Hybrid SFNet Model for Bone Fracture Detection and Classification Using ML/DL.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Orthopedics

Background:

• Manual bone fracture diagnosis from X-rays is time-consuming and requires expert interpretation.
• Machine learning (ML) and deep learning (DL) offer advanced solutions for medical image analysis.
• Accurate and efficient fracture detection is crucial for timely patient treatment.

Purpose of the Study:

• To develop a novel deep learning model for automated bone fracture diagnosis.
• To improve the accuracy and efficiency of fracture detection in X-ray images.
• To investigate the impact of an improved Canny edge algorithm on fracture localization and diagnosis.

Main Methods:

• Proposed a hybrid two-scale sequential deep learning model named SFNet (Scale Fracture Network).
• Integrated an improved Canny edge algorithm to precisely localize fracture regions.
• Fed grayscale and Canny edge images into the SFNet for training and evaluation.
• Compared SFNet performance against state-of-the-art deep convolutional neural network (CNN) models.

Main Results:

• SFNet combined with the Canny edge algorithm (SFNet + Canny) achieved superior diagnostic performance.
• Achieved an accuracy of 99.12%, F1-score of 99%, and recall of 100% for bone fracture diagnosis.
• Demonstrated reduced computation time compared to other deep CNN models.
• The Canny edge algorithm significantly enhanced the performance of the CNN model.

Conclusions:

• The proposed SFNet model with the Canny edge algorithm offers a highly efficient and accurate solution for automated bone fracture diagnosis.
• Integrating edge detection techniques improves the localization and identification of fractures in medical imaging.
• This AI-driven approach has the potential to streamline the diagnostic workflow and improve patient outcomes.