Research Of Motion Estimation Search Algoirthm Based On H.264

H.264video coding standard was jointly launched by ITU-T VCEG and ISO/IECMPEG, also known as MPEG-4Part10or Advanced Video Coding. H.264standard hashigher compression performance and good network compatibility, gained widespreadapplication. Motion estimation is a very important part of the H.264standard and is used toreduce the time redundancy of continuous frames, which can more effectively to encodevideo sequence. But motion estimation takes large quantities of operation time, which is oneof the bottlenecks affect the coding efficiency and real-time performance, it has become ahot research topic in the field of video coding.This paper analyses H.264video coding technique and then studies the searchalgorithm for the motion estimation. The full search algorithm is improved using the searchtheory of sub-pixel, the Interlaced Hops Small Diamond search algorithm is proposed. Theexperimental results show that compared with the full search algorithm, Interlaced HopsSmall Diamond search algorithm to ensure the same PSNR and bit rate, can reduce morethan57%motion estimation time whether for small motion, medium motion or high motionsequences, verify the effectiveness of the algorithm.Aiming at the existing problem of UMHexagonS algorithm, an improvedUMHexagonS algorithm is proposed in this paper. The algorithm makes full use of thecenter-biased characteristics of motion vector and motion homogeneity, symmetrical-crosssearch pattern and multi-octagon search pattern with increasing sequence step are used toreduce the useless search point. The searching direction of multi-octagon is adaptivelydetermined according to the change trend of motion vector, and5×5full search is replacedwith16points numbersign search, and improves the coarse position of motion search.Theexperimental results show that compared with UMHexagonS algorithm, the improvedalgorithm can reduce about18.53%motion estimation time with only negligible change ofPSNR and bit rate, especially for high motion video sequences, it can reduce25.17%motionestimation time.