摘要
针对错误隐藏过程中传统边界匹配算法对边缘匹配的局限性,提出一种基于时域与空域平滑性的边界匹配算法。该算法引入加权的边界匹配误差:时域误差用于衡量时域的连续性;空域误差则反映边缘的连续性。实验结果表明,该算法有效解决了存在多方向边缘穿越受损区域情况下最优矢量的选择问题,使得恢复后的图像在边界处具有较好的连续性。
Aiming at the limitation of classical boundary matching algorithm for temporal error concealment, a new boundary matching algorithm based on both the spatial and temporal smoothness is presented. The proposed algorithm introduces a weighted distortion function, in which the temporal distortion is used to measure temporal continuity and the spatial distortion to reflect the continuity of boundary. Experiments show that the motion vector of the damaged region with multi-directional edge can be well recovered with the proposed algorithm.
出处
《电子技术应用》
北大核心
2008年第1期120-122,共3页
Application of Electronic Technique
基金
国家863高技术研究发展计划项目(2006AA01Z127)
国家自然科学基金项目(60572152)
陕西省自然科学基金项目(2005F26)
关键词
错误隐藏
运动矢量
边界匹配算法
边缘匹配
error concealment
motion vector
boundary matching algorithm
edge matching
