A deep skin cancer classification approach using image and structured information

Abstract

In this paper we describe an approach to implement a hybrid skin cancer detection machine learning model that uses the combination of two models. The final concatenated model makes prediction based on the information it gains from the two models. The first model being the Convolutional Neural Network model that learns from the image’s dataset and the second Artificial Neural Network model that learns from the textual data associated with the images. Hence the final concatenated model uses these models to make predictions.