Bone Fracture Detection using Region-Based Convolutional Neural Network

Bone fracture detection is an essential process in medical field to ensure accurate diagnosis and treatment. This study focusses on novel approach to fracture diagnosis in X Ray images that uses complex machine learning algorithms. As early diagnosis results in an effective treatment, finding bone fractures in medical imaging became very crucial. This study utilizes a substantial dataset of annotated X-ray images featuring various types of fractures. Existing methods for fracture detection are reviewed and compared to the proposed approach. We provide a neural network-based method for identifying bone fractures in X-ray images using Region-Based Convolutional Neural Network (R-CNN) and You Only Look Once (YOLOV8) which are well-known object detection methods that we employ to find fractures. For testing and training we use a large collection of annotated X-Ray images of patients with different kinds of fractures. We evaluate the performance of our method on this dataset and compare it with state-of-the-art methods. Our results demonstrate that the approach we proposed yields superior precision and recall than alternative approaches, with high accuracy of $\mathbf{9 6. 5 \%}$ and $\mathbf{9 0 \%}$ respectively. The recommended method can be reliable tool for radiologists, who can use it to detect bone fractures more rapidly and accurately, leading to better patient outcomes.