Ophora-160K: Ophthalmic Surgical Video Instruction Dataset

160,185 video clip-instruction pair samples from 9,819 ophthalmic surgical videos, covering multiple procedure types. Designed for text-guided surgical video generation and understanding. Published at MICCAI 2025.

At a glance

Field	Value
Short name	ophora
Full name	Ophora-160K: Ophthalmic Surgical Video Instruction Dataset
Primary category	surgical_video
Contained modalities	surgical_video, text
Tasks	classification, phase_recognition
Samples	160,185
Classes	Not reported (Not reported)
Splits	train
Size	500.0 GB
Source-stated terms	Unknown — no license stated
Normalized terms	unknown
Descriptive screening label	Unknown or unclear; do not assume permission
Terms scope	unknown
Access friction	self_service_authenticated
Route backend	HuggingFace Hub
Availability	available (checked 2026-07-21)
Acquisition support	standard_platform_supported
Legacy sample-loader status	Standard loader included

Notes

162,185 video clip-instruction pairs from 9,819 source videos (avg clip length ~5.5 s). Estimated ~500 GB total. Video clip-instruction pair dataset for text-guided surgical video generation. May require HF_TOKEN for gated access. Set in .env.

Access preflight and acquisition

CLI

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

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

Python

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

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

Upstream page: huggingface.co/datasets

Source-term evidence: huggingface.co/datasets

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

Citation

BibTeX

@misc{ophora,
  title  = { Ophora-160K: Ophthalmic Surgical Video Instruction Dataset },
  note   = { Ophora: Text-Guided Ophthalmic Surgical Video Generation. MICCAI 2025. arXiv:2505.07449. HuggingFace: https://huggingface.co/datasets/General-Medical-AI/Ophora-160K — GitHub: https://github.com/uni-medical/Ophora },
  year   = { 2025 },
  url    = { https://huggingface.co/datasets/General-Medical-AI/Ophora-160K },
}

Plain text

Ophora: Text-Guided Ophthalmic Surgical Video Generation. MICCAI 2025. arXiv:2505.07449. HuggingFace: https://huggingface.co/datasets/General-Medical-AI/Ophora-160K — GitHub: https://github.com/uni-medical/Ophora

Source-stated terms

• Raw source string: Unknown — no license stated
• Normalized category: unknown
• Apparent scope: unknown
• Descriptive screening label: Unknown or unclear; do not assume permission

⚠️ 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

• lmod_plus: LMOD+ Multimodal Ophthalmology Benchmark (32,633 records, unknown)
• lmod_cataract_1k_cot: Cataract-1K Surgical Analysis Chain-of-Thought Dataset (count not reported records, mit)
• fundus_cc_2_5m: Fundus-CC-2.5M Text Corpus (2,500,000 records, unknown)
• ocular_chat_vqa: OcularChat-VQA: AREDS-Derived Patient-Physician Dialogue Dataset (844,000 records, cc-by-nc-sa)
• fundus_105k: Fundus-105K Text Dataset (105,000 records, unknown)
• eyecare_100k: Eyecare-100K: Multimodal Ophthalmology VQA Corpus (102,000 records, unknown)
• angioreport: AngioReport Fundus Angiography Report Dataset (55,361 records, unknown)
• ophthalmology_mcqa_v3: Ophthalmology-MCQA-v3 (51,745 records, unknown)