English

Quickshift++: Provably Good Initializations for Sample-Based Mean Shift

Machine Learning 2018-05-22 v1 Machine Learning

Abstract

We provide initial seedings to the Quick Shift clustering algorithm, which approximate the locally high-density regions of the data. Such seedings act as more stable and expressive cluster-cores than the singleton modes found by Quick Shift. We establish statistical consistency guarantees for this modification. We then show strong clustering performance on real datasets as well as promising applications to image segmentation.

Keywords

Cite

@article{arxiv.1805.07909,
  title  = {Quickshift++: Provably Good Initializations for Sample-Based Mean Shift},
  author = {Heinrich Jiang and Jennifer Jang and Samory Kpotufe},
  journal= {arXiv preprint arXiv:1805.07909},
  year   = {2018}
}

Comments

ICML 2018. Code release: https://github.com/google/quickshift

R2 v1 2026-06-23T02:02:18.283Z