A Framework of Scalar Quantization Multiwavelet Transform for Very Low Bit Rate Video Coding
JZ Chen,JL Zhou,SS Yu,JH Zheng,C Zhang,L Zhong
2003-01-01
Abstract:A very low bit rate video coding framework in which motion correlation between successive video frames is exploited in the multiwavelet transform domain is described in this paper. Some new techniques, such as spatial prediction in intra coding, 4Y4 discrete integer transform, 1/2 and 1/4 pixels interpolation, loop filtering to remove blocky artifacts, and content adaptive binary arithmetic coding (CABAC) are used in H.26L. The testing results show that H.26L can greatly outperform MPEG-4 ASP in both PSNR and Visual quality. However, the H.26L encoder costs too much time for CABAC is very time-consuming. Whereas, only four or two types of symbol are generated by zero tree wavelet coding so that the entropy coding can cost less time than CABAC. Accordingly, multiwavelet transform, and a novel adaptive quantization scheme are applied to the prediction error. Experimental results reveal 0.2-0.5dB outperform in coded PSNR at low bit rates over the state-of-the-art H.26L recommendation, with a considerable improvement in subjective reconstruction quality, while simultaneously supporting a scalable representation.