Optimization On Motion Estimation Algorithms And Research On Entropy Coding Based On H.264

H.264/AVC was developed as a joint project of the ITU-T SG16 Q.6 Video Coding Experts Group(VCEG) and the ISO/IEC JTC1/SC29/WG11 Moving Picture Experts Group(MPEG),which has a higher compression encoding efficiency than MPEG-4 and H.263.The research of this thesis is based on the video coding techniques of international video coding Standard H.264.This thesis introduces the basic theory of video coding,the development and technical characteristics of two series of international video coding standards at first,then discusses the system framework,encoding process and several core techniques of video coding section of H.264 which include intra prediction, inter prediction,and entropy coding.The major research work of this thesis on motion estimation and entropy coding of H.264 was presented, including:Researching kinds of classical motion estimation algorithms such as Full search,Three steps search,Diamond search and so on.Then research of UMHexagonS algorithm was done deeply. The search pattern of UMHexagonS algorithm is symmetrical with the search center, which can avoid local optimization but has no a clear direction. The probability of local optimization is very small for stationary or slow sequence when UMHexagonS algorithm spends more search points.So,a spatial-adaptive method is proposed based on the characters of the neighbors'motion,which can identify a clear direction for current block.Experimental results show that without loss of PSNR the proposed method can decrease 27.2% search points at best and 5.5% at least as well as 13.2% ME-time at best and 4.8% at least comparing with UMHexagonS algorithm.Researching Context-adaptive Variable Length Coding and Context-based Adaptive Binary Arithmetic Coding deeply.The analysis between CAVCL and CABAC shows that CABAC has a higher compression rate than CAVLC, which is proved by our experiments.