AIROGS: AI for Robust Glaucoma Screening

~113,893 color fundus images labelled as referable glaucoma (RG), no referable glaucoma (NRG), or ungradable. Large-scale, multi-ethnic, multi-site screening dataset.

At a glance

Field	Value
Short name	airogs
Full name	AIROGS: AI for Robust Glaucoma Screening
Primary category	fundus
Contained modalities	fundus
Tasks	classification
Samples	113,893
Classes	2 (NRG, RG)
Splits	train, test
Size	40.0 GB
Source-stated terms	CC BY-NC-ND 4.0
Normalized terms	cc-by-nc-nd
Descriptive screening label	Explicit noncommercial clause recorded; check source
Terms scope	dataset_files
Access friction	anonymous_direct
Route backend	Direct HTTP
Availability	available (checked 2026-07-21)
Acquisition support	guided_instructions_only
Legacy sample-loader status	Standard loader included

Notes

Training set (~101k images) on Zenodo. Test set distributed through Grand Challenge. Requires free account for Grand Challenge.

Access preflight and acquisition

CLI

# Read-only preflight
eyehub download airogs --data-dir ./data --dry-run --json

# Explicit transfer, only when preflight reports supported behavior
eyehub download airogs --data-dir ./data

Python

from eyedatahub.acquisition import preflight_dataset
from eyedatahub.datasets.registry import REGISTRY

ds = REGISTRY.get_dataset('airogs')
print(preflight_dataset(ds, './data'))  # no transfer

Upstream page: zenodo.org/records

Source-term evidence: zenodo.org/records

Loader example

This entry includes a standard DatasetSample loader.

from pathlib import Path
from eyedatahub.datasets.registry import REGISTRY

data_dir = Path('~/.eyedatahub/data').expanduser()
ds = REGISTRY.get_dataset('airogs')
samples = ds.load(data_dir, split='test')
for s in samples[:5]:
    print(s.sample_id, s.label, s.image_path)

Citation

BibTeX

@misc{airogs,
  title  = { AIROGS: AI for Robust Glaucoma Screening },
  note   = { De Vente et al., 'AIROGS: Artificial Intelligence for Robust Glaucoma Screening Challenge', TMI 2023 },
  year   = { 2023 },
  url    = { https://zenodo.org/records/5793241 },
}

Plain text

De Vente et al., 'AIROGS: Artificial Intelligence for Robust Glaucoma Screening Challenge', TMI 2023.

Source-stated terms

• Raw source string: CC BY-NC-ND 4.0
• Normalized category: cc-by-nc-nd
• Apparent scope: dataset_files
• Descriptive screening label: Explicit noncommercial clause recorded; check source

⚠️ Source-stated terms, scope, and normalized labels are curation metadata, not legal advice or a permission finding. Review the current official source before transfer or reuse.

Related datasets with shared modalities

• eyecare_100k: Eyecare-100K: Multimodal Ophthalmology VQA Corpus (102,000 records, unknown)
• justraigs: JustRAIGS: Just Referral AI Glaucoma Screening Dataset (101,442 records, cc-by-nc-nd)
• eyepacs: EyePACS — Diabetic Retinopathy Detection (Kaggle 2015) (88,702 records, research-only)
• multieye: MultiEYE: OCT-Enhanced Fundus Multi-Disease Benchmark (58,036 records, mit)
• angioreport: AngioReport Fundus Angiography Report Dataset (55,361 records, unknown)
• ffa_ir: FFA-IR Medical Report Dataset (47,247 records, unknown)
• bidr: BiDR: Diabetic Retinopathy Diagnosis Dataset (35,126 records, unknown)
• dr_arranged: Diabetic Retinopathy Arranged Dataset (Tianchi 93926) (35,126 records, cc-by-nc-sa)