摘要
提出了一种基于均值偏移的彩色图像分割算法。首先阐述了在CIE-LUV均匀彩色模型下均值偏移算法的基本原理,然后给出了在图像分割中的具体实现方法:选定一个像素,在适当的空间窗和色彩窗限定的特征空间中寻找模式点,实现窗口中心从选定点到模式点的偏移,重复此过程,直到找到稳定的模式点并用模式点的色彩值代替该像素,遍历所有像素,最终达到对所有像素进行聚类。通过两幅图像对算法进行检验,实验结果证明该算法对彩色图像具有良好的分割效果。
A color image segmentation algorithm based on mean shift was proposed. The basic principle of the mean shift algorithm under CIE-LUV color mode was demonstrated. The color image segmentation process methods were given. That is to select a pixel, find the mode in a given feature window and let the center of the window shift to the mode, repeat this operation until find a convergent mode, run this procedure for all the pixels and finally obtain the segmented image. Experiment results on two different color images show that the proposed algorithm can get excellent segmentation.
出处
《计算机应用》
CSCD
北大核心
2006年第7期1605-1606,1610,共3页
journal of Computer Applications
关键词
颜色空间
均值偏移
图像分割
color space
mean shift
image segmentation
