Melanoma diagnosis from dermoscopy images using artificial neural network

Among all skin cancers, melanoma is the most serious and unpredictable type of skin cancer although it is less common. Up to now, skin biopsy is the most reliable way of diagnosing melanoma. To avoid this invasive and costly biopsy, melanoma detection from dermoscopy images has been introduced for last few decades. But it is very challenging due to low interclass variance between melanoma and non-melanoma images, and high intraclass variance in melanoma images. A new approach for diagnosing melanoma skin cancer from dermoscopy images based on fundamental ABCD (Asymmetry, Border, Color, and Diameter) rule associated with shape, size and color properties of the images is presented in this paper. Two new features related to area and perimeter of the lesion image are proposed in this paper along with the other existing features which are distinguishing between melanoma and benign images. Dull razor algorithm is applied for black hair removal from the input images and Chan-Vese method is employed for segmentation. The extracted features are applied to an ANN model for training and finally detecting melanoma images from the input images. 98% overall accuracy is achieved in this approach. This promising result would be able to assist dermatologist for making decision clinically.