WAVELETS ANALYSIS (USING LIFTING SCHEME ( AND ITS APPLICATIONS TO IMAGE COMPRESSION

Abstract

In this thesis, analysis and synthesis of various Wavelet Transform [1,7,10] techniques with and without lifting scheme [15] has been studied and implemented. Further the implementation and analysis of performance of a “Discrete Wavelet Transform based embedded image codec” in MATLAB has been presented. This codec performs compression and decompression of still gray images in an efficient manner. The image codec consists of two major parts .The first one is encoder, which performs compression by applying Wavelet Transform [1,7,10] on image followed by quantization and encoding. Another part is decoder, which decompresses the image, and reconstructs it by de-quantization, decoding and Inverse wavelet transform. Main data loss occurs during quantization and encoding phase, hence for efficient quantization and encoding a SPIHT (Set Partitioning in Hierarchical Trees) coder [3] based on progressive transmission and embedded coding is implemented. The performance of the designed image codec is evaluated for different wavelets filters [6] at different bit rate. The coding gain of SPIHT coder in terms of distortion measures MSE and PSNR over different compression ratios i.e. bit rate is evaluated and an analysis is made. Further the effect on performance of the image codec is also analyzed for different wavelet filters at different levels of decomposition.