AI-powered retinal analysis

Diabetic retinopathy, macular edema and more, from one fundus photograph

Three analyses in one platform: an ICDR diabetic retinopathy grade with vessel and lesion maps, a diabetic macular edema grade with a hard-exudate map, and a beta screen for other retinal diseases. Upload one or many fundus photographs and download a printable report for each analysis. Built on the lab's fractional enhancement and vascular-geometry research.

3Analyses: DR · DME · multi-disease
ICDR 0–4DR grading scale
SMS HospitalLive clinical pilot
1Patent published
1 · ChooseDR · DME · Multi-disease (beta)
2 · UploadOne or more fundus photographs
3 · ResultGrade, maps and referral advice
4 · ReportPrintable PDF with patient details
TimeA few seconds per image

Three analyses, one platform

A screening aid for ophthalmologists and diabetic eye-screening programmes. Pick the analysis you need; each runs its own model on the images you give it. Results are always reviewed by a qualified ophthalmologist.

Diabetic retinopathy

DR grading and segmentation

A five-stage diabetic retinopathy grade on the ICDR 0–4 scale (No DR to PDR) with a confidence score and referral advice.

  • Retinal vessel map and binary vessel segmentation
  • Lesion maps for microaneurysms, haemorrhages, hard and soft exudates (FLA-Net)
  • Both eyes or several images in one go
Diabetic macular edema

DME grading with GRACE-Net

Grades each image as no DME, or hard exudates away from or near the fovea, with the probability of each grade.

  • Hard-exudate overlay on the fundus image
  • Presence scores for MA, HE, EX and SE
  • Protocol-aware fusion of a direct and a grading-rule prediction, with an agreement check
Multi-disease · beta

Look-alike disease screen

A research preview that looks past diabetic retinopathy for conditions that can be mistaken for it.

  • AMD, myopic changes, retinal vein occlusion and chorioretinal scarring
  • Referable-DR and macular-edema probabilities
  • Overlays of DR lesions and other-disease signs

Built for screening clinics

The same workflow for all three analyses, from upload to report.

1

Many images at once

Upload both eyes or a whole set of fundus photographs. Results open with a summary table and a full panel for every image.

2

A report for every analysis

Add patient details and download a printable PDF report for DR, DME or the multi-disease screen, with the images and findings.

3

Maps you can check

Every result shows the overlay it rests on, such as vessels, lesions or exudates, next to the original photograph.

4

No account needed

Run any analysis as a guest. Sign in to keep your DR reports in My Reports.

5

Works on any device

A web app that runs in the browser on desktops, tablets and phones, with nothing to install.

6

Research-backed

Built on the lab's fractional-order image enhancement and retinal vessel research, published in peer-reviewed journals.

Coming soon GlaucomaCataractOCT segmentation & analysisFFA analysisHypertensive retinopathyRetinopathy of prematurityRetinal detachmentKeratoconus
Clinical validation in progress

Real-world testing at SMS Hospital, Jaipur

The Directorate of Medical & Health Services, Government of Rajasthan (State Health Society, NCO) approved the pilot implementation and clinical validation of ChakshuRakshak AI for screening diabetic retinopathy and diabetic macular edema, in collaboration with the Department of Medical, Health & Family Welfare.

  • 1
    SMS Hospital, Jaipur · first two monthsFundus-camera screening with ChakshuRakshak AI and clinical validation against ophthalmologists. Running now.
  • 2
    District Hospital, Tonk · next phaseThe pilot continues at the district level after the SMS phase.

Weekly progress reports go to the State Directorate. Final diagnosis and treatment stay with qualified ophthalmologists; the AI assists screening.

The MAITRI Lab team screening patients with a fundus camera and ChakshuRakshak AI at SMS Hospital, Jaipur, and a team photo in the ophthalmology department
Screening and validation with the ophthalmology team at SMS Hospital, Jaipur.

Where we have presented it

ChakshuRakshak AI was showcased at four state and national events in 2026 and recognised for its impact as a scalable AI healthcare solution.

Rajasthan DigiFest 2026Live demo and exhibition
Regional AI Impact Conference 2026Exhibition stall
MNIT Industry-Institute Conclave 2026Poster and demo
State-Level National Technology Day 2026Exhibition

Research behind it

Papers and the patent that ChakshuRakshak AI is built on.

Patent

Published 28-02-2026

AI-Driven Framework for Diabetic Retinopathy Grading and Related Methods

Priyanka Harjule, Mukesh Delu, Rajesh Kumar, Kushal Gajjar · Reg. No. 202611024093 · Controller General of Patents, Designs & Trade Marks, Govt. of India

Papers

Try it or work with us

Hospitals, screening programmes and researchers interested in ChakshuRakshak AI can try the demo or write to us at info@chakshurakshakai.com or priyanka.maths@mnit.ac.in.