A Computer-Aided Brain Tumor Detection Integrating Ensemble Classifiers with Data Augmentation and VGG16 Feature Extraction

Abstract

Early detection of brain tumors is important to increase the rate of complete recovery from it without risking the lives of patients. Nowadays, the medical domain aims to use magnetic resonance to achieve early detection of Brain Tumors (BT), as 40 out of 100 people survive their cancer for 1 year or more[6], therefore the early detection of the tumors helps in the recovery. Magnetic resonance imaging (MRI) and X-Ray images are used in the early diagnosis of BT to eliminate its spreading. In this paper, we build an ensemble classifier model that integrates data augmentation with the VGG16 deep-learning feature extraction model for early detection of multi-class brain tumor types of patient infection. We perform the BT classification using the proposed model on a dataset that has a multiclass classification (Glioma tumor, Meningioma tumor, No tumor, and Pituitary tumor), it will classify the type of the tumor if it exists in the MRI. Our model results in an accuracy of 96.8% using the proposed model.