Detection and Differentiation of Skin Cancer from Rashes

Abstract

Skin cancer is a highly contagious disease that spreads over the human body very rapidly within a short period of time. Skin cancer and rashes may seem to be similar to each other in the eyes of a dermatologist so it is not easy to determine whether the skin is affected by rashes or skin cancer. Rashes are generic terms used by people and doctors to denote the changes in the skin such as skin infections, skin allergies, skin diseases. Dermatologist identifies skin cancer from rashes when skin does not heal over medication for a prolonged period of time and continues to spread rapidly to other parts of the skin meanwhile knowing the differences between rashes and skin cancer can help a person seek the necessary help or avoid anxiety about a noncancerous rash. The objective of the paper is to differentiate skin cancer from rashes and helps in detecting skin cancer in human skin. Pervious researches were successful in classifying the skin cancer types using image classification and this modal detects and differentiates skin cancer images from rashes images by using CNN[Convolutional neural network] and classify the images as skin cancer or rashes. The model is able to classify the image as skin cancer affected image or rashes image and obtained an average accuracy of 80.2% for 20 epochs.