Detection of skin lesions by fuzzy entropy based texel identification

This paper proposes automated detection of skin lesions by unsupervised feature based clustering based on a new fuzzy entropy function for characterizing texture. The parameterized entropy function is optimized using the Bacterial Foraging algorithm. The clustering of the entropy function of the image is done using the popular Fuzzy C-means algorithm (FCM). The experimental results obtained after the clustering process indicate a very good segregation of texture clusters with satisfactory visual results. The results also provide us with the normalized entropy values needed for texel identification.