Skip to content

jiaxiZeng/Parameterized-Cost-Volume-for-Stereo-Matching

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Parameterized Cost Volume for Stereo Matching (ICCV2023)

This repository is the implementation of Parameterized Cost Volume for Stereo Matching.

We thank the authors of RAFT-Stereo, as our code is built on top of their project.

Data

To evaluate/train this model, you need to download the required datasets,

and organize them as following:

├── datasets
    ├── FlyingThings3D
        ├── frames_cleanpass
        ├── frames_finalpass
        ├── disparity
    ├── Monkaa
        ├── frames_cleanpass
        ├── frames_finalpass
        ├── disparity
    ├── Driving
        ├── frames_cleanpass
        ├── frames_finalpass
        ├── disparity
    ├── KITTI
        ├──Kitti12
           ├── testing
           ├── training
        ├──Kitti15
           ├── testing
           ├── training
    ├── Middlebury
        ├── trainingF
        ├── trainingH
        ├── trainingQ
        ├── testF
        ├── testH
        ├── testQ

The datasets directory need to be placed on the root of this project.

Environment

CUDA 11.3

python 3.8.12

pytorch 1.12.0

pip install scipy
pip install tqdm
pip install tensorboard
pip install opt_einsum
pip install imageio
pip install opencv-python
pip install scikit-image
pip install einops
pip install wandb
pip install matplotlib

Train

bash ./train_sceneflow.sh

Evaluate

You need to create a directory called pcv_ckpts in the root directory of this project, then download chechkpoint to the directory, and run the following code:

bash ./test_sceneflow.sh

Citation

@inproceedings{zeng2023parameterized,
  title={Parameterized Cost Volume for Stereo Matching},
  author={Zeng, Jiaxi and Yao, Chengtang and Yu, Lidong and Wu, Yuwei and Jia, Yunde},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={18347--18357},
  year={2023}
}

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published