Related Experiment Video
Updated: Apr 15, 2026
12:28
Quantification of Vascular Parameters in Whole Mount Retinas of Mice with Non-Proliferative and Proliferative Retinopathies
Published on: March 12, 2022
4.5K
Retinal vessel segmentation: an efficient graph cut approach with retinex and local phase
Yitian Zhao1, Yonghuai Liu2, Xiangqian Wu3
1Department of Eye and Vision Science, University of Liverpool, Liverpool, United Kingdom. School of Mechatronical Engineering, Beijing Institute of Technology, Beijing, China.
Plos One
|April 2, 2015
Summary
This study introduces an automated framework for retinal vessel segmentation, significantly improving accuracy and efficiency in detecting diseases. The novel approach enhances diagnostic capabilities for various retinal and systemic conditions.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computer Vision
Background:
- Automated detection of retinal vessels is crucial for understanding disease mechanisms and diagnosis.
- Existing methods face challenges with image inhomogeneity and low contrast of thin vessels.
Purpose of the Study:
- To propose a novel framework for accurate and efficient retinal vasculature segmentation.
- To improve the understanding and diagnosis of retinal and systemic diseases through enhanced imaging analysis.
Main Methods:
- A three-component framework: Retinex-based inhomogeneity correction, local phase-based vessel enhancement, and graph cut-based active contour segmentation.
- The method was applied to four public retinal image datasets (color fundus photography and fluorescein angiography).
Main Results:
- Each component demonstrated expected performance levels.
- The overall framework outperformed widely used unsupervised and supervised methods.
- Achieved high sensitivity (0.744), specificity (0.978), and accuracy (0.953) on the DRIVE dataset, comparable to manual annotations.
Conclusions:
- The proposed framework offers a significant advancement in automated retinal vessel segmentation.
- The enhanced accuracy and efficiency support improved diagnosis and treatment of retinal and systemic diseases.
- The method's performance is robust across different retinal imaging datasets.
