Automated Multiclass Fracture Detection and Classification Using YOLO: Advancing Real-Time Diagnostic Accuracy

The role of fracture detection in medical diagnostics is very much related to timely intervention and minimized complications. This paper develops the YOLO-Object Detection Framework for auto-multiclass fracture detection/classification. YOLO's real-time detection capability as well as efficiency make it strong candidate to aid radiologists in clinical workflows. YOLOv11 model was tested, obtaining an F1 score of 0.78 on a dataset of annotated X-ray and CT images covering a wide range of fracture types. The comparison of performance with the latest state-of-the-art models, YOLOv9 and YOLOv10 is done, to benchmark improvements in detection accuracy and processing speed. Results indicate that YOLO is effective at detecting and classifying fractures with low computational overhead, which makes it suitable for real-world applications. This study establishes that YOLO has much to contribute toward improving the efficiency and lowering the workload on the part of radiologists for detection. However, overlapping features and model generalization issues across imaging modalities present areas of improvement to come through dataset expansion and possibly further improved training techniques that pave the way for detecting fractures in healthcare.