Computer-Aided Diagnostic System for Classification and Segmentation of Brain Tumors Using Image Feature Processing, Deep Learning, and Convolutional Neural Network

This research aims at the detection of the tumor clusters, found in the brain and classifying the type of tumor using Convolutional Neural Network (CNN) using MR Images of the patient. The proposed technique/mechanism consists of several phases, namely, Acquisition, Refining, Segmentation, and finally the Classification. The image refining process includes several sub-processes such as Noise Removal and Edge Detection. Further, based upon the input of the end-user, the class variance value gets calculated from the extracted features for segmentation and gets stored in a matrix called the convolutional pattern. The developed system classifies the type of tumor that either it is malignant or benign using Neural Network and Deep Learning Algorithms. The Idea of this project is to understand how we can develop industry grade, doctor acceptable, and diagnosable correct; an engineered mechanism for evaluating tumor presence in the subject so that faster and better measures can be taken to provide a good cure to the patient at the early stage.