视频编码器中DCT域系数的变换编码研究  

Transform coding of DCT coefficients in video encoder

作　　者：王中元[1] 胡瑞敏[1] 田纲[1] 李明[1] 

机构地区：[1]武汉大学国家多媒体软件工程技术研究中心,湖北武汉430072

出　　处：《通信学报》2008年第6期133-136,共4页Journal on Communications

基　　金：国家自然科学基金资助项目(60472040)~~

摘　　要：通过分析视频编码器中DCT域变换系数的实验数据,发现它们之间存在较强的冗余度,所以提出对系数进行子带分解,并对每个子带展开进一步变换编码的思想。设计了对宏块中每个变换块的子带进行重组,以使互有冗余的系数更集中的子带划分方法;为消除二次变换计算的误差失配,采用Hadamard变换作为分析工具,并导出了整数正交Hadamard变换公式。实验结果表明,在基本不增加失真度的前提下,二次变换编码去掉了视频信号的频谱相关性,进一步提高了视频信号的压缩效率。工作有望为正在制订的国际视频编码标准H.265提供一种新的思路。The redundancy among discrete cosine transfonn（DCT） coefficients in video encoder was found very strong, a method for subband decomposition and transform coding was proposed. Each 8 × 8 DCT spectrum date of macroblock was regrouped to gather the same frequency bands. The Hadamard transform function was used and orthogonal integer Ha- damard transform formula was derived. Simulation results shows that the proposed approach reduces the spectrum redun- dancy and improves the compression efficiency further compared with conventional DCT coder, however the PSNR loss are almost invisible. This work is expected to contribute to the under drafted international video compression standard H.265.

关 键 词：DCT系数 视频编码 整数Hadamard变换 

分 类 号：TN919.81[电子电信—通信与信息系统]