Aligned Monocular Depth Estimation for Dynamic Videos
Permutation-Equivariant Visual Geometry Learning