UNIT IV

Image Registration

Presented by
Mrs .S.Maria Seraphin Sujitha M.E,
Asst. Prof, ECE Dept,
St.Xavier’s Catholic College of
Engineering,
Chunkankadai.
 UNIT IV REGISTRATION AND IMAGE
FUSION
Registration - Preprocessing, Feature selection -
points, lines, regions and templates Feature
correspondence - Point pattern matching, Line
matching, Region matching, Template matching.
Transformation functions - Similarity
transformation and Affine Transformation.
Resampling – Nearest Neighbour and Cubic Splines.
Image Fusion - Overview of image fusion, pixel
fusion, wavelet based fusion -region based fusion.
 Image registration
 Image registration is the process of aligning two or more images of the same
scene. This process involves designating one image as the reference image,
also called the fixed image, and applying geometric transformations or local
displacements to the other images so that they align with the reference.

 Registration is the determination of a geometrical transformation that aligns

points in one view of an object with corresponding points in another view of
that object or another object. (or)

 Image registration is the process of transforming different sets of data into

one coordinate system. Data may be multiple photographs, data from
different sensors, times, depths, or viewpoints.

 Image registration seeks to align images taken in different times, or taken

from different modalities
 Cont..
Registration has applications especially in
 Medical Imaging
 Remote Sensing
 Entertainment
 Computer Vision
 Military
 Automatic Target Recognition
 Registration Algorithm classification

1. Intensity-based vs feature-based
2. Transformation based
3. Single- vs multi-modality methods
4. Similarity measures for image registration
5. Spatial vs frequency domain methods

One of the images is referred to as the moving or source and the others are
referred to the target, fixed or sensed images.
Image registration involves spatially transforming the source/moving
image(s) to align with the target image.
The reference frame in the target image is stationary, while the other datasets
are transformed to match to the target.
Intensity-based methods compare intensity patterns in images via correlation
metrics, while feature-based methods find correspondence between image
features such as points, lines, and contours.
 Transformation models

The first broad category of transformation models includes

1. Linear transformations, which include rotation, scaling, translation, and

other affine transforms.
Linear transformations are global in nature, thus, they cannot model local
geometric differences between images.
2.The second category of transformations allow 'elastic' or 'nonrigid'
transformations. These transformations are capable of locally warping the
target image to align with the reference image.
Nonrigid transformations include radial basis functions (thin-plate or surface
splines, multiquadrics, and compactly-supported transformations [3]), physical
continuum models (viscous fluids), and large deformation models (
diffeomorphisms).
 Spatial vs frequency domain methods

• Spatial methods operate in the image domain,

matching intensity patterns or features in images.

• Frequency-domain methods find the

transformation parameters for registration of the
images while working in the transform domain.
Such methods work for simple transformations,
such as translation, rotation, and scaling.
 Single- vs multi-modality methods

• Single-modality methods tend to register

images in the same modality acquired by the
same scanner/sensor type.
• Multi-modality registration methods tended
to register images acquired by different
scanner/sensor types.
 Transformation functions-Affine
transformation
• Image registration methods and algorithm provide the transformation of a
source image to target image.
• Target image form an atlas.
• Ideal cases the transformation from source to target image should be
reversible.
• Forward transformation f to map to source image A to the target image B.
• Reverse transformation g to
map to target image B to source
image A.
Cont..
 Cont…
• Suppose (x, y) and (x', y') are two spatial coordinate
systems input space output space .
• A geometric transformation T that maps the input
space (x, y ) to output space (x', y').
• i.e (x’, y’)=T(x, y)
T is called a forward transformation or forward mapping.
(x, y)= T-1 (x’, y’)
T-1 is called a inverse transformation or inverse mapping
 Geometrical transformations
 Geometric transformations modify the spatial
relationship between pixels in an image.
The images can be shifted, rotated, or stretched
in a variety of ways
 change digital video resolution
 correct distortions caused by viewing
geometry
 align multiple images of the same scene
 Affine transformation

Rigid body transformation is based on translation

and rotation operations.
 Two images of equal dimensions are registered by
applying a pixel by pixel transformation consistently
throughout the image space.
 In three dimensions, a rigid transformations based
mapping of a point vector x to x’ is defined by
x’=Rx+t
R=Ration operation
t=Translation operation
Cont..
a)Translation along x-axis by p:
x‘=x+p
y’=y
z’=z
b)Translation along y-axis by q:
x‘=x
y’=y+q
z’=z

c)Translation along z-axis by r:

x‘=x
y’=y
z’=z+r
d)Rotation along x-axis by θ:
x‘=x
y’=y
z’= -ysin

e)Rotation along y-axis by :

x‘=x
y’=y
z’= xsin

f)Rotation along z-axis by :

x‘=x
y’= -xsin
z’=z
 Cont…
Rotation operation R for x-y-z rotational order of
operation can be given as
 Cont..
• Affine transformation is a special case of rigid body transformation.
• It includes translation, rotation and scaling operation.
Scaling operation

x‘=ax
y’=by
z’=
Where a,b and c are the scaling parameter along x,y and z direction.
The affine transformation can be expressed as
x’=Ax
• A=Affine matrix that includes translation, rotation, scaling and
Operation.
• Overall mapping can be expressed
2-D Affine Transformations
 3-D Affine Transformations

• The last column must contain [0 0 0 1].

Example
Example
 Feature based registration
Similarity transformation

 Similarity measures for image registration

 Image similarities are broadly used in medical imaging.
 An image similarity measure quantifies the degree of similarity between
intensity patterns in two images.
 The choice of an image similarity measure depends on the modality of the
images to be registered
 Common examples of image similarity measures include

1)Mutual information and normalized mutual information are the most popular
image similarity measures for registration of multimodality images.
2)Cross-correlation, sum of squared intensity differences and ratio image
uniformity are commonly used for registration of images in the same modality.
• THANK YOU