分析研究了图象分割质量的评价方法,将模糊集合的概念应用到分割评价方法中,引入模糊度测度.提出了一个简单有效的映射函数,可以快速有效地将图象从空间域变换到模糊性质域.对几种具有实际意义的图象分割质量评价参数及模糊度做了分析,提出了一个综合评价函数.实验结果表明:引入的模糊度比较准确地反映了分割图象的质量,其反映的分割图象质量差别与人的视觉效果基本一致.

Since different segmentation algorithms works well for different applications, quality evaluation of image segmentation is indispensable, and thus we introduce fuzzy measure into quality evaluation algorithm of image segmentation and propose a function that could change an image to the field of fuzzy property effectively. Moreover, a composite evaluation function that includes region uniformity, gray contrast, shape parameter and fuzzy measure is presented. The experiment results indicate that fuzzy measure can evaluate image segmentation quality correctly.