Related Experiment Video
Updated: Sep 25, 2025
07:23
Retinal Vascular Reactivity as Assessed by Optical Coherence Tomography Angiography
Published on: March 26, 2020
7.6K
U-shaped Retinal Vessel Segmentation Based on Adaptive Aggregation of Feature Information
Liming Liang1, Jun Feng2, Longsong Zhou2
1School of Electrical Engineering and Automation, Jiangxi University of Science and Technology, Ganzhou, 341000, Jiangxi, China. 9119890012@jxust.edu.cn.
Interdisciplinary Sciences, Computational Life Sciences
|April 29, 2022
Summary
This study introduces a novel U-shaped network for enhanced retinal blood vessel segmentation, improving microvessel identification. The adaptive aggregation module boosts segmentation accuracy and model robustness in ophthalmic diagnostics.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computer Vision
Background:
- Accurate detection of retinal blood vessels is crucial for diagnosing ophthalmic diseases.
- Existing algorithms struggle with precise microvessel segmentation.
Purpose of the Study:
- To develop a more accurate retinal blood vessel segmentation algorithm.
- To enhance the identification of microvessels using a novel network architecture.
Main Methods:
- Proposed a U-shaped network incorporating an adaptive aggregation module and DenseASPP.
- Implemented a feature selection module to strengthen feature transmission.
- Utilized a joint loss function for effective network training.
Main Results:
- Achieved high performance on public datasets (DRIVE, STARE, CHASE_DB1).
- Reported sensitivity, accuracy, and AUC values exceeding 83%, 95%, and 98% respectively.
- Demonstrated improved robustness in segmenting retinal blood vessels.
Conclusions:
- The proposed U-shaped network with adaptive feature aggregation significantly improves retinal blood vessel segmentation.
- The method enhances the ability to identify challenging microvessels.
- This approach offers a promising tool for clinical diagnosis of eye conditions.
