Related Experiment Video
Updated: Aug 30, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.9K
Particle Swarm Optimization and Salp Swarm Algorithm for the Segmentation of Diabetic Retinal Blood Vessel Images
Liwei Deng1, Shanshan Liu1, Xiaofei Wang2
1Heilongjiang Provincial Key Laboratory of Complex Intelligent System and Integration, School of Automation, Harbin University of Science and Technology, Harbin 150080, China.
Computational Intelligence and Neuroscience
|September 2, 2022
Summary
A novel method improves diabetic retinopathy diagnosis by accurately segmenting retinal blood vessels using enhanced particle swarm and salp swarm algorithms. This approach offers more efficient and precise medical image analysis for early detection.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computational Intelligence
Background:
- Diabetes mellitus is a growing global health concern with increasing incidence.
- Diabetic retinopathy, a complication of diabetes, affects retinal blood vessels, making early detection crucial.
- Segmentation of complex retinal vessel images, especially terminal vessels, remains a challenge in medical image processing.
Purpose of the Study:
- To propose a new segmentation method for diabetic retinal vessel images.
- To enhance the accuracy and efficiency of diabetic retinopathy detection through improved image segmentation.
- To address the difficulties in identifying fine blood vessels in retinal images.
Main Methods:
- Utilized Gaussian filtering to enhance main retinal blood vessels.
- Employed top-hat transform to strengthen terminal retinal vessels.
- Integrated and normalized enhanced images.
- Applied improved particle swarm optimization and salp swarm algorithms for multi-threshold segmentation.
Main Results:
- The proposed improved particle swarm-salp swarm algorithm demonstrated superior efficiency in segmenting diabetic retinal blood vessel images.
- The method achieved better threshold selection compared to existing techniques.
- Comprehensive evaluation using metrics like SSIM, PSNR, accuracy, Dice, and Jaccard confirmed the method's effectiveness.
Conclusions:
- The developed vascular segmentation method significantly improves the accuracy of medical image processing for diabetic retinopathy.
- The combined particle swarm and salp swarm algorithm offers a more efficient and effective approach to retinal vessel segmentation.
- This technique has the potential to reduce computational load and enhance diagnostic capabilities in clinical settings.
