一种基于分层率失真优化的容错性视频转码算法  

Error Resilient Video Transcoding Based on Layered Rate Distortion Optimization

作　　者：韩镇[1] 胡瑞敏[1,2] 陈皓[2] 李明[2] 

机构地区：[1]武汉大学计算机学院,武汉430072 [2]武汉大学国家多媒体软件工程技术研究中心,武汉430072

出　　处：《计算机学报》2009年第6期1152-1161,共10页Chinese Journal of Computers

基　　金：国家自然科学基金(60472040);部属应用创新计划(2005YYCXHBST114)资助~~

摘　　要：容错性视频转码能够在已压缩视频流中嵌入容错性工具,以增强视频流面向无线信道的抗误码能力.但是容错性工具的嵌入在降低失真的同时也会增加码率,因此需要进行率失真优化.针对这一问题,文中提出了一种基于分层率失真优化的容错性转码算法.该算法通过在帧层和宏块层分别对重同步标记的插入和帧内宏块的刷新进行不同粒度的优化来提高算法的灵活性,通过在帧层考虑帧内宏块刷新、帧内预测和运动矢量预测的影响来提高算法的精确性,通过在宏块层减少候选模式的数量来降低算法的复杂度.实验表明,该算法与基于单层率失真优化的算法相比,信噪比可获得0.6～1.1dB的增益,复杂度最多可降低25%.Error resilient video transcoding can insert error resilient tools in the compressed video to enhance error resilience of the video over wireless channels. However the error resilience insertion increases bit rate while reducing distortion. So the rate distortion optimization is required. In order to solve this problem, this paper proposes an error resilient video transcoding algorithm based on layered rate distortion optimization. This algorithm improves the flexibility by performing rate distortion optimization with different granularity for synchronization marker insertion and intra macroblock refresh on the frame layer and macroblock layer respectively. Moreover this algorithm improves the accuracy by accounting for the intra refresh, intra prediction, motion vector prediction on the frame layer and reduces the complexity by reducing the candidate macroblock modes on the macroblock layer. The simulation results show that the proposed algorithm improves PSNR by 0. 6-1. 1dB and reduces the complexity by about 25% compared with the algorithms based on single layer rate distortion optimization.

关 键 词：视频转码 容错性 率失真优化 重同步标记插入 帧内宏块刷新 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]