Fracturespot: YOLO-Powered X-Ray Detection System

Abstract

Bone fracture detection is a crucial aspect of early cancer diagnosis, and deep learning-based object detection models have shown promising results in medical imaging. This study explores the application of YOLO (You Only Look Once) models for fast and reliable bone fracture detection. The dataset used in this research is the Bone Fracture Detection dataset from Roboflow, which includes annotated medical images to train and evaluate the models. Four YOLO variants—YOLOv5x6, YOLOv5s6, YOLOv8n, and YOLOv9n—are implemented and compared based on their efficiency and ability to detect fractures accurately. Performance evaluation metrics such as precision, recall, and mean Average Precision (mAP) are used to assess model effectiveness. Among the tested models, YOLOv8n achieved the highest precision of 90.9%, demonstrating its superior capability in detecting bone fractures accurately. These models are optimized for real-time medical image analysis, ensuring quick and precise fracture identification. The study aims to identify the most effective YOLO variant for detecting bone fractures, balancing speed and reliability. The results demonstrate that advanced YOLO architectures significantly improve early fracture detection, aiding in timely diagnosis and treatment planning. This research contributes to the biomedical field by enhancing automated fracture detection methods, potentially reducing human error and improving healthcare outcomes.

Introduction

Bone fractures, including benign and malignant types (bone cancers), require early and accurate detection to improve treatment outcomes. Traditional diagnostic methods like X-rays, CT, and MRI depend on expert interpretation but can be time-consuming and error-prone.

Recent advances in AI, particularly deep learning using Convolutional Neural Networks (CNNs), have transformed medical imaging by enabling automated, real-time, and precise detection. The YOLO (You Only Look Once) family of models stands out for its speed and accuracy in object detection, making it highly suitable for bone fracture detection.

This study explores applying several YOLO variants (YOLOv5x6, YOLOv5s6, YOLOv8n, and YOLOv9n) to detect and localize bone fractures from X-ray images using a publicly available annotated dataset. Techniques such as image preprocessing, augmentation, and training optimize model performance. Results show that these models achieve high precision, recall, and mean average precision (mAP), with YOLOv5s6 and YOLOv8n performing best in different metrics.

The integration of YOLO models into real-time medical workflows can reduce diagnostic errors, speed up fracture detection, and aid early intervention, thus enhancing healthcare delivery and patient outcomes.

Conclusion

Bone fracture detection plays a critical role in early cancer diagnosis, where timely identification can significantly improve treatment outcomes. Deep learning-based object detection models, particularly YOLO (You Only Look Once), have demonstrated their effectiveness in medical imaging for rapid and precise fracture identification. This study explores the application of YOLOv5x6, YOLOv5s6, YOLOv8n, and YOLOv9n to develop an efficient bone fracture detection system. The Bone Fracture Detection dataset from Roboflow is used to train and evaluate these models, ensuring robust learning and accurate predictions. The dataset undergoes preprocessing techniques like augmentation and normalization to enhance detection performance. Each YOLO variant is assessed for its efficiency, detection speed, and reliability in fracture localization. The results indicate that advanced YOLO architectures significantly improve the accuracy and speed of bone fracture detection, making them suitable for real-time medical applications. Performance metrics such as precision, recall, and mean Average Precision (mAP) were used for evaluation, with YOLOv8n achieving the highest precision of 90.9%. By automating the detection process, this system reduces dependency on manual diagnosis, minimizing human errors and assisting healthcare professionals in early cancer identification. The study highlights the potential of deep learning in revolutionizing medical diagnostics, offering a practical and effective solution for fracture detection. Future workcan focus on integrating this system with clinical workflows, further refining model accuracy and real-world applicability. Expanding the dataset with diverse medical images and optimizing model architectures can enhance system robustness. This research contributes to the advancement of AI-driven healthcare solutions, improving diagnostic reliability and patient outcomes.

