harvardairobotics/Harvard-GF

---
license: cc-by-nc-nd-4.0
task_categories:
- image-classification
modality:
- image
language:
- en
tags:
- medical
- ophthalmology
- fairness
- glaucoma
- OCT
- retinal-nerve
pretty_name: Harvard-GF
size_categories:
- 1K<n<10K
---

# 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](https://ieeexplore.ieee.org/abstract/document/10472539).

## 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](https://creativecommons.org/licenses/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](https://creativecommons.org/licenses/by-nc-nd/4.0/) — non-commercial research only
- **Paper:** [IEEE TMI 2024](https://ieeexplore.ieee.org/abstract/document/10472539)
- **Contact:** [e-mail], [e-mail]

### 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:**

```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