Datasets:

harvardairobotics
/

Harvard-FairSeg

filename stringlengths 14 14	age float64 10.1 97.9	gender stringclasses 2 values	race stringclasses 3 values	ethnicity stringclasses 3 values	language stringclasses 4 values	maritalstatus stringclasses 6 values	use stringclasses 2 values
data_00001.npz	76.29	male	white	non-hispanic	english	unknown	training
data_00002.npz	57.92	female	white	unknown	english	single	training
data_00003.npz	35.19	female	white	non-hispanic	english	single	training
data_00004.npz	59.3	male	white	non-hispanic	english	married or partnered	training
data_00005.npz	47.92	female	white	hispanic	other languages	married or partnered	training
data_00006.npz	64.52	male	black	non-hispanic	english	married or partnered	training
data_00007.npz	71.18	female	white	non-hispanic	english	married or partnered	training
data_00008.npz	19.97	male	white	non-hispanic	english	single	training
data_00009.npz	51.13	male	black	non-hispanic	english	single	training
data_00010.npz	72.92	female	white	non-hispanic	english	married or partnered	training
data_00011.npz	77.05	male	white	non-hispanic	english	widowed	training
data_00012.npz	47.77	female	white	non-hispanic	english	married or partnered	training
data_00013.npz	54.81	female	white	hispanic	spanish	married or partnered	training
data_00014.npz	68.8	male	white	non-hispanic	english	married or partnered	training
data_00015.npz	58.73	male	asian	non-hispanic	english	married or partnered	training
data_00016.npz	55.72	male	black	non-hispanic	english	single	training
data_00017.npz	71.46	male	white	non-hispanic	english	married or partnered	training
data_00018.npz	33.21	male	white	non-hispanic	english	married or partnered	training
data_00019.npz	26.43	male	white	non-hispanic	english	single	training
data_00020.npz	79.89	male	white	non-hispanic	english	married or partnered	training
data_00021.npz	69.02	female	white	non-hispanic	english	married or partnered	training
data_00022.npz	64.33	female	white	non-hispanic	english	married or partnered	training
data_00023.npz	57.92	male	white	non-hispanic	english	married or partnered	training
data_00024.npz	54.87	female	white	non-hispanic	english	married or partnered	training
data_00025.npz	70.12	male	white	non-hispanic	english	married or partnered	training
data_00026.npz	57.42	female	white	non-hispanic	english	single	training
data_00027.npz	48.08	male	asian	non-hispanic	english	married or partnered	training
data_00028.npz	52.28	female	white	non-hispanic	english	married or partnered	training
data_00029.npz	68.1	female	white	non-hispanic	english	single	training
data_00030.npz	77.16	male	white	non-hispanic	english	married or partnered	training
data_00031.npz	78.5	male	white	hispanic	english	married or partnered	training
data_00032.npz	63.86	female	white	non-hispanic	english	married or partnered	training
data_00033.npz	46.44	female	white	non-hispanic	english	divorced	training
data_00034.npz	57.82	female	white	non-hispanic	english	divorced	training
data_00035.npz	53.1	female	white	non-hispanic	english	single	training
data_00036.npz	51.22	male	white	unknown	english	unknown	training
data_00037.npz	67.65	female	white	non-hispanic	english	single	training
data_00038.npz	51.87	female	black	unknown	english	single	training
data_00039.npz	77.49	male	white	non-hispanic	english	married or partnered	training
data_00040.npz	67.48	female	white	non-hispanic	english	widowed	training
data_00041.npz	59.94	female	black	non-hispanic	english	single	training
data_00042.npz	64.2	female	white	non-hispanic	english	married or partnered	training
data_00043.npz	63.02	male	white	non-hispanic	english	married or partnered	training
data_00044.npz	56.16	female	white	non-hispanic	english	single	training
data_00045.npz	80.54	female	white	non-hispanic	english	married or partnered	training
data_00046.npz	53.55	female	white	unknown	english	single	training
data_00047.npz	67.76	female	black	unknown	english	single	training
data_00048.npz	73.51	male	white	non-hispanic	english	married or partnered	training
