A New Look at Clustering Large Datasets

Summary

• Descriptor counts provide useful information that can be leveraged for clustering large datasets.



• Performance gains versus N² appear feasible.



• More empirical results needed to really quantify.