Learning Efficient Photometric Feature Transform for Multi-View Stereo

Kaizhang Kang, Cihui Xie, Ruisheng Zhu, Xiaohe Ma, Ping Tan, Hongzhi Wu, Kun Zhou

ICCV 2021 pp. 5956-5965

doi:10.1109/ICCV48922.2021.00590 /iccv/2021/kang2021iccv-learning/

Abstract

We present a novel framework to learn to convert the per-pixel photometric information at each view into spatially distinctive and view-invariant low-level features, which can be plugged into existing multi-view stereo pipeline for enhanced 3D reconstruction. Both the illumination conditions during acquisition and the subsequent per-pixel feature transform can be jointly optimized in a differentiable fashion. Our framework automatically adapts to and makes efficient use of the geometric information available in different forms of input data. High-quality 3D reconstructions of a variety of challenging objects are demonstrated on the data captured with an illumination multiplexing device, as well as a point light. Our results compare favorably with state-of-the-art techniques.

PDF ICCV Semantic Scholar

Cite

Text

Kang et al. "Learning Efficient Photometric Feature Transform for Multi-View Stereo." International Conference on Computer Vision, 2021. doi:10.1109/ICCV48922.2021.00590

Markdown

[Kang et al. "Learning Efficient Photometric Feature Transform for Multi-View Stereo." International Conference on Computer Vision, 2021.](https://mlanthology.org/iccv/2021/kang2021iccv-learning/) doi:10.1109/ICCV48922.2021.00590

BibTeX

@inproceedings{kang2021iccv-learning,
  title     = {{Learning Efficient Photometric Feature Transform for Multi-View Stereo}},
  author    = {Kang, Kaizhang and Xie, Cihui and Zhu, Ruisheng and Ma, Xiaohe and Tan, Ping and Wu, Hongzhi and Zhou, Kun},
  booktitle = {International Conference on Computer Vision},
  year      = {2021},
  pages     = {5956-5965},
  doi       = {10.1109/ICCV48922.2021.00590},
  url       = {https://mlanthology.org/iccv/2021/kang2021iccv-learning/}
}