Fast Motion Search Algorithm In Avs-m

In the last ten years, video coding technology plays a key role in areas like digital TV, HDTV, video conference, and multimedia communications etc. AVS-M video coding standard is developed by the Audio Video Coding Standard Working Group of China. AVS-M standard provides a good trade-off solution between complexity and coding efficiency and introduces a number of advanced video coding techniques. It also significantly enhances ability of error resilience and network adaptation. Research on the AVS-M video coding standard is of great importance in many fields. The thesis focuses on the encoding algorithm including motion estimation and fast block match algorithm used in AVS-M framework. First principles of video coding and key features of AVS-M standard are introduced and then normally used motion estimation and compensation fast search technologies are researched. After that, based on these fast search algorithms several reformative algorithms are proposed, the following are included:(1)Threshold-adaptive early termination algorithm: when the cost of the block matching is smaller than the threshold which is adaptive according to the pictures to be coded, it is believed that the block is the best-matching block of the being coded block and motion estimation can be stopped.(2)Block-matching simplifying algorithm: the block-matching criterion is simplified.(3)Search-range-adaptive algorithm: because of the difference of the motion degree of the being coded pictures, we can use the information of the coded pictures to predict the needed search range of the next frame or macroblock adaptively. Simulation results show that the proposed algorithms are much better than other fast search methods whether or not the image sequence contained fast or slow motion. The computational payload is much depressed without evident quality loss.