DeMe-Net Model to Diagnose Skin Cancer Disease from Skin Lesion

Abstract

Skin cancer is a quite common disease in the world and Melanoma is one of the serious types of skin cancer. It begins in melanocyte cells. Skin irritation or tanning due to UV radiation triggers changes in melanocytes, leading to uncontrolled growth of cells. It can potentially develop and spread if it is not treated early. The biopsy is a traditionally used method to diagnose skin cancer in which a sample of the skin lesion is sent to a laboratory for testing. This methodology of diagnosis is troublesome, tedious, and time-consuming. Deep Learning, a branch of artificial intelligence (AI) is one of the most promising researches and innovation areas in the field of medical imaging. In this research, we propose a DeMe-Net model to diagnose skin cancer with deep learning techniques. The overall accuracy of the model is 98.33% on the testing data and 99.85% on the training dataset.