google / quickshift

A clustering algorithm that first finds the high-density regions (cluster-cores) of the data and then clusters the remaining points by hill-climbing. Such seedings act as more stable and expressive cluster-cores than the singleton modes found by popular algorithm such as mean shift. (https://arxiv.org/abs/1805.07909)
68Updated 6 years ago

Alternatives and similar repositories for quickshift:

Users that are interested in quickshift are comparing it to the libraries listed below