Abstract
The transmission of video stream over the error-prone channel is an indispensable subject to discuss in the present days. With the widespread of high-ratio video compression standard (such as H264/MPEG-4 AVC), high-definition video is the latest trend of real-time video stream. However, once the packet loss happened, it can cause serious error propagations during the transmission process. To solve this problem, we need to use the error concealment techniques, which utilize the information of remaining spatial and temporal relatively, to reconstruct the corrupted data. In our thesis, we introduce a game theoretic error concealment structure. Game theory is helpful in choosing the best strategy from two block modes for the lost macroblocks. The visual quality is the payoff in this game, which is evaluated from the PSNR (peak signal to noise ratio) value, and the computational time is the cost. We produce the best visual quality under the constraint of limited time. Edge information reflects the complexity of a region in a frame. We can use it as the criterion to make decision between temporal error concealment and spatial error concealment. With proper selection of error concealing method, we can have a good balance between the visual quality and the computational time.