Noise Estimation and Image Compression using Singular Value Decomposition

Abstract

We require to transmit and store many images in our day to day life through different applications .Less is the size of image , lower is the price required for transmission and storage. So we require to go for various data compression techniques to minimize the amount of space required by the image . The possible way to tackle this kind of problem is by using (SVD) where an whole image is considered as a matrix and then various kind of operations are performed on that matrix. One good approach to do this is to apply Singular Value Decomposition (SVD) on the matrix obtained from that image. It works as, digital image is given to SVD. The image can be refactored into three metrices by using SVD. The process of refactoration is carried out using singular values, the orignal image is represented by smaller values, which results in reducing of the storage space required by the image. Our goal here is to get a low size image while getting the important features which explain the original image properly but in reduced size. SVD can be used to any arbitrary, square, reversible and non-reversible matrix of m × n size.