Multi-view low-rank sparse subspace clustering

PATTERN RECOGNITION

卷:73

Article

Brbic, Maria¹  Kopriva, Ivica¹ 

[1] Rudjer Boskovic Inst, Lab Machine Learning & Knowledge Representat, Div Elect, Bijenicka Cesta 54, Zagreb 10000, Croatia

关键词: Subspace clustering;    Multi-view data;    Low-rank;    Sparsity;    Alternating direction method of multipliers;    Reproducing kernel Hilbert space;   

DOI  :  10.1016/j.patcog.2017.08.024

来源: Elsevier

【 摘 要 】

Most existing approaches address multi-view subspace clustering problem by constructing the affinity matrix on each view separately and afterwards propose how to extend spectral clustering algorithm to handle multi-view data. This paper presents an approach to multi-view subspace clustering that learns a joint subspace representation by constructing affinity matrix shared among all views. Relying on the importance of both low-rank and sparsity constraints in the construction of the affinity matrix, we introduce the objective that balances between the agreement across different views, while at the same time encourages sparsity and low-rankness of the solution. Related low-rank and sparsity constrained optimization problem, is for each view solved using the alternating direction method of multipliers. Furthermore, we extend our approach to cluster data drawn from nonlinear subspaces by solving the corresponding problem in a reproducing kernel Hilbert space. The proposed algorithm outperforms state-of-the-art niulti-view subspace clustering algorithms on one synthetic and four real-world datasets. (C) 2017 Elsevier Ltd. All rights reserved.

【 授权许可】

Free