Robust Ensemble Manifold Projective Non-Negative Matrix Factorization for Image Representation
Projective non-negative matrix factorization (PNMF) as a variant of NMF has received considerable attention. However, the existing PNMF methods can be further improved from two aspects. On the one hand, the square loss function that is intended to measure the reconstruction error is sensitive to noi...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9261478/ |
| _version_ | 1818934633001123840 |
|---|---|
| author | Peng Luo Xilong Qu Lina Tan Xiaoliang Xie Weijin Jiang Lirong Huang Wai Hung Ip Kai Leung Yung |
| author_facet | Peng Luo Xilong Qu Lina Tan Xiaoliang Xie Weijin Jiang Lirong Huang Wai Hung Ip Kai Leung Yung |
| author_sort | Peng Luo |
| collection | DOAJ |
| description | Projective non-negative matrix factorization (PNMF) as a variant of NMF has received considerable attention. However, the existing PNMF methods can be further improved from two aspects. On the one hand, the square loss function that is intended to measure the reconstruction error is sensitive to noise. On the other hand, it is non-trivial to estimate the intrinsic manifold of the feature space in a principal manner. So current paper is an attempt that has proposed a new method named as robust ensemble manifold projective non-negative matrix factorization (REPNMF) for image representation. Specifically, REPNMF not only assesses the influence of noise by imposing a spare noise matrix for image reconstruction, but it also assumes that the intrinsic manifold exists in a convex hull of certain pre-given manifold candidates. We aim to remove noise from the data and find the optimized combination of candidate manifolds to approximate the intrinsic manifold simultaneously. We develop iterative multiplicative updating rules for the optimization of REPNMF along with its convergence proof. The experimental results on four image datasets verify that REPNMF is superior as compare to other related state-of-the-art methods. |
| first_indexed | 2024-12-20T05:07:22Z |
| format | Article |
| id | doaj.art-e1a0dbb99ca549dcb4c0f69ebf2c8cd6 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-20T05:07:22Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-e1a0dbb99ca549dcb4c0f69ebf2c8cd62022-12-21T19:52:22ZengIEEEIEEE Access2169-35362020-01-01821778121779010.1109/ACCESS.2020.30383839261478Robust Ensemble Manifold Projective Non-Negative Matrix Factorization for Image RepresentationPeng Luo0https://orcid.org/0000-0002-7806-0850Xilong Qu1Lina Tan2Xiaoliang Xie3Weijin Jiang4Lirong Huang5Wai Hung Ip6https://orcid.org/0000-0001-6609-0713Kai Leung Yung7School of Mathematics and Statistics, Hunan University of Technology and Business, Changsha, ChinaSchool of Information Technology and Management, Hunan University of Finance and Economics, Changsha, ChinaSchool of Computer and Information Engineering, Hunan University of Technology and Business, Changsha, ChinaSchool of Mathematics and Statistics, Hunan University of Technology and Business, Changsha, ChinaSchool of Computer and Information Engineering, Hunan University of Technology and Business, Changsha, ChinaSchool of Information Technology and Management, Hunan University of Finance and Economics, Changsha, ChinaDepartment of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong KongDepartment of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong KongProjective non-negative matrix factorization (PNMF) as a variant of NMF has received considerable attention. However, the existing PNMF methods can be further improved from two aspects. On the one hand, the square loss function that is intended to measure the reconstruction error is sensitive to noise. On the other hand, it is non-trivial to estimate the intrinsic manifold of the feature space in a principal manner. So current paper is an attempt that has proposed a new method named as robust ensemble manifold projective non-negative matrix factorization (REPNMF) for image representation. Specifically, REPNMF not only assesses the influence of noise by imposing a spare noise matrix for image reconstruction, but it also assumes that the intrinsic manifold exists in a convex hull of certain pre-given manifold candidates. We aim to remove noise from the data and find the optimized combination of candidate manifolds to approximate the intrinsic manifold simultaneously. We develop iterative multiplicative updating rules for the optimization of REPNMF along with its convergence proof. The experimental results on four image datasets verify that REPNMF is superior as compare to other related state-of-the-art methods.https://ieeexplore.ieee.org/document/9261478/Non-negative matrix factorizationprojection recoveryimage representationensemble manifold learning |
| spellingShingle | Peng Luo Xilong Qu Lina Tan Xiaoliang Xie Weijin Jiang Lirong Huang Wai Hung Ip Kai Leung Yung Robust Ensemble Manifold Projective Non-Negative Matrix Factorization for Image Representation IEEE Access Non-negative matrix factorization projection recovery image representation ensemble manifold learning |
| title | Robust Ensemble Manifold Projective Non-Negative Matrix Factorization for Image Representation |
| title_full | Robust Ensemble Manifold Projective Non-Negative Matrix Factorization for Image Representation |
| title_fullStr | Robust Ensemble Manifold Projective Non-Negative Matrix Factorization for Image Representation |
| title_full_unstemmed | Robust Ensemble Manifold Projective Non-Negative Matrix Factorization for Image Representation |
| title_short | Robust Ensemble Manifold Projective Non-Negative Matrix Factorization for Image Representation |
| title_sort | robust ensemble manifold projective non negative matrix factorization for image representation |
| topic | Non-negative matrix factorization projection recovery image representation ensemble manifold learning |
| url | https://ieeexplore.ieee.org/document/9261478/ |
| work_keys_str_mv | AT pengluo robustensemblemanifoldprojectivenonnegativematrixfactorizationforimagerepresentation AT xilongqu robustensemblemanifoldprojectivenonnegativematrixfactorizationforimagerepresentation AT linatan robustensemblemanifoldprojectivenonnegativematrixfactorizationforimagerepresentation AT xiaoliangxie robustensemblemanifoldprojectivenonnegativematrixfactorizationforimagerepresentation AT weijinjiang robustensemblemanifoldprojectivenonnegativematrixfactorizationforimagerepresentation AT lironghuang robustensemblemanifoldprojectivenonnegativematrixfactorizationforimagerepresentation AT waihungip robustensemblemanifoldprojectivenonnegativematrixfactorizationforimagerepresentation AT kaileungyung robustensemblemanifoldprojectivenonnegativematrixfactorizationforimagerepresentation |