基于运动区域划分的H.264低复杂度模式选择  

Low Complexity Mode Selection for H.264 Based on Motion Region Partitions

作　　者：韦耿[1] 刘文予[1] 李鹏飞[1] 

机构地区：[1]武汉光电国家实验室(筹)华中科技大学电子与信息工程系,武汉430074

出　　处：《计算机辅助设计与图形学学报》2008年第1期93-98,共6页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(60572063);教育部博士点基金(20040487009)

摘　　要：为降低H.264编码器的复杂度,提出基于运动区域划分的模式选择算法.首先利用已编码宏块的运动信息识别运动对象的内部区域和边界区域,然后利用已编码宏块的模式信息对边界区域的运动剧烈程度和细节复杂程度进行进一步的划分,从而将视频内容划分为运动性质不同的区域,并选择合适的模式对其进行编码.实验结果表明,该算法平均节省编码时间达47.95%,有利于实时应用;同时PSNR平均下降约0.06dB,不影响主观视觉效果.An efficient mode selection algorithm is proposed in the paper to reduce the complexity of H.264 encoder. By the algorithm, firstly the boundary and the interior of the motion object are identified with the motion vector information. Then the boundary regions are classified into different types of regions based on the coded mode information of the macroblocks （MBs） surrounding the current MB. And at last, the various motion regions with appropriate modes will be processed according to their characteristics. Experimental results demonstrate that the proposed algorithm can save the encoding time up to 47.95 % on average, to meet the requirement of real time applications, with a sacrifice of 0.06 dB degradation on PSNR generated but with no effect to the visual quality produced.

关 键 词：模式选择 低复杂度 区域划分 H.264 

分 类 号：TN919.8[电子电信—通信与信息系统]