Digital Image Processing

Morphological Image Processing

1
 Introduction

► Morphology: a branch of biology that deals with the form

and structure of animals and plants

► Morphological image processing is used to extract image

components for representation and description of region
shape, such as boundaries, skeletons, and the convex hull

11/17/2023 2
 Types of Morphological
Operations
► Erosion

► Dilation

► Opening

► Closing

11/17/2023 3
Motivation: Morphological Operations
Blood vessel detection in retinal images

4
 Application: Cell segmentation

11/17/2023 5
 Preliminaries (1)

► Reflection

The reflection of a set B, denoted B , is defined as

  {w | w  b, for b  B}
B

► Translation

The translation of a set B by point z  ( z1 , z2 ), denoted ( B) Z ,

is defined as
( B) Z  {c | c  b  z, for b  B}

11/17/2023 6
 Example: Reflection and Translation

11/17/2023 7
 Preliminaries (2)

► Structure elements (SE)

Small sets or sub-images used to probe an image under study for

properties of interest

OR

A particular shape/structure used to deal with similar shape/varying

regions of an image

11/17/2023 8
 Examples: Structuring Elements (1)

origin

11/17/2023 9
 Examples: Structuring Elements (2)
Accommodate the
entire structuring
elements when its Origin of B visits
origin is on the every element of A
border of the
original set A
At each location of
the origin of B, if B
is completely
contained in A,
then the location is
a member of the
new set, otherwise
it is not a member
of the new set.

11/17/2023 10
 Erosion

With A and B as sets in Z 2 , the erosion of A by B , denoted A B,

defined
A B  z | ( B) Z  A
The set of all points z such that B, translated by z , is contained by A, i.e.,
retain all elements of image A which completely contain the SE B, rest are
maked zero

A B  z | ( B ) Z  A   c

Take all possible translations of SE B, and take its intersection with the non-image,
comp A. If for any image element this is a disjoint set, it will remain part of the erosion
result and all image elements for which its not a disjoint set, it would be discarded, i.e.
11/17/2023 11
marked zero
Erosion:
Example

11/17/2023 12
Example
of
Erosion
(1)

11/17/2023 13
Example
of
Erosion
(2)

11/17/2023 14
 Dilation

2
With A and B as sets in Z , the dilation of A by B,
denoted A  B, is defined as

 
 A
A  B= z | B
z

 and A
The set of all displacements z, the translated B
overlap by at least one element, and all such points which
allow at least one overlap, will become part of A.
 
A  B  z |  B

  A  A
z  
Every common element between all the possible translations and reflectance
of 11/17/2023
SE B and image A (should not be disjoint) will become the part of the image A, i.e.
15
A
will contain the common elements
 Dilation:
Example

11/17/2023 16
 Examples of Dilation (1)

11/17/2023 17
 Examples of Dilation (2)

11/17/2023 18
 Duality

► Erosion and dilation are duals of each other with respect to

set complementation and reflection

 A  B
c 
 A B
c

and
 A  B
c 
 Ac  B

11/17/2023 19
 Duality

► Erosion and dilation are duals of each other with respect to

set complementation and reflection

 A  B  z |  B Z  A
c c

 z |  B Z  A  
c c

 z |  B Z  Ac  

 Ac  B

11/17/2023 20
 Duality

► Erosion and dilation are duals of each other with respect to

set complementation and reflection

    A  
c
 A  B
c 
 z| B
Z

 z |  B 
 A 
c

Z


 Ac  B

11/17/2023 21
 Opening and Closing

► Opening generally smoothes the contour of an object,

breaks narrow necks, and eliminates thin bulges

► Closing tends to smooth sections of contours but it

generates fuses narrow breaks and long thin gulfs,
eliminates small holes, and fills gaps in the contour

Opening->An erosion followed by a dilation

Closing->An dilation followed by a erosion
11/17/2023 22
Motivation: Morphological Operations
Blood vessel detection in retinal images

23
 Opening
The opening of set A by structuring element B,
denoted A  B, is defined as
A B   A  B B
OR
A  B   B Z |  B Z  A

Union of all translations of B that fit entirely in A.

