MAP inference with MRF by graduated non-convexity and concavity procedure

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

8 Scopus citations

Abstract

In this paper we generalize the recently proposed graduated non-convexity and concavity procedure (GNCCP) to approximately solve the maximum a posteriori (MAP) inference problem with the Markov random field (MRF). Unlike the commonly used graph cuts or loopy brief propagation, the GNCCP based MAP algorithm is widely applicable to any types of graphical models with any types of potentials, and is very easy to use in practice. Our preliminary experimental comparisons witness its state-of-the-art performance.

Original languageEnglish
Title of host publicationNeural Information Processing - 21st International Conference, ICONIP 2014, Proceedings
EditorsChu Kiong Loo, Keem Siah Yap, Kok Wai Wong, Andrew Teoh, Kaizhu Huang
PublisherSpringer Verlag
Pages404-412
Number of pages9
ISBN (Electronic)9783319126395
DOIs
StatePublished - 2014
Externally publishedYes
Event21st International Conference on Neural Information Processing, ICONIP 2014 - Kuching, Malaysia
Duration: 3 Nov 20146 Nov 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8835
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference21st International Conference on Neural Information Processing, ICONIP 2014
Country/TerritoryMalaysia
CityKuching
Period3/11/146/11/14

Keywords

  • Energy maximization
  • GNCCP
  • Markov random field

Fingerprint

Dive into the research topics of 'MAP inference with MRF by graduated non-convexity and concavity procedure'. Together they form a unique fingerprint.

Cite this