Hamid Reza Vaezi Joze - Computer Engineering Department at Sharif University of Technology
Mansoor Jamzad - Computer Engineering Department at Sharif University of Technology

Many fuzzy clustering based techniques for image segmentation do not incorporate special relationships of the pixels, while most of fuzzy rule-based image segmentation tend to be application dependent.Generic fuzzy rule-based image segmentation algorithm (GFRIS) and its extensions introduced application independent techniques for fuzzy monochromatic image segmentation. These techniques define membership functions using predefined segmentation (manually or by clustering) and also they do not consider noise in the segmented region. This paper addresses the aforementioned problems by proposing an iterative fuzzy rule-based multi-chromatic image segmentation which is application independent and can iteratively produced membership functions in noisy images. A qualitative comparison is made between segmented result of this method and the popular fuzzy c-means (FCM) applied to noisy colored images. The result shows significant improvements of this method over FCM.

Color Segmentation ,Colored Image,Fuzzy Rule Base ,Image Segmentation