Enhanced zerotree wavelet transform image coding exploiting similarities inside subbands

Abstract

This work presents a method for wavelet image coding employing similarity between and within subbands to increase the efficiency of data compression. This method is based on zerotrees of successive-approximation quantization of wavelet coefficients and arithmetic coding, with distinct coefficient scanning directions, variable maximum histogram frequency count and new coding conditioning. For coding 256/spl times/256 and 512/spl times/512 Lena test images, the PSNR results are, respectively, 37.43 and 40.24 dB, at 1 bit/pixel, which compare with the highest reported compression ratios.