L2, 1 Regularized Correntropy for Robust Feature Selection

He, Ran; Tan, Tieniu; Wang, Liang; Zheng, Wei-Shi

doi:10.1109/CVPR.2012.6247966

L2, 1 Regularized Correntropy for Robust Feature Selection

Ran He, Tieniu Tan, Liang Wang, Wei-Shi Zheng

CVPR 2012 pp. 2504-2511

doi:10.1109/CVPR.2012.6247966 /cvpr/2012/he2012cvpr-l/

Abstract

In this paper, we study the problem of robust feature extraction based on l 2 , 1 regularized correntropy in both theoretical and algorithmic manner. In theoretical part, we point out that an l 2 , 1 -norm minimization can be justified from the viewpoint of half-quadratic (HQ) optimization, which facilitates convergence study and algorithmic development. In particular, a general formulation is accordingly proposed to unify l 1 -norm and l 2 , 1 -norm minimization within a common framework. In algorithmic part, we propose an l 2 , 1 regularized correntropy algorithm to extract informative features meanwhile to remove outliers from training data. A new alternate minimization algorithm is also developed to optimize the non-convex correntropy objective. In terms of face recognition, we apply the proposed method to obtain an appearance-based model, called Sparse-Fisherfaces. Extensive experiments show that our method can select robust and sparse features, and outperforms several state-of-the-art subspace methods on largescale and open face recognition datasets.

CVPR Semantic Scholar

Cite

Text

He et al. "L2, 1 Regularized Correntropy for Robust Feature Selection." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012. doi:10.1109/CVPR.2012.6247966

Markdown

[He et al. "L2, 1 Regularized Correntropy for Robust Feature Selection." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012.](https://mlanthology.org/cvpr/2012/he2012cvpr-l/) doi:10.1109/CVPR.2012.6247966

BibTeX

@inproceedings{he2012cvpr-l,
  title     = {{L2, 1 Regularized Correntropy for Robust Feature Selection}},
  author    = {He, Ran and Tan, Tieniu and Wang, Liang and Zheng, Wei-Shi},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2012},
  pages     = {2504-2511},
  doi       = {10.1109/CVPR.2012.6247966},
  url       = {https://mlanthology.org/cvpr/2012/he2012cvpr-l/}
}