摘要
针对H.264帧内错误隐藏忽略边缘纹理信息的问题,提出了一种改进算法。首先对丢失块的相邻块的像素作边缘检测,统计相关像素的个数,确定相邻块的边缘方向;然后通过距离和相关像素进行加权,以估计出丢失块每个像素的插值方向;最后,通过对宏块状态标志进行检测,选择插值方法,完成插值运算。在H.264标准测试平台JM86上对该算法进行验证,结果表明:在像素丢失块达50%的恶劣情况下,该算法的PSNR值比空间像素加权算法提高了2.11dB。
Aiming at the question of neglectinged edge texture information in intra frame error concealment of H. 264,this paper presented an improved algorithm. Firstly,detected the edge of border block pixels,counted the number of relevant pixels and confirmed the edge direction of border block. Then,estimated each pixel's interpolation direction of lose block according to distance and number of relevant pixels. Finally,chose the interpolation method of pixels and completed the interpolation operation based macro block status flag. Implemented it in JM86 of H. 264 standard test platform. Experimental results show that this algorithm can increase PSNR 2. 11 dB than space pixels weight algorithm when 50% blocks of a frame are lost.
出处
《计算机应用研究》
CSCD
北大核心
2010年第12期4798-4800,共3页
Application Research of Computers
基金
信号与信息处理重庆市重点实验室建设资助项目(CSTC
2009CA2003)
关键词
错误隐藏
边缘检测
方向加权和插值
仿真测试
error concealment
edge detection
directional weight and interpolation
simulation testing