Analyzing Impact of Wavelet Thresholding for Image Compression

Abstract

Image compression plays a pivotal role in the applications of digital TV transmission, teleconferencing, remote sensing images, archiving medical images and multimedia communications. A digital image suffers from extensive storage and transmission resource requirements. Discrete Wavelet Transform technique has gained popularity owing to its ability of resolving less significant image details, thereby reducing size of an image. In this work, bi-orthogonal (bior1.3), reverse bi-orthogonal (rbio1.5) and discrete meyer wavelet (dmey) wavelets at various decomposition levels have been engaged to achieve compression. For truncation of wavelet transformed coefficients, images of true colour are tested against global and level dependent thresholding. A comparative analysis in terms of compression ratio, bits per pixel, PSNR and MSE is carried out using images from a text, facial and medical domains