Inter Prediction In Video Coding

Video coding technology for compression has been a main research topic, and the block-based hybrid video coding scheme is the most widely used coding framework. In order to obtain more efficient performance, several dual hypothesis prediction technologies based on the hybrid coding scheme are proposed in this thesis.This thesis first introduces the background of video coding, including the necessity and feasibility of video coding, and introduces several video coding standard Formulated by video coding standard workgroup along with the development process. Meanwhile, a brief comparison and analysis of the currently most cutting-edge video coding standard is also presented.Subsequently, this thesis gives a detailed analysis of one key technology in video coding-inter prediction technology. On the basis of the theoretical analyses to inter prediction technology, it gives introduction and analysis of several technologies that influence the performance of inter prediction. Furthermore, this thesis deeply studies the multiple hypothesis prediction technology and proposes several dual hypothesis prediction technologies, including linear dual hypothesis prediction mode and weighted skip mode based on the statistical characteristics of dual-hypothesis prediction. Experimental results show that2.9%bit-rate savings on average is achieved by linear dual hypothesis prediction mode and weighted skip mode. Finally, it comes to the summary and outlook.