In this paper, the conventional H.264 Rate Distortion (RD) optimization algorithm is extended to improve error resilience for video transmission over error-prone channels. Novel RD cost functions are derived for both ...
详细信息
ISBN:
(纸本)9781612843490
In this paper, the conventional H.264 Rate Distortion (RD) optimization algorithm is extended to improve error resilience for video transmission over error-prone channels. Novel RD cost functions are derived for both motion estimation and mode decision by separating the end-to-end distortion into independent items. The modified cost functions involve additional items which are very simple and can be easily computed in a recursive way, hence low computation complexity can be achieved. Moreover, the proposed scheme maintains complete decoder compatibility and high RD performance. Extensive simulation results show that significant performance improvements are gained over compared algorithms. The computation complexity issue is also discussed.
This paper presents a novel error resilience scheme for wavelet scalable video coding. We use Wyner-Ziv codec to produce extra bits protecting the important parts of the embedded video streaming. At the same time thes...
详细信息
This paper presents a novel error resilience scheme for wavelet scalable video coding. We use Wyner-Ziv codec to produce extra bits protecting the important parts of the embedded video streaming. At the same time these bits also as the second description of important parts are transmitted over auxiliary channel to the receiver for error resilience. The errors in the embedded video streaming can be corrected by Wyner-Ziv description which regards the decoded frame as side information. Moreover, Wyner-Ziv decoder utilizes a coarse estimated version of the corrupted parts exploiting frame correlation in wavelet video decoder to generate a refine version. Simulation results show that our proposed method can achieve much better performance compared with Forward error Correction code. Meanwhile, this errorresilient algorithm can achieve 2-3 dB PSNR gains over the motion compensation error concealment.
This paper presents a novel error resilience scheme for wavelet scalable video *** use Wyner-Ziv codec to produce extra bits protecting the important parts of the embedded video *** the same time these bits also as th...
详细信息
This paper presents a novel error resilience scheme for wavelet scalable video *** use Wyner-Ziv codec to produce extra bits protecting the important parts of the embedded video *** the same time these bits also as the second description of important parts are transmitted over auxiliary channel to the receiver for error *** errors in the embedded video streaming can be corrected by Wyner-Ziv description which regards the decoded frame as side information. Moreover,Wyner-Ziv decoder utilizes a coarse estimated version of the corrupted parts exploiting frame correlation in wavelet video decoder to generate a refine *** results show that our proposed method can achieve much better performance compared with Forward error Correction code. Meanwhile,this errorresilient algorithm can achieve 2～3dB PSNR gains over the motion compensation error concealment.
An errorresilient algorithm based on error estimation is proposed for the latest video coding standard H.264. The algorithm first estimates the error propagation over the video frames, incorporating the channel condi...
详细信息
An errorresilient algorithm based on error estimation is proposed for the latest video coding standard H.264. The algorithm first estimates the error propagation over the video frames, incorporating the channel condition. Then it chooses intra coded macroblock adaptively according to the severity of macroblock corruption to stop the further accumulation of error. In the environment where the backward channel is available, the proposed algorithm is capable of utilizing the real macroblock error information from the decoder's feedback, tracking the error propagation, estimating the error more precisely, resulting in higher resilience performance. Simulation shows that under various channel conditions, the proposed algorithm outperforms the random intra update algorithm by 1.3-2.5 dB in the peek average signal to noise ratio of the reconstructed image. At the same time, it only requires rough channel condition estimation, and incurs relatively low computation and storage increase, which makes it feasible for implementation.
暂无评论