Consistency and Diversity Induced Tensorized Multi-View Subspace Clustering

Tensorizedmulti-view subspace clustering has attracted intensive attention to achieve promising clustering performance by effectively modeling both consistency and high-order correlation structure from multiple views. However, existing tensorized multi-view subspace clustering methods often neglect...

Full description

Saved in:
Bibliographic Details
Published inIEEE transactions on emerging topics in computational intelligence pp. 1 - 12
Main Authors Xiao, Chunming, Huang, Yonghui, Huang, Haonan, Zhao, Qibin, Zhou, Guoxu
Format Journal Article
LanguageEnglish
Published IEEE 24.06.2024
Subjects
Automation
Clustering methods
Consistency
Correlation
Data models
diversity
Linear programming
Matrix decomposition
multi-view subspace clustering
schatten <named-content xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:ali="http://www.niso.org/schemas/ali/1.0/" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" content-type="math" xlink:type="simple"> <inline-formula> <tex-math notation="LaTeX"> p</tex-math> </inline-formula> </named-content>-norm
tensor
Tensors
Online AccessGet full text

Cover

More Information
Summary:Tensorizedmulti-view subspace clustering has attracted intensive attention to achieve promising clustering performance by effectively modeling both consistency and high-order correlation structure from multiple views. However, existing tensorized multi-view subspace clustering methods often neglect the prevalent diversity among different views, which are specific attributes unique to each view. In this paper, we focus on simultaneously exploiting the multi-view consistency and diversity information in a tensorized multi-view subspace framework. In the diversity part, an additional position-aware exclusivity term is introduced to explore the unique features of each view to enhance feature complementarity between different views. Meanwhile, we explore inter-view similarity by minimizing the tensor Schatten p -norm, which well captures both consistency and high-order correlation information of multi-view data. Then, the objective function can be efficiently optimized by the alternating direction method of multipliers. Extensive experiments on eight benchmark datasets demonstrate that the proposed method outperforms the state-of-the-art multi-view clustering methods.
ISSN:2471-285X
2471-285X
DOI:10.1109/TETCI.2024.3412999