A NOVEL APPROACH FOR MELANOMA SEGMENTATION USING INFORMATION THEORY

Abstract

Skin cancer, melanoma has become a precarious issue in European continents. It can be easily cured if diagnosed and excised at very early stages. In this article we have focused our research on the lesion extraction from melanoma sample using entropy based approaches. Various ideal and general entropy functions viz Havrda, Shannon, Renyi, Kapur, Vajda respectively, are investigated for the evaluation of a randomness measure at each possible threshold in grey range. The results retrieved shows an outperforming behavior of the general entropy measure of Havrda function as compared to any other existing entropy functions.