Fair-Eye Net

Gold definitionUpdated Apr 2, 2026

Definition

Fair-Eye Net is a multimodal AI system for glaucoma detection and follow-up, integrating diverse clinical data with a dual-stream architecture. It employs an uncertainty-aware gating strategy and a fairness constraint to improve diagnostic accuracy, ensure equitable outcomes, and provide early risk alerts.

At a glance

Executive summary

Fair-Eye Net is an AI system designed to improve glaucoma detection and follow-up by combining various eye exam data and patient demographics. It aims to provide accurate diagnoses, early risk warnings, and ensure fair outcomes across different patient groups, especially reducing disparities in diagnosis for disadvantaged populations.

TL;DR

Fair-Eye Net is an AI tool that uses multiple types of eye scan data to detect glaucoma early, track its progression, and ensure fair diagnoses for everyone.

Key points

Integrates multimodal clinical data (fundus photos, OCT, VF, demographics) via a dual-stream heterogeneous fusion architecture with uncertainty-aware hierarchical gating.
Addresses subjective, fragmented, and inequitable glaucoma screening and follow-up, preventing irreversible vision loss through early detection and consistent care.
Used by ophthalmologists, optometrists, and healthcare providers for glaucoma screening, diagnosis, longitudinal follow-up, and risk alerting.
Unlike current methods relying on single tests or loosely linked examinations, Fair-Eye Net offers an integrated, multimodal, and fair AI-driven approach.
Focuses on developing fair and reliable multimodal AI systems for medical diagnosis, particularly in addressing health disparities and ensuring equitable access to advanced care.

Use cases

Early Glaucoma Screening: Identifying individuals at risk of glaucoma from a diverse patient population using integrated clinical data, even in areas with limited specialist access.
Personalized Longitudinal Follow-up: Monitoring glaucoma progression over time by integrating new data points and providing 3-12 months early risk alerts for individual patients.
Reducing Diagnostic Disparities: Ensuring equitable diagnosis and preventing missed cases in racial or disadvantaged subgroups by actively mitigating false-negativity disparities.
Remote or Tele-ophthalmology: Facilitating consistent and high-quality glaucoma assessment in underserved regions by providing an AI-powered diagnostic aid.