一种基于区域的彩色图像分割方法

A Method of Region-based Color Image Segmentation

提出一种基于区域的彩色图像分割方法,该方法首先选用适当的彩色空间对图像中的每个像素抽取颜色、纹理及空间位置等综合特征,形成基于像素的综合特征空间;利用模糊C均值聚类方法,在综合特征空间中进行聚类,利用模糊熵的原理获得最佳聚类的簇数目,得到初步的区域分割,最后利用连接原理对图像区域进一步分割。该方法还提供了丰富的区域特征。

This paper introduces a method of region- based color image segmentation. This method first extracts color, texture, and location features for each pixel to form integrative feature vectors by selecting suitable color space. Then, the integrative feature vectors are clustering with the fuzzy c- means clustering ,and the amount and the center of clusterings are decided by the proposed fuzzy entropy method. Each pixel is grouped and labeled according to its membership degree. Finally, the regions with the same label are segmented again according to the neighbor connection theory for pixels and a lot of the features which describe the regions are provided.