BPENet: Boundary perception enhancement network for retinal vessel and coronary angiogram segmentation
Related Experiment Video
Updated: Jul 8, 2026
Retinal Vascular Reactivity as Assessed by Optical Coherence Tomography Angiography
Published on: March 26, 2020
Linfeng Kong1,2, Yun Wu1,2
1State Key Laboratory of Public Big Data, Guizhou University, Guiyang, China.
Insights
This study introduces BPENet, a novel network for improved vessel segmentation. BPENet enhances boundary detection, leading to more accurate results for retinal and coronary artery diseases.
Area of Science:
- Medical Imaging
- Computer Vision
- Biomedical Engineering
Background:
- Accurate vessel segmentation is vital for diagnosing diseases like retinal and coronary artery conditions.
- Existing segmentation methods struggle with precise vessel boundary delineation, posing a challenge for clinical applications.
Purpose of the Study:
- To improve vessel segmentation accuracy by addressing the challenge of poor boundary segmentation.
- To introduce a novel network, the Boundary Perception Enhancement Network (BPENet), for enhanced vessel segmentation.
Main Methods:
- BPENet integrates a boundary detection auxiliary task with the main segmentation task.
- Key modules include Feature Extraction Enhancement (FE2) for feature extraction, Boundary Enhancement (BE) for boundary perception, and Deep Feature Aggregation (DFA) for multi-layer feature utilization.
Main Results:
- BPENet demonstrated superior performance compared to existing mainstream methods on retinal vessel datasets (DRIVE, CHASE-DB1, STARE) and coronary angiogram datasets (DCA1, XCAD).
- The proposed modules effectively enhanced feature extraction, boundary perception, and feature aggregation.
Conclusions:
- BPENet offers a reliable solution for accurate vessel segmentation, particularly in challenging cases with unclear boundaries.
- The network is expected to aid clinicians in the diagnosis and treatment of various vascular diseases.
Abstract:
Vessel segmentation is crucial for assisting in the diagnosis and treatment of a range of diseases, such as retinal diseases and coronary artery diseases. However, current vessel segmentation methods often face the problem of poor vessel boundary segmentation, making the vessel segmentation task still challenging. This study aims to alleviate this problem, thereby improving the vessel segmentation results. By introducing a boundary detection auxiliary task to the vessel segmentation main task, we propose a boundary perception enhancement network (BPENet) for retinal vessel and coronary angiogram segmentation. Among them, to enhance vessel feature extraction capability, BPENet introduces the feature extraction enhancement (FE2) module. To enhance vessel boundary perception capability, BPENet introduces the boundary enhancement (BE) module. In addition, to fully leverage feature information from different layers, BPENet introduces the deep feature aggregation (DFA) module. Experimental results on retinal vessel datasets (DRIVE, CHASE-DB1, and STARE) and coronary angiogram datasets (DCA1 and XCAD) show that BPENet outperforms the existing mainstream segmentation methods. BPENet is expected to provide a reliable vessel segmentation result for doctors during the diagnosis and treatment of related diseases.
