Bone Fracture Detection and Classification using Machine Learning

This project presents an advanced deep learning approach for automated bone fracture detection and classification using X-ray images. A YOLOv11-based convolutional neural network was trained on the GRAZPEDWRI-DX pediatric wrist dataset to distinguish between normal and fractured bones and classify fracture types. The model achieved a high [email protected] of 77.1% with an inference time of just 2.4 ms, demonstrating both accuracy and real-time capability. Comparative analysis with YOLOv8, YOLOv9-E, and YOLOv10 highlights YOLOv11's superior performance. The work emphasizes clinical relevance, suggesting potential integration into diagnostic workflows to support radiologists and improve patient outcomes.