3D Room Layout Recovery Generalizing Across Manhattan and Non-Manhattan Worlds

Haijing Jia, Hong Yi, Hirochika Fujiki, Hengzhi Zhang, Wei Wang, Makoto Odamaki

CVPRW 2022 pp. 5188-5197

doi:10.1109/CVPRW56347.2022.00567 /cvprw/2022/jia2022cvprw-3d/

Abstract

Recent 3D room layout recovery approaches mostly concentrate on Manhattan layouts, where the vertical walls are orthogonal with respect to each other, even though there are many rooms with non-Manhattan layouts in the real world. This paper presents a room layout recovery method generalizing across Manhattan and non-Manhattan worlds. Without introducing additional supervision, we extend current Manhattan layout recovery methods by predicting an extra surface normal feature, which is further used for an adaptive post-processing to reconstruct layouts of arbitrary shapes. Experimental results show that our method has a great improvement on non-Manhattan layouts while being capable of generalizing across Manhattan and non-Manhattan layouts.

CVPRW Semantic Scholar

Cite

Text

Jia et al. "3D Room Layout Recovery Generalizing Across Manhattan and Non-Manhattan Worlds." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022. doi:10.1109/CVPRW56347.2022.00567

Markdown

[Jia et al. "3D Room Layout Recovery Generalizing Across Manhattan and Non-Manhattan Worlds." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022.](https://mlanthology.org/cvprw/2022/jia2022cvprw-3d/) doi:10.1109/CVPRW56347.2022.00567

BibTeX

@inproceedings{jia2022cvprw-3d,
  title     = {{3D Room Layout Recovery Generalizing Across Manhattan and Non-Manhattan Worlds}},
  author    = {Jia, Haijing and Yi, Hong and Fujiki, Hirochika and Zhang, Hengzhi and Wang, Wei and Odamaki, Makoto},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
  year      = {2022},
  pages     = {5188-5197},
  doi       = {10.1109/CVPRW56347.2022.00567},
  url       = {https://mlanthology.org/cvprw/2022/jia2022cvprw-3d/}
}