An integrated application of multiple description transform coding and error concealment for error-resilient video streaming

Multiple description coding has been shown to provide improved error resiliency for video streaming over error-prone channels. In a typical multiple description codec, the decoder simply reproduces the signal using the received description(s). The optimal reconstruction is obtained from the available description in the sense that the mean squared error is minimized. However, this reconstruction operation does not consider the unique spatial and temporal properties underlying a video signal. In this paper, we propose a joint multiple description transform coding and statistical error concealment algorithm that adopts a maximum a posteriori estimation approach to compensate for the loss of descriptions. Our method considers not only the correlation between descriptions but also the spatial/temporal redundancy in video signals. Simulation results with several video sequences show that the proposed algorithm results in significant visual improvements over a conventional multiple description transform decoder.