One-pass View-unaligned Clustering

Wenhua Dong, Xiaojun Wu, Zhenhua Feng, Sara Atito Ali Ahmed, Muhammad Awais, Josef Kittler

Abstract

Given a set of multi-view instances, the prevailing assumption in most existing clustering approaches is that they are complete and exhibit cross-view alignment. However, this assumption is often unrealistic. In such scenarios, it could be satisfied at the cost of data pre-processing, but this would be complex and inconsistent with practical applications. Therefore, developing more effective solutions for the View-unaligned Problem (VuP) is highly desirable. Several pioneering works have tackled the partially VuP, yet handling fully VuP remains a challenge due to the reliance on partially pre-aligned instances. In this paper, we propose One-pass View-unaligned Clustering (OpVuC) that simultaneously aligns and clusters instances in a unified framework. Specifically, we align shuffled instances with a selected template using an innovative global-local alignment scheme based on the notion of geometric invariance and separate the fully aligned instances using a relaxed $k$ -means algorithm. The proposed OpVuC method can handle VuP at any alignment level without requiring any pre-aligned instances. Extensive experiments conducted on several benchmark datasets demonstrate the effectiveness and merits of the proposed OpVuC method.
BibTeX
@article{dong2024one,
  author = {Wenhua Dong and Xiaojun Wu and Zhenhua Feng and Sara Atito Ali Ahmed and Muhammad Awais and Josef Kittler},
  title = {{One-pass View-unaligned Clustering}},
  journal = {IEEE Trans. Multimedia},
  year = {2024},
  url = {https://openreview.net/forum?id=pHU0Xv8HRP},
}

Loading related content...

Figure