11/17/2023 24
 Closing

The closing of set A by structuring element B,

denoted AB, is defined as
AB   A  B   B

11/17/2023 25
 SE could not fit Helps

Examples in this region separate two

regions

Opening: Removes completely the regions that can not contained the SE,
breaks thin connections, smoothens object contours
Closing:
11/17/2023 Smoothens the object contours, Joins narrow breaks, fills long 26

thin gulfs and fills holes smaller than the SE

 Example: Opening

11/17/2023 27
 Example: Closing

Smoothens edges, joins components attached through small

edges/gaps​
11/17/2023 28
11/17/2023 29
 Duality of Opening and Closing

► Opening and closing are duals of each other with respect

to set complementation and reflection

 AB 
c )
 (A  B
c

 A B
c )
 ( A B
c

11/17/2023 30
 The Properties of Opening and Closing

► Properties of Opening
(a) A  B is a subset (subimage) of A
(b) if C is a subset of D, then C  B is a subset of D  B
(c) ( A  B)  B  A  B
► Properties of Closing

(a) A is subset (subimage) of AB

(b) If C is a subset of D, then C B is a subset of DB
(c) ( AB )B  AB
11/17/2023 31
11/17/2023 32
 The Hit-or-Miss Transformation

Its useful to match specified configuration of pixels in an image such as

isolated foreground pixels

if B is structuring element pair,

B=(B1 , B2 )

B   B1 , B2 
B1 : object
B2 : background
A  B   A  B1   ( Ac  B2 )
11/17/2023 33
 Example:
► AIM-
How to identify
the locations
Of following
cross-shaped pixel
configuration

11/17/2023 34
Example
:

► How to locate upper left corner pixels of square objects in first image
using hit-or-miss transformation ?
► i.e., we want to locate foreground pixels that have east and south
neighbors (these are “hits”) and
► that have no north, northwest, west or southwest neighbors (these are
“misses”)
11/17/2023 35
 Some Basic Morphological Algorithms (1)

► Boundary Extraction
The boundary of a set A, can be obtained by first eroding A
by B and then performing the set difference between A
and its erosion.

 ( A)  A   A  B 

11/17/2023 36
 Example 1

11/17/2023 37
 Example 2

11/17/2023 38
 Some Basic Morphological Algorithms (2)

► Hole Filling
A hole may be defined as a background region surrounded
by a connected border of foreground pixels.

Let A denote a set whose elements are 8-connected

boundaries, each boundary enclosing a background region
(i.e., a hole). Given a point in each hole, the objective is to
fill all the holes with 1s.

11/17/2023 39
 Some Basic Morphological Algorithms (2)

► Hole Filling
1. Forming an array X0 of 0s (the same size as the array
containing A), except the locations in X0 corresponding to
the given point in each hole, which we set to 1. Let A be
the original image and B the Structural Element

2. Xk = (Xk-1 + B) Ac k=1,2,3,…

Stop the iteration if Xk = Xk-1

11/17/2023 40
 Example

11/17/2023 41
11/17/2023 42
 Some Basic Morphological Algorithms (3)

► Extraction of Connected Components

Central to many automated image analysis applications.

Let A be a set containing one or more connected

components, and form an array X0 (of the same size as the
array containing A) whose elements are 0s, except at each
location known to correspond to a point in each connected
component in A, which is set to 1.

11/17/2023 43
 Some Basic Morphological Algorithms (3)

► Extraction of Connected Components

Central to many automated image analysis applications.

X k  ( X k 1  B )  A
B : structuring element

until X k  X k -1

11/17/2023 44
11/17/2023 45
11/17/2023 46
 Some Basic Morphological Algorithms (5)

► Thinning
The thinning of a set A by a structuring element B, defined

A  B  A  ( A * B)
 A  ( A * B) c

11/17/2023 47
 Some Basic Morphological Algorithms (5)

► A more useful expression for thinning A symmetrically is

based on a sequence of structuring elements:

  
B  B1
, B 2
, B 3
,..., B n

where B i is a rotated version of B i-1

The thinning of A by a sequence of structuring element {B}

A  {B}  ((...(( A  B1 )  B 2 )...)  B n )

11/17/2023 48
11/17/2023 49
 Some Basic Morphological Algorithms (6)

► Thickening:

The thickening is defined by the expression

A  B  A   A* B

The thickening of A by a sequence of structuring element {B}

A  {B}  ((...(( A  B1 )  B 2 )...)  B n )

In practice, the usual procedure is to thin the background of the set

and then complement the result.

