Fig. 3
From: Intelligent detection and grading diagnosis of fresh rib fractures based on deep learning
Rib fracture detection in CT images using a YOLO-based model with Exponential Moving Average (EMA). The input CT images are resized to 640px × 640px and processed through the model to generate feature maps. These maps are used to produce a class probability map and a bounding map, which are then passed through Non-Maximum Suppression (NMS) to yield the final detection of the fracture location