Egocentric video capturing full household tasks from initiation to completion, with hands visible in most frames and per-action narration. Supports training on Ego4D and EgoSchema.
Available Now · Sample Available
Egocentric Dataset
First-person video of complete household task workflows, narrated by the performer.
Volume
0
Volume
Scalable Volume
Coverage
Video Understanding
Targets
Modality & Delivery
VideoAudio→.mp4
Formats
What You Receive
Coverage
Package Structure
<clip_id>/One folder per egocentric clip
- video.mp4The first-person video segment
- narration.jsonTime-aligned narrations and action labels
- annotations/Frame-level annotations (objects, hands, gaze)
Differentiators
- Captures full first-person household workflows with visible hands and narrated actions.
Record Schema
Results
Methodology & Quality
Methodology
SME involvement
Domain-expert curated with up to L2 review
Stressors
Long-form temporal understanding of embodied household activity.
Design criteria
- Record tasks from initiation through completion rather than isolated clips.
Build assumptions
Quality approach
Known limitations
Customization & Scalability
Custom sets available. Additional volume in progress
Get Started
Request a Sample
Share your target domains and the OTS team will follow up with sample data and next steps.