Intelligent Bone Fracture Recognition Framework: Convolutional Neural Network(CNN): Note: Sub-titles are not captured in Xplore and should not be used

In this study, we introduce a method to diagnose bone disease using advances in machine learning (ML) and deep learning (DL) technologies. Our scheme combines several different technologies with convolutional neural networks (CNN) to distinguish healthy bones. This new hybrid model represents a breakthrough in bone diagnosis and provides a significant performance in computation time compared to existing deep CNN models. The main innovation is that we use an improved sharp-edge algorithm to accurately identify cracks in X-ray images. Then the grayscale image and the corresponding image are input into the algorithm for training and evaluation. We perform a performance test evaluating our method against a state-of-the-art deep CNN model using skeletal muscle data. Our results showed the best performance in fracture diagnosis, achieving irregularity, F1 score, and recall of 96.12%, 99%, and 98%, respectively. This demonstrates the effectiveness of combining algorithms to improve CNN performance, demonstrating the potential of our method to revolutionize bone diagnosis in clinical practice.