Robust Recovery of Corrupted Low-RankMatrix by Implicit Regularizers
Authors:
- 1 October 2013
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 36 (4) , 770-783
Abstract
Low-rank matrix recovery algorithms aim to recover a corrupted low-rank matrix with sparse errors. However, corrupted errors may not be sparse in real-world problems and the relationship between ℓ 1 regularizer on noise and robust M-estimators is still unknown. This paper proposes a general robust framework for low-rank matrix recovery via implicit regularizers of robust M-estimators, which are derived from convex conjugacy and can be used to model arbitrarily corrupted errors. Based on the additive form of half-quadratic optimization, proximity operators of implicit regularizers are developed such that both low-rank structure and corrupted errors can be alternately recovered. In particular, the dual relationship between the absolute function in ℓ 1 regularizer and Huber M-estimator is studied, which establishes a connection between robust low-rank matrix recovery methods and M-estimators based robust principal component analysis methods. Extensive experiments on synthetic and real-world data sets corroborate our claims and verify the robustness of the proposed framework.Keywords
This publication has 40 references indexed in Scilit:
- The generalized trace-norm and its application to structure-from-motion problems2011
- Low-rank Matrix Recovery via Iteratively Reweighted Least Squares MinimizationSIAM Journal on Optimization, 2011
- Robust Matrix Decomposition With Sparse CorruptionsIEEE Transactions on Information Theory, 2011
- Recovering Low-Rank Matrices From Few Coefficients in Any BasisIEEE Transactions on Information Theory, 2011
- Maximum Correntropy Criterion for Robust Face RecognitionIEEE Transactions on Pattern Analysis and Machine Intelligence, 2010
- The Power of Convex Relaxation: Near-Optimal Matrix CompletionIEEE Transactions on Information Theory, 2010
- Uniqueness of Low-Rank Matrix Completion by Rigidity TheorySIAM Journal on Matrix Analysis and Applications, 2010
- Sparse and low-rank matrix decompositions2009
- R1-PCA2006
- Parameter estimation techniques: a tutorial with application to conic fittingImage and Vision Computing, 1997