The H.264 Motion Estimation Technology

Block matching motion estimation (BMME) plays a very important role in video coding. The performance of BMME greatly affects the quality of the encoded bit stream.Full Search (FS)is the optimal algorithm for BMME. But the tremendous computation payload of it is for practical applications,especially for real-time unaffordable applications.To overcome this problem,a lot of fast algorithms were developed.An adaptive fast motion-estimation algorithm based on H.264/MPEG 4-AVC is researched in this paper and motion-vectors of the nearby macro-blocks is taken advantage of to find the original searching central position and determines the size of the searching zone. Compared with the former algorithms with fixed searching zone,this algorithm is more flexible and can combine with any other fast motion estimation algorithm.Based on an adaptive search range,the proposed algorithm makes use of the statistical characteristics between successive frames and the spatial and temporal correlations among the motion vectors of adjacent blocks to speed the process of motion estimation while maintaining the PSNR with no lose.Experimental results show that the algorithm can reduce computation amount variously by different motion-sequence under the premise don't reduce the PNSR.The simulation results show that the algorithm is able to fit for all types of video sequences adaptively in spite of the degree of the motions.Compared with other algorithms,it is more than 280 times faster than FS (Full Se arch)and more than 1.8 times faster than 3SS f3 Steps Search)in average with no PSNR lose.And it is a better one than the traditional fast motion estimation algorithms.such as N3SS (New 3 Steps Search)and 4SS (4 Steps Se arch).