Computer Engineering and Applications ›› 2006, Vol. 42 ›› Issue (12): 6-.

• 博士论坛 • Previous Articles     Next Articles

Reduction Method of Blocking Artifacts based on Markov Random Field


  1. 暨南大学电子工程系
  • Received:2006-01-17 Revised:1900-01-01 Online:2006-04-21 Published:2006-04-21
  • Contact: sunny




  1. 暨南大学电子工程系
  • 通讯作者: 石敏 littledeer littledeer

Abstract: The block-based discrete cosine transform(BDCT) has been a main compression technique in most of the international standards for image and video compression. One of the major drawbacks in BDCT is that the decoded image differs from visible blocking artifacts at low bit rates, which reduces the image visual quality. In this paper, a new deblocking algorithm based on maximum a posteriori estimation of Markov Random Field (MRF-MAP) is proposed in order to reduce blocking artifacts as many as possible, while preserving edge information. The mathematical expression is deduced to set the threshold for Huber function of MRF. The threshold obtained by this expression can provide good tradeoff between protecting image edge and improving image quality. The experiment results show the effectiveness of the proposed algorithm.

Key words: blocking artifacts, wavelet transform, Markov Random Field, maximum a posteriori estimation

摘要: 基于块离散余弦变换(BDCT)是国际上大多数图像、视频压缩标准的核心部分,它的主要缺点是在低比特率时其恢复图像的块边界上会出现明显可见的方块效应,降低了图像的视觉质量。本文提出了一种新的基于马尔可夫随机场最大后验估计(MRF-MAP)的块效应消除算法,目的是实现在尽量消除块效应的同时充分地保护图像的边缘信息。文中通过线性回归,给出了设置MRF算法中的Huber函数阈值的数学公式,由该公式得到的阈值可以在保护图像边缘和提高图像质量之间有较好的折中。仿真结果验证了本文算法的有效性。

关键词: 块效应, 小波变换, 马尔可夫随机场, 最大后验估计