QoE Evaluation And Optimization Of Real-time Streaming Video

With the service of streaming media video, such as video conference and IPTV, the video QoE of end-user is influenced by many factors,such as packet loss, delay and encoded mode etc. Now the research focus has become how to correctly evaluate the QoE of real-time streaming media video and improve it. Combined with the standard of current video codec and the research methods of QoE, the thesis includes the QoE evaluation model of real-time streaming video. Besides, a novel strategy to improve the QoE is proposed further in this thesis. The main contents are as follow:1. In order to quantify the importance of the video data information, we focus on the coding prediction mode and QP in researching the standard of video-coding, H.264. Meanwhile, the influencing factors and evaluation methods of QoE are investigated, including the QoE controlling methods of streaming video.2. We analyze the importance of the information extracted from video data at the time of image rebuilt. We also research the influence of Quantization Parameters on the video distortion. At the same time, combined with the HVS(Human Visual System), the influence of video content complexity on the end-user QoE is measured. Finally, a no-reference QoE evaluation model for real-time streaming media video is proposed, and the results demonstrate the effectiveness of the model.3. We investigate the main QoE controlling mechanisms of the streaming media video, including FEC(Forward Error Correction), ARQ(Automatic Repeat re Quest) and HARQ(Hybrid Automatic Retransmission re Quest). Combined with the characteristics of real-time streaming video applications, it is more adapted for this scenario to use the method of FEC. Based on FEC and H.264, a layered protection strategy of the streaming media video data is proposed. By quantifying the importance of H.264 video data, the strategy makes sure that the more important the video data is, the higher priority it has to be transmitted. The experimental results show that the QoE scores of streaming video increase by 2.15 dB.