Research On The Acceleration Algorithms Of High Efficiency Video Coding Standard

With the continuous development of multimedia technologies, video communication in People’s Daily communication proportion increases year by year, and gradually becomes an indispensable communication tool in the information age. In recent years, video compression technology is rapidly developing because of the upgrading demand of video applications. In order to deal with the digital video formats in higher resolution or higher frame rate, the international standards organization ISO/IEC and ITU-T, working together in a partnership known as the Joint Collaborative Team on Video Coding(JCT-VC), released a new generation of video coding standard, namely, High Efficiency Video Coding(HEVC). Compared with the former generation of H.264/AVC standard, the coding efficiency of HEVC standard increases more than one times under the same coding conditions. In addition, it also significantly improves the video data parallel processing ability. However, the first application of many advanced coding technologies for HEVC standard leads to a sharp computational complexity increasement in the HEVC encoder, which has seriously hindered the popularization and popularization of it. Therefore, the research on the optimization of the HEVC encoder and the related fast algorithms becomes an urgent problem to be solved.To solve the above problems, this research focused on the improvement of algorithm of HEVC encoder technology for intra prediction coding and inters prediction coding. The mainly work of the paper is proposed as follows: 1. Fast algorithm research of HEVC intra prediction unit size determination.The intra prediction utilize the correlation between adjacent pixels in one frame, predict the current blocks to be encoded with the surrounding encoded pixels, thereby eliminating the original spatial redundancy of video signal. However, the intra prediction mode selection process is very complex because the HEVC encoder needs to divide the coding unit(CU) recursively. In order to reduce the complexity of intra prediction coding process, this paper firstly makes a pretreatment in each CTU unit of the video frames, we can reduce the CTU’s texture-complexity through building the corresponding saliency map; Secondly, the optimal PU size is determined via a scheme that implements an information entropy comparison among sub-blocks of saliency maps. Finally, the partitioning results in saliency maps are applied to the original CTU units, obtaining the proper partitioning result, which achieves the research on correlation algorithm of HEVC intra prediction unit size determination 2. Research on fast algorithm of HEVC inter prediction mode selection based on motion features.Inter prediction is based on the temporal correlation between adjacent frames in the video, using the pixel blocks in the encoded reference frame, and predicting the pixel block in the current frame by motion estimation and motion compensation techniques. In order to improve the inter prediction coding speed, first of all, a depth selection algorithm based on frame difference is proposed by using the difference between adjacent frames of HEVC, through non-zero detection rate of LCU level decision and non-LCU decision based on frame difference dispersion(FDD), each CTU will be contained in a reasonable coding depth allocation; secondly, we proposed an inter prediction mode selection method based on regional movement features(RMF), divide the CU into three kinds of motion areas in the optimal coding depth, reducing the candidate inter prediction modes in the traversal patterns and the computational complexity in HEVC encoder.In order to validate the performance of this algorithm, this paper build an experiment platform based on fast algorithm research of HEVC intra prediction unit size determination and research on fast algorithm of HEVC inter prediction mode selection. Through the comparision test between our proposed algorithm and HEVC standard, the comparation test between the coding performance in other previous fast algorithm and the algorithm proposed in the paper, we can make the conclusion that we proposed the acceleration algorithm of intra prediction coding and inter prediction coding of HEVC. Under the premise of guaranteeing the coding quality, the efficiency of the HEVC encoder is improved, and effectively reduces the computational complexity of HEVC encoding, which achieves the research on the acceleration algorithms of High Efficiency Video Coding Standard. Experimental results show that the proposed fast intra prediction algorithm achieves the 37.9% coding time saving with 0.62% bit-rate loss; while the proposed fast inter prediction algorithm achieves the 51.6% and 48.5% total coding time savings with 0.89% and 0.83% bit-rate loss under random access and low delay conditions. Above all, we can verify the effectiveness and practicability of our algorithm from two aspects of theory and experimental results.