Multilevel spectral partitioning for efficient image segmentation and tracking

David Tolliver; Robert T. Collins; Simon Baker

doi:10.1109/ACVMOT.2005.83

Multilevel spectral partitioning for efficient image segmentation and tracking

David Tolliver, Robert T. Collins, Simon Baker

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

4 Scopus citations

Abstract

An efficient multilevel method for solving normalized cut image segmentation problems is presented. The method uses the lattice geometry of images to define a set of coarsened graph partitioning problems. This problem hierarchy provides a framework for rapidly estimating the eigenvectors of normalized graph Laplacians. Within this framework, a coarse solution obtained with a standard eigensolver is propagated to increasingly fine problem instances and refined using subspace iterations. Results are presented for image segmentation and tracking problems. The computational cost of the multilevel method is an order of magnitude lower than current sampling techniques and results in more stable image segmentations.

Original language	English (US)
Title of host publication	Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005
Pages	414-420
Number of pages	7
DOIs	https://doi.org/10.1109/ACVMOT.2005.83
State	Published - 2007
Event	7th IEEE Workshop on Applications of Computer Vision, WACV 2005 - Breckenridge, CO, United States Duration: Jan 5 2005 → Jan 7 2005

Publication series

Name	Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005

Other

Other	7th IEEE Workshop on Applications of Computer Vision, WACV 2005
Country/Territory	United States
City	Breckenridge, CO
Period	1/5/05 → 1/7/05

All Science Journal Classification (ASJC) codes

Computer Vision and Pattern Recognition
Software

Access to Document

10.1109/ACVMOT.2005.83

Cite this

@inproceedings{b5a8ab4c1e0740b3bf36467b366503bd,

title = "Multilevel spectral partitioning for efficient image segmentation and tracking",

abstract = "An efficient multilevel method for solving normalized cut image segmentation problems is presented. The method uses the lattice geometry of images to define a set of coarsened graph partitioning problems. This problem hierarchy provides a framework for rapidly estimating the eigenvectors of normalized graph Laplacians. Within this framework, a coarse solution obtained with a standard eigensolver is propagated to increasingly fine problem instances and refined using subspace iterations. Results are presented for image segmentation and tracking problems. The computational cost of the multilevel method is an order of magnitude lower than current sampling techniques and results in more stable image segmentations.",

author = "David Tolliver and Collins, {Robert T.} and Simon Baker",

year = "2007",

doi = "10.1109/ACVMOT.2005.83",

language = "English (US)",

isbn = "0769522718",

series = "Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005",

pages = "414--420",

booktitle = "Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005",

note = "7th IEEE Workshop on Applications of Computer Vision, WACV 2005 ; Conference date: 05-01-2005 Through 07-01-2005",

}

Tolliver, D, Collins, RT & Baker, S 2007, Multilevel spectral partitioning for efficient image segmentation and tracking. in Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005., 4129511, Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005, pp. 414-420, 7th IEEE Workshop on Applications of Computer Vision, WACV 2005, Breckenridge, CO, United States, 1/5/05. https://doi.org/10.1109/ACVMOT.2005.83

Multilevel spectral partitioning for efficient image segmentation and tracking. / Tolliver, David; Collins, Robert T.; Baker, Simon.
Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005. 2007. p. 414-420 4129511 (Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Multilevel spectral partitioning for efficient image segmentation and tracking

AU - Tolliver, David

AU - Collins, Robert T.

AU - Baker, Simon

PY - 2007

Y1 - 2007

N2 - An efficient multilevel method for solving normalized cut image segmentation problems is presented. The method uses the lattice geometry of images to define a set of coarsened graph partitioning problems. This problem hierarchy provides a framework for rapidly estimating the eigenvectors of normalized graph Laplacians. Within this framework, a coarse solution obtained with a standard eigensolver is propagated to increasingly fine problem instances and refined using subspace iterations. Results are presented for image segmentation and tracking problems. The computational cost of the multilevel method is an order of magnitude lower than current sampling techniques and results in more stable image segmentations.

AB - An efficient multilevel method for solving normalized cut image segmentation problems is presented. The method uses the lattice geometry of images to define a set of coarsened graph partitioning problems. This problem hierarchy provides a framework for rapidly estimating the eigenvectors of normalized graph Laplacians. Within this framework, a coarse solution obtained with a standard eigensolver is propagated to increasingly fine problem instances and refined using subspace iterations. Results are presented for image segmentation and tracking problems. The computational cost of the multilevel method is an order of magnitude lower than current sampling techniques and results in more stable image segmentations.

UR - http://www.scopus.com/inward/record.url?scp=35348839580&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35348839580&partnerID=8YFLogxK

U2 - 10.1109/ACVMOT.2005.83

DO - 10.1109/ACVMOT.2005.83

M3 - Conference contribution

AN - SCOPUS:35348839580

SN - 0769522718

SN - 9780769522715

T3 - Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005

SP - 414

EP - 420

BT - Proceedings - Seventh IEEE Workshop on Applications of Computer Vision, WACV 2005

T2 - 7th IEEE Workshop on Applications of Computer Vision, WACV 2005

Y2 - 5 January 2005 through 7 January 2005

ER -

Multilevel spectral partitioning for efficient image segmentation and tracking

Abstract

Publication series

Other

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this