计算机工程 ›› 2008, Vol. 34 ›› Issue (14): 222-224.doi: 10.3969/j.issn.1000-3428.2008.14.079

• 多媒体技术及应用 • 上一篇    下一篇

基于MRF-MAP的运动矢量恢复差错掩盖算法

许志良1,王建华1,谢胜利2   

  1. (1. 深圳信息职业技术学院,深圳 518029;2. 华南理工大学电子与信息学院,广州 510641)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-07-20 发布日期:2008-07-20

Error Concealment Algorithm for Motion Vector Recovery Based on MRF-MAP

XU Zhi-liang1, WANG Jian-hua1, XIE Sheng-li2   

  1. (1. Shenzhen Institute of Information Technology, Shenzhen 518029; 2. College of Electronic & Information Engineering, South China University of Technology, Guangzhou 510641)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-07-20 Published:2008-07-20

摘要: 针对帧间编码模式下出现图像块丢失的情况,提出一种有效的时间域运动矢量恢复差错掩盖算法。把运动矢量场建模为高斯马尔科夫随机场,对丢失图像块的运动矢量采用最大后验概率方法恢复,其权值能够根据空间和时间信息而自适应选择。仿真结果表明,该算法在客观和主观上都能获得高质量的图像。

关键词: 运动矢量, 马尔可夫随机场, 最大后验概率, 差错掩盖

Abstract: Error concealment is an attractive approach to combat channel errors for video transmission. Aiming at recovering the damaged macroblocks which belong to inter-coded model, an effective error concealment algorithm for motion vector recovery based on MRF-MAP is proposed in this paper. The motion vectors field is modeled as Gauss-Markov Random Field(GMRF). The motion vectors of the damaged image macroblocks can be recovered adaptively by Maximum A Posteriori(MAP), and the weight is selected adaptively based on the spatial information and temporal information. Simulation results show that the proposed method offers significant improvement on both objective PSNR measurement and subjective visual quality of restored video sequence.

Key words: motion vector, Markov Random Field(MRF), Maximum A Posteriori(MAP), error concealment

中图分类号: