Research Of Intra-prediction Mode Selection Algorithm In AVS And HEVC

With the fast development of social and economic, more and more multimedia devices appear in front of us, greatly change the way that we work and live. The multimedia device is equipped with a flood of data, audio and video information of which the proportion column maximum. So how to compress audio and video data processing vast amounts of multimedia technology development have become a major problem. The amount of digital data equipped with audio and video information is not easy to use and transport, so people urgently want audio and video coding technology to solve this problem.Although multimedia data contains a lot of data, many of the data is redundant, so it can be compressed. Video compression is that the redundant information of the audio and video data is removed. In other words, it represent the amount of information with the less data.Now the main audio and video coding standard AVS (Audio Video coding Standard) as a new generation of video coding standards is independently developed by our country. It is not only close to the performance and H.264, complexity is reduced by half, and royalties greatly reduced. International the latest standards high-efficiency video coding standard HEVC (High Efficiency Video coding Standard), which is an international standards organization ISO (International Organization for Standardization) and the International Telecommunication Union ITU-T (International Telegraph and Telephone Consultative Committee),the two organizations to jointly launch. HEVC improve in the respective series of modules and achieve a two-fold performance encoding of H.264. In AVS and HEVC intra prediction mode selection algorithm and method are based on sum of absolute differences SAD (Sum of Absolute Differences) or the rate distortion optimization method RDO (Rate-Distortion Optimization) to calculate the cost value, and use the RDO internal error squares and methods SSD (Sum of Squared Differences), these methods are based on mathematical calculations, and not effectively considering the subjectivity of the human visual system, so the elected best prediction mode is not very accurate sometimes.Structural similarity SSIM (Structural Similarity) is a recently proposed to image quality evaluation criteria, which take full account of the image brightness, contrast and structure information, more suitable for the human visual system. A large number of studies show that SAD and SSD method appliy to image relatively flat area, while SSIM is more suitable for evaluation of image detail more areas. In summary, this paper will introduce the SSIM into the AVS and the HEVC intra prediction mode selection algorithm, together with the SAD and RDO as the mode selection criteria values.By testing a large number of YUV video sequences,the experiment show that the PSNR of the reconstructed frames by new algorithms is improved than the original algorithm to some extent, and the rate is decreased to a certain extent. That will show that introduce the SSIM into AVS and HEVC intra prediction mode selection algorithm can effectively improve the AVS and HEVC encoder encoding performance and efficiency.