基于交互势函数和均场参数估计的分层MRF模型的纹理图像分割

Texture image segmentation using hierarchical MRF model based on the interactive potential function and mean-field parameter estimation

针对分层马尔科夫模型在用期望最大(EM)算法进行参数估计时,隐变量之间相互作用导致求期望值较难的问题,将均场理论应用到GMRF模型的系数估计中,使得模型的参数可以在不使用窗函数的情况下仅通过简单的线性方程即可求出。而对于固定势函数和变权重势函数不能表达图像区域间节点交互关系的缺点,提出了一种基于贝叶斯传播算法的交互势函数。试验结果表明:本文算法分割后的图像不仅具有良好的区域性,而且区域内部平滑,改善了传统小波域分层马尔科夫模型在分割区域内部存在混分的现象。

To solve the difficult problem of expectations caused by interaction between hidden variables when using Expectation-maximization（EM）algorithm to estimate parameters for hierarchical Markov Random Fields（MRF）model,the mean-field theory is introduced into Gaussian-MRF（GMRF）model.Parameters can be estimated easily through simple linear equation in case of without window function.An interactive potential function based on Bayesian belief propagation algorithm is proposed to change the situation that the fixed or variable weighted potential function can not express the interaction of image regions.Experiments demonstrate that the proposed method not only has good regional classification but also smoothly internal region.In addition,the mixed and confused phenomenon of traditional hierarchical MRF is improved in wavelet domain.