A Domain Independent Approach for Learning-Based Monocular Depth Estimation

M.Mancini, G.Costante, P.Valigi, T.A. Ciarfuglia, J.Delmerico, D.Scaramuzza

Supplementary Material

[README: Models and benchmarks guide]

Models

[Download (374 MB)]

KITTI benchmark

[Download (350 MB)]

Zurich Forest benchmark

[Download (1,8 GB)]

Datasets

Depth GT is stored as a 8-bit png file, with values ranged between 0 and 255. You can get metric depth values by unnormalization (a 255 pixel value corresponds to 39.75 meters)

UVD train set

UVD test set

FVD train set

FVD test set