A Novel Single-Sample Retinal Vessel Segmentation Method Based on Grey Relational Analysis

Area of Science:

• Ophthalmology
• Medical Imaging
• Computer Vision

Background:

• Accurate retinal vessel segmentation is crucial for diagnosing and treating various eye diseases.
• Limited annotated datasets pose a significant challenge for developing robust segmentation models.
• Grey theory offers a promising approach for handling problems with scarce data and poor information.

Purpose of the Study:

• To propose a novel grey relational-based method for accurate retinal vessel segmentation.
• To address the limitations of small and unlabeled datasets in medical image analysis.
• To enhance computer-aided diagnosis systems for retinal diseases.

Main Methods:

• Developed a noise-adaptive discrimination filtering algorithm based on grey relational analysis (NADF-GRA) for image enhancement.
• Designed a threshold segmentation model based on grey relational analysis (TS-GRA) for segmenting enhanced vessel images.
• Implemented a post-processing stage including hole filling and isolated pixel removal for refined segmentation.

Main Results:

• The proposed method demonstrated high performance on public datasets (DRIVE, STARE, HRF).
• Achieved an average accuracy of 96.03% and specificity of 98.51% on the DRIVE dataset.
• Showcased superior performance compared to current mainstream retinal vessel segmentation methods.

Conclusions:

• The novel grey relational-based method offers an effective solution for retinal vessel segmentation, especially with limited data.
• The method provides accurate and robust segmentation results, outperforming existing techniques.
• This approach has the potential to significantly advance computer-aided diagnosis in ophthalmology.