed data [6, 5], and for target tracking in infrared images [65]. Most applications, however, are concerned with the ﬁltering of medical images [26, 28, 29, 155, 244, 248, 264, 270, 308, 321, 355, 386, 393, 431, 434, 437, 444]. Some of these applications will be investigated in more detail in Chapter 5. Besides such speciﬁc problem solutions, nonlinear diﬀusion ﬁlters can be found in commercial software packages such as the medical visualization tool Analyze.12
1.4
Methods of diﬀusion–reaction type
This section investigates variational frameworks, in which diﬀusion–reaction equations or coupled systems of them are interpreted as steepest descent minimizers of suitable energy functionals. This idea connects diﬀusion methods to edge detection and segmentation ideas. Besides the variational interpretation there exist other interesting theoretical frameworks for diﬀusion ﬁlters such as the Markov random ﬁeld and mean ﬁeld annealing context [152, 153, 247, 251, 328, 387], robust statistics [41], and deterministic interactive particle models [279]. Their discussion, however, would lead us beyond the scope of this book.
1.4.1
Single diﬀusion–reaction equations
Nordstr¨m [311] has suggested to obtain a reconstruction u of a degraded image o f by minimizing the energy functional Ef (u, w) :=
Ω
β ·(u−f )2 + w·∇u2 + λ2 ·(w−ln w) dx.
(1.52)
The parameters β and λ are positive weights and w : Ω → [0, 1] gives a fuzzy edge representation: in the interior of a region, w approaches 1 while at edges, w is close to 0 (as we shall see below). The ﬁrst summand of E punishes deviations of u from f (deviation cost), the second term detects unsmoothness of u within each region (stabilizing cost), and the last one measures the extend of edges (edge cost). Cost terms of these three types are typical for variational image restoration methods. The corresponding Euler equations to this energy functional are given by 0 = β ·(u−f ) − div (w∇u), 0 = λ
2 1 ·(1− w )
(1.53) (1.54)
+ ∇u ,
2
Analyze is a registered trademark of Mayo Medical Ventures, 200 First Street SW, Rochester, MN 55905, U.S.A.
12
28
CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS
equipped with a homogeneous Neumann boundary condition for u. Solving (1.54) for w gives w= 1 . 1 + ∇u2/λ2 (1.55)
We recognize that w is identical with the Perona–Malik diﬀusivity g(∇u2) introduced in (1.32). Therefore, (1.53) can be regarded as the steadystate equation of ∂t u = div (g(∇u2) ∇u) + β(f −u). (1.56) This equation can also be obtained directly as the descent method of the functional Ff (u) :=
Ω
β ·(u−f )2 + λ2 ·ln 1+ ∇u λ2
2
dx.
(1.57)
The diﬀusion–reaction equation (1.56) consists of the Perona–Malik process with an additional bias term β ·(f −u). One of Nordstr¨m’s motivations for introo ducing this term was to free the user from the diﬃculty of specifying an appropriate stopping time for the Perona–Malik process. However, it is evident that the Nordstr¨m model just shifts the problem of o specifying a stopping time T to the problem of determining β. So it seems to be a matter of taste which formulation is preferred. People interested in image restoration usually prefer the reaction term, while for scalespace researchers it is more natural to have a constant steadystate as the simplest image representation. Nordstr¨m’s method may suﬀer from the same illposedness problems as the o underlying Perona–Malik equation, and it is not hard to verify that the energy functional (1.57) is nonconvex. Therefore, it can possess numerous local minima, and the process (1.56) with f as initial condition does not necessarily converge to a global minimum. Similar diﬃculties may also arise in other diﬀusion–reaction models, where convergence results have not yet been established [152, 186]. A popular possibility to avoid these illposedness and convergence problems is to renounce edgeenhancing diﬀusivities in order to end up with (nonquadratic) convex functionals [43, 88, 110, 367, 391]. In this case the frameworks of convex optimization and monotone operators are applicable, ensuring wellposedness and stability of a standard ﬁniteelement approximation [367]. Diﬀusion–reaction approaches have been applied to edge detection [367, 391], to the restoration of inverse scattering images [263], to SPECT [88] and vascular reconstruction in medical imaging [102, 325], and to optic ﬂow [368, 111] and stereo problems [343]. They can be extended to vectorvalued images [369] and to cornerpreserving smoothing of curves [136, 323]. Diﬀusion–reaction methods with constant diﬀusivities have also been used for local contrast normalization in images [330].
1.4 METHODS OF DIFFUSION–REACTION TYPE
29
1.4.2
Coupled systems of diﬀusion–reaction equations
Mumford and Shah [295, 296] have proposed to obtain a segmented image u from f by minimizing the functional Ef (u, K) = β
Ω
(u−f )2 dx +
Ω\K
∇u2 dx + αK
(1.58)
with nonnegative parameters α and β. The discontinuity set K consists of the edges, and its onedimensional Hausdorﬀ measure K gives the total edge length. Like the Nordstr¨m functional (1.52), this expression consists of three cost terms: o the ﬁrst one is the deviation cost, the second one gives the stabilizing cost, and the third one represents the edge cost. The Mumford–Shah functional can be regarded as a continuous version of the Markov random ﬁeld method of Geman and Geman [154] and the weak membrane model of Blake and Zisserman [42]. Related approaches are also used to model materials with two phases and a free interface. The fact that (1.58) leads to a free discontinuity problem causes many challenging theoretical questions [249]. The book of Morel and Solimini [292] covers a very detailed analysis of this functional. Although the existence of a global minimizer with a closed edge set K has been established [108, 17], uniqueness is in general not true [292, pp. 197–198]. Regularity results for K in terms of (at least) C 1 arcs have recently been obtained [18, 19, 20, 48, 49, 107]. The concept of energy functionals for segmenting images oﬀers the practical advantage that it provides a framework for comparing the quality of two segmentations. On the other hand, (1.58) exhibits also some shortcomings, e.g. the problem that sigmoidlike edges produce multiple segmentation boundaries (oversegmentation, staircasing eﬀect) [377]. Another drawback results from the fact that the Mumford–Shah functional allows only singularities which are typical for minimal surfaces: Corners or Tjunctions are not possible and segments meet at triple points with 120o angle [296]. In order to avoid such problems, modiﬁcations of the Mumford–Shah functional have been proposed by Shah [379]. An aﬃne invariant generalization of (1.58) is investigated in [32, 31] and applied to aﬃne invariant texture segmentation [31, 33], and a Mumford–Shah functional for curves can be found in [323]. Since many algorithms in image processing can be restated as versions of the Mumford–Shah functional [292] and since it is a prototype of a free discontinuity problem it is instructive to study this variational problem in more detail. Numerical complications arise from the fact that the Mumford–Shah functional has numerous local minima. Global minimizers such as the simulated annealing method used by Geman and Geman [154] are extremely slow. Hence, one searches for fast (suboptimal) deterministic strategies, e.g. pyramidal regiongrowing algorithms [3, 239].
Related equations are also studied in [398]. but a maximum–minimum principle and a local stability proof have been established. Shah investigates diﬀusion–reaction systems for matching stereo images [378].61) ∂t w = c∆w − + (1−w) 4c with homogeneous Neumann boundary conditions. texture generation [406. They are based on Turing’s pattern formation model [405]. The system of Richardson and Mitter is used for edge detection [341]. He replaces the functional (1. . PARTIAL DIFFERENTIAL EQUATIONS Another important class of numerical methods is based on the idea to approximate the discontinuity set K by a smooth function w. Experiments indicate that it converges to a stable solution. Since (1. 440] and halftoning [382]. this approach gives rise to the same theoretical questions as (1. We may for instance study the functional Ff (u. Their ﬁnite diﬀerence algorithms run on a parallel transputer network. Wellposedness results for this system have not been obtained up to now. apply coupled diﬀusionreaction equations to image sequence analysis.60) resembles the Nordstr¨m o process (1.56). They obtained pronounced edges by replacing such a term by its Perona–Malik counterpart div (g(∇u2) ∇u).61). w) := Ω β ·(u−f )2 + w 2 ·∇u2 + α· c∇w2 + (1−w) 4c 2 dx (1. (1. 336] observed that this solution looks fairly blurred since the equations contain diﬀusion terms such as ∆u. see [22] for more details). Proesmans et al.59) with a positive parameter c specifying the “edge width”. 382]. Minimizing Ff corresponds to the gradient descent equations ∂t u = div (w 2 ∇u) + β ·(f −u).60) (1. and which are not connected to Perona–Malik or Mumford– Shah ideas. which is close to 0 near edges of u and which approximates 1 elsewhere. similar problems can arise: The functional Ff is not jointly convex in u and v.30 CHAPTER 1. It should also be mentioned that there exist reaction–diﬀusion systems which have been applied to image restoration [334. Another diﬀusion–reaction system is studied by Shah [375. Equations of this type are investigated by Richardson and Mitter [341]. This results in ﬁnding an equilibrium state between two competing processes. so it may have many local minima and a gradient descent algorithm may get trapped in a poor local minimum. [337. while Proesmans et al. 338]. 376].60). 335.58) by two coupled convex energy functionals and applies gradient descent. Of course. Ambrosio and Tortorelli proved that this functional converges to the Mumford–Shah functional for c → 0 (in the sense of Γconvergence. vectorvalued images and stereo vision [336. w ∇u2 α (1.
except for 1. 184] for an overview. they are invariant under monotone greylevel rescalings. quality control. mineralogy.5. and many others. Its applications cover biology. (1. see e. Morphology is usually described in terms of algebraic set theory.5 Classic morphological processes Morphology is an approach to image analysis based on shapes.g.e.64) We observe that for this type of morphological operations only greylevel sets matter. closed sets X ⊂ IR2 whose boundaries are Jordan curves [16]. they can be viewed as curve or shape deformations. Greyscale morphology generalizes these ideas [274] by decomposing an image f into its level sets {Xa f. (1. Binary morphological operations aﬀect only the boundary curve of the shape and. This section surveys the basic ideas and elementary operations of binary and greyscale morphology. f (x) ≥ a}. [280. afterwards it was extended to greyscale images by regarding level sets as shapes. 1. where Xa f := {x ∈ IR2 . [25] and Alvarez et al.62) χ(x) := 0 else. presents its PDE representations for images and curves. therefore.6. medical diagnostics. Its mathematical formalization goes back to the group around Matheron and Serra. radar and remote sensing. science of material. Nevertheless.6 and some modiﬁcations . (1.1. PDE formulations for classic morphological processes have been discovered recently by Brockett and Maragos [60]. 371. The theory had ﬁrst been developed for binary images. i.5. 181. histology. van den Boomgaard [50]. Arehart et al. we identify a shape X with its characteristic function 1 if x ∈ X. both working at ENS des Mines de Paris in Fontainebleau. Henceforth. a ∈ IR}.63) A binary morphological operation A can be extended to some greyscale image f by deﬁning Xa (Af ) := A(Xa f ) ∀ a ∈ IR. and generalizations are sketched which comprise the morphological equivalent of Gaussian scalespace. and summarizes the results concerning wellposedness and scalespace properties. As a consequence.1 Binary and greyscale morphology Binary morphology considers shapes (silhouettes). This morphological invariance (greyscale invariance) is characteristic for all methods we shall study in Section 1. [12]. Afterwards numerical aspects of the PDE formulation of these processes are discussed.5 and 1.5 CLASSIC MORPHOLOGICAL PROCESSES 31 1.
while closing smoothes by eliminating small holes. y ∈ B}. (1. t) . PARTIAL DIFFERENTIAL EQUATIONS in 1. (1. t) .70) ∂t u(x. or ellipses. t) = inf y.3 Continuousscale morphology Let us consider a convex structuring element tB with a scaling parameter t > 0..66) These names can be easily motivated when considering a shape in a binary image and a discshaped structuring element.32 CHAPTER 1. for applications like edge detection or image restoration. while erosion shrinks them. fusing narrow breaks and ﬁlling gaps at the contours [181]. y∈B (1. Typical shapes for B are discs. On the other hand. ∇u(x. Moreover. 13 Henceforth. dilation and erosion with a structuring element B. y ∈ B}. we frequently use the simpliﬁed notation u(t) instead of u(.65) (1. In the preceding shape interpretation opening smoothes the shape by breaking narrow isthmuses and eliminating small islands.67) (1. Dilation and erosion form the basis for constructing other morphological processes.6.69) (1. y∈B with f as initial condition [12. for instance opening and closing: opening: closing: (f ◦ B) (x) := ((f ⊖ B) ⊕ B) (x).5. It is a very desirable property in all cases where brightness changes of the illumination occur or where one wants to be independent of the speciﬁc contrast range of the camera. ∇u(x. (f ⊕ B) (x) := sup {f (x−y). t) = sup y. can be shown to be equivalent to solving ∂t u(x.5. 360]. respectively13 . calculating u(t) = f ⊕ tB and u(t) = f ⊖ tB. The two basic morphological operations. in some cases isolines may give inadequate information about the depicted physical object boundaries.2 Basic operations Classic morphology analyses a shape by matching it with a socalled structuring element. Then. t) . squares. 1. In this case dilation blows up its boundaries. are deﬁned for a greyscale image f ∈ L∞ (IR2 ) by [60] dilation: erosion: (f ⊖ B) (x) := inf {f (x+y). contrast provides important information which should be taken into account. 1.68) (f • B) (x) := ((f ⊕ B) ⊖ B) (x). a bounded set B ⊂ IR2 .5.
which may be a function of its curvature κ := det(Cp . This can be illustrated by considering a smooth Jordan curve C : [0. Such a process is also named grassﬁre ﬂow or prairie ﬂow. Connection to curve evolution Morphological PDEs such as (1.77) (1.76) is called the Eulerian formulation of the curve evolution (1. He simulated grassﬁre ﬂow by a selfconstructed opticomechanical device. t) (1.5 (a) presents the temporal evolution of a test image under such a continuousscale dilation. because it is written in terms of a ﬁxed coordinate system. they describe the curve evolutions ∂t C = ± n. t) = x1 (p.74) (1.78) This equation moves level sets in normal direction with constant speed. .72) correspond to the simple cases β = ±1. y ≤ 1} one obtains ∂t u = ∇u. erosion) of f with a disc of radius t and centre 0 as structuring element. 2π] × [0. 16] ∂t u = β(curv(u)) · ∇u. 33 (1.71) (1. ∂t C = β(κ) · n.71) or (1. Hence. where the curvature of u is curv(u) := div ∇u . (1.73) where p is the parametrization and t is the evolution parameter.72) are closely related to shape and curve evolutions. Its importance for shape analysis in biological vision has already been pointed out in the sixties by Blum [47]. (1. The corresponding evolution for u is given by [319.Cpp ) : Cp 3 C(p. It is closely related to the Huygens principle for wave propagation [25].g. We observe that (1. C(p. ∞) → IR2 . t) in an image u(x.1. 0) = C0 (p). ∇u (1.76) Sometimes the image evolution (1. 362.75) One can embed the curve C(p. Figure 5. The solution u(t) is the dilation (resp.71) and (1. t) in such a way that C is just a level curve of u.5 CLASSIC MORPHOLOGICAL PROCESSES By choosing e. B := {y ∈ IR2 . t) x2 (p.72) ∂t u = −∇u. We assume that C evolves in outer normal direction n with speed β.74).
t) which fulﬁls the maximum–minimum principle inf f ≤ u(x.70) possess a unique global viscosity solution u(x. which is also based on local extrema instead of zerocrossings. ∞).5.76) is provided by the theory of viscosity solutions [103]. that for an initial value f ∈ BUC(IR2 ) := {ϕ ∈ L∞ (IR2 )  ϕ is uniformly continuous on IR2 } (1. t) ≤ sup f 2 IR IR 2 on IR2 × [0. Guichard. Similar results have been found by van den Boomgaard and Smeulders [53]. The advantage of this analysis is that it allows us to treat shapes with singularities such as corners. one smoothing axiom (comparison principle) and additional invariance requirements (greylevel shift invariance.79) the equations (1. (1. entropy solutions. Jackway and Deriche [206. where the classical solution concept does not apply.(1. but a unique weak solution in the viscosity sense still exists. (1. they conjecture a causality property where singularities play a role similar to zerocrossings in Gaussian scalespace. morphological invariance). a twodimensional scalespace . PARTIAL DIFFERENTIAL EQUATIONS 1. Moreover. A complete scalespace interpretation is due to Alvarez. It can be shown [90. 207] prove a causality theorem for the dilation– erosion scalespace.69). Hence.78) may develop singularities and intersections even for smooth initial data. They establish that under erosion the number of local minima is decreasing. while dilation reduces the number of local maxima. concepts of jump conditions.81) 1.34 CHAPTER 1. g the corresponding solutions u(t). locality and regularity). The location of these extrema is preserved during their whole lifetime. it is L∞ stable: for two diﬀerent initial images f . 103]. A suitable framework for the image evolution equation (1.5. 129.4 Theoretical results Equations such as (1.5 Scalespace properties Brockett and Maragos [60] pointed out that the convexity of B is suﬃcient to ensure the semigroup property of the corresponding dilations and erosions. v(t) satisfy u(t)−v(t) L∞ (IR2 ) ≤ f −g L∞ (IR2 ) .80) Moreover. and shocks have to be applied to this shape evolution [228]. Lions and Morel [12]: They prove that under three architectural assumptions (semigroup property. This establishes an important architectural scalespace property. invariance under rotations and translations.
5. dilation and erosion belong to the class (1.87) u(x.82) are fulﬁlled. t) = − x2 4t (t > 0).82). thus being good candidates for morphological scalespaces.14 1.86) (1. t) is a weak solution of ∂t u = ∇u2 . A dilation of an image f with a structuring function b : IR2 → IR is deﬁned as (f ⊕ b) (x) := sup {f (x−y) + b(y)}.5 (b). Indeed.6 Generalizations It is possible to extend morphology with a structuring element to morphology with nonﬂat structuring functions. since one can recover dilation with a structuring element B by considering the ﬂat structuring function b(x) := 0 −∞ (x ∈ B). 14 Invariance under rotations is only satisﬁed for a disc centered in 0 as structuring element. (1. (1.85) It can be shown [50. The temporal evolution of a test image under this process is illustrated in Figure 5. in [12] it is shown that the converse is true as well: all axioms that lead to (1. curv(u)) 35 (1.82) Clearly. 51] and Jackway [206] proposed to dilate an image f (x) with quadratic structuring functions of type b(x. . 53] that the result u(x. but we gain an interesting insight into a process which has very much in common with Gaussian scalespace. 51]. 0) = f (x). t) = a exp( x ) are the 4t only rotationally symmetric kernels which are separable with respect to convolution.65). (1.83) This is a generalization of deﬁnition (1.5 CLASSIC MORPHOLOGICAL PROCESSES equation has the following form: ∂t u = ∇u F (t.84) Van den Boomgaard [50. 2 In analogy to the fact that Gaussiantype functions k(x. y∈IR2 (1. In this case we have to renounce invariance under monotone grey level transformations. (x ∈ B). van den Boomgaard proves that the quadratic structuring functions b(x.1. t) are the only rotationally invariant structuring functions which are separable with respect to dilation [50.
88) where h is the pixel size.36 CHAPTER 1. 360. To illustrate the basic idea with a simple example. (b) The time t plays the role of a continuous scale parameter. Overviews of these level set approaches and their various applications can be found in [372. and it is possible to get results with subpixel accuracy. The widelyused Osher–Sethian schemes [319] are based on the idea to derive numerical methods for such equations from techniques for hyperbolic conservation laws. For morphological operations with ﬂat structuring elements. 218. Level set methods possess two advantages over classical settheoretic schemes for dilation/erosion [25.86) as limiting case. 25. 66]: (a) They give excellent results for nondigitally scalable structuring elements whose shapes cannot be represented correctly on a discrete grid. they can be implemented very eﬃciently by applying onedimensional operations. PARTIAL DIFFERENTIAL EQUATIONS A useful tool for understanding this similarity and many other analogies between morphology and linear systems theory is the slope transform. A fast algorithm is described by van den Boomgaard [51]. the size of a structuring element need not be multiples of the pixel size.0 h 2 + max un − un i i+1 . It has been discovered simultaneously by Dorst and van den Boomgaard [119] and by Maragos [275] in slightly diﬀering formulations. and un denotes a discrete i approximation of u(ih. 374]. τ is the time step size.5.0 h 2 .7 Numerical aspects Dilations or erosions with quadratic structuring functions are separable and. for instance discs or ellipses. and suﬀer from the problem of coping with singularities and topological changes [319.86) and Gaussian scalespace has also triggered Florack and Maas [142] to study a oneparameter family of isomorphisms of linear diﬀusion which reveals (1. . 360]. (1. nτ ). This generalization of the Legendre transform is the morphological equivalent of the Fourier transform. The close relation between (1. Therefore. let us restrict ourselves to the onedimensional dilation equation ∂t u = ∂x u. For this reason it is useful to discretize the corresponding image evolution equations. Schemes for dilation or erosion which are based on curve evolution turn out be be diﬃcult to handle: they require prohibitive small time steps. 1. the situation is more complicated. A ﬁrstorder upwind Osher–Sethian scheme for this process is given by n un+1 − ui i = τ min un − un i−1 i . thus.
and – although they cannot be written in conservation form – they reveal interesting relations to diﬀusion processes.8 Applications Continuousscale morphology has been applied to shapefromshading problems. In this subsection we shall discuss these PDEs. 1. 385].6 Curvaturebased morphological processes Besides providing a useful reinterpretation of classic continuousscale morphology. 1. (b) Dissipative eﬀects such as blurring of discontinuities occur. they reveal also two disadvantages: (a) They are slower than settheoretic morphological schemes. 1.9) can be rewritten as ∂t u = ∂ηη u + ∂ξξ u. gridless image halftoning. Applications outside the ﬁeld of image processing and computer vision include for instance shape oﬀsets in CAD and path planning in robotics. The ﬁrst term on the righthand side of (1. ENO schemes obtain both highorder accuracy in smooth regions and sharp shock transitions [183]. 37 To address the ﬁrst problem.1 Meancurvature ﬁltering In order to motivate our ﬁrst curvaturebased morphological PDE. distance transformations. 276].6 CURVATUREBASED MORPHOLOGICAL PROCESSES However.89) where the unit vectors η and ξ are parallel and perpendicular to ∇u. Blurring of discontinuities can be minimized by applying shockcapturing techniques such as highorder ENO15 schemes [395. 435. 15 . possible generalizations. 373]. (1.89) describes smoothing along the ﬂowlines. speedup techniques for shape evolution have been proposed which use only points close to the curve at every time step [8. Overviews and suitable references can be found in [233. respectively.5. These processes are curvaturebased. let us recall that the linear diﬀusion equation (1. numerical aspects. By adapting the stencil for derivative approximations to the local smoothness of the solution. Two important representatives of this class are mean curvature motion and its aﬃne invariant counterpart.6. while the second one smoothes along isophotes. the PDE approach has led to the discovery of new morphological operators. When we want to smooth ENO means essentially nonoscillatory. and applications.1. and skeletonization.
p v(p. grid generation. The corresponding curve evolution ∂t C(p. Cp 3 Processes of this type have ﬁrst been studied by Brakke in 1978 [54]. t) denote the Euclidean arclength of C(p. t) · n(p. (1. i.94) is named (mean) curvature motion (MCM). PARTIAL DIFFERENTIAL EQUATIONS the image anisotropically along its isophotes.90) can also be written as ∂t u = u2 2 ux1 x1 − 2ux1 ux2 ux1x2 + u2 1 ux2 x2 x x u2 1 + u2 2 x x 1 = ∆u − ∇u. Intrinsic heat ﬂow Interestingly. mean curvature motion can be regarded as the limit process when classic morphological operators such as median ﬁltering are iteratively applied. mappings x → Rx + b (1.95) is also called geometric heat equation or Euclidean shortening ﬂow.38 CHAPTER 1. the derivation of minimal surfaces. 0) = f (x).97) .Cpp ) . The subsequent discussions shall clarify these names. t) (1.96) where Cρ := ∂ρ C. (1. t) = κ(p. t).e. (1. and many other applications. Hess(u)∇u ∇u2 = ∇u curv(u). we can neglect the ﬁrst term and end up with the problem ∂t u = ∂ξξ u. It is invariant under Euclidean transformations.92) (1. t) := 0 Cρ (ρ. crystal growth. 374] and the references therein for an overview. Let v(p. Figure 5.93) (1.95) shows that (1. The Euclidean arclength is characterized by Cv  = 1. t) dρ. i.91) By straightforward calculations one veriﬁes that (1. u(x. see [372.5 (c) and 5. Equation (1. They arise in ﬂame propagation.e.11 (a) present examples for mean curvature ﬁltering. there exists a further connection between linear diﬀusion and motion by curvature. The importance of MCM in image processing became only recently clear: As nicely explained in a paper by Guichard and Morel [173]. equation (1.90) (1.94) ∇u Since curv(u) = div ∇u is the curvature of u (mean curvature for dimensions ≥ 3).90) propagates isophotes in inner normal direction with a velocity that is given by their curvature κ = det(Cp .
98) we recognize that curvature motion can be regarded as Euclidean invariant diﬀusion of isophotes: ∂t C(p. equation (1.99) is called Euclidean shortening ﬂow [150].6 CURVATUREBASED MORPHOLOGICAL PROCESSES 39 where R ∈ IR2×2 denotes a rotation matrix and b ∈ IR2 is a translation vector.1. Moreover. since under all ﬂows Ct = Cqq the Euclidean arclength parametrization q(p) := v(p) is the fastest way to shrink the Euclidean perimeter Cp  dp. Such a shape inclusion principle implies in connection with (1. it can be shown that. if one isophote is enclosed by another. p.5.90) has a unique viscosity solution which is L∞ stable and satisﬁes a maximum–minimum principle [12]. 227]. that have been mentioned in 1. This shows that the 2 . since it is independent of the curve parametrization. When studying the evolution of isophotes under MCM. it has been shown in [188.100) that it takes the time T = 1 σ 2 to remove all isophotes within a circle of radius σ. T ). this ordering is preserved [129. t) = ∂vv C(p. a point with a circle as limiting shape. t) · n(p. Scalespace interpretation A shape scalespace interpretation for curve evolution under Euclidean heat ﬂow is studied by Kimia and Siddiqi [227]. the curve shrinks to a circular point. [71]. and smoothing properties follow from the fact that the total curvature decreases. the reader should be aware of the fact that – although this equation looks like a linear onedimensional heat equation – it is in fact nonlinear. 14) that κ(p. a nonconvex one becomes convex and. t).g. However. i. Guichard. (1. Moreover.99) This geometric heat equation is intrinsic. Theoretical results For the evolution of a smooth curve under its curvature. MCM belongs to the class of morphological scalespaces which satisfy the general axioms of Alvarez. As an image evolution. Since it is wellknown from diﬀerential geometry (see e. Lions and Morel [12].5. It is based on results of Evans and Spruck [129]. since the arclength v is again a function of the curve. They establish the semigroup property as architectural quality. equation (1. A convex curve remains convex. for an initial image f ∈ BUC(IR2 ). (1. (1. for t → T .100) In analogy to the dilation/erosion case it can be shown that.e. t) = ∂vv C(p. The time for shrinking a circle of radius σ to a point is given by 1 T = 2 σ2 . the number of extrema and inﬂection points of the curvature is nonincreasing. 169] that a smooth solution exists for some ﬁnite time interval [0. 151. t).
Aﬃne invariant intrinsic diﬀusion In analogy to the Euclidean invariant heat ﬂow. 363] constructed an aﬃne invariant ﬂow by replacing the Euclidean arclength v(p. Such an aﬃne arclength was proposed by Blaschke [44. there exist certain problems which also require invariance with respect to aﬃne transformations. t) that is invariant with respect to aﬃne transformations with det(A) = 1. t) := 0 det Cρ (ρ. This property is characteristic for all scalespaces of the Alvarez–Guichard–Lions–Morel axiomatic and distinguishes them from nonlinear diﬀusion ﬁlters.14)). Sapiro and Tannenbaum [362. t).101) where b ∈ IR2 denotes a translation vector and the matrix A ∈ IR2×2 is invertible. Cρρ (ρ. t) in (1. PARTIAL DIFFERENTIAL EQUATIONS relation between temporal and spatial scale for MCM is the same as for linear diﬀusion ﬁltering (cf. Css ) = 1. pp. (1. t) = κ(p. two level sets cannot move closer to one another than they were initially [129. Aﬃne transformations arise as shape distortions of planar objects when being observed from a large distance under diﬀerent angles.103) C(p.105) . · n(p. and it can be calculated as p s(p. t) 1 3 1 3 · n(p. Hence.102) By virtue of ∂ss C(p. 12–15] in 1923. t) = κ(p. 0) = C0 (p). t) we obtain the aﬃne invariant heat ﬂow ∂t C(p.104) (1.40 CHAPTER 1.2 Aﬃne invariant ﬁltering Motivation Although Euclidean invariant smoothing methods are suﬃcient in many applications. contrast cannot be enhanced. (1. t) (1.99) by an “arclength” s(p. It is characterized by det(Cs . Moreover. 1. A (full) aﬃne transformation is a mapping x → Ax + b (1.6. 227]. (1. t) 1 3 dρ. t).
aﬃne morphological scalespace (AMSS). Guichard.109) (c) The time for shrinking a circle of radius σ to a point is T = 3 4 σ3.106) we have the same results as for MCM and dilation/erosion concerning wellposedness of a viscosity solution which satisﬁes a maximum–minimum principle [12]. the aﬃne arclength parametrization q(p) := s(p) is the fastest way to shrink the aﬃne perimeter L(t) := det Cp (p.1. with three exceptions [363]: (a) Closed curves shrink to points with an ellipse as limiting shape (elliptical points). . Besides the name aﬃne invariant heat ﬂow. t) 1 3 dp. 5.106) (1. Lions and Morel [12] via an axiomatic scalespace approach. Theoretical results The curve evolution properties of aﬃne invariant heat ﬂow can be shown to be the same as in the Euclidean invariant case. 4 (1. under all ﬂows Ct = Cqq .6 CURVATUREBASED MORPHOLOGICAL PROCESSES Aﬃne invariant image evolution 41 When regarding the curve C(p.110) For the image evolution equation (1.107) (1. This image evolution equation has been discovered independently of and simultaneously with the curve evolution approach of Sapiro and Tannenbaum by Alvarez. The temporal evolution of an image under such an evolution resembles mean curvature motion. see Fig. Cpp (p. (1. and fundamental equation in image processing.108) 2 ux2 ux1 x1 − 2ux1 ux2 ux1 x2 + u2 1 ux2 x2 x where ξ is the direction perpendicular to ∇u. we shall brieﬂy sketch this reasoning below. After having mentioned some theoretical results. t). t). this equation is also called aﬃne shortening ﬂow. (b) The name aﬃne shortening ﬂow reﬂects the fact that. 2 1 1 3 1 3 (1.6(a). we end up with the evolution equation ∂t u = ∇u curv(u) = 3 = ∇u 3 uξξ . t) as a levelline of an image u(x.
Moreover. The scalespace reasoning of Sapiro and Tannenbaum investigates properties of the curve evolution. 352. Bruckstein and Shaked [62]. (1. a shape inclusion principle holds. This geodesic curvature ﬂow and other evolutions. can be regarded as steepest descent methods of energy functionals which have been proposed by Polyakov in the context of string theory [235]. a13 x1 + a23 x2 + a33 a13 x1 + a23 x2 + a33 ⊤ (1. Guichard. 113].112) with A = (aij ) ∈ IR3×3 and det A = 1.3 Generalizations In order to analyse planar shapes in a way that does not depend on their location in IR3 .or lengthpreserving equations [188. A) ≥ 0. It turns out that intrinsic heatequationlike formulations for the projective group are more complicated than the Euclidean and aﬃne invariant ones.42 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS Scalespace properties Alvarez. x2 )⊤ → a11 x1 + a21 x2 + a31 a12 x1 + a22 x2 + a32 . A study of heat ﬂows which are invariant under subgroups of the projective group can be found in [314.6.82) an additional aﬃne invariance axiom: For every invertible A ∈ IR2×2 and for all t ≥ 0. 366]. 133]. and Dibos [112. Simpliﬁcations of this axiomatic and related axioms for shape scalespaces can be found in [16]. see [362] and the references therein.106) is unique (up to temporal rescalings) when imposing on the scalespace axioms for (1. Lions and Morel [12] proved that (1. Euclidean and aﬃne invariant curve evolutions can also be modiﬁed in order to obtain area. [314]. 113]. one requires a multiscale analysis which is invariant under a general projective mapping (x1 . 131. 150. . 1.111) For this reason they call the AMSS equation also fundamental equation in image analysis. It is invariant to the bending (isometric mapping) of the surface. both for scalar and vectorvalued images. Olver et al. there exists a rescaled time t′ (t. An intrinsic heat ﬂow for images painted on surfaces has been investigated by Kimmel [232]. and that there is some evidence that they do not reveal the same smoothing scalespace properties [314]. 24] they point out that the Euclidean absolute curvature decreases as well as the number of extrema and inﬂection points of curvature. Research in this direction has been carried out by Faugeras and Keriven [130. such that Tt (Af ) = A(Tt′ f ) ∀ f ∈ BUC(IR2 ). Based on results of [229.
1. 304] approximate . and Moisan [289]. they are very fast and they e are entirely invariant under monotone greyscale transformations. 15. [285]. Niessen et al. Generalizations of MCM or AMSS to 3D images are investigated in [91. Cohignac et al. Lions and Morel to colour images are discussed in [82]. Merriman et al.1. On the one hand. 79]. Related discrete curve evolutions are analysed in [135.6 CURVATUREBASED MORPHOLOGICAL PROCESSES 43 Recently it has been suggested to modify AMSS such that any evolution at Tor Xjunctions of isophotes is inhibited [75]. Most direct approximations of morphological image evolution equations are based on the Osher–Sethian schemes [319.7. 298. it is diﬃcult to ﬁnd consistent approximations on a pixel grid which have good rotational invariance. this leads to an explicit ﬁnite diﬀerence method which approximates the spatial derivatives by central diﬀerences. respectively. A comparison of diﬀerent methods of these classes can be found in [95]. 359]. In this case two principal curvatures occur. 267. 374]. Such a ﬁltering intends to simplify the image while preserving its “semantical atoms” in the sense of Kanizsa [219]. 372. 362]. Since they can have diﬀerent sign. 95. Bruckstein et al. topological changes such as the splitting of conconvex structures may occur. whose corners belong to circles and ellipses. 1.5.6. This problem is reminescent of the creation of new extrema in diﬀusion scalespaces when going from 1D to 2D. A comparative evaluation of these approaches has been carried out by Lucido et al. convergence to polygones. Ruuth [347]. on the other hand. 288]. is established. Convergent settheoretic morphological schemes for MCM and AMSS have been proposed by Catt´ et al.4 Numerical aspects Due to the equivalence between curve evolution and morphological image processing PDEs. and possibilities to generalize the axiomatic of Alvarez. settheoretic morphological schemes and approximation schemes for the Eulerian formulation. the question arises of how to combine them to a process which simpliﬁes arbitrary surfaces without creating singularities. Curve evolution schemes can reveal perfect aﬃne invariance. [80. Guichard. Aﬃne scalespace axiomatics for image sequences (movies) have been established in [12. This is essentially the same tradeoﬀ as for circular structuring elements in classical settheoretic morphology. 287. In [61] discrete analogues of MCM and AMSS for the evolution of planar polygons are introduced. we have three main classes of numerical methods: curve evolution schemes. Curve evolution schemes are investigated by Mokhtarian and Mackworth [291]. 316]. 78. Diﬀerent variants of these schemes have been proposed in order to get better rotational invariance. [267]. In the case of MCM or AMSS. higher stability or less dissipative eﬀects [10. [95]. In complete analogy to the behaviour of the continuous equations. cf. In contrast to the 2D situation. [305. [61].
In order to solve the resulting linear system of equations he applies symmetric Gauß–Seidel iterations.16 MegaWave2 has been developed by Jacques Froment and Sylvain Parrino. CEREMADE. renounce consistency with the original equation as well as rotational invariance: round shapes evolve into polygonal structures. If one wants to have a good numerical approximation of aﬃne invariance. Nicolet and Sp¨ hler [301] investigate a consistent fully implicit scheme for MCM u leading to a nonlinear system of equations. An inherent problem of all ﬁnite diﬀerence schemes for morphological image evolutions are their dissipative eﬀects which create additional blurring of discontinuities. one can decompose the image into binary level sets. one has to decrease the temporal step size signiﬁcantly below the step size of experimental stability [16. It is solved by means of nonlinear Gauß– Seidel iterations. cf. More information can be found under http://www. 1. For AMSS an additional constraint appears: the behaviour of this equation is highly nonlocal.44 CHAPTER 1.ceremade. and run a ﬁnite diﬀerence method on them. 1.dauphine. PARTIAL DIFFERENTIAL EQUATIONS the spatial derivatives by Gaussian derivatives which are calculated in the Fourier domain.6. Concerning stability one observes that all these explicit schemes can violate a discrete maximum–minimum principle and require small time steps to be experimentally stable. 218]. 146]. and assembles the ﬁnal image by superimposing all evolved level sets [75]. MumfordShah segmentation.fr. 16 . and active contour models (cf. As a remedy. Comparing it with the explicit scheme they report a tradeoﬀ between the larger time step size and the higher computational eﬀort per step. map them into Lipschitzcontinuous images by applying a distance transformation.2.6) is available under the name MegaWave2. University Paris IX. since aﬃne invariance implies that circles are equivalent to ellipses of any arbitrary eccentricity. AMSS and many other modern techniques such as wavelets.3. Afterwards one extracts the processed level sets as the zerolevel sets of the evolved images. The natural price one has to pay for the excellent results is a fairly high computational eﬀort. 13. Using Gaussian derivatives for MCM or AMSS permits larger time steps for large kernels [304]. France. however. A consistent semiimplicit approximation of MCM which discretizes the ﬁrstorder spatial derivatives explicitly and the secondorder derivatives implicitly has been proposed by Alvarez [10]. Schemes of this type. One way to achieve unconditional L∞ stability for MCM is to approximate uξξ by suitable linear combinations of onedimensional secondorder derivatives along grid directions and to apply an implicit ﬁnite diﬀerence scheme [95. 75775 Paris Cedex 16. A software package which contains implementations of MCM. but their calculation in the Fourier domain is computationally expensive and aliasing may lead to oscillatory solutions.
