Hybrid ResNet50-SVM Framework for Detecting Brain Tumors from MRI Images

Abstract

Brain tumors are abnormal cell proliferations that may develop within the brain and can be either non-cancerous (benign) or cancerous (malignant). They might arise primarily in the brain or spread to it from other regions through metastasis. The task of classifying brain images obtained from Magnetic Resonance Imaging (MRI) has gained significant importance in medical research. Recent studies increasingly adopt machine learning approaches to build predictive and diagnostic tools for healthcare applications. This study proposes a method for brain tumor detection using various MRI brain scans. Features are extracted by employing the ResNet50 deep convolutional neural network architecture. Subsequently, classification models based on Support Vector Machines (SVM) are implemented to perform tumor prediction.