Research On Key Technology And Optimization In Video Coding

In more recent decades, information technology keeps developing constantly. The applicationof multimedia has become an indispensable part of people work life. In the multimedia information,more than 80% information the people get from outside is gained through the vision, so videoinformation is the most important. Due to restricted by the current transmission bandwidth andstorage capacity, the information amount is huge, we must compress the video information in orderto obtain high-definition video. Thus, video coding technology arises. In the past 30 years, theISO/IEC and ITU-T of two international standard organization developed a series of video codingstandard. Among them, H.264/AVC is the latest international video coding standard, and comparedwith the previous standard, H.264/AVC receives excellent code performance. The performance gainof H.264/AVC is calculated by adding the cost of complexity, its computational complexity of theencoding is approximately three times than the H.263, and two times about the decoding than theH.263. So it is necessary to use the optimization algorithm for some key technology, throughreducing computational complexity, making H.264/AVC more practical applications.First of all, this paper simply describes the video coding theory, summarizes from thefeasibility, the basic principle and the development of video coding standard, then introduces theH.264 coding technology and Open source software platform, and mainly describes two H.264 keytechnology—intra prediction mode decision and Rate Distortion Optimization(RDO) strategy.This paper makes the H.264 standard as the research goal, and the main research includes thefollowing three aspects: (1) We unscramble H.264 Open source software X264, and comprehend thetechnology—intra prediction mode decision and inter prediction mode decision in depth. It canprovide software support for latter research. (2) Through the use of edge detection algorithm basedon block, we improve the intra prediction mode decision algorithm. The experiment result shows,edge has a certain correlation with intra prediction mode, and on condition that the reconstructedimage quality is not decreased obviously, fast algorithm saves more than 70% time than the originalalgorithm. (3) We research RDO technology and RDO algorithm. On the basis of summarizingH.264 classic RDO algorithm, we make the image quality assessment standard based on thestructure similarity(SSIM) use in distortion evaluation of video coding RDO strategy, put forwardand design an high efficiency coding mode decision strategy which are applied in H.264 intraprediction mode decision. The experiment results show that, the RDO strategy based on SSIM hasthe better Rate Distortion curve (R-D curve) performance, that is to say on a given coding bit rateconditions, the reconstructed image quality is better.