Enhanced Robust Tensor Multi-View Spectral Clustering via Nonconvex Low-Rank and Sparse Regularization
IEEE Signal Processing Letters · 2026 · 33 · 4306-4310
DOI: 10.1109/lsp.2026.3721053 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Yongjie Wei und 3 weiteren Autor:innen. Die Arbeit nennt 37 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Joint Guidance Project of Heilongjiang Provincial Natural Science Foundation.
Abstract (Originalsprache)
Multi-view spectral clustering aims to exploit the consistency and complementarity of multiple views to learn a reliable affinity graph. However, existing tensor-based methods often suffer from inflexible low-rank modeling, limited robustness to complex noise, and insufficient exploitation of within-view structure. To address these issues, we propose a Tensor Markov chain with Schatten- $p$ and Log-det regularization for Multi-view Spectral Clustering (TMSL-MSC). Specifically, a latent affinity tensor is constructed from diverse views, where a weighted tensor Schatten $p$ -norm is employed to characterize the global low-rank property. Meanwhile, an $\ell _{2,\log }$ -norm is introduced to suppress noise and outliers, and a log-determinant regularization term is used to retain the intrinsic geometric characteristics of each view. In the following spectral clustering process, a consistent affinity matrix is obtained, and an efficient optimization algorithm is designed to tackle the established model. Comprehensive experiments on several typical multi-view image benchmark datasets illustrate that our method attains more competitive clustering performance than existing advanced approaches, which is verified by the metrics ACC, NMI and F-score.
Förderung
- National Natural Science Foundation of China
- Joint Guidance Project of Heilongjiang Provincial Natural Science Foundation
Zitieren
APA
Wei, Y., Lan, X., Yang, M., & Kang, Z. (2026). Enhanced Robust Tensor Multi-View Spectral Clustering via Nonconvex Low-Rank and Sparse Regularization. IEEE Signal Processing Letters, 33, 4306-4310. https://doi.org/10.1109/lsp.2026.3721053
Vancouver
Wei Y, Lan X, Yang M, Kang Z. Enhanced Robust Tensor Multi-View Spectral Clustering via Nonconvex Low-Rank and Sparse Regularization. IEEE Signal Processing Letters. 2026;33:4306-4310. doi:10.1109/lsp.2026.3721053
BibTeX
@article{Wei2026,
author = {Wei, Yongjie and Lan, Xinyu and Yang, Ming and Kang, Zhao},
title = {Enhanced Robust Tensor Multi-View Spectral Clustering via Nonconvex Low-Rank and Sparse Regularization},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {4306--4310},
doi = {10.1109/lsp.2026.3721053}
}
EUVIMED