结合GLWT和LBP提取纹理特征的图像分割

Combined with GLWT and LBP extract texture features for image segmentation

提出一种提取图像纹理特征的新方法:结合广义局部沃尔什变换(GLWT)和局部二值模式(LBP)。首先说明广义局部沃尔什变换的定义,然后计算并分析广义局部沃尔什变换系数,并选取识别性能较好的2阶矩作为纹理特征。结合LBP空间纹理特征和灰度特征,然后根据纹理特征采用模糊C均值算法对像素进行聚类。实验结果表明,结合广义局部沃尔什变换和LBP的纹理特征具有更好的鉴别性能,且计算简单。

A new group of texture features are presented in this paper： based on generalized local Walsh transform （GLWT） and local binary pattern（LBP） . The definition of GLWT is given. Then calculate and analyze the GLWT coefficients. And the 2nd order moment are selected as texture features. Combined with LBP and the gray space Characteristics, then all the pixels are clustered by fuzzy C means algorithm. The experimental results reveal that the texture features presented by Combined with GLWT and LBP have the best discriminating performance, and simple calculation.