A closed form solution to robust subspace estimation and clustering

Authors:
P. Favaro;R. Vidal;A. Ravichandran
Affiliations:
-;-;Univ. of California, Los Angeles, CA, USA
Venue:
CVPR '11 Proceedings of the 2011 IEEE Conference on Computer Vision and Pattern Recognition
Year:
2011

Citing 0
Cited 10

A general framework for subspace detection in unordered multidimensional data

Pattern Recognition
A fast tri-factorization method for low-rank matrix recovery and completion

Pattern Recognition
An efficient matrix factorization based low-rank representation for subspace clustering

Pattern Recognition
Multi-scale clustering of frame-to-frame correspondences for motion segmentation

ECCV'12 Proceedings of the 12th European conference on Computer Vision - Volume Part II
Towards feature-based situation assessment for airport apron video surveillance

Proceedings of the 15th international conference on Theoretical Foundations of Computer Vision: outdoor and large-scale real-world scene analysis
Analyzing the subspace structure of related images: concurrent segmentation of image sets

ECCV'12 Proceedings of the 12th European conference on Computer Vision - Volume Part IV
Low-rank representation based discriminative projection for robust feature extraction

Neurocomputing
Learning dictionary from signals under global sparsity constraint

Neurocomputing
An efficient matrix bi-factorization alternative optimization method for low-rank matrix recovery and completion

Neural Networks
Robust subspace discovery via relaxed rank minimization

Neural Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

We consider the problem of fitting one or more subspaces to a collection of data points drawn from the subspaces and corrupted by noise/outliers. We pose this problem as a rank minimization problem, where the goal is to decompose the corrupted data matrix as the sum of a clean, self-expressive, low-rank dictionary plus a matrix of noise/outliers. Our key contribution is to show that, for noisy data, this non-convex problem can be solved very efficiently and in closed form from the SVD of the noisy data matrix. Remarkably, this is true for both one or more subspaces. An important difference with respect to existing methods is that our framework results in a polynomial thresholding of the singular values with minimal shrinkage. Indeed, a particular case of our framework in the case of a single subspace leads to classical PCA, which requires no shrinkage. In the case of multiple subspaces, our framework provides an affinity matrix that can be used to cluster the data according to the sub-spaces. In the case of data corrupted by outliers, a closed-form solution appears elusive. We thus use an augmented Lagrangian optimization framework, which requires a combination of our proposed polynomial thresholding operator with the more traditional shrinkage-thresholding operator.