Directed Acyclic Graph Continuous Max-Flow Image Segmentation for Unconstrained Label Orderings

Authors

John S.H. Baxter, Robarts Research Institute
Martin Rajchl, Imperial College London
A. Jonathan McLeod, Robarts Research Institute
Jing Yuan, Xidian University
Terry M. Peters, Robarts Research InstituteFollow

Document Type

Article

Publication Date

7-1-2017

Journal

International Journal of Computer Vision

Volume

123

Issue

3

First Page

415

Last Page

434

URL with Digital Object Identifier

10.1007/s11263-017-0994-x

Abstract

© 2017, Springer Science+Business Media New York. Label ordering, the specification of subset–superset relationships for segmentation labels, has been of increasing interest in image segmentation as they allow for complex regions to be represented as a collection of simple parts. Recent advances in continuous max-flow segmentation have widely expanded the possible label orderings from binary background/foreground problems to extendable frameworks in which the label ordering can be specified. This article presents Directed Acyclic Graph Max-Flow image segmentation which is flexible enough to incorporate any label ordering without constraints. This framework uses augmented Lagrangian multipliers and primal–dual optimization to develop a highly parallelized solver implemented using GPGPU. This framework is validated on synthetic, natural, and medical images illustrating its general applicability.