A Hybrid Deep Learning Method with Transfer Learning and Support Vector Machine (SVM) for Brain Tumor Classification from MRI Images

Abstract

Accurate determination of the classification and severity of a cerebral neoplasm is imperative in formulating a course of therapy and predicting the likely outcome. Magnetic resonance imaging (MRI) has emerged as the preferred imaging modality for the diagnosis of brain tumors, owing to its exceptional sensitivity and specificity. The present study introduces a novel approach that combines deep learning and machine learning techniques to classify brain tumors from MRI images. The proposed model incorporates Convolutional Neural Network (CNN) layers into the Inception v3 base model for the purpose of feature extraction. The utilization of support vector machines (SVM) in classification has resulted in an output layer of the model that attains a maximum accuracy of 99.48% in the classification of a dataset comprising 5712 MRI images. The model’s performance was assessed using different evaluation metrics such as sensitivity, specificity, and F1 score. The findings indicate that the hybrid model suggested for the classification of brain tumors exhibits superior performance compared to existing state-of-the-art techniques. The clinical applicability potential of the model is underscored by its high precision and robustness. The proposed hybrid model’s precision and effectiveness in categorizing brain lesions from MRI images indicate the potential for the advancement of machine learning algorithms in the field of medical image analysis. The present study showcases the importance of deep learning and machine learning models in augmenting the precision and effectiveness of brain tumor diagnosis, thereby potentially influencing patient outcomes in a meaningful manner. viii