315] is used). and which is called entropy scalespace or reaction–diﬀusion scalespace. for instance an analysis by means of Gabor functions [72. . since the theory of viscosity solutions is no longer applicable. This longtime forgotten method is similar to the Perona–Malik process (1. 260]. depending on the average grey value within a certain neighbourhood. If one wants to stay within the morphological framework one can combine diﬀerent morphological processes.6. 127]. They evolved the boundary curve by means of a backward Euclidean shortening ﬂow with a stabilizing bias term as in (1. Another modiﬁcation results from omitting the factor ∇u in the mean curvature motion (1. cf. 304] and blob detection [157. Wellposedness results are not known up to now. [126. 0) or max(κ. unless an “aﬃne invariant gradient” [231. 5. see also Fig. 230.2.7. 364.6 CURVATUREBASED MORPHOLOGICAL PROCESSES 45 1.1) or global PDEs for histogramme enhancement [358].1. shock ﬁltering ([14]. 132]. The potential of MCM for shape segmentation [290] has even been used for classifying chrysanthemum leaves [1]. 16]. This socalled min–max ﬂow produces a restored image as steadystate solution and reveals good denoising properties. If one aims to use these equations for image restoration one usually modiﬁes them by multiplying them with a term that reduces smoothing at highcontrast edges [13. 1.94). This leads to a process which is useful for analysing components of shape [228. 301] in medical images and to terrain matching [268]. Adapting them to tasks such as texture enhancement requires more sophisticated and more global feature descriptors than the gradient. Introducing a reaction term as in [102] allows to attract the solution to speciﬁed grey values which can be used as quantization levels [11]. 0). 453]. 304]. 234]. Gabor – the inventor of optical holography and the socalled Gabor functions – proposed a deblurring algorithm based on combining MCM with backward smoothing along ﬂowlines [149. Recently Steiner et al. Malladi and Sethian [272] propose to replace MCM by a technique in which the motion of level curves is based on either min(κ. All these preceding modiﬁcations are at the expense of renouncing the morphological invariance of the genuine operators (and also aﬃne invariance in the case of [364. 97. 365. 383. 304].g. 384. It is interesting to note that. MCM or AMSS have also been combined with other processes such as linear diﬀusion [13]. In order to improve images.56). already in 1965.6 (b). 365. MCM and AMSS have also been applied to denoising [305. for instance MCM and dilation/erosion. 96.7. see e. proposed a method for caricaturelike shape exaggeration [394].5 Applications The special invariances of AMSS are useful for shape recognition tasks [12. and for texture discrimination [265. for corner detection [15].36) for large image gradients. This corresponds to nonlinear diﬀusion ﬁlters and a restoration method by total variation minimization [345] which shall be described in 1.
6 Active contour models One of the main applications of curve evolution ideas appears in the context of active contour models (deformable models). x2 (s))⊤ which minimizes the functional E(C(s)) = C(s) α 2 Cs (s)2 + β 2 Css (s)2 − γ ∇f (C(s))2 ds. For this reason. ∂t (1. Witkin and Terzopoulos proposed the ﬁrst active contour model in a journal paper in 1988 [221].113) The ﬁrst summand is a membrane term causing the curve to shrink. (1. it is also called an explicit model. PARTIAL DIFFERENTIAL EQUATIONS 1. β and γ serve as weights of these expressions. while 3D active models are sometimes named active surfaces or active blobs. snake models incorporate many ideas ranging from energy minimization over curve evolution to the Perona–Malik ﬁlter and diﬀusion–reaction models.). It is therefore not surprising that they play an important role in several generalization and uniﬁcation attempts [356. while the third one represents the external (image) energy. 452]. 380. Active contour models can be used to search image features in an interactive way..114) which can be approximated by ﬁnite diﬀerences.6. tumour. Since this model makes direct use of the snake contour. the second one is a rigidity term which encourages a straight contour. The nonnegative parameters α. and an external energy term attracting the result to salient image features. and the third term pushes the contour to high gradients of the image f . We observe that terms 1 and 2 describe the internal energy. . Explicit snakes Kass.113) by steepest descent gives ∂C = αCss − βCssss − γ∇(∇f 2 ). Usually. the energy functional consists of two parts: an internal energy fraction which controls the smoothness of the result. Especially for assisting the segmentation of medical images they have become very popular [282]: The expert user gives a good initial guess of an interesting contour (organ. Their snakes can be thought of as an energyminimizing spline. Minimizing the functional (1.. which is attracted by image features such as edges.46 CHAPTER 1. Apart from these practical merits. the result will depend on the choice of the initial curve and a good segmentation requires an initial curve which is close to the ﬁnal segment. Such a snake is represented by a curve C(s) = (x1 (s). 2D versions are also called snakes. The main . which will be carried the rest of the way by some energy minimization. A 3D version of such an active contour model is presented in [401].
Implicit snakes In 1993 some of the inherent diﬃculties of explicit snakes have been solved by replacing them by a socalled implicit model. i. Hence. 17 .1. T ). 435].32): it becomes small for large ∇fσ  = ∇Kσ ∗f . • ν∇u describes motion in normal direction. Moreover. β and γ. An advantage of implicit snake models is their topological ﬂexibility: The contour may split. and Caselles and Coll investigated related approaches for image sequences [74]. it slows down the snake as soon as it approaches signiﬁcant edges of the original image f . This socalled balloon force [93] is required for pushing a level set into concave regions. the balloon force ν.94).115) have the following meaning: • ∇u div (∇u/∇u) is the curvature term of MCM which smoothes level sets. see (1. they use essentially only one remaining parameter. Catt´. Then u0 is evolved under a PDE which includes knowledge about the original image f : ∂t u = g(∇fσ 2 ) ∇u div ∇u +ν . On Recently McInerley and Terzopoulos have proposed modiﬁed explicit deformable models which allow topological changes [281. The idea behind implicit snakes is to embed the initial curve C0 (s) as a zero level curve into a function u0 : IR2 → IR. It was discovered by Caselles. Sethian and Vemuri [273]. and later on by Malladi. For this model Caselles et al. a compensation for the property of MCM to create convex regions.115) This evolution is stopped at some time T . it is also sometimes not easy to ﬁnd a good balance between the parameters α. when the process does hardly alter anymore. Whitaker and Chen developed similar implicit active contour models for 3D images [436. ∇u (1. 283]. • g is a stopping function such as the Perona–Malik diﬀusivity (1. e Coll and Dibos [73]. could prove wellposedness in the viscosity sense [73].e. The terms in (1.6 CURVATUREBASED MORPHOLOGICAL PROCESSES 47 disadvantage of the preceding model is its topological rigidity: a classical explicit snake cannot split in order to segment several objects simultaneously17 . for instance by using the distance transformation. In practice. dilation or erosion depending on the sign of ν. and the ﬁnal contour C is extracted as the zero level curve of u(x. This allows simultaneous segmentation of multiple objects.
∇u (1. They have been proposed simultaneously by Caselles.118) by g(∇uσ 2 ). 271].118) This active contour model is parameterfree. A theoretical analysis of geodesic snakes concerning existence.48 CHAPTER 1. Kumar. the process does not really stop at the desired result. Recently also an aﬃne invariant analogue to geodesic active contours has been proposed [315]. PARTIAL DIFFERENTIAL EQUATIONS the other hand. Techniques which can be related to geodesic active contours have also been used for solving 3D vision problems such as stereo [134] and motion analysis [322]. In order to address the initialization problem. but often a speed term νg(∇fσ 2 )∇u is added to achieve faster and more stable attraction to edges. and it is diﬃcult to interpret implicit snakes in terms of energy minimization.116) where g denotes again a Perona–Malik diﬀusivity of type (1. Then the .113) by E(C) = C(s) α 2 Cs (s)2 − γ g(∇fσ (C)2 ) ds (1. Embedding the initial curve as a level set of some image u0 and applying a descent method to the corresponding EulerLagrange equation leads to the image evolution PDE ∂t u = ∇u div g(∇fσ 2 ) ∇u . They name this technique reaction–diﬀusion bubbles. 226]. Tek and Kimia use implicit active contours starting from randomly chosen seed points. both in 2D [399] and in 3D [400]. These snakes replace the contour energy (1. Geodesic snakes Geodesic snakes make a synthesis of explicit and implicit snake ideas. Olver. Kimmel and Sapiro [76] and by Kichenassamy. Selfsnakes The properties of geodesic snakes induced Sapiro to use a related technique for image enhancement [357]: he replaced g(∇fσ 2 ) in (1. 226. (1. Under some additional assumptions they derive that minimizing (1. and extensions to 3D images are studied in [77.117) This is nothing else than ﬁnding a curve of minimal distance (geodesic) with respect to some imageinduced metric.116) is equivalent to searching min C C(s) g ∇fσ (C(s))2 Cs (s) ds. Tannenbaum and Yezzi [226]. it only slows down.32). uniqueness and stability of a viscosity solution can be found in [76.
123) plays an important role for shock calculations. Grossmann and Friedman [125]. 1. cf. Chen. This would be useful to restore discontinuities such as edges. 422] for related ideas. Generalizations of selfsnakes to vectorvalued images [357. It maps the image grey values to gas dynamical parameters and solves the compressible Euler equations using shockcapturing total variation diminishing (TVD) techniques based on Godunov’s method. we observe striking similarities with the Perona–Malik process from Section 1. (1. Vemuri and Wong [89] established existence and stability of a unique viscosity solution to a modiﬁed selfsnake process. The restored image is given by the steadystate of ∂t u = ∇u div g(∇uσ 2 ) ∇u + β ∇u (f −u) ∇u (β > 0).6(c). Their model contains a reaction term which inhibits smoothing at edges and keeps the ﬁltered image u close to the original image f .121) Thus. however.1. see also [414. 18 . 361] can be obtained using Di Zenzo’s ﬁrst fundamental form for colour images [114]. it cannot be excluded that a selfsnake without spatial regularization reveals the same illposedness problems as the Perona–Malik ﬁlter [356]. Although this equation cannot be cast in divergence form.18 Another image enhancement method that is close in spirit is due to Eidelman. one may ask if it is possible to apply related ideas to image processing.119) (1.120) = g(∇u2) uξξ + ∇⊤ g(∇u2) ∇u with ξ ⊥ ∇u. Below we shall focus on two important TVbased image restoration techniques which have been pioneered by Osher and Rudin: TVpreserving methods and techniques which are TVminimizing subject to certain constraints.122) The temporal evolution of a regularized selfsnake without reaction term is depicted in Fig.3. (1. For σ = 0 this gives ∂t u = ∇u div g(∇u2) ∇u ∇u (1. 5. [380].7 Total variation methods T V (u) := Ω Inspired by observations from ﬂuid dynamics where the total variation (TV) ∇u dx (1.7 TOTAL VARIATION METHODS 49 snake becomes a selfsnake no longer underlying external image forces.1: the latter can be written as ∂t u = g(∇u2) ∆u + ∇⊤ g(∇u2) ∇u. For σ > 0.
By means of our knowledge from morphological processes. From a practical point of view. They prove that their semiimplicit ﬁnitediﬀerence scheme has a unique solution which satisﬁes a maximum–minimum principle. 0) = f (x). PARTIAL DIFFERENTIAL EQUATIONS 1.1 TVpreserving methods In 1990 Osher and Rudin have proposed to restore blurred images by shock ﬁltering [317]. Alvarez and Mazorra [14] replace the operator L(u) = uηη in (1. blocky . 1.50 CHAPTER 1. sgn(F (u)) = sgn(u).7. both in the continuous and discrete case. Here. (1.g. we recognize that this ﬁlter aims to produce a ﬂow ﬁeld that is directed from the interior of a region towards its edges where it develops shocks.2 TVminimizing methods Total variation is good for quantifying the simplicity of an image since it measures oscillations without unduly punishing discontinuities. the goal is to obtain a piecewise constant steadystate solution with discontinuities only at the edges of the initial image. Osher and Rudin have also proposed another TVpreserving deblurring technique [318]. the Laplacian L(u) = ∆u or the secondorder directional derivative L(u) = uηη with η ∇u. It solves the linear diﬀusion equation backwards in time under the regularizing constraint that the total variation remains constant. For this reason. This stabilization can be realized by keeping local extrema ﬁxed during the whole evolution.124) with F (u) := sgn(u) arises as the PDE formulation of a classical image enhancement algorithm by Kramer and Bruckner [243]. Kramer and Bruckner proved in 1975 that their Ndimensional discrete scheme converges after a ﬁnite number of iterations to a state where each point is a local extremum. For the twodimensional case not many theoretical results are available except for a discrete maximum–minimum principle. These ﬁlters calculate the restored image as the steadystate solution of the problem ∂t u = −∇u F (L(u)). For this reason.7. and L(u) is a secondorder elliptic operator whose zerocrossings correspond to edges.124) (1.124) by a Gaussiansmoothed version L(Kσ ∗u) = Kσ ∗uηη and supplement the resulting equation with a noiseeliminating mean curvature process. e.125) u(x. It has been shown that a onedimensional version of this ﬁlter preserves the total variation and satisﬁes a maximum–minimum principle. Recently van den Boomgaard [52] pointed out that the 1D version of (1. TVpreserving methods suﬀer from the problem that ﬂuctuations due to noise do also create shocks. Thus.
19 To ﬁx ideas.130) is equivalent to the constrained TV minimization. Moreover.56) is not intended to satisfy the noise constraint exactly [346].7 TOTAL VARIATION METHODS 51 images (consisting only of a few almost piecewise constant segments) reveal very small total variation. let us study an example.129). In [345] a gradient descent method is proposed to solve (1. PDE methods can be applied: A solution of (1.127) (1. but – in contrast to TV approaches – equation (1. Osher and Fatemi [345] have proposed to minimize the total variation under constraints which reﬂect assumptions about noise. Ω (1.56). which relates TVminimizing techniques to MCM.128) veriﬁes necessarily the Euler equation div ∇u − µ − λ(u−f ) = 0 ∇u (1.128) u dx = Ω Ω f dx. Interestingly.129) is identical with the curvature.129) looks similar to the steadystate equation of the diﬀusion–reaction equation (1.126)–(1. One may also reformulate the constrained TV minimization as an unconstrained problem [83]: The penalized least square problem min u 1 u−f 2 2 L2 (Ω) +α Ω ∇u dx (1. Given an image f with additive noise of zero mean and known variance σ 2 .126) subject to (u−f )2 dx = σ 2 . 19 Related ideas have also been developed by Geman and Reynolds [153]. if α is related to the Lagrange 1 multiplier λ via α = λ . we seek a restoration u satisfying min u Ω ∇u dx (1. In order to restore noisy blocky images. the divergence term in (1.129) with homogeneous Neumann boundary conditions. Rudin. .1. In order to solve this constrained variational problem. It uses an explicit ﬁnite diﬀerence scheme with central and onesided spatial diﬀerences and adapts the Lagrange multiplier by means of the gradient projection method of Rosen. The (unknown) Lagrange multipliers µ and λ have to be determined in such a way that the constraints are fulﬁlled. (1.
• TVminimizing methods have also been used for estimating discontinuous blurring kernels such as motion or outoffocus blur from a degraded image.52 CHAPTER 1. To overcome the problem that the total variation integral contains the nondiﬀerentiable argument ∇u. problems of type (1. and to enhancing pictures from confocal microscopy [409] and tomography [115. 118. 253]. to improving material from criminal and civil investigations as court evidence [344]. They are useful for enhancing reconstruction algorithms for inverse scattering problems [37]. they establish relations between TVminimizing methods and nonlinear diﬀusion techniques [397]. and they have established uniqueness. By adapting α to the local image structure. PARTIAL DIFFERENTIAL EQUATIONS In recent years. 83]. 83]. 205. or numerical analysis [2. Total variation methods have been applied to restoring images of military relevance [345. This leads to TVbased blind deconvolution algorithms [86]. 410. as ∇u ranges from 0 to ∞. . 409]. 85. Much research is done in order to ﬁnd eﬃcient numerical methods for which convergence can be established. one applies regularization strategies or techniques from nonsmooth optimization. 117. 344. see [46]. 116. noise with blur. Chambolle and Lions [83] have extended the existence proof to noncompact operators (which comprises also the situation without blur). (1.131) where p(∇u) decreases from 2 to 1. • The tendency of TVminimizing to create piecewise constant structures can cause undesired eﬀects such as the creation of staircases at sigmoidlike edges [116.130) as a scale parameter [396]. it has been proposed to minimize the L1 norm of expressions containing also higherorder derivatives [344. where the generalized TV norm is chosen as the l2 norm of the TV norms of the separate channels. Recently. 396].130) have attracted much interest from mathematicians working on inverse problems. 115. 253. 253. 87. As a remedy. • Strong and Chan have identiﬁed the parameter α in (1. or other types of noise [262. 84. [262] and Dobson and Santosa [115] have shown the existence of BV(Ω)solutions for problems of this type. 262. 204. 346. Lions et al. and the idea of L1 norm minimization has also led to improved optic ﬂow algorithms [245]. 83]. 346. • They have been applied to colour images [45]. TVminimizing methods have been generalized in diﬀerent ways: • The constrained TVminimization idea is frequently adapted to other constraints such as blur. Another possibility is to consider the constrained minimization of B(u) := Ω ∇u p(∇u) dx. optimization.
we shall see that anisotropic nonlinear diﬀusion processes can share many advantages of the scalespace and the image enhancement world. or the use of explicit schemes causes restrictive step size limitations. In particular. they possess the undesirable property that they do not permit contrast enhancement and that they may blur and delocalize structures. Other interesting image restoration topics have found less attention. The goal of the subsequent chapters is to develop a theory for nonlinear anisotropic diﬀusion ﬁlters which addresses some of the abovementioned shortcomings.8 Conclusions and further scope of the book Although we have seen that there exists a large variety of PDEbased scalespace and image restoration methods which oﬀer many advantages. for instance unsolved wellposedness questions or the search for eﬃcient minimizers of nonconvex or nondiﬀerentiable functionals. For both scalespace and restoration methods many questions concerning their discrete realizations are still open: discrete scalespace results are frequently missing. for which the results hold. for instance enhancement of coherent ﬂowlike structures in textures. speciﬁc anisotropic models are presented which permit applications beyond segmentation and edge enhancement tasks. The latter comprises for instance semiimplicit techniques for which unconditional stability in the L∞ norm is proved. Both scalespace and wellposedness properties carry over from the continuous to the semidiscrete and discrete setting. Finally. Pure restoration methods such as diﬀusion–reaction equations or TVbased techniques do allow contrast enhancement and lead to stable structures but can suﬀer from theoretical or practical problems. most imageenhancing PDE methods focus on edge detection and segmentation problems. They shall serve as a motivation for the theory which will be explored in the subsequent chapters. The general framework. for some applications. . Linear diﬀusion and morphological scalespaces are wellposed and have a solid axiomatic foundation.7 TOTAL VARIATION METHODS 53 1. we have also become aware of some limitations. includes also linear and isotropic nonlinear diﬀusion ﬁlters. minimization algorithms can get trapped in a poor local minimum. A scalespace interpretation is presented which does not exclude contrast enhancement. Moreover. and wellposedness results are established.1. On the other hand.
54 CHAPTER 1. PARTIAL DIFFERENTIAL EQUATIONS .
where uσ (x.1) (2.. 0) = f (x) Ω × (0. 2.3. . Ω. a1 ) × (0. We shall see that.. we adapt the diﬀusion process to the structure tensor. one would usually let it depend on the edge estimator ∇uσ (cf.3) Hereby. it is possible to establish wellposedness and regularity results and to prove a maximum–minimum principle. t) := (Kσ ∗ u(. After sketching invariances of the resulting scalespace. besides the extremum principle.Chapter 2 Continuous diﬀusion ﬁltering This chapter presents a general continuous model for anisotropic diﬀusion ﬁlters. n = 0 on on on u(x.1 Basic ﬁlter structure Let us consider a rectangular image domain Ω := (0. 1. which may be obtained by mirroring ˜ at Γ (cf. [81]).4) and u denotes an extension of u from Ω to IR2 . Γ × (0. ∞). 55 . (2. To this end. Since the proof does not require any monotony assumption it applies also to contrastenhancing diﬀusion processes. a wellknown tool for analysing local orientation. t)) (x) ˜ (σ > 0) (2.2) (2. we focus on analysing its smoothing properties. a2 ) with boundary Γ := ∂Ω and let an image be represented by a mapping f ∈ L∞ (Ω). In order to adapt the diﬀusion tensor D ∈ IR2×2 to the local image structure. the Euclidean scalar product on IR2 . Under fairly weak assumptions on the class of ﬁlters. ∞). a large class of associated Lyapunov functionals plays an important role in this context [414. The class of anisotropic diﬀusion ﬁlters we are concerned with is represented by the initial boundary value problem ∂t u = div (D ∇u) D∇u. n denotes the outer normal and . analyses its theoretical properties and gives a scalespace interpretation. 415].3).
we reconsider the vectorvalued structure descriptor ∇uσ within a matrix framework. shape cues [256. This will be presented next. Averaging this orientation information can be accomplished by convolving J0 (∇uσ ) componentwise with a Gaussian Kρ . Matrices of this type are useful for many diﬀerent tasks. but point in opposite directions. pp.56 CHAPTER 2. 147–153]. [168. 182. we shall choose a more general structure descriptor which comprises the edge detector ∇uσ . pp. 219–258]. corners and Tjunctions [145.5) σ (2. 220]. The structure tensor – also called interest operator. Related approaches can also be found in [39. for instance for analysing ﬂowlike textures [340. we need methods which take into account how the orientation of the (smoothed) gradient changes within the vicinity of any investigated point. 2. 40. v2 with v1 j22 − j11 + 1 j11 +j22 ± 2 The corresponding eigenvalues µ1 and µ2 are given by µ1. Let us focus on some aspects which are of importance in our case. they describe the average contrast in the . 309]. (2. In order to identify features such as corners or to measure the local coherence of structures. As they integrate the variation of the grey values within a neighbourhood of size O(ρ). 31]. In order to study orientations instead of directions. 310. This gives the structure tensor Jρ (∇uσ ) := Kρ ∗ (∇uσ ⊗ ∇uσ ) (ρ ≥ 0). CONTINUOUS DIFFUSION FILTERING However. 349–382] and spatio–temporal image sequences [209.6) It is not hard to verify that the symmetric matrix Jρ = j11 j12 is positive semidefj12 j22 inite and possesses orthonormal eigenvectors v1 .2 The structure tensor has an orthonormal basis of eigenvectors v1 . but also allows to extract more information. The matrix J0 resulting from the tensor product J0 (∇uσ ) := ∇uσ ⊗ ∇uσ := ∇uσ ∇uT (2.2 = 2j12 . scatter matrix or (windowed second) moment tensor – is an important representative of this class. To this end.8) where the + sign belongs to µ1 . we have to identify gradients which diﬀer only by their sign: they share the same orientation.7) (2. pp. 2 (j11 −j22 )2 + 4j12 2 (j11 −j22 )2 + 4j12 . v2 with v1 ∇uσ and v2 ⊥ ∇uσ . The corresponding eigenvalues ∇uσ 2 and 0 give just the contrast (the squared gradient) in the eigendirections.
we denote by Cp (X.3 Theoretical results In order to discuss wellposedness results. T ]) such that T 2 H1 (Ω) u L2 (0.T ]. An example which illustrates the advantages of the structure tensor for analysing coherent patterns can be found in Figure 5. T ]. µ1 and µ2 can be used as descriptors of local structure: Constant areas are characterized by µ1 = µ2 = 0. (2. 2. the coherence direction. C([0. corners can be identiﬁed by µ1 ≥ µ2 ≫ 0. Furthermore. we observe that v1 is the orientation with the highest grey value ﬂuctuations.L2 (Ω)) := max u(t) [0. T . L2 (Ω)) is deﬁned as the space of continuous functions u : [0. The parameter σ is called local scale or noise scale. Now we can give a precise formulation of the problem we are concerned with. strongly measurable on [0.10) and identify it with its dual space. Let H1 (Ω) be the Sobolev space of functions u(x) ∈ L2 (Ω) with all distributional derivatives of ﬁrst order being in L2 (Ω).12) As usual.T ] L2 (Ω) .T . and the expression 2 (µ1 − µ2 )2 = (j11 −j22 )2 + 4j12 (2. straight edges give µ1 ≫ µ2 = 0.2. H1(Ω)) be the space of functions u. By virtue of µ1 ≥ µ2 ≥ 0. THEORETICAL RESULTS 57 eigendirections. and v2 gives the preferred local orientation.9) becomes large for anisotropic structures.H1 (Ω)) := u(t) 0 dt 1/2 < ∞. T ] → L2 (Ω) supplemented with the norm u C([0. see Section 5. We need the following prerequisites: .2. (2. the integration scale ρ should reﬂect the characteristic window size over which the orientation is to be analysed. Y ) the set of Cp mappings from X to Y .10 (d). Let L2 (0. Thus. We equip H1 (Ω) with the norm u := u 2 L2 (Ω) 2 1/2 H1 (Ω) + i=1 ∂xi u 2 L2 (Ω) (2. Presmoothing in order to obtain ∇uσ makes the structure tensor insensitive to noise and irrelevant details of scales smaller than O(σ). It is a measure of the local coherence.3.11) In a similar way. T ] with range in H1 (Ω) (for the Lebesgue measure dt on [0. let us ﬁrst recall some useful notations.
L2 (Ω)) ∩ L2 (0. and σ. T . 1 . L2(Ω) .13) (2. t) ≤ b on Ω × (0. T ]. This topic will not be addressed here.14) ¯ Moreover. This solution depends continuously on f with respect to . (2. T .1 regularity. and it fulﬁls the extremum principle a ≤ u(x. (C3) Uniform positive deﬁniteness: ¯ For all w ∈ L∞ (Ω. 414]. 0) = f (x) D(Jρ (∇uσ ))∇u. and consider the problem Ω Ω ∂t u = div (D(Jρ(∇uσ )) ∇u) u(x. Ω × (0. ∂t u ∈ L (0. extremum principle) The problem (Pc ) has a unique solution u(x. b := ess sup f . T > 0. CONTINUOUS DIFFUSION FILTERING Assume that f ∈ L∞ (Ω).58 CHAPTER 2. (Pc ) Under these assumptions the following theorem. Let a := ess inf f . H (Ω)). (C2) Symmetry: d12 (J) = d21 (J) for all symmetric matrices J ∈ IR2×2 . IR2 ) with w(x) ≤ K on Ω. ρ ≥ 0. n = 0 on on on Ω. IR2×2 ). T ]. 2 1 (2.15) For a complete wellposedness proof one also has to establish stability with respect to perturbations of the diﬀusion equation. which generalizes and extends results from [81. T ]). T ]. u ∈ C∞ (Ω×(0. can be proved. Γ × (0. where the diﬀusion tensor D = (dij ) satisﬁes the following properties: (C1) Smoothness: D ∈ C∞ (IR2×2 . there exists a positive lower bound ν(K) for the eigenvalues of D(Jρ (w)). t) in the distributional sense which satisﬁes u ∈ C([0. T ]. Theorem 1 (Wellposedness. H1 (Ω)).
We restrict ourselves to proving only the maximum principle. t) − b) dx. t)−b) ∂t u(x. Lions. L2 (Ω)) to the strong solution of (Pc ). Since W (0. i. T . p. we have G(u(x. Smoothness follows from classical bootstrap arguments and the general theory of parabolic equations [246]. The minimum principle follows from the maximum principle when being applied to the initial datum −f . ϕ(t) := Ω H(u(x. H1(Ω)). one deduces that u(t) ∈ H2 (Ω) for all t > 0. H1(Ω)) dt 2 2 W (0. (b) Extremum principle In order to prove a maximum–minimum principle. T . T ] × Ω).e. By the Cauchy–Schwarz inequality. L (Ω)). The basic idea of the uniqueness proof consists of using energy estimates for the diﬀerence of two solutions. T ]. Finally an iterative linear scheme is investigated. u = U(u). THEORETICAL RESULTS Proof: 59 (a) Existence. T ]. 0] and 0 < G′ (s) ≤ C on (0. convex and weakly compact subset W0 of into itself. we just sketch the basic ideas of this proof. uniqueness and regularity are straightforward anisotropic extensions of the proof for the isotropic case studied by Catt´. Now. By iterating. T . s ∈ IR. T ) := w ∈ L2 (0. t) dx ≤ C · u(t)−b L2 (Ω) Ω · ∂t u(t) L2 (Ω) . t ∈ [0. [58]. we utilize Stampacchia’s truncation method (cf. we deﬁne s H(s) := 0 G(σ) dσ. U reveals a ﬁxed point u ∈ W0 .3. Let G ∈ C1 (IR) be a function with G(s) = 0 on (−∞. one can establish that u is a ¯ strong solution of (Pc ) and u ∈ C∞ ((0. such that the Gronwall–Bellman inequality can be applied. whose solution is shown to converge in C([0. Then. dw ∈ L2 (0. Morel and e Coll [81]. 211). T ) is contained in L (0. Existence can be proved using Schauder’s ﬁxed point theorem. uniqueness and regularity Existence. ∞) for some constant C. Therefore. Then one shows that U is a weakly continuous mapping from a nonempty. Since u(t) ∈ H1 (Ω) for all t > 0.2. with compact inclusion. One considers the solution U(w) of a distributional linear version of (Pc ) where D depends on some function w instead of u. uniqueness follows from the fact that both solutions start with the same initial values.
Hence. and (2. Due to the smoothness of u for t > 0. 2 (2. (2.17) tends to 0 = ϕ(0) for t → 0+ which proves the continuity of ϕ(t) in 0.16) we have C u(t)−f 2 2 L2 (Ω) . t) − b ≤ 0 almost everywhere (a. In the same way as in the uniqueness proof in [81]. T ]. Therefore. w the corresponding solutions.13). T ] ≤ c · ∇u(t) 2 L2 (Ω) · u(t)−w(t) 2 L2 (Ω) . n dS =0 = Γ − Ω G (u−b) ∇u. ϕ ≥ 0 on [0. we obtain u(x. and we get dϕ = dt = Ω G(u−b) ∂t u dx Ω G(u−b) div (D(Jρ (∇uσ )) ∇u) dx G(u−b) D(Jρ (∇uσ )) ∇u. for all t ∈ [0. T ].60 CHAPTER 2.17) Since u ∈ C([0. it follows that ϕ ≡ 0 on [0. (c) Continuous dependence on the initial image Let f. pp. h ∈ L∞ (Ω) be two initial values and u. the righthand side of (2. we ﬁnally end up with the assertion ¯ u(x. T ].e. T ]. t) ≤ b on Ω × (0. ϕ is diﬀerentiable for t > 0. By means of H(s) ≤ 0 ≤ ϕ(t) ≤ Ω C 2 s. . H(u(x.) on Ω. Applying the Gronwall–Bellman lemma [57. L2 (Ω)). D(Jρ(∇uσ )) ∇u dx ≥0 ≥0 ′ ≤ 0. CONTINUOUS DIFFUSION FILTERING and by virtue of (2. one shows that there exists some constant c > 0 such that d dt u(t)−w(t) 2 L2 (Ω) ϕ(0) = 0.16).14) we know that the righthand side of this estimate exists. 156–137] yields u(t)−w(t) 2 L2 (Ω) ≤ f −h 2 L2 (Ω) · exp c · t 0 ∇u(s) 2 L2 (Ω) ds . T ]. Now from ϕ ∈ C[0. t)−f (x)) dx ≤ (2.
More restrictive requirements – for instance f ∈ BUC(IR2 ) in order to apply the theory of viscosity solutions – are not necessary in our case. ∇uσ is also bounded. s)))∇u(x. such that t 0 T ∇u(s) ∇u(s) T 2 L2 (Ω) ds ≤ 2 L2 (Ω) ds 0 1 ≤ ν = 1 ν 1 ν 0 Ω T ∇u(x.H1 (Ω)) L2 (0. let ǫ > 0 and choose δ := ǫ · exp Then for f −h L2 (Ω) −c u 2ν L2 (0. s) · div D(Jρ (∇uσ (x. T ].13). Thus. we know that the righthand of this estimate exists. and prerequisite (C3) implies that there exists some constant ν = ν(σ. Now.14). (b) Moreover.H1 (Ω)) ∂t u L2 (0. our proof does not require any monotony assumption. This has the advantage that contrastenhancing processes are permitted as well. Remarks: (a) We observe a strong smoothing eﬀect which is characteristic for many diffusion processes: under fairly weak assumptions on the initial image (f ∈ L∞ (Ω)) we obtain an inﬁnitely often diﬀerentiable solution for arbitrary small positive times.T . 2 which proves the continuous dependence on the initial data. By virtue of (2. s) dx ds u(x. < δ. T ]. s).T . Chapter 5 will illustrate this by presenting examples where contrast is enhanced. THEORETICAL RESULTS 61 ¯ By means of the extremum principle we know that u is bounded on Ω×[0. f L∞ (Ω) ) > 0. s) dx ds L2 (Ω) 0 Ω T ≤ ≤ u(s) 0 ∂t u(s) ∂t u L2 (Ω) ds 1 u ν L2 (0. the preceding results imply L2 (Ω) u(t)−w(t) <ǫ ∀ t ∈ [0.3.T .2.T . (2. s)))∇u(x. . D(Jρ(∇uσ (x.H1 (Ω)) .H1 (Ω)) .
1 Invariances Let u(x. t). we shall not focus on further investigations of architectural requirements like recursivity. This does not contradict contrast enhancement: In the case of two similar images.2. if we start for instance with an image within the range [0. This property is responsible for the uniform positive deﬁniteness of the diﬀusion tensor. regularity and locality. we turn to a more crucial task. 255].18) where u(t) := u(. as these qualities do not distinguish nonlinear diﬀusion scalespaces from other ones. spatiotemporal image sequences or slices from medical CT or MRI sequences.4 Scalespace properties Let us now investigate scalespace properties of the class (Pc ) and juxtapose the results to other scalespaces. informationreducing image transformation. Much of the terminology used below is borrowed from [12].2 (d) The extremum principle oﬀers the practical advantage that. where one leads to contrast enhancement and the other not. 2. t) be the unique solution of (Pc ) and deﬁne the scalespace operator Tt by Tt f := u(t). we will never obtain results with grey value such as 257. Afterwards. t ≥ 0. (2. (e) The wellposedness results are essentially based on the fact that the regularization by convolution with a Gaussian allows to estimate ∇uσ L∞ (Ω) by u L∞ (Ω) . It is also closely related to smoothing scalespace properties. the regularization damps the enhancement process in such a way that both images do not diﬀer much after ﬁltering.62 CHAPTER 2.4. The properties we discuss now illustrate that an invariance of Tt with respect to some image transformation P is characterized by the fact that Tt and P commute. 2.4. To this end. This is of importance when considering stereo image pairs. as we shall see in 2. 2 . CONTINUOUS DIFFUSION FILTERING (c) The continuous dependence of the solution on the initial image has significant practical impact as it ensures stability with respect to perturbations of the original image. namely the question in which sense our evolution equation – which may allow contrast enhancement – can still be considered as a smoothing.. We start with brieﬂy discussing invariances. since we know that similar images remain similar after ﬁltering.
Average grey level invariance Average grey level invariance is a further property in which diﬀusion scalespaces diﬀer from morphological scalespaces.20) Reverse contrast invariance From D(Jρ (−∇uσ )) = D(Jρ (∇uσ )). Moreover. the evolution PDEs of the latter ones are not of divergence form and do not preserve the mean grey value. (2. we have Tt (0) = 0. a constant function is not aﬀected by diﬀusion ﬁltering. it follows that Tt (−f ) = −Tt (f ) ∀ t ≥ 0. (2. In general. A constant average grey level is essential for scalespace based segmentation algorithms such as the hyperstack [307. and the ﬁltered images will also be shifted by the same constant.23) u(x. but not of u. It is also a desirable quality for applications in medical imaging where grey values measure physical qualities of the depicted object. we may shift the grey level range by an arbitrary constant C. When reversing the contrast. The average grey level 1 µ := f (x) dx Ω Ω (2.19) (2. Then the Cauchy–Schwarz inequality . SCALESPACE PROPERTIES Grey level shift invariance 63 Since the diﬀusion tensor is only a function of Jρ (∇uσ ). Tt (f + C) = Tt (f ) + C ∀ t ≥ 0. (2. Proposition 1 (Conservation of average grey value).2. 408]. the role of dilation and erosion has to be exchanged as well.21) This property is not fulﬁlled by classical morphological scalespace equations like dilation and erosion.22) is not aﬀected by nonlinear diﬀusion ﬁltering: 1 Ω Proof: Deﬁne I(t) := Ω Tt f dx = µ Ω ∀ t > 0.4. t) dx for all t ≥ 0. for instance proton densities in MR images. Therefore.
T ]. (2. CONTINUOUS DIFFUSION FILTERING I(t)−I(0) = 2 Ω u(x. (2. it makes no diﬀerence whether the orthogonal transformation is applied before or after diﬀusion ﬁltering: Tt (Rf ) = R(Tt f ) ∀ t ≥ 0. n dS = 0.27) . it also makes sense to investigate Euclidean transformations of an image.25) When studying diﬀusion ﬁltering as a pure initial value problem in the domain IR . Average grey level invariance may be described by commuting operators. For t > 0. when introducing an averaging operator M : L1 (Ω) → L1 (Ω) which maps f to a constant image with the same mean grey level: 1 f (x) dx ∀ y ∈ Ω. If we apply R to f by deﬁning Rf (x) := f (Rx). This leads us to translation and isometry invariance.24) (Mf )(y) := Ω Ω Then Proposition 1 and grey level shift invariance imply that the order of M and Tt can be exchanged: M(Tt f ) = Tt (Mf ) ∀ t ≥ 0.64 implies CHAPTER 2. then the eigenvalues of the diﬀusion tensor are unaltered and any eigenvector v is transformed into Rv. Since u ∈ C([0. Thus. t)−f (x) dx ≤ Ω1/2 u(t)−f L2 (Ω) .26) This is a consequence of the fact that the diﬀusion tensor depends on Jρ (∇uσ ) solely. (2. the preceding inequality gives the continuity of I(t) in 0. (2. Theorem 1. I(t) must be constant for all t ≥ 0. the divergence theorem and the boundary conditions yield dI = dt ∂t u dx = Ω Γ D(Jρ (∇uσ ))∇u. L (Ω)). Then diﬀusion ﬁltering fulﬁls Tt (τh f ) = τh (Tt f ) ∀ t ≥ 0. but not explicitly on x. 2 Translation invariance Deﬁne a translation τh by (τh f )(x) := f (x + h). 2 Hence. Isometry invariance Let R ∈ IR2×2 be an orthogonal transformation.
isointensity linking through the scales is possible and a structure at a coarse scale can (in principle) be traced back to the original image (causality). θ))) is bounded. As in the linear diﬀusion case. (2. cf.2 Informationreducing properties Nonenhancement of local extrema Koenderink [240] required that a scalespace evolution should not create new level curves when increasing the scale parameter. Thus. the ﬁrst term of the righthand side of (2.28) A suﬃcient condition for the causality equation (2. for instance.28)..30) (2. if ξ is a local maximum. Suppose that ξ ∈ Ω is a local extremum of u(. ∂t u(ξ.4. For this reason.32) vanishes in (ξ. He showed that this constraint can be written as sgn(∂t u) = sgn(∆u). (2. he imposed that at spatial extrema with nonvanishing determinant of the Hessian isophotes in scalespace are upwards convex. and ∂t u < 0 if ξ is a maximum. [30] in the context of linear diﬀusion ﬁltering. Nonenhancement of local extrema has ﬁrst been used by Babaud et al. ∂t u(ξ. θ) < 0. if ξ is a local minimum. θ). If this is satisﬁed.3 Theorem 2 (Nonenhancement of local extrema). 1.2. θ) with nonvanishing Hessian. This implication is easily seen: In the ﬁrst case.28) to hold is requiring that local extrema with positive or negative deﬁnite Hessians are not enhanced: an extremum in ξ at time θ satisﬁes ∂t u > 0 if ξ is a minimum. nonenhancement of local extrema generally does not imply that their number is nonincreasing. Let u be the unique solution of (Pc ) and consider some θ > 0.29) giving immediately the causality requirement (2. (2. However.4.31) Proof: Let D(Jρ (∇uσ )) =: (dij (Jρ (∇uσ ))). it is also satisﬁed by nonlinear diﬀusion scalespaces. ∆u = trace(Hess(u)) = η1 + η2 > 0.2.5 and [342]. η2 of the Hessian Hess(u) are positive. θ) = 0 and ∂xi dij (Jρ (∇uσ (ξ. as we shall see now.32) Since ∇u(ξ. θ) > 0. the eigenvalues η1 . (2. Then. SCALESPACE PROPERTIES 65 2. 3 . Then we have 2 2 2 2 ∂t u = i=1 j=1 ∂xi dij (Jρ (∇uσ )) ∂xj u + i=1 j=1 dij (Jρ (∇uσ )) ∂xi xj u.
