Significance-linked connected component analysis for very low bit-rate wavelet video coding

In recent years, a tremendous success in wavelet image coding has been achi eved. It is mainly attributed to innovative strategies for data organizatio n and representation of wavelet-transformed images. However, there hare bee n only a few successful attempts in wavelet video coding. The most successf ul is perhaps Sarnoff Corp.'s zerotree entropy (ZTE) video coder, In this p aper, a novel hybrid wavelet video coding algorithm termed video significan ce-linked connected component analysis (VSLCCA) is developed for very low h it-rate applications. There also has been empirical evidence that wavelet t ransform combined with those innovative data organization and representatio n strategies can be an invaluable asset in very low bit-rate video coding a s long as motion compensated error frames are ensured to be free of blockin g effect or coherent, In the proposed VSLCCA codec, first, fine-tuned motion estimation based on the H.263 Recommendation is developed to reduce temporal redundancy, and ex haustive overlapped block motion compensation is utilized to ensure coheren cy in motion compensated error frames. Second, wavelet transform is applied to each coherent motion compensated error frame to attain global energy co mpaction. Third, significant fields of wavelet-transformed error frames are organized and represented as significance-linked connected components so t hat both the within-subband clustering and the cross-scale dependency are e xploited, Last, the horizontal and vertical components of motion vectors ar e encoded separately using adaptive arithmetic coding while significant wav elet coefficients are encoded in bit-plane order by using high order Markov source modeling and adaptive arithmetic coding, Experimental results on eight standard MPEG-4 test sequences show that for intraframe coding, on average the proposed codec exceeds H.263 and ZTE in p eak signal-to-noise ratio by as much as 2.07 and 1.38 dB at 28 kbits, respe ctively. For entire sequence coding, VSLCCA is superior to H.263 and ZTE by 0.35 and 0.71 dB on average, respectively.