基于高斯马尔可夫随机场混合模型的纹理图像分割

Textured Image Segmentation Based on Gauss Markov Random Field Mixture Model

针对以高斯马尔可夫随机场中的邻域像素互作用参数为特征、以高斯混合模型为分类器的二步纹理图像分割方法,提出一个两者相互结合的一步模型———高斯马尔可夫随机场混合模型,并给出其EM算法的迭代计算公式。利用该模型进行纹理图像的分割实验,发现该算法在纹理图像分割的精度上比前者有较大程度的提高。

In the field of textured image segmentation, some propose a two step textured image segmentation method by taking interacting parameter of neighbor pixel in Gauss Markov Random Field as features, taking Gauss Mixture Model as classifier. However, we propose a one step model—Gauss Markov Random Field Mixture Model and give their iterative formulae of EM algorithm. In the experiments of textured image segmentation, our algorithm is better than the former in accuracy of textured image segmentation.