H．264／AVC中基于感兴趣区域的计算资源的优化分配算法  

Optimal computing resources allocation strategy based on region-of-interest in H.264/AVC

作　　者：王晓龙[1] 崔慧娟[1] 唐昆[1] 

机构地区：[1]清华大学电子工程系,清华信息科学与技术国家实验室(筹),北京100084

出　　处：《清华大学学报（自然科学版）》2011年第11期1747-1750,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(60572081)

摘　　要：H.264/AVC视频编码标准在获得更高压缩效率的同时是以编码复杂度的提升为代价的。视频编码算法的高计算复杂度对于实时视频通信以及计算资源有限的应用场合都是巨大的挑战。该文提出了一种基于感兴趣区域的计算资源的优化分配算法。该方法利用人眼对图像不同区域敏感度的差异,根据不同区域的特点确定不同的计算复杂度级别和编码参数集合,将更多的计算资源优先分配给感兴趣区域,实现了计算资源的优化分配。实验结果表明:与标准算法相比,该算法在保证图像主观质量的同时减少了35%左右的编码时间,而客观质量仅下降平均0.2dB左右。The higher compression efficiency of H. 264/AVC comes at the expense of increased computational complexity. The computational complexity poses a great challenge for real-time video communications or computational resources-constrained applications. To solve these problems, an optimal computing resources allocation strategy was developed based on the region of interest （ROI） for video coding. The system takes advantage of the different sensitivities of human eyes in different regions to determine different levels of computational complexity and coding parameters based on the different region characteristics to allocate more computational resources to the region of interest. The encoding complexity is optimized with results showing that the method reduces the computational load by about 35% with objective quality degradation of about 0.2 dB while providing a similar subjective feeling compared with the standard method.

关 键 词：视频编码 计算复杂度 感兴趣区域 

分 类 号：TN919.8[电子电信—通信与信息系统]