Research On Video Communication And Scalable Video Coding Technology

Since ninety nineties, networked digital video applications are becoming more and more important in our everyday life along with Internet's developing. Storage is no longer an obstacle for video deployment. The target of video coding is changing from storage to transmission. Low bit-rate video coding, error resilience video coding, layered video coding and scalable video coding is becoming the researching hotspot. This dissertation makes a comprehensive research on scalable video coding based on the IP network communication. Mainly including the following content:1. Research on Error Resilient Video Coding algorithmAn efficient rate-distortion (RD) optimized macroblock mode decision algorithm for H.264-Scalable video coding (SVC) extension in packet loss environment is presented. The end-to-end distortion is estimated according to the hierarchy B picture coding structure of SVC and the error-propagated distortion stored in a distortion map. A parameter on temporal level is added to the RD optimized macroblock mode decision algorithm in current H.264-SVC test model to improve the error resilience performance of the encoded scalable streams. The simulation results show that the proposed method outperforms the original algorithm in current H.264-SVC test model.2. Research on FGS technologyThe FGS coding with adaptive reference (AR-FGS) scheme uses the weighted sum of the base layer reference picture and the enhancement FGS layer reference picture to form high quality reference, thereby it produces higher coding efficiency. In this paper, an improved AR-FGS coding method is presented. Leaky factor a is adjusted depending on the macroblock type of the current block and motion vector difference between the motion vector of the current block and those of the surrounding blocks. The simulation results show that the proposed method outperforms the original AR-FGS scheme.3. Unequal error protection using redundant picture for H.264-SVCVideo redundant coding is used for the video transmission based on IP network if the feedback channel isn't useful. A bitrate allocation optimized coding algorithm of redundant pictures (RL-RPC) based on H.264-SVC encoding structure was proposed to efficiently prevent temporal error propagation and to improve the error resilience of encoded bit stream. The bitrate allocation of redundant pictures is optimized based on the importance of the primary pictures with different space layers and time levels. The algorithm adds little additional codec complexity.4. Research on error conceal method of H.264-SVCSpace-layered H.264-SVC bitstream can make use of the correlation of different space layer picture when error concealment. Four error concealment algorithms for H.264-SVC are analyzed in details. It makes sense of practice applications.5. Research and realization on application system of scalable bitstreamA system of scalable bitstream is realized for typical application based on the research on video transport technology on IP network. The system can encode, extract and decode the scalable video. It can provide best-effort Internet and Diff Serve network environments. Different video encoding algorithms can be assessed on the system.