Learning Robust and Discriminative Subspace With Low-Rank Constraints

IEEE Transactions on Neural Networks and Learning Systems
Sheng Li, Yun Fu

Abstract

In this paper, we aim at learning robust and discriminative subspaces from noisy data. Subspace learning is widely used in extracting discriminative features for classification. However, when data are contaminated with severe noise, the performance of most existing subspace learning methods would be limited. Recent advances in low-rank modeling provide effective solutions for removing noise or outliers contained in sample sets, which motivates us to take advantage of low-rank constraints in order to exploit robust and discriminative subspace for classification. In particular, we present a discriminative subspace learning method called the supervised regularization-based robust subspace (SRRS) approach, by incorporating the low-rank constraint. SRRS seeks low-rank representations from the noisy data, and learns a discriminative subspace from the recovered clean data jointly. A supervised regularization function is designed to make use of the class label information, and therefore to enhance the discriminability of subspace. Our approach is formulated as a constrained rank-minimization problem. We design an inexact augmented Lagrange multiplier optimization algorithm to solve it. Unlike the existing sparse representation and low-...Continue Reading

References

Mar 4, 2005·IEEE Transactions on Pattern Analysis and Machine Intelligence·Xiaogang Wang, Xiaoou Tang
Mar 8, 2005·IEEE Transactions on Pattern Analysis and Machine Intelligence·Xiaofei HeHong-Jiang Zhang
May 7, 2005·IEEE Transactions on Pattern Analysis and Machine Intelligence·Kuang-Chih LeeDavid J Kriegman
Aug 5, 2006·IEEE Transactions on Pattern Analysis and Machine Intelligence·Manli Zhu, Aleix M Martinez
Dec 27, 2008·IEEE Transactions on Pattern Analysis and Machine Intelligence·John WrightYi Ma
Apr 11, 2012·IEEE Transactions on Pattern Analysis and Machine Intelligence·Guangcan LiuYi Ma
Jun 13, 2012·IEEE Transactions on Pattern Analysis and Machine Intelligence·Xiaowei ZhouWeichuan Yu
Sep 14, 2012·Neural Computation·Guangcan Liu, Shuicheng Yan
Jan 5, 2013·IEEE Transactions on Pattern Analysis and Machine Intelligence·Tingting MuSophia Ananiadou
Jan 1, 1991·Journal of Cognitive Neuroscience·M Turk, A Pentland
Mar 1, 2013·IEEE Transactions on Neural Networks and Learning Systems·Yue DengSanqing Hu

❮ Previous
Next ❯

Citations

Jan 17, 2017·IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society·Zheng ZhangGuo-Sen Xie
Feb 7, 2017·IEEE Transactions on Neural Networks and Learning Systems·Ming ShaoYun Fu
Apr 25, 2017·IEEE Transactions on Neural Networks and Learning Systems·Kailing GuoDong Xu
Aug 22, 2018·IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society·Zhengming Ding, Yun Fu
Jul 12, 2018·IEEE Transactions on Neural Networks and Learning Systems· Bo Han Sai-Fu Fung
Jul 11, 2018·IEEE Transactions on Neural Networks and Learning Systems· Zheng Zhang Jian Yang
Jun 26, 2018·Neural Networks : the Official Journal of the International Neural Network Society·Qiaolin YeMeem Naiem

❮ Previous
Next ❯

Related Concepts

Related Feeds

Cajal Bodies & Gems

Cajal bodies or coiled bodies are dense foci of coilin protein. Gemini of Cajal bodies, or gems, are microscopically similar to Cajal bodies. It is believed that Cajal bodies play important roles in RNA processing while gems assist the Cajal bodies. Find the latest research on Cajal bodies and gems here.

Related Papers

IEEE Transactions on Cybernetics
Jinhui Chen, Jian Yang
IEEE Transactions on Neural Networks and Learning Systems
Liping Jing Tieyong Zeng
IEEE Transactions on Systems, Man, and Cybernetics. Part B, Cybernetics : a Publication of the IEEE Systems, Man, and Cybernetics Society
Peng ZhangCarlotta Domeniconi
IEEE Transactions on Cybernetics
Xiaobing PeiChuanbo Chen
IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
Yun Fu, Thomas S Huang
© 2021 Meta ULC. All rights reserved