Comparative Study of Wavelets for Image Compression with Embedded Zerotree Algorithm

摘要

This paper presents study of different wavelets using the Embedded Zerotree Wavelet (EZW) algorithm, and their performance is analyzed for the application of image compression. The EZW is specially designed algorithm which uses zero tree property of wavelet transformed image to arrange the coefficients. These coefficients gives the progressively improved image information in order to predetermined threshold. We have used Haar, Daubechies, Bi-orthogonal, Coiflet, and Symlets to perform discrete wavelet transform of a grayscale image. The effect of wavelet families has been analyzed on test images using measuring parameter: mean square error (MSE), peak signal-to-noise ratio (PSNR), maximum error, and compression ratio (CR). It is observed that using EZW algorithm, Coiflet and Symlet wavelet families produce uniform results in terms of MSE and PSNR.