Glossary
Hand pose estimation
Recovering 3D positions of hand and finger joints over time.
Definition
Hand pose estimation recovers the 3D positions of hand and finger joints across a clip, usually as a per-frame skeleton. For manipulation it is the difference between knowing a hand is present and knowing exactly how it is shaped around an object at contact.
01
Why it matters
Manipulation policies need the fine geometry of grasp and release. RGB alone leaves the model to infer 3D hand structure it cannot see reliably through occlusion and motion blur.
02
How Firsthand handles it
Firsthand delivers 3D hand pose as handJoints[2][21][3] in metric units in the head-camera frame, with a visibility flag per joint, aligned to depth and RGB.