Detecting Subtypes of Brain Tumors using Resnet50 with Transfer Learning

Abstract

To classify a Brain Tumor dataset consisting of 3064 T1 weighted contrast-enhanced brain MR (Magnetic Resonance) images, we used the CNN (Convolutional Neural Network) model ResNet50 Neural Network, one of the most popular deep learning architectures, in the proposed framework. We graded (classified) the brain tumors into three classes (Glioma, Meningioma, and Pituitary Tumor). Additionally, we will apply Transfer Learning to reduce the amount of data needed, improve neural network performance (most of the time), and save training time. Using a refined ResNet50 Neural Network architecture, our Brain Tumor Detector achieves over 95% accuracy by using the technique of Transfer Learning.