Implementation of a Classification System of Calcaneal Fracture with Deep Learning Techniques

Calcaneus fracture is the most common fracture in all types of Tarsal fracture. Early and accurate diagnosis is essential for prompt treatment. This research aims to develop a deep learning system that can detect calcaneal fractures in lateral X-ray images promptly. We’ve collected a dataset of calcaneal X-ray images, categorized as either fracture or non-fracture, and employed data augmentation techniques to expand this dataset. We’ve trained a Deep Residual Neural Network (ResNet) model for binary fracture classification. To enhance model interpretability and help non-deep learning experts understand how the model predicts. We’ve utilized Grad-CAM to generate heatmaps. It highlights crucial image regions, making the predictions more intuitive.