Open-Source Periorbital Segmentation Dataset

Cropped periorbital images with multi-structure masks for the iris, sclera, eyelid, caruncle, and eyebrow.

At a glance

Field	Value
Short name	periorbital_segmentation
Full name	Open-Source Periorbital Segmentation Dataset
Primary category	external_eye
Contained modalities	external_eye
Tasks	segmentation, measurement
Samples	2,842
Classes	Not reported (Not reported)
Splits	all
Size	0.19 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	Zenodo
Availability	available (checked 2026-07-21)
Acquisition support	standard_platform_supported
Legacy sample-loader status	Metadata and access only

Notes

Five annotators labeled crops derived from the Chicago Face Database and CelebAMask-HQ. The Zenodo record is CC BY 4.0, but users should also verify the terms of both upstream face datasets.

Access preflight and acquisition

CLI

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

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

Python

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

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

Upstream page: https://doi.org/10.5281/zenodo.13916845

Source-term evidence: https://doi.org/10.5281/zenodo.13916845

Loader status

This catalog record provides metadata and access instructions, but it does not yet include a standard DatasetSample loader. Inspect the source file structure or contribute a loader before using it in a training pipeline.

Citation

BibTeX

@misc{periorbital_segmentation,
  title  = { Open-Source Periorbital Segmentation Dataset },
  note   = { Nahass GR, Koehler E, Tomaras N, et al. Open-Source Periorbital Segmentation Dataset for Ophthalmic Applications. Ophthalmology Science. 2025;5:100757. doi:10.1016/j.xops.2025.100757 },
  year   = { 2025 },
  url    = { https://doi.org/10.5281/zenodo.13916845 },
}

Plain text

Nahass GR, Koehler E, Tomaras N, et al. Open-Source Periorbital Segmentation Dataset for Ophthalmic Applications. Ophthalmology Science. 2025;5:100757. doi:10.1016/j.xops.2025.100757

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

• eyecare_100k: Eyecare-100K: Multimodal Ophthalmology VQA Corpus (102,000 records, unknown)
• lmod_plus: LMOD+ Multimodal Ophthalmology Benchmark (32,633 records, unknown)
• x_pcr: X-PCR Ophthalmology Progressive Clinical Reasoning Benchmark (18,700 records, unknown)
• popeye_nir: PopEYE Infrared Ocular Image Dataset (14,976 records, cc-by)
• mcoa: MCOA: Multimodal Corneal Opacity Assessment Dataset (6,664 records, cc-by)
• tear_meniscus: Multicentre Tear Meniscus Segmentation Dataset (3,432 records, cc-by)
• uveitis_smote: Image Dataset on Eye Diseases Classification with Symptoms and SMOTE Validation (3,245 records, cc-by)
• eed_astig: EED-Astig Pediatric External-Eye Dataset (3,088 records, research-only)