Objective:

To develop a method for accurate nerve segmentation in ultrasound images by integrating diffusion processing and edge detection.

Approach:

• Methodology: A diffusion-guided edge-aware multi-channel segmentation approach was proposed, utilizing raw ultrasound images, diffusion-processed features, and edge information. The EdgeFusion-U-Net model was developed to integrate these features.

Key Findings:

• The proposed method achieved Dice scores of 94.62% and 93.91% on two benchmark datasets.
• Baseline methods without diffusion denoising and edge fusion had Dice scores around 83% and 84%.
• Incorporating diffusion filter-based speckle noise removal and edge-based feature fusion improved boundary detection.

Interpretation:

The integration of diffusion processing and edge detection enhances the accuracy of nerve segmentation in ultrasound imaging.

Limitations:

• The study does not address the scalability of the proposed method in clinical settings.
• Further validation on diverse datasets may be necessary to confirm generalizability.

Conclusion:

The proposed technique supports accurate nerve identification in ultrasound images.