11/17/2023 50
 Some Basic Morphological Algorithms (6)

11/17/2023 51
 Some Basic Morphological Algorithms (4)

► Convex Hull
A set A is said to be convex if the straight line segment
joining any two points in A lies entirely within A.

The convex hull H or of an arbitrary set S is the smallest

convex set containing S.

11/17/2023 52
 Some Basic Morphological Algorithms (4)

► Convex Hull
Let B i , i  1, 2, 3, 4, represent the four structuring elements.
The procedure consists of implementing the equation:
X ki  ( X k 1 * B i )  A
i  1, 2,3, 4 and k  1, 2,3,...
with X 0i  A.
When the procedure converges, or X ki  X ki 1 , let Di  X ki ,
the convex hull of A is
4
C ( A)   D i
i 1

11/17/2023 53
11/17/2023 54
11/17/2023 55
 Some Basic Morphological Algorithms (7)

► Skeletons

A skeleton, S ( A) of a set A has the following properties

a. if z is a point of S ( A) and ( D) z is the largest disk
centered at z and contained in A, one cannot find a
larger disk containing ( D) z and included in A.
The disk ( D) z is called a maximum disk.
b. The disk ( D) z touches the boundary of A at two or
more different places.

11/17/2023 56
 Some Basic Morphological Algorithms (7)

11/17/2023 57
 Some Basic Morphological Algorithms (7)

The skeleton of A can be expressed in terms of

erosion and openings.
K
S ( A)   Sk ( A)
k 0

with K  max{k | A  kB   };
S k ( A)  ( A  kB )  ( A  kB )  B
where B is a structuring element, and
( A  kB)  ((..(( A  B)  B)  ...)  B)
k successive erosions of A.
11/17/2023 58
11/17/2023 59
11/17/2023 60
 Some Basic Morphological Algorithms (7)

A can be reconstructed from the subsets by using

K
A   ( S k ( A)  kB)
k 0

where S k ( A)  kB denotes k successive dilations of A.

( S k ( A)  kB)  ((...(( S k ( A)  B)  B)...  B)

11/17/2023 61
11/17/2023 62
 Some Basic Morphological Algorithms (8)

► Pruning

a. Thinning and skeletonizing tend to leave parasitic components

b. Pruning methods are essential complement to thinning and
skeletonizing procedures

11/17/2023 63
Pruning:
Example
X 1  A  {B}

11/17/2023 64
Pruning:
Example
X 2    X1 * Bk 
8

k 1

11/17/2023 65
Pruning:
Example

X3   X2  H  A
H : 3  3 structuring element

11/17/2023 66
Pruning:
Example

X 4  X1  X 3

11/17/2023 67
Pruning:
Example

11/17/2023 68
 Some Basic Morphological Algorithms (9)

► Morphological Reconstruction

It involves two images and a structuring element

a. One image contains the starting points for the
transformation (The image is called marker)
b. Another image (mask) constrains the transformation
c. The structuring element is used to define connectivity

11/17/2023 69
 Morphological Reconstruction: Geodesic
Dilation

Let F denote the marker image and G the mask image,

F  G. The geodesic dilation of size 1 of the marker image
with respect to the mask, denoted by DG(1) ( F ), is defined as
DG(1) ( F )   F  B   G

The geodesic dilation of size n of the marker image F

with respect to G, denoted by DG( n ) ( F ), is defined as
DG( n ) ( F )  DG(1) ( F )  DG( n 1) ( F ) 
with DG(0) ( F )  F .
11/17/2023 70
11/17/2023 71
 Morphological Reconstruction: Geodesic
Erosion

Let F denote the marker image and G the mask image.

