AEGIS Ocular Station — Eye-Disease Screening (P15) · Case Study

What this is

The AEGIS Ocular Station reads a retinal fundus photo once and runs it past four dedicated-data specialists at the same time, then routes the finding to the right expert. It is honest by construction: each specialist stands on its own validated metric, and any case that lands in the grey zone is flagged for a clinician, never auto-called. AEGIS scores; a clinician confirms.

Read this first — what is built, and what it is not. This is a feasibility research programme: four single-disease fundus screeners, each trained and validated on dedicated public data, unified by one engine and fronted by a public learning game. Every number below is an honestly-reported validation ROC-AUC on held-out images — not a clinical-trial result. AEGIS Ocular does not diagnose. A human is always in command.

The specialists — honest metrics

Each dedicated-data specialist raised its ODIR baseline. The scores are reported plainly, strong and weaker alike.

Specialist	Disease	Champion	Validation ROC-AUC
H	Hypertensive retinopathy	HistGB	0.993
D	Diabetic retinopathy (referable)	CatBoost	0.977
A	Age-related macular degeneration	CatBoost	0.953
G	Glaucoma	CatBoost	0.899
Myopia	Pathological myopia (ODIR base plate)	—	0.979
Cataract	Cataract (ODIR base plate)	—	0.904

Honest note: the four dedicated specialists (D/A/G/H) were each trained on disease-specific public datasets and each broke its weaker ODIR baseline; scores are leak-certified (exact + perceptual-hash dedup). Any uncertain case in the 0.40–0.60 grey zone is deferred to a clinician rather than force a call.

▶ Play the live “Spot the Disease” game →

The live station lets you try to beat the AI on real example fundus photos — some eyes look perfectly normal but are actually diseased. A learning game and screening aid, never a diagnosis.

We’re not eye specialists — help us make AEGIS better through more in-depth training

Let’s be honest: AEGIS is a small research programme, not an eye clinic, and we don’t claim to be ophthalmologists. The system only gets better when it sees more real examples. You can already try it right now — the live station runs on real de-identified fundus photos from public ophthalmology datasets, so you can test AEGIS on real eyes without sending anything. And if you’d be willing to share your own de-identified retinal fundus image, you’d genuinely be helping us test and improve it — and we’ll show you what the model sees, purely as an illustration, never a diagnosis. A few honest ground rules first; sending an image means you accept them. Thank you for your support!

A few honest ground rules before you send anything.
1. Not a medical device and not a diagnosis. AEGIS is a research and educational demonstration. Nothing it outputs is medical advice, a diagnosis, a screening result, or a substitute for a qualified clinician. No doctor–patient relationship is created.
2. Do not send identifiable patient data. Remove all names, IDs, dates of birth, and any protected health information (PHI) before sending. Only submit an image you have the right to share.
3. You release AEGIS from liability. By submitting, you agree to indemnify and hold harmless AEGIS, Dr Loh Kah Meng, and anyone associated with the programme from any and all claims, losses, or liability arising from your use of, or reliance on, anything AEGIS provides. You use it entirely at your own risk.
4. Always consult a licensed clinician. For any real eye or health concern, see a qualified doctor or ophthalmologist. Do not delay or disregard professional medical advice because of anything here.

✉ Contact AEGIS to submit an image

One programme, one honest promise. AEGIS augments, it never replaces. Every specialist reports its own real number; grey-zone cases defer to a person; nothing is auto-called. Dedicated to Professor Dhanjoo N. Ghista, whose nondimensional-index lineage this work continues.