Retinal vessel ratio estimation aids early retinopathy screening accuracy
A Stage-Wise Explainable Pipeline for Arteriolar-to-Venular Ratio Estimation and Retinopathy Screening
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Changes in the width of tiny blood vessels in the eye can signal health problems like retinopathy. The authors developed a step-by-step method that analyzes eye images to measure the ratio between artery and vein widths, helping to detect signs of this eye disease. Their approach breaks the process into clear parts, making it easier to see where any mistakes happen. They tested it on a small set of images and found it was very accurate, though more testing is needed on larger groups.
What this means in practice
- •For ophthalmology imaging teams: Process retinal images to estimate vessel diameter ratios with interpretable outputs aiding early detection of retinopathy.
- •For medical device developers: Integrate transparent vessel-calibre measurement and disease screening steps into retinal imaging devices to improve diagnostic workflows.
Tested on one dataset.
Authors
Sharan Sharmeela Karuppannan, Yoga Sri Varshan, Subin Sahayam, Umarani Jayaraman
Abstract
The arteriolar-to-venular diameter ratio (AVR) is a clinically motivated marker of retinal vascular health. We present a five-stage pipeline for AVR estimation and exploratory retinopathy screening from fundus images, with inspectable intermediate outputs at each stage. The pipeline combines vessel and artery-vein masks, optic-disc localization, vessel-calibre estimation using a minimal-chord boundary walk, and AVR computation from dominant arterial and venular branches. An explicit threshold-based rule produces the screening decision, enabling errors to be traced to individual processing stages. On an exploratory cohort of 30 HRF images, we report 96.7% accuracy, an F1-score of 0.968 and ROC-AUC of 0.978. Ablation studies and an exploratory comparison with VascX examine the effects of vessel pairing, calibre measurement and aggregation. Interpretation is limited by the small cohort, proxy disease labels and absence of independent reference calibre measurements. The pipeline provides a transparent framework for investigating retinal vascular biomarkers, with further validation needed on larger, clinically graded cohorts.