The geodesic erosion of size 1 of the marker image with
respect to the mask, denoted by EG(1) ( F ), is defined as
EG(1) ( F )   F  B   G

The geodesic erosion of size n of the marker image F

with respect to G, denoted by EG( n ) ( F ), is defined as
EG( n ) ( F )  EG(1) ( F )  EG( n 1) ( F ) 
with EG(0) ( F )  F .
11/17/2023 72
11/17/2023 73
 Morphological Reconstruction by Dilation

Morphological reconstruction by dialtion of a mask image

G from a marker image F , denoted RGD ( F ), is defined as
the geodesic dilation of F with respect to G, iterated until
stability is achieved; that is,

RGD ( F )  DG( k ) ( F )

with k such that DG( k ) ( F )  DG( k 1) ( F ).

11/17/2023 74
11/17/2023 75
 Morphological Reconstruction by Erosion

Morphological reconstruction by erosion of a mask image

G from a marker image F , denoted RGE ( F ), is defined as
the geodesic erosion of F with respect to G, iterated until
stability is achieved; that is,

RGE ( F )  EG( k ) ( F )

with k such that EG( k ) ( F )  EG( k 1) ( F ).

11/17/2023 76
 Opening by Reconstruction

The opening by reconstruction of size n of an image F is

defined as the reconstruction by dilation of F from the
erosion of size n of F ; that is

OR( n ) ( F )  RFD  F  nB 

where  F  nB  indicates n erosions of F by B.

11/17/2023 77
11/17/2023 78
 Filling Holes

Let I ( x, y ) denote a binary image and suppose that we

form a marker image F that is 0 everywhere, except at
the image border, where it is set to 1- I ; that is

1  I ( x, y ) if ( x, y ) is on the border of I
F ( x, y )  
 0 otherwise
then
c
H   R ( F ) 
D
Ic

11/17/2023 79
 SE : 3  3 1s.

11/17/2023 80
 c
H   R ( F ) 
D
Ic
11/17/2023 81
 Border Clearing
It can be used to screen images so that only complete objects
remain for further processing; it can be used as a singal that
partial objects are present in the field of view.

The original image is used as the mask and the following

marker image:
 I ( x, y ) if ( x, y ) is on the border of I
F ( x, y )  
 0 otherwise
X  I  RID ( F )

11/17/2023 82
11/17/2023 83
 Summary (1)

11/17/2023 84
 Summary (2)

11/17/2023 85
11/17/2023 86
11/17/2023 87
 Gray-Scale Morphology

f ( x, y ) : gray-scale image
b( x, y ): structuring element

11/17/2023 88
 Gray-Scale Morphology: Erosion and Dilation
by Flat Structuring

 f  b ( x, y)  (min
s ,t )b
 f ( x  s, y  t )

 f  b ( x, y)  max
( s ,t )b
 f ( x  s, y  t )

11/17/2023 89
11/17/2023 90
 Gray-Scale Morphology: Erosion and Dilation
by Nonflat Structuring

 f  bN  ( x, y )  (min
s ,t )b
 f ( x  s, y  t )  bN ( s, t )

 f  bN  ( x, y )  max
( s ,t )b
 f ( x  s, y  t )  bN ( s, t )

11/17/2023 91
 Duality: Erosion and Dilation

 c 
 f b 
c
( x, y )   f  b  ( x, y )
 

where f c   f ( x, y ) and b  b(  x,  y )
 c 
 f b 
c
  f b
 

 
c
f  b  ( f c
 b)

11/17/2023 92
 Opening and Closing

f b   f  b  b
f  b   f  b  b

 
 f b 
c
 f b   f b
c

 
 f b
c
 f b   f b
c

11/17/2023 93
11/17/2023 94
 Properties of Gray-scale Opening

(a ) f  bf
(b) if f1f 2 , then  f1  b    f 2  b 
(c )  f bb  f b

where er denotes e is a subset of r and also

e( x, y )  r ( x, y ).

11/17/2023 95
 Properties of Gray-scale Closing

(a) f f b
(b) if f1f 2 , then  f1 b    f 2 b 
(c )  f b b  f b

11/17/2023 96
11/17/2023 97
 Morphological Smoothing

► Opening suppresses bright details smaller than the

specified SE, and closing suppresses dark details.

► Opening and closing are used often in combination as

morphological filters for image smoothing and noise
removal.

