Prediction-based loss recovery for frame-level streaming video

Forward error correction (FEC) is a common error control technique to recover transmission losses with low delay overhead. However, enhancing or optimizing the loss recovery capability for video streaming results in the increased FEC processing latencies to violate the stringent delay constraints of video communications. Accordingly, a low-delay FEC scheme is proposed in this paper to minimize the additional FEC processing delay while achieving nearly the same video quality that can be obtained by the optimal FEC. To support the online FEC rate allocation, the proposed FEC scheme extends the optimal frame-level FEC with the frame size prediction technique to imitate the optimal allocation parameters of previous video data. The simulation results show that the proposed prediction-based FEC scheme can maintain low end-to-end delay and significantly improves the video streaming quality for delay-sensitive video.