Convex Color Image Segmentation with Optimal Transport Distances

Abstract

This work is about the use of regularized optimal-transport distances for convex, histogram-based image segmentation. In the considered framework, fixed exemplar histograms define a prior on the statistical features of the two regions in competition. In this paper, we investigate the use of various transport-based cost functions as discrepancy measures and rely on a primal-dual algorithm to solve the obtained convex optimization problem.

arXiv commentA short version of this report has been submitted to the Fifth International Conference on Scale Space and Variational Methods in Computer Vision (SSVM) 2015

References in corpus (1)

• A Smoothed Dual Approach for Variational Wasserstein Problems

Cited by in corpus (8)

• Scaling Algorithms for Unbalanced Transport Problems
• Gradient descent algorithms for Bures-Wasserstein barycenters
• Transport-based analysis, modeling, and learning from signal and data distributions
• A Smoothed Dual Approach for Variational Wasserstein Problems
• Verification in the Loop: Correct-by-Construction Control Learning with Reach-avoid Guarantees
• Relaxed Earth Mover's Distances for Chain- and Tree-connected Spaces and their use as a Loss Function in Deep Learning
• Transport based Graph Kernels
• Coupling Matrix Manifolds and Their Applications in Optimal Transport