Related Experiment Video
Updated: Aug 8, 2025
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
Automatic vessel crossing and bifurcation detection based on multi-attention network vessel segmentation and directed
Gengyuan Wang1, Yuancong Huang2, Ke Ma2
1State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, 510060, China; School of Life Sciences, South China University of Technology, Guangzhou, 510006, Guangdong, China.
Insights
This study introduces a new neural network to automatically identify and classify blood vessel intersections and bifurcations in retinal images. This advancement aids in diagnosing eye and systemic diseases by improving vascular network analysis.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computer Vision
Background:
- Accurate analysis of the retinal vascular tree is crucial for diagnosing ophthalmic and systemic diseases.
- Identifying intersections and bifurcations in retinal vessels is challenging but vital for understanding vascular morphology and network complexity.
Purpose of the Study:
- To develop an automated method for segmenting the vascular network in color fundus images.
- To accurately detect and classify vascular intersections and bifurcations using a novel neural network approach.
Main Methods:
- A directed graph search-based multi-attentive neural network was employed for vascular segmentation.
- Multi-dimensional attention mechanisms integrated local and global features for scale-adaptive target structure focus.
- Vascular topology and connectivity were represented using a directed graphical model.
- Local geometric information was utilized to decompose the vascular tree and classify feature points.
Main Results:
- The method achieved an F1-score of 0.863 on the DRIVE dataset and 0.764 on the IOSTAR dataset for detection points.
- Average accuracy for classification points reached 0.914 on DRIVE and 0.854 on IOSTAR.
- The proposed approach demonstrated superior performance compared to existing state-of-the-art methods in feature point detection and classification.
Conclusions:
- The novel directed graph search-based multi-attentive neural network effectively segments retinal vasculature and identifies key feature points.
- This method offers a significant improvement for automated retinal biomarker analysis, aiding in disease diagnosis.
- The approach shows strong potential for clinical applications in ophthalmology and systemic disease monitoring.
Abstract:
Analysis of the vascular tree is the basic premise to automatically diagnose retinal biomarkers associated with ophthalmic and systemic diseases, among which accurate identification of intersection and bifurcation points is quite challenging but important for disentangling complex vascular network and tracking vessel morphology. In this paper, we present a novel directed graph search-based multi-attentive neural network approach to automatically segment the vascular network and separate intersections and bifurcations from color fundus images. Our approach uses multi-dimensional attention to adaptively integrate local features and their global dependencies while learning to focus on target structures at different scales to generate binary vascular maps. A directed graphical representation of the vascular network is constructed to represent the topology and spatial connectivity of the vascular structures. Using local geometric information including color difference, diameter, and angle, the complex vascular tree is decomposed into multiple sub-trees to finally classify and label vascular feature points. The proposed method has been tested on the DRIVE dataset and the IOSTAR dataset containing 40 images and 30 images, respectively, with 0.863 and 0.764 F1-score of detection points and average accuracy of 0.914 and 0.854 for classification points. These results demonstrate the superiority of our proposed method outperforming state-of-the-art methods in feature point detection and classification.
