Skin lesion classification using fourier descriptors of lesion borders

Abstract

Between cancer types, Melanom is the one of the most deadliest form. Although, dermoscopy increased rate of diagnosis of the disease, most of the time detection of the disease depends on experience of physicist. There is great effort to diagnosis of the disease using computerized techniques. In this paper, for that purpose, we propose new method that uses fourier descriptors. After the segmentation of dermoscopic images, they are classified using extracted fourier descriptors of lesion border.