bi1997972026other
EgoVerse Annotation Full (v06)
Egocentric VR hand-tracking corpus with per-frame hand pose, wrist/camera trajectories, and language narratives paired with head-mounted video clips.
Downloads908
Episodes49650
Hours987.0
Why This Matters for Physical AI
This egocentric hand-tracking dataset from VR environments provides dense per-frame pose annotations and language narratives that enable training of visuomotor models for understanding human hand manipulation and embodied task execution in first-person perspective.
Technical Profile
- Modalities
- rgbhand_posecamera_trajectorylanguage
- Environment
- vr
- Episodes
- 49650
- Total Hours
- 987.0
- Data Format
- parquet
- Annotation Types
- hand_posecamera_poselanguage_narratives
- License
- other
Access
Need custom rgb data?
Claru builds purpose-built datasets for vr applications with dense human annotations and quality assurance.
Request a Sample Pack