Semi-global matching: A principled derivation in terms of message passing

Amnon Drory, Carsten Haubold*, Shai Avidan, Fred A. Hamprecht

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

35 Scopus citations

Abstract

Semi-global matching, originally introduced in the context of dense stereo, is a very successful heuristic to minimize the energy of a pairwise multi-label Markov Random Field defined on a grid. We offer the first principled explanation of this empirically successful algorithm, and clarify its exact relation to belief propagation and tree-reweighted message passing. One outcome of this new connection is an uncertainty measure for the MAP label of a variable in a Markov Random Field.

Original languageEnglish
Title of host publicationPattern Recognition - 36th German Conference, GCPR 2014, Proceedings
EditorsJoachim Hornegger, Xiaoyi Jiang, Joachim Hornegger, Joachim Hornegger, Reinhard Koch
PublisherSpringer Verlag
Pages43-53
Number of pages11
ISBN (Electronic)9783319117515
DOIs
StatePublished - 2014
Event36th German Conference on Pattern Recognition, GCPR 2014 - Münster, Germany
Duration: 2 Sep 20145 Sep 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8753
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference36th German Conference on Pattern Recognition, GCPR 2014
Country/TerritoryGermany
CityMünster
Period2/09/145/09/14

Funding

FundersFunder number
HGS MathComp Graduate School
Google

    Fingerprint

    Dive into the research topics of 'Semi-global matching: A principled derivation in terms of message passing'. Together they form a unique fingerprint.

    Cite this