Another way to establish causality is via the extremum principle (2. namely a constant image with the same average grey value as the original one. Hence. 2). Lyapunov functionals and behaviour for t → ∞ Since scalespaces are intended to subsequently simplify an image. it is desirable that. This ensures its stability. one proceeds in the same way utilizing the positive deﬁniteness of the Hessian. θ) is a local maximum where H := Hess(u(ξ. see [189] for more details. θ))) is positive deﬁnite. for t → ∞. It should be noted that nonenhancement of local extrema is just one possibility to end up with Koenderink’s causality requirement. λ2 being the positive eigenvalues of D. θ) = trace (DH) = trace (S T DS S T HS) = trace (ΛB) 2 = i=1 λi bii < 0. and by the invariance of the trace with respect to orthogonal transformations it follows that ∂t u(ξ. Now. B := (bij ) := S T HS are negative deﬁnite. The following theorem states that anisotropic diﬀusion ﬁltering always leads to a constant steadystate. thus. If ξ is a local minimum of u(x. there exists an orthogonal matrix S ∈ IR2×2 such that S T DS = diag(λ1 . . 182– 183]. pp. λ2 ) =: Λ with λ1 . Then we have bii < 0 (i = 1. This is due to the class of Lyapunov functionals associated with the diﬀusion process. let us assume that (ξ.15) following Hummel’s reasoning.66 CHAPTER 2. θ)) and. which do violate this principle. Although possibly behaving like backward diﬀusion across edges. nonlinear diﬀusion is always in the forward region at extrema. we obtain the simplest possible image representation. CONTINUOUS DIFFUSION FILTERING We know that the diﬀusion tensor D := D(Jρ (∇uσ (ξ. θ). 2 Nonenhancement of local extrema distinguishes anisotropic diﬀusion from classical contrast enhancing methods such as highfrequency emphasis [163.
V ∈ C[0. µ and M be deﬁned as in (Pc ).33) is a Lyapunov functional: (i) (ii) Φ(u(t)) ≥ Φ(Mf ) for all t ≥ 0. t→∞ lim u(t) − Mf Lp (Ω) =0 for ≤ = Ω r(u(x. (2. Using the average grey level invariance and Jensen’s inequality we obtain. .24).22) and (2. t)) dx (2. b]. b]. respectively. T ] p ∈ [1. t)) dx dy (2. we know that r is convex on [a. t)) dx Ω 1 Ω 1 r Ω Ω Ω u(x. ∞).2. b] with r ′′ ≥ 0 on [a. t) = µ is uniform on Ω. b]. b] with r ′′ ≥ 0 on [a. Φ(Mf ) = Ω t→∞ ¯ In the 1D case.e.34) = Φ(u(t)).e. then V (t) = Φ(u(t)) is a strict Lyapunov functional: (iii) (iv) (v) Φ(u(t)) = Φ(Mf ) ¯ on Ω a. Then the following properties are valid: (a) (Lyapunov functionals) For all r ∈ C2 [a. if r ′′ > 0 on [a.4. on Ω ¯ If t > 0. (b) (Convergence) (i) (ii) Proof: (a) (i) Since r ∈ C2 [a. SCALESPACE PROPERTIES 67 Theorem 3 (Lyapunov functionals and behaviour for t → ∞). Suppose that u is the solution of (Pc ) and let a. the function V (t) := Φ(u(t)) := Ω r(u(x. Moreover. b. the convergence lim u(x. on Ω and ¯ u(t) = Mf on Ω × (0. ⇐⇒ u(t) = Mf u(t) = Mf for T > 0 ⇐⇒ (if t > 0) (if t = 0) V (0) = V (T ) f = Mf a. ∞) and V ′ (t) ≤ 0 for all t > 0. b]. t) dx dy r(u(x. ∞) ∩ C1 (0. for all t ≥ 0. then V ′ (t) = 0 if and only if u(t) = Mf on Ω.
t))−r(f (x)) ≤ L u(x. L2 (Ω)). b]. we know that V is diﬀerentiable for t > 0 and V ′ (t) = Ω r ′ (u) ut dx. t)−f (x) is veriﬁed a. Ω) and α := 1 Ω1  u dx = Ω1 1 Ω2  u dx =: β. the divergence theorem yields V ′ (t) = Ω r ′ (u) div (D(Jρ(∇uσ ))∇u) dx r ′ (u) D(Jρ(∇uσ ))∇u. then u(t) is continuous in Ω.e. T ]. Then. CONTINUOUS DIFFUSION FILTERING (ii) Let us start by proving the continuity of V (t) in 0. Let us now show that equality in ¯ the estimate (2. Ω2  ∈ (0.34) implies that u(t) = const. b] and 1 r Ω Ω u dx = r Ω2  Ω1  α+ β Ω Ω . the limit t → 0+ gives the announced continuity in 0. Thanks to the maximum–minimum principle. the Lipschitz condition r(u(x. ¯ If t > 0. By Theorem 1 and the boundedness of r ′ on [a. To this end. assume ¯ that u is not constant on Ω. From this and the Cauchy–Schwarz inequality. n dS Γ =0 = − Ω r ′′ (u) ∇u.b] such that for all t > 0. (iii) Let Φ(u(t)) = Φ(Mf ). and by virtue of u ∈ C([0. by the continuity of u. D(Jρ(∇uσ ))∇u dx ≥0 ≥0 ≤ 0. we get V (t)−V (0) ≤ Ω1/2 r(u(t))−r(f ) ≤ Ω1/2 L u(t)−f L2(Ω) L2 (Ω) . b] it follows that r is strictly convex on [a. there exists a partition Ω = Ω1 ∪ Ω2 with Ω1 .68 CHAPTER 2. we may choose a constant L := max r ′(s) s∈[a. Thus. on Ω. Ω2 From r ′′ > 0 on [a. on Ω.
D(Jρ(∇uσ (x.) on Ω implies Φ(u(t)) = Φ(Mf ). ∞). for t > 0. t) = const. V ′ (t) = − r ′′ (u) ∇u. t)))∇u(x. D(Jρ(∇uσ ))∇u = 0 on Ω. SCALESPACE PROPERTIES Ω1  Ω1  r(α) + r(β) Ω Ω 1 1 r(u) dx + r(u) dx ≤ Ω Ω < Ω1 Ω2 69 = 1 Ω r(u) dx.e. Ω . we conclude that Φ(u(θ)) = Φ(Mf ) for all θ > 0. on Ω. Then from 0 = V ′ (t) = − r ′′ (u(x. t) dx >0 Ω and the smoothness of u we obtain ¯ ∇u. on Ω.2. the limit θ → 0+ ﬁnally yields u(0) = Mf a. By the uniform boundedness of D. L2 (Ω)). θ) − µ dx ≤ Ω1/2 u(θ) − Mf L2 (Ω) = 0.34) we observe that. for all x ∈ Ω. D(Jρ(∇uσ ))∇u ¯ on Ω × (0. Then. this yields u(x. it is obvious that u(t) = Mf (a. Thus. Thus.e.4. T ]. Thanks to the ¯ average grey value invariance we ﬁnally obtain u(t) = Mf on Ω. let u(x. such that ν ∇u2 ≤ ∇u. on Ω. Conversely. there exists some constant ν > 0. the Cauchy–Schwarz inequality gives u(x. Conversely. t) = 0 a. ¯ Φ(u(t)) = Φ(Mf ) implies that u(t) = const. (iv) Let t > 0 and V ′ (t) = 0. So let us turn to the case t = 0. t) = µ on Ω. and the average grey level invariance ﬁnally gives u(x.e. we have ∇u(x. t) = µ on Ω. From (i) and (ii). t). we have u(θ) = Mf for all θ > 0. t)) ∇u(x. Ω If we utilize this result in the estimate (2. Ω Since u ∈ C([0. D(Jρ(∇uσ ))∇u dx = 0. For θ > 0. Due to the continuity of ∇u.
we know that v(t) f −Mf L∞ (Ω) . T ]. This proves the assertion. we ﬁnd some ν > 0 such that 1d ( v 2 dt 2 L2 (Ω) ) ≤ −ν ∇v 2 L2 (Ω) . Since V is decreasing. CONTINUOUS DIFFUSION FILTERING (v) Suppose that V (T ) = V (0). T ]. and H¨lder’s inequality gives. (b) (i) By the grey level shift invariance we know that v := u−Mf satisﬁes the diﬀusion equation as well. we have V (t) = const.70 CHAPTER 2. there exists some x0 with v(x0 ) = 0. L2 (Ω)). Ω Ω Since ∇vσ is bounded. T ]. q ≥ 2. the limit t → 0+ yields f = Mf a. for 1 ≤ p < q < ∞. For t > 0. we have V ′ (t) = 0. Thus. the Cauchy–Schwarz inequality gives f − Mf  dx ≤ Ω1/2 f − u(t) L2 (Ω) . on [0. Ω As u ∈ C([0. Let ǫ > 0. This yields d v dt 2 L2 (Ω) ≤ −2ν C0 v 2 L2 (Ω) and hence the exponential decay of v L2 (Ω) to 0. By the maximum principle. Poincar´’s e inequality (cf. We multiply this equation by v.e. p. 122]) may be applied giving v 2 L2 (Ω) ≤ C0 ∇v 2 L2 (Ω) with some constant C0 = C0 (Ω) > 0. T ]. for q ∈ IN. [9.) on Ω holds for all t ∈ [0. Conversely. . we get v(t) q Lq (Ω) L∞ (Ω) is bounded by ≤ f −Mf q−2 L∞ (Ω) · v(t) 2 L2 (Ω) → 0. and use the divergence theorem to obtain vvt dx = − ∇v. integrate. Therefore. Now. it is evident that V (0) = V (T ). if u(t) = Mf (a. Then for any t ∈ [ǫ. on Ω. and part (iv) implies that u(t) = Mf on Ω.e. D(Jρ(∇vσ ))∇v dx. o v(t) Lp (Ω) ≤ Ω(1/p)−(1/q) · v(t) Lq (Ω) → 0.
∞) ∩ C1 (0. 9–12]. p. a V (t) = −2 ′ u2 D(Jρ (∂x uσ )) dx x 0 a ≤ −2ν ≤ 0. From part (a) we know that V (t) := a 0 u2 (x. Hence. Let Ω = (0. 144] we know that the embedding from H1 (0. for t > 0. we can generalize and adapt methods from [202] to our case. there exists a subsequence (tij ) → ∞ and some u with ¯ u(tij ) → u in C0.36) L2 (Ω) → 0. pp. (2. a]. Equations (2.2.37) Part (a) tells us that u(t)−Mf p p (Ω) is a Lyapunov function for p ≥ 2.35) and (2. the space of H¨ldercontinuous functions on [0. Hence. ∞) the mean value theorem implies ∃ ti ∈ (i. (2. Thus.36) yield ux (ti ) u2 dx x 0 (2. ¯ This also gives u(tij ) → u in L2 (0. a]. a) into C0. a] o 1 [7. t) dx is nonincreasing and bounded from below. Since we already know from (b)(i) ¯ 2 that u(tij ) → Mf in L (0. it follows that u = Mf . the sequence (V (i))i∈IN converges. i+1) : V ′ (ti ) = V (i + 1) − V (i). L Thus. Therefore. a). a). Thus. By virtue of the Rellich– Kondrachov theorem [7. u(ti ) is a bounded sequence in H1 (0. ¯ j→∞ lim u(tij ) − Mf L∞ (Ω) = 0. a).α [0. SCALESPACE PROPERTIES 71 (ii) To prove uniform convergence in the onedimensional setting.α [0.4. 2 ). is compact for α ∈ (0.35) Thanks to the uniform positive deﬁniteness of D there exists some ν > 0 such that. (ti )i∈IN → ∞ and from the convergence of (V (i))i∈IN it follows that V ′ (ti ) → 0. a). u(t)−Mf L∞ (Ω) = lim u(t)−Mf Lp (Ω) p→∞ . Since V ∈ C[0.
r(s) := (s−µ)2n and r(s) := s ln s. Consider the intensity in an image f as a random variable Zf with distribution .37) we conclude that t→∞ L∞ (Ω) exists and from lim u(t)−Mf L∞ (Ω) = 0. lim u(t)−Mf t→∞ (2. t) = µ 2 ¯ is uniform on Ω. Moreover.72 CHAPTER 2. the competition evolves in a certain direction all the time: although backward diﬀusion may be locally superior. t)) dx. t) − µ)2n dx for all n ∈ IN. As a special case of (a) it follows that the energy u(t) 2 2 (Ω) is reduced by L diﬀusion. Corollary 1 (Special Lyapunov functionals). Since the class (Pc ) does not forbid contrast enhancement it admits processes where forward diﬀusion has to compete with backward diﬀusion. Corollary 1 oﬀers multiple possibilities of how to interpret nonlinear anisotropic diﬀusion ﬁltering as a smoothing transformation.22). Part (b) gives a probabilistic interpretation of anisotropic diﬀusion ﬁltering. Let us have a closer look at what might be the meaning of this global result in the context of image processing. Then the following functions are decreasing for t ∈ [0. 1 Ω Ω M2n [u(t)] := (u(x. CONTINUOUS DIFFUSION FILTERING is also nonincreasing. the preceding theorem gives the following corollary. Let u be the solution of (Pc ) and a and µ be deﬁned as in (Pc ) and (2. the global result – denoted by the Lyapunov functional – becomes permanently better for forward diﬀusion. Theorem 3 is of importance as it states that the regularization by convolving with Kσ tames the backward diﬀusion in such a way that forward diﬀusion wins in the long run. Therefore. The smoothness of u establishes ﬁnally that the convergence t→∞ lim u(x. if a > 0. t) ln(u(x. (c) H[u(t)] := Ω u(x. ∞): (a) (b) u(t) Lp (Ω) for all p ≥ 2. Considering the Lyapunov functions associated with r(s) := sp . respectively.
anisotropic diﬀusion does really simplify the original image in a steady way. In a classic scalespace representation. t) ln(u(x. If we choose another probabilistic model of images.2. (2. Ff (z) is the probability that an arbitrary grey value Zf of f does not exceed z. µ is equal to the expected value EZu(t) := IR z dFu(t) (z). S[u(t)] := − u(x. The interpretation of the entropy in terms of Lyapunov functionals carries also over to generalized entropies. By increasing t. Hence. local eﬀects such as edge enhancement. It is a common tool for constructing measures for the relative smoothness of the intensity distribution. (2. From all the previous considerations. In particular. in spite of possible contrastenhancing properties. the process reaches the state with the lowest possible information.4. i. Since anisotropic diﬀusion ﬁlters increase the entropy the corresponding scalespace embeds the genuine image f into a family of subsequently likelier versions of it which contain less information. t)) dx Ω (2.4 Then. a measure of uncertainty and missing information [63]. the time t plays the role of the scale parameter.40) is called the entropy of u(t). SCALESPACE PROPERTIES 73 Ff (z). we recognize that. Higher moments are more diﬃcult to interpret. Moreover.e. are overcompensated by smoothing in other areas. which object to increase central moments. Let us ﬁnally point out another interpretation of the Lyapunov functionals. By the average grey level invariance.38) and it follows that M2n [u(t)] is just the even central moment z−EZu(t) IR 2n dFu(t) (z). we may regard it also as a twodimensional density. All decreasing even moments demonstrate that the image becomes smoother during diﬀusion ﬁltering. see [390] for more details. Provided the initial image f is strictly positive on Ω. for t → ∞. it helps to explain the success of nonlinear diﬀusion ﬁltering as a preprocessing step for subsampling as observed in [144]. 414–415]. namely a constant image. although they do provide important information for tasks like texture discrimination [163.39) The second central moment (the variance) characterizes the spread of the intensity about its mean. The fourth moment is frequently used to describe the relative ﬂatness of the grey value distribution. This informationreducing property indicates that anisotropic diﬀusion might be generally useful in the context of image compression. . one transforms the image from a local to a more global 4 Without loss of generality we omit the normalization. then part (c) characterizes the informationtheoretical side of our scalespace. pp.
74 CHAPTER 2. This can be achieved by taking some Lyapunov function Φ(u(t)) and investigating the expression Ψ(u(t)) := Φ(f ) − Φ(u(t)) . We have seen in Chapter 1 that. CONTINUOUS DIFFUSION FILTERING representation. it is possible to associate with the evolution time a corresponding spatial scale. Φ(f ) − Φ(Mf ) (2. Experiments in this direction can be found in [431. one can only deﬁne an average measure for the globality of the representation. In the nonlinear diﬀusion case. 308]. however. for linear diﬀusion scalespaces and morphological scalespaces. . Since the smoothing is nonuniform. the situation is more complicated. Prescribing a certain value for Ψ provides us with an aposteriori criterion for the stopping time of the nonlinear diﬀusion process.41) We observe that Ψ(t) increases from 0 to 1. It gives the average globality of u(t) and its value can be used to measure the distance of u(t) from the initial state f and the ﬁnal state Mf .
We shall investigate conditions under which one can establish similar properties as in the continuous setting concerning wellposedness. N ≥ 2.r−1. In order to specify the requirements for our semidiscrete ﬁlter class we ﬁrst recall a useful deﬁnition of irreducible matrices [407.. j ∈ J there exist k0 .. pp...N represent the grey values at each pixel.. Afterwards we shall discuss whether it is possible to obtain such ﬁlters from spatial discretizations of the continuous models that have been investigated in Chapter 2. j = 1.. 3.. N} by J.kr ∈ J with k0 = i and kr = j such that akp kp+1 = 0 for p = 0. and convergence to a constant steadystate.Chapter 3 Semidiscrete diﬀusion ﬁltering The goal of this chapter is to study a semidiscrete framework for diﬀusion scalespaces where the image is sampled on a ﬁnite grid and the scale parameter is continuous. We denote the index set {1. extremum principles. whose components fj . 18–20]. .... Lyapunov functions.. A matrix A = (aij ) ∈ IRN ×N is called irreducible if for any i.. average grey level invariance.1 The general model A discrete image can be regarded as a vector f ∈ IRN . The semidiscrete problem class (Ps ) we are concerned with is deﬁned in the following way: 75 .. We will see that there exists a ﬁnite stencil on which a diﬀerence approximation of the spatial derivatives are in accordance with the semidiscrete scalespace framework. Deﬁnition 1 (Irreducibility)... This leads to a system of nonlinear ordinary diﬀerential equations (ODEs).
. Find a function u ∈ C1 ([0. 3. j∈J ∀ u ∈ IR . where A = (aij ) has the following properties: (S2) symmetry: (S1) Lipschitzcontinuity of A ∈ C(IRN . IRN ×N ) for every bounded (Ps ) subset of IRN .76 CHAPTER 3. aij (u) ≥ 0 ∀ i = j. The existence of Lyapunov functions can be established by means of (S2)–(S4). while average grey value invariance uses (S2) and (S3). IRN ) which satisﬁes an initial value problem of type du = A(u) u. (S2) and (S3) correspond to the speciﬁc structure of the divergence expression with a symmetric diﬀusion tensor D. ∀ u ∈ IRN .2 Theoretical results Before we can establish scalespace results. it is of importance to ensure the existence of a unique solution. respectively. while (S4) and (S5) play a similar role as the nonnegativity of the eigenvalues of D and its uniform positive deﬁniteness. and strict Lyapunov functions and the convergence to a constant steadystate require (S5) in addition to (S2)–(S4). SEMIDISCRETE DIFFUSION FILTERING Let f ∈ IRN . (S1) is needed for wellposedness. (S5) irreducibility for all u ∈ IRN . This is done in the theorem below which also states the continuous dependence of the solution and a maximum–minimum principle. aij (u) = aji (u) ∀ i. This indicates that these properties reveal some interesting parallels to the continuous setting from Chapter 2: In both cases we need smoothness assumptions to ensure wellposedness. ∀ u ∈ IRN . dt u(0) = f. ∞). the proof of a maximum–minimum principle involves (S3) and (S4). Not all of these requirements are necessary for every theoretical result below. j ∈ J. N (S3) vanishing row sums: (S4) nonnegative oﬀdiagonals: aij (u) = 0 ∀ i ∈ J.
u) ∈ B0 . θ]. (b) Maximum–minimum principle We prove only the maximum principle. (3. j∈J ∀ i ∈ J. Assume that the problem (Ps ) has a unique solution on [0. u) t ∈ [t0 . In order to prove existence and uniqueness of a solution of (Ps ) in R0 := (t. extremum principle). u−f c ∞ ≤ β ⊂ B0 we have to show that φ(t. First we show that the derivative of the largest component of u(t) is nonpositive for every .3) b := max fj . this follows directly from the fact that A is Lipschitzcontinuous on {u ∈ IRN  u−f ∞ ≤ β}. T ] × u ∈ IRN u ∞ ≤ f ∞ +β . u) := ψ(u) := A(u) u is continuous on B0 := [0. u) ∞ ≤c ∀ (t. T ].1) (3.2.2) (3. This solution depends continuously on the initial value and the righthand side of the ODE system. u) satisﬁes a global Lipschitz condition on R0 with respect to u. Moreover. Evidently. For every T > 0 the problem (Ps ) has a unique solution u(t) ∈ C1 ([0. and it satisﬁes the extremum principle a ≤ ui (t) ≤ b where a := min fj . ∀ t ∈ [0. o Let t0 := 0 and β > 0. THEORETICAL RESULTS 77 Theorem 4 (Wellposedness. However. j∈J Proof: (a) Local existence and uniqueness Local existence and uniqueness are proved by showing that our problem satisﬁes the requirements of the Picard–Lindel¨f theorem [432. by the compactness of B0 there exists some c > 0 with φ(t. t0 +min( β . 59]. since it is a composition of continuous functions.3. T ]. since the proof for the minimum principle is analogous. T )]. φ(t. IRN ). p.
5) and the smoothness of u we may ﬁnd a ϑp ∈ (0. let us consider some p ∈ P . SEMIDISCRETE DIFFUSION FILTERING t ∈ [0. for t = ϑ. uε (t) := u(t) − . t2 ).6) Next. Due to (3. by (3. (3. t1 ). let P := {p ∈ J  uεp (0) = max uεj (0)}. j∈J and the continuity of u there exists some t1 ∈ (0. ϑp ). (3. akj (u) j∈J = (3. ϑp ) max uεp(t) < max uεj (0) p∈P j∈J ∀ t ∈ (0. εt Moreover. θ) such that i∈J\P max uεi(t) < max uεj (0) j∈J ∀ t ∈ [0. . Let ε > 0 and set εt . we have uεp (t) < uεp (0) and.7) . [201]). Then. ∀ t ∈ (0. θ]. duk dt = j∈J akj (u) uj akk (u) uk + j∈J\{k} = ≤ (S4) akj (u) uj ≥0 ≤uk uk · 0. If we keep j∈J this k ﬁxed we obtain.5) max uεi (0) < max uεj (0). Let uk (ϑ) := max uj (ϑ) for some arbitrary ϑ ∈ [0. . (3. it follows that p∈P ∀ t ∈ [0.4) Let us now prove that this implies a maximum principle (cf.78 CHAPTER 3. for t2 := min ϑp .4). θ]. θ) with duεp (t) < 0 dt Thus. j∈J duεp (0) = dt By means of i∈J\P dup (0) − ε < 0 dt ≤0 ∀ p ∈ P.
c . Together with u = lim uε and the continuity of u this yields the announced ε→0 maximum principle max uj (t) ≤ max uj (0) j∈J j∈J ∀ t ∈ [0. t3 ) with duεk uεk (t3 ) − uεk (t4 ) (t5 ) = dt t3 − t4 (3. t2 ).3.8) Now we prove that this estimate can be extended to the case t ∈ (0. we know that the problem (Ps ) has a unique solution u(t) for t ∈ [t0 . implies that we ﬁnd a t5 ∈ (t4 . Using the notations and results from (a). (3. dt ≤0 Due to the continuity of there exists some t4 ∈ (0. t0 ). (c) Global existence and uniqueness Global existence and uniqueness follow from local existence and uniqueness when being combined with the extremum principle. θ) such that max uεj (t3 ) = max uεj (0). however.10) duεk (t) < 0 dt The mean value theorem.8) must be valid on the entire interval (0.4) gives duεk (t3 ) = dt du dt ∀ t ∈ (0. j∈J j∈J Let uεk := max uεj (t3 ).9) > 0. for t0 := min(t1 . assume the opposite is true. Then the minimality of t3 yields j∈J uεk (t) < uεk (t3 ) and inequality (3. (3. t3 ). THEORETICAL RESULTS Hence.10). t3 ]. the estimates (3. by virtue of the intermediate value theorem.7) give max uεj (t) < max uεj (0) j∈J j∈J 79 ∀ t ∈ (0. θ). Then. (3. t0 + min( β . θ].9) duk (t3 ) − ε < 0. θ). which contradicts (3. Hence.2. (3. T )].6) and (3. To this end. there exists some t3 which is the smallest time in (0. t3 ) with ∀ t ∈ (t4 .
dt ˜ u(0) = f ˜ . (d) Continuous dependence Let u(t) be the solution of du = φ(t. u) ∈ B1 . the existence of a unique solution can be extended to the entire interval [0. u) t ∈ [t1 . min( 2β . T ). p. u) is continuous. Using the same considerations as in (a) one shows that φ is Lipschitzcontinuous on R1 := (t. t1 +min( β . the considered problem has a unique solution on [t1 . Hence. with respect to its second argument. t) = ψ(u) = A(u)u. T )]. c Therefore. T ] × u ∈ IRN u ∞ ≤ g ∞ +β . and by the extremum principle we know that B1 ⊂ B0 . φ(t. 93] ensure that for every ε > 0 there exists a δ > 0 such that the solution u of ˜ the perturbed problem d˜ u ˜ ˜ = φ(t. and that there exists some α > 0 such that φ(t. u−g c ∞ ≤β . φ(t. SEMIDISCRETE DIFFUSION FILTERING Now let t1 := t0 + min( β . v) t ∈ [0. dt u(0) = f for t ∈ [0. t1 + min( β . g := u(t1 ). it is suﬃcient to prove that φ(t. T ] as well. with the same c as in (a). (Ps ) reveals a unique solution on [0. In order to show that u(t) depends continuously on the initial data and the righthand side of the ODE system. u) ∞ ≤c ∀ (t. T ] and φ(u. T ]. Hence. As a consequence. u).80 CHAPTER 3. u) satisﬁes a global Lipschitz condition on Sα := (t. T )]. T )]. u). T ]. u) = A(u)u is continuous on B1 := [0. v−u ∞ ≤α . In this case the results in [412. by iterating c this reasoning. Clearly. dt u(t1 ) = g. the extremum principle is valid on [0. and. and consider the problem c du = A(u) u.
3 Scalespace properties It is evident that properties such as grey level shift invariance or reverse contrast invariance are automatically satisﬁed by every consistent semidiscrete approximation of the continuous ﬁlter class (Pc ). (3. v) ˜ f −f ∞ ∞ 81 < δ. v) − φ(t. Thus. the global Lipschitz condition on Sα follows direcly from the fact that A is Lipschitzcontinuous on {v ∈ IRN  v−u ∞ ≤ α}. < δ for v−u ∞ < α exists in [0. for t ≥ 0. On the other hand. ∞) and concludes the proof. translation invariance only makes sense for translations in grid direction with multiples of the grid size.11) ∀ t ≥ 0. . Proposition 2 (Conservation of average grey value).3. and isometry invariance is satisﬁed for consistent schemes up to an discretization error. k∈J j∈J j∈J duj = dt j∈J ajk (u) uk = j∈J k∈J which shows that uj (t) is constant on [0. So let us focus on average grey level invariance now.12) Proof: By virtue of (S2) and (S3) we have j∈J ajk (u) = 0 for all k ∈ J.3. 2 3. SCALESPACE PROPERTIES ˜ with continuous φ and ˜ φ(t. ajk (u) uk = 0. T ] and satisﬁes the inequality u(t) − u(t) ˜ ∞ < ε. Similar to the the local existence and uniqueness proof. 2 This property is in complete accordance with the result for the continuous ﬁlter class. The average grey level 1 µ := fj N j∈J is not aﬀected by the semidiscrete diﬀusion ﬁlter: 1 N uj (t) = µ j∈J (3.
and (3.3).82 CHAPTER 3. then V (t) = Φ(u(t)) is a strict Lyapunov function: (iii) (iv) Φ(u) = Φ(c) V ′ (t) = 0 ⇐⇒ u=c u=c ⇐⇒ (b) (Convergence) lim u(t) = c. . let a... Let u(t) be the solution of (Ps ). SEMIDISCRETE DIFFUSION FILTERING Similar to the continuous setting. Theorem 5 (Lyapunov functions and behaviour for t → ∞). Then the following properties are valid: (a) (Lyapunov functions) For all r ∈ C1 [a. Average grey level invariance and this convexity yield. t→∞ Proof: (a) (i) Since r ′ is increasing on [a.. µ.11).13) = Φ(u). b] with increasing r ′ on [a. b]. µ)⊤ ∈ IRN . for all t ≥ 0. Moreover. .2). 1 r Φ(c) = uj i=1 j=1 N 1 ≤ N i=1 N N N N N j=1 r(uj ) = j=1 r(uj ) (3. V ∈ C1 [0. b]. (3. and µ be deﬁned as in (3. the function V (t) := Φ(u(t)) := i∈J r(ui (t)) is a Lyapunov function: (i) (ii) Φ(u) ≥ Φ(c) for all t ≥ 0. b] we know that r is convex on [a. if r ′ is strictly increasing on [a. ∞) and V ′ (t) ≤ 0 for all t ≥ 0. b]. it is possible to ﬁnd a large class of Lyapunov functions which establish smoothing scalespace properties and ensure that the image tends to a constant steadystate with the same average grey level as the initial image. b. and let c := (µ. respectively.
η := j=1 j=j0 uj 1 . SCALESPACE PROPERTIES 83 (ii) Since u ∈ C1 ([0. b]. 1− N 1 N Then.14) = Since r ′ is increasing.i+k (u) (ui+k −ui) r ′ (ui ) ai+k. we get r 1 uj i=1 N N = r 1 N 1 ≤ N < N 1 η N 1 r(η) r(uj0 ) + 1− N N 1 r(uj0 ) + r(uj ) j=1 N 1 uj + N 0 1− j=j0 = 1 r(uj ). we know that r is strictly convex. it follows that V ∈ C1 [0. we always have (ui+k − ui ) r ′ (ui ) − r ′ (ui+k ) ≤ 0.14) implies that V ′ (t) ≤ 0 for t ≥ 0.3. Hence. IRN ) and r ∈ C1 [a.i+k (u) (ui+k −ui) r ′ (ui)−r ′ (ui+k ) .3. ∞). To this end. Using the prerequisites (S2) and (S3) we get V ′ (t) N = (S3) i=1 N N dui ′ r (ui ) dt aij (u) (uj −ui) r ′ (ui) N i−1 = i=1 j=1 N = i=1 + j=i+1 j=1 = + (S2) N N −i i=1 k=1 N N −i i=1 k=1 N N −i i=1 k=1 aij (u) (uj −ui) r ′ (ui ) ai. suppose that ui0 := min ui < max uj =: uj0 and let i j N With this and (S4).i (u) (ui −ui+k ) r ′ (ui+k ) ai. (3. (iii) Let us ﬁrst prove that equality in the estimate (3.13) implies that all components of u are equal. b]. Since r ′ is strictly increasing on [a. ∞). j=1 N . η < uj0 . equation (3.
t→∞ Now assume that η > 0. Then from the representation (3. . Since u(t)−c is bounded from above by α := f −c we have u(t)−c ≤ α ∀ t ≥ 0.84 CHAPTER 3. ≤0 and by virtue of the symmetry of A(u) it follows that aij (u) (uj −ui ) r ′ (ui)−r ′ (uj ) = 0 ∀ i. As r ′ is strictly increasing we have.... for p = 0. (3.i+k (u) (ui+k −ui ) r ′ (ui)−r ′ (ui+k ) .. x−c < β. Consider the Lyapunov function V (t) := Φ(u(t)) := u(t)−c2 . Since i0 and j0 are arbitrary. which results from the choice r(s) := (s−µ)2 .. Conversely. j0 ∈ J. and the average grey level invariance gives u = c.. Φ(x) < η ∀ x ∈ IRN ..13) implies that u1 = . . p = 0. Conversely. j ∈ J.. for all i ∈ J. (ukp −ukp+1 ) r ′ (ukp+1 )−r ′(ukp ) From this and (3. (b) The convergence proof is based on classical Lyapunov reasonings..kr ∈ J with k0 = i0 . we obtain ui = const. we know that lim V (t) =: η exists and η ≥ 0.. = 0 ⇐⇒ ukp = ukp+1 . SEMIDISCRETE DIFFUSION FILTERING This shows that equality in (3..14) we have 0 = V ′ (t) = N N −i i=1 k=1 ai.15) Now consider two arbitrary i0 . kr = j0 .15) we get ui0 = uk0 = uk1 = . By virtue of the grey level shift invariance we conclude that u = c.. From (3. = ukr = uj0 . The irreducibility of A(u) implies that there exist k0 . [180] for an introduction to these techniques. for all i ∈ J.g.. √ η). for β ∈ (0.16) By virtue of Φ(x) = x − c2 we know that. it is trivial that Φ(u) = Φ(c) for u = c. see e. and akp kp+1 (u) = 0. (3. = uN . Since V (t) is decreasing and bounded from below by 0. let ui = const. This proves the ﬁrst implication. (iv) Let V ′ (t) = 0..r − 1. r − 1..14) we immediately conclude that V ′ (t) = 0.
∞): (a) (b) (c) u(t) p for all p ≥ 2.17) we have u(t) ∈ {x ∈ IRN  β ≤ x−c ≤ α} =: S ∀ t ≥ 0. SCALESPACE PROPERTIES Let w. .3. β < α.18) is increasing with respect to t. According to (a)(iii) this yields lim u(t) = c. ∀ t ≥ 0. the compactness of S. Hence the assumption η > ρ is wrong and we must have η = ρ. 85 (3. it follows V (t) = V (0) + t→∞ t 0 V ′ (θ) dθ ≤ V (0) − tM which implies lim V (t) = −∞ and.17) By (a)(ii).11). and β > 0 there exists some M > 0 such that V ′ (t) ≤ −M ∀ t ≥ 0. Since all pnorms (p ≥ 2) and all central moments are decreasing. j=1 uj (t) ln(uj (t)).l. Therefore. we observe that the semidiscrete setting reveals smoothing scalespace properties which are closely related to the continuous case. contradicts (a)(i). and obtain the following corollary. 1 N N j=1 M2n [u(t)] := H[u(t)] := N (uj (t) − µ)2n for all n ∈ IN.16) and (3.2) and (3.g. Let u be the solution of (Ps ) and a and µ be deﬁned as in (3. Then the following functions are decreasing for t ∈ [0. r(s) := (s−µ)2n and r(s) := s ln s. while the discrete entropy N S[u(t)] := − uj (t) ln(uj (t)) j=1 (3.3. thus. t→∞ 2 As in the continuous case. respectively. we can consider the Lyapunov functions associated with r(s) := sp . if a > 0. Corollary 2 (Special Lyapunov functions).o.(iv). Since Φ(u(t)) ≥ η we conclude that u(t)−c ≥ β So from (3.
2 and the grid sizes h1 . Then a consistent spatial discretization of the isotropic diﬀusion equation with homogeneous Neumann boundary conditions can be written as 2 duk gl + gk (ul − uk ). respectively. j). First we shall verify that this is easily done for isotropic models. a2 ) be discretized by a grid of N = n1 · n2 pixels such that a pixel (i. A constructive existence proof is presented showing that for a suﬃciently large stencil it is always possible to ﬁnd such a nonnegative discretization. In the anisotropic case. IR2×2 ). we may replace the diﬀusion tensor D(Jρ (∇uσ )) by some scalarvalued function g(Jρ (∇uσ )).. To this end. For this reason. These pixels can be numbered by means of an arbitrary bijection p: {1. j). .. Let us now verify that a standard FD space discretization of an isotropic variant of (Pc ) leads to a semidiscrete ﬁlter satisfying the requirements (S1)–(S5).20) 1 ) h2 . n1 } × {1. j) with 1 ≤ i ≤ n1 and 1 ≤ j ≤ n2 represents the location (xi .4.. 3. N}.. (3. . This concept is illustrated by investigating the situation on a (3×3)stencil in detail. n2 } → {1... respectively. dt (3..1 Isotropic case Let the rectangle Ω = (0.. Next. yj ) where xi := (i − 1 ) h1 . the mixed derivative terms make it more diﬃcult to ensure nonnegative oﬀdiagonal elements. The structure tensor requires the calculations of convolutions with ∇Kσ and Kρ . h2 are given by h1 := a1 /n1 and h2 := a2 /n2 . however. j) is represented by a single index p(i.21) Thus..86 CHAPTER 3.19) (3. (3. a1 ) × (0. .22) = 2 dt n=1 l∈Nn (k) 2hl where Nn (k) consists of the one or two neighbours of pixel k along the nth coordinate axis (boundary pixels have only one neighbour) and gk := g ((H(u))k ).23) .22) becomes du = A(u) u.4 Relation to continuous models In this section we investigate whether it is possible to use spatial discretizations of the continuous ﬁlter class (Pc ) in order to construct semidiscrete diﬀusion models satisfying (S1)–(S5). consider some pixel k = p(i. In the spatially discrete case this comes down to speciﬁc vector–matrix multiplications. In vector–matrix notation (3. pixel (i. we may approximate the structure tensor by some matrix H(u) = (hij (u)) where H ∈ C∞ (IRN . 2 yj := (j − (3. SEMIDISCRETE DIFFUSION FILTERING 3.
