Convex Histogram-Based Joint Image Segmentation with Regularized Optimal Transport Cost
This work addresses image segmentation challenges in computer vision, offering a versatile method that is incremental as it builds on existing optimal transport theory.
The authors tackled the problem of multiple image segmentation by developing a convex framework that uses regularized optimal transport to match feature histograms between image regions and reference models, achieving efficient solutions with primal-dual algorithms for both multi-phase segmentation and co-segmentation.
We investigate in this work a versatile convex framework for multiple image segmentation, relying on the regularized optimal mass transport theory. In this setting, several transport cost functions are considered and used to match statistical distributions of features. In practice, global multidimensional histograms are estimated from the segmented image regions, and are compared to referring models that are either fixed histograms given a priori, or directly inferred in the non-supervised case. The different convex problems studied are solved efficiently using primal-dual algorithms. The proposed approach is generic and enables multi-phase segmentation as well as co-segmentation of multiple images.