Taihong Xiao, Jinwei Yuan, Deqing Sun, Qifei Wang, Xin-Yu Zhang, Kehan Xu, Ming-Hsuan Yang
Please cite our paper if you find it useful to your research.
@inproceedings{xiao2020learnable,
title={Learnable cost volume using the cayley representation},
author={Xiao, Taihong and Yuan, Jinwei and Sun, Deqing and Wang, Qifei and Zhang, Xin-Yu and Xu, Kehan and Yang, Ming-Hsuan},
booktitle={European Conference on Computer Vision (ECCV)},
pages={483--499},
year={2020},
organization={Springer}
}
The learnable cost volume can be easily implemented using either pytorch or tensorflow.
For the pytorch implementation, please refer to corr.py
This code is modified for RAFT+LCV.
Compare the classes LearnableCorrBlock
with CorrBlock
to know the differences between learnable cost volume and vanilla cost volume.
For the tensorflow implementation, please refer to network.py
for details.
This code is modified for DDFlow+LCV.