YOLO-v9-YOLO-v11: Brain Tumor Performance Analysis Using MRI Images

Abstract

Early and accurate detection of brain tumors in Magnetic Resonance Imaging (MRI) scans is essential for effective treatment and improved patient survival. Although MRI is a widely used diagnostic tool, the manual interpretation of these images is often time-consuming and prone to variability among medical professionals. In biomedical imaging, deep learning methods have been increasingly explored to address these challenges, particularly for automated tumor detection and segmentation. This study compares three advanced object detection models—YOLO-v9, YOLO-v10, and YOLO-v11—evaluated on a dedicated brain tumor MRI dataset. Performance was assessed using multiple metrics, including precision, recall, F1-score, Intersection over Union (IoU), and mean Average Precision (mAP) at 0.5:0.95. In addition, Non-Maximum Suppression (NMS) and frames-per-second (FPS) rates were examined to determine efficiency and feasibility for real-time inference. The results demonstrate that all three YOLO-based architectures achieve high accuracy and efficiency, highlighting their suitability for medical image analysis. Among them, YOLO-v10 achieved the best balance of performance, recording the highest precision (0.95), a substantial recall value (0.88), and superior mAP scores. These outcomes confirm YOLO-v10’s advantage in providing reliable and consistent detection compared to YOLO-v9 and YOLO-v11. Overall, the findings underscore the potential of advanced YOLO models in supporting clinical decision-making by enabling faster, more precise, and automated brain tumor detection, thereby contributing to improved diagnostic reliability and patient outcomes.