Awesome
MC-Stereo (3DV 2024)
This repository contains the source code for our paper:
MC-Stereo: Multi-Peak Lookup and Cascade Search Range for Stereo Matching<br/> Miaojie Feng, Junda Cheng, Hao jia, Longliang Liu, Gangwei Xu, Xin Yang<br/>
<img src="./images/MC-Stereo.png"> <img src="./images/vis_kitti.png">Comparision
| Method | KITTI-2012<br>(2-noc) | KITTI-2012<br>(2-noc-ref) | KITTI-2015<br>(D1-all) |
|---|---|---|---|
| ACVNet | 1.83 | 11.42 | 1.65 |
| RAFT-Stereo | - | - | 1.91 |
| IGEV-Stereo | 1.71 | 7.29 | 1.59 |
| CREStereo | 1.72 | 9.71 | 1.69 |
| MC-Stereo(Ours) | 1.55 | 6.82 | 1.55 |
Requirements
conda create -n MC_Stereo python=3.7
conda activate MC_Stereo
conda install pytorch==1.10.0 torchvision==0.11.0 torchaudio==0.10.0 cudatoolkit=11.3 -c pytorch -c conda-forge
pip install opencv-python
pip install scikit-image
pip install tensorboard
pip install tqdm
pip install timm==0.5.4
Required Data
To evaluate/train MC-Stereo, you will need to download the required datasets.
Evaluation
Pretrained models can be downloaded from google drive.
To evaluate on Scene Flow, run
sh evaluate.sh
Training
To train on Scene Flow, run
sh train_sceneflow.sh
To train on KITTI, run
sh train_kitti.sh
To train on ETH3D, run
sh train_eth3d.sh
Submission
For submission to the KITTI benchmark, run
python save_disp.py
Citation
If you find our work useful in your research, please consider citing our paper:
@article{feng2023mc,
title={MC-Stereo: Multi-peak Lookup and Cascade Search Range for Stereo Matching},
author={Feng, Miaojie and Cheng, Junda and Jia, Hao and Liu, Longliang and Xu, Gangwei and Yang, Xin},
journal={arXiv preprint arXiv:2311.02340},
year={2023}
}
Contact
Please feel free to contact me (Miaojie) at fmj@hust.edu.cn.
Acknowledgements
This project is heavily based on RAFT-Stereo, We thank the original authors for their excellent work.