Sourcing egocentric kitchen video

Quick facts

Closest public corpus

EPIC-KITCHENS-100 — 100 hours Full HD across 45 kitchens in 4 cities, 90,000 action segments (2022 release; HD-EPIC validation March 2025)

License

EPIC-KITCHENS is CC BY-NC 4.0 — non-commercial only, blocks commercial training.

Adjacent corpus

Ego4D — 3,670 hours general daily-life egocentric (subset of which is kitchen activity); requires Data Use Agreement.

Commercial gap

Buyers needing the right kitchen layout, appliance set, recipe, or commercial license cannot start from public corpora alone.

Spec checklist

Appliance manifest, recipe/task labels, hand visibility, contributor consent, commercial license, accepted/rejected episode samples.

Comparison

Dataset requirements

Buyers should specify first-person cooking, cleaning, object handling, and cabinet interaction video, accepted scenes in residential kitchens with varied layouts and appliances, a minimum of 10 hours of useful captured volume, consent rules, and the exact metadata package: task label, appliance type, object set, consent artifact, and clip boundary ^[1]. The Datasheets framework spells out which dataset-documentation questions matter before any commercial training program begins.

"The machine learning community currently has no standardized process for documenting datasets, which can lead to severe consequences in high-stakes domains."

Best-fit buyers

The strongest fit is household robotics, VLA, and world-model teams ^[3]. It can also work as a smaller eval set — typically 100 to 500 episodes — before a larger net-new capture program.

Egocentric kitchen video sample package

A credible egocentric kitchen video supplier should provide a sample package that includes raw files, a manifest, capture context, and these critical metadata fields: task label, appliance type, object set, consent artifact, and clip boundary ^[4]. Across at least 10 representative episodes, the buyer should be able to inspect whether first-person cooking, cleaning, object handling, and cabinet interaction video actually appears in residential kitchens with varied layouts and appliances, not just trust a verbal description of the inventory.

Egocentric kitchen video licensing check

The licensing review for egocentric kitchen video dataset should confirm whether the data is off-the-shelf or net-new, whether it can be used for commercial model training, whether contributors or sites consented, and whether the supplier can reproduce the same rights package for the full delivery ^[5]. A well-scoped buyer typically reviews 3 to 5 supplier deliveries before approving scale; without those checks, an apparently useful dataset can become a legal or procurement blocker.

Related pages

Use these to move from category-level context into specific task, dataset, format, and comparison detail.

External references and source context

1. Datasheets for Datasets

   Supports the dataset-requirements framework dimensions: dataset motivation, composition, collection process, recommended uses, and license review.

   arXiv ↩
2. Datasheets for Datasets

   Datasheets for Datasets defines the consent, provenance, and intended-use questions buyers must ask before commercial training; quoted verbatim in the dataset-requirements section.

   arXiv ↩
3. Open X-Embodiment: Robotic Learning Datasets and RT-X Models

   Open X-Embodiment establishes the cross-embodiment robotics pretraining baseline — a useful reference for buyer fit when mapping a deployment-specific dataset onto a generalist policy.

   arXiv ↩
4. Data Cards: Purposeful and Transparent Dataset Documentation for Responsible AI

   Data Cards capture dataset origins, development, intent, and ethical considerations buyers can attach to each delivered batch for procurement audit.

   arXiv ↩
5. encord

   Commercial vendors deliver licensed dataset collection programs with explicit contributor consent, rights, and per-batch documentation buyers can audit before scale.

   encord.com ↩

FAQ