Noisy Dual Principal Component Pursuit

Tianyu Ding, Zhihui Zhu, Tianjiao Ding, Yunchen Yang, Rene Vidal, Manolis Tsakiris, Daniel Robinson
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:1617-1625, 2019.

Abstract

Dual Principal Component Pursuit (DPCP) is a recently proposed non-convex optimization based method for learning subspaces of high relative dimension from noiseless datasets contaminated by as many outliers as the square of the number of inliers. Experimentally, DPCP has proved to be robust to noise and outperform the popular RANSAC on 3D vision tasks such as road plane detection and relative poses estimation from three views. This paper extends the global optimality and convergence theory of DPCP to the case of data corrupted by noise, and further demonstrates its robustness using synthetic and real data.

Cite this Paper


BibTeX
@InProceedings{pmlr-v97-ding19b, title = {Noisy Dual Principal Component Pursuit}, author = {Ding, Tianyu and Zhu, Zhihui and Ding, Tianjiao and Yang, Yunchen and Vidal, Rene and Tsakiris, Manolis and Robinson, Daniel}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {1617--1625}, year = {2019}, editor = {Kamalika Chaudhuri and Ruslan Salakhutdinov}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/ding19b/ding19b.pdf}, url = { http://proceedings.mlr.press/v97/ding19b.html }, abstract = {Dual Principal Component Pursuit (DPCP) is a recently proposed non-convex optimization based method for learning subspaces of high relative dimension from noiseless datasets contaminated by as many outliers as the square of the number of inliers. Experimentally, DPCP has proved to be robust to noise and outperform the popular RANSAC on 3D vision tasks such as road plane detection and relative poses estimation from three views. This paper extends the global optimality and convergence theory of DPCP to the case of data corrupted by noise, and further demonstrates its robustness using synthetic and real data.} }
Endnote
%0 Conference Paper %T Noisy Dual Principal Component Pursuit %A Tianyu Ding %A Zhihui Zhu %A Tianjiao Ding %A Yunchen Yang %A Rene Vidal %A Manolis Tsakiris %A Daniel Robinson %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-ding19b %I PMLR %P 1617--1625 %U http://proceedings.mlr.press/v97/ding19b.html %V 97 %X Dual Principal Component Pursuit (DPCP) is a recently proposed non-convex optimization based method for learning subspaces of high relative dimension from noiseless datasets contaminated by as many outliers as the square of the number of inliers. Experimentally, DPCP has proved to be robust to noise and outperform the popular RANSAC on 3D vision tasks such as road plane detection and relative poses estimation from three views. This paper extends the global optimality and convergence theory of DPCP to the case of data corrupted by noise, and further demonstrates its robustness using synthetic and real data.
APA
Ding, T., Zhu, Z., Ding, T., Yang, Y., Vidal, R., Tsakiris, M. & Robinson, D.. (2019). Noisy Dual Principal Component Pursuit. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:1617-1625 Available from http://proceedings.mlr.press/v97/ding19b.html .

Related Material