References

[1] Lu, W., Zhang, W., Liu, Y., Xu, L., Fan, Y., Meng, Z., &Jia, Q. (2024). MFP-YOLO: a multi-scale feature perception network for CT bone metastasis detection. Medical & Biological Engineering & Computing, 1-12. [2] Ahmed, A., Imran, A. S., Manaf, A., Kastrati, Z., &Daudpota, S. M. (2024). Enhancing wrist abnormality detection with yolo: Analysis of state-of-the-art single-stage detection models. Biomedical Signal Processing and Control, 93, 106144. [3] Zou, J., &Arshad, M. R. (2024). Detection of whole body bone fractures based on improved yolov7. Biomedical Signal Processing and Control, 91, 105995. [4] Zhou, T., Wang, H., Chen, K., Zhang, Z., Chai, W., & Lu, H. (2025). Mandible-YOLO: The fracture region is detected only once. Biomedical Signal Processing and Control, 106, 107724. [5] Verma, A., Kumar, V., & Yadav, R. K. (2024, March). Humerus bone fracture detection utilizing YOLOv4 algorithm: a deep learning approach. In 2024 2nd International Conference on Disruptive Technologies (ICDT) (pp. 1191-1196). IEEE. [6] Sharma, P. (2025). Bone age estimation with HS-optimized Resnet and Yolo for child growth disorder. Expert Systems with Applications, 259, 125160. [7] Tao, X., Zhao, X., Liu, H., Wang, J., Tian, C., Liu, L., ...& Liu, Y. (2024). Automatic recognition of concealed fish bones under laryngoscopy: a practical AI model based on YOLO?V5. The Laryngoscope, 134(5), 2162-2169. [8] Ragab, M. G., Abdulkader, S. J., Muneer, A., Alqushaibi, A., Sumiea, E. H., Qureshi, R., ...&Alhussian, H. (2024). A comprehensive systematic review of YOLO for medical object detection (2018 to 2023). IEEE Access. [9] Ahmed, A., Imran, A. S., Manaf, A., Kastrati, Z., &Daudpota, S. M. (2024). Enhancing Wrist Fracture Detection with YOLO. arXiv preprint arXiv:2407.12597. [10] Likitha, M., &Shidaganti, G. I. (2024, August). Image Segmentation Techniques for Bone Cancer Identification in X-ray and MRI Imagery. In 2024 Second International Conference on Networks, Multimedia and Information Technology (NMITCON) (pp. 1-7). IEEE. [11] Pattabiraman, V., Sudhan, K. H. H., &Logeshwari, V. (2024, October). Bone Fracture Detection using Region-Based Convolutional Neural Network. In 2024 IEEE International Conference on Computer Vision and Machine Intelligence (CVMI) (pp. 1-6). IEEE. [12] Erdem, F., Gitto, S., Fusco, S., Bausano, M. V., Serpi, F., Albano, D., ...&Sconfienza, L. M. (2024). Automated detection of bone lesions using CT and MRI: a systematic review. La radiologiamedica, 1-8. [13] Alhussainan, N. F., Ben Youssef, B., & Ben Ismail, M. M. (2024). A deep learning approach for brain fracture firmness detection based on five different YOLO Versions: YOLOv3–YOLOv7. Computation, 12(3), 44. [14] Das, S., Bhattachya, D., &Biswas, T. Detection of bone fractures along with other abnormalities in wrist x-ray images using enhanced-yolo11. Available at SSRN 5056626. [15] Yao, S., Huang, Y., Wang, X., Zhang, Y., Paixao, I. C., Wang, Z., ...& Song, J. (2025). A Radiograph Dataset for the Classification, Localization, and Segmentation of Primary Bone Fractures. Scientific Data, 12(1), 88. [16] Parvin, S., &Rahman, A. (2024). A real-time human bone fracture detection and classification from multi-modal images using deep learning technique. Applied Intelligence, 54(19), 9269-9285. [17] Dharani, P., Kiranmayi, G., Chautharia, J., &Kalyani, C. H. (2024). Pediatric Fracture Detection with X-ray Images Using yolov8. image, 24(05). [18] Murrad, B. G., Mohsin, A. N., Al-Obaidi, R. H., Albaaji, G. F., Ali, A. A., Hamzah, M. S., ... & Al-Sharifi, H. K. (2024). An AI-Driven Framework for Detecting Bone Fractures in Orthopedic Therapy. ACS Biomaterials Science & Engineering, 11(1), 577-585. [19] Zhou, T., Wang, H., Du, Y., Liu, F., Guo, Y., & Lu, H. (2024). M3YOLOv5: Feature enhanced YOLOv5 model for mandibular fracture detection. Computers in Biology and Medicine, 173, 108291. [20] Alshahrani, A., &Alsairafi, A. (2024). Bone Fracture Classification using Convolutional Neural Networks from X-ray Images. Engineering, Technology & Applied Science Research, 14(5), 16640-16645.

Copyright

Copyright © 2025 Mr. Ch. Sunil, Thottempudi Harsha Vardhini, Shaik Afsha Shireen, Abdul Sattar, Gadila Surya Revanth. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.