基于多信息融合的纹理分割新方法

为了提高纹理分割的边缘准确性和区域一致性，该文综合考虑了纹理的统计和随机特征，提出了一种摹于GMRF和Gabor滤波得到的纹理特征进行信息融合后再进行纹理分割的新方法。实验结果表明该算法适应各种不同类型的纹理图像，与传统的只利用纹理的一种特征进行分割方法相比，该方法在边缘准确性以及区域一致性等方面均有明显的改善。

Texture Segmentation Based on Multi-feature Fusion

To improve the accuracy of boundary locations and region homogeneity in texture image segmentation, novel adaptive texture segmentation is proposed. In this model, a texture is considered containing both structural and stochastic feature components. These two components are represented by gaussian markov random field (GMRF) model and Gabor wavelet filters. Experimental results are presented and demonstrated that the new adaptive model can better represent a wide variety of textures and hence can lead to better texture segmentation results.