摘要
传统的视频差错掩盖技术在处理含有剧烈或复杂运动的丢失宏块时效果不理想,该文在对立体视频编码预测方式统计分析的基础上,利用丢失宏块周围正确接收块和已掩盖块的运动和视差信息,将当前丢失宏块进行判别分类,根据不同类别做不同的处理,这就对丢失宏块的恢复更具有针对性,进而提高差错掩盖的质量和效率。实验结果表明,所提方法能够改善视频剧烈或复杂运动区域的掩盖质量。
Traditional video error concealment techniques perform poorly in handling loss MacroBlocks（MDs） with intense or complicated motion. Based on the statistic analysis of stereoscopic video coding prediction mode, this paper classifies the loss MBs into different categories according to the motion and disparity information of their neighboring correctly received or concealed MBs, and uses different methods to conceal different kinds of MBs. Therefor the recovery of loss MBs is more purposeful and efficient. Simulation results show that the proposed method can improve the concealment quality of intense or complicated motion region.
出处
《电子与信息学报》
EI
CSCD
北大核心
2012年第12期2809-2815,共7页
Journal of Electronics & Information Technology
基金
国家863计划项目(2009AA01A336)资助课题
关键词
立体视频
差错掩盖
运动补偿
视差补偿
子块掩盖
Stereoscopic video
Error concealment
Motion compensation
Disparity compensation
Subblock concealment
