CUSTOM GRADEVIDEO
Egocentric Task Demonstrations
First-person video of real-world task completion with synchronized audio, hand-tracking, and environmental metadata for robotics and embodied AI.
- Frame rate
- Up to 60fps
- Quality
- Multi-layer QC
- Availability
- Sample clips on request
[ OVERVIEW ]
First-person video capturing how humans actually complete real-world tasks: cooking, repair, assembly, navigation, craft, and household work. Recordings include synchronized audio narration, hand-tracking data, environmental metadata, and step-level task annotations. Built for teams training embodied AI agents, robotic manipulation policies, and task-understanding models. Every contributor consents to derivative-work usage with demographic metadata on file.
[ KEY HIGHLIGHTS ]
- First-person perspective with synchronized audio narration
- Hand-tracking and keypoint data where the task and lighting allow
- Step-level task annotations with verb-object labels per action
- Environmental metadata: location class, lighting, clutter, object count
- Full 1080p at up to 60fps with standard or action-camera capture
- Diverse task domains: cooking, repair, assembly, navigation, craft
- Licensed per task domain or across the full egocentric corpus
[ TECHNICAL SPECIFICATIONS ]
- Files
- MP4 H.264 at 1080p/30-60fps with synchronized WAV audio; action-camera variants available
- Annotations
- Step-level task labels · verb-object pairs · hand-tracking data · environmental tags
- Schema
- COCO-style JSON · Ego4D-compatible schema · custom task taxonomies on request
- Licensing
- Commercial training rights · derivative-work rights signed · per-domain or full-corpus
More from the catalog.
Explore the full catalog, or scope a custom build matched to your brief.