Image labeling based on graphical models using wasserstein messages and geometric assignment

• Verfasst von: Hühnerbein, Ruben [VerfasserIn]
• Verfasst von: Savarino, Fabrizio [VerfasserIn]
• Verfasst von: Åström, Freddie [VerfasserIn]
• Verfasst von: Schnörr, Christoph [VerfasserIn]
• Titel: Image labeling based on graphical models using wasserstein messages and geometric assignment
• Verf.angabe: Ruben Hühnerbein, Fabrizio Savarino, Freddie Åström, and Christoph Schnörr
• E-Jahr: 2018
• Jahr: May 24, 2018
• Umfang: 46 S.
• Fussnoten: Gesehen am 29.08.2018
• Titel Quelle: Enthalten in: Society for Industrial and Applied MathematicsSIAM journal on imaging sciences
• Ort Quelle: Philadelphia, Pa. : SIAM, 2008
• Jahr Quelle: 2018
• Band/Heft Quelle: 11(2018), 2, Seite 1317-1362
• ISSN Quelle: 1936-4954
• DOI: doi:10.1137/17M1150669
• Datenträger: Online-Ressource
• Sprache: eng
• K10plus-PPN: 1580522300

Abstract

We introduce a novel approach to Maximum A Posteriori (MAP) inference based on discrete graphical models. By utilizing local Wasserstein distances for coupling assignment measures across edges of the underlying graph, a given discrete objective function is smoothly approximated and restricted to the assignment manifold. A corresponding multiplicative update scheme combines in a single process (i) geometric integration of the resulting Riemannian gradient flow, and (ii) rounding to integral solutions that represent valid labelings. Throughout this process, local marginalization constraints known from the established LP relaxation are satisfied, whereas the smooth geometric setting results in rapidly converging iterations that can be carried out in parallel for every edge.