An Algorithm for Automatic Rib Fracture Recognition Combined with nnU-Net and DenseNet

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Thoracic Trauma

Background:

• Rib fractures are common thoracic injuries, often presenting with multiple fracture sites.
• Accurate and rapid identification of all rib fracture regions is critical for effective patient treatment.
• Current diagnostic methods can be time-consuming and prone to errors.

Purpose of the Study:

• To develop and evaluate a novel two-stage deep learning model for automated rib fracture recognition.
• To enhance the accuracy and efficiency of identifying rib fracture regions in clinical settings.
• To reduce false-positive and false-negative rates in rib fracture detection.

Main Methods:

• A two-stage deep learning approach using nnU-Net was implemented.
• The first stage involved a segmentation model to identify candidate rib fracture regions.
• The second stage utilized a classification model to confirm fractures within segmented regions.

Main Results:

• The proposed two-stage deep learning model significantly improved the accuracy of rib fracture recognition.
• The model demonstrated a reduction in both false-positive and false-negative rates for fracture detection.
• The system showed potential for better assisting medical professionals in identifying fracture regions.

Conclusions:

• The two-stage deep learning model offers a promising solution for accurate and efficient rib fracture recognition.
• This AI-driven approach can enhance clinical decision-making and improve patient care for thoracic trauma.
• Further integration of such models can optimize the diagnostic workflow for rib fractures.