MVSMamba: Multi-View Stereo with State Space Model
Authors: Jianfei Jiang, Qiankun Liu*, Hongyuan Liu, Haochen Yu, Liyong Wang, Jiansheng Chen, Huimin Ma*
Institute: University of Science and Technology Beijing
NeurIPS 2025

• 2025-12-04: Code and pre-trained model release !

conda create -n mvsmamba python=3.10.8
conda activate mvsmamba
pip install torch==1.13.1+cu117 torchvision==0.14.1+cu117 --extra-index-url https://download.pytorch.org/whl/cu117
pip install -r requirements.txt

Please refer to RRT-MVS.

You need to download extra Rectified_raw data for high-resolution training.

To train the model on DTU, specify DTU_TRAINING in ./scripts/train_dtu.sh first and then run:

bash scripts/train_dtu.sh

After training, you will get model checkpoints in ./checkpoints/dtu.

For DTU testing, just run:

bash scripts/test_dtu.sh

For TNT evaluation, just run:

bash scripts/test_tnt_inter.sh

bash scripts/test_tnt_adv.sh

For quantitative evaluation, you can upload your point clouds to Tanks and Temples benchmark.

If you find this work useful in your research, please consider citing the following:

@inproceedings{mvsmamba,
  title={MVSMamba: Multi-View Stereo with State Space Model},
  author={Jiang, Jianfei and Liu, Qiankun and Liu, Hongyuan and Yu, Haochen and Wang, Liyong and Chen, Jiansheng and Ma, Huimin},
  booktitle={The Thirty-ninth Annual Conference on Neural Information Processing Systems}
}

Our work is partially based on these opening source works ET-MVSNet, JamMa, and EfficientVMamba. We appreciate their contributions to the MVS community.