(3. (b) The ﬁlter class (Pc ) is not the only family which leads to semidiscrete ﬁlters satisfying (S1)–(S5)..24) and the symmetry of the neighbourhood relation: l ∈ Nn (k) ⇐⇒ k ∈ Nn (l). Remarks: (a) We observe that (S1)–(S5) are properties which are valid for all arbitrary pixel numberings.r−1. thus.. This proves (S5)..24) that akk < 0. (l = k).r−1.. (S4) holds. This is due to the fact that the extremum principle limits the modulus of discrete gradient approximations. such that kq and kq+1 are neighbours for q = 0. Then. By the construction of A it is also evident that all row sums vanish.3.. For k = l. 1. thus.. gkq + gkq+1 > 0 2h2 n . we have A ∈ C∞ (IRN .e. IR2×2 ) and g ∈ C∞ (IR2×2 ). (S3) is satisﬁed.. The symmetry of A follows directly from (3. a semidiscrete version of the Perona– Malik ﬁlter – which is to a certain degree illposed in the continuous setting (cf.kr . RELATION TO CONTINUOUS MODELS where the matrix A(u) = (akl (u))kl is given by gk +gl 2h2 n 2 n=1 l∈Nn (k) gk +gl 2h2 n 87 (l ∈ Nn (k)). Then we have to ﬁnd k0 .24) akl := − 0 Let us now verify that (S1)–(S5) are fulﬁlled. we already know from (3..3. 2}. They study an image evolution under an ODE system with a discontinuous right hand side. it follows that akl ≥ 0 for all k = l and. In order to show that A is irreducible.. the spatial discretization implicitly causes a regularization. These results are also in accordance with a recent paper by Pollak et al.. akq kq+1 = for some n ∈ {1. i.. we may choose any arbitrary path k0 .1) – also satisﬁes (S1)–(S5) and.4. Hence..kr ∈ J with k0 = k and kr = l such that akq kq+1 = 0 for q = 0. reveals all the discussed wellposedness and scalespace properties [425]. since g is positive. which has some interesting relations to the limit case of a semidiscrete Perona–Malik model. let us consider two arbitrary pixels k and l. If k = l. [332]. Interestingly.... Moreover.. Since H ∈ C∞ (IRN . (else). They also report stable behaviour of their process. In this case we have the trivial path k = k0 = kr = l. This proves (S1). IRN ×N ).
−1}). the nonnegativity of all oﬀdiagonal elements of A(u). Now we show that for 2 2 a suitable m there exists a stencil direction βk . ... π ] into 4m−2 subin2 2 tervals Ii .2 Anisotropic case If one wishes to transfer the results from the isotropic case to the general anisotropic setting the main diﬃculty arises from the fact that... Theorem 6 (Existence of a nonnegative discretization).. where ψ ∈ (− π . βi +βi+1 > π ). i ∈ Jm : (− π . Now let Jm := {1. 2 (i = 2m−1)... (−2m + 1 ≤ i < −m). and It is not hard to verify that βi ∈ Ii for i ∈ Jm . 2m−1} and deﬁne a partition of (− π . The “boundary pixels” of this stencil deﬁne 4m principal orientations βi ∈ (− π . 2m according to βi := arctan arccot arccot ih2 mh1 (2m−i)h1 mh2 (i−2m)h1 mh2 (i ≤ m)..4.88 CHAPTER 3. 2 2 i = −2m+1. π ] = 2 2 where −1 i=−2m+1 2m−1 (θi . (m < i ≤ 2m). it is not obvious how to ensure (S4). . i=1 =:Ii 0 1 2 π 4 π 2 π 2 arctan θi := 2 cot βi −tan βi+1 2 cot βi −tan βi+1 + 1 arctan 2 (i = 0).. 2m−2}. due to the mixed derivative expressions. Proof: Let us consider some m ∈ IN and the corresponding (2m+1)×(2m+1)stencil.. . cos ψ)⊤ . 2 (i ∈ {1. ... . (i ∈ {1. k ∈ Jm such that the splitting div (D ∇u) = ∂eβ0 α0 ∂eβ0u + ∂eβk αk ∂eβku + ∂eβ2m α2m ∂eβ2mu (3. Then there exists some m(κ) ∈ IN such that div (D ∇u) reveals a secondorder nonnegative FD discretization on a (2m+1)×(2m+1)stencil. 2m−2}. βi +βi+1 < π ). SEMIDISCRETE DIFFUSION FILTERING 3. π ].. θi ].. π ].. sin ψ)⊤ and (− sin ψ.... βi +βi+1 = π ). Let D ∈ IR2×2 be symmetric positive deﬁnite with a spectral condition number κ. 2 (i ∈ {1. The theorem below states that this is always possible for a suﬃciently large stencil. Let λ1 ≥ λ2 > 0 be the eigenvalues of D with corresponding eigenvectors (cos ψ.25) θi := −θ−i (i ∈ {−2m+1. θi+1 ] ∪ =:Ii (θi−1 . 2m−2}. . .
So let us now verify (a)–(c).m (3. c − b tan βk ≥ 0.25) arises in a natural way. γ1 := αk . ϕ1 := βk where ψ ∈ Ik .3. m→∞ i∈Jm = ∞. This can be done by proving the following properties: (a) Let ψ ∈ Ik and D = a b . (a) In order to use subsequent indices. let ϕ0 := 0.27) (k ∈ {1. Once these assertions are proved a nonnegative secondorder discretization of (3.. θk−1 (k = 1). 2m−2}).26) (b) Inequality (3.. (k ∈ {2. ϕ1 and ϕ2 we obtain the linear system div 2 2 i=0 γi cos2 ϕi i=0 γi sin ϕi cos ϕi γi sin ϕi cos ϕi i=0 2 i=0 2 γi sin2 ϕi ∇u . Then (3.. and γ2 := α2m . β2m . αk . sin βi )⊤ reveals nonnegative “directional diﬀusivities” α0 . . 2m−1}). Furthermore. (3.25) 2 requires that div = + ∂ ∂x ∂ ∂y a b b c 2 ∇u = ∂ i=0 ∂eϕi 2 γi ∂u ∂eϕi cos ϕi γi (ux cos ϕi + uy sin ϕi ) i=0 2 sin ϕi γi(ux cos ϕi + uy sin ϕi ) i=0 = By comparing the coeﬃcients and using the deﬁnition of ϕ0 .25) is bc possible if min a − b cot βk .m θk 1 (θ + βk ) 2 k 1 β 2 k =: κk..26) is satisﬁed for λ1 ≤ min cot(ρk −βk ) tan ρk . (k = 2m−1).. βk . cot(βk −ηk ) cot ηk λ2 with ρk := ηk := (c) lim min κi. α2m along the stencil orientations β0 . as we shall see at the end of this chapter. . and ϕ2 := π . RELATION TO CONTINUOUS MODELS 89 with eβi := (cos βi . a γ0 1 cos2 βk 0 0 sin βk cos βk 0 γ1 = b c γ2 0 sin2 βk 1 ..4. Then a nonnegative splitting of type (3. let γ0 := α0 .
29) (3. γ1 = sin βk cos βk γ2 = c − b tan βk . Now. (3. π ].90 CHAPTER 3. B(ϕ) and the continuity of B(ϕ) and C(ϕ) we may extend (3. (3. θk ). λ1 − λ2 ≥ 0.28) (3. (b) Let λ1 λ2 ≤ κk. λ1 cos2 ψ + λ2 sin2 ψ (λ1 −λ2 ) sin ψ cos ψ (λ1 −λ2 ) sin ψ cos ψ λ1 sin2 ψ + λ2 cos2 ψ . c − b tan βk ≥ 0. and since ψ. βk ]. In particular.31) ϕ∈(βk . C(ϕ) := sin2 ϕ + sin ϕ cos ϕ cot βk we get λ1 C(ρk ) ≤ cot(ρk −βk ) tan ρk = − = λ2 B(ρk ) Since B(ϕ) < 0 on (βk .30) From the structure of the eigenvalues and eigenvectors of D it is easily seen that b = (λ1 −λ2 ) sin ψ cos ψ.m and consider the case 0 < βk < π . 2 ∀ ϕ ∈ [− π . since ψ ∈ Ik . b . γ1 is always nonnegative. π ) we have 2 λ1 B(ϕ) + λ2 C(ϕ) ≥ 0 Because of B(ϕ) ≥ 0 C(ϕ) ≥ 0 ∀ ϕ ∈ [0. 2 ∀ ϕ ∈ (βk . we have 0 ≤ λ1 B(ψ) + λ2 C(ψ) By the representation a b b c = = cos ψ − sin ψ sin ψ cos ψ λ1 0 0 λ2 cos ψ sin ψ − sin ψ cos ψ = (λ1 cos2 ψ + λ2 sin2 ψ) − (λ1 −λ2 ) sin ψ cos ψ cot βk . By deﬁning 2 B(ϕ) := cos2 ϕ − sin ϕ cos ϕ cot βk . θk ].31) to the entire interval Ik = (θk−1 . In order to satisfy the nonnegativity of γ0 and γ2 we need that min a − b cot βk . SEMIDISCRETE DIFFUSION FILTERING which has the unique solution γ0 = a − b cot βk . Thus.θk ) min − C(ϕ) . βk ∈ Ik we conclude that ψ and βk belong to the same quadrant.
ρi = θi . 1 − gm (i) the function gm (x) is bounded and attains its global maximum xm := − 1 + 6 1 6 1 + 12 m2 h2 1 . yields cot ρi = tan ρi 1 cot βi − tan βi+1 + (cot βi − tan βi+1 )2 + 4 . 2 1 = − cot βi + tan βi+1 + (cot βi − tan βi+1 )2 + 4 . and the deﬁnition of θi implies that cot ρi − tan ρi = cot βi − tan βi+1 . h2 2 . −1 Let us now assume that 1 ≤ i ≤ m−1. respectively. mh1 mh1 = . RELATION TO CONTINUOUS MODELS we recognize that this is just the desired condition a − b cot βk ≥ 0. Solving for cot ρi and tan ρi . ih2 1 4m2 i m2 h1 +i(i+1) h2 2 h h 2 1 =: 1 =: fm (i). In an analogous way one veriﬁes that λ1 ≤ cot(βk −ηk ) cot ηk λ2 =⇒ c − b tan βk ≥ 0. 2 By means of these results we obtain cot(ρi −βi ) tan ρi = cot βi + tan ρi = 1+ cot βi − cot ρi 2 (cot βi +tan βi+1 )2 (cot βi −tan βi+1 )2 +4 .32) For the case − π < βk < 0 a similar reasoning can be applied leading also to 2 (3.32). Then we have tan βi+1 = cot βi This gives (cot βi + tan βi+1 )2 = (cot βi − tan βi+1 )2 + 4 1− For m > in 1 h2 2 h1 (i + 1) h2 .4. Then. 91 (3. (c) Let us ﬁrst consider the case 1 ≤ i ≤ 2m−2.3.
= cot It is not hard to verify that for −2m + 1 ≤ i ≤ −1 the preceding results carry over. one establishes that m→∞ lim i∈Jm min cot(βi −ηi ) cot ηi = ∞. Hence. For i = 2m−1 we have cot ρ2m−1 −β2m−1 tan ρ2m−1 π β2m−1 π β2m−1 tan − + 4 2 4 2 π β2m−1 = tan2 + 4 2 → ∞ for m → ∞. For m ≤ i ≤ 2m−2 similar calculations show that by means of tan βi+1 = cot βi one obtains m≤i≤2m−2 mh2 .33) Now. min cot(ρi −βi ) tan ρi fm (xm ) − 1 → ∞ for m → ∞. SEMIDISCRETE DIFFUSION FILTERING Thus. (2m−i−1) h1 (2m−i) h1 = mh2 min cot(ρi −βi ) tan ρi → ∞ for m → ∞. (3.34) we ﬁnally end up with the assertion m→∞ lim i∈Jm min κi. 1 → 1+ 1 − gm (xm ) ≥ 1+ 2 for m → ∞. for 1 ≤ i ≤ m−1. in a similar way as above.92 CHAPTER 3.34) From (3. (3. 2 .m = ∞.33) and (3. m→∞ lim i∈Jm min cot(ρi −βi ) tan ρi = ∞. gm (i) ≤ gm (xm ) → 0+ which yields fm (i) ≤ This gives 1≤i≤m−1 for m → ∞.
π ] =: I−1 ∪ I1 . Therefore. 0] ∪ (0. 2 2 2 2 . (c) For a speciﬁed diﬀusion tensor function D it is possible to give apriori estimates for the required stencil size: using the extremum principle it is not hard to show that ∇uσ (x. As a remedy.4.3. RELATION TO CONTINUOUS MODELS Remarks: 93 (a) We observe that the preceding existence proof is constructive. Since m = 1 we have a partition of (− π . b and c may be functions of Jρ (∇uσ ). we may expect to ﬁnd stricter estimates than those given in the proof. even if we use only 3 directions. Thanks to the uniform positive deﬁniteness of D there exists an upper limit for the spectral condition number of D. This condition limit can be used to ﬁx a suitable stencil size. π ] into 4m−2 = 2 subintervals: 2 2 (− π . Crandall and Lions [104] propose to study a convergent sequence of regularizations which can be approximated on a ﬁnite stencil. a (2m+1)×(2m+1)stencil reveals much more directions than those 4m that are induced by the 8m “boundary pixels”. where D= a b b c and a. Moreover. only three directions are suﬃcient to guarantee a nonnegative directional splitting. most of the stencil coeﬃcients can be set to zero. π ] = (− π . Let us now illustrate the ideas in the proof of Theorem 6 by applying them to a practical example: We want to ﬁnd a nonnegative spatial discretization of div (D∇u) on a (3×3)stencil. t) ≤ 4 f √ L∞ (Ω) 2π σ ¯ on Ω × (0. (d) The existence of a nonnegative directional splitting distinguishes the ﬁlter class (Pc ) from morphological anisotropic equations such as mean curvature motion. where the notations from Chapter 2 have been used. In this case it has been proved that it is impossible to ﬁnd a nonnegative directional splitting on a ﬁnite stencil [13]. ∞). unless m is very small. These estimates might be improved further by admitting more than 3 directions. Thus. t) = (∇Kσ ∗u)(x. (b) Especially for large m.
1 = (3. h1 . 2 π β θ1 + β1 = + .8284. κ1.1 = κ−1. h2 h1 First we focus on the case ψ ∈ I1 where (cos ψ. 2h1 h2 which restricts the upper condition number for a nonnegative discretization with ψ ∈ I1 to 1 + sin β 1 + cos β .94 CHAPTER 3.30) we obtain as expressions for the directional diﬀusivities α−1 = b − b · h2 + h2 1 2 .35) κ1.1 := min . ρ1 = 2 4 2 β . 1 − sin β 1 − cos β α0 = a − b · α1 α2 h1 . In this case β = π gives 4 √ 1 √ 1+ 2 2 √ = 3 + 2 2 ≈ 5. SEMIDISCRETE DIFFUSION FILTERING I−1 and I1 belong to the grid angles β−1 = arctan − β1 = arctan h2 . h1 =: β. 1 − sin β Thanks to the symmetry we obtain the same condition restriction for ψ ∈ I−1 . (3. η1 = 2 Therefore. 2 1 − cos β = 1 + sin β .28)–(3.36) 1 1− 2 2 By virtue of (3. h2 h2 + h2 2 . These bounds on the condition number attain their maximal value for h1 = h2 . sin ψ) denotes the eigenvector to the larger eigenvalue λ1 of D. we get cot(ρ1 −β1 ) tan ρ1 = cot cot(β1 −η1 ) cot η1 π β π β tan − + 4 2 4 2 β 1 + cos β = cot2 = . = b + b · 1 2h1 h2 h2 = c − b · . With the notations from the proof of Theorem 6 we obtain π θ1 = .
j  2h1 h2 − bi−1.j +ci.3.j+1 +bi+1.j +2ai.g.j  2h1 h2 bi−1.j +ai. . j) is the index of some inner pixel (i.j 4h1 h All nonvanishing entries of the pth row of A(u) are represented in this stencil. where p(i.j+1 +2bi. for instance.j−1 +bi−1.g. yj ). [294. His derivation is based on the diophantine problem of approximating irrationals by rational numbers.j 4h1 h ci. y) ∂y u + ∂y b(x.j−1 +bi−1.j −b2i.j−1 4h1 h2 + bi−1.j +bi. j).j+1 4h1 h2 − bi−1. j). The approach presented here extends these results to ∂x a(x.j +bi+1. Recently Kocan described an interesting alternative to obtain upper bounds for the stencil size as a function of the condition number [238]. y) ∂xy u + c(x. y) ∂x u + ∂y c(x.j−1 −bi+1.j +ai+1. e.j+1 4h1 h2 + bi.j−1 +2ci.j +b2i.j  2h1 h2 + bi.j+1 −bi−1.j 2h2 2 − bi.j −b2i.j +b2i. y) ∂yy u.j  2h1 h2 − ci. The problem of ﬁnding nonnegative diﬀerence approximations to elliptic expressions with mixed derivatives has a long history.j+1 +ci.j  2h1 h2 bi+1.j 4h1 h ci.4.j +bi. RELATION TO CONTINUOUS MODELS 95 This induces in a natural way the following secondorder discretization for div (D∇u): bi−1. see e.j+1 +bi+1. 120.j+1 2h2 2 ai+1. j+1)) of A(u).j−1 −bi+1.j−1 +bi.j+1 −bi−1. Usually it is studied for the expression a(x.j−1 4h1 h2 + bi. p(i−1.j+1 +bi. The other notations should be clear from the context as well. bi.j−1 +bi+1.j 2h2 1 − bi−1.j 4h1 h − ai−1.j +bi.j 2h2 1 − bi+1.j +ai. Thus.j 2h2 2 − bi. y) ∂xx u + 2b(x. 170].j−1 +ci.j 2h2 1 ai−1. y) ∂y u b and establishes the relation between the condition number of a c and the nonb negativity of the diﬀerence operator. the upper left stencil entry gives the element (p(i.j+1 4h1 h2 bi+1. y) ∂x u + ∂x b(x.j+1 +bi+1.j denotes a ﬁnite diﬀerence approximation of b(Jρ (∇uσ )) at some grid point (xi .j−1 4h1 h2 + bi.j−1 +bi.
SEMIDISCRETE DIFFUSION FILTERING .96 CHAPTER 3.
4. (D6) positive diagonal: . u(k+1) = Q(u(k) ) u(k) . uniqueness.. ∀ v ∈ IRN . N ≥ 2. j ∈ J. We shall see that this class comprises αsemiimplicit discretizations of the semidiscrete ﬁlter class (Ps ) as well as certain variants of them which are based on an additive operator splitting. maximumminimum principle. qii (v) > 0 ∀ i ∈ J.Chapter 4 Discrete diﬀusion ﬁltering This chapter presents a discrete class of diﬀusion processes for which one can establish similar properties as in the semidiscrete case concerning existence. 97 ∀ v ∈ IRN . Q ∈ C(IRN . average grey level invariance. We consider the following discrete ﬁlter class (Pd ): Let f ∈ IRN . N where Q = (qij ) has the following properties: (D1) continuity in its argument: (D2) symmetry: (D3) unit row sum: (D4) nonnegativity: qij (v) = qji (v) ∀ i. and denote the index set {1. j∈J qij (v) = 1 ∀ i ∈ J. Calculate a sequence (u(k) )k∈IN0 of processed versions of f by means of u(0) = f.. IRN ×N ).1 The general model As in Chapter 3 we regard a discrete image as a vector f ∈ IRN . continuous dependence of the solution on the initial image. ∀ v ∈ IR . N} by J. . j ∈ J. Lyapunov sequences and convergence to a constant steadystate. (D5) irreducibility for all v ∈ IRN . ∀ k ∈ IN0 . ∀ v ∈ IRN . (Pd ) qij (v) ≥ 0 ∀ i..
223].3) b := max fj .2) (4. for all i ∈ J and k ∈ IN0 . u(k) depends continuously on f . Then. if Q is stochastic and i∈J qij = 1 for all j ∈ J. then Q is doubly stochastic. The existence of Lyapunov sequences is a consequence of (D2)–(D4).2 Theoretical results It is obvious that for a ﬁxed ﬁlter belonging to the class (Pd ) every initial image f ∈ IRN generates a unique sequence (u(k) )k∈IN0 . Proposition 3 (Extremum principle). it is evident that fully discrete results are of importance since. (k) a ≤ ui ≤ b ∀ i ∈ J. j∈J (4. while the proof of the extremum principle requires (D3) and (D4). and the convergence to a constant steadystate utilizes (D2)–(D5). (b) The requirements (D1)–(D5) have a similar meaning as their semidiscrete counterparts (S1)–(S5). This indicates that our discrete diﬀusion processes are related to the theory of Markov chains [370. for every ﬁnite k.98 Remarks: CHAPTER 4. Indeed. the following inequalities hold: (i) ui (k+1) = j∈J qij (u(k) )uj (k) (D4) ≤ max u(k) m m∈J j∈J qij (u(k) ) = max u(k) . Therefore. strict Lyapunov sequences need (D5) and (D6) in addition to (D2)–(D4). ∀ k ∈ IN0 . j∈J Proof: The maximum–minimum principle follows directly from the fact that. Moreover. in practice. by means of (D1) we know that. m m∈J (D3) . Let f ∈ IRN and let (u(k) )k∈IN0 be the sequence of ﬁltered images according to (Pd ). (D1) immediately gives wellposedness results. scalespace evolutions are evaluated exclusively at a ﬁnite number of scales. and average grey value invariance is based on (D2) and (D3). DISCRETE DIFFUSION FILTERING (a) Although the basic idea behind scalespaces is to have a continuous scale parameter.1) where a := min fj . (4. Moreover. let us now prove a maximum–minimum principle. 4. (c) Nonnegative matrices Q = (qij ) ∈ IRN ×N satisfying j∈J qij = 1 for all i ∈ J are also called stochastic matrices.
. Proposition 4 (Conservation of average grey value).. Assume that (u(k) )k∈IN0 satisﬁes the requirements of (Pd ). However. (k) k ∈ IN0 is a Lyapunov sequence: . Then the following properties are fulﬁlled: (a) (Lyapunov sequences) For all convex r ∈ C[a. . µ.4. and (4.4) ∀ k ∈ IN0 . µ)⊤ ∈ IRN . (4. As one might expect. This socalled redistribution property [164] ensures that. ui i∈J (k+1) = i∈J j∈J qij (u(k) )uj (k) = j∈J i∈J qij (u(k) ) uj (k) = j∈J uj . the class (Pd ) allows an interpretation as a transformation which is smoothing in terms of Lyapunov sequences. SCALESPACE PROPERTIES (ii) ui (k+1) 99 qij (u(k) ) = min u(k) . which generates the Lyapunov sequences. for all k ∈ IN0 .3. The average grey level 1 fj µ := N j∈J is not aﬀected by the discrete diﬀusion ﬁlter: 1 N uj = µ j∈J (k) (4. These functions ensure that u(k) converges to a constant image as k → ∞. we need less regularity than in the semidiscrete case: The convex function r.3 Scalespace properties All statements from Chapter 3 with respect to invariances are valid in the discrete framework as well. and let c := (µ. let a. m m∈J (D3) = j∈J qij (u(k) )uj (k) (D4) ≥ min u(k) m m∈J j∈J 2 4. Below we focus on proving average grey level invariance. b. and µ be deﬁned as in (4. respectively.2). needs only to be continuous.. b] the sequence V (k) := Φ(u(k) ) := i∈J r(ui ). 2 (k) which proves the proposition.5) Proof: By virtue of (D2) and (D3) we have i∈J qij (u(k) ) = 1 for all j ∈ J and k ∈ IN0 . but no more diﬀerentiable. (4.4). Theorem 7 (Lyapunov sequences and behaviour for k → ∞).3).
and the prerequisites (D2) and (D3) we obtain N V (k+1) − V (k) = i=1 conv.7) N qij (u(k) ) r(uj ) − r(ui ) (k) aij (u(k) ) r(uj ) i=1 j=1 N N (D3) = i=1 j=1 aij (u(k) ) r(uj ) − r(ui ) (k) (k) .7) Using the convexity of r. the preceding deﬁnition. DISCRETE DIFFUSION FILTERING Φ(u(k) ) ≥ Φ(c) ∀ k ∈ IN0 V (k+1) − V (k) ≤ 0 ∀ k ∈ IN0 Moreover.100 (i) (ii) CHAPTER 4. j ∈ J we deﬁne aij (u(k) ) := qij (u(k) ) − 1 qij (u(k) ) (i = j) (i = j). k→∞ Proof: (a) (i) Average grey level invariance and the convexity of r give N Φ(c) = i=1 N ≤ = i=1 N j=1 r 1 N 1 (k) uj j=1 N N (k) j=1 N r(uj ) (4. then V (k) = Φ(u(k) ) is a strict Lyapunov sequence: (iii) (iv) Φ(u(k) ) = Φ(c) V (k+1) − V (k) = 0 ⇐⇒ u(k) = c u(k) = c ⇐⇒ (b) (Convergence) lim u(k) = c. (4.6) (k) r(uj ) = Φ(u(k) ). r N j=1 N j=1 N i=1 N (k) (k) qij (u(k) ) uj − r(ui ) (k) (k) ≤ = (4. (ii) For i. if r is strictly convex.
. To this end. . we obtain V (k+1) − V (k) N = i=1 i=n r N N j=1 + r (k) qij (u(k) ) uj − j=1 (k) qnj (u(k) ) uj (k) r(ui ) − r(u(k)) n . (iii) This part of the proof can be shown in exactly the same manner as in the semidiscrete case (Chapter 3. the nonnegativity of Q(u(k) ) gives qnm (u(k) ) > 0. Moreover.. (iv) In order to verify the ﬁrst implication. lr = j0 and qlp lp+1 = 0 for p = 0. Theorem 5): Equality in the estimate (4.i+m (u(k) ) r(ui+m ) − r(ui ) (4. we ﬁnd l0 .8) (k) (k) (k) (k) = 0.m qnj (u(k) ) uj + qnn (u(k) ) u(k) + qnm (u(k) ) u(k) n m (k) < j=1 j=n. Hence. r − 1. there exists some p0 ∈ {0.6) holds due to the strict convexity of r if and only if u(k) = c.. Together with the strict convexity of r these properties lead to r = N (k) qnj (u(k) ) uj N j=1 (k) r N j=1 j=n..3. . lr ∈ J with l0 = i0 . (k) If we combine this with the results in (4.. by the irreducibility of Q(u(k) )... SCALESPACE PROPERTIES = + (D2) N N −i i=1 m=1 N N −i i=1 m=1 101 ai+m. and (k) u(k) = un .m qnj (u(k) ) r(uj ) + qnn (u(k) ) r(u(k)) + qnm (u(k) ) r(u(k)) n m N = j=1 qnj (u(k) ) r(uj ). r − 1} such that n := lp0 . = uN . qnm (u(k) ) = 0.8).. let us start with a proof that (k) (k) V (k+1) = V (k) implies u1 = .. . m := lp0 +1 .. (k) Then.. assume that u(k) is not constant: ui0 := min ui i∈J (k) (k) < max uj j∈J (k) =: uj0 .4. m and by (D6) we have qnn (u(k) ) > 0.i (u(k) ) r(ui ) − r(ui+m ) ai.
let u(k) = c. Conversely. = uN . we can argue exactly in the same way as in the semidiscrete case if we replace Lyapunov functions by Lyapunov sequences and integrals by sums..102 CHAPTER 4. i=1 (c) In order to prove convergence to a constant steadystate. Theorem 5 for more details. 2 In analogy to the semidiscrete case the preceding theorem comprises many Lyapunov functions which demonstrate the informationreducing qualities of our ﬁlter class.8) N j=1 = 0. (k) (k) for all n ∈ IN. 4.2) and (4. DISCRETE DIFFUSION FILTERING N < i=1 (4.4 4. qij (u(k) ) r(uj ) − r(ui ) (k) (k) (k) (k) This establishes that V (k+1) = V (k) implies u1 = . we conclude that u(k) = c. if a > 0. See Chapter 3. To this end. Let (u(k) )k∈IN0 be a diﬀusion sequence according to (Pd ). Then the following functions are decreasing in k: (a) (b) (c) u(k) p for all p ≥ 1. decreasing central moments and increasing entropy is evident. we immediately obtain the following corollary.4). By means of prerequisite (D3) we obtain N V (k+1) − V (k) = i=1 r N j=1 qij (u(k) )µ − N r(µ) = 0. we regard u(k) as an . r(s) := (s − µ)2n and r(s) := s ln s. Then. by virtue of the grey value invariance.4. 1 N N j=1 (k) M2n [u(k) ] := H[u(k) ] := N (uj − µ)2n ln(uj ).1 Relation to semidiscrete models Semiimplicit schemes Let us now investigate in which sense our discrete ﬁlter class covers in a natural way time discretizations of semidiscrete ﬁlters. Corollary 3 (Special Lyapunov sequences). and let a and µ be deﬁned as in (4. j=1 uj An interpretation of these results in terms of decreasing energy. Choosing the convex functions r(s) := sp ..
1). Now assume that α > 0. Then B is strictly diagonally dominant. 1]. the argument u(k) is suppressed frequently since the considerations below are valid for all u(k) ∈ IRN . while the linear remainder is discretized in an αimplicit manner. Theorem 8 (Scalespace interpretation for αsemiimplicit schemes). since bii = 1 − ατ aii = 1 + ατ (S3) aij > ατ j∈J j=i j∈J j=i aij = (S4) j∈J j=i bij  ∀ i ∈ J. They reveal the advantage that the linear implicit part ensures good stability properties. max aii (u(k) ) i∈J (4. The theorem below states that this class of schemes is covered by the discrete framework. then B = I and hence invertible. for which we have established scalespace results. Then the αsemiimplicit scheme u(k+1) − u(k) = A(u(k) ) αu(k+1) + (1−α)u(k) τ fulﬁls the prerequisites (D1)–(D6) for discrete diﬀusion models provided that τ ≤ 1 (1−α) max aii (u(k) ) i∈J (4.1.9) can be written as B(u(k) ) u(k+1) = C(u(k) ) u(k) we ﬁrst have to show that B(u(k) ) is invertible for all u(k) ∈ IRN .4. Henceforth. where I ∈ IRN denotes the unit matrix. and let A = (aij ) : IRN → IRN ×N satisfy the requirements (S1)–(S5) of section 3. where τ denotes the time step size. In the explicit case (α = 0) the properties (D1)–(D6) hold for τ < 1 . Proof: Let B(u(k) ) := (bij (u(k) )) := I − ατ A(u(k) ).4. We consider a ﬁnite diﬀerence scheme with two time levels where the operator A – which depends nonlinearly on u – is evaluated in an explicit way.9) (4. Such schemes are called αsemiimplicit. C(u(k) ) := (cij (u(k) )) := I + (1−α)τ A(u(k) ). Let α ∈ [0.10) for α ∈ (0. τ > 0. RELATION TO SEMIDISCRETE MODELS 103 approximation of the solution u of (Ps ) at time t = kτ . . If α = 0.11) and the semiimplicit case (α = 1) satisﬁes (D1)–(D6) unconditionally. Since (4. while the explicit evaluation of the nonlinear terms avoids the necessity to solve nonlinear systems of equations.
14) = Hv = v qij = 1 for all i ∈ J. By means of (S3) we obtain bij = 1 = j∈J j∈J This also shows that bii > 0 for all i ∈ J. cij ∀ i ∈ J.12) is equivalent to Bv = v = Cv. (4. we ﬁrst check the nonnegativity of C. Thus. from Qv = HCv we conclude that j∈J (4. In order to show that (D4) is fulﬁlled.. Thanks to the fact that B is irreducibly diagonally dominant. 1)⊤ ∈ IRN . we know from [407.14) (4. Then (4. Q := (qij ) := B −1 C exists and by (S1) it follows that Q ∈ C(IRN . DISCRETE DIFFUSION FILTERING The requirement (D2) is not hard to satisfy: Since B −1 and C are symmetric and reveal the same set of eigenvectors – namely those of A – it follows that Q = B −1 C is symmetric as well. IRN ×N ). however. For 0 ≤ α < 1. C(u(k) ) is nonnegative if τ ≤ 1 =: τα (u(k) ). This proves (D3). bij ≤ 0 for all i = j. j ∈ J.. and the invertibility of B gives v = B −1 v = Hv. (k) ) (1−α) max aii (u i∈J (S3) (S4) aij j∈J j=i (S4).. (4. .13) (4.(S5) < 0 ∀ i ∈ J. Therefore. 85] that B −1 =: H =: (hij ) exists and hij > 0 for all i.104 CHAPTER 4. For i = j we have cij = (1−α)τ aij ≥ 0.13) (4. p. This proves (D1). and bii > 0 for all i ∈ J.12) Let v := (1.15) . If α = 1 we have cii = 1 for all i ∈ J. and by the structure of the oﬀdiagonal elements of B we observe that the irreducibility of A implies the irreducibility of B. Let us now verify (D3). The diagonal entries yield cii = 1 + (1−α)τ aii . nonnegativity of C is not automatically guaranteed: Using (S3)–(S5) we obtain aii = − Hence.
11) shows that they reveal the most prohibitive time step size restrictions. j ∈ J with i = j we know that aij (u(k) ) > 0 implies cij (u(k) ) > 0. Explicit twolevel schemes are comprised by the choice α = 0. This means that u(k) belongs to the compact set {v ∈ IRN v ∞ ≤ f ∞ } for all k ∈ IN0 .4. v ∈ IRN .12) we obtain qij = k∈J hik ckj ≥ min hik · k∈J >0 ckj > 0 k∈J =1 ∀ i. IR ) it follows that Kf := max aii (v) i ∈ J. and by the positivity of H we have qij > 0 for all i. Next let us consider the case 0 < α < 1 and τ ≤ τα (u(k) ). and (4. (b) The conditions (4. If α = 1. then C = I.10) and (4. Now for τ < 1 max aii (u(k) ) i∈J it follows that cii (u(k) ) > 0 for all i ∈ J and. and (4. Since the semiimplicit scheme fulﬁls (D1)–(D6) we know by Theorem 3 that the solution obeys an extremum principle. Equation (4. j ∈ J. the symmetry of C. 2 Remarks: (a) We have seen that the discrete ﬁlter class (Pd ) – although at ﬁrst glance looking like a pure explicit discretization – covers the αsemiimplicit case as well. j ∈ J. for α = 0.11) can be satisﬁed by means of an apriori estimate. the positivity of H.4. By N N ×N A ∈ C(IR . choosing τ ≤ 1 (1−α) Kf . In all the abovementioned cases the time step size restrictions for ensuring irreducibility imply that all diagonal elements of Q(u(k) ) are positive. Thus. Now we want to prove (D5). This establishes (D6). Using this information. v ∞ ≤ f ∞ exists. Q is irreducible. Finally. we know that the nonnegativity of C implies the nonnegativity of Q = HC. which establishes the irreducibility of Q. we have Q = C. the irreducibility of A(u(k) ) carries over to Q(u(k) ). For i. Thus. RELATION TO SEMIDISCRETE MODELS 105 Since H is nonnegative.15) shows that Kf > 0. thus. Then we know that C is nonnegative.
11) holds.4. Multigrid methods [59. and their convergence can be guaranteed for the special structure of A. Modiﬁcations in order to reduce these problems [122] are quite diﬃcult to implement. DISCRETE DIFFUSION FILTERING ensures that (4. a large linear system of equations has to be solved.10) is always satisﬁed. this advantage usually overcompensates for the additional eﬀort of resolving a linear system. one may for . Compared to explicit schemes. i+2p2 +2p]. (e) By the explicit discretization of the nonlinear operator A it follows that all schemes in the preceding theorem are of ﬁrst order in time. and positive deﬁnite. that in the ith row these entries can be found within the positions [i. Faster iterative methods such as preconditioned conjugate gradient algorithms [348] need signiﬁcantly more storage. i−2p2 −2p] to [i. Classical methods such as Gauß–Seidel or SOR [447] are easy to code. In spite of the fact that the nonlinearity is discretized in an explicit way they are absolutely stable in the maximum norm. In general. if one insists in secondorder schemes. however. the matrix reveals a much larger bandwidth. if it results from a ﬁnite diﬀerence discretization on a (2p+1)×(2p+1)stencil it contains at most 4p2 +4p+1 nonvanishing entries per row. Usually. which can become prohibitive for large images. since in image processing one is in general more interested in maintaining qualitative properties such as maximum principles or invariances rather than having an accurate approximation of the continuous equation. This should not give rise to concern. Iterative algorithms appear to be better suited. Unfortunately. A typical problem of iterative methods is that their convergence slows down for larger τ .2. they converge rather slowly. they do not need additional storage. Its system matrix is symmetric. (d) For α = 1 we obtain semiimplicit schemes which do not suﬀer from time step size restrictions. it is also sparse: For instance. another possibility will be studied in Section 4. 1 Kf (c) If α > 0. and τ < guarantees that (4. Applying standard direct algorithms such as Gaussian elimination would destroy the zeros within the band and would lead to an immense storage and computation eﬀort. diagonally dominant. 179] are one possibility to circumvent these diﬃculties. One should not expect. and they inherit the scalespace properties from the semidiscrete setting regardless of the step size.106 CHAPTER 4. However. since this increases the condition number of the system matrix.
Hence. Since this can be numerically expensive.18) By means of a Taylor expansion one can verify that.4.4.16) such that the m linear systems with system matrices (I −mατ Al (u(k) )). 4.. for α > 0. Suppose we know a splitting m A(u(k) ) = l=1 Al (u(k) ). Nonnegativity of Q(u(k) ) may also be achieved using spatial discretizations where A(u(k) ) has negative oﬀdiagonal elements (see [55] for examples). 2 τ This scheme satisﬁes the properties (D1)–(D6) for τ ≤ 2/Kf ... the preceding αsemiimplicit schemes require to solve a linear system with the system matrix (I − ατ A(u(k) )).4. (4. it would be nice to have an eﬃcient alternative. . RELATION TO SEMIDISCRETE MODELS 107 instance use the predictor–corrector approach by Douglas and Jones [121]: u(k+1/2) − u(k) = A(u(k) ) u(k+1/2) . τ /2 u(k+1) − u(k) 1 = A(u(k+1/2) ) 1 u(k+1) + 2 u(k) . although both schemes are not identical. (f) The assumptions (S1)–(S5) are suﬃcient conditions for the αsemiimplicit scheme to fulﬁl (D1)–(D6). but they are not necessary. from a numerical viewpoint.17) its additive operator splitting (AOS) variant [424] u(k+1) = 1 m m l=1 I − αmτ Al (u(k) ) −1 I + (1−α)mτ Al (u(k) ) u(k) . they have the same approximation order in space and time. they are both consistent approximations to the semidiscrete ODE system from (Ps ).2 AOS schemes We have seen that. l = 1. Then it is advantageous to study instead of the αsemiimplicit scheme m u (k+1) = I − ατ Al (u ) l=1 (k) −1 m I + (1−α)τ l=1 Al (u(k) ) u(k) (4. (4.m can be solved more eﬃciently. The following theorem clariﬁes the conditions under which AOS schemes create discrete scalespaces..
