Research And Self-adaptive Motion Estimation Algorithm Of Inter Prediction In HEVC

The video compressing coding technology has stepped into the high way ofdevelopment with the innovation of information technology and the improvement ofdevices’ processing competency. In order to meet the emerging demand andchallenges, guided by the two major international standard organizations, the nextgeneration of video compression coding standard, known as HEVC (High EfficiencyVideo Coding), has been in the three-year stage of developing. HEVC standard will bepublished in January,2013.Compared with H.264/AVC, HEVC still utilizes the hybrid video codingstructure with many technical improvements and algorithm optimizations. Thosechanges offer obviously better performance in the process of high resolution video.The experimental results demonstrated the superiority of HEVC in videocompression.This thesis has made researches on the algorithms of HEVC and compared thecoding performance of HEVC and H.264/AVC. Experiments under differentresolution and motion features have been conducted and the analysis of theexperimental results testified the outperformance of coding capability andcompressing efficiency of HEVC.With the foundation of the relevant work and the research of the coding structureof HEVC, we proposed a self-adaptive motion estimation algorithm based on motionfeatures of image. By utilizing the algorithm, the motion vectors of the neighbor PUsand previous frame can be extracted and processed to self-adaptively prediction themotion search range of current PU. Experimental results showed that the proposedalgorithm can save40%motion estimation time under LC configuration, whichgreatly improved the performance and feasibility of HEVC.