基于小波和高斯-马尔可夫随机场的纹理分割

Method of texture segmentation based on wavelet-transform and GMRF

为了提高纹理分割的准确性和区域一致性,降低分割的错误率,在文献的基础上,提出了一种基于小波和高斯-马尔可夫随机场(GMRF)的纹理分割方法。该方法首先对图象进行Gabor小波分解,得到一系列分辨率不同的子图象,然后采用基于GMR的K-均值聚类算法从最低分辨率图象进行聚类,一直到最高分辨率为止,这样就得到一个原始图象的初始分割,最后引入特征加权算法,进行后分割,得到最终分割结果,并对仿真结果与文献的算法进行了比较,表明该算法是比较有效的。

In order to improve the accuracy and region homogeneity as well as to reduce the error rate in texture segmentation, the paper proposes a novel approach based on wavelet and gaussian markov random field (GMRF). The original image is first processed with Gabor wavelet-transform, then processes with GRMF-clustering algorithm for getting the pre-segmented image, and at last uses feature weighting for processing the above pre-segmented image. As a result, the present approach shows visible improvements both in reducing the segmentation error and in improving the precision, comparing to the method which is proposed in the paper [5].