1.18) fulﬁls the prerequisites (D1)–(D6) of discrete diﬀusion scalespaces provided that τ < 1 . p. biil > 0 for all i. 1]. (S3) aijl > αmτ j∈J j=i aijl = (S4) j∈J j=i bijl  ∀ i ∈ J.108 CHAPTER 4. and bijl ≤ 0 for i = j. 1)... DISCRETE DIFFUSION FILTERING Theorem 9 (Scalespace interpretation for AOS schemes). assume that A(u) = m Al (u) l=1 is irreducible for all u ∈ IRN . and let Al = (aijl )ij : IRN → IRN ×N . the properties (D1)–(D6) are unconditionally satisﬁed. In the same way as in the proof of Theorem 8 one shows that Bl−1 Cl has only unit row sums for all l. The symmetry property (D2) of Q results directly from the fact that Bl−1 and Cl are symmetric and share their eigenvectors with those of Al . τ > 0. and that for each Al there exists a permutation matrix Pl ∈ IRN ×N such that Pl Al PlT is block diagonal and irreducible within each block. (1−α) m max aiil (u(k) ) i. Since Al (u) is continuous in u by virtue of (S1).m satisfy the requirements (S1)–(S4) of section 3. Moreover. Then the following holds: For α ∈ (0. it follows that Q(u) := 1 m m l=1 Bl−1 (u) Cl (u) is also continuous in u. and (D3) is satisﬁed.19) In the semiimplicit case (α = 1). Let α ∈ (0. (S4) . Let Bl (u(k) ) := (bijl (u(k) ))ij := I − αmτ Al (u(k) ). l = 1. Thus. This proves (D1). a suﬃcient condition for proving (D4) is to ensure that Cl is nonnegative for all l. the AOS scheme (4. we utilize that Bl is strictly diagonally dominant. Bl is invertible because of its strict diagonal dominance: biil = 1 + αmτ j∈J j=i Cl (u(k) ) := (cijl (u(k) ))ij := I + (1−α)mτ Al (u(k) ). 192] that Hl := Bl−1 is nonnegative in all components.l (4.. For i = j we have cijl = (1−α)mτ aijl ≥ 0. Thus. the row sums of Q are 1 as well. To verify (D4).. Proof: The reasoning is similar to the proof of Theorem 8. Under these circumstances it follows from [284.
As a consequence. Then there exists an l0 ∈ {1. Denoting Bl−1 by Hl = (hijl )ij .4..l guarantees that ciil > 0 ∀ i ∈ J. RELATION TO SEMIDISCRETE MODELS The diagonal entries yield ciil = 1 + (1−α)mτ aiil .21) also proves (D6): By virtue of (4. Each block is irreducible and strictly diagonally dominant with a positive diagonal and nonpositive oﬀdiagonals. nonnegativity of C is not automatically guaranteed: Since Al satisﬁes (S3) and (S4). it is also possible to construct a predictor–corrector scheme of Douglas–Jones type [121] within the AOS framework: u(k+1/2) = u(k+1) 1 m m l=1 m l=1 1 I − 2 mτ Al (u(k) ) −1 u(k) . −1 1 = m 1 I − 2 mτ Al (u(k+1/2) ) I + 1 mτ Al (u(k+1/2) ) u(k) . This can be seen as follows: There exist permutation matrices Pl . j0 ∈ J. Moreover. Thus. . however...15) we have aii < 0 for all i ∈ J..j0 ) 0 ≥0 ≥0 >0 >0 (k) Recapitulating. Moreover. m such that Pl Bl PlT is block diagonal. l = 1. the irreducibility of Q.. m (l. p. Thus. . ai0 j0 = 0 =⇒ qi0 j0 > 0. a theorem by Varga [407. for all i. we show now that ai0 j0 l0 = 0 implies hi0 j0 l0 > 0. 2 Remarks: (a) In analogy to the unsplit αsemiimplicit schemes... 2 .21) Thus. For 0 < α < 1. and (D5) is proved.20) this yields qi0 j0 = 1 hi nl cnj0 l + hi0 j0 l0 cj0 j0 l0 > 0. for τ < τα (u ). we know that aiil ≤ 0. by (4.n)=(l0 . Again. Therefore. ∀ l = 1. Q must have a positive diagonal. Suppose that ai0 j0 = 0 for some i0 . (4. ai0 j0 l0 = 0 implies hi0 j0 l0 > 0. 85] ensures that the inverse of each block contains only positive elements. .15) it follows that for every i there exists an l with aiil < 0. Together with (4. m. (4..4. 109 If α = 1 we have cii = 1 for all i ∈ J.20) Next we prove (D5). (4. m} such that ai0 j0 l0 = 0.. the irreducibility of A carries over to Q. this means that. requiring 1 =: τα (u(k) ) τ < (1−α) m max aiil (u(k) ) i. the case α = 1 is especially interesting. because no time step size restriction occurs.
(c) The result u(k+1) of an AOS scheme can be regarded as the average of m ﬁlters of type vl (k+1) := I − αmτ Al (u(k) ) (k+1) −1 I + (1−α)mτ Al (u(k) ) u(k) (l = 1. 354. v ∈ IRN . This violates requirement (D2) for discrete scalespaces. In practice. 286.. m}. l ∈ {1. Moreover.. a natural operator splitting A = m Al results from a decoml=1 position of the divergence expression into onedimensional terms of type ∂xl (g(∇uσ 2 ) ∂xl u) (l = 1. this means that these schemes produce diﬀerent results if the image is rotated by 90 degrees. .. the result of multiplicative splittings will depend on the order of the operators. [81]. v ∞ ≤ f ∞ . DISCRETE DIFFUSION FILTERING It satisﬁes (D1)–(D6) for τ < 2/Kf . Since vl . In this case. where Kf is determined by the apriori estimate Kf := m max aiil (v) i ∈ J. They are multiplicative splittings. m do not commute. m)..2 we have seen that this ﬁlter is based on the PDE ∂t u = div (g(∇uσ 2 ) ∇u).. l = 1... the corresponding linear systems can be . it is possible to distribute their computation to diﬀerent processors of a parallel machine. However.. (d) AOS schemes with α = 1 have been presented in [424. Hence... 277... Since in the general nonlinear case the split operators Al .m can be calculated independently from each other.. In e Section 1. m). . the AOS–Douglas–Jones scheme is only ﬁrst order accurate in time.. This separation is very eﬃcient: There exist permutation matrices Pl (pixel orderings) such that Pl Al PlT is block diagonal and each block is diagonally dominant and tridiagonal. (b) The fact that AOS schemes use an additive splitting ensures that all coordinate axes are treated in exactly the same manner. . 430] as eﬃcient discretizations of the isotropic nonlinear diﬀusion ﬁlter of Catt´ et al.3. most multiplicative splittings lead to a nonsymmetric matrix Q(u(k) ). 442].. This is in contrast to the various conventional splitting techniques from the literature [120.110 CHAPTER 4.. A typical representative is m u (k+1) = l=1 I − τ Al (u(k) ) −1 u(k) .. l = 1.
The resulting forward substitution and backward elimination can be regarded as a recursive ﬁlter.4. RELATION TO SEMIDISCRETE MODELS 111 solved in linear eﬀort by means of a simple Gaussian algorithm.4. .2. then a splitting of such a 2D ﬁlter into 4 onedimensional diﬀusion processes acting along the 4 stencil directions satisﬁes all prerequisites of Theorem 9. A parallel implementation assigning these tridiagonal subsystems to diﬀerent processors is described in [431]. If it fulﬁls (S1)–(S5).4. The denoising of a medical 3D ultrasound data set with 138 × 208 × 138 voxels on an SGI Power Challenge XL with eight 195 MHz R10000 processors was possible in less than 1 minute. (e) The idea to base AOS schemes on decompositions into onedimensional operators can also be generalized to anisotropic diﬀusion ﬁlters: Consider for instance the discretization on a (3 × 3)stencil at the end of Section 3.
DISCRETE DIFFUSION FILTERING .112 CHAPTER 4.
Chapter 5 Examples and applications The scalespace theory from Chapters 2–4 also covers methods such as linear or nonlinear isotropic diﬀusion ﬁltering. Dilation with a ﬂat structuring element is approximated by an OsherSethian scheme of type (1. for which many interesting applications have already been mentioned in Chapter 1. and MCM. Therefore. 1. we shall pursue two diﬀerent objectives: • smoothing with simultaneous edgeenhancement.1 and ∆t := 0. On an HP 9000/889 workstation it takes less than 0. 113 . 5].3 seconds. In order to compare anisotropic diﬀusion to other methods. In order to demonstrate the ﬂexibility of anisotropic diﬀusion ﬁltering. the goal of the present chapter is to show that a generalization to anisotropic models with diﬀusion tensors depending on the structure tensor oﬀers novel properties and application ﬁelds. and dilation with a quadratic structuring function is performed in a noniterative way using van den Boomgaard’s algorithm from [51].5.06 seconds per iteration. Thus. AMSS or dilation with a disc require approximately 0. morphological scalespaces and modiﬁcations of them have been discretized as well.01.3 CPU seconds to calculate one nonlinear diﬀusion step for a 256 × 256 image. Typical dilations with a quadratic structuring function take less than 0. For MCM and AMSS this is achieved by means of explicit FD schemes (cf. we focus mainly on these anisotropic techniques and juxtapose the results to other methods. respectively. All calculations for diﬀusion ﬁltering are performed using semiimplicit FD schemes with time steps ∆t ∈ [2.6. • smoothing with enhancement of coherent ﬂowlike textures.88) with ∆t := 0.4) with ∆t := 0.
λ2 depends on the desired goal of the ﬁlter.114 CHAPTER 5. λ] and decreasing for s ∈ (λ.1 Edgeenhancing diﬀusion Filter design In accordance with the notations in 2.1 5. see 1. Since ∇uσ remains bounded on Ω × [0. This behaviour may be accomplished by the following choice (m ∈ IN.1. the preceding ﬁlter strategy can be regarded as an anisotropic regularization of the Perona–Malik model.2. Thus. 2. then one can reduce the diﬀusivity λ1 perpendicular to edges the more the higher the contrast µ1 is.3. we know that the uniform positive deﬁniteness of D is automatically satisﬁed by this ﬁlter. λ2 := 1 with g(s) := (5. Since in this section we are only interested in edgeenhancing diﬀusion we may set the integration scale ρ of the structure tensor equal to 0. ∞) and µ1 = ∇uσ 2 . v2 the corresponding orthonormal eigenvectors. 1 1 − exp −Cm (s/λ)m (s ≤ 0) (s > 0).2) This exponentially decreasing function is chosen in order to fulﬁl the smoothness requirement stated in (Pc ). It depicts the temporal evolution of a . cf.3) 5. λ > 0): λ1 (µ1 ) := g(µ1). EXAMPLES AND APPLICATIONS 5.2 Applications Figure 5. Applications which require nonvanishing integration scales shall be studied in section 5.1 illustrates that anisotropic diﬀusion ﬁltering is still capable of possessing the contrastenhancing properties of the Perona–Malik ﬁlter (provided the regularization parameter σ is not too large).2.3 and [415]. let µ1 . Since the diﬀusion tensor should reﬂect the local image structure it ought to be chosen in such a way that it reveals the same set of eigenvectors v1 . The choice of the corresponding eigenvalues λ1 . (5. v2 as Jρ . Cm > 0. and v1 . The choice m := 4 (which implies C4 = 3.3.31488) gives visually good results and is used exclusively in the examples below.1) (5. If one wants to smooth preferably within each region and aims to inhibit diffusion across edges. ∞).1. The constant Cm is calculated in such a way that the ﬂux Φ(s) = sg(s) is increasing for s ∈ [0. µ2 with µ1 ≥ µ2 be the eigenvalues of the structure tensor Jρ .
λ = 3. 125. σ = 2. 625. Ω = (0. 15625.1. 256)2 . From top left to bottom right: t = 0. EDGEENHANCING DIFFUSION 115 Figure 5.6. 78125. 3125.1: Anisotropic diﬀusion ﬁltering of a Gaussiantype function.5. .
This indicates that. 423]. all images in the present work are depicted in such a way that the lowest value is black and the highest one appears white.2(c) shows the eﬀect when applying the isotropic nonlinear diﬀusion equation [81] ∂t u = div (g(∇uσ 2 )∇u) (5. 15. Except for Figs.4) with g as in (5.1. nonlinear isotropic diﬀusion does not lead to correspondence problems which are characteristic for linear ﬁltering. 102]. a theoretically and practically rather diﬃcult problem. 5. the steepness of edges decreases very slowly until the gradient reaches a value where no backward diﬀusion is possible anymore. 5. 5. It combines the good noise eliminating properties of linear diﬀusion with the stable edge structure of nonlinear isotropic ﬁltering. Edge enhancement is caused by the fact that. in practice. Besides the fact that edges get smoothed so that they are harder to identify. where contrast enhancement is to be demonstrated.6. 99. They reveal a range within the interval [0. 255] and all pixels have unit length in both directions.3–5. 1 . 5. The edge location remains stable over a very long time interval. Let us now compare the denoising properties of diﬀerent diﬀusion ﬁlters. Figure 5.1 It can be observed that two regions with almost constant grey value evolve which are separated by a fairly steep edge. 16. smoothing within each region is strongly preferred to diﬀusion between the two adjacent regions. but we have to pay a price: the image becomes completely blurred. the drastically reduced diﬀusivity at edges is also responsible for the drawback that noise at edges is preserved.3–5.2(d) demonstrates that nonlinear anisotropic ﬁltering shares the advantages of both methods. where an MRI slice of a human head is processed [414. EXAMPLES AND APPLICATIONS Gaussianlike function and its isolines. Due to the permitted smoothing along edges. corners get more rounded than in the nonlinear isotropic case. the determination of a suitable stopping time is not a critical problem. In Fig.2(a) consists of a triangle and a rectangle with 70 % of all pixels being completely degraded by noise. Then the image converges quickly towards a constant image. Test images of this type have been used to study the behaviour of ﬁlters such as [13. The scalespace behaviour of diﬀerent PDEbased methods is juxtaposed in Figures 5. Fig.116 CHAPTER 5. once they are identiﬁed at a coarse scale. After the process of contrast enhancement is concluded.6.3). due to the rapidly decreasing diﬀusivity. however. On the other hand. Figure 5.2(b) we observe that linear diﬀusion ﬁltering is capable of removing all noise. This image is taken from the software package MegaWave. they have to be traced back in order to ﬁnd their true location. Since edges are hardly aﬀected by this process. the correspondence problem appears: edges become dislocated. Thus.
(5.5) .3(b) shows the evolution under ∂t u = div (g(∇f 2) ∇u). Fig. (c) Bottom Left: Nonlinear isotropic diﬀusion.3(a)) does not only blur all structures in an equal amount but also dislocates them more and more with increasing scale. σ = 3. 5.5. σ = 3. t = 80.5. (a) Top Left: Test image. Again we observe that linear diﬀusion (Fig. 5. A ﬁrst step to reduce these problems is to adapt the diﬀusivity to the gradient of the initial image f [147]. λ = 3. 128)2 . (d) Bottom Right: Nonlinear anisotropic diﬀusion. Ω = (0. λ = 3. t = 80.2: Restoration properties of diﬀusion ﬁlters. EDGEENHANCING DIFFUSION 117 Figure 5. (b) Top Right: Linear diﬀusion.1. t = 80.5.
6) is used. edges remain better localized and their blurring is reduced. As in Figure 5.4(a) where the regularized isotropic nonlinear diﬀusion ﬁlter (5. All structures are extremely welllocalized and the results are segmentationlike. 5.7) Figure 5. This leads to the nonlinear diﬀusion equation [326] ∂t u = div (g(∇u2) ∇u). On the other hand. Fig. and it is no longer possible to enhance the contour of the entire head. This is illustrated in Figure 5.4(b)). the temporal evolution of these models generates a . A positive consequence of this slight shrinking eﬀect is the fact that small or elongated and thin structures are better eliminated than in the isotropic case. 1.3) is applied.4) with the diﬀusivity (5. (5. however.6) and leads to a nonmonotone ﬂux function. Thus.4(c)) permits diﬀusion along edges and inhibits smoothing across them. (5. we recognize that most of the depicted “segments” coincide with semantically correct objects that one would expect at these scales. Anisotropic nonlinear diﬀusion (Fig. also small structures exist over a long range of scales if they diﬀer from their vicinity by a suﬃciently large contrast.3(c) shows how such a nonlinear feedback is useful to increase the edge localization in a signiﬁcant way: Structures remain welllocalized as long as they can be recognized. At the chin we observe that this equation is indeed capable of enhancing edges. which can be seen at the nose. before it converges to a constant image. t) instead of the original image f (x). but this also leads to stronger blurring of large structures. EXAMPLES AND APPLICATIONS where a diﬀusivity of type [88] g(∇f 2) := 1 1 + ∇f 2 /λ2 (λ > 0). Compared with homogeneous linear diﬀusion. 5. The latter problem can be avoided using a diﬀusivity which decreases faster than (5. Unlike diﬀusion–reaction models aiming to yield one segmentationlike result for t → ∞ (cf. On the other hand. The absolute contrast at edges. for large t the ﬁltered image reveals some artifacts which reﬂect the diﬀerential structure of the initial image.118 CHAPTER 5. becomes smaller. Also blurring at edges is reduced very much.2(d). Finally the image turns into a silhouette of the head. this causes a stronger rounding of structures. One can try to make this ﬁlter faster and more insensitive to smallsize structures by increasing the regularizing Gaussian kernel size σ (cf. The tendency to produce piecewise almost constant regions indicates that diffusion scalespaces with nonmonotone ﬂux are ideal preprocessing tools for segmentation. A natural idea to reduce the artifacts of inhomogeneous linear diﬀusion ﬁltering would be to introduce a feedback in the process by adapting the diﬀusivity g to the gradient of the actual image u(x.4).
1. This is not true for the morphological ﬁlters and their modiﬁcations which are depicted in Fig. for the disc. The traditional opinion that the evolution parameter t of scalespaces should be related to the spatial scale reﬂects the assumption that a scalespace analysis should be uncommitted. which gives a hierarchy of structures and allows to extract the relevant information from a certain scale.5. The contrast and noise parameters λ and σ give the user the liberty to adapt nonlinear diﬀusion scalespaces to the desired purpose in order to reward interesting features with a longer lifetime.5. 5. It is not very diﬃcult to guess the shape of the structuring function from the scalespace evolution. Therefore. than a descriptor of spatial scale. and by ∂t u = ∇u2 (5. the time t is rather a parameter of importance.3 and 1. Figure 5. 207]. however. Suitable values for them should result in a natural way from the speciﬁc problem.8) . The basic idea of scalespaces. about the contrast of semantically important structures) into the evolution process. In both cases the number of local maxima is decreasing.g. both linear and nonlinear diﬀusion ﬁlters preserve the average grey level of the image.5(a) and (b) show the result under continuousscale dilation with a ﬂat discshaped structuring element and a quadratic structuring function. and that they usually do not preserve the average grey level. From Section 1. Besides these speciﬁc features of nonlinear diﬀusion scalespaces it should be mentioned that.9) for the quadratic structuring function. It should be noted that contrast enhancement is a local phenomenon which cannot be replaced by simple global rescalings of the grey value range. Nonlinear diﬀusion ﬁltering renounces this requirement by allowing to incorporate apriori information (e. The contrastenhancing quality distinguishes nonlinear diﬀusion ﬁlters from most other scalespaces. (5.5 and 5.5. EDGEENHANCING DIFFUSION 119 complete hierarchical family of segmentationlike images. due to the homogeneous Neumann boundary condition and the divergence form. with respect to the speciﬁed task.6. The fact that the maximum with the largest grey value will dominate at the end shows that these processes can be sensitive to noise (maxima might be caused by noise). is maintained: to provide a family of subsequently simpliﬁed versions of the original image. respectively. and maxima keep their location in scalespace until they disappear [206. In this sense.6 we know that their evolution equations are given by ∂t u = ∇u. it is generally not possible to obtain similar segmentationlike results by just rescaling the grey values from a scalespace which is only contrastreducing.
200. t = 0. EXAMPLES AND APPLICATIONS Figure 5. 236)2. 400. 12. 200.5. Top: Original image.120 CHAPTER 5. (c) Right Column: Nonlinear isotropic diﬀusion with the Charbonnier diﬀusivity (λ = 3). (a) Left Column: Linear diﬀusion. 70. top to bottom: t = 0. 150. (b) Middle Column: Inhomogeneous linear diﬀusion (λ = 8). 70. .3: Evolution of an MRI slice under diﬀerent PDEs. t = 0. 600. 50. Ω = (0.
σ = 4). 3000. . σ = 1). 25000. t = 0. (c) Right Column: Edgeenhancing anisotropic diﬀusion (λ = 3. 500000.4: Evolution of an MRI slice under diﬀerent PDEs. 250. t = 0. EDGEENHANCING DIFFUSION 121 Figure 5. 236)2.1. σ = 1). 1500. (a) Left Column: Isotropic nonlinear diﬀusion (λ = 3. Ω = (0. 40. 400. t = 0. 875. (b) Middle Column: Isotropic nonlinear diﬀusion (λ = 3.5. 7000000. Top: Original image.
25. 236)2. t = 0. t = 0. 4. 70.5: Evolution of an MRI slice under diﬀerent PDEs. 275. 1275. 1.122 CHAPTER 5. . (b) Middle Column: Dilation with a quadratic structuring function. EXAMPLES AND APPLICATIONS Figure 5. (a) Left Column: Dilation with a disc. 0. Top: Original image. 20. 10. Ω = (0. t = 0. (c) Right Column: Mean curvature motion. 4.
(a) Left Column: Aﬃne morphological scalespace. (b) Middle Column: Modiﬁed mean curvature motion (λ = 3. 50. t = 0. 236)2. 350. 20. 140. EDGEENHANCING DIFFUSION 123 Figure 5. 5000. (c) Right Column: Selfsnakes (λ = 3. Top: Original image. t = 0.5. Ω = (0. . t = 0. 100.1. 1500. 600. σ = 1). 40000. σ = 1).6: Evolution of an MRI slice under diﬀerent PDEs.
6 we have seen that processes of this type are called selfsnakes [357].32). while highcurved structures (e. This is at the expense of withdrawing morphology in terms of invariance under monotone greyscale transformations. The corresponding evolution is depicted in Fig. the nose) exhibit roundings at an earlier stage. As a consequence.90) depicted in Fig. EXAMPLES AND APPLICATIONS A completely diﬀerent morphological evolution is given by the mean curvature motion (1. ∇u (5. Since they diﬀer from isotropic nonlinear diﬀusion ﬁlters by the ∇u terms inside and outside the divergence expression.124 CHAPTER 5. Since MCM shrinks level lines with a velocity that is proportional to their curvature. for a comparable elimination of small structure.11) In Section 1.6.108) shown in Fig. We observe that structures remain much better localized than in the original MCM.6(a). 5. let us investigate the modiﬁed MCM ∂t u = g(∇uσ 2 ) ∇u div ∇u ∇u (5. see e. the advantage of having aﬃne invariance may counterbalance the correspondence problem in certain applications. As a simple prototype for this idea.6(b). On the other hand. This also explains its excellent noise elimination qualities. the chin.g. Nevertheless.g. but they are not designed to act contrastenhancing. the results appear less segmentationlike than those for nonlinear diﬀusion ﬁltering. Since it takes the time T = 3 s 3 to remove 4 1 all isolines within a circle of radius s – in contrast to T = 2 s2 for MCM – we see that. Using g(∇uσ 2 ) inside the divergence expression leads to ∂t u = ∇u div g(∇uσ 2 ) ∇u . One possibility is to use the damping function outside the divergence expression. 364. 5. they will not preserve the average . 304]. 365. the experiments give evidence that this process is probably not contrastenhancing. the head looks almost like a ball. Processes of this type are studied in [13. lowcurved object boundaries are less aﬀected by this process. the correspondence problem is more severe than for MCM. A similar behaviour can be observed for the aﬃne invariant morphological scale4 space (1. 5. This is in accordance with the theory which predicts convergence of all closed level lines to circular points. the shrinking eﬀect of large structures is stronger for AMSS than for MCM. Both MCM and AMSS give signiﬁcantly sharper edges than linear diﬀusion ﬁltering. After some time. Since the AMSS involves no additional parameters and oﬀers more invariances than other scalespaces. Thus.10) with g(∇uσ 2 ) as in (1. it is ideal for uncommitted image analysis and shape recognition.5(c). however. One possibility to reduce the correspondence problem of morphological scalespaces is to attenuate the curve evolution at highcontrast edges.
namely clouds and stripes.8(a) depicts a wood surface possessing one defect. equation (5. This behaviour which resembles MCM is not surprising if one compares (1.7: Preprocessing of a fabric image. one has to process the image in such a way that quality relevant features become better visible und unimportant structures disappear. For a suitable parameter choice. If one aims to automize this evaluation.6(c) indicates that g(∇uσ 2 ) gives similar edgeenhancing eﬀects as in a nonlinear diﬀusion ﬁlter. Anisotropic diﬀusion ﬁlters are capable of visualizing both qualityrelevant features simultaneously (Fig. Ω = (0. The quality of a fabric is determined by two criteria. σ = 2.4) can be applied with good success (Fig. whereas stripes are an anisotropic phenomenon caused by adjacent ﬁbres pointing in the same direction. Fig. but one can observe a stronger tendency to create circular structures. (b) Right: Anisotropic diﬀusion. grey value. λ = 4. EDGEENHANCING DIFFUSION 125 Figure 5. 5. 5. For furniture production it is of importance to classify the quality of wood surfaces.2. whereas stripes can be enhanced by a special nonlinear diﬀusion ﬁlter which is designed for closing interrupted lines and which shall be discussed in Section 5.7). 257)2. t = 240. Let us now study two applications of nonlinear diﬀusion ﬁltering in computer aided quality control (CAQ): the grading of fabrics and wood surfaces (see also [413]).1. Clouds result from isotropic inhomogeneities of the density distribution. they perform isotropic smoothing at clouds and diﬀuse in an anisotropic way along ﬁbres in order to enhance them. 5.5. . (a) Left: Fabric. To visualize this defect. However.120) with (1. 5.121). one can use a fast pyramid algorithm based on linear diﬀusion ﬁltering for the clouds [417]. if one wants to visualize both features separately.8(b)). The evolution in Fig.
which are given by the two white longitudinal objects in the centre. Ω = (0.9(d). it is intended to automize this feature extraction step by a segmentation algorithm. Figure 5. 5. Fig.12) It has been obtained by a MegaWave programme using a hierarchical region growing algorithm due to Koepﬂer et al. EXAMPLES AND APPLICATIONS Figure 5.9(b)) prior to segmenting it. σ = 2. (a) Left: Wood surface.58): Ef (u. 5. As is seen in Fig. For detecting Alzheimer’s disease one is interested in determining the ratio between the ventricle areas. In order to make the diagnosis more objective and reliable. 5.9(a) gives an example for possible medical applications of nonlinear diffusion ﬁltering as a preprocessing tool for segmentation (see also [415] for another example). and the entire head area. It depicts an MRI slice of the human head. one gets a better segmentation when processing the original image slightly by means of nonlinear diﬀusion ﬁltering (Fig. (5. In [413] it is demonstrated how a modiﬁed anisotropic diﬀusion process yields even more accurate results with less roundings at the corners. 256)2. (b) Right: Isotropic nonlinear diﬀusion. λ = 4. t = 2000.9(c) shows a segmentation according to the following simpliﬁcation of the Mumford– Shah functional (1.126 CHAPTER 5. . K) = Ω (u−f )2 dx + αK. [239].8: Defect detection in wood.
one should smooth preferably . t = 2. α = 8192. (b) Top Right: Diﬀusionﬁltered. COHERENCEENHANCING DIFFUSION 127 Figure 5. α = 8192.2. 256)2.1. (a) Top Left: Head.1 Coherenceenhancing diﬀusion Filter design In this section we shall investigate how the structure tensor information can be used to design anisotropic diﬀusion scalespaces which enhance the coherence of ﬂowlike textures [418].5. (d) Bottom Right: Segmented ﬁltered image.9: Preprocessing of an MRI slice. v2 the corresponding orthonormal eigenvectors.2. Let again µ1 . and v1 . This requires a nonvanishing integration scale ρ. µ2 with µ1 ≥ µ2 be the eigenvalues of Jρ . σ = 0. As in 5. (c) Bottom Left: Segmented original image. 5.5. If one wants to enhance coherent structures. Ω = (0.1 the diﬀusion tensor D(Jρ (∇uσ )) ought to possess the same set of eigenvectors as Jρ (∇uσ ).2 5. λ = 5.
σ = 0. σ = 5.128 CHAPTER 5. Ω = (0. σ = 0. EXAMPLES AND APPLICATIONS Figure 5. (a) Top Left: Original ﬁngerprint. (c) Bottom Left: Orientation of smoothed gradient. m ∈ IN): λ1 := α. ρ = 4. (d) Bottom Right: Structure tensor orientation.5. λ2 := α α + (1−α) exp −C (µ1−µ2 )2m if µ1 = µ2 . along the coherence direction v2 with a diﬀusivity λ2 which increases with respect to the coherence (µ1 −µ2 )2 . This may be achieved by the following choice for the eigenvalues of the diﬀusion tensor (C > 0.5. 256)2. (b) Top Right: Orientation of smoothed gradient. else.10: Local orientation in a ﬁngerprint image. where the exponential function was chosen to ensure the smoothness of D and the .
averages the gradient orientation instead of its direction.2 All examples below are calculated using C := 1. ﬁlters of this type are not regularizations of the Perona–Malik process: the limit σ → 0.1): ∂t u = uξξ = ∇u curv(u) (5. (b) Right: Coherenceenhancing anisotropic diﬀusion. We observe very high ﬂuctuations in the local orientation. Therefore. while vertical structures are represented in white. The structure tensor. ρ = 4.13) with ξ being the direction perpendicular to ∇u.10(d)). 1) keeps D(Jρ (∇uσ )) uniformly positive deﬁnite.2 Applications Figure 5. To illustrate how the result of anisotropic PDE methods depends on the direction in which they smooth. however.10(a).2. and α := 0.5. 5.5. t = 20. 5. (a) Left: Meancurvature motion. the results in (c) are much worse than in (b). 5.10(b)).10 illustrates the advantages of local orientation analysis by means of the structure tensor. 1. In order to detect the local orientation of the ﬁngerprint depicted in Fig. Since MCM smoothes by propagating level lines in inner normal direction we recognize that its smoothing diEvidently. ρ → 0 leads to a linear diﬀusion process with constant diﬀusivity α. small positive parameter α ∈ (0. t = 5. σ = 0.001. the gradient orientation of a slightly smoothed image has been calculated (Fig. let us recall the example of mean curvature motion (cf. m := 1.6. 2 .11: Anisotropic equations applied to the ﬁngerprint image. Horizontally oriented structures appear black.2. COHERENCEENHANCING DIFFUSION 129 Figure 5. 5. This is the reason for the reliable estimates of local orientation that can be obtained with this method (Fig. When applying a larger smoothing kernel it is clear that adjacent gradients having the same orientation but opposite direction cancel out.
130 CHAPTER 5. the locations of the semantically important singularities in the ﬁngerprint remain the same. . see Fig. although this method is in a complete anisotropic spirit.5.11(b). biases the diﬀusive ﬂux towards the coherence orientation v2 and is therefore wellsuited for closing interrupted lines in coherent ﬂowlike textures. 5. we should not expect it to be capable of closing interrupted linelike structures. The proposed anisotropic diﬀusion ﬁlter. (d) Bottom Right: t = 640. (b) Top Right: t = 20. however. ρ = 2). The results in Fig.11(a) conﬁrm this impression. 5. Due to its reduced diﬀusivity at noncoherent structures. This is an important prerequisite that any image processing method has to satisfy if it is to be applied to ﬁngerprint analysis.12: Scalespace behaviour of coherenceenhancing diﬀusion (σ = 0. rection depends exclusively on ∇u. Ω = (0. 257)2. (c) Bottom Left: t = 120. (a) Top Left: Original fabric image. Thus. EXAMPLES AND APPLICATIONS Figure 5.
13 illustrates the potential of CED for medical applications. a diﬃcult problem for the automatic grading of nonwovens [299]. There is evidence that the trabecular formation is for a great deal determined by the external load.7. COHERENCEENHANCING DIFFUSION 131 Figure 5.5. quantifying the rate of progression of rheumatism and osteoporosis. such as the placement or removal of metal implants. From Figure 5. the trabeculae. 5. 300) × (0.13(b). Examples are the control of recovery after surgical procedures. It depicts a human bone. the determination of leftright deviations of symmetry in the load or establishing optimal load corrections by physiotherapy. . 186).(c) we observe that CED is capable of enhancing the trabecular structures in order to ease their subsequent orientation analysis.2. Ω = (0. Figure 5. t = 16. ρ = 6. (b) Bottom Left: Filtered by coherenceenhancing anisotropic diﬀusion. σ = 0. Figure 5. Its internal structure has a distinctive texture through the presence of tiny elongated bony structural elements.5.13: (a) Top: High resolution slipring CT scan of a femural bone. (c) Bottom Right: Dito with t = 128. The temporal behaviour of this diﬀusion ﬁlter seems to be appropriate for visualizing coherent ﬁbre agglomerations (stripes) at diﬀerent scales.12 depicts the scalespace behaviour of coherenceenhancing anisotropic diﬀusion applied to the fabric image from Fig. For this reason the trabecular structure constitutes an important clinical parameter in orthopedics.
5.132 CHAPTER 5. e Paris. which are not typical texture images. 429].16 [160. 5. EXAMPLES AND APPLICATIONS Figure 5. Due to the established . 5.15.14: Image restoration using coherenceenhancing anisotropic diﬀusion. Interestingly. 5.15(d)). Muse´ d’Orsay). which indicates that this van Gogh painting possesses a uniform “texture scale” reﬂecting the characteristic painting style of the artist. 419]. all ﬁlter parameters are ﬁxed except for ρ. 275). To this end. It is depicted in Fig. Ω = (0.15(b) shows that a value for ρ which is too small does not lead to the visually dominant coherence orientation and creates structures with a lot of undesired ﬂuctuations. Increasing the value for ρ improves the image signiﬁcantly (Fig.5. e σ = 0.15(c)). We observe that the diﬀusion ﬁlter can close interrupted lines and enhance the ﬂowlike character which is typical for van Gogh paintings. (a) Left: “Selfportrait” by van Gogh (SaintR´my. 5. ρ = 4. Fig. but still reveal a ﬂowlike character. The next painting we are concerned with is called “Road with Cypress and Star” [162. In order to demonstrate the inﬂuence of the integration scale ρ. t = 6.14 shows the restoration properties of coherenceenhancing anisotropic diﬀusion when being applied to a selfportrait of the artist [161]. Fig. 1889. 215) × (0. 5. (b) Right: Filtered. we shall process impressionistic paintings by Vincent van Gogh. In a last example the temporal evolution of ﬂowlike images is illustrated by virtue of the “Starry Night” painting in Fig. Let us now investigate the impact of coherenceenhancing diﬀusion on images. a further increasing of ρ does hardly alter this result (Fig.
(c) Bottom Left: ρ = 4. (b) Top Right: o u Filtered with ρ = 1. Rijksmuseum Kr¨ller–M¨ ller). 1890. Otterlo.5. 203) × (0. COHERENCEENHANCING DIFFUSION 133 Figure 5. (a) Top Left: “Road with Cypress and Star” by van Gogh (AuverssurOise. 290). Ω = (0.5. t = 8).2. (d) Bottom Right: ρ = 6.15: Impact of the integration scale on coherenceenhancing anisotropic diﬀusion (σ = 0. .
a constant image with the same average grey value as the original one. 199).16: Scalespace properties of coherenceenhancing anisotropic diﬀusion (σ = 0. EXAMPLES AND APPLICATIONS Figure 5. however. is maintained for a very long time. scalespace properties. (c) Bottom Left: t = 64. Ω = (0. the image becomes gradually simpler in many aspects. New York.134 CHAPTER 5. before it ﬁnally will tend to its simplest representation. e 1889. The ﬂowlike character. The Museum of Modern Art). 255) × (0. (d) Bottom Right: t = 512. ρ = 4).5. (b) Top Right: t = 8. (a) Top Left: “Starry Night” by van Gogh (SaintR´my. .3 3 Results for AMSS ﬁltering of this image can be found in [305].
invariances. the goal of Chapters 2–5 has been to present a scalespace framework for nonlinear diﬀusion ﬁltering which does not require any monotony assumption (comparison principle). and convergence to a constant steadystate have been established. besides the fact that many global smoothing scalespace properties are maintained. the prerequisites in all three settings reveal many similarities and. semidiscrete and discrete setting. Rather than deducing a unique equation from ﬁrst principles we have analysed wellposedness and scalespace properties of a general family of regularized anisotropic diﬀusion ﬁlters. For this reason it is possible to design scalespaces with restoration properties giving segmentationlike results. for instance concerning size and contrast of especially interesting features. informationreducing transformations. continuous dependence of the solution on the initial image. representatives of the semidiscrete class can be obtained by suitable spatial discretizations of the continuous class. The degree of freedom within the proposed class of ﬁlters can be used to tailor the ﬁlters towards speciﬁc restoration tasks. as a consequence. Lyapunov functionals. maximum–minimum principles. Prerequisites have been stated under which one can prove wellposedness and scalespace results in the continuous. Existence and uniqueness results. Therefore. these scalespaces do not need to be uncommitted. These global smoothing properties do not contradict seemingly opposite local eﬀects such as edge enhancement.Chapter 6 Conclusions and perspectives While Chapter 1 has given a general overview of PDEbased smoothing and restoration methods. 135 . On the other hand. new possibilities with respect to image restoration appear. The large class of Lyapunov functionals permits to regard these ﬁlters in many ways as simplifying. they give the user the liberty to incorporate apriori knowledge. We have seen that. Each of these frameworks is selfcontained and does not require the others. while representatives of the discrete class may arise from time discretizations of semidiscrete ﬁlters.
