Our significant extension version of IGEV, named IGEV++, is available at Paper, Code

IGEV-Stereo (CVPR 2023)

This repository contains the source code for our paper:

Iterative Geometry Encoding Volume for Stereo Matching
Gangwei Xu, Xianqi Wang, Xiaohuan Ding, Xin Yang

📢 News

2024-12-30: We add bfloat16 training to prevent potential NAN issues during the training process.

Demos

Pretrained models can be downloaded from google drive

We assume the downloaded pretrained weights are located under the pretrained_models directory.

You can demo a trained model on pairs of images. To predict stereo for Middlebury, run

python demo_imgs.py \
--restore_ckpt pretrained_models/sceneflow/sceneflow.pth \
-l=path/to/your/left_imgs \
-r=path/to/your/right_imgs

or you can demo a trained model pairs of images for a video, run:

python demo_video.py \
--restore_ckpt pretrained_models/sceneflow/sceneflow.pth \
-l=path/to/your/left_imgs \
-r=path/to/your/right_imgs

To save the disparity values as .npy files, run any of the demos with the --save_numpy flag.

Comparison with RAFT-Stereo

Method	KITTI 2012 (3-noc)	KITTI 2015 (D1-all)	Memory (G)	Runtime (s)
RAFT-Stereo	1.30 %	1.82 %	1.02	0.38
IGEV-Stereo	1.12 %	1.59 %	0.66	0.18

Environment

• NVIDIA RTX 3090
• Python 3.8

Create a virtual environment and activate it.

conda create -n IGEV python=3.8
conda activate IGEV

Dependencies

bash env.sh

Alternatively, you can install a higher version of PyTorch that supports bfloat16 training.

bash env_bfloat16.sh

Required Data

To evaluate/train IGEV-Stereo, you will need to download the required datasets.

• Scene Flow (Includes FlyingThings3D, Driving & Monkaa)
• KITTI
• Middlebury
• ETH3D

By default core/stereo_datasets.py will search for the datasets in these locations.

├── /data
    ├── sceneflow
        ├── frames_finalpass
            ├── TRAIN
                ├── A
                ├── ...
                ├── 15mm_focallength
                ├── ...
                ├── funnyworld_augmented0_x2
                ├── ...
            ├── TEST
        ├── disparity
    ├── KITTI
        ├── KITTI_2012
            ├── training
            ├── testing
            ├── vkitti
        ├── KITTI_2015
            ├── training
            ├── testing
            ├── vkitti
    ├── Middlebury
        ├── trainingH
        ├── trainingH_GT
    ├── ETH3D
        ├── two_view_training
        ├── two_view_training_gt
    ├── DTU_data
        ├── dtu_train
        ├── dtu_test

You should replace the default path with your own.

Evaluation

To evaluate on Scene Flow or Middlebury or ETH3D, run

python evaluate_stereo.py --restore_ckpt ./pretrained_models/sceneflow/sceneflow.pth --dataset sceneflow

or

python evaluate_stereo.py --restore_ckpt ./pretrained_models/sceneflow/sceneflow.pth --dataset middlebury_H

or

python evaluate_stereo.py --restore_ckpt ./pretrained_models/sceneflow/sceneflow.pth --dataset eth3d

Training

To train on Scene Flow, run

python train_stereo.py --logdir ./checkpoints/sceneflow

To train on KITTI, run

python train_stereo.py --logdir ./checkpoints/kitti --restore_ckpt ./pretrained_models/sceneflow/sceneflow.pth --train_datasets kitti

Bfloat16 Training

NaN values during training: If you encounter NaN values in your training, this is likely due to overflow when using float16. This can happen when large gradients or high activation values exceed the range represented by float16. To fix this:

-Try switching to bfloat16 by using --precision_dtype bfloat16.

-Alternatively, you can use float32 precision by setting --precision_dtype float32.

Training with bfloat16

1. Before you start training, make sure you have hardware that supports bfloat16 and the right environment set up for mixed precision training.

bash env_bfloat16.sh

2. Then you can train the model with bfloat16 precision:

python train_stereo.py --mixed_precision --precision_dtype bfloat16

Submission

For submission to the KITTI benchmark, run

python save_disp.py

MVS training and evaluation

To train on DTU, run

python train_mvs.py

To evaluate on DTU, run

python evaluate_mvs.py

Citation

If you find our work useful in your research, please consider citing our paper:

@inproceedings{xu2023iterative,
  title={Iterative Geometry Encoding Volume for Stereo Matching},
  author={Xu, Gangwei and Wang, Xianqi and Ding, Xiaohuan and Yang, Xin},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={21919--21928},
  year={2023}
}

@article{xu2025igev++,
  title={Igev++: Iterative multi-range geometry encoding volumes for stereo matching},
  author={Xu, Gangwei and Wang, Xianqi and Zhang, Zhaoxing and Cheng, Junda and Liao, Chunyuan and Yang, Xin},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year={2025},
  publisher={IEEE}
}

Acknowledgements

This project is based on RAFT-Stereo, and CoEx. We thank the original authors for their excellent works.