Dyna-DM: Dynamic Object-aware Self-supervised Monocular Depth Maps

Dyna-DM: Dynamic Object-aware Self-supervised Monocular Depth Maps

[PDF]

Install

The models were trained using CUDA 11.1, Python 3.7.x (conda environment), and PyTorch 1.8.0.

Create a conda environment with the PyTorch library:

conda create -n my_env python=3.7.4 pytorch=1.8.0 torchvision torchaudio cudatoolkit=11.1 -c pytorch -c nvidia
conda activate my_env

Install prerequisite packages listed in requirements.txt:

pip3 install -r requirements.txt

Also, ensure to install torch-scatter and torch-sparse:

pip3 install torch-scatter==2.0.8 torch-sparse==0.6.12 -f https://pytorch-geometric.com/whl/torch-1.8.0+cu111.html

Datasets

We use the datasets provided by Insta-DM and evaluate the model with the KITTI Eigen Split using the raw KITTI dataset.

Models

Pretrained models for CityScape and KITTI+CityScape are provided here, where KITTI+CityScape is trained on both CityScape and KITTI and leads to the greatest depth estimations.

Training

The models can be trained on the KITTI dataset by running:

bash scripts/train_kt.sh

Also, the models can be trained on the CityScape dataset by running:

bash scripts/train_cs.sh

The hyperparameters are defined in each script file and set at their defaults as stated in the paper.

Evaluation

We evaluate the models by running:

bash scripts/run_eigen_test.sh

References

• Insta-DM (AAAI 2021, our baseline framework)

• Struct2Depth (AAAI 2019, object scale loss)

• SC-SfMLearner (NeurIPS 2019)