PERFORMANCE COMPARISON OF TRANSFORM BASED IMAGE COMPRESSION TECHNIQUES

Abstract

Image compression is among the most widely used technologies in the area of digital image processing. It has enabled the multimedia revolution to occur. Rising bandwidth needs for communication of images in the field of mobile, computer and internet environments has led to the advances in the field of image compression. Digital transform based compression method have relatively high speed, low memory requirements and provide complete reversibility. Transform based image compression methods works on converting spatial domain of an image into a domain, that presents image constituents according to their importance, further followed by application of suitable coding methods to eliminate data redundancy. This dissertation is mainly concentrated on the implementation of the four transform based image compression techniques which are JPEG using Discrete Cosine Transform. JIEG 2000 using Discrete Wavelet Transform (DWT) and SPII-IT based on DWT. Joint source channel coding (JSCC) based transform image compression technique is proposed here to add error robustness to the traditional SPIt-IT technique. The results of these techniques in terms of their Peak signal to noise ratio (PSNR) and Mean Structural similarity index metric (MSSIM) and Bits per pixels are presented. in ideal transmission system without any noise, JIEG 2000 is the best among the four techniques in terms of its PSNR and MSSIM. For AWGN corrupted noise channel, PSNR and MSSIM performance of JSCC SPII-IT is best among the four techniques over a range of SNR of the channel.