data_00049.npz	37.1	female	white	non-hispanic	english	divorced	training
data_00050.npz	63.44	male	white	non-hispanic	english	married or partnered	training
data_00051.npz	39.12	male	white	non-hispanic	english	single	training
data_00052.npz	43.48	female	white	non-hispanic	spanish	married or partnered	training
data_00053.npz	69.58	female	black	non-hispanic	english	widowed	training
data_00054.npz	39.49	female	white	non-hispanic	english	single	training
data_00055.npz	26.66	male	white	non-hispanic	english	single	training
data_00056.npz	70.37	male	white	non-hispanic	english	married or partnered	training
data_00057.npz	56.29	male	white	non-hispanic	english	single	training
data_00058.npz	58.69	male	white	unknown	english	divorced	training
data_00059.npz	55.25	female	white	non-hispanic	english	married or partnered	training
data_00060.npz	29.55	female	black	non-hispanic	english	married or partnered	training
data_00061.npz	84.77	female	white	non-hispanic	english	married or partnered	training
data_00062.npz	49.58	male	white	non-hispanic	english	married or partnered	training
data_00063.npz	74.48	female	white	non-hispanic	english	married or partnered	training
data_00064.npz	65.09	female	white	non-hispanic	english	married or partnered	training
data_00065.npz	51.36	female	white	non-hispanic	english	married or partnered	training
data_00066.npz	69.77	male	white	non-hispanic	english	married or partnered	training
data_00067.npz	66.03	female	white	non-hispanic	english	single	training
data_00068.npz	15.36	male	white	non-hispanic	english	single	training
data_00069.npz	60.29	male	white	non-hispanic	english	married or partnered	training
data_00070.npz	67.93	male	white	non-hispanic	english	single	training
data_00071.npz	62.53	male	white	non-hispanic	english	married or partnered	training
data_00072.npz	54.48	female	white	non-hispanic	english	single	training
data_00073.npz	42.22	male	white	non-hispanic	english	married or partnered	training
data_00074.npz	75.48	female	white	non-hispanic	unknown	single	training
data_00075.npz	45.06	male	asian	non-hispanic	english	single	training
data_00076.npz	54.52	male	white	non-hispanic	english	married or partnered	training
data_00077.npz	75.88	female	white	non-hispanic	english	married or partnered	training
data_00078.npz	69.09	male	black	non-hispanic	english	single	training
data_00079.npz	75.97	female	white	non-hispanic	english	married or partnered	training
data_00080.npz	58.42	female	white	non-hispanic	english	single	training
data_00081.npz	31.72	female	black	non-hispanic	english	married or partnered	training
data_00082.npz	54.18	male	white	non-hispanic	english	married or partnered	training
data_00083.npz	44.38	female	white	non-hispanic	other languages	married or partnered	training
data_00084.npz	53.95	female	white	non-hispanic	english	married or partnered	training
data_00085.npz	76.4	male	white	non-hispanic	english	married or partnered	training
data_00086.npz	70.27	male	white	non-hispanic	english	married or partnered	training
data_00087.npz	39.35	male	white	non-hispanic	english	divorced	training
data_00088.npz	83.83	female	white	unknown	english	unknown	training
data_00089.npz	66.87	female	white	non-hispanic	english	unknown	training
data_00090.npz	58.21	female	white	non-hispanic	english	married or partnered	training
data_00091.npz	51.39	female	black	non-hispanic	other languages	single	training
data_00092.npz	76.67	male	white	unknown	english	single	training
data_00093.npz	76.51	female	black	non-hispanic	english	married or partnered	training
data_00094.npz	18.08	female	white	non-hispanic	english	single	training
data_00095.npz	77.22	female	white	non-hispanic	english	widowed	training
data_00096.npz	40.45	female	white	non-hispanic	english	married or partnered	training
data_00097.npz	52.62	female	white	non-hispanic	english	married or partnered	training
data_00098.npz	46.32	female	black	non-hispanic	english	divorced	training
data_00099.npz	78.68	male	white	non-hispanic	english	divorced	training
data_00100.npz	58.87	male	white	non-hispanic	english	married or partnered	training

