Block-Diagonal Constrained Low-Rank and Sparse Graph for Discriminant Analysis of Image Data
Recently, low-rank and sparse model-based dimensionality reduction (DR) methods have aroused lots of interest. In this paper, we propose an effective supervised DR technique named block-diagonal constrained low-rank and sparse-based embedding (BLSE). BLSE has two steps, i.e., block-diagonal constrai...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2017-06-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | http://www.mdpi.com/1424-8220/17/7/1475 |
| _version_ | 1798039755115462656 |
|---|---|
| author | Tan Guo Xiaoheng Tan Lei Zhang Chaochen Xie Lu Deng |
| author_facet | Tan Guo Xiaoheng Tan Lei Zhang Chaochen Xie Lu Deng |
| author_sort | Tan Guo |
| collection | DOAJ |
| description | Recently, low-rank and sparse model-based dimensionality reduction (DR) methods have aroused lots of interest. In this paper, we propose an effective supervised DR technique named block-diagonal constrained low-rank and sparse-based embedding (BLSE). BLSE has two steps, i.e., block-diagonal constrained low-rank and sparse representation (BLSR) and block-diagonal constrained low-rank and sparse graph embedding (BLSGE). Firstly, the BLSR model is developed to reveal the intrinsic intra-class and inter-class adjacent relationships as well as the local neighborhood relations and global structure of data. Particularly, there are mainly three items considered in BLSR. First, a sparse constraint is required to discover the local data structure. Second, a low-rank criterion is incorporated to capture the global structure in data. Third, a block-diagonal regularization is imposed on the representation to promote discrimination between different classes. Based on BLSR, informative and discriminative intra-class and inter-class graphs are constructed. With the graphs, BLSGE seeks a low-dimensional embedding subspace by simultaneously minimizing the intra-class scatter and maximizing the inter-class scatter. Experiments on public benchmark face and object image datasets demonstrate the effectiveness of the proposed approach. |
| first_indexed | 2024-04-11T21:58:07Z |
| format | Article |
| id | doaj.art-f4b6ed7ad81747539f6b74e720e5101a |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-11T21:58:07Z |
| publishDate | 2017-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-f4b6ed7ad81747539f6b74e720e5101a2022-12-22T04:01:03ZengMDPI AGSensors1424-82202017-06-01177147510.3390/s17071475s17071475Block-Diagonal Constrained Low-Rank and Sparse Graph for Discriminant Analysis of Image DataTan Guo0Xiaoheng Tan1Lei Zhang2Chaochen Xie3Lu Deng4College of Communication Engineering, Chongqing University, Chongqing 400044, ChinaCollege of Communication Engineering, Chongqing University, Chongqing 400044, ChinaCollege of Communication Engineering, Chongqing University, Chongqing 400044, ChinaCollege of Communication Engineering, Chongqing University, Chongqing 400044, ChinaCollege of Communication Engineering, Chongqing University, Chongqing 400044, ChinaRecently, low-rank and sparse model-based dimensionality reduction (DR) methods have aroused lots of interest. In this paper, we propose an effective supervised DR technique named block-diagonal constrained low-rank and sparse-based embedding (BLSE). BLSE has two steps, i.e., block-diagonal constrained low-rank and sparse representation (BLSR) and block-diagonal constrained low-rank and sparse graph embedding (BLSGE). Firstly, the BLSR model is developed to reveal the intrinsic intra-class and inter-class adjacent relationships as well as the local neighborhood relations and global structure of data. Particularly, there are mainly three items considered in BLSR. First, a sparse constraint is required to discover the local data structure. Second, a low-rank criterion is incorporated to capture the global structure in data. Third, a block-diagonal regularization is imposed on the representation to promote discrimination between different classes. Based on BLSR, informative and discriminative intra-class and inter-class graphs are constructed. With the graphs, BLSGE seeks a low-dimensional embedding subspace by simultaneously minimizing the intra-class scatter and maximizing the inter-class scatter. Experiments on public benchmark face and object image datasets demonstrate the effectiveness of the proposed approach.http://www.mdpi.com/1424-8220/17/7/1475dimensionality reductionlow-rank representationsparse representationblock-diagonalimage classification |
| spellingShingle | Tan Guo Xiaoheng Tan Lei Zhang Chaochen Xie Lu Deng Block-Diagonal Constrained Low-Rank and Sparse Graph for Discriminant Analysis of Image Data Sensors dimensionality reduction low-rank representation sparse representation block-diagonal image classification |
| title | Block-Diagonal Constrained Low-Rank and Sparse Graph for Discriminant Analysis of Image Data |
| title_full | Block-Diagonal Constrained Low-Rank and Sparse Graph for Discriminant Analysis of Image Data |
| title_fullStr | Block-Diagonal Constrained Low-Rank and Sparse Graph for Discriminant Analysis of Image Data |
| title_full_unstemmed | Block-Diagonal Constrained Low-Rank and Sparse Graph for Discriminant Analysis of Image Data |
| title_short | Block-Diagonal Constrained Low-Rank and Sparse Graph for Discriminant Analysis of Image Data |
| title_sort | block diagonal constrained low rank and sparse graph for discriminant analysis of image data |
| topic | dimensionality reduction low-rank representation sparse representation block-diagonal image classification |
| url | http://www.mdpi.com/1424-8220/17/7/1475 |
| work_keys_str_mv | AT tanguo blockdiagonalconstrainedlowrankandsparsegraphfordiscriminantanalysisofimagedata AT xiaohengtan blockdiagonalconstrainedlowrankandsparsegraphfordiscriminantanalysisofimagedata AT leizhang blockdiagonalconstrainedlowrankandsparsegraphfordiscriminantanalysisofimagedata AT chaochenxie blockdiagonalconstrainedlowrankandsparsegraphfordiscriminantanalysisofimagedata AT ludeng blockdiagonalconstrainedlowrankandsparsegraphfordiscriminantanalysisofimagedata |