Related Experiment Video
Updated: Jun 25, 2025
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
389
TD Swin-UNet: Texture-Driven Swin-UNet with Enhanced Boundary-Wise Perception for Retinal Vessel Segmentation
Angran Li1, Mingzhu Sun1, Zengshuo Wang1
1College of Artificial Intelligence, Nankai University, Tianjin 300350, China.
Bioengineering (Basel, Switzerland)
|May 25, 2024
Summary
This study introduces a novel texture-driven Swin-UNet for improved retinal vessel segmentation. The enhanced model accurately identifies blood vessel boundaries, aiding in ophthalmological diagnoses.
Area of Science:
- Medical Image Analysis
- Ophthalmology
- Computer Vision
Background:
- Accurate retinal vessel segmentation is vital for diagnosing and monitoring eye diseases.
- Existing segmentation methods struggle with complex vessel boundaries and texture features.
Purpose of the Study:
- To develop an advanced deep learning model for precise retinal blood vessel segmentation.
- To enhance the accuracy of vessel boundary detection in retinal images.
Main Methods:
- Proposed a texture-driven Swin-UNet incorporating a Cross-level Texture Complementary Module (CTCM) for feature fusion.
- Introduced a Pixel-wise Texture Swin Block (PT Swin Block) for improved boundary localization.
- Utilized an improved Hausdorff distance loss function to refine boundary segmentation accuracy.
Main Results:
- The proposed model demonstrated superior performance on DRIVE and CHASEDB1 datasets.
- Achieved significant improvements in Accuracy (ACC), Sensitivity (SE), Specificity (SP), and F1 score (F1).
- Markedly enhanced the precision of retinal blood vessel boundary segmentation.
Conclusions:
- The texture-driven Swin-UNet effectively addresses limitations in current retinal vessel segmentation techniques.
- The model offers a promising tool for clinical applications in ophthalmology.
- This approach significantly advances the accuracy of medical image analysis for eye conditions.
