OculoScope: Fairer AI in Ophthalmology Dataset

16,530 ultra-widefield fundus images from 8,405+ patients (age 0–90) annotated for 38 ophthalmic diseases and 67 fundus features. Released alongside the FairerOPTH study on sexism and ageism in ophthalmic AI (Nature Communications 2024).

At a glance

Field	Value
Short name	oculoscope
Full name	OculoScope: Fairer AI in Ophthalmology Dataset
Primary category	uwf_fundus
Contained modalities	uwf_fundus, tabular
Tasks	classification, multilabel
Samples	16,530
Classes	38 (Not reported)
Splits	train, test
Size	20.0 GB
Source-stated terms	CC BY 4.0
Normalized terms	cc-by
Descriptive screening label	Standard label without an explicit NC clause; not a permission finding
Terms scope	dataset_files
Access friction	anonymous_direct
Route backend	Figshare
Availability	available (checked 2026-07-21)
Acquisition support	guided_instructions_only
Legacy sample-loader status	Standard loader included

Notes

Also annotates 67 fine-grained fundus features in addition to 38 disease-level labels. Full class list in dataset CSV.

Access preflight and acquisition

CLI

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

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

Python

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

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

Upstream page: Figshare private-share page

Source-term evidence: Figshare private-share page

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('oculoscope')
samples = ds.load(data_dir, split='test')
for s in samples[:5]:
    print(s.sample_id, s.label, s.image_path)

Citation

BibTeX

@misc{oculoscope,
  title  = { OculoScope: Fairer AI in Ophthalmology Dataset },
  note   = { FairerOPTH Study — OculoScope Dataset. Nature Communications 2024. https://www.nature.com/articles/s41467-024-48972-0 — Data: https://figshare.com/s/926c2c2ef9e77ab5eb9d },
  year   = { 2024 },
  url    = { https://figshare.com/s/926c2c2ef9e77ab5eb9d },
}

Plain text

FairerOPTH Study — OculoScope Dataset. Nature Communications 2024. https://www.nature.com/articles/s41467-024-48972-0 — Data: https://figshare.com/s/926c2c2ef9e77ab5eb9d

Source-stated terms

• Raw source string: CC BY 4.0
• Normalized category: cc-by
• Apparent scope: dataset_files
• Descriptive screening label: Standard label without an explicit NC clause; not a permission finding

⚠️ 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

• ocular_chat_vqa: OcularChat-VQA: AREDS-Derived Patient-Physician Dialogue Dataset (844,000 records, cc-by-nc-sa)
• lmod_plus: LMOD+ Multimodal Ophthalmology Benchmark (32,633 records, unknown)
• mmrdr: MMRDR: Multi-Modal Retinal Diabetic Retinopathy Dataset (24,460 records, cc-by)
• tsukazaki_uwf: Tsukazaki Hospital UWF Fundus Dataset (13,047 records, research-only)
• leops_erg: LEOPs Light-Adapted Electroretinogram and Oscillatory Potentials Dataset (9,743 records, cc-by)
• olives: OLIVES: Ophthalmic Labels for Investigating Visual Eye Semantics (9,408 records, cc-by)
• birdshot_wide: Birdshot-WIDE Widefield Fundus Dataset (6,352 records, cc-by)
• fprm_retina: FPRM Multimodal Eye Imaging and Psychological Assessment Dataset (3,361 records, research-only)