TRANSFORMATION BASED IMAGE CODING

Abstract

Image require substantial storage and transmission resources, as these resources being limited in nature image compression techniques provide us a way advantageous to reduce these requirements. In brief image compression methods can be classified into 2 types. One is lossy methods and other one is loss less methods. For the first type of method, i.e. lossy method, some information from the original image is lost. While in other case, i.e. in lossless compression methods there is a perfect reproduction of the original image. Most popular available lossy image compression methods use a transform-based scheme in which image is transformed in to other domain. The most commonly used transforms today are the DCT, the wavelet transform, e.t.c. Here I have chosen wavelet transform to compress image because of multiresolution property of wavelets. After getting wavelet coefficients I had applied SPIHT (set partitioning in hierarchical trees) algorithm on wavelet coefficients to compress image. The SPIHT algorithm is one of the most successful algorithm for wavelet image compression. The type of wavelet used is important for compression.Here I have used biorthogonal wavelets. This is because by using biorthogonal wavelets perfect reconstruction is possible. Here 14 different biorthogonal wavelets and Haar wavelet are evaluated for their ability to compress image