基于Markov随机场的串行错误隐藏算法研究

Error Concealment Algorithms Based on Markov Random Field

Markov随机场是分析物理现场空间和背景依赖性的概率理论分支,在图像传输丢包后的错误隐藏过程中有很高的适应性;Markov算法在恢复丢失宏块的过程中,会利用已恢复像素作为参考像素来恢复其余像素,具有比传统空域错误隐藏更好的恢复效果;通过对4邻域Markov随机场算法的改进,该算法在并未大量增加计算复杂度的基础上具有更理想的恢复效果,具有很高的实际应用价值;实验结果表明,与传统空域错误隐藏算法和同类算法比较,该算法恢复效果更加良好。

Markov random field is to analyze the physical field space and background--dependent theory of probability, when the image transmission lost package use Markov to error concealment have very high adaptability. Markov algorithm in a process to recover the lost macro block, use the has been restored pixels as reference pixels to recover the remaining pixels, more better than the traditional airspace er- ror concealment. Through to tbe 4 neighborhood Markov random field improved, this algorithm isn＇t increase the computing complexity and has more better effect, it has a high application value. The experimental results show that, compared with the traditional airspace error con- cealment and similar algorithm this algorithm restore the effect is more good.