English

Spatial Transformer K-Means

Machine Learning 2022-02-17 v1 Computer Vision and Pattern Recognition

Abstract

K-means defines one of the most employed centroid-based clustering algorithms with performances tied to the data's embedding. Intricate data embeddings have been designed to push KK-means performances at the cost of reduced theoretical guarantees and interpretability of the results. Instead, we propose preserving the intrinsic data space and augment K-means with a similarity measure invariant to non-rigid transformations. This enables (i) the reduction of intrinsic nuisances associated with the data, reducing the complexity of the clustering task and increasing performances and producing state-of-the-art results, (ii) clustering in the input space of the data, leading to a fully interpretable clustering algorithm, and (iii) the benefit of convergence guarantees.

Keywords

Cite

@article{arxiv.2202.07829,
  title  = {Spatial Transformer K-Means},
  author = {Romain Cosentino and Randall Balestriero and Yanis Bahroun and Anirvan Sengupta and Richard Baraniuk and Behnaam Aazhang},
  journal= {arXiv preprint arXiv:2202.07829},
  year   = {2022}
}

Comments

arXiv admin note: substantial text overlap with arXiv:2012.09743

R2 v1 2026-06-24T09:40:10.948Z