Bootstrap Deep Spectral Clustering with Optimal Transport

Spectral clustering is a leading clustering method. Two of its major shortcomings are the disjoint optimization process and the limited representation capacity. To address these issues, we propose a deep spectral clustering model (named BootSC), which jointly learns all stages of spectral clustering...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on multimedia pp. 1 - 14
Main Authors: Guo, Wengang, Ye, Wei, Chen, Chunchun, Sun, Xin, Bohm, Christian, Plant, Claudia, Rahardja, Susanto
Format: Journal Article
Language:English
Published: IEEE 2025
Subjects:
Clustering algorithms
Deep clustering
Pattern recognition
Spectral clustering
Unsupervised learning
ISSN:1520-9210, 1941-0077
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Spectral clustering is a leading clustering method. Two of its major shortcomings are the disjoint optimization process and the limited representation capacity. To address these issues, we propose a deep spectral clustering model (named BootSC), which jointly learns all stages of spectral clustering-affinity matrix construction, spectral embedding, and <inline-formula><tex-math notation="LaTeX">k</tex-math></inline-formula>-means clustering-using a single network in an end-to-end manner. BootSC leverages effective and efficient optimal-transport-derived supervision to bootstrap the affinity matrix and the cluster assignment matrix. Moreover, a semantically-consistent orthogonal re-parameterization technique is introduced to orthogonalize spectral embeddings, significantly enhancing the discrimination capability. Experimental results indicate that BootSC achieves state-of-the-art clustering performance. For example, it accomplishes a notable 16% NMI improvement over the runner-up method on the challenging ImageNet-Dogs dataset. Our code is available at https://github.com/spdj2271/BootSC .
ISSN:1520-9210
1941-0077
DOI:10.1109/TMM.2025.3623492