• The various possibilities to include semilocal or global information constitute another future perspective. . The scalespace stack of these ﬁlters appears to be wellsuited to extract semantically important information with respect to a speciﬁed task. • In contrast to the linear diﬀusion case.136 CHAPTER 6. the general availability of aﬀordable colour scanners and printers. the fact that these ﬁlters are steered by the structure tensor instead of the regularized gradient allows to adapt them to more sophisticated tasks such as the enhancement of coherent ﬂowlike structures. This ﬁeld oﬀers a lot of challenging mathematical questions. In view of these results. It is a ﬂexible and mathematically sound class of methods which ties the advantages of two worlds: scalespace analysis and image restoration. however. we can expect a lot of new results in the near future. anisotropic diﬀusion deserves to be regarded as much more than an adhoc strategy for transforming a degraded image into a more pleasant looking one. This could lead to speciﬁcally tuned ﬁlters for topics such as perceptual grouping. Gabor ﬁlters. New ﬁlter models might arise using other structure descriptors than the (regularized) gradient or the structure tensor. are likely to consist of straightforward extensions of topics presented in this text: • While the theory and the examples in the present book focus on 2D greyscale images. CONCLUSIONS AND PERSPECTIVES The analysed class comprises linear diﬀusion ﬁltering and the nonlinear isotropic model of Catt´ et al. that nonlinear diﬀusion ﬁltering is a young ﬁeld which has certainly not reached its ﬁnal state yet. it would certainly be useful to better understand the deep structure in nonlinear diﬀusion processes. Some of its future developments. but also novel ape proaches have been proposed: The use of diﬀusion tensors instead of scalarvalued diﬀusivities puts us in a position to design real anisotropic diﬀusion processes which may reveal advantages at noisy edges. Coherenceenhancing anisotropic diﬀusion is only a ﬁrst step in this direction. Some of the references in Chapter 1 point in directions how this can be accomplished. the relation between structures at diﬀerent scales has rarely been exploited in the nonlinear context. Although this problem is less severe. since the avoidance of correspondence problems was one of the key motivations to study nonlinear scalespaces. Interesting candidates could be wavelets. Last but not least. Thus. or steerable ﬁlters. [81] and Whitaker and Pizer [438]. It is clear. it is evident that most of its results can easily be generalized to higher dimensions and vectorvalued images. and the wish to integrate information from diﬀerent channels. The need for such extensions grows with the rapid progress in the development of faster computers. however.
there still remains a lot of work to be done. 444]. for instance a physician who wants to denoise ultrasound images. Thus. In this context it would be helpful to have software packages. Novel. more research on ﬁnding some guidelines for automatic parameter determination for a task at hand would encourage also people without a speciﬁc image processing background to apply nonlinear diﬀusion ﬁlters. Especially combinations with concepts such as data compression. where diﬀerent nonlinear diﬀusion ﬁlters are implemented in an eﬃcient way. may be frightened by this perspective. Since these parameters have a rather natural meaning. • The price one has to pay for the ﬂexibility of nonlinear diﬀusion ﬁltering is the speciﬁcation of some parameters. 431. and which are easy to use for everyone. splitting and multigrid techniques. tomographic reconstruction techniques. or grid adaptation strategies. Several useful suggestions for parameter adaptation can already be found in [328. Somebody with another primary interest. or neural networks for learning apriori information might lead to novel application areas for these techniques. 270.137 • Most people working in computer vision do not have a speciﬁc knowledge on numerical methods for PDEs. Thus. but ineﬃcient explicit (Euler forward) schemes. • There are not yet many studies which explore the potential of nonlinear diffusion ﬁltering when being combined with other image processing techniques. It would be nice if this book has inspired its readers to contribute to the solution of some of the remaining open problems. 36. this is not a very diﬃcult problem for someone with experience in computer vision. As a consequence. . more timecritical application areas could be explored by applying implicit schemes. the most widelyused numerical methods for nonlinear diﬀusion ﬁltering are still the simple. segmentation algorithms.
CONCLUSIONS AND PERSPECTIVES .138 CHAPTER 6.
1993. A. Berlin. Los Alamitos. F. wavelets and PDEs. Vol. 1994. Vol. ICAOS ’96: Images. Vol. Florack. 200–207. [12] L. May 26–29. Acton. Deriche. J. 12th Int. Rational Mech. Image Processing (ICIP–96.).. Mokhtarian.. M. Vol. J. Springer. 16–19. 219. [6] S. Reliable classiﬁcation of chrysanthemum leaves through curvature scale space. Geoscience and Remote Sensing Symposium (IGARSS ’92. Vol. [11] L. Proc. A. 1993). 1252. 1992. Vol. Koenderink. 1996). IEEE Computer Society Press. Anal.. Guichard. 3. Abbasi.Bibliography [1] S. M. Vol. Numer. Arch.R. Academic Press. Proc. Gagalowicz.A. Alvarez. Math. 118. 284–295. 1217–1229. [3] A. 1975. Comput. Conf. Houston. 1996. C. 478–482. Viergever (Eds. ter Haar Romeny. I. Image quantization using reaction–diﬀusion equations. [14] L. L. Nov. Kittler. [2] R. [13] L. 845–847. Fourth Int. Conf. Acar. Adams. 139 .A. SIAM J. Springer. Acton. Alvarez.M. Bovik. F. J. 590–605. Sobolev spaces. 1994). 199–257. [4] S. [8] D. 1994. 123.T. Lecture Notes in Control and Information Sciences. Morel.T. M. Vol. Morel. Discrete models for energyminimizing segmentation. A mean ﬁeld solution to anisotropic edge detection of remotely sensed data. 13–16.M. J. P. [7] R. Anisotropic diﬀusion pyramids for image segmentation. Lecture Notes in Computer Science. 10. 1992. Images and PDE’s. Signal and image restoration using shock ﬁlters and anisotropic diﬀusion. 1992. 3–14. 1994. Ackah–Miezan. Adalsteinsson. Alvarez. London.M. 1992). IEEE Int. Springer. Edge enhancement of infrared imagery by way of the anisotropic diﬀusion pyramid. Alt. on Computer Vision (ICCV ’93. Anal. Inverse Problems. Alvarez.T. 1997. [9] H. Lions.. A fast level set method for propagating interfaces. Vol. 865–868. SIAM J. New York. Proc. Lineare Funktionalanalysis. J. Vol. J. J. Image Processing (ICIP–94. Image selective smoothing and edge detection by nonlinear diﬀusion.. [5] S. J. IEEE Computer Society Press. Scalespace theory in computer vision. 269–277. Conf. [10] L. Los Alamitos. P. II. Sept. Vogel Analysis of bounded variation penalty methods for illposed problems. Herlin. Esclarin. Jaﬀr´. 1997. Lausanne. Vol. Proc. B. M. Crawford. May 11–14.M.O.M. 31. 1996. Mazorra.).L. 57.L. 1995. 845–866. Anal. Morel e (Eds. 1. Appl. Berger. J. 1993. 2. SIAM J. Austin. 29. R. Acton. Crawford.W. Phys. L. IEEE Int. Alvarez. Lions. Berlin. Axioms and fundamental equations in image processing. Numer.C. Sethian. Berlin. 153–175.
M. [25] A. A compactness theorem for a special class of functions of bounded variations. 3B. W. 1995. 35017 Las Palmas. Anisotropic ﬁltering of MRI data based upon image s gradient histogram. Angenent. Comput. Scuola Normale Superiore. to appear in Int. Arridge. Italy. Simmons. Ital. Ambrosio. Scuola Normale Superiore.O. I. 1986. . 999–1036. Duda.). Vol. 133. A. Sporring. Ambrosio. Math. Koenderink.. Springer. Ital. J. (EMB ’96. N. Pure Appl. Vol. Berlin. Nielsen. Multispectral probabilistic diﬀusion using Bayesian classiﬁcation.140 BIBLIOGRAPHY [15] L. ˚str¨m. Uniqueness of the Gaussian kernel for scale space ﬁltering. 1994. ter Haar Romeny. J. Vincent. M. Annals of Mathematics. D.. 1995. Kimia. Annals of Mathematics. M. Proc. V. [26] S. Angenent. 1993). Boll. 26–33. [21] L. Fourth Int. 129–136. 56126 Pisa. 1–59. 56126 Pisa. Parabolic equations on curves and surfaces. 857–881. L. Preprint CVGMT96040105. 451–483.. Scalespace theory in computer vision. Pattern Anal. Parabolic equations on curves and surfaces. M. Sr´mek. IEEE Trans. Stochastic analysis of image acquisition and scalespace smoothing.. CDROM document N21. [19] L. Pallara. Piazza dei Cavalieri. Aﬃne morphological multiscale analysis of corners and multiple junctions. Departamento de Inform´tica y Sistemas. Partial regularity of free discontinuity sets I. Bajla. Universidad de Las a Palmas. Intersections. Berlin. 56126 Pisa. [28] I. Mathematical morphology: The Hamilton–Jacobi connection. Preprint CVGMT96040106. D. 1993.G. Ambrosio. Acta Numerica. Gaussian scalespace theory. submitted to Annali della Scuola Normale Superiore di Pisa. Kluwer. Florack. Fusco. 1993. Ambrosio. R.. F. Piazza dei Cavalieri. Amsterdam. 1990. Pallara. Oct. Morel. Vision. Babaud.). 7. [20] L. Scuola Normale Superiore. Witkin. Part II. [27] K. 132. 1992. Baudin. 8. Morales. M. Vol. 1995. Math. and generalized solutions. ˇ a [29] I. IEEE Computer Society Press. A o J. 3. on Computer Vision (ICCV ’93. Pallara. Partial regularity of free discontinuity sets II. Math. Berlin. B. D. J. 224–235. Chetverikov. 1990. Approximation of functionals depending on jumps by elliptic functionals via Γconvergence. 719. Vol. Kropatsch (Eds. 1252. P. Higher regularity of solutions of free discontinuity problems. Dept. 90–97. Intell. Mach. 1990. Los Alamitos. Arehart. Florack. 43. May 11–14. Ambrosio. 1997. Lecture Notes in Computer Science. Spain. Dordrecht. [16] L. Alvarez. blowup. Maruˇiak. Formalization and computational aspects of image analysis. 18th Annual Int.M. Tortorelli. Boll. Vol. N. Italy. Vol. Part I. Geometrydriven diﬀusion smoothing of the MR brain images using a a novel variable conductance. Dept. Comm. Conf. Conf.R. [17] L. Preprint CVGMT96040107. Johansen (Eds. 1991. Dept. Math. 105–123. Tortorelli. Alvarez. Holl¨nder. L. Vol.). 1997. [30] J. Report No. [22] L. Vol. [24] S. Fusco. Un. [18] L. Piazza dei Cavalieri. Bajla. submitted to Annali della Scuola Normale Superiore di Pisa.B. Viergever (Eds. Computer analysis of images and patterns.. Lecture Notes in Computer Science. Math. 1996). Italy.. Campus Universitario de Taﬁra.P. [23] S. 1994. Springer. Curves with pintegrable curvature. Proc. Un. Math. On the approximation of functionals depending on jumps by elliptic functionals. L. Heyden.B. 171–215. V. 31 – Nov. B. Ambrosio. A. A. D. of the IEEE Engineering in Medicine and Biology Soc. 9402. 215–219.
Appl. Proc. 474–477.O. Vol. Marimont. a Deriche. Vol. 1991. Dal Passo. G. 1923. 1996. 1987. Image Processing (ICIP–95. London. Benhamouda. Kleinman. Vorlesungen uber Diﬀerentialgeometrie II. van den Berg. Herlin. Scalespace theory in computer vision. U. Conf. [38] F. e Lecture Notes in Control and Information Sciences. Springer. Springer. Vol. Vol. Dept. L5–L10. 1994. A coarsening ﬁnite element strategy in image selective smoothing. Wiklund. 56. University of Illes Baleares.). 1995). master thesis. Physics Express Letters. Edge focusing. B¨nsch. [42] A. D. 1996. Aﬃne invariant segmentation by variational a method. Robust anisotropic diﬀusion: connections between robust statistics. CA 90024. Sapiro. P. Jaﬀr´. [32] C. Intell. 1252. 775–790. Texture segmentation by variational methods. Black. M. G. Dept. Aubert. Springer. Big¨ n. Berlin. Caselles. Germany. Washington. 24. 1996. 1997. University of Freiburg. Bergholm. Barenblatt. J.F. A degenerate pseudoparabolic regularization of a nonlinear forward–backward heat equation arising in the theory of heat and mass exchange in stably stratiﬁed turbulent shear ﬂow. Ballester. A total variation enhanced modiﬁed gradient algorithm for proﬁle reconstruction. Nonlinear image processing: e modelling and fast algorithm for regularization with edge detection. Oct. Florack.BIBLIOGRAPHY 141 [31] C. of Mathematics. Viergever (Eds.. IEEE Trans.). Anal. Palma de Mallorca. Koenderink. and anisotropic diﬀusion. D. J.. 1414–1439. University of Kaiserslautern. Mikula. Big¨ n. Faculty of Mathematics. a Preprint No. Vol. A. . Math. Barlaud. L. Ph.). 323–326. ICAOS ’96: Images. P. to appear in Computation and Visualization in Science.S. Ballester. 1995. of Mathematics. Multidimensional orientation estimation with apu plications to texture analysis and optical ﬂow. M. I. T. V. IEEE Int. Proc. Heeger. June 8–11. Visual reconstruction. [37] P. [41] M. Image Proc. Conf. Parameter adaptation for nonlinear diﬀusion in image processing. [35] G. M. M.M.M. [43] L. Berlin. Blaschke. Intell. [40] J. 726–741. Spain. Charbonnier.A.. wavelets and PDEs. [44] W. IEEE Trans. 18/1996. Pattern Anal. M. R. Blomgren. of California. Zisserman. 67653 Kaiserslautern. thesis. [39] J. Box 3049. SIAM J. M. line processing. Cambridge (Mass. 79104 Freiburg. Inverse Problems. Blanc–F´raud. R. on Computer Vision (ICCV ’87.I.. 1996. Univ. Pattern Anal. Lecture Notes in Computer Science.H. 187–193. SIAM J. First u Int. pp. to appear in IEEE Trans.. 11.J. Mach. 9. B. Gonz´lez. 433–438. London. 1987. Ughi. 22–25. 27. 219. 1987. Ballester Nicolau. K. R. IEEE Computer Society Press. ter Haar Romeny. 13. 1995. [33] C. Los Angeles. Granlund. Blake. [34] E. 294–325. Bertsch. Vol. G. [36] B. An aﬃne invariant model for image segmentation: Mathematical analysis and applications. Optimal orientation detection of linear symmetry. G. Gonz´lez. 1987).E. Vol. J. 1.H. Berger. Germany.D. Vol.O. Color TV: total variation methods for restoration of vector valued images. M. 1995. 1993. J. Morel (Eds. MIT Press. Math. ¨ [45] P. Mach. CAM Report 965. Washington. Granlund. Chan. Hebelstr.
Paris. Brauer.J. [62] A. A. Sapiro. Philadelphia. IEEE Southwest Symposium on Image Analysis and Interpetation (SSIAI ’96. Ph. Luk (Ed.F. 16. 1967. The morphological structure of images: The differential equations of morphological scalespace. [55] J. Bruckstein. van den Boomgaard. 1992. 362–380.. architectures. Bruckstein. . G. 1967. Mathematical morphology: Extensions towards computer vision. 1994. Paris. Vol. Morphology. Acad.D. 219–236. 1997. 18. 1964. R.A.L. [47] H. Intell. 1995. Imag. SPIE Vol. Charact´risation des minima globaux de la fonctionelle du Mumford–Shah en e e segmentation d’images. 349–367. [59] W. Comp.). Masson. Math. Clarendon. [65] M. van den Boomgaard. C. Vol. Blomgren. S´rie I.M. [63] B. [49] A. New monotone type approximations for elliptic problems. Maragos. Briggs. S. [53] R. [51] R. Fast ﬁlter transforms for image processing. 1995. 3162. Proc. R.). Artif. 3377–3386. D.T. Vol. Target tracking using the anisotropic diﬀusion pyramid. P. Extensions to total variation denoising. Brockett. 1275–1279. Blum. thesis. Vision. Burt. Op´rateurs maximaux monotones et semigroupes de contractions dans les ese paces de Hilbert. T. IEEE Trans.W. Sci. [48] A. V. A transformation for extracting new descriptors of shape. 1996). IEEE Trans. 1101–1113.. Bonnet. Princeton. 9. Sci. 42. Evolution equations for continuousscale morphological ﬁltering. Vol. Shaked.D. Sur la r´gularit´ des bords des minima de la fonctionelle de Mumford–Shah. J. oral presentation. Computer Graphics and Image Processing. [58] H. 20–51. D. 1994. F. Wathen–Dunn (Ed.H.E. 991–1014. April 8–9. t. Vierde Serie. The morphological equivalent of the Gauss convolution. to appear. Ph. Acad. Evolution of planar polygons. S´rie I. 7. Vol. Burton.M. 1992. van den Boomgaard. 1992. 321. 1996. e [50] R. J. Macaulay (Eds. May 10–13.142 BIBLIOGRAPHY [46] P. 1973. Benjamin. [61] A. SIAM. Deel 10. [54] K. and implementations VII. [56] F. Mach. 1997.I. Pattern Recogn. Amsterdam. Models for the perception of speech and visual form. 1991. Int. The motion of a surface by its mean curvature. Brakke. On projective invariant smoothing and evolution of planar curves and polygons. Brezis. e e C. 1995. Mulet. Press. North Holland. Brezis. Cambridge. Ordinary diﬀerential equations: A ﬁrst course. [64] P. Copenhagen. Vol. Chan. Shaked. Math.T. P. 1987. van den Boomgaard. [60] R. Bonnet. Nohel. 16. A multigrid tutorial. Paris. Smeulders. Maximum entropy in action. Princeton University Press. 1978. Acton. W. School on Classical ScaleSpace Theory. t.. [52] R.A. [57] H. New York. Signal Processing. University of Amsterdam. Advanced signal processing: algorithms. San Antonio. Intell. Hubbard. Oxford. B. 225–240. J. Bramble. 1981. M. 1121–1126. Analyse fonctionelle.). 321. The Netherlands. Nieuw Archief Voor Wiskunde. Buck. Pattern Anal.
Math. Mathematical morphology and its applications to image and signal processing. I. Image selective smoothing and edge detection e by nonlinear diﬀusion. [74] V. Caselles. London. R. Braunschweig. on Artiﬁcial Intelligence (IJCAI ’89. SIAM J. Sapiro. Comput. Vol. Adaptive image smoothing respecting feature directions. [82] A.). F. Sapiro. Vol. Detroit. 1. Vieweg. Caselles. ICAOS ’96: e Images. Proc. B. Butt (Eds. Vol. Butt.). Ferretti. 1996. 1585–1590. Diﬀerentialgeometrie von Kurven und Fl¨chen. Coll. Math. [77] V.A. 61–79. 1199–1246. R. Numer.D. 1988. Jaﬀr´. 2445–2456.A. Pattern Recognition..S. Pyramidal architectures for computer vision. Maragos. [75] V.A. SIAM J. Conf.. [72] R. P. G. Intell.M.). Catt´. J.M. 20–25. J. Anal. 33. Numer. Koenderink. 1989). Aug. ter Haar Romeny. 1993. Vol. Pattern Anal. B. Canny. 29. [78] V. Morel (Eds. 182–193. 301. Chambolle. J.. 16–20. Anal. Image Processing (ICIP–94. B. Koepﬂer. Kluwer. Caselles. Schafer.L. Dibos. 1895–1909. M. 219. Caselles. Coll. 1994. Proc. Princeton. 2. A geometric model for active contours in image e processing. 66. 8. Coll. Austin.. Vol. [67] L. Int. Vol. 1996.. [69] J. Zhong. Dept. T. Catt´. What is the best causal scalespace for 3D images?. F. Geodesic active contours.. J. Coll. Math. a 1983. C. Plenum Press. Scalespace versus topographic map for natural images. IEEE Trans.). 32. 22. Morel. Lecture Notes in Computer Science. F. Springer. [81] F. A computational approach to edge detection. 423–451. Lecture Notes in Computer Science. Vol. Comparison of multiscale morphology approaches: PDE implemented via curve evolution versus Chamfer distance transform. [70] V. T. 1996.. 1994. Berlin. 1994). Berger. Kimmel. Mach. Maragos. Numer. New York. 597–605. . Scalespace theory in computer vision. IEEE Computer Society Press. 77. Springer. 1996. Minimal surfaces: a geometric three dimensional segmentation approach. S. Herlin. Convergence of iterated aﬃne and morphological ﬁlters by nonlinear semigroup e theory. Preprint. [79] F. Dibos. 31–40. 13–16. J. Joint Conf. Florack. Vol. M. 1989.M. G. Some notes on repeated averaging smoothing. Lions. 56. [76] V. Vol. Caselles. wavelets and PDEs. Kimmel.O. 125–133. L. G. Image Proc. NJ 08544–5263. Cai. J. Vision.. Vol. 1252. P. Vol.W. Numer.D. 1995. Vol. Cai.. P. R. 1997. R. SIAM J. Berlin. Nov. of Civil Engineering. U. Vol. Springer. 11th Int. Appl. C. Numer. Anal. A morphological scheme for mean curvature motion and e applications to anisotropic diﬀusion and motion of level sets. 679–698. submitted to IEEE Trans. [80] F. SIAM J. 1997. 1986. Cantoni. Catt´. Kittler (Ed. M. 1992. Los Alamitos. Deriche. Viergever (Eds. M. Morel. do Carmo. [71] M. 1–31. 29–49. Snakes in movement. 1996. Carmona. 1997. [68] L. Boston.P. [73] V. Caselles. Sbert. A “small leakage” model for diﬀusion smoothing of image data. Princeton University. Catt´.BIBLIOGRAPHY 143 [66] M. Sbert. Partial diﬀerential equations and image processing. Lecture Notes in Control and Information Sciences. IEEE Int. J.
Cohignac. Comput. U. 1994. of California. J. London. Cohignac. Lopez. Proc. FL 32611. S. S.S. [88] P. [95] T. Integral and local aﬃne invariant parameter and application to shape recognition.A. Nov. Computing minimal surfaces via level set curvature ﬂow. IMACS Press. Guichard. R. Charbonnier. Math. 2567. 1995. 167–188.K. Conf. F.S. Pattern Recognition (ICPR 12. Vol. CVGIP: Image Understanding. Jerusalem. 1991 (see also J. A.). 75775 Paris Cedex 16. Deriche. IEEE Int. [84] R. Vol. Wong.. On active contour models and balloons. 1993). Chan. 53. 1994. Chan. Los Angeles. diﬀusion and resonance. Chopp. Los Alamitos. G. P. Luk (Ed. Mulet. [85] T. Uniqueness and existence of viscosity solutions of generalized mean curvature ﬂow equations. Wavelets: A tutorial in theory and applications. Lions. IEEE Computer Society Press.M. Chen. Chan. Image Processing (ICIP–94. [91] D.L..F. thesis LBL30685. Chambolle. 241–252.144 BIBLIOGRAPHY [83] A. Herlin. Physics Division. Investigative and trial image processing.F.M. [90] Y. Morel. 2. J. 76. . Lawrence Berkeley Laboratory. 1992. Morel (Eds. 106. S.H. Image Proc. Vol. University of Florida. CEREMADE. A nonlinear primal–dual method for totalvariation based image restoration.).A. Total variation blind deconvolution.M. Springer. L.K. Chan. Vol. 311–329. U. Two deterministic halfe quadratic regularization algorithms for computed imaging. J. 428–456. Goto. Gainesville. Proc.F. H. 314–325.. SPIE Vol. [96] T.O. Vol. Proc. 1992. CAM Report 9645. Numer. Diﬀerential Geometry.I. J.L.A.K. T. M. Chen.D. Eve. 749–786. Advanced signal processing algorithms. Austin. Wong. 2563. Vemuri. 12th Int.A. M. 36. e ICAOS ’96: Images.H. to appear in IEEE Trans. Lopez. Rudin. Jaﬀr´. Univ. Second IMACS Symposium on Iterative Methods. 77–91.M. 219. F.. Conf. Perception and Psychophysics. New York. 33. Cosine transform based preconditioners for total variation minimization problems in image processing.M. Vol. 1995. CA 94720. Dept.K. submitted to SIAM J. Wang. F. P. 214–222. Y. C. Scalespace and aﬃne invariant recognition of occluded shapes. Morel. 1991. Chan.). Morel. Blanc–F´raud. of Mathematics. Appl. [94] M. 1997..F. [89] Y. Vol. Grossberg. L. [86] T. 9254. Vassilevski (Eds. Chan. C. Continuation method for total variation minimizing problems in image processing.C. B. Place du Mar´chal de Lattre de Tassigny. 1984. SPIE Vol. P. 1994). e France. [97] T. Math.. Cohignac. Oct. Cohen. Chui. [92] C. 164– 168. Report UFCISETR97015. [93] L. G. 1994). 9–13. U. 168–172. C. Image denoising and segmentation via nonlinear diffusion. R. 1995. Zhou. C. Berkeley. CISE Dept.H. 1996. Berger. 211–218. CA 90024. Bramble (Eds. S. I. Numerical analysis of the fundamental equation of image processing. Vol. [87] T. Image recovery via total variation minimization and related problems. wavelets and PDEs. boundaries. Margenov. Lecture Notes in Control and Information Sciences. 1991. Golub.). Phys. Ph. 1997. J. Los Alamitos. Preprint No. Giga. IEEE Computer Society Press.T. Barlaud. Academic Press. 1996. University of California.D.S. Cohen. L. Neural dynamics of brightness perception: Features. J.G. 13–16. Universit´ e Paris IX – Dauphine. Aubert.
El Ayyadi. 1995. 108. Dobson. Second Int. 1989. 7–11. 290–295. 1996). P. Teubner. 115. Numer. Bull. A. Proc. Inverse Problems. Vol. Appl. [103] M. Recovery of blocky images from noisy and blurred data. 1995. Aug.. 1996. Ong S. [107] G. e [113] F. Rat. Germain. Vol. 1995. Conf. Dibos.C.H. (Eds.H. and Image Processing. User’s guide to viscosity solutions of second order partial diﬀerential equations. [116] D. [105] J. 1131–1139. Lausanne. 1995. 1992. Sept. 61. Computer Vision (ACCV ’95. Proc. Computer Vision. David. Paris. 1996. Convergent diﬀerence schemes for nonlinear parabolic equations and mean curvature motion. [101] G. Vol. Cottet. M. Crandall. P. Analyse multi´chelle invariante par transformations projectives. [104] M. Diﬀusion approximation on neural networks and applications for image processing. Opticalﬂow estimation while preserving its discontinuities: A variational approach. Vol.H. Image Processing (ICIP–96. Stuttgart. Kornprobst.. 1992. Vol. C 1 arcs for minimizers of the Mumford–Shah functional. G. IEEE Int. Faugeras. Vol. 1–67. Dobson. 56. 1994. Am. Traitement du Signal. on Mathematics in Industry. Cottet. Di Zenzo. Vol. Singapore Conf. Comp. 403–406. [111] R.G. J. Crandall. [106] I. R. S. Vol. Image processing through reaction combined with nonlinear diﬀusion. A note on the gradient of a multiimage. 659–673. Les EDP en traitement des images et vision par ordinateur. Math.C. P. De Giorgi. 783–888. 56.H. Anal. [110] R. Proc. C. Ishii. 1996).. Math. Leaci. Conf. Local Morse theory for solutions to the heat equation and Gaussian blurring. Appl. 195–218. 1996. Mech. [99] G. e Sci. O.. 1181–1198.L. J. (Singapore. Sept.). 1992). [115] D. Singapore.H. 1992. Image Processing (ICIP–96. 1995).D. Sept. Santosa. F. Ma. No. 1993. 1997. Cottet.G. Lions. 1. Ang Y. Existence theorem for a minimum problem with a free discontinuity set. Damon. Soc. [109] R. Carriero. 1996.. 6. [102] G. Vol. Acad. Hodnett (Ed. Dibos.). 263–267. 1. [114] S. Cottet. Vol. 1986. 1996). 116–125. Santosa. Lausanne. 317–334. Recursively implementing the Gaussian and its derivatives. Math.. Deriche. Nonlinear PDE operators with memory terms for image processing. Srinivasan. Proc. Math. 917–921. 33. 16–19. Deriche. 3. Proc. SIAM. An image enhancement technique for electrical impedance tomography.BIBLIOGRAPHY 145 [98] G. Projective invariant multiscale analysis. Neural networks: Continuous approach and applications to image processing. Aubert. Vol. 25–30. Conf. F. 10.. . Proc. Vol. Deriche. [112] F. 2. Arch. Diﬀerential Equations. IEEE Int. Vol. Vol.H. L. Daubechies. SIAM J. 320. Pattern Recognition (ICPR 13. Vol. 3–9. Ten lectures on wavelets. on Image Proc. SIAM J. Nonlinear diﬀusion for early vision. 13th Int. Biological Systems.L. F. t. [100] G. 13. 481–483. 485–488. Vienna. 27. Graphics. 17–41. Cong. Sixth European Conf. 16–19. [108] E. Second Asian Conf. M. Lions. Philadelphia. 75. 1996. 1996. 368–401. Math. December 5–8. V. S´rie I. A. 1992. I.
1994. Montana State University. Faugeras. 1991. Jaﬀr´. M. Part I.. . Oxford. Anal.G. Jones. 371– 391. Lecture Notes in Computer Science. Diﬀerential Geometry. A diﬀerential geometric approach to anisotropic diﬀusion. Math. Stevenson (Eds. Berlin. S. of Mathematical Sciences. Convergence of an iterative method for total variation denoising. I. [121] J. C.E. ElFallah. 11–46. Analysis of regularized total variation penalty methods for denoising. B. Bozeman. Ford. Evans. Comput. Dept.). 1994. R. 1963. 1995).L. [118] D. Eidelman. Conf. IEEE Int. Preprint.F. [133] O. 1996. 1993. Lecture e Notes in Control and Information Sciences.O. Faugeras. Springer. Reid. Chenke. Vol. Springer. Applications of invariance in computer vision. 219. 1996. 601–617. Workshop on Geometrical Modelling and Invariants for Computer Vision (Xi’an. Neubauer.). [130] O. Math. 23. G. Morel (Eds. 1991. Scalespaces and aﬃne curvature. G. Keriven. 825. Faugeras. [123] S. [119] L. Herlin.. Dordrecht. Soc. Dobson. Eberly. Forsyth (Eds.S. Vol. 13–16. A. 38. 6. 2182.M. [128] H. ElFallah. D. Sur l’´volution de courbes simples du plan projectif r´el.. Inverse Problems. Dzhu Magazieva. O. Proc. Duﬀ. Hanke. MT 597170240. J. A. ICAOS ’96: Images. B. Maths. Numer. Faugeras. The evolution of mean curvature in image ﬁltering.S. Zisserman. Vol. M. Numer. 12. [127] A. 195–204.E. U. Spruck. Dobson. R. 1994. 1994. U. Acad. Motion of level sets by mean curvature I. 4. 11. Sci. Grossmann. On predictor–corrector methods for nonlinear parabolic diﬀerential equations. Appl. [126] A. S´rie I. SIAM J. Vol. Regularization of inverse problems. Los Alamitos. 428–453. R.146 BIBLIOGRAPHY [117] D. Phys. A. 45–49. Clarendon.. Vol. 298–302. Vogel.K. Vol. 1983.). 439–450.). Xidian University Press. wavelets and PDEs.S. Mundy. Geometrydriven diﬀusion in computer vision. 1994).. e [131] O. J. J. Image and video processing II.K. London. ter Haar Romeny (Ed. R. Rajala. A. R.C.M. Douglas. Erisman. Friedman. 565–570. J. [132] O. Vol. 317. Nonlinear signal processing using integration of ﬂuid dynamics equations. Kluwer. 79–98.A. 33. 1567. 49–63. Berger. Parabolic and hyperbolic problems. Deriche. e e Paris. Math. 1994. van den Boomgaard. Vol. Mohr (Eds. On projective plane curve evolution. R. Ford. Nov. Douglas. Dorst.A. A. W.R. Austin.L.W. R. 1964. C. A general formulation of alternating direction methods. Tescher (Ed. Scherzer. J. 1. Signal Processing. IEEE Computer Society Press.C. Proc. Morphological signal processing and the slope transform. SPIE Vol. Xi’an.S.). [122] I. 1995. Kluwer. April 27–30.E. Dordrecht. [120] J. J. 66–73. J. No. Cartan’s moving frame method and its application to the geometry and evolution of curves in the Euclidean. 1996. SPIE Vol.I. Engl. aﬃne and projective planes. Numerical study of a partial diﬀerential equation. Keriven. Nonlinear adaptive image ﬁltering based on inhomogeneous diﬀusion and diﬀerential geometry. t.): Applications of digital image processing XIV. Image Processing (ICIP–94. Gunn. [124] D.. Ind. Vol. 1986. W. 1995.I.M. 635–681. to appear.R. Direct methods for sparse systems. [125] S. [129] L. J.
Franke. [141] L. 16–19. 18. Viergever (Eds.A.R. Comput.J. Geometrydriven curve evolution. University of Kaiserslautern. Proc. 1997.E. J. 800. Conf. Faugeras. Math. Vol. Berlin. 1252.J. Vienna.M. Florack. 1992. B. Berlin. Ford. Koenderink. Image processing using a wavelet algorithm for nonlinear diﬀuo sion. Vol. 1996). Salden. [142] L. L.A. Vol.). D. ISPRS Intercommission Conf. J. P. A fast operator for detection and precise location of distinct o u points. Laboratory Investigation. Vol. Image Vision Comput. B. R. A medial description of greyscale image structure by gradientlimited diffusion. Pattern Recognition (ICPR 13. Dordrecht. 51–62. 376–388. Dept. 1997. Van Gool. Discrete active models and applications. [144] G. M. Fr¨hlich. E. 279–294. 1997. A. Nonlinear scalespace.M. Vol. Hering (Eds.M. P.O.A. Viergever. M. Scalespace theory in computer vision. Gabor.J. The Gaussian scalespace paradigm and the multiscale local jet. 281–305. Springer.O. Computer Vision – ECCV ’94. Rosenfeld. Anisotropic diﬀusion based on mean curvature motion: A computational study.M. Haußecker. Florack. Proc. 1996 IEEE Int. 25–30. Koenderink. E. Seydel. Eklundh (Ed. J.J.L. 1996. Springer. 1808. 1997. Schwartz. Weickert. Sept. Gage.. ter Haar Romeny.J. Aug. R. Image structure. Conf. of Computer Science. 3. San Francisco. The Netherlands. I.M. J. ter Haar Romeny. Pauwels. 13. Pattern Recognition. Adaptive nonlinear ﬁltering for nonlinear diﬀusion approximation in image processing. 817–835.J. Report No. F. R. M. A. corners and centres of circular features. 1996.J. Lausanne.H..). On areapreserving evolution equation for plane curves. Proc. Vol. L. H. J. 1986. [136] P. Florack. Geißler. ter Haar Romeny.A. Berlin. [135] S. 10.M. Vol. Chen. SPIE Vol. H. Koenderink. Estes. Lecture Notes in Computer Science. Florack. Vol. Vision. Image Processing (ICIP–96. 1996. 105–117. 272–283. Scale and the diﬀerential structure of images. IEEE Int. G¨ lch. M. E.J. Kluwer. Contemp.S. Florack. Neumann. R.A. 61–75. Data. F¨rstner. Visualization in biomedical computing ’92. Springer. Germany. ter Haar Romeny. 1992). Vol. 165–168. Viergever. Padualaan 14.). 1994. Scalespace analysis for image sampling and interpolation.J. Lecture Notes in Computer Science. 1996). Florack. 67653 Kaiserslautern. 3584 CH Utrecht. 1992. on Fast Processing of Photogrammetric Data (Interlaken. [138] L. [148] J. Fejes. [139] L. Fritsch. Vol. Fischl. Laboratory of Technomathematics. Fiddelaers. Proc. 13th Int. Viergever. 104. Technical Report UUCS199725. Information theory in electron microscopy. Acoustics. [143] L. [147] D. and images.M. 14. Robb (Ed. J. [140] L. 47–54. Koenderink. 1996. Maas. [146] T.J. [137] B. [149] D. 30.BIBLIOGRAPHY 147 [134] O. 427–432. Mustererkena nung 1996. June 2–4.. Florack. B. 1987. Pseudolinear scalespace theory: towards the integration of linear and morphological scalespace paradigms. 1995. J¨hne. Vol. Box 3049.M. 469–472. 1965.). 1994. Speech and Signal Processing (ICASSP–92. J. Keriven. H. March 23–26. R. [150] M. 1992. B. . Conf. B. models. Level set methods and the stereo problem. 276–280. 801–807. 23. 1987). [145] M. Image Vision Comput. Int.
1997. u [157] G. 69–96. Goshtasby. Scale and segmentation of greylevel images using maximum gradient paths. [172] L. The heat equation shrinking convex plane curves. The Museum of Modern Art. Israel. Gourlay.A. Grayson.D. 1992. Paris. Reading. 1987. L.C. Box 3049. 1965. Diﬀerential Geometry. New York. Bizais. G. 1890. Image Vision Comput. 367–383. 398–407. e e [162] V. Vol. Greenspan. ETH– Zentrum. P. 360–365. 1984. 279. J. Vol. P. A. Gibbs distributions. 14. Comput. M. 1992. [170] D. 67653 Kaiserslautern. Knutsson. Vol. SaintR´my. Scaleimprecision space.F. van Gogh. On edge focusing. Int. 1985. On nonnegative diﬀerence analogues of elliptic diﬀerential equations. Rijksmuseum Kr¨ller–M¨ ller. 15. Mach. Switzerland. 221–232. Wintz. Vol. [159] D. Kikinis. [165] A. IEEE Workshop on VLSI Signal Processing. Dordrecht. . P. 6. J. AuverssurOise. Stochastic relaxation. Intell. 247–256. Gonzalez. Granlund. van Gogh. Signal processing for computer vision. 7. Preprint No. Image Vision Comput. Germany. D. u IEEE Trans. Colchester. F. Kluwer. 1992. Geman. Gorenﬂo. Finding line structures by multivalued nonlinear diﬀusion of feature e maps. A common framework for image segmentation. Berger. Information Processing in Medical Imaging. IEEE Trans. 1991. Image Vision Comput. A. [166] A. Image Science Division. A VLSIarchitecture for parallel nonlinear diﬀusion with applications in vision. H. Implicit convolution. 1889. G. Franklin Inst. G.P. Jolesz.H. W1000 Berlin 33. 1977. Nonlinear anisotropic ﬁltering of MRI data. 8092 Z¨ rich. Medical Imaging. Graham. Diﬀerential Geometry. 1889. 1994. 227–243. Gerig. Conservative diﬀerence schemes for diﬀusion problems. Geman. 165–176. Snow removal: A noisestripping process for TV signals. Information Theory. R. 369–398. 129–144. Starry night. Vision. The heat equation shrinks embedded plane curves to round points. IRE Trans.. Yuille. Gerig. R. 39. J. Image Vision Comput.148 BIBLIOGRAPHY [151] M. 12. G. van Gogh. Vol. [154] S. e [161] V. Detection and characterization of unsharp blobs e by curve evolution. F. Griﬃn. o u [163] R. Six. Pattern Anal. Oosterlinck. 11. 26. [155] G. 3... Vol. J. FB Mathematik. [152] D. Jain. 23. 1995. Sz´kely.S. Report. [153] D.D. IEEE Trans. [160] V. Vol. 6. A. FU Berlin. and the Bayesian restoration of images. Geman. G. Vol. Gage. Mach. Van Gool. 8. 10. Vol. Vol.C. H. Catthoor. P. Giuiliani. master thesis. Directed diﬀusion equations and edge detection. 1962.. K¨ bler. Arnimallee 2–6. 1994. Geiger. Di Paola (Eds.O. 15–23. [167] R. 389–402. Barillot. 1991. Road with cypress and star. Pattern Anal. [156] G. 1993. Y. [171] L. Sz´kely. of Mathematics. R. Addison–Wesley. Reynolds. Vol.R. Proc.). Vol. O. De Man. Hamilton. Dept. Otterlo. Kluwer. Selfportrait. Vol. Digital image processing. Robinson. University of Kaiserslautern.. [168] G. Griﬃn. 1986. Muse´ d’Orsay.. Gerig. Intell. C. 285–314. SaintR´my.C. Dordrecht. Communication Technology Laboratory. Gijbels. 721–741. [158] T.. Constrained restoration and the recovery of discontinuities. 1987. [169] M. [164] R. Germany. 1995.
