Related Experiment Video
Updated: Oct 15, 2025
07:15
Transient Optical Clearing Using Absorbing Molecules for Ex Vivo and In Vivo Imaging
Published on: July 11, 2025
696
Construction and application of color fundus image segmentation algorithm based on Multi-Scale local combined global
Yanjie Hao1, Hongbo Xie2, Rong Qiu3
1Yanjie Hao, Associate Chief Physician, Department of Ophthalmology, Jiaozhou Central Hospital of Qingdao, Qingdao, 266300, Shandong, China.
Pakistan Journal of Medical Sciences
|October 29, 2021
Summary
A novel retinal blood vessel segmentation method improves accuracy by combining multi-scale linear detection with local and global enhancement. This approach enhances the extraction of small blood vessels in retinal images.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computer Vision
Background:
- Accurate segmentation of retinal blood vessels is crucial for diagnosing various eye diseases.
- Existing methods often struggle with the accurate extraction of small, delicate blood vessels.
Purpose of the Study:
- To develop an improved retinal blood vessel segmentation method.
- To address the challenge of low accuracy in extracting small retinal blood vessels.
Main Methods:
- A multi-scale linear detector was developed, comprising small-scale and large-scale components.
- The small-scale detector processed locally enhanced images, while the large-scale detector processed globally enhanced images.
- Fusion of responses from different scales generated the final retinal vascular structure.
Main Results:
- The proposed algorithm achieved high average vascular accuracy rates of 96.62% on the STARE database and 96.45% on the DRIVE database.
- Average true positive rates were 75.52% (STARE) and 83.07% (DRIVE).
Conclusions:
- The developed method demonstrates high segmentation accuracy for retinal blood vessels.
- This technique offers superior blood vessel segmentation results, particularly for small vessels.
