Performance Comparison of Deep Learning Models in Brain Tumor Classification

Abstract

Accurate and timely detection of brain tumors is critical for successful treatment. Magnetic Resonance Imaging (MRI) is an essential tool that provides invaluable information for the recognition of different types of brain tumors such as glioma, meningioma, pituitary tumors and benign entities. However, distinguishing between these tumor types and taking preventive measures poses a significant challenge in the classification of brain tumors. Compared to traditional disease detection methods, artificial intelligence-based computer applications offer significant contributions to brain tumor detection. In particular, deep learning methods, which have gained popularity in disease detection through the analysis of medical images, play a critical role in this process. Several deep learning techniques have been reported in the literature for brain tumor classification. In this study, the YOLOv8s-cls model is used to detect brain tumors from MRI scans. The proposed model showed a high success rate of 98.7% accuracy during the experimental studies. The results show that the YOLOv8 model not only outperforms existing methods but also proves to be an effective approach for image classification.