一种基于残差分析的H.264/AVC帧间块尺寸选择快速预测算法  

Residual Texture Based Fast Block-Size Selection for Inter-frame Coding in H.264/AVC

作　　者：王正宁[1] 彭强[1] 诸昌钤[1] 

机构地区：[1]西南交通大学计算机与通信工程学院,成都610031

出　　处：《中国图象图形学报》2005年第12期1584-1590,共7页Journal of Image and Graphics

基　　金：铁道部科技研究开发项目(2003X040-A)

摘　　要：为了进一步减小宏块的帧间预测误差,新的视频编码标准H.264/AVC采用了灵活多样的块划分类型,块划分尺寸包含16×16到4×4共7种情况。更准确的块匹配带来了编码效率的提高,同时也使计算复杂度急剧增加,当采用率失真全搜索算法遍历所有可能的块划分尺寸组合时更是如此。这一问题也成了影响H.264/AVC实际应用的因素之一,针对这一问题提出了一种基于残差纹理分析的快速块尺寸划分算法。首先通过一次宏块级的运动搜索预处理得到各个4×4块运动补偿后的残差;然后根据得到的残差纹理大小及分布情况对帧间块的划分尺寸做出快速选择。大量的实验验证了该算法在几乎不牺牲图像质量和压缩效率的基础上将这一过程的计算复杂度平均减少了50%。One of the new features adopted in video coding standard H. 264/AVC is the utilization of flexible block sizes ranging from 16 × 16 to 4 × 4 in inter-frame coding. The aim is to reduce the error due to fixed block size prediction within a macroblock. However, this feature requires the encoder to compute the rate-distortion cost for every available block size. As a result, the complexity and computation load of video coding increase drastically, which make it very difficult for practical applications. In this paper we proposed a residual texture based fast block size selection algorithm for inter-frame coding. Finally, we do a motion extimation（ME） for a macroblock and get the residual, then we predict the block size by analyzing the residual texture. Extensive simulations verify that the proposed method speeds up the block-size selection procedure by 50% without sacrificing picture quality and compression ratio. A comparison between the proposed algorithm and other algorithms is also orovided.

关 键 词：视频编码 H.264/AVC 块划分 预测残差 

分 类 号：TN941.1[电子电信—信号与信息处理]