Search papers, labs, and topics across Lattice.
1
0
2
6
A single generative model recovers temporally consistent 3D hand motion from video by jointly learning 2D-3D pose dynamics, bypassing unreliable per-frame estimates.
