Skin Cancer Detection and Intensity Calculator using Deep Learning

Abstract

With time, a lot of genetic and metabolic abnormalities are now known to cause cancer, which is typically fatal. Any body part may become infected by cancerous cells, which can be fatal. Skin cancer is a condition which is pretty common and is gradually becoming more and more common in the world. There are several different types of skin cancers like Squamous and Basal cell carcinomas, as well as Melanoma known to be very lethal, this might result in fatal endings. Therefore, skin cancer diagnosis is very pivotal. Deep Learning is a technique that can aid in the prediction of skin cancer and can thus help save lives with early detection. The convolutional neural network (CNN) is an extraordinary algorithm in deep learning which outshines other traditional approaches. These networks help programmers in extracting spatially conserved discriminatory attributes from pictures, which can in turn be utilized for medical picture classification and area detecting. Using just image pixels and labels of diagnosis as inputs, this approach suggested a CNN architecture in this study to categorize skin lesions. In this study, the problem is approached with a CNN algorithm with 5 layers to categorize skin cancer utilizing picture pixels and labels of diagnosis as input data. After classification, the image is passed to another model which uses OpenCV color prediction to calculate the intensity of the tumor.