Automated Diabetic Screening via Anterior Segment Ocular Imaging: A Deep Learning and Explainable AI Approach

2026-03-15

AI Summary

This study addresses the limited accessibility of traditional diabetic retinopathy screening—reliant on fundus photography and constrained by equipment availability and specialist expertise—in primary care settings. It presents the first systematic validation of using visible biomarkers in anterior segment ocular images (e.g., iris, sclera, conjunctiva) for automated diabetes classification. The authors propose a preprocessing pipeline integrating self-supervised learning (SimCLR), specular reflection removal, and CLAHE enhancement, and evaluate five state-of-the-art architectures. Among them, EfficientNet-V2-S combined with self-supervised pretraining achieves an F1-score of 98.21%, with perfect (100%) precision in identifying non-diabetic cases, significantly outperforming ImageNet-pretrained baselines. This work establishes a novel, non-invasive, and resource-efficient paradigm for diabetes screening in low-resource environments.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multi-instance/Multi-view Learning
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Security and Privacy: Data transparency and provenance

Abstract

Diabetic retinopathy screening traditionally relies on fundus photography, requiring specialized equipment and expertise often unavailable in primary care and resource limited settings. We developed and validated a deep learning (DL) system for automated diabetic classification using anterior segment ocular imaging a readily accessible alternative utilizing standard photography equipment. The system leverages visible biomarkers in the iris, sclera, and conjunctiva that correlate with systemic diabetic status. We systematically evaluated five contemporary architectures (EfficientNet-V2-S with self-supervised learning (SSL), Vision Transformer, Swin Transformer, ConvNeXt-Base, and ResNet-50) on 2,640 clinically annotated anterior segment images spanning Normal, Controlled Diabetic, and Uncontrolled Diabetic categories. A tailored preprocessing pipeline combining specular reflection mitigation and contrast limited adaptive histogram equalization (CLAHE) was implemented to enhance subtle vascular and textural patterns critical for classification. SSL using SimCLR on domain specific ocular images substantially improved model performance.EfficientNet-V2-S with SSL achieved optimal performance with an F1-score of 98.21%, precision of 97.90%, and recall of 98.55% a substantial improvement over ImageNet only initialization (94.63% F1). Notably, the model attained near perfect precision (100%) for Normal classification, critical for minimizing unnecessary clinical referrals.

Problem

• Diabetic screening
• Anterior segment imaging
• Deep learning
• Explainable AI
• Resource-limited settings

Innovation

• Anterior Segment Imaging
• Deep Learning
• Self-Supervised Learning
• Explainable AI
• Diabetic Screening