Error resilience coding based on adaptive redundancy allocation and parameter estimation

Weiwei Xu,Xiang Tian,Yaowu Chen

DOI: https://doi.org/10.13245/j.hust.140614

2014-01-01

Abstract:Quality degradation occurs during transmission of video streaming over the error-prone net-work .A fast error resilience coding method was proposed to solve this ,different amount of redundant pictures were adaptively adopted ,and the optimized method was proposed to decline computational cost on calculating redundant picture number .By analyzing the feature of distortion and bit consump-tion ,the statistical module of rate-distortion cost was formulated as the function of encoding parame-ters ,and then the encoding parameters of redundant pictures with different redundant picture counts were limited based on the model of rate-distortion cost .The encoding parameters ,distortion and bit consumption of neighbor frames were used for the updating of statistical model .The optimized num-ber of redundant picture was finally chosen by estimating the influence of redundant picture number on the total rate-distortion cost .From experimental results ,the average PSNR (power signal-to-noise ra-tio) of the proposed method outperforms the state-of-the-art redundant coding method by about 1 dB over error prone channel .