Low computational complexity enhanced zerotree coding for wavelet-based image compression

The embedded zerotree wavelet (EZW) algorithm, introduced by J.M. Shapiro and extented by A. Said and W.A. Pearlman, has proven to be a computationally simple and efficient method for image compression. In the current study, we propose a novel algorithm to improve the performance of EZW coding. The proposed method, called enhanced zerotree coding (EZC), is based on two new techniques: adaptive multi-subband decomposition (AMSD) and band flag scheme (BFS). The purpose of AMSD is to change the statistics of transformed coefficients so that the coding performance in peak signal-to-noise ratio (PSNR) can be elevated at a lower bit rate. In addition, BFS is used to reduce execution time in finding zerotrees. In BFS the tree depths are controlled, therefore, many unnecessary comparison operations can be skipped. Experimental results show that the proposed algorithm improves the performance of EZW coding and requires low computational complexity. In addition, the property of embedded coding is preserved, which enables a progressive transmission.