A Method of Image Segmentation Based on the Fast Fuzzy C-Means Clustering and Rough Sets
The conventional Fuzzy C-Means (FCM) clustering algorithm has been widely used in automated image segmentation. However, it has two limitations,one is the convergent speed is very slow.the other is the segmentation is uncertainty because the images object pixels and background pixels have silimar characteristic value and membership grade,which leads to the discontinuousness and vagueness of image boundary regions. A new algorithm is proposed to restrain the limtations of FCM in this paper.At first the images are segmented by the fast FCM clustering,then applying upper approximation and lower approximation of Rough sets theory to describe the object and background of image, the image can be segmetnted accutately by introducing rough entropy to choice the suitable threshold. Experimental results indicate that the segmentations to the four types of images are perfect.
fast fuzzy c-means clustering rough sets rough entropy image segmentation
Yunsong Li Gang Zhang
Electronic information department Zhengzhou electronic power college Zhengzhou,China
国际会议
三峡
英文
2094-2098
2012-05-18（万方平台首次上网日期，不代表论文的发表时间）