Segmentation of color image contour using fuzzy cluster analysis

Segmentation of edge contour is an important part of color image processing, however the existence of gradual change of color and edge scatters will badly affect the result of image processing. A classification algorithm based on fuzzy cluster analysis is put forward in this article for segmentation of edge contour. On the basis of the color aberration analysis in HSI color space, fuzzy similarity matrix is constructed and then made equivalent. Combined with the transitivity character of fuzzy equivalent matrix, the algorithm is further improved by adding special constraint conditions according to the characters of sudden change of edge scatters in the small-scale area. The algorithm is proved valuable and effective by experimental analysis to well solve the problems of color´s gradual change and edge scatters, and to get good result of color image processing.