EgoDex
Apple
Egocentric tabletop manipulation video paired with hand, upper-body and head poses and language annotations.
What you can explore
Inspect human manipulation sequences and pose annotations as an input to imitation or representation learning.
Before you use it
Human pose observations are not robot action commands. The full release is large; inspect a sample or the test split first. Dataset terms restrict commercial reuse and derivatives; code terms are separate.
Original contributors
Ryan Hoque, Peide Huang, David J. Yoon, Mouli Sivapurapu and Jian Zhang
License / access: Dataset: CC BY-NC-ND · code: Apple sample-code license
Dataset, annotations & access terms
Keep exploring
DexCap
Hand motion capture and scene observations for learning dexterous manipulation from human demonstrations.
HOT3D
Multi-view egocentric recordings of human hands interacting with objects, with tracking annotations.
HandEdit
A dataset and benchmark for replacing visible human hands or arms with specified robot embodiments in egocentric images.
Editorial reference · checked 2026-09-22