Glossary
Egocentric video
Definition
Egocentric video is footage recorded from a head-mounted camera worn by the person performing a task, so the camera shares the actor’s viewpoint and moves with their head. For embodied AI it matters because the observation distribution matches what a robot with a head or chest camera actually sees.
01
Why it matters
A policy trained on third-person footage sees a viewpoint it will never occupy at runtime. Egocentric video puts hands entering frame from below, occlusion during grasp, and motion blur during reach into the training distribution, which is exactly what an embodied agent must handle.
02
How Firsthand handles it
Firsthand captures egocentric video from a head-mounted global-shutter camera plus two wrist cameras, all hardware-triggered onto one clock, so the first-person stream is aligned with depth, pose and action labels rather than recorded in isolation.
Related terms