Research On Inter Prediction In AVS-M

Our country has good industrial foundation in audio and video domain. However, because of the lack of mastery of the important techniques and standards, related corporations have long been enslaved to foreign corporations or organizations. To change this disadvantageous situation,"Digital Audio and Video Codec Technology Standard Work Group"(AVS) was established in Beijing on June 21,2002. AVS establish our country's video and audio coding standards, called Information Technology Advanced Audio and Video Coding Standards. AVS(Audio Video Coding Standard) is the first standard which is put forward by our country independently. AVS-M(mobile video)is the seventh part of AVS. It aims at next generation multimedia applications. And it has significance in multimedia communication, digital multimedia broadcast and portable consumer electronics etc.The design of AVS-M is based on conventional block-based motion-compensation hybrid video coding concepts. For the enhancement of the coding efficiency, AVS-M adopts new tools as followed: new intra frame prediction, variable block-size with seven block sizes in motion prediction, multiple reference picture, fractional-pixel accuracy for motion estimation, integer transform and loop filter. With these features, AVS-M can provide improved performance close to the newest international video coding standard H.264/AVC. At the same time, the complexity of the AVS-M codec is increased remarkably. Accordingly, the real-time video codec demands higher-powered processor and faster-speed algorithms. So the main endeavor of this paper is to achieve maximal processing speed and better video compression quality through optimization for inter prediction, in the AVS-M encoding process. The major works of the thesis are as follows.To reduce the complexity of inter prediction, a fast inter prediction mode decision algorithm is proposed. It classifies all modes by both homogeneity and temporal correlation of macroblock based on the statistic analysis of the results of the optimal modes. A adaptive threshold is contained by using the high correlation in temporal domains of video sequence. A selective block- size classes searching algorithm and a selective intra-coding mode searching algorithm are proposed, based on the adaptive threshold. So the computation of inter mode decision can be decreased efficiency and the coding speed can be enhanced.Motion estimation is the most important module in the AVS-M encoder. A fast fractional pixel motion estimation (ME) algorithm is proposed in order to reduce the computation of fractional pixel ME. More motion vector predictions are utilized to enhance the accuracy. Based on the single peak characteristic of fractional pixel error matching function, the matching distortion direction is predicted to exclude the little possible points. It also adopts adaptive threshold decision to stride over unnecessary search. It can efficiently decrease the number of search points and come to a good matching.