11/17/2023 98
 Morphological Smoothing

11/17/2023 99
 Morphological Gradient

► Dilation and erosion can be used in combination with

image subtraction to obtain the morphological gradient of
an image, denoted by g,

g  ( f  b)  ( f  b)
► The edges are enhanced and the contribution of the
homogeneous areas are suppressed, thus producing a
“derivative-like” (gradient) effect.

11/17/2023 100
 Morphological Gradient

11/17/2023 101
 Top-hat and Bottom-hat Transformations

► The top-hat transformation of a grayscale image f is

defined as f minus its opening:

That ( f )  f  ( f  b)
► The bottom-hat transformation of a grayscale image f is
defined as its closing minus f:

Bhat ( f )  ( f  b)  f

11/17/2023 102
 Top-hat and Bottom-hat Transformations

► One of the principal applications of these transformations

is in removing objects from an image by using structuring
element in the opening or closing operation

11/17/2023 103
 Example of Using Top-hat Transformation in
Segmentation

11/17/2023 104
 Granulometry

► Granulometry deals with determining the size of

distribution of particles in an image

► Opening operations of a particular size should have the

most effect on regions of the input image that contain
particles of similar size

► For each opening, the sum (surface area) of the pixel

values in the opening is computed

11/17/2023 105
 Example

11/17/2023 106
11/17/2023 107
 Textual Segmentation

► Segmentation: the process of subdividing an image into

regions.

11/17/2023 108
 Textual Segmentation

11/17/2023 109
 Gray-Scale Morphological Reconstruction (1)

► Let f and g denote the marker and mask image with the
same size, respectively and f ≤ g.
The geodesic dilation of size 1 of f with respect to g is
defined as

Dg(1) ( f )   f  g   g
where  denotes the point-wise minimum operator.
The geodesic dilation of size n of f with respect to g is
defined as

Dg( n ) ( f )  Dg(1)  Dg( n 1) ( f )  with Dg(0) ( f )  f

11/17/2023 110
 Gray-Scale Morphological Reconstruction (2)

► The geodesic erosion of size 1 of f with respect to g is

defined as

Eg(1) ( f )   f  g   g
where  denotes the point-wise maximum operator.

The geodesic erosion of size n of f with respect to g is

defined as

Eg( n ) ( f )  Eg(1)  Eg( n 1) ( f )  with Eg(0) ( f )  f

11/17/2023 111
 Gray-Scale Morphological Reconstruction (3)

► The morphological reconstruction by dilation of a gray-

scale mask image g by a gray-scale marker image f, is
defined as the geodesic dilation of f with respect to g,
iterated until stability is reached, that is,
RgD ( f )  Dg( k )  f

with k such that Dg( k )  f   Dg( k 1)  f 
The morphological reconstruction by erosion of g by f is
defined as
RgE ( f )  Eg( k )  f 
with k such that Eg( k )  f   Eg( k 1)  f 
11/17/2023 112
 Gray-Scale Morphological Reconstruction (4)

► The opening by reconstruction of size n of an image f is

defined as the reconstruction by dilation of f from the
erosion of size n of f; that is,

OR( n ) ( f )  R Df  f  nb 

The closing by reconstruction of size n of an image f is

defined as the reconstruction by erosion of f from the
dilation of size n of f; that is,

CR( n ) ( f )  R Ef  f  nb
11/17/2023 113
11/17/2023 114
 Steps in the Example

1. Opening by reconstruction of the original image using a horizontal

line of size 1x71 pixels in the erosion operation
OR( n ) ( f )  R Df  f  nb 
2. Subtract the opening by reconstruction from original image

f '  f  OR( n ) ( f )
3. Opening by reconstruction of the f’ using a vertical line of size 11x1
pixels

f 1  OR( n ) ( f ')  R Df  f ' nb '

4. Dilate f1 with a line SE of size 1x21, get f2.

11/17/2023 115
 Steps in the Example

5. Calculate the minimum between the dilated image f2

and and f’, get f3.

6. By using f3 as a marker and the dilated image f2 as

the mask,

R Df 2 ( f 3)  D (f k2)  f 3
with k such that D (f k2)  f 3  D (f k21)  f 3

11/17/2023 116