https://github.com/prinsphield/lcv
Learnable Cost Volume Using the Cayley Representation
https://github.com/prinsphield/lcv
Last synced: 3 months ago
JSON representation
Learnable Cost Volume Using the Cayley Representation
- Host: GitHub
- URL: https://github.com/prinsphield/lcv
- Owner: Prinsphield
- License: mit
- Created: 2021-03-29T23:36:26.000Z (about 4 years ago)
- Default Branch: main
- Last Pushed: 2021-03-30T00:42:34.000Z (about 4 years ago)
- Last Synced: 2025-01-11T00:33:19.471Z (4 months ago)
- Language: Python
- Homepage: https://arxiv.org/abs/2007.11431
- Size: 8.79 KB
- Stars: 3
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Learnable Cost Volume Using the Cayley Representation
Taihong Xiao, Jinwei Yuan, Deqing Sun, Qifei Wang, Xin-Yu Zhang, Kehan Xu, Ming-Hsuan Yang
Please cite our [paper](https://arxiv.org/abs/2007.11431) 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}
}
```
## Introduction
The learnable cost volume can be easily implemented using either pytorch or tensorflow.
For the pytorch implementation, please refer to [`corr.py`](https://github.com/Prinsphield/LCV/blob/main/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`](https://github.com/Prinsphield/LCV/blob/main/network.py#L122-L162) for details.
This code is modified for DDFlow+LCV.