Skip to content
conftrace_
2015 NeurIPS NeurIPS 2015

Planar Ultrametrics for Image Segmentation

Abstract

We study the problem of hierarchical clustering on planar graphs. We formulate this in terms of finding the closest ultrametric to a specified set of distances and solve it using an LP relaxation that leverages minimum cost perfect matching as a subroutine to efficiently explore the space of planar partitions. We apply our algorithm to the problem of hierarchical image segmentation.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Computer Vision and Data Science & Analytics and Mathematics & Optimization
🐣 Hot Topic Early Bird - hierarchical clustering

Authors

Julian E Yarkony, Charless Fowlkes

Topics

Machine Learning > Core Methods > Clustering Computer Vision > Processing > Image Segmentation Data Science & Analytics > Applications > Clustering Mathematics & Optimization > Optimization > Linear Programming

Keywords

image segmentation graph matching hierarchical clustering linear programming relaxation planar graph minimum cost perfect matching planar ultrametric hierarchical image segmentation
Download PDF

More from NeurIPS 2015

All NeurIPS 2015 papers →
On-the-Job Learning with Bayesian Decision Theory
Spatial Transformer Networks
Precision-Recall-Gain Curves: PR Analysis Done Right
Sparse Local Embeddings for Extreme Multi-label Classification
Super-Resolution Off the Grid
Automatic Variational Inference in Stan