Brain Tumor MRI Images Classification Using Fine-Tuned Deep Learning Models

Abstract

Accurate identification of brain tumors plays a crucial role in diagnosis and treatment planning. Magnetic Resonance Imaging (MRI) is one of the used methods in the detection of brain tumors, as it offers safety and provides good images. In this study, a fine-tuned deep-learning model for brain tumor classification was applied. In this respect, a publicly available dataset that consists of brain MRI images, which were labeled Glioma, Meningioma, Pituitary, and no tumor has been used. The methodology involved resizing the images, labeling the classes, shuffling the data, and splitting it into training and testing datasets. After that, two CNN models, ResNet101 and Xception, were fine-tuned on this dataset. The ResNet101 model evaluated an accuracy of 96%, while the Xception model evaluated an accuracy of 97%. These findings show how effectively one can identify these types of brain tumors using fine-tuned deep-learning models.