PERFORMANCE BASED IMAGE COMPRESSION

Abstract

The rapid growth of digital imaging applications, including desktop publishing, multimedia, teleconferencing, and high-definition television (HDTV) has increased the need for effective and standardized image compression techniques. The purpose of image compression is to achieve a very low bit rate representation, while preserving a high visual quality of decompressed images. As use and reliance on computers continues to grow, so does the need for efficient ways of storing large amount of data. For example, someone with a web page or online catalog that uses dozens or hundreds of images will more than likely need to use some form of image compression to store those images. The purpose of image compression is to achieve a very low bit rate representation, while preserving a high visual quality of decompressed images. Compression reduces the storage and finds its potency and limitations. Transmission burdens of raw information by reducing the ubiquitous redundancy without losing its entropy significantly. The image manipulation that occupies a significant position in multimedia technology necessitated the development of JPEG compression technique, which has proved its usefulness Until recently, to minimize the blocking artifact, inherently present in JPEG at higher compression ratios, JPEG2000 is devised that makes use of wavelet function. In this work, a new approach to JPEG compression technique is proposed that enhanced the compression performances in comparison with aforesaid JPEG techniques. The new technique considers both Discrete Cosine Transform (DCT) based (DCT, SVD, BTC) and Discrete Wavelet Transform (DCT) based (PYRAMID, EZW) methods in the transformation and reconstruction sides for best performed algorithm. A rigorous comparison of the various compressions through quality components (PSNR, MSE). iv The proposed Algorithm select the best possible algorithm based on the decision parameter for image to achieve low mean square error (MSE), better peak signal to noise ratio (PSNR), a high Compression ratio (CR), while preserving good fidelity of decompressed image. MATLAB codes have been developed for all the possible combinations, separately.