Dataset Card: Harvard-GF

Dataset Summary

Harvard-GF (Harvard Glaucoma Fairness) is a retinal nerve disease dataset for fairness learning in glaucoma detection, featuring both 2D and 3D OCT imaging data with balanced racial groups. It contains 3,300 samples from 3,300 patients with equal representation across Asian, Black, and White racial groups — a unique design addressing the doubled glaucoma prevalence observed in Black patients compared to other races.

This dataset was introduced in IEEE Transactions on Medical Imaging 2024: Harvard Glaucoma Fairness: A Retinal Nerve Disease Dataset for Fairness Learning and Fair Identity Normalization.

Dataset Details

Dataset Description

Field	Value
Institution	Department of Ophthalmology, Harvard Medical School
Task	Glaucoma detection
Modality	OCT RNFLT maps (2D), OCT B-scans (3D), visual field (MD, TDs)
Scale	3,300 patients, 3,300 OCT RNFLT maps
Image size	200 × 200 (RNFLT map), 200 × 200 × 200 (B-scans)
Splits	2,100 train / 300 validation / 900 test
License	CC BY-NC-ND 4.0

• Curated by: Yan Luo, Yu Tian, Min Shi, Louis R. Pasquale, Lucy Q. Shen, Nazlee Zebardast, Tobias Elze, Mengyu Wang
• License: CC BY-NC-ND 4.0 — non-commercial research only
• Paper: IEEE TMI 2024
• Contact: harvardophai@gmail.com, harvardairobotics@gmail.com

Data Fields

Each subject is stored as a .npz file containing:

Field	Description
rnflt	OCT retinal nerve fiber layer thickness (RNFLT) map, size 200 × 200
oct_bscans	3D OCT B-scans image, size 200 × 200 × 200
glaucoma	Glaucomatous status: 0 = non-glaucoma, 1 = glaucoma
md	Mean deviation value of visual field
tds	52 total deviation values of visual field
age	Patient age
male	Gender: 0 = Female, 1 = Male
race	0 = Asian, 1 = Black or African American, 2 = White or Caucasian
ethnicity	0 = Non-Hispanic, 1 = Hispanic, -1 = Unknown
language	0 = English, 1 = Spanish, 2 = Other, -1 = Unknown
maritalstatus	0 = Married/Civil Union/Life Partner, 1 = Single, 2 = Divorced, 3 = Widowed, 4 = Legally Separated, -1 = Unknown

Demographics

A key feature of Harvard-GF is its balanced racial composition: equal numbers of Asian, Black, and White patients are included across splits, enabling rigorous racial fairness evaluation without class imbalance.

Uses

Direct Use

• Fairness benchmarking for glaucoma detection models using 2D and 3D OCT imaging
• Racial and gender fairness analysis in ophthalmic AI
• Development and evaluation of fairness learning methods (e.g., fair identity normalization)
• Equity-scaled performance measurement across demographic subgroups

Out-of-Scope Use

Clinical decisions, patient care, or any commercial application. This dataset shall not be used for clinical decisions at any time.

Access

The "Harvard" designation indicates the dataset originates from the Department of Ophthalmology at Harvard Medical School. It does not imply endorsement, sponsorship, or legal responsibility by Harvard University or Harvard Medical School.

Citation

BibTeX:

@article{10472539,
  author={Luo, Yan and Tian, Yu and Shi, Min and Pasquale, Louis R. and Shen, Lucy Q. and Zebardast, Nazlee and Elze, Tobias and Wang, Mengyu},
  journal={IEEE Transactions on Medical Imaging},
  title={Harvard Glaucoma Fairness: A Retinal Nerve Disease Dataset for Fairness Learning and Fair Identity Normalization},
  year={2024},
  volume={43},
  number={7},
  pages={2623-2633},
  doi={10.1109/TMI.2024.3377552}
}

APA:

Luo, Y., Tian, Y., Shi, M., Pasquale, L. R., Shen, L. Q., Zebardast, N., Elze, T., & Wang, M. (2024). Harvard Glaucoma Fairness: A Retinal Nerve Disease Dataset for Fairness Learning and Fair Identity Normalization. IEEE Transactions on Medical Imaging, 43(7), 2623–2633. https://doi.org/10.1109/TMI.2024.3377552