IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo
Abstract
We present IterMVS, a new data-driven method for high-resolution multi-view stereo. We propose a novel GRU-based estimator that encodes pixel-wise probability distributions of depth in its hidden state. Ingesting multi-scale matching information, our model refines these distributions over multiple iterations and infers depth and confidence. To extract the depth maps, we combine traditional classification and regression in a novel manner. We verify the efficiency and effectiveness of our method on DTU, Tanks&Temples and ETH3D. While being the most efficient method in both memory and run-time, our model achieves competitive performance on DTU and better generalization ability on Tanks&Temples as well as ETH3D than most state-of-the-art methods. Code is available at https://github.com/FangjinhuaWang/IterMVS.
Cite
Text
Wang et al. "IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo." Conference on Computer Vision and Pattern Recognition, 2022. doi:10.1109/CVPR52688.2022.00841Markdown
[Wang et al. "IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo." Conference on Computer Vision and Pattern Recognition, 2022.](https://mlanthology.org/cvpr/2022/wang2022cvpr-itermvs/) doi:10.1109/CVPR52688.2022.00841BibTeX
@inproceedings{wang2022cvpr-itermvs,
title = {{IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo}},
author = {Wang, Fangjinhua and Galliani, Silvano and Vogel, Christoph and Pollefeys, Marc},
booktitle = {Conference on Computer Vision and Pattern Recognition},
year = {2022},
pages = {8606-8615},
doi = {10.1109/CVPR52688.2022.00841},
url = {https://mlanthology.org/cvpr/2022/wang2022cvpr-itermvs/}
}