Search papers, labs, and topics across Lattice.
University of T眉bingen
1
0
2
Reconstructing complete, animatable 3D avatars from heavily occluded YouTube videos is now possible, thanks to a hallucination-as-supervision pipeline using diffusion models.
