Novel algorithm for CABAC based on motion vector difference
Kangshun Li, Jing Jin, Feng Wang · Jisuanji yingyong yanjiu · 2012
In H.264/AVC video compression standard,context-based adaptive binary arithmetic coding(CABAC) is mainly used in main profile and has high compression efficiency.At the beginning,this paper introduced and analyzed the theory of CABAC algorithm,the context model selection theory of each component which belonged to MVD,then it proposed an optimization CABAC algorithm(CABAC_1 algorithm) based on motion vector difference(MVD) which were full use of the relevance between the current block MV and the current block MVD,the relevance between the current block MVD and the adjacent encoded block MVD and the relevance between the vertical and horizontal component of MVD.Compared with the standard CABAC algorithm,the CABAC_1 algorithm not only can effectively decrease the encoding time and rate and ensure the visual quality of coding sequence,but also can effectively save the code stream(bit streams) of the inter coding frames.The experiments show that it can save about an average of 10% bit streams of P frames and about an average of 5% bit streams of B frames by using CABAC_1.Therefore,CABAC_1 algorithm is an effective optimization algorithm.