Regularized kernel discriminant analysis with a robust kernel for face recognition and verification

We propose a robust approach to discriminant kernel-based feature extraction for face recognition and verification. We show, for the first time, how to perform the eigen analysis of the within-class scatter matrix directly in the feature space. This eigen analysis provides the eigenspectrum of its r...

Full description

Bibliographic Details
Main Authors: Zafeiriou, Stefanos, Tzimiropoulos, Georgios, Petrou, Maria, Stathaki, Tania
Format: Article
Published: Institute of Electrical and Electronics Engineers 2012
Online Access:https://eprints.nottingham.ac.uk/31423/
_version_ 1848794198060302336
author Zafeiriou, Stefanos
Tzimiropoulos, Georgios
Petrou, Maria
Stathaki, Tania
author_facet Zafeiriou, Stefanos
Tzimiropoulos, Georgios
Petrou, Maria
Stathaki, Tania
author_sort Zafeiriou, Stefanos
building Nottingham Research Data Repository
collection Online Access
description We propose a robust approach to discriminant kernel-based feature extraction for face recognition and verification. We show, for the first time, how to perform the eigen analysis of the within-class scatter matrix directly in the feature space. This eigen analysis provides the eigenspectrum of its range space and the corresponding eigenvectors as well as the eigenvectors spanning its null space. Based on our analysis, we propose a kernel discriminant analysis (KDA) which combines eigenspectrum regularization with a feature-level scheme (ER-KDA). Finally, we combine the proposed ER-KDA with a nonlinear robust kernel particularly suitable for face recognition/verification applications which require robustness against outliers caused by occlusions and illumination changes. We applied the proposed framework to several popular databases (Yale, AR, XM2VTS) and achieved state-of-the-art performance for most of our experiments.
first_indexed 2025-11-14T19:12:22Z
format Article
id nottingham-31423
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:12:22Z
publishDate 2012
publisher Institute of Electrical and Electronics Engineers
recordtype eprints
repository_type Digital Repository
spelling nottingham-314232020-05-04T20:21:51Z https://eprints.nottingham.ac.uk/31423/ Regularized kernel discriminant analysis with a robust kernel for face recognition and verification Zafeiriou, Stefanos Tzimiropoulos, Georgios Petrou, Maria Stathaki, Tania We propose a robust approach to discriminant kernel-based feature extraction for face recognition and verification. We show, for the first time, how to perform the eigen analysis of the within-class scatter matrix directly in the feature space. This eigen analysis provides the eigenspectrum of its range space and the corresponding eigenvectors as well as the eigenvectors spanning its null space. Based on our analysis, we propose a kernel discriminant analysis (KDA) which combines eigenspectrum regularization with a feature-level scheme (ER-KDA). Finally, we combine the proposed ER-KDA with a nonlinear robust kernel particularly suitable for face recognition/verification applications which require robustness against outliers caused by occlusions and illumination changes. We applied the proposed framework to several popular databases (Yale, AR, XM2VTS) and achieved state-of-the-art performance for most of our experiments. Institute of Electrical and Electronics Engineers 2012-03 Article PeerReviewed Zafeiriou, Stefanos, Tzimiropoulos, Georgios, Petrou, Maria and Stathaki, Tania (2012) Regularized kernel discriminant analysis with a robust kernel for face recognition and verification. IEEE Transactions on Neural Networks and Learning Systems, 23 (3). pp. 526-534. ISSN 2162-237X http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6129513 doi:10.1109/TNNLS.2011.2182058 doi:10.1109/TNNLS.2011.2182058
spellingShingle Zafeiriou, Stefanos
Tzimiropoulos, Georgios
Petrou, Maria
Stathaki, Tania
Regularized kernel discriminant analysis with a robust kernel for face recognition and verification
title Regularized kernel discriminant analysis with a robust kernel for face recognition and verification
title_full Regularized kernel discriminant analysis with a robust kernel for face recognition and verification
title_fullStr Regularized kernel discriminant analysis with a robust kernel for face recognition and verification
title_full_unstemmed Regularized kernel discriminant analysis with a robust kernel for face recognition and verification
title_short Regularized kernel discriminant analysis with a robust kernel for face recognition and verification
title_sort regularized kernel discriminant analysis with a robust kernel for face recognition and verification
url https://eprints.nottingham.ac.uk/31423/
https://eprints.nottingham.ac.uk/31423/
https://eprints.nottingham.ac.uk/31423/