EZW Algorithm Using Visual Weighting in the Decomposition and DPCM

摘要

Some modifications of the EZW algorithm are described in the paper. First, the EZW algorithms using bi-orthogonal wavelet decomposition and quincunx wavelet decomposition are compared. After studying the statistics of wavelet coefficient. a DPCM quantization is used for coding the low pass comonent. In the EZW algorithm using biorthogonal wavelet transform the coefficients are also weighted using a visual weighting function. The visual percetpion of the result seems a little better but he PSNR is little worse. As a conclusion, this modification adds some complexity to the algorithm without a very significant result. However, modifications in the standard Shapiro EZW algorithm shed light on excellent performance of the recent Said and Pearlman EZW coder.