Santa Barbara. Watson (Eds. [182] C. D. M. B. Image Proc. Proc. IEEE Trans. 1994. ter Haar Romeny. R. Basic theory of pattern observation. Variational image segmentation using boundary functions. L. Hellwig.C. 204–209. Florack. U. Mathematical morphology: A modern approach in image processing based on algebra and geometry. [176] B. IEEE Computer Society Press. Universit´ Paris IX – Dauphine. in preparation. M. Haralick.. 9535. Proc. Wilting. 1988).A. J. Hackbusch. [180] W. Preprint No. 1988. L. J. Conf. Image analysis using mathematical morphology. Niessen. Comput. Flow by mean curvature of convex surfaces into spheres. Soc. [175] B.. Proc. Representations based on zerocrossings in scale space. Wilson (Eds.). 1984. Partial diﬀerential equations and image iterative ﬁltering.).J.J. Vol. G. Johansen (Eds.A. Springer. 139–148. Duncan. J. M. J. Partial diﬀerential equations. Zhuang. Computer Vision and Pattern Recognition (CVPR ’86. Multigrid methods and applications. CEREMADE. Kluwer. L. 37.M. 75775 Paris Cedex 16. Iijima. Nielsen. Soc.. 1997. Nov. Harris. Phys. Papers of Technical Group on Automata and Automatic Control. Fouth Alvey Vision Conf. Amer. Kluwer.M. 237–266. Manjunath. Lange. 1977.. Huisken.J. Intell. New York. Hummel. [189] R. Chakravarthy. Enquist. 1994). 9607. Existence of inﬁnitely many solutions for a forward–backward heat equation. 1983. [186] G. Gaussian scalespace theory. 2111–2130. Acoustics. Frontend vision and multiscale image analysis: Introduction to scalespace theory. [187] K. Teubner. Image Processing (ICIP–94. 71. M. Stuttgart. Diﬀerential Geometry. 1987.BIBLIOGRAPHY 149 [173] F.M. Manchester. Conf. Wilting. 1994. 31 – Sept. IECE. Vol. 1. 1987. Austin. Mach. (AVC 88. Geometrydriven diﬀusion in computer vision.M. S. Theorie und Anwendung der direkten Methode von Ljapunov. Kenney. CA 93106. J. A combined corner and edge detector. 1995. . Berlin. Vol. Florack.A. J. 20. [177] B. IEEE Trans. [183] A. [190] R. Vol. Stephens. Math. N. J. Hewer. [174] B. ter Haar Romeny.A. 1959 (in Japanese). Bookstein. 1989. 2. Hahn. University of California. 1986. 1996. Report No. 1986).M.G. SPIE Vol. de Swart. o Trans. Japan. to appear in I.S. F. Vol. Reconstructions from zerocrossings in scale space. Dordrecht. Aug.S. Speech and Signal Processing. 299–316.M. 231–303. ter Haar Romeny (Ed.M. Los Alamitos. C. [179] W.). 37. W. Miami Beach.M. [181] R. 1995. 3–19. 1985. June 22–26.S. III. Washington..J. Morel. Hummel. Place du Mar´chal e e de Lattre de Tassigny. H¨llig. Duﬀ.). P. Oxford University Press. X. Vol. The state of the art in numerical analysis. [185] G. [178] B. Dordrecht. Deblurring Gaussian blur. submitted to IEEE Trans. [188] G. Diﬀerential structure of images: accuracy of representation. Harten. B. Viergever. ECE Dept. Vol. Osher. Florack.A. Sternberg. SIAM Review. France. Guichard. Heijmans. 278. ter Haar Romeny. 1959.L. IEEE Int. S. Moniot. [191] T. Sporring. 1994. Applications of scalespace theory. Dec. Pattern Anal. 13–16. 21–25. 147–152. ter Haar Romeny. 1–36.S. Uniformly high order accurrate essentially nonoscillatory schemes. 532–550.M.M. [184] H. Springer. S. Dordrecht. Mathematical methods in medical imaging III. IEEE Comp. IEEE Computer Society Press. Kluwer.R. 9. 2299.
T. Cambridge. IECE. K. Mach. J¨gersand. Spatiotemporal image processing. Morphological scalespace. IEEE Trans. Vol. Meth. . Ito. a Springer. Technische Universit¨t Berlin. IECE. [198] T. Basic theory on normalization of a pattern (for the case of a typical onedimensional pattern). Iijima. Pattern Anal. 1. H. Morishita Publishing. [193] T. Illner. Fachbereich Mathematik (3). Proc. Ito. Appl. Math. Germany. Conf. An active set strategy for image restoration based on the augmented Lagrangian formulation. Math. Papers of Technical Group on Automata and Automatic Control. Joint Conf. Straße des 17. 10623 Berlin. C. H. [199] T. convex optimization in Hilbert spaces. [209] B. Mori. Preprint No. Augmented Lagrangian methods for nonsmooth. Iijima. Appl.. Sci. A theory of character recognition by pattern matching method. [197] T. 1989 (in Japanese). No. 26. [201] R. 50–56.150 BIBLIOGRAPHY [192] T. Meth. 751. No. Neunzert. Systems. Straße des 17. Iijima. Theory of pattern recognition. Washington. Juni 136. On directed diﬀusion with measurable background. Vol. 1972). 16.. Observation theory of twodimensional visual patterns. Jackway. 15–22. Coronasha.T. 252–255. 11th Int. Bulletin of the Electrotechnical Laboratory. 30 – Sept. 1993. Tie. Mori. 1963 (in Japanese). 70–77. Controls. 1992. First Int. Iijima. Berlin. 1973 (in English). 1. 3–5. Proc. [196] T. 368–388. K. The Hague. [202] R. Vol. Jackway. 4. 10623 Berlin.. Oct. Pattern Recognition (IJCPR 1. 681–690. 6. Scalespace properties of the multiscale morphological dilation– erosion. Vol. [195] T. 1979. 2. Iijima. A theoretical study of pattern identiﬁcation by matching method. 1972 (in English). IEEE Computer Society Press. Sci.. Juni 136. 1994. Preprint No. Nov. K. Vol. M. Math. Deriche. 42–48. 38–51. Oct. Global existence for twovelocity models of the Boltzmann equation. 123– 134. Pattern Recognition Issue. 1993. [208] M. Series of Basic Information Technology. Japan. 3. 1992). Illner. 187–193. a [206] P. Fifth Int. Meth. Vol. Germany. Lecture Notes in Computer Science. Illner. Appl. Iijima. 1995). Kunisch. 195–202. Fachbereich Mathematik (3). Pattern recognition. Theory of pattern recognition. Iijima. Japan. Genchi. 1962 (in Japanese). Conf. 1973). Oct. Relative entropy maximization and directed diﬀusion equations. Vol. 1994. 1971 (in English). Studies on Information and Control. 1962 (in Japanese). 545–554. 18. 1993. Computers. on Computer Vision (ICCV ’95. 1973 (in Japanese). Vol. 1996. Proc. Kunisch. H. [204] K. K. Sci. Aug. Iijima. Pattern Recognition (ICPR 11. Iijima. Basic equation of ﬁgure and observational transformation. Technische Universit¨t Berlin. [194] T. Basic theory on normalization of twodimensional visual pattern. [200] T. 409/1994. [207] P. First USA–Japan Computer Conference (Tokyo. 410/1994. Los Alamitos. June 20–23. J¨hne. Proc. a [205] K. Vol. [203] R. J. Genchi. 1995. Saliency maps and attention selection in scale and spatial coordinates: an a information theoretic approach. 1963 (in English). Electronics and Communications in Japan. Intell. 16.
1988. [228] B.G. Kemeny. Tannenbaum.W. Morel (Eds. Rosenfeld. Imag. Deriche. S. [211] P. Denmark.D. Nonlinear diﬀusions and hyperbolic smoothing for edge enhancement. 1979. Eighth Int.. Dept. Vol. Vol. Kass. University of Cologne. [217] J. [216] J. John. Jolion. Berlin. Proceedings of the workshop on Gaussian scalespace theory. Snakes: Active contour models. 1987. Springer. 1978. Mikula. Sandini (Ed. G. Pattern Recognition (ICPR 8. 1986. Anal. Viergever (Eds. Appl. Math. K. J. Conformal curvature ﬂows: from phase transitions to active vision. Faugeras (Ed. c Koenderink. A. Berger. 588. M. [212] P. Lecture Notes in Computer Science. Kutev. Princeton. Math. A. Mikula. Kluwer. J. Germany. 1995. Germany. [219] G. Slowed anisotropic diﬀusion. J.L. of Computer Science. 1960. Andersen. O. Vol. Vol. Geometric heat equation and nonlinear diﬀusion of shapes and images. [223] J. 1990. Springer.). Wavelets and PDEs. Kawohl. Dordrecht. Vol. A.O. Witkin. Mech. 37. Math. 1252. 1996.B.). 119–124. 321–331. Herlin.W. Preprint. Kumar. New York. Skelboe. Appl. K. [227] B.B. SIAM J. Olver. K. Kichenassamy. 1994. A. [214] J. 27–31. London. Vol.M. B. [224] S. D. J. Vol. Kichenassamy. Lecture Notes in Computer Science. [215] D. 395–410. Praeger. Yezzi. Johansen. 1994.G. [225] S. L. Kaˇur. 427. 1992. Toward a computational theory of shape: An overview. Lecture Notes in Computer Science. J. Terzopoulos. Snell. Berlin. ICAOS ’96: Ime ages. 17. Kass. 275–301. Rat. Van Nostrand. Kaﬁtz. Morphologische Diﬀerentialgleichungen in der Bildverarbeitung. Finite Markov chains. 1996. [213] F. Computer Vision – ECCV ’90. A. Siddiqi. J.O.). 402–407. Comput. 134. Grue. [222] B. P. Jones.BIBLIOGRAPHY 151 [210] P. Florack. Paris. 57–67. 50923 Cologne. Analyzing oriented patterns. Conf. 1996.. 1. Graphics. . Computer vision – ECCV ’92. Report DIKU 96/19. to appear. Vol. Vol. I. On the classiﬁcation of toppoints in scalespace. Jaﬀr´. 47–59. Springer. Tannenbaum. 1997. Int. University of Kaiserslautern. P. Vision. Johansen (Ed. D.). Kaˇur. Springer. The Perona–Malik paradox. Proc. A pyramid framework for early vision. J. A computational framework for determining stereo correspondence from a set of linear spatial ﬁlters. Vol. Springer. R. Computer Vision and Image Understanding. New York. of Mathematics. 362–385. 2100 Copenhagen. Mathematical Institute. 357–360. M. Kichenassamy. Kanizsa. Johansen. Zucker. [226] S. 1986).). 1996. K. 1996. 219. Universitetsparken 1. Berlin. 64. Organization in vision. Computer Vision. N. J. Dept. Kimia. Num. A. [221] M. 67653 Kaiserslautern. Malik. Vision. Scalespace theory in computer vision. and Image Processing. Kimia. Oct. Maximum and comparison principles for anisotropic diﬀusion. Representing signals by their toppoints in scale space. University of Copenhagen. 4. Witkin. [218] H. S. Box 3049. Lecture Notes in Control and Information Sciences.. Partial diﬀerential equations. ter Haar Romeny. Arch. A. 305–322. 215–217. [220] M. Solution of nonlinear diﬀusion appearing in image smoothing and c edge detection. 1997.M. diploma thesis.
Appl. van Doorn. I. 1252. Sochen. 1995.J. Vol. 350–355. IEEE Trans. Florack. C. Vienna. 1984. Images as embedding maps and minimal surfaces: Movies. and deformations I: The components of twodimensional shape and the reaction–diﬀusion space. 159–162. Preprint LBNL39640. 58. [238] M. L. Vol. N. Vol. 367–375. The Hague. Kramer. 1997.M. Malladi. R. [244] K. Pattern Recognition (ICPR 11. The classical case. Aﬃne diﬀerential signatures for gray level images of planar shape. Proc. [239] G.). J. University of California. Ayache. 1968. Springer. Computer Vision and Pattern Recognition (CVPR ’97. 7.. Kimmel. 1988. J. 388–397. Kimia. 1992). 45–49. Iterations of a nonlinear transformation for enhancement of digital images. [235] R. . N. A multiscale algorithm for image segmentation by variational method. 282–299. 72. 1992. Numer. Scalespace theory in computer vision. Math. Cybern. 11th Int. 30 – Sept. J. [241] J.N. Vol. Kin. Viergever (Eds. 1983. Anal. 50. Sato.A.J. [243] H. Kimmel. Berkeley. J. Kimmel. Anal. Bruckner. 3. Florack. SIAM J. Vision. C. 345–348. Biol.E. Cybern. 1987. M. Lopez. Kimmel. G. Solonnikov. 1992. Zucker. Vol.). Vol. Kumar. S.152 BIBLIOGRAPHY [229] B. 212–223. Los Alamitos.. Tannenbaum. G. Pattern Recognition (ICPR 13. R. Conf. Scale space ﬁltering on spherical pattern. Vol. Scaletime. 31. Scalespace theory in computer vision. [232] R. 1975. [233] R. Math. [230] B. A. N. 1997. Directional anisotropic diﬀusion applied to segmentation of vessels in 3D images. A. Biol.J. Ural’ceva.. Proc. 1996. Kimia. Koenderink. Koenderink. IEEE Trans. Knutsson. Springer. Conf. Soc. 363–370. Vol. Comput. Viergever (Eds. Providence. Image Proc. 1996. A. Krissian. [234] R. 1997. Shapes. V. Aug. [246] O.A. Optic ﬂow: a curve evolution approach. Kocan. Int. Representation of local geometry in the visual system. [231] R. J. 163. Malandain. Vol. Ladyˇenskaja. Cybern. Koenderink.J. 53–58. S. 638–641. The structure of images.A.P. IEEE Comp. Lawrence Berkeley National Laboratory. Malladi. CA 94720. Berlin. 1996). American Mathematical Society. Kiryati. Tannenbaum. Approximation of viscosity solutions of elliptic partial diﬀerential equations on minimal grids. Wilson. Balas. A. and volumetric medical images. 13th Int. Analyzing and synthesizing images by evolving curves with the OsherSethian method. Bruckstein. 598–610. 189–224. Vol. N. 31. Lecture Notes in Computer Science. ter Haar Romeny. 1994. 1996. Granlund. Berlin. Pattern Recognition.W. shocks. A. Lecture Notes in Computer Science. R. [237] H. Tannenbaum. Vol. Vol. Vol..Zucker. N. Proc. [242] J. U. On the evolution of curves via a function of curvature. Anisotropic nonstationary image estimation and its applications: Part I – Restoration of noisy images. Morel. [240] J. 438–458. Biol. 15. Kimmel. Int.R. June 17–19. 25–30... 1996.J.B.S. Conf. Vol. 1997. L. Numer. [236] G. 24. J. 1252.. A. Aug. IEEE Computer Society Press. Comm. M. Vision. Intrinsic scalespace for images on surfaces: The geodesic curvature ﬂow.. color. G. B. 73–92. Vol. Koenderink. ter Haar Romeny. San Juan. [245] A. 1997). Koepﬂer. J. M. 5. 37–56. Sochen.. Comput.M.H. Koenderink. B. Linear and quasilinear equations of z parabolic type.B. 55. On the geometry of texture.
[253] Y. Annual Int. Scalespace theory in computer vision. S.M. Vol. S. [261] P.. A multiresolution hierarchical approach to image segmentation based on intensity extrema. 2167. 1997. J. 1994. F. Image Vision Comput. Computer vision – ECCV ’96. Sitta.. J. Suite 101. [249] A. M. 1994. The relationship between diﬀusion ﬁltering and Gibbs smoothing. Vol. Vol. R. [257] T. 5. 14. 7. Denoising and deblurring images using constrained nonlinear partial diﬀerential equations. Variational problems with a free discontinuity set. Santosa. Image processing. Lindeberg. of Biomedical Engineering. Santa Monica. 12. U. 1–8. A. Kluwer. Inverse Problems. Shapeadapted smoothing in estimation of 3D depth cues from aﬃne distortions of local 2D brightness structure. Vol. L. University of North Carolina. [263] P. Lions. J. T. Vol. [248] C. G˚ arding. 1977. Chapel Hill. Chen. [250] A. IEEE Trans. 1029–1037. 1990. Lindeberg. Appl. M. A computational algorithm for minimizing total variation in image enhancement. A clinical tool for enhancement of portal images. Dordrecht. Math. 27. 1994. IEEE Trans. Zucker. Mach. NC 27599. M. 1994. personal note. Man and Cybernetics. Rosenman. 1994. Loew (Ed. Lindeberg. Scalespace with causal time direction. C. Mach. Li. Johansen (Eds. Dept. 403–410. Geometrydriven diﬀusion in computer vision. 17. Springer.I. Boston. Systems. Pattern Anal. 1990. Solimini. B. [251] S. F.. Vol. 1996. A new regularization scheme for inverse e scattering. Lindenbaum. 1993. Pattern Anal. Intell. Lalush. 27. S.S. Inc.H. IEEE Trans. Sgallari. Lifshitz. [256] T. 2800 – 28 St. B. 147– 154. Vol. Iterative enhancement of noisy images. 467–475. On discontinuityadaptive smoothness priors in computer vision. of the IEEE Engineering in Medicine and Biology Society. IEEE Trans. Axiomatic derivation of image processing models. Intell. 415– 434. Nielsen. On Gabor’s contribution to image enhancement. 234–254. Dordrecht. S.Z. Berlin.. BlancF´raud. Chen. M. Conf.). Buxton. Meth. 229–240. Lindeberg. [254] L. Nonlinear diﬀusion with multiple edginess thresholds. Leaci. D. [259] T. Cipolla o (Eds..L. 13. Preprint.M. Lions.S. Vol. Vol. Barlaud. 1064. 1995. Pattern Recognition. Li.). 987–995..).. L. Improvements to the anisotropic diﬀusion model for 2D echo image processing.. SPIE Vol. Pizer.A.BIBLIOGRAPHY 153 [247] D. Gaussian scalespace theory.L. [252] X. 435–442. Sci. Lindeberg. Mod. [262] P. Fischer. Lobel. Li. . ter Haar Romeny (Ed. 12. 1996. CA 90405. Lamberti.. Pattern Recognition. 4.A. Pattern Anal. Volume I. 1997. 1872–1873.H. Proc. Intell. 15. Vol.). 1997. IEEE Trans. Vol. P. [258] T. 529–540. Image Proc. Pichot. Kluwer. Rosenfeld. Vol. 75–97. A. 576–586. Bruckstein. Fagerstr¨m. 543–550. Florack.M. M. [264] M. L.S. Kluwer. Osher. [260] M. Lev. [255] T.. On the axiomatic foundations of linear scalespace. 1994. U. 1994. Rudin. J. Cognitech. Sporring. Mach. Lecture Notes in Computer Science. Loew. 1992. Scalespace for discrete signals.
Marchuk. 5. New York.K. Axiomatization of shape analysis and applications to texture hyperdiscrimination. Shape modeling with front propagation: A level set approach. 1996. 1. 55. Pattern Anal. Deriche. IEEE Trans. 1996).L. 11. Maragos. 57–77. L. on Computer Vision (ICCV ’95. 197–462. 633–636. 1993. Computer Vision and Pattern Recognition (CVPR ’93. [268] L. 16–19. Malladi. Stuttgart. [281] T. BP 93. Glick. Vemuri. Vol.. CVGIP: Graphical Models and Image Processing. 1996. Soc. Conf. Report No. Interim report of the project DPH 6473/91 – Coupling of NavierStokes and Boltzmann regions. 1989. Vol. IEEE Computer Society Press. Conf. Internal Report. [267] L. [266] A. Theory of edge detection. Lopez. 17. Lucido. Intell.C. 187–217. e e e e 3192. King. 1975.. Image processing: ﬂows under min/max curvature and mean curvature. Terzopoulos. Z. Morphological systems: Slope transforms and max–min diﬀerence and diﬀerential equations. J. 1994. June 15–17. 1993. H.154 BIBLIOGRAPHY [265] C. 646–647. S. 207. June 20–23. 91–108. 1996. Sept. Caselles. 2. [273] R. Image Proc. E. IEEE Trans. [272] R. Louis. [275] P. [282] T. V. Sapiro. Matheron. Vol. Fifth Int. 127–141. Sur quelques sch´mas num´riques de e e r´solution d’´quations aux d´riv´es partielles pour le traitement d’images. Mach. ROBOTVIS. Nuclear Sci. Topologically adaptable snakes. .G. 2004 route des Lucioles.L. France. New York. Alvarez. R. Ciarlet. V. 840–845. 1996. Graphical Models and Image Processing. 58. Neunzert. 1994. Random sets and integral geometry. J. Laboratory of Technomathematics. [271] R. Cambridge. Soc. Lions (Eds. June 21–22. A representation theory for morphological image and signal processing. Vol. 1993. Splitting and alternating direction methods. G. Luo. INRIA. 1996). Conf. Zhang. [280] G. J. Maragos. Handbook of numerical analysis..S. Proc. IEEE Computer Society Press. Wiley. McInerney. J. J. D. Kimmel. IEEE Int. Proc. 408–417. Opderbecke. 586–599. Intell. Sethian. Signal Processing.I. Sethian. Vol. [279] J. [277] G. 1995. London Ser. R.A. A geometric approach to segmentation and analysis of 3D medical images. Terzopoulos. P. 1989. 1995. I. Vol. 1996. Diﬀerential morphology and image processing. J. Proc. 06902 Sophia–Antipolis Cedex. Vol. IEEE Workshop on Mathematical Models in Biomedical Image Analysis (MMBIA ’96. 1997. Marroquin. V. D.A.M. Vol. IEEE Trans. Teubner. Malladi. San Francisco. Los Alamitos. B. Proc. 2800–2806. [270] D. 244–252. Lausanne. Rigaud. Image Processing (ICIP–96. IEEE Comp. Malladi. Local geometry variable conductance diﬀusion for postreconstruction ﬁltering. Deterministic interactive particle models for image processing and computer graphics.). Pattern Anal. Medical Image Analysis. Vol. M.. Marr. Inverse und schlecht gestellte Probleme. 38. 1993). Vol. [276] P. Hildreth. Lukschin. 158–175. Los Alamitos. 922–937. 1995). Los Alamitos. D. Deformable models in medical image analysis: A survey. Struckmeier. [274] P.A. Proc. [269] A. B. Lucido. [278] D. Maragos. An integrated multiscale approach for terrain referenced underwater navigation. Adalsteinsson. R. 1980. Deriche. IEEE Trans. 41. McInerney. 1990. Sethian. Kaiserslautern. Rigaud. J. Morel.A. IEEE Computer Society Press. R. Mach. Vol.
J. a a a T. [297] M. Rjasanowa. Motion of multiple junctions: a level set approach. 320. Phys. S. Viergever (Eds. 1997. IEEE Int. R. 449–462. [287] L. Neunzert. Mitchell. Morton.M. [286] A. Nagao.). D. IEEE Computer Society Press. 1. Pattern Anal. Schunck (Eds. 1994. e Paris. France. June 19–23. u u Springer. J. Washington. Rudin. W. Shah. Image Processing (ICIP–94. R.H. Mumford. Proc. 1994). . E.I. Investigative and trial image processing. Solimini. Meis. 14. L. Mathematics and Physics. [285] B. K. San Francisco. Boundary detection by minimizing functionals. Threedimensional shape representation from curvature dependent surface evolution. [298] P.). Sci. W. Berlin. Comm. Los Alamitos. [299] H. Variational methods in image segmentation. Multiscale contour segmentation. 1980. 112. Mathematik – Schl¨ sseltechnologie f¨ r die Zukunft. 1995. I. H. Computer Vision and Pattern Recognition (CVPR ’85. J. 1989. J¨ger. Nov. e e 75775 Paris Cedex 16. S. Soc. Wasow. Mackworth. Springer. Terzopoulos. Springer. 2567. Marcowitz. [292] J. 394–407. a 1994. Conf. Vol. Math. 1995. Preprint No. Scalespace theory in computer vision. Bramble (Eds. 13–16. IEEE Computer Society Press. Comput. D. Mumford. 334–363. [290] F. A. Lecture Notes o in Computer Science. Weickert. Numerical solution of partial diﬀerential equations. Mathematische Werkzeuge in o der Bildverarbeitung zur Qualit¨tsbeurteilung von Oberﬂ¨chen. Shah. Kimia. Conf. L. J. [293] K. [291] F. 1985.R. A theory of multiscale. Koenderink. 6–10. U. [294] T. M¨sges (Eds. Lecture Notes in Computer Science. 1979. Lohmann. B.). Moisan. Vol. Mokhtarian. Mokhtarian. IEEE Trans. CEREMADE. 1205. R¨sch. [296] D. Claus. Acad. Mach. 1252. Universit´ Paris IX – Dauphine. Numerische Behandlung partieller Diﬀerentialgleichungen. McInerney. Berlin. Berlin. 9628. 1997.K.W. J. Neskovic. Vol. CVRMed–MRCAS ’97. Vol. Place du Mar´chal de Lattre de Tassigny. B. Vol. Mayers. e [288] L. R. Hoﬀmann.F. Cambridge. Analyse multi´chelle de ﬁlms pour la reconstruction du relief. 42.K.. 296–307. Intell. IEEE Comp. 9. Moisan. 31. 1992.S. Motzkin. 1994. Vol. Morel. Wiley. Perspective invariant movie analysis for depth recovery. D. Matsuyama. [295] D.. J. 1994.B.). curvaturebased shape representation for planar curves. C. 253–259. Vol.F. 1996. S´rie I. Osher. 84–94. 1985). J. Berlin. Vol. S.BIBLIOGRAPHY 155 [283] T. 1997. Computer Graphics and Image Processing. Chichester. M. J. Austin. 1952. 23–32. 279–284. [289] L. 1978. 789–805. Moisan. Birkh¨user. ter Haar Romeny. Cambridge University Press. 22–26. Grimson. Optimal approximations by piecewise smooth functions and associated variational problems. SPIE Vol. Proc. The ﬁnite diﬀerence method in partial diﬀerential equations. Merriman.. Pure Appl. Medical image segmentation using topologically adaptable surfaces. Boston. [284] T. Griﬃths. 577–685. Springer. t. Bence. Aﬃne plane curve evolution: a fully consistent scheme. B. Edge preserving smoothing. K. Troccaz. Florack. T. On the approximation of linear elliptic diﬀerential equations by diﬀerence equations with positive coeﬃcients.
IEEE Computer Society Press. Conf. Deriche. T. [301] E. L. 187–205. IEEE Trans. L. [304] W.A.G. Proc. Olver. 1997. A. G. Math. Denmark. J. 1996. S´rie I. Dordrecht. San Francisco. Vision. 7. Dept. 1993. Sapiro. J.J.). Appl. ter Haar Romeny (Ed. Dirilten. scalespace and edge detection ﬁlters.L. 11. e [315] P. Lecture Notes in Computer Science. Viergever. thesis. u [302] M. Bombay. Classiﬁcation and uniqueness of invariant geometric ﬂows. Nonlinear multiscale representations for image segmentation. 1994. R. 318–327. Dordrecht. H. R. u o diploma thesis. Nonlinear image ﬁltering with edge and corner enhancement. Aﬃne invariant detection: edges. Invariant geometric evolutions of surfaces and volumetric smoothing. Nielsen. Florack. Proc. Nielsen. [309] M.F. Dordrecht. 319. 1997. active contours. A. 1997. Geometrydriven diﬀusion in computer vision.D. 22. 176–194. Nitzberg. M. Sapiro. Kluwer. 1992. Weickert. Multiscale optic ﬂow and stereo. Computers. and segments. Niessen. B. 1995. IEEE Comp. A discrete analogue of the Weierstrass transform. Universitetsparken 1. [316] P. [308] W. Viergever. Johansen (Eds. Proc. 191–200. Tannenbaum. Image Vision Comput. Filtering. P. Vol. P.. Sixth Int. Niessen. 662. 596–604.. in press. Florack. [313] O. Tannenbaum. R. 14. Mach. Gaussian scalespace theory. Olver. J.J.A. B. Paris. [307] W. IEEE Trans. Nielsen. M. Shiota. Kluwer. C. ter Haar Romeny. 392–410. M.J. Berlin. Vol. Weickert. K. University of Copenhagen. A nonlinear transformation for digital picture processing.A. Vol. M. Imag. 1996). 57.M. Numerical analysis of geometrydriven diﬀusion equations. Florack. Florack. Regularization. Charakterisierung von unscharfen blobf¨rmigen 3D Strukturen.J. 291–307. Niessen. G. Acad. T. B. of Computer Science. L.M. Nordstr¨m. Soc. SIAM J. Intell. Multiscale watershed segmentation. A general framework for geometrydriven evolution equations. 1998). Los Alamitos. Image Science Division. Ph. M. . 826–833. 233–245. Shiota. 31–42. Newman. G. L.). 1996. J. 66. [305] W. Sci. Nitzberg.A. Int. segmentation and depth. Switzerland. [306] W.156 BIBLIOGRAPHY [300] T. 1990. Computer Vision and Image Understanding. M. 1973.M. Biased anisotropic diﬀusion .J. Olsen. D. Niessen. 2100 Copenhagen. [312] E. Vol. Kluwer. Math. t. Vol. Computer Vision and Pattern Recognition (CVPR ’96. Conf. 869–873. Viergever. Vol. ETH– Zentrum. 1997. Nielsen. Sp¨ hler.a uniﬁed regularization and diﬀusion apo proach to edge detection. [303] M. J. Vol. [314] P. From paradigm to algorithms in computer vision. Vol.J. Tannenbaum. Mumford. Vincken. 21. Communication Technology Laboratory. Vision. Jan. [310] M.L. Sporring.. June 18–20. Pattern Anal.). ter Haar Romeny. 1997. Norman. 1997. Vincken. Vol. 8. K. 1994. on Computer Vision (ICCV ’98. Sapiro. Viergever. Math. [311] N. Nicolet. A.M. Comput. Sporring. 8092 Z¨ rich. M. Johansen (Eds. Soc. Gaussian scalespace theory.J.. Olver. Springer. J. Maas. Threedimensional MR brain segmentation.J. 339–344. 4–7. Report DIKU 95/8. Amer.J. 1960. Niessen. 520–525.
[333] E.). Sixth Int. A PDEbased levelset approach for detection and tracking of moving objects. Vol.D.. June 7–9. Lecture Notes in Computer Science. Workshop on Computer Vision (Miami Beach.J. Japan. R. Perona. Ph. Tartagni. Van Gool.BIBLIOGRAPHY 157 [317] S. Paragios. P. 13–16. 1996. Bombay. Fiddelaers. 691–701. 2. Vol. P. 16–22. 1994. [324] E.A. Anisotropic diﬀusion. Springer. Conf. [320] N. 4–7.J. ter Haar Romeny. Fiddelaers. J. in press. Mach. Jan. 1–5. Koenderink. Circuits and Systems (ISCAS–88. Tescher (Ed. Ibaraki. Dordrecht. L. 629–639. Report No.S.). 1567. Willsky. [323] E. SPIE Vol. R. Dordrecht. Oceanogr.J. 1997. 1988. Kluwer. A. The generation of salinity ﬁnestructure by vertical diﬀusion. J. J. IEEE Trans.G. Osher. A network for multiscale image segmentation. Ph. Scalespace theory in computer vision. [319] S. 1998). 1101–1105. Perona. [326] P. Payot. Proc. 79. 1995. Enhancement of planar shape through optimization of functionals for curves. [328] P. Scale space and edge detection using anisotropic diﬀusion. 12. An extended class of scaleinvariant and recursive scale space ﬁlters. Malik. Symp. 1993. Scale space and edge detection using anisotropic diﬀusion. [322] N.. Los Alamitos. Deriche. 1995. Petrowski. Perona. IEEE Computer Society Press. IEEE Computer Society Press. Electrotechnical Laboratory. IEEE Trans. 414–431. 17. Mach. Intell. [318] S. Pollak. Washington. Espoo. Vorlesungen uber partielle Diﬀerentialgleichungen. Image Processing (ICIP–94. Moons. 1991.). Soc. Pauwels. Van Gool. Malik. H.S. Vol. Sethian. 1988. Modelbased extraction of geometric structure from digital images. J. F. L. J. Phys. J. IEEE Int. Osher. 1. ¨ [332] I. T. 27. Proc. Preteux. 919–949. Anal. Proc. 1981 (in Japanese). J. Osher. 1955. Utrecht University. Numer. J. Posmentier. 200–211. 30 – Dec. Phys. 1987). Mach. Krim. Pauwels. Leipzig. . thesis. [321] K. IEEE Trans. SIAM J.J. IEEE Comp. Teubner. 12–49. Proc. 114. 47–57. Otsu. 1988). [325] E.. thesis. B. Umezono. Vol. [331] I. 1994).L. Shiota. Comput. Shocks and other nonlinear ﬁltering applied to image processing. Rudin. Pattern Anal. IEEE Int. Kluwer. B. [330] P. Intell. Diﬀusion networks for onchip image contrast normalization. Grangeat. T. Intell. Ottenberg. Perona. Perona. 7. 818.): Applications of digital image processing XIV. Vol. Geometrydriven diﬀusion in computer vision. Threedimensional image reconstruction in radiation and nuclear medicine. 1252. Mathematical studies on feature extraction in pattern recognition. Nov. ter Haar Romeny (Ed. L. Malik. 1987. Pattern Anal.. Guillemaud. A. P. 1990. Malik. 72–92. Nov. Rudin. Florack. Y. Austin. An adaptive and constrained model for 3D Xray vascular reconstruction. Sakuramura. M. Vol. 2565–2568.. [327] P.. 1994. 17. 1990. Scale space analysis by stabilized inverse diﬀusion equations. Pattern Anal. Niiharigun. L. Vol. 298–300. Viergever (Eds.D. Amans (Eds. 1977. on Computer Vision (ICCV ’98. The Netherlands. Vol.G. Trousset. J. Conf. [329] P. M.M. Fronts propagating with curvaturedependent speed: Algorithms based on Hamilton–Jacobi formulations. Berlin. Featureoriented image enhancement using shock ﬁlters. L.
Eklundh (Ed. Suite 101. Van Gool. 5. Schunck. Rao. Preprint.M. Physica D. S. Cipolla (Eds. A. project c work. 18–20 July 1989).B.J. [338] M.I. J.). A.. Osher. [347] S. L. T. Fatemi. 1993.I. Pauwels. 82–91. Morel (Eds. 1994. 13. Osher. wavelets and PDEs. Continuation and bifurcations: Numerical techniques and applications. Inc. Adaptive smoothing: a general tool for early vision. S. IEEE Int. [344] L.J. 191–228. Computer Vision and Pattern Recognition (CVPR ’93. Computing oriented texture ﬁelds.).J. Computer vision – ECCV ’96. A. 2800 – 28 St. Oosterlinck. Santa Monica. [349] P. and distortion in the variational formulation. Moons. Proesmans. J. IEE. University of Kaiserslautern. Proesmans. 1996. Nonlinear total variation based noise removal algorithms. Vol. M. E. E. New York. Approximation.). 439–451. Jaﬀr´. Los Alamitos. Mach. Conf. I. Proc. Vol. 1996. (Eds. Determination of optical ﬂow and its discontinuities using nonlinear diﬀusion. B. G. Volume I. Vol. P. London. An algorithm for generating motion by mean curvature. Robert. U.M. Oosterlinck.I.. Price. [335] C. Vol. Berger. [339] I.B. Lecture Notes in Computer Science. E. Geometrydriven diﬀusion in computer vision. Applications of reaction–diﬀusion equations to image processing. Springer. . IEEE Trans. Pattern Anal. D´partement G´nie Math´matique. 801.. Chen. IEEE Comp. e Lecture Notes in Control and Information Sciences. 1. B. Kluwer. Springer. R. Pauwels. Van Gool.J. Computing with reaction–diﬀusion systems: applications in image processing. 157–185.).158 BIBLIOGRAPHY [334] C. Nonlinear anisotropic diﬀusion ﬁltering of 3D images. 1992. [345] L. 1995. Vision. Deriche. 1992. Austin. Pauwels. 1994. Vol.R. J. 1994). Boston. Mitter. [348] Y. Rudin. 1064. Los Alamitos. L. Vol. Math.J.H. SaintMarc. Kluwer. CVGIP: Graphical Models and Image Processing. Rudin. E. 1989. 379–387. J. Intell. Osher. 1991.S. Van Gool. Deriche. 53. Lecture Notes in Computer Science. Conf. [346] L. B. 1993). IEEE Computer Society Press. R. ICAOS ’96: Images. 1996. L.G. IEE Third Int. Generic evolution of edges on families of diﬀused greyvalue surfaces. Dordrecht.). 1990. Computer Vision – ECCV ’94. 31–35. Vol. Fu. 259–268. computation. 1994. [342] J. Total variation based restoration of noisy. Oosterlinck. 1994. 49–53. [340] A. Wambacq. 60. Kluwer. P. Conf. 219. Gar¸on. ter Haar Romeny (Ed.M. 13–16.. Springer. Coupled geometrydriven diﬀusion equations for lowlevel vision. Saad.S. B. [343] L. 514–529. Nov. D. London. [341] T. Price. CA 90405. PWS Publishing. Buxton.O. Image Processing and its Applications (Warwick. J. Berlin.A. Proc. 295–304. [337] M. Soc. Geometrydriven diﬀusion in computer vision. Oosterlinck. Image enhancement using nonlinear diﬀusion. Rieger. Medioni.O. Wambacq. Rudin. ter Haar Romeny (Ed. 169–190. 1990. R. 680–681. [336] M. Vol. Image Processing (ICIP–94. INSA de Rouen and Laboratory of Technoe e e mathematics. Richardson. Cognitech. Dense depth map reconstruction: A minimization and regularization approach which preserves discontinuities. Proesmans. Roose et al. Ruuth. Dordrecht. P. Total variation based image restoration with free local constraints. blurred images. S. 1994. Rambaux. S. C. Imag. 207–217. June 15–17. Iterative methods for sparse linear systems. Berlin. A.). Herlin. IEEE Computer Society Press.
