Skip to content

Latest commit

 

History

History
46 lines (41 loc) · 1.04 KB

README.md

File metadata and controls

46 lines (41 loc) · 1.04 KB
Raw

Efficient Edge-Preserving Multi-View Stereo Network for Depth Estimation (AAAI 2023 Oral)

Environment

  • PyTorch 1.4.0
  • Python 3.7
  • open3d 0.9.0.0
  • numpy 1.20.3

Scripts

1. train on DTU

bash scripts/train_dtu.sh

2. test on DTU

bash scripts/test_dtu.sh

3. finetune on BlendedMVS

bash scripts/train_blended.sh

4. test on Tanks and Temple

bash scripts/test_tt.sh

5. test on ETH3D

bash scripts/test_eth.sh

Citation

@inproceedings{Su2023epnet,
  title={Efficient Edge-Preserving Multi-View Stereo Network for Depth Estimation},
  author={Su, Wanjuan and Tao, Wenbing},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  year={2023},
  number={2},
  pages={2348--2356},
  number={37},
}

Acknowledge

Our work is partially based on these opening source work: Casmvsnet and Vis-MVSNet. Thanks for their contributions to the MVS community.