This is a personal reimplementation of PWC-Net [1] using PyTorch. Should you be making use of this work, please cite the paper accordingly. Also, make sure to adhere to the licensing terms of the authors. Should you be making use of this particular implementation, please acknowledge it appropriately [2].

For the original version of this work, please see: https://github.com/NVlabs/PWC-Net
Another optical flow implementation from me: https://github.com/sniklaus/pytorch-liteflownet
And another optical flow implementation from me: https://github.com/sniklaus/pytorch-unflow
Yet another optical flow implementation from me: https://github.com/sniklaus/pytorch-spynet

The authors of PWC-Net are thankfully already providing a reference implementation in PyTorch. However, its initial version did not reach the performance of the original Caffe version. This is why I created this repositroy, in which I replicated the performance of the official Caffe version by utilizing its weights.

The official PyTorch implementation has adopted my approach of using the Caffe weights since then, which is why they are all performing equally well now. Many people have reported issues with CUDA when trying to get the official PyTorch version to run though, while my reimplementaiton does not seem to be subject to such problems.

To download the pre-trained models, run bash download.bash. These originate from the original authors, I just converted them to PyTorch.

The correlation layer is implemented in CUDA using CuPy, which is why CuPy is a required dependency. It can be installed using pip install cupy or alternatively using one of the provided binary packages as outlined in the CuPy repository.

To run it on your own pair of images, use the following command. You can choose between two models, please make sure to see their paper / the code for more details.

python run.py --model default --first ./images/first.png --second ./images/second.png --out ./out.flo

I am afraid that I cannot guarantee that this reimplementation is correct. However, it produced results identical to the Caffe implementation of the original authors in the examples that I tried. Please feel free to contribute to this repository by submitting issues and pull requests.

As stated in the licensing terms of the authors of the paper, the models are free for non-commercial share-alike purpose. Please make sure to further consult their licensing terms.

[1]  @inproceedings{Sun_CVPR_2018,
         author = {Deqing Sun and Xiaodong Yang and Ming-Yu Liu and Jan Kautz},
         title = {{PWC-Net}: {CNNs} for Optical Flow Using Pyramid, Warping, and Cost Volume},
         booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},
         year = {2018}
     }

[2]  @misc{pytorch-pwc,
         author = {Simon Niklaus},
         title = {A Reimplementation of {PWC-Net} Using {PyTorch}},
         year = {2018},
         howpublished = {\url{https://github.com/sniklaus/pytorch-pwc}}
    }