Vol. 17. 1996). Aﬃne invariant scalespace. Intell. Tannenbaum. Report 3DCV 9228. [364] G. Vision. On aﬃne plane curve evolution. Int. IEEE Trans. Trucco (Eds. Kimmel.02. 1996). [354] A. Dept. [357] G. Mach. Tannenbaum. Vol. 1996. Lausanne. Lausanne. 1996. MN 55455. 141–143. Sapiro. 1994. S´nchez–Ortiz. Vol. Sapiro. Tannenbaum. A. Sapiro. V. Edinburgh.B. Sapiro. Seventh British Machine Vision Conference (BMVC 1996. 9–12. submitted to IEEE Trans. U. A. Utrecht University.M. Sapiro. 1993. 5. Int. Sept. [362] G. 817–820. Los Alamitos. Shaked. 1994). [359] G. Comput. 1993. 1994. IEEE Int. J. Image structure generating normalised geometric scale spaces. Vol. IEEE Trans. Nov. IEEE Int. A subdivision scheme for continuousscale Bsplines and aﬃneinvariant progressive smoothing. Image Processing (ICIP–96.M. [363] G. Vol. Bruckstein. 1992. Math. 23–40. Conf. J. Vision. IEEE Int. [366] G. 477–480. Funct. Diﬀerential Equations. Implementing continuousscale morphology via curve evolution. 1994. 238–268. 1997. B.. R.S.I. Vision. [367] C. Sept. Medical Imaging. 3584 CX Utrecht.H. From active contours to anisotropic diﬀusion: connections between basic PDEs in image processing. [358] G. A. 1. Tannenbaum. The Netherlands.. Samarskij. Anal.H. Burger. Vol.M.D. Volume Image Processing (VIP ’93. Knowledgebased anisotropic diﬀusion of a vectorvalued 4dimensional cardiac MR images.BIBLIOGRAPHY 159 [350] A.B. 16–19. Bruckstein. to appear. Image Proc. [352] A. R. P. M.A. Vol. Anisotropic diﬀusion of multivalued images with applications to color ﬁltering. A. Sapiro. A. texture and multiscale image segmentation. Conf. BMVA Press.222. You. University of Minnesota. Utrecht. D. 26. 7. J. Vector (self ) snakes: a geometric framework for color. Linear scalespace theory from physical principles. B. Ph. [355] G. 25–44. Cohen. Imag. Experiments on geometric image enhancement. A. Vol. Proc. Conf. Proc. 11. of Electical Engineering. B. Rueckert. Heidelberglaan 100. 13–16. Sapiro. 67–72. Austin. [361] G. Proc. Sapiro. The Netherlands. 1363–1372. Room E. D. 1995. Proc. Salden. Math. Dynamic scalespace paradigms. 1996). D. 1996. Ringach. 1582–1586. Edge preserving geometric smoothing of MRI data. Vol. Sapiro.L. The nonlinear rescaling process. J. 3D Computer Vision. Image Processing (ICIP–94. Pattern Anal. Sapiro. Fisher.A. Conf. 1994. J. Proc. ter Haar Romeny. Vol 135. 1993. Salden. Area and length preserving geometric invariant scalespaces. 1. Math..H. J. Sapiro. Viergever. Pattern Recognition. 2. [351] A. Caselles. A. Histogram modiﬁcation via diﬀerential equations. Unique reconstruction of piecewise smooth images by minimizing strictly cono vex nonquadratic functionals.M. 1984. 1996. 1997.H.A. Imag.L. M. 189–198. Technical Report. Image Processing (ICIP–96. Leipzig. 79–120. 16–19. M. Vol. Kimia. [356] G. Theorie der Diﬀerenzenverfahren. Minneapolis. Tannenbaum. June 2–4. IEEE Computer Society Press. thesis. E. . 4. 472–476.). [365] G. Y. Imag. Viergever. [360] G. 119. Kaveh. Geest & Portig. [353] A. AZU.A. 1993). A. Sept. ter Haar Romeny. Schn¨rr.. Salden. Vision. Salden.
A.160 BIBLIOGRAPHY [368] C. 1137–1150. IEEE Comp. Berger. Acad. and applications of level set methods for propagating interfaces. 34–40. Sci. 1996). IEEE Computer Society Press. Sethian. A fast marching level set method for monotonically advancing fronts. Shah. Properties of energyminimizing segmentations. IEEE Comp. Springer. 1982. Computer Vision and Pattern Recognition (CVPR ’96. 202–207. Proc. II. Acta Numerica. 1996. Los Alamitos. London. Maui. M. . Proc. IEEE Comp. June 3–6. Graphical Models and Image Processing. Conf.B. [380] J. Vol. 99–111. Pattern Anal. 1993.O.B. Parts of visual form: computational aspects. [381] J.A. San Francisco.W. 644–647. S. 1994. Vol. 1996. B. [383] K. IEEE Computer Society Press. Control and Optimization.). Segmentation by nonlinear diﬀusion. Morel (Eds. Computer Vision and Pattern Recognition (CVPR ’96. 1993). [384] K. 1992). Seneta. Soc. Soc. Siddiqi. B. algorithms. [370] E. Schn¨rr. Picard. on Computer Vision and Pattern Recognition (CVPR ’91. Piecewise smooth approximations of functions. Mlattice: From morphogenesis to image processing. Shah. Siddiqi. 1996. Vol. 9–13. A common framework for curve evolution. [369] C. London. Proc. Shah. 1992. 507–513. 12th Int. IEEE Computer Society Press. San Francisco. [385] K. Vol. Springer. New York. Conf. SIAM J.W. Los Alamitos. 1996). Proc. Champaign.. Conf. I. 1981. Sethian. IEEE Comp. 2. Cambridge. 136–142. Calc. A nonlinear diﬀusion model for discontinuous disparity and halfocclusions in stereo. 1996. Shu. Nat.B. A. [377] J. 315–328. 1996. Vol. A shock grammar for recognition. Castan. Shen. 1995. Intell.. Siddiqi. computation and curve evolution. 1996. Los Alamitos. ICAOS ’96: Images. IEEE Comp. IEEE Trans. Mach. 219. Soc. Proc. Segmentation by nonlinear diﬀusion. IEEE Computer Society Press.A. Los Alamitos. Proc. in press. [379] J. IEEE Trans. Cambridge University Press. Var. Level set methods. Segmentation of visual motion by minimizing convex nonquadratic functiono als. New York. 1994). Image Proc. 1992. [382] A. Herlin. segmentation and anisotropic diﬀusion. Nonnegative matrices and Markov chains.. [375] J. Shah. 201–207. [371] J. 30. 1591–1595. Kimia. 93. Jerusalem. 1992). [374] J. Shah. Academic Press. wavelets and PDEs. J. 17. Geometric shockcapturing ENO schemes for subpixel interpolation. 309–395. Graphical Models and Image Processing. R. e Lecture Notes in Control and Information Sciences. 112–133. Computer Vision and Pattern Recognition (CVPR ’92. An optimal linear operator for step edge detection. Convex variational segmentation of multichannel images. June 18–20. Conf. Image analysis and mathematical morphology. o Deriche. [372] J. [376] J. June 15–17. J. 1996. Sethian. Soc. C. Soc. Sherstinsky. Kimia. Conf. Kimia. Proc. Los Alamitos. Serra. Los Alamitos.M. Schn¨rr. Jaﬀr´. 239–251. Pattern Recognition (ICPR 12. 1991. Vol.. Computer Vision and Pattern Recognition (CVPR ’93. 54. June 18–20. R. IEEE Computer Society Press. 1994. B. June 15–18. Vol. Conf. [373] J. IEEE Computer Society Press. Oct. 661–663. 5. A review of the theory. Shah. [378] J. Vol. 1992.
U. Gaussian scalespace theory. 1995. Intell. Dept. R. [387] W. DISKI 123. [398] S. M. Vol. Whitaker. CAM Report 9646. 13th Int. Scalespace and boundary detection in ultrasonic imaging using nonlinear signaladaptive anisotropic diﬀusion. A. Dept.F. Lecture Notes in Computer Science. University of Hamburg.B. 1996. Vol. Computer Vision and Image Understanding. Volumetric segmentation of medical images by threedimensional bubbles. 13th Int.. U. B. Memo 671. 1995).I.H.M. Teboul.. J. L. 2167. 1996. Florack. ter Haar Romeny. 36. of Mathematics. CA 90024. Van der Linden. Fifth Int. Kimia. Conclusions from the commodity expert project. IEEE Computer Society Press. 1996). Berlin. Spatially and scale adaptive total variation based regularization and anisotropic diﬀusion in image processing. [399] H. Image processing. T. Conf. Chan. 1997. Sprengel. On generalized entropies and scalespace. Florack. E. Scalespace theory in computer vision. CA 90024. Sporring. 1252.S. Osher–Sethian and ENO schemes. Vol. Strong. 620–624. A. Stansﬁeld. A. Weickert.M. Los Angeles. B. Raman. Aubert. Univ. IEEE Trans. Sankt Augustin. P.S. 15. Constraints on deformable models: Recovering 3D shape and nonrigid motion. 1988. Proc. A. Han. Image relaxation: Restoration and feature extraction. 25–30. Tek. Sporring. Scheunders. 65. Vol. Cambridge. A. Chan. SPIE Vol.. 116–127. Artif.S. Conf. 91–123. 25–30. Magnetic Resonance Imaging. of Computer Science. J. P. thesis.). Shockcapturing numerical methods for viscosity solutions of certain PDEs in computer vision: The Godunov. Pizer. 1994. 523–527. [397] D. Van Dyck. Vol. Brown University. Tek. Kimia. 900–904. MA. Bruckstein. of Mathematics. Variational approach for edgee preserving regularization using coupled PDEs. Massachusetts Inst. Los Angeles. 1996. of California. Planar shape enhancement and exaggeration. Kimmel.B. [400] H. Watershedbased segmentation of 3D MR data for volume quantization. Koenderink. 156–162.F. M. Image Proc. U. M.A. Germany. . Kimia. 246–258. IEEE Trans. of Technology. on Computer Vision (ICCV ’95. Olstad. Vienna. Proc. Snyder. The entropy of scalespace.A. Artiﬁcial Intelligence Lab. Pattern Anal. S. Conf. Pattern Recognition (ICPR 13. 53–64. [392] J. [395] P. 1996. 1996). [394] A.M. Mach. Kluwer. Sporring. Inﬁx Verlag. Ph. Springer. Stoll. Aug.. [388] J. March 1998. B. Viergever (Eds. Kass. Pattern Recognition (ICPR 13.. [401] D. Vienna.B.BIBLIOGRAPHY 161 [386] J. [396] D. Vol. Witkin.S. Division of Engineering. M. C. Steen. R. Proc. Dept. Dordrecht. Blanc–F´raud. Intell. 1994. G. 1996. Barlaud. Entwurf und Analyse nichtlinearer Diﬀusionsverfahren f¨r die Bildveraru beitung. Univ. Vol.S. Steiner. Bilbro.I. Image segmentation by reaction–diﬀusion bubbles. R. [391] R. in press. Verhoye. Sijbers. Y. 679–688. Los Alamitos.L. G. D. Cambridge. A. CAM Report 967. Technical Report LEMS132. U. M. 1995. 02912. 1997.).. Terzopoulos. Loew (Ed. Nielsen. 17. B. [393] E. [389] J. Aug. B. L. 1997. L. Relation of regularization parameter and scale in total variation based image denoising.D.)... Shu. 1997. [390] J. of California.W. Providence. M. Johansen (Eds. T.A. 1980. 1995.A. Strong. June 20–23.
Box 3049. 227–238.O. Comput. Pattern Anal. Pattern Anal. Weickert. Gaussian scalespace theory. 221–236.). Herlin. Springer. Vol. London B. 1970.D. 17. Oman. Matrix iterative analysis.R. A. IEEE Trans. Angew. Lecture Notes in Computer Science.T. Lecture Notes in Control and Information Sciences.. 148–163. 1994. 219. Computer Graphics. Nonlinear scalespaces. Scalespace properties of nonlinear diﬀusion ﬁltering with a diﬀusion tensor. Image recovery using the anisotropic diﬀusion equation. P. Royal Soc.L. Gew¨hnliche Diﬀerentialgleichungen. University of Kaiserslautern. Vol. Anisotropic diﬀusion ﬁlters for image processing based quality control. Vol. Torre. Jaﬀr´. 283–286. 1976. [412] W. 2563. Weickert. IEEE Trans. Florack. 1962. 1952. [414] J.. [403] F. Foundations and applications of nonlinear anisotropic diﬀusion ﬁltering. Philosoph. Tait. H. 1995. 1991. 1996. [408] K. [406] G.R. 1997.A. 25. 111–118. T. Diﬀerential and integral inequalities. V. 37–72. Springer. M. Weickert. 289–298. Mech. Intell. Ph.A. Vol. Berlin. 67653 Kaiserslautern. Berlin. Theoretical foundations of anisotropic diﬀusion in image processing. ICAOS ’96: Images. [420] J. P. Turing. Springer. Z. Germany.). Dept. Trans. 237. 1994. Generating textures on arbitrary surfaces using reaction–diﬀusion. Vol. 355–362. Walter. Proc. 1. J. Hlav´ˇ. [404] V. Solutions of illposed problems. Johansen (Eds.O. Vol. SIAM J. Anisotropic diﬀusion in image processing. 1996. Dordrecht. [421] J. Iterative methods for total variation denoising. 1996. Nonlinear diﬀusion scalespaces: From the continuous to the discrete setting. Teubner. Nielsen. J. Suppl. Report No. Laboratory of Technomathematics. [407] R. Vol. Vogel. Turk. SPIE Vol. e wavelets and PDEs. Probabilistic multiscale image segmentation.E. Vincken. Berlin. [416] J. Neunzert (Ed.. Weickert. o [413] J. On edge detection. Koster. The chemical basis of morphogenesis. [415] J. IEEE Trans. 1996. Multiscale texture enhancement. M. R. 19. 1997. [411] W. M. on Mathematics in Industry. 11. Box 3049. University of Kaiserslautern. Stuttgart.). Vol. Computer analysis ac of images and patterns. Chichester. [419] J.E.). Vogel. M.. Fast numerical methods for total variation minimization in image reconstruction. S´ra (Eds. 67653 Kaiserslautern. P. Intell. 1995. Weickert. 1996. Suppl. thesis. Oman. R. Berger. 230–237. J. F. Sci.E. L. Poggio. Arsenin. Weickert. Deriche. M. 8. 1996. 76. Image Proc.M. Englewood Cliﬀs. 359–367. Primicerio (Eds. A model for the cloudiness of fabrics. V. 258–265. Computing. 110. Kluwer. New York. ˇa [418] J.162 BIBLIOGRAPHY [402] A.S.A. 970. London. Varga. Fasano. Seventh European Conf. Walter. Springer. Germany.). Weickert. Vol. 4. Mach. No. Math. Tikhonov. K. Sporring. Weickert. of Mathematics. A.E. I. M. Advanced signal processing algorithms. Luk (Ed. Progress in industrial mathematics at ECMI 94. Viergever.. Torkamani–Azar. [409] C. Wiley. . 1977. Mach. [410] C.O. 1996. Weickert. 221–234. 1573–1578. 5.). Morel (Eds. 1986. [417] J. [405] A.M. 109–120. Prentice Hall. Wiley–Teubner.
A review of nonlinear diﬀusion ﬁltering. Image processing. Viergever. Scalespace theory in computer vision. CVGIP: Image Understanding. 1995). to appear. R. ter Haar Romeny. M. April a 21–25. Denmark. B. Gmitro (Eds. G. 340–352. Kluwer. Vol. 239–244. 1. [437] R.M. 1997). K. Conf. Orientation analysis by coherenceenhancing diﬀusion. SPIE Vol.). 57.).). 260–271. 2100 Copenhagen. Kluwer. Sanfeliu. 1996. B. A.A. On the history of Gaussian scalespace axiomatics. Benhamouda. W. Proc. Weickert. ter Haar Romeny. Jan. 96–103. B. Dept. Solina.J. Vectorvalued diﬀusion. Geometry limited diﬀusion in the characterization of geometric patches in images. Scalespace theory in computer vision. University of Copenhagen. Whitaker.T.M. Embedded active surfaces for volume visualization.BIBLIOGRAPHY 163 [422] J. 1994. 822–827. Vol.). 3–28. 1996).F. Algorithms for implicit deformable models. Zuiderveld. [428] J. Florack. 465–468. Weickert. R. Koenderink. Lausanne. L. 1997. 1997. Los Alamitos. to appear. ter Haar Romeny. Klette. Santa Barbara. Niessen. 1996). Weickert.M. Whitaker. Bajcsy (Eds. 93–134. 1995.G. Lecture Notes in Computer Science. Oct. Ishikawa. Viergever. Nielsen. Lecture Notes in Computer Science. 1986. Dagstuhl. 1997). Springer. Berlin. Characterizing ﬁrst and secondorder patches using geometrylimited diffusion. 111–120. submitted. IEEE Computer Society Press. o [433] R. P. Kropatsch. Scalespace has been discovered in Japan. B. [425] J. Florack. June 20–23. Florack. Conservative image transformations with restoration and scalespace properties. 1993. A. 1997. 26–29. Barcelona.J. ter Haar Romeny (Ed. Real World Computing (RWC ’97. ter Haar Romeny. Springer. Ishikawa. L. [424] J. [429] J. J.). A. Pattern Recognition and Image Analysis (VII NSPRIA. [436] R. Lopez. B.. M. Whitaker. Cambridge. Coherenceenhancing diﬀusion of colour images. Weickert.T. Weickert. F. Conf. M. Springer. Gew¨hnliche Diﬀerentialgleichungen. of Computer Science. B. [430] J. . 1993. Weickert.). Dordrecht. W. Theoretical foundations of computer vision (TFCV ’96. ter Haar Romeny. Imiya. 1997. Gaussian scalespace theory. Berlin. S. 45–59. 1252.M.). Lecture Notes in Computer Science. on Computer Vision (ICCV ’95. J. Arndt. 1997). Imiya. IEEE Int. Vol. J.A. Springer.J. March 1998. Proc. [431] J. Weickert. [423] J.T. [434] R. 1997. 1997 IEEE International Conference on Image Processing (ICIP–97. A. [426] J.T. Information processing in medical imaging. [432] H. Springer. H. B. L.J.M. D. van Enk. Berlin. Sept. Werner. Vol.T. 1252. Tokyo.). Gerig. Sporring. Vol. [427] J. Johansen (Eds. Geometrydriven diﬀusion in computer vision. A semidiscrete nonlinear scalespace theory and its relation to the Perona–Malik paradox. Weickert. 29–31. 1997. Symp. Whitaker. 1994. [435] R. M. B. H. A. Parallel implementations of AOS schemes: A fast way of nonlinear diﬀusion ﬁltering. Image Proc. 149–167. Villanueva. in press. Weickert.H. Weickert. 16–19. Fifth Int. 687. Viergever (Eds. M. March 18–22. Barrett. W. J. ter Haar Romeny. Loew (Ed. 2167. Chen. Whitaker. Berlin. Wien. Vol. Vitri` (Eds. Eﬃcient and reliable schemes for nonlinear diﬀusion ﬁltering. Proc. Viergever (Eds.H. Dordrecht. IEEE Trans. Universitetsparken 1. S. J. Recursive separable schemes for nonlinear diﬀusion ﬁlters. Koenderink. Proc. Report DIKU 97/18. 1. M. Image Processing (ICIP– 96.
Image Processing (ICIP–96. 1983. IEEE Trans. 15–25.M. Conf. Vol. Pizer. New York. Sept. B. Young. 1996. You. Barrett.S. You. A. 1539–1553. 1996. 1995. Lausanne. Region competition: unifying snakes. 1996. 1019–1022. P. Int. 1996. Pattern Anal. A. Vol. Signal matching through scalespace.). A multiscale approach to nonuniform diﬀusion. Tannenbaum. X. Witkin. Spatial Vision. Vision. [440] A. Proc.Z. Eighth Int. [449] R. Image Processing (ICIP–96.T. 16–19. D. Intell. 1993. B. 1987. 1996).T. Pattern Anal. Mach. [443] G. New York. Using statistical pattern recognition techniques to control variable conductance diﬀusion. M. [446] Y. Zhu. Signal Processing. W. Anisotropic blind image restoration. Iterative solution of large linear systems. CVGIP: Image Understanding.. Retinal mechanisms. 1996. L. M. 1971. 1993. Vol. Vol. 1993. Joint Conf. A. [442] N. Terzopoulos. IEEE Int. M. Recursive implementation of the Gaussian ﬁlter. IEEE Trans. Structure adaptive anisotropic ﬁltering. 25–39. Vol. Witkin. D. L. Kass. Vol. Intell. [448] I. Basu.L. theory and applications. 16–19.164 BIBLIOGRAPHY [438] R.H. Scalespace ﬁltering.. Berlin. Yuille. Yoo. Scaling theorems for zero crossings. 884–900.R. [441] A. Dobbins.M. . IEEE Int. T. 687. [451] D. Li (Eds. van Vliet. S.P. Sept. The Gaussian derivative model for spatial vision: I. Firmin.L. Image Vision Comput. [450] A. 57. [452] S. IEEE Trans. Vol. Witkin. 1. 4. Karlsruhe.. 2. Image Proc. Kaveh. World Scientiﬁc. Information processing in medical imaging. Coggins. on Artiﬁcial Intelligence (IJCAI ’83.N. 459–471. A. Singapore. Lecture Notes in Computer Science. Behavioral analysis of anisotropic diﬀusion in image processing. Proc. Poggio. [439] A. S. The method of fractional steps: the solution of problems of mathematical physics in several variables. 1986. 18. 99–110.J. [447] D. [445] Y. region growing. Young. Li. Zucker. Computer Graphics.M. Xu. No. Lausanne. Yang. Conf. 1996). Proc. Academic Press. Whitaker. 2. Aug. Vol. 299–308. Springer. Vol.). Reaction–diﬀusion textures.C. Yuille. 8. 1991. Gmitro (Eds. 1971. 461–464. Springer. 25.W. A. Yanenko. Kass. Young. Kaveh. Vol.N. 383–386. 1. 2. Vol.A. Kimia. Vol. Tannenbaum.A. 1983). H. 135–145. Burger. Zhao. M. Mach. A new implementation of discrete multiscale ﬁltering. Comput.F. 44. [444] T. 134–144.. 8–12. 14. From curve detection to shape description: an outline. Computer vision: systems. Vol. energy/Bayes/MDL for multiband image segmentation.L. J. Underwood. Iverson. 139–151. 273–293. [453] S.L. 5. J. A. 1987.
116. 114–126. 98. 16 convergence. 128 coherenceenhancing diﬀusion. 118. 25 adaptive smoothing. 48 invariant gradient. 13 bias term. 56. see additive operator splitting applications. 45. 130 condition number. 13 invariance. 44. 15 causality. 46 adaptive mesh coarsening. 105. 12–14. see aﬃne morphological scalespace anisotropic diﬀusion. 53. 102 circular point. 72. 124 Mumford–Shah functional. 30. 61. 52. 62. 21. 24. 45 invariant heat ﬂow. 28. 124 CPU times. 40 Euclidean. 2 continuous dependence. 26. see diﬀusion colour images. 110 apriori knowledge. 43 deblurring. 44–49 surfaces. 95. 61. see diﬀusion. 38 areapreserving ﬂows. 37. 10 corners. see transformation AMSS. 113–126 contrast parameter. 47. 19. see vectorvalued images commuting operators. 74 average grey level invariance. 42 average globality. 61 CAD. see wellposedness continuous diﬀusion. 5. 11. 48 basic equation of ﬁgure. see shortening ﬂow transformation. 22. 50. 52. 28. 46. 3. 30. 135 computer aided quality control. see diﬀusion AOS. 52 blob detection. 99. 107–111 aﬃne arclength. 40. 34. 5. 29. 119. 32 coherence. 26. 20. see scalespace balloon force. see perimeter shortening ﬂow. 129–134 arclength aﬃne. 88. 12. 52 165 . 33 curve evolution. see diﬀusion contrast enhancement. see invariance invariant geodesic snakes. 13. see arclength Gaussian scalespace. 28.Index apriori estimates. coherenceenhancing central moments. 106 conjugate gradient methods. 45 BUC. 36. 57. 49 blind image restoration. 38–43. 27. 128 coherence direction. see invariance axiomatics. 82. 73. 28. 5. 50. 15. 134 convolution theorem. 106. 29 morphological scalespace. 34. 29 perimeter. 57. 19. 5. 53. 67. 85. 124. 76. 116. 37 Canny edge detector. 135 active blobs. 26. 113 crystal growth. 47 curve evolution schemes. 26. 65 CED. 106 continuity. 62 comparison principle. 125–126. 97 continuity equation. 14. 33 of an image. 93. 46 contour models. 40 invariant texture segmentation. 40–45. 33. 116 correspondence problem. 39 closing. 14 additive operator splitting. 57. 113. 38 curvature of a curve.
166
deep structure, 11, 136 defect detection, see computer aided quality control defocusing, 7 deviation cost, 27, 29 diﬀerencesofGaussians, 5 diﬀerential inequalities, 17 diﬀerential invariants, 5 diﬀusion anisotropic, 2, 13, 22, 24, 55–74, 87–95, 111, 113–137 backward, 5, 50, 116 coherenceenhancing, 24, 127–134 continuous, 55–74 directed, 14 discrete, 97–111 edgeenhancing, 23, 113–126 equation, 2 forward–backward, 15, 45, 66, 72 homogeneous, 2 inhomogeneous, 2, 117 isotropic, 2–22, 86–87, 110, 116 linear, 2–14, 36, 116, 125 nonlinear, 2, 14–27, 52, 55–137 on a sphere, 12 physical background, 2 semidiscrete, 75–95 tensor, 2, 13, 22, 58, 62, 88, 95, 114, 127, 135 diﬀusion–reaction, 27–30, 118 single equations, 27 systems, 28 diﬀusivity, 2, 15, 16, 47, 48, 116 digitally scalable, 36 dilation, 32, 35, 47, 63, 113, 119 direction, 56 directional diﬀusivities, 89, 94 directional splitting, 88–95, 107–111 discrete analogue AMSS, MCM, 43 Gaussian, 10 linear diﬀusion, 10 nonlinear diﬀusion, 25, 97–111 discrete diﬀusion, see diﬀusion dissipative eﬀects, 37, 43, 44 distance transformation, 37, 47 DoGs, see diﬀerencesofGaussians doubly stochastic matrix, 98 edge cost, 27, 29
INDEX
edge detection, 5, 15, 24, 28, 30, 50, 53, 55, 57 edge enhancement, see contrast enhancement edgeenhancing diﬀusion, see diﬀusion elliptic point, 41 energy, 72, 85, 102 energy functional explicit snakes, 46 for curves, 28, 29 geodesic snakes, 48 Mumford–Shah, 28, 44, 126 Nordstr¨m, 27, 30, 51 o Perona–Malik, 18 Polyakov, 42 TVminimization, 51 ENO schemes, 37 entropy, 11, 73, 85, 102 generalized, 11, 73 entropy scalespace, see scalespace, reaction– diﬀusion erosion, 32, 47, 63 Euclidean arclength, see arclength perimeter, see perimeter shortening ﬂow, see shortening ﬂow transformation, see transformation Eulerian formulation, 33 existence, see wellposedness expected value, 73 explicit schemes, 11, 19, 25, 43, 53, 103, 105, 106, 113, 137 external energy, 46 extremum principle, 4, 6, 18–20, 30, 34, 39, 41, 44, 50, 58, 62, 66, 76, 77, 87, 93, 98, 105 fabrics, 125, 130 FD, see ﬁnite diﬀerences feature vectors, 24 Fick’s law, 2 ﬁlter design, 114, 127 ﬁngerprint enhancement, 26, 129 ﬁnite diﬀerences, 11, 25, 30, 43, 46, 50, 51, 85–95, 102–111, 113 ﬁnite elements, 25, 28 ﬂame propagation, 38 ﬂowline, 16 ﬂux, 2, 15, 22, 114 focusofattention, 11 forensic applications, 52
INDEX
Fourier transformation, 3, 10, 35, 44 free discontinuity problems, 29 fundamental equation in image processing, 41, 42 Gabor ﬁlters, 45, 136 Gabor’s restoration method, 45 gas dynamics, 49 gauge coordinates, 16 Gaussian derivatives, 4, 43, 44 elimination, 106 kernel, 3, 35 pyramid, 11 scalespace, see scalespace smoothing, 2–14, 56 Gauß–Seidel method, 44, 106 generalized entropy, see entropy generalized ﬁgure, 13 geodesic, 48 geodesic curvature ﬂow, 42 geometric heat equation, 38, 39 grassﬁre ﬂow, 33 grey level shift invariance, see invariance greyscale invariance, see invariance halftoning, 30, 37 hardware realizations, 11, 13, 25, 33 heat equation, 2 histogramme enhancement, 28, 45 Huygens principle, 33 hyperstack, 63 hysteresis thresholding, 5, 15 Iijima’s axiomatic, 7 illposedness, 4, 17, 28–30, 49, 53 image colour, see vectorvalued images discrete, 75 greyscale, 3, 55 higher dimensional, see threedimensional images vectorvalued, see vectorvalued images image compression, 73, 137 image enhancement, 14–30, 45, 48–53, 55– 74, 113–134 image restoration, see image enhancement image sequences, 11, 24, 30, 47, 48, 56, 62 immediate localization, 19 implicit schemes, 11, 44, 137
167
information theory, 73 integral geometry, 10 integration scale, 57, 114, 127, 132 interest operator, see structure tensor internal energy, 46 intrinsic heat ﬂows, 39, 40, 42 invariance aﬃne, 42–45 average grey level, 13, 63, 76, 81, 98, 99, 119 grey level shift, 62, 81 greyscale, 31, 43, 45, 124 isometry, 64, 81 morphological, see invariance, greyscale reverse contrast, 63, 81 rotational, 13, 21, 35, 43 scale, 8 translation, 8, 64, 81 irreducibility, 75, 76, 97 isometry invariance, see invariance isophote, 16 Japanese scalespace research, 7, 13 jet space, 24 junctions, 5, 29, 42, 56 Kramer–Bruckner ﬁlter, 50 LaplacianofGaussian, 5 lengthpreserving ﬂows, 42 level set, 31, 44 level set methods, 36 line detection, 26 linear diﬀusion, see diﬀusion linear systems, 106, 107, 110 linearity, 8 Lipschitzcontinuity, 76 local extrema creation, 12, 43, 65 noncreation, 34, 39, 119 nonenhancement, 65 local scale, 57 LoG, see LaplacianofGaussian Lyapunov functionals, 66–74, 135 functions, 76, 82–85 sequences, 98–102 Markov chains, 98 Markov processes, 11
168
Markov random ﬁelds, 27, 29 Marr–Hildreth operator, 5 maximal monotone operator, 17 maximum–minimum principle, see extremum principle MCM, see mean curvature motion mean curvature motion, 37–40, 42–45, 47, 50, 51, 93, 113, 119, 125, 129 backward, 45 mean ﬁeld annealing, 27 median ﬁltering, 38 medical imaging, 27, 28, 45, 46, 52, 62, 63, 111, 116–126, 131 MegaWave, 44, 116, 126 min–max ﬂow, 45 minimal surfaces, 29, 38 monotonicity preservation, 19 morphological invariance, see invariance, greyscale morphology, 30–49, 119–124 binary, 31 classical, 30–37 continuousscale, 32 curvaturebased, 37–49 greyscale, 31 multigrid, 11, 25, 106, 137 multiplicative operator splittings, 110 Mumford–Shah functional, see energy functional neural networks, 25, 137 noise elimination, 12, 21, 23, 45, 116, 124 noise scale, 21, 57 nonmaxima suppression, 5, 15 noncreation of local extrema, see local extrema nonenhancement of local extrema, see local extrema nonlinear diﬀusion, see diﬀusion nonnegativity, 76, 88, 97, 107 Nordstr¨m functional, see energy functiono als numerical aspects, 10, 25, 36, 43, 51, 85–95, 102–111, 113 oceanography, 18 ODE, see ordinary diﬀerential equations opening, 32 optic ﬂow, 11, 28, 52 optical character recognition, 13
INDEX
optimization convex, 28 nonconvex, 53 ordinary diﬀerential equations, 75–95 orientation, 56, 129 oscillations, 10, 19, 21, 44, 50 Osher–Sethian schemes, 36, 113 oversegmentation, 29 parallel computing, 25, 26, 30, 110, 111 parameter determination, 17, 46, 114, 119, 129, 137 partial diﬀerential equations, 1–53 particle methods, 27 path planning, 37 PDE, see partial diﬀerential equations perceptual grouping, 136 perimeter aﬃne, 41 Euclidean, 39 Perona–Malik ﬁlter, 14–20, 28, 30, 45, 49, 114 basic idea, 15 edge enhancement, 15 illposedness, 17 regularizations, 20–24 scalespace properties, 19 semidiscrete, 87 piecewise smoothing, 19 pixel numbering, 87 positivity, 97 positivity preservation, 8 postprocessing, 26 prairie ﬂow, 33 predictor–corrector schemes, 107, 109 preprocessing, 118, 126 pseudospectral methods, 25 pyramids, 11, 25, 125 quantization, 45 random variable, 72 reaction–diﬀusion bubbles, 48 reaction–diﬀusion scalespace, see scalespace recursive ﬁlters, 10, 11, 110, 111 recursivity, 6 redistribution property, 99 region growing, 29, 126 regularity, 58, 61
44. 41. 47. 6. 88–95 stereo. 10 semidiscrete nonlinear diﬀusion. 124 software. see diﬀusion semigroup property. 62–74 architectural principles. 6 axioms. 52 Stampacchia’s truncation method. 26. 39 morphological equivalent of Gaussian scalespace. 46 explicit. 34. 27 terrain matching. 22 shapefromshading. 47 selfsnakes. 25. 38 silhouette. 42 reaction–diﬀusion. 27. 11. 97. 34. 98 stopping time. 11. see structure tensor second moment matrix. 58. 124 segmentation. 20 reverse contrast invariance. 39. 28. 25. 72. 36. 35 smoothness. 45. 36. 58. 4. 44. 50. 35 nonlinear diﬀusion. 35 subpixel accuracy. 113 semidiscrete. 34 169 separability. 27 row sum. 25. 116 string theory. 7. 36 subsampling. 45 shape inclusion principle. 6 scalespace aﬃne Gaussian. 4. 99. 41 anisotropic diﬀusion. 37 slope transform. 43. 48. 52. 27 rescaling. 62–74. 110 settheoretic schemes. 109 stabilizing cost. 37. 102–111. 76. 7 mean curvature. 50 ﬁlters. 44. 7 linear diﬀusion. 99–111 for curves. 118 projective. 41 Euclidean. 32 structuring function. 34 discrete nonlinear diﬀusion. 39 shapeadapted Gaussian smoothing. 62 stochastic matrix. 49 creation. 48. 6. 136 structuring element. see snakes semiimplicit schemes. 10 semidiscrete analogue linear diﬀusion. 81– 95 smoothing principles. 29. 42. 63. 62 scatter matrix. 45 oﬀset. 126. 75–95 semidiscrete diﬀusion. 45 . 28. 50. 13 aﬃne morphological. 76. 37 shock capturing schemes. 56 exaggeration. 116 linearity principle. 11 scale–imprecision space. 8. 97 target tracking. 128 snakes. 42 Gaussian. 114. 53. 48 implicit. 62. 37 recognition. 56. 49. 6. 97 scale invariance. see convergence steerable ﬁlters. 74. 137 SOR. 50 shortening ﬂow aﬃne. 43 shape. 93. 30. see invariance robust statistics. 136 stencil size. 27. 116. 21. 114 remote sensing. 118.INDEX regularization. 76. 31 simulated annealing. 39. 25. 42 for image sequences. 10. 6 invariance principles. 59 steadystate. 43 for shapes. 53 dilation–erosion. 10. 46 geodesic. 7 general concept. 27. see invariance scale selection. 10 nonlinear diﬀusion. 106 stability. 18. 35. 29 skeletonization. 29 staircasing eﬀect. 44. 11. 31 cues. 13. 106. 45 semidiscrete linear. see structure tensor segmentation. 86. 25. 137 selfsnakes. 73 symmetry. 42 structure tensor.
50 preserving methods. 52. 33 wavelets. 24. 5. 38 projective. 24. 39. 76. 127 generation. 58. 34. 43. 42. 58. 73 enhancement. 47–49. 42 translation invariance. 24. 30 TV. 125 zerocrossings. see invariance Turing’s pattern formation model. 129 uniqueness. 53. 62. 50 INDEX . 25. 28. 46–48. 136 viscosity solution. 41. 43.170 texture analysis. 45. 41. 77. 47 toppoint. 30 segmentation. 47–49. 136 weak membrane model. 34. 93. see wellposedness van Gogh. 49 minimizing methods. see total variation uniform positive deﬁniteness. 62. 73 vectorvalued images. 29 threedimensional images. 135 wood surfaces. 23. 49 transformation aﬃne. 98. 17. 27. 4. 29 wellposedness. 29. 30. 34. 45. 52. 56 discrimination. 61 visual system. 12 total variation. 20. 136 topological changes. 15. 11. 131 variance. 76. 46. 5. 28. 44. 43. 49. 40 Euclidean. 45. 26.
.
Their ﬂexibility allows to combine smoothing properties with image enhancement qualities.G.Anisotropic Diﬀusion in Image Processing Joachim Weickert University of Copenhagen Many recent techniques for digital image enhancement and multiscale image representations are based on nonlinear partial diﬀerential equations. B. A large number of references enables the reader to acquire an uptodate overview of the original literature. and it describes in a systematic way their theoretical foundations. but also for the algorithmically important semidiscrete and fully discrete settings. This book gives an introduction to the main ideas behind these methods. The presented examples range from applications in medical image analysis to problems in computer aided quality control. The central emphasis is on anisotropic nonlinear diﬀusion ﬁlters. and applications. Teubner Stuttgart . numerical aspects. A general framework is explored covering wellposedness and scalespace results not only for the continuous.
This action might not be possible to undo. Are you sure you want to continue?