Segmentation of skin lesions using an improved FLICM method

Abstract

In this paper a modified fuzzy approach is introduced to diagnose the skin damages in dermoscopy images. In this method, firstly the level of brightness on images is arranged by colored contrast modification; afterward, the edge of area is achieved by applying FLICM algorithm, which is modified by concept of complex Gaussian model approximation and FCM. Efficiency of this method is evaluated on real dermoscopy images which are taken from skin damages with different color and size. The presented parameter evaluation and their results are compared with the newest method of level set partitioning. Increasing amount of partitioning sensitivity in comparison with reliable methods, demonstrate the efficiency of the proposed method and its application in cad systems.