Related Experiment Video
Updated: May 14, 2026
Quantification of Vascular Parameters in Whole Mount Retinas of Mice with Non-Proliferative and Proliferative Retinopathies
Published on: March 12, 2022
Simultaneously identifying all true vessels from segmented retinal images
Qiangfeng Peter Lau1, Mong Li Lee, Wynne Hsu
1Department of Computer Science, National University of Singapore, 117417 Singapore. plau@comp.nus.edu.sg
Insights
Accurate identification of retinal blood vessels is crucial for diagnosing cardiovascular disease risk. This study presents a novel graph-based method to precisely identify true vessels in retinal images, improving diagnostic accuracy.
Area of Science:
- Ophthalmology
- Medical Imaging
- Cardiovascular Disease Research
Background:
- Retinal blood vessel morphology measurements are linked to cardiovascular disease risk.
- Incorrect vessel identification can lead to significant measurement variations and misdiagnosis.
Purpose of the Study:
- To develop an automated postprocessing method for accurate true vessel identification in retinal images.
- To improve the reliability of retinal vascular measurements for clinical diagnosis.
Main Methods:
- Modeling segmented vascular structures as a vessel segment graph.
- Formulating vessel identification as an optimal forest finding problem with constraints.
- Developing and evaluating a method to solve this optimization problem on a large dataset.
Main Results:
- Achieved 98.9% pixel precision and 98.7% recall for true vessels in clean images.
- Demonstrated robustness of the method even with noisy segmented retinal images.
- Validated on a real-world dataset of 2,446 retinal images.
Conclusions:
- The proposed graph-based method accurately identifies true retinal vessels.
- This approach enhances the reliability of morphological measurements for cardiovascular risk assessment.
- The method shows significant potential for improving clinical diagnosis based on retinal imaging.
Abstract:
Measurements of retinal blood vessel morphology have been shown to be related to the risk of cardiovascular diseases. The wrong identification of vessels may result in a large variation of these measurements, leading to a wrong clinical diagnosis. In this paper, we address the problem of automatically identifying true vessels as a postprocessing step to vascular structure segmentation. We model the segmented vascular structure as a vessel segment graph and formulate the problem of identifying vessels as one of finding the optimal forest in the graph given a set of constraints. We design a method to solve this optimization problem and evaluate it on a large real-world dataset of 2,446 retinal images. Experiment results are analyzed with respect to actual measurements of vessel morphology. The results show that the proposed approach is able to achieve 98.9% pixel precision and 98.7% recall of the true vessels for clean segmented retinal images, and remains robust even when the segmented image is noisy.
