Real-Time Elbow Fracture Detection on Mobile Devices: Performance and Limitations
This study investigates the feasibility and performance of deploying YOLOv11-based elbow fracture detection models on mobile devices for real-time clinical diagnosis. Motivated 3by clinicians’ poor diagnostic accuracy (54.4%) in interpreting elbow radiographs, we 4developed and evaluated a smartphone application capable of performing inference via 5three pathways: photo library analysis, direct camera capture, and live video streaming. Two YOLOv11 models were trained on approximately 1,100 elbow X-ray images using systematic hyperparameter optimisation and deployed in both FP16 and FP32 quantisation formats. On digital radiographs, Model 1 achieved strong performance with 93.4% 9accuracy, 92.7% F1-score, and 69.3% mAP@50, demonstrating computational efficiency 10with RAM usage below 0.29GB and CPU consumption under 25%. However, substantial 11performance degradation occurred during real-world camera-based testing, with F1-scores 12declining to 31 − 60.3% for static photographs and 28.8 − 43.1% for live detection. Independent validation on 214 external images yielded moderate classification accuracy (64 − 66%), highlighting generalisation challenges that may stem from limited dataset diversity. The study demonstrates that while YOLOv11 achieves clinically relevant accuracy 16on curated digital radiographs with mobile-friendly computational requirements, current 17training paradigms prove insufficient for robust camera-based deployment. Severe domain shift effects, environmental sensitivity, and class imbalance issues (fracture detection F1: 1969.2 − 81.6% versus non-fracture: 93.7 − 94%) represent critical barriers to clinical implementation. These findings emphasise that exciting laboratory metrics do not guarantee real-world utility and establish that safe clinical deployment requires diverse datasets that incorporate varied acquisition conditions, environmental compensation strategies, and enhanced fracture-detection sensitivity to meet patient safety standards.
