Embedded image coding using zeroblocks of subband/wavelet coefficients and context modeling

In this paper, we present a new embedded wavelet image coding system using quadtree splitting and context modeling. It features low computational complexity and high compression efficiency, thanks to joint utilization of two powerful embedded coding techniques-set partitioning and context modeling. With effective exploitation of strong statistical dependencies among the quadtree nodes built up from subband coefficients, the proposed algorithm substantially improves the coding efficiency of the existing set-partition coders for both lossy and lossless image compression. For example, our experimental results show that the new algorithm respectively outperforms zerotree-based SPIHT and quadtree-based SPECK by 0.7 dB and 1.0 dB at 0.5 bpp on an average.