Wenyan Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Wenyan Wang has a total of 3 co-authors. Here are the top co-authors by number of shared publications:
- Jianping Chen
School of Computer Science Guangzhou Maritime University Guangzhou China
1 shared publication - Jie Xu
School of Computer Science Guangzhou Maritime University Guangzhou China
1 shared publication - Sicong Gong
Faculty of Automation, Guangdong University of Technology, No. 100, West Outer Ring Road, Guangzhou Higher Education Mega Center, Panyu District, Guangzhou 510006, P. R. China
1 shared publication
Latest Publications
- 22 November 2024
- Published by World Scientific Pub Co Pte Ltd in International Journal of Wavelets, Multiresolution and Information Processing
- Vol. 23 (02)
Local multi-scale structure discriminant analysis
The Local Multi-Scale Discriminant Analysis (LMSDA) is a novel dimensionality reduction technique that proficiently maps high-dimensional data into a low-dimensional space. It seamlessly integrates the collaborative representation framework with dual constraint optimization in intrinsic and penalty graphs. The intrinsic graph captures intricate local manifold structures by leveraging multi-scale reconstruction relationships, ensuring class cohesion during reduction. Meanwhile, the penalty graph sharpens discriminability by maximizing inter-class disparities, fostering clearer class boundaries. This dual strategy not only preserves the essence of intra-class data but also bolsters its discriminative powers. Extensive experiments on AR, CMU_PIE, and Yale B datasets affirm LMSDA’s superiority over contemporary methods, underscoring its potential for diverse applications.
- 22 November 2024
- Published by World Scientific Pub Co Pte Ltd in International Journal of Wavelets, Multiresolution and Information Processing
- Vol. 23 (02)
Local multi-scale structure discriminant analysis
The Local Multi-Scale Discriminant Analysis (LMSDA) is a novel dimensionality reduction technique that proficiently maps high-dimensional data into a low-dimensional space. It seamlessly integrates the collaborative representation framework with dual constraint optimization in intrinsic and penalty graphs. The intrinsic graph captures intricate local manifold structures by leveraging multi-scale reconstruction relationships, ensuring class cohesion during reduction. Meanwhile, the penalty graph sharpens discriminability by maximizing inter-class disparities, fostering clearer class boundaries. This dual strategy not only preserves the essence of intra-class data but also bolsters its discriminative powers. Extensive experiments on AR, CMU_PIE, and Yale B datasets affirm LMSDA’s superiority over contemporary methods, underscoring its potential for diverse applications.