Higher-order clique reduction in binary graph cut

Research output: Chapter in Book/Report/Conference proceedingConference contribution

93 Citations (Scopus)

Abstract

We introduce a new technique that can reduce any higher-order Markov random field with binary labels into a first-order one that has the same minima as the original. Moreover, we combine the reduction with the fusion-move and QPBO algorithms to optimize higher-order multi-label problems. While many vision problems today are formulated as energy minimization problems, they have mostly been limited to using first-order energies, which consist of unary and pairwise clique potentials, with a few exceptions that consider triples. This is because of the lack of efficient algorithms to optimize energies with higher-order interactions. Our algorithm challenges this restriction that limits the representational power of the models, so that higherorder energies can be used to capture the rich statistics of natural scenes. To demonstrate the algorithm, we minimize a third-order energy, which allows clique potentials with up to four pixels, in an image restoration problem. The problem uses the Fields of Experts model, a learned spatial prior of natural images that has been used to test two belief propagation algorithms capable of optimizing higher-order energies. The results show that the algorithm exceeds the BP algorithms in both optimization performance and speed.

Original languageEnglish
Title of host publication2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009
Pages2993-3000
Number of pages8
DOIs
Publication statusPublished - 2009
Externally publishedYes
Event2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009 - Miami, FL
Duration: 2009 Jun 202009 Jun 25

Other

Other2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009
CityMiami, FL
Period09/6/2009/6/25

Fingerprint

Labels
Image reconstruction
Fusion reactions
Pixels
Statistics

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Biomedical Engineering

Cite this

Ishikawa, H. (2009). Higher-order clique reduction in binary graph cut. In 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009 (pp. 2993-3000). [5206689] https://doi.org/10.1109/CVPRW.2009.5206689

Higher-order clique reduction in binary graph cut. / Ishikawa, Hiroshi.

2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009. 2009. p. 2993-3000 5206689.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ishikawa, H 2009, Higher-order clique reduction in binary graph cut. in 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009., 5206689, pp. 2993-3000, 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009, Miami, FL, 09/6/20. https://doi.org/10.1109/CVPRW.2009.5206689
Ishikawa H. Higher-order clique reduction in binary graph cut. In 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009. 2009. p. 2993-3000. 5206689 https://doi.org/10.1109/CVPRW.2009.5206689
Ishikawa, Hiroshi. / Higher-order clique reduction in binary graph cut. 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009. 2009. pp. 2993-3000
@inproceedings{4c98146efacf430a885a9206a92a246c,
title = "Higher-order clique reduction in binary graph cut",
abstract = "We introduce a new technique that can reduce any higher-order Markov random field with binary labels into a first-order one that has the same minima as the original. Moreover, we combine the reduction with the fusion-move and QPBO algorithms to optimize higher-order multi-label problems. While many vision problems today are formulated as energy minimization problems, they have mostly been limited to using first-order energies, which consist of unary and pairwise clique potentials, with a few exceptions that consider triples. This is because of the lack of efficient algorithms to optimize energies with higher-order interactions. Our algorithm challenges this restriction that limits the representational power of the models, so that higherorder energies can be used to capture the rich statistics of natural scenes. To demonstrate the algorithm, we minimize a third-order energy, which allows clique potentials with up to four pixels, in an image restoration problem. The problem uses the Fields of Experts model, a learned spatial prior of natural images that has been used to test two belief propagation algorithms capable of optimizing higher-order energies. The results show that the algorithm exceeds the BP algorithms in both optimization performance and speed.",
author = "Hiroshi Ishikawa",
year = "2009",
doi = "10.1109/CVPRW.2009.5206689",
language = "English",
isbn = "9781424439935",
pages = "2993--3000",
booktitle = "2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009",

}

TY - GEN

T1 - Higher-order clique reduction in binary graph cut

AU - Ishikawa, Hiroshi

PY - 2009

Y1 - 2009

N2 - We introduce a new technique that can reduce any higher-order Markov random field with binary labels into a first-order one that has the same minima as the original. Moreover, we combine the reduction with the fusion-move and QPBO algorithms to optimize higher-order multi-label problems. While many vision problems today are formulated as energy minimization problems, they have mostly been limited to using first-order energies, which consist of unary and pairwise clique potentials, with a few exceptions that consider triples. This is because of the lack of efficient algorithms to optimize energies with higher-order interactions. Our algorithm challenges this restriction that limits the representational power of the models, so that higherorder energies can be used to capture the rich statistics of natural scenes. To demonstrate the algorithm, we minimize a third-order energy, which allows clique potentials with up to four pixels, in an image restoration problem. The problem uses the Fields of Experts model, a learned spatial prior of natural images that has been used to test two belief propagation algorithms capable of optimizing higher-order energies. The results show that the algorithm exceeds the BP algorithms in both optimization performance and speed.

AB - We introduce a new technique that can reduce any higher-order Markov random field with binary labels into a first-order one that has the same minima as the original. Moreover, we combine the reduction with the fusion-move and QPBO algorithms to optimize higher-order multi-label problems. While many vision problems today are formulated as energy minimization problems, they have mostly been limited to using first-order energies, which consist of unary and pairwise clique potentials, with a few exceptions that consider triples. This is because of the lack of efficient algorithms to optimize energies with higher-order interactions. Our algorithm challenges this restriction that limits the representational power of the models, so that higherorder energies can be used to capture the rich statistics of natural scenes. To demonstrate the algorithm, we minimize a third-order energy, which allows clique potentials with up to four pixels, in an image restoration problem. The problem uses the Fields of Experts model, a learned spatial prior of natural images that has been used to test two belief propagation algorithms capable of optimizing higher-order energies. The results show that the algorithm exceeds the BP algorithms in both optimization performance and speed.

UR - http://www.scopus.com/inward/record.url?scp=70450161205&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70450161205&partnerID=8YFLogxK

U2 - 10.1109/CVPRW.2009.5206689

DO - 10.1109/CVPRW.2009.5206689

M3 - Conference contribution

AN - SCOPUS:70450161205

SN - 9781424439935

SP - 2993

EP - 3000

BT - 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009

ER -