Search papers, labs, and topics across Lattice.
1
1
3
LiftFormer achieves state-of-the-art monocular depth estimation by cleverly mapping image features into depth-oriented geometric subspaces, sidestepping the ill-posed nature of direct depth prediction.
