A novel general kernel-based non-negative matrix factorisation approach for face recognition

Kernel-based non-negative matrix factorisation (KNMF) is a promising nonlinear approach for image data representation using non-negative features. However, most of the KNMF algorithms are developed via a specific kernel function and thus fail to adopt other kinds of kernels. Also, they have to learn...

Full description

Saved in:
Bibliographic Details
Main Authors: Wen-Sheng Chen, Xiya Ge, Binbin Pan
Format: Article
Language:English
Published: Taylor & Francis Group 2022-12-01
Series:Connection Science
Subjects:
Online Access:http://dx.doi.org/10.1080/09540091.2021.1988904
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849402000089808896
author Wen-Sheng Chen
Xiya Ge
Binbin Pan
author_facet Wen-Sheng Chen
Xiya Ge
Binbin Pan
author_sort Wen-Sheng Chen
collection DOAJ
description Kernel-based non-negative matrix factorisation (KNMF) is a promising nonlinear approach for image data representation using non-negative features. However, most of the KNMF algorithms are developed via a specific kernel function and thus fail to adopt other kinds of kernels. Also, they have to learn pre-image inaccurately that may influence the reliability of the method. To address these problems of KNMF, this paper proposes a novel general kernel-based non-negative matrix factorisation (GKBNNMF) method. It not only avoids pre-image learning but also is suitable for any kernel functions as well. We assume that the mapped basis images fall within the cone spanned by the mapped training data, allowing us to use arbitrary kernel function in the algorithm. The symmetric NMF strategy is exploited on kernel matrix to establish our general kernel NMF model. The proposed algorithm is proven to be convergent. The facial image datasets are selected to evaluate the performance of our method. Compared with some state-of-the-art approaches, the experimental results demonstrate that our proposed is both effective and robust.
format Article
id doaj-art-8bdf69c36ada4a9caae229f5e3cae278
institution Kabale University
issn 0954-0091
1360-0494
language English
publishDate 2022-12-01
publisher Taylor & Francis Group
record_format Article
series Connection Science
spelling doaj-art-8bdf69c36ada4a9caae229f5e3cae2782025-08-20T03:37:38ZengTaylor & Francis GroupConnection Science0954-00911360-04942022-12-0134178581010.1080/09540091.2021.19889041988904A novel general kernel-based non-negative matrix factorisation approach for face recognitionWen-Sheng Chen0Xiya Ge1Binbin Pan2Shenzhen UniversityShenzhen UniversityShenzhen UniversityKernel-based non-negative matrix factorisation (KNMF) is a promising nonlinear approach for image data representation using non-negative features. However, most of the KNMF algorithms are developed via a specific kernel function and thus fail to adopt other kinds of kernels. Also, they have to learn pre-image inaccurately that may influence the reliability of the method. To address these problems of KNMF, this paper proposes a novel general kernel-based non-negative matrix factorisation (GKBNNMF) method. It not only avoids pre-image learning but also is suitable for any kernel functions as well. We assume that the mapped basis images fall within the cone spanned by the mapped training data, allowing us to use arbitrary kernel function in the algorithm. The symmetric NMF strategy is exploited on kernel matrix to establish our general kernel NMF model. The proposed algorithm is proven to be convergent. The facial image datasets are selected to evaluate the performance of our method. Compared with some state-of-the-art approaches, the experimental results demonstrate that our proposed is both effective and robust.http://dx.doi.org/10.1080/09540091.2021.1988904non-negative matrix factorisationkernel functionface recognition
spellingShingle Wen-Sheng Chen
Xiya Ge
Binbin Pan
A novel general kernel-based non-negative matrix factorisation approach for face recognition
Connection Science
non-negative matrix factorisation
kernel function
face recognition
title A novel general kernel-based non-negative matrix factorisation approach for face recognition
title_full A novel general kernel-based non-negative matrix factorisation approach for face recognition
title_fullStr A novel general kernel-based non-negative matrix factorisation approach for face recognition
title_full_unstemmed A novel general kernel-based non-negative matrix factorisation approach for face recognition
title_short A novel general kernel-based non-negative matrix factorisation approach for face recognition
title_sort novel general kernel based non negative matrix factorisation approach for face recognition
topic non-negative matrix factorisation
kernel function
face recognition
url http://dx.doi.org/10.1080/09540091.2021.1988904
work_keys_str_mv AT wenshengchen anovelgeneralkernelbasednonnegativematrixfactorisationapproachforfacerecognition
AT xiyage anovelgeneralkernelbasednonnegativematrixfactorisationapproachforfacerecognition
AT binbinpan anovelgeneralkernelbasednonnegativematrixfactorisationapproachforfacerecognition
AT wenshengchen novelgeneralkernelbasednonnegativematrixfactorisationapproachforfacerecognition
AT xiyage novelgeneralkernelbasednonnegativematrixfactorisationapproachforfacerecognition
AT binbinpan novelgeneralkernelbasednonnegativematrixfactorisationapproachforfacerecognition