Study On Video Error Concealment Technology Based On AVS-M

Multimedia communication service is the core technology of next generation Internet and 3G mobile communication system, among them, the most important video communication has become one of the research focus in industry and academe. Due to the interference in transmission channel, network congestion and delay, most of the wired and wireless networks couldn't ensure to get the Quality of Service (QoS) at present, and the loss of video data during transmission is inevitable. Error concealment based on decoder neither needs to modify the source encoder at the coder terminal, nor influences transmission rate. Compared with other error resilient methods, it is more suitable for the application of real-time video communication. Therefore, the research on error resilient and error concealment technology has become particularly important in the video communication system.AVS standard is short for Information Technology Advanced Audio and Video Coding Standards, and it is the second generation source coding standard that belongs to our country's independent intellectual property. AVS-M is the seventh part of the AVS standard, which is oriented to satisfy the need of compression technology of moving images, and it is applied in video communication service of wireless mobile network and IP network. Consequently, the research on error concealment technology has practical significance, which can enhance the robustness of video transmission, and meet the need of transmission quality in all kinds of applications.The existing video error concealment algorithms(especially for H.264/AVC) are mainly discussed in this thesis and have been implemented on the AVS-M platform. Comparing with the experiment results, the optimal error concealment scheme is proposed for AVS-M as follows:(1)Spatial error concealment algorithm. These existing error concealment technologies have been implemented in this thesis, including bilinear interpolation algorithm(BI), directional interpolation algorithm(DI), Content -Adaptive Spatial Error Concealment(CASEC), Refined Content-Adaptive Spatial Error Concealment (RCA). By comparing and evaluating these algorithms with different standard video sequences, spatial error concealment algorithm that is suitable for AVS-M standard is put forward in the end of this part.(2)Temporal error concealment algorithm. This paper introduces the realization methods of these algorithms, including zero motion vector algorithm(TR), side matching algorithm(SMA), boundary matching algorithm (BMA) and two step multi-weighted boundary matching algorithm(TMBMA). Compared with other algorithms, it is proved by TMBMA simulation result that satisfactory concealment performance can be achieved.(3) Based on the study of the last two chapters, this paper designs an optimal error concealment scheme called ISEC(Integrated Spatio-temporal Error Concealment Approach) for AVS-M standard. It is according to a certain circumstance when intra I frame or inter P frame is lost, and has further improved the quality of the video decoder.