Video coding with optimal inter/intra-mode switching for packetloss resilience

摘要

Resilience to packet loss is a critical requirement in predictivenvideo coding for transmission over packet-switched networks, since thenprediction loop propagates errors and causes substantial degradation innvideo quality. This work proposes an algorithm to optimally estimate thenoverall distortion of decoder frame reconstruction due to quantization,nerror propagation, and error concealment. The method recursivelyncomputes the total decoder distortion at pixel level precision tonaccurately account for spatial and temporal error propagation. Thenaccuracy of the estimate is demonstrated via simulation results. Thenestimate is integrated into a rate-distortion (RD)-based framework fornoptimal switching between intra-coding and inter-coding modes pernmacroblock. The cost in computational complexity is modest. Thenframework is further extended to optimally exploitnfeedback/acknowledgment information from the receiver/network.nSimulation results both with and without a feedback channel demonstratenthat precise distortion estimation enables the coder to achievensubstantial and consistent gains in PSNR over known state-of-the-art RD-nand non-RD-based mode switching methods