Computer-Aided Diagnosis System for Colorectal Cancer

Colon cancer is the third most common cancer worldwide. It is usually developed from colon polyps. Although polyps are initially benign, they can become malignant over time, so screening and early detection are of great importance. The goal of the research is to develop a Computer-Aided Diagnosis System that can help in the prevention and detection of colon cancer. The proposed diagnosis system has two components. The first part examines a novel dataset composed by both numeric (blood and urine analysis) and qualitative data (including tumor data, complications, incidents, etc.) and establishes a prediction. The second part applies image processing to a colonoscopy result for accurate tumour detection. The colonoscopy images are grouped using a Convolutional Neural Network (CNN) and segments the colonoscopy using a Region-based Convolutional Neural Network (RCNN). The system is designed to help specialists increasing the accuracy of diagnosis as well as to minimize the diagnosis time, thereby reducing the number of misdiagnosed or late diagnosed patients.