Deep Convolutional Neural Network (DCNN) for Skin Cancer Classification

Skin cancer is one of the most threatening types of cancer, with an increasing rates throughout the decade. Detecting and classifying skin cancer in its early stages provides better chances for treatment. In the recent years, Convolutional Neural Networks (CNNs) emerged as a powerful solution that aids the diagnosis of skin cancer. In this paper, Human Against Machine (HAM) 10000 dataset is used to demonstrate skin cancer classification strategy. VGG16, VGG19, and a Deep CNN proposed in this paper are implemented, trained, and evaluated. The dataset pre-processing steps and methodology are illustrated, and the network parameters and training process are explained. The performance of all three networks are compared in terms of the average overall accuracy and loss.