Code repository for "Stable View Synthesis".

Install the following Python packages in your Python environment

- numpy (1.19.1)
- scikit-image (0.15.0)
- pillow (7.2.0)
- torch
- torchvision (0.7.0)
- torch-scatter (1.6)
- torch-sparse (1.6)
- torch-geometric (1.6)
- torch-sparse (1.6)
- open3d (0.11)
- python-opencv
- matplotlib (3.2.x)
- pandas (1.0.x)

To compile the Python extensions you will also need Eigen and cmake.

Clone the repository and initialize the submodule

git clone https://github.com/intel-isl/StableViewSynthesis.git
cd StableViewSynthesis
git submodule update --init --recursive

Finally, build the Python extensions

cd ext/preprocess
cmake -DCMAKE_BUILD_TYPE=Release .
make 

cd ../mytorch
python setup.py build_ext --inplace

Tested with Ubuntu 18.04 and macOS Catalina.

Make sure you adapted the paths in config.py to point to the downloaded data!

cd experiments

Then run the evaluation via

python exp.py --net resunet3.16_penone.dirs.avg.seq+9+1+unet+5+2+16.single+mlpdir+mean+3+64+16 --cmd eval --iter last --eval-dsets tat-subseq

This will run the pretrained network on the four Tanks and Temples sequences.

To train the network from scratch you can run

python exp.py --net resunet3.16_penone.dirs.avg.seq+9+1+unet+5+2+16.single+mlpdir+mean+3+64+16 --cmd retrain

See FreeViewSynthesis.

Please cite our paper if you find this work useful.

@inproceedings{Riegler2021SVS,
  title={Stable View Synthesis},
  author={Riegler, Gernot and Koltun, Vladlen},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  year={2021}
}