Explainable AI and Deep Learning techniques for Colon Cancer Detection

Abstract

Cancer is the most common disease in the world. Cancer is a disease in which some somatic cells grow out of control and spread to other parts of the body. Of the types of cancer, colon cancer is the third most frequent cancer in the world. Artificial Intelligence is being used to diagnose cancer and deep learning techniques are used for prediction. The motive of this paper is to propose a model to implement a deep learning-based Convolution Neural Network (CNN), InceptionResNetV2 Network, and InceptionV3 transfer learning model to predict cancer and non-cancerous. The model helps in the classification and prediction of two types of colon cancer: colon adenocarcinoma and colon benign cancer. We experimented with the histopathological image data set. Along with the deep learning model, Explainable AI (XAI) technique is used for the representation of the black-box model of deep learning.