End of preview. Expand in Data Studio

Dataset Card: Harvard-FairSeg

Dataset Summary

Harvard-FairSeg is a large-scale ophthalmology dataset for studying fairness in medical image segmentation. It contains 10,000 SLO fundus images with pixel-wise optic disc and cup segmentation masks, paired with comprehensive demographic annotations. The dataset is designed to benchmark and improve demographic equity in segmentation models, including foundation models such as SAM (Segment Anything Model).

This dataset was introduced at ICLR 2024: FairSeg: A Large-Scale Medical Image Segmentation Dataset for Fairness Learning Using Segment Anything Model with Fair Error-Bound Scaling.

Dataset Details

Dataset Description

Field	Value
Institution	Department of Ophthalmology, Harvard Medical School
Task	Optic disc and cup segmentation
Modality	Scanning Laser Ophthalmoscopy (SLO) fundus images
Scale	10,000 patients, 10,000 images
Annotation	Pixel-wise disc and cup masks
License	CC BY-NC-ND 4.0

Curated by: Yu Tian, Min Shi, Yan Luo, Ava Kouhana, Tobias Elze, Mengyu Wang
License: CC BY-NC-ND 4.0 — non-commercial research only
Paper: ICLR 2024
Contact: harvardophai@gmail.com, harvardairobotics@gmail.com

Data Fields

Each subject is stored as a .npz file (data_00001.npz … data_10000.npz) containing:

Field	Description
`slo_fundus`	Scanning Laser Ophthalmoscopy (SLO) fundus image
`disc_cup_mask`	Pixel-wise optic disc and cup segmentation mask
`age`	Patient age
`gender`	`0` = Female, `1` = Male
`race`	`0` = Asian, `1` = Black, `2` = White
`ethnicity`	`0` = Non-Hispanic, `1` = Hispanic, `-1` = Unknown
`language`	`0` = English, `1` = Spanish, `2` = Other, `-1` = Unknown
`maritalstatus`	`0` = Married/Partnered, `1` = Single, `2` = Divorced, `3` = Widowed, `4` = Legally Separated, `-1` = Unknown

A metadata overview is provided in data_summary.csv under the ReadMe folder.

Demographics

6 demographic attributes are annotated per patient: age, gender, race, ethnicity, preferred language, and marital status.

Uses

Direct Use

Fairness benchmarking for medical image segmentation models, including evaluation of demographic disparities in optic disc and cup segmentation. Intended for use with segmentation architectures such as SAMed and TransUNet.

Out-of-Scope Use

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

Citation

BibTeX:

@inproceedings{tianfairseg,
  title={FairSeg: A Large-Scale Medical Image Segmentation Dataset for Fairness Learning Using Segment Anything Model with Fair Error-Bound Scaling},
  author={Tian, Yu and Shi, Min and Luo, Yan and Kouhana, Ava and Elze, Tobias and Wang, Mengyu},
  booktitle={The Twelfth International Conference on Learning Representations}
}

APA:

Tian, Y., Shi, M., Luo, Y., Kouhana, A., Elze, T., & Wang, M. (2024). FairSeg: A Large-Scale Medical Image Segmentation Dataset for Fairness Learning Using Segment Anything Model with Fair Error-Bound Scaling. The Twelfth International Conference on Learning Representations (ICLR 2024).

Downloads last month: 2

Size of the auto-converted Parquet files:

151 kB

Number of rows:

10,000