摘要
In this article, a spatio-temporal post-processing error concealment algorithm designed initially for a H.264 video-streaming scheme over packet-lossy networks has been presented. It aims at optimizing subjective quality of restored video and the conventional objective metric, peak signal-to-noise ratio （PSNR）, as well, under the constraints of low delay and computational complexity, which are critical to real-time applications and portable devices having limited resources. Specifically, it takes into consideration physical property of motion to achieve more meaningful perceptual video quality. Further, a content-adaptive bilinear spatial interpolation approach and a temporal error concealment approach are combined under a unified boundary match criterion based on texture and motion activity analysis. Extensive experiments have demonstrated that the proposal not only result in better reconstruction, objectively and subjectively, than the reference software model benchmark, but also results in better robustness to different video sequences.
In this article, a spatio-temporal post-processing error concealment algorithm designed initially for a H.264 video-streaming scheme over packet-lossy networks has been presented. It aims at optimizing subjective quality of restored video and the conventional objective metric, peak signal-to-noise ratio （PSNR）, as well, under the constraints of low delay and computational complexity, which are critical to real-time applications and portable devices having limited resources. Specifically, it takes into consideration physical property of motion to achieve more meaningful perceptual video quality. Further, a content-adaptive bilinear spatial interpolation approach and a temporal error concealment approach are combined under a unified boundary match criterion based on texture and motion activity analysis. Extensive experiments have demonstrated that the proposal not only result in better reconstruction, objectively and subjectively, than the reference software model benchmark, but also results in better robustness to different video sequences.
