ML Anthology

Multibody Grouping via Orthogonal Subspace Decomposition

Ying Wu, Zhengyou Zhang, Thomas S. Huang, John Y. Lin
CVPR 2001 pp. II:252-257
doi:10.1109/CVPR.2001.990968 /cvpr/2001/wu2001cvpr-multibody/

Abstract

Multibody structure from motion could be solved by the factorization approach. However, the noise measurements would make the segmentation difficult when analyzing the shape interaction matrix. This paper presents an orthogonal subspace decomposition and grouping technique to approach such a problem. We decompose the object shape spaces into signal subspaces and noise subspaces. We show that the signal subspaces of the object shape spaces are orthogonal to each other. Instead of using the shape interaction matrix contaminated by noise, we introduce the shape signal subspace distance matrix for shape space grouping. Outliers could be easily identified by this approach. The robustness of the proposed approach lies in the fact that the shape space decomposition alleviates the influence of noise, and has been verified with extensive experiments.

CVPR Semantic Scholar

Cite

Text

Wu et al. "Multibody Grouping via Orthogonal Subspace Decomposition." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2001. doi:10.1109/CVPR.2001.990968

Markdown

[Wu et al. "Multibody Grouping via Orthogonal Subspace Decomposition." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2001.](https://mlanthology.org/cvpr/2001/wu2001cvpr-multibody/) doi:10.1109/CVPR.2001.990968

BibTeX

@inproceedings{wu2001cvpr-multibody,
  title     = {{Multibody Grouping via Orthogonal Subspace Decomposition}},
  author    = {Wu, Ying and Zhang, Zhengyou and Huang, Thomas S. and Lin, John Y.},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2001},
  pages     = {II:252-257},
  doi       = {10.1109/CVPR.2001.990968},
  url       = {https://mlanthology.org/cvpr/2001/wu2001cvpr-multibody/}
}