Monodepth2-from-scratch

This my simplified re-implementation of monodepth2 in pytorch

Monodepth2 is self-supervised depth estimation network.

This model was trained on Kitti-raw dataset. I used Eigen split used by Zhou et.al.

Requirements

pytorch=1.11.0+cu113
tensorboard=2.7.0

Dataset format

Root
├───2011_09_26
│ ├───2011_09_26_drive_0001_sync
│ │ ├───image_02
│ │ │ └───data
│ │ ├───image_03
│ │ │ └───data
.
.
│ ├───2011_10_03_drive_0058_sync
│ │ ├───image_02
│ │ │ └───data
│ │ ├───image_03
│ │ │ └───data

Results

Tensorboard outputs

Original authors and paper:

Digging into Self-Supervised Monocular Depth Prediction

Clément Godard, Oisin Mac Aodha, Michael Firman and Gabriel J. Brostow

ICCV 2019 (arXiv pdf)

Credits

@article{monodepth2,
  title     = {Digging into Self-Supervised Monocular Depth Prediction},
  author    = {Cl{\'{e}}ment Godard and
               Oisin {Mac Aodha} and
               Michael Firman and
               Gabriel J. Brostow},
  booktitle = {The International Conference on Computer Vision (ICCV)},
  month = {October},
year = {2019}
}