Related Experiment Video
Updated: Dec 7, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.2K
Retinal blood vessel segmentation based on Densely Connected U-Net.
Yin Lin Cheng1,2, Meng Nan Ma1,2, Liang Jun Zhang1
1School of Biomedical Engineering, Sun Yat-sen University, Guangzhou 510006, China.
Mathematical Biosciences and Engineering : MBE
|September 29, 2020
Summary
This study introduces an improved U-Net model for enhanced retinal blood vessel segmentation, achieving higher accuracy, especially for smaller vessels. The method shows superior performance on public datasets, aiding in ophthalmic disease evaluation.
Area of Science:
- Medical Imaging Analysis
- Computer Vision in Healthcare
- Ophthalmology Diagnostics
Background:
- Accurate retinal blood vessel segmentation is crucial for diagnosing ophthalmic diseases.
- Existing methods face challenges in segmenting fine blood vessels.
- U-Net architecture is a common baseline for image segmentation tasks.
Purpose of the Study:
- To propose a novel U-Net architecture for improved retinal blood vessel segmentation.
- To enhance the accuracy of segmenting small and delicate blood vessels.
- To evaluate the proposed method's effectiveness on public datasets.
Main Methods:
- A modified U-Net architecture incorporating dense blocks was developed.
- Dense blocks allow feature reuse from all preceding layers, improving information flow.
- The model was trained and validated on the DRIVE and CHASE_DB1 datasets.
Main Results:
- The proposed method achieved high accuracy on both datasets (DRIVE: Acc=0.9559, AUC=0.9793; CHASE_DB1: Acc=0.9488, AUC=0.9785).
- Demonstrated superior performance compared to existing state-of-the-art methods.
- Showed particular effectiveness in segmenting small retinal blood vessels.
Conclusions:
- The enhanced U-Net architecture with dense blocks significantly improves retinal blood vessel segmentation accuracy.
- The method offers a promising tool for the detection and evaluation of ophthalmic diseases.
- The findings support the clinical utility of advanced deep learning models in ophthalmology.
