一种H.264视频流自适应率失真优化编码算法  

An adaptive RD-optimal coding algorithm based on H.264 video stream

作　　者：李虓江[1] 陈抗生[2] 

机构地区：[1]华信邮电咨询设计研究院有限公司,浙江杭州310014 [2]浙江大学信息与电子工程学系,浙江杭州310027

出　　处：《电路与系统学报》2005年第6期115-118,92,共5页Journal of Circuits and Systems

摘　　要：为了提高编码视频流在丢包网络环境中的抗误码性能,目前比较常用的是采用帧内刷新算法。在率失真框架之内的帧内优化编码刷新算法,则被认为是更为直接和有效的解决办法。在视频编码标准H.264/JVT中采用的就是这种算法。然而由于没有考虑到信道丢包率对编码器端进行仿真解码次数的影响,从而导致在进行率失真优化编码时的计算量较大,编码耗时较长,严重影响了编码器的编码效率。基于以上分析,提出了一种改进的自适应率失真优化编码算法。将H.264标准率失真优化编码算法中计算解码器端视频帧期望失真度的代数平均值算法,改进为加权平均值算法。仿真实验表明,提出的算法可根据信道丢包率和模拟信道状态个数信息来自适应地决定编码器端进行仿真解码计算的次数,从而有效降低H.264标准率失真优化编码算法中的计算冗余和计算复杂度,节省编码耗时。在模拟信道状态个数默认为30个时,本算法最多可节省近55%的编码耗时。Intra refresh scheme is one of the most effective solutions to improve robust character of the coded video in packet-loss environment. More direct and complete solution is to use optimal intra refresh scheme within rate-distortion framework, which has been used in H.264/JVT standard. However this scheme does not consider the influence of packet loss rate on the computer times of simulative decoding in encoder. So the coding efficiency of encoder is decreased due to spending more coding time. In the proposed algorithm, weighted average algorithm is used to compute estimated distortion in decoder instead of algebraic average algorithm used in H.264/JVT standard algorithm. Simulation have shown that proposed scheme can adaptively choose the computer times of simulative decoding in encoder based on channel packet loss rate and simulative channel state information. Using proposed algorithm, the coding time can be saved up to nearly 55 percent under the same computing accuracy as H.264/JVT standard algorithm.

关 键 词：H.264建议 率失真优化编码 加权平均算法 自适应编码 

分 类 号：TN919.8[电子电信—通信与信息系统]