Color image processing control using fuzzy performance indicators

The aim of this paper is to propose an automatic segmentation method for color images. It is based on the evaluation of some performance indicators which are linked to the action parameters of the fuzzy c-means segmentation process. This is performed by using symbolic representations and a set of fuzzy rules. The setting-up of these rules is a translation of the expert strategy trying to get the best segmentation. Experimental results on textile images are presented to validate this approach.