Scribd
Upload
120 views2 pages

SVD and KL Transform-1

The document discusses two linear transforms: 1) SVD (Singular Value Decomposition) transform - decomposes a matrix into three matrices based on eigenvalues and eigenvectors. It is used for image processing tasks like compression, restoration, and recognition. 2) KL (Karhunen-Loeve) transform - also called the dimension reduction technique, depends on eigenvalues and eigenvectors of the covariance matrix. It transforms random signals/images to represent higher information in fewer dimensions and has applications in data reduction, rotation, and decorrelation.

Uploaded by

Deepak Singh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
120 views2 pages

SVD and KL Transform-1

The document discusses two linear transforms: 1) SVD (Singular Value Decomposition) transform - decomposes a matrix into three matrices based on eigenvalues and eigenvectors. It is used for image processing tasks like compression, restoration, and recognition. 2) KL (Karhunen-Loeve) transform - also called the dimension reduction technique, depends on eigenvalues and eigenvectors of the covariance matrix. It transforms random signals/images to represent higher information in fewer dimensions and has applications in data reduction, rotation, and decorrelation.

Uploaded by

Deepak Singh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

SVD AND KL TRANSFORM

SVD TRANSFORM

 SVD stand for Singular Value Decomposition Transform.


 Depends on the concept of Eigen values and Eigen vector.
 Data Dependent.
 The Singular Value Decomposition is widely used in image processing.
 SVD is an effective numerical analysis tool used to analyze matrices. In
SVD transformation, a matrix can be decomposed into three matrices
that are of the same size as the original matrix.
 From the view point of linear algebra, an image is an array of non
negative scalar entries that can be regarded as a matrix.
 If A is an square matrix then it satisfy a equation

Image
compression with SVD

AX= λ X

Where λ= Eigen value

X=Eigen vector

 In image processing Eigen Value Eigen Vector are useful to reduce matrix.
 If A is image matrix so it reduced by a transform by
D=N^T*A*N
 The column of matrix N are the normalized Eigen vector of image A.
 An image file can be compressed in many ways and one such way to compress a JPEG image file
is through Singular Value Decomposition.
 This JPEG method is commonly used for photographs.

APPLICATION OF S.V.D. TRANSFORM:-

 ✓ The process of Singular Value Decomposition can be used in many applications, including
watermarking an image, computing w least squares, and optimal prediction
 Image Restoration
 Image Compression
 Object Recognition

K.L.TRANSFORM :-

 KL TRANSFORM stand for Karhunen Loeve Transform.


 K.L. Transform is also depends upon the concept of Eigen Value and Eigen Vector.
 It is used for transforming given set of measurement to new set of featured.
 Featured show higher information packing properties.
 This technique is also know as dimensionally reduction technique.
 KL. Transform is applies to random signals/images and has wide applications in data reduction,
rotation and data de correlation applications.
 Multispectral image show larger co-relation therefore K.L Transform is commonly used.
 Now consider a population random vector

 X= [x1 x2 x3……….xn]
 The mean vector of the population is defined as:
Mx=E(x)-[ml m2........mn]^T
 The covariance matrix of the population is defined as
C=E((x-Mx)(x-Mx)^T)
 For M vectors of a random population, where M is large enough.

Mx=1/M2 xk

 Let A be a matrix whose rows are formed from the eigenvectors of the covariance matrix C of
the population.
 They are ordered so that the first row of A is the eigenvector corresponding to the largest Eigen
value, and the last row the eigenvector corresponding to the smallest Eigen value.
 We define the following Transform:
y=A(x-Mx)
 It is called the Karhunen-Loeve transform.

You might also like