Bayesian structure from motion

D. A. Forsyth, S. Ioffe, J. Haddon

Research output: Contribution to conferencePaper

Abstract

We formulate structure from motion as a Bayesian inference problem, and use a Markov chain Monte Carlo sampler to sample the posterior on this problem. This results in a method that can identify both small and large tracker errors, and yields reconstructions that are stable in the presence of these errors. Furthermore, the method gives detailed information on the range of ambiguities in structure given a particular dataset, and requires no special geometric formulation to cope with degenerate situations. Motion segmentation is obtained by a layer of discrete variables associating a point with an object. We demonstrate a sampler that successfully samples an approximation to the marginal on this domain, producing a relatively unambiguous segmentation.

Original languageEnglish (US)
Pages660-665
Number of pages6
StatePublished - Dec 1 1999
EventProceedings of the 1999 7th IEEE International Conference on Computer Vision (ICCV'99) - Kerkyra, Greece
Duration: Sep 20 1999Sep 27 1999

Other

OtherProceedings of the 1999 7th IEEE International Conference on Computer Vision (ICCV'99)
CityKerkyra, Greece
Period9/20/999/27/99

Fingerprint

Markov processes

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition

Cite this

Forsyth, D. A., Ioffe, S., & Haddon, J. (1999). Bayesian structure from motion. 660-665. Paper presented at Proceedings of the 1999 7th IEEE International Conference on Computer Vision (ICCV'99), Kerkyra, Greece, .

Bayesian structure from motion. / Forsyth, D. A.; Ioffe, S.; Haddon, J.

1999. 660-665 Paper presented at Proceedings of the 1999 7th IEEE International Conference on Computer Vision (ICCV'99), Kerkyra, Greece, .

Research output: Contribution to conferencePaper

Forsyth, DA, Ioffe, S & Haddon, J 1999, 'Bayesian structure from motion', Paper presented at Proceedings of the 1999 7th IEEE International Conference on Computer Vision (ICCV'99), Kerkyra, Greece, 9/20/99 - 9/27/99 pp. 660-665.
Forsyth DA, Ioffe S, Haddon J. Bayesian structure from motion. 1999. Paper presented at Proceedings of the 1999 7th IEEE International Conference on Computer Vision (ICCV'99), Kerkyra, Greece, .
Forsyth, D. A. ; Ioffe, S. ; Haddon, J. / Bayesian structure from motion. Paper presented at Proceedings of the 1999 7th IEEE International Conference on Computer Vision (ICCV'99), Kerkyra, Greece, .6 p.
@conference{d454c452a4d142c2b4880a6eb98f255b,
title = "Bayesian structure from motion",
abstract = "We formulate structure from motion as a Bayesian inference problem, and use a Markov chain Monte Carlo sampler to sample the posterior on this problem. This results in a method that can identify both small and large tracker errors, and yields reconstructions that are stable in the presence of these errors. Furthermore, the method gives detailed information on the range of ambiguities in structure given a particular dataset, and requires no special geometric formulation to cope with degenerate situations. Motion segmentation is obtained by a layer of discrete variables associating a point with an object. We demonstrate a sampler that successfully samples an approximation to the marginal on this domain, producing a relatively unambiguous segmentation.",
author = "Forsyth, {D. A.} and S. Ioffe and J. Haddon",
year = "1999",
month = "12",
day = "1",
language = "English (US)",
pages = "660--665",
note = "Proceedings of the 1999 7th IEEE International Conference on Computer Vision (ICCV'99) ; Conference date: 20-09-1999 Through 27-09-1999",

}

TY - CONF

T1 - Bayesian structure from motion

AU - Forsyth, D. A.

AU - Ioffe, S.

AU - Haddon, J.

PY - 1999/12/1

Y1 - 1999/12/1

N2 - We formulate structure from motion as a Bayesian inference problem, and use a Markov chain Monte Carlo sampler to sample the posterior on this problem. This results in a method that can identify both small and large tracker errors, and yields reconstructions that are stable in the presence of these errors. Furthermore, the method gives detailed information on the range of ambiguities in structure given a particular dataset, and requires no special geometric formulation to cope with degenerate situations. Motion segmentation is obtained by a layer of discrete variables associating a point with an object. We demonstrate a sampler that successfully samples an approximation to the marginal on this domain, producing a relatively unambiguous segmentation.

AB - We formulate structure from motion as a Bayesian inference problem, and use a Markov chain Monte Carlo sampler to sample the posterior on this problem. This results in a method that can identify both small and large tracker errors, and yields reconstructions that are stable in the presence of these errors. Furthermore, the method gives detailed information on the range of ambiguities in structure given a particular dataset, and requires no special geometric formulation to cope with degenerate situations. Motion segmentation is obtained by a layer of discrete variables associating a point with an object. We demonstrate a sampler that successfully samples an approximation to the marginal on this domain, producing a relatively unambiguous segmentation.

UR - http://www.scopus.com/inward/record.url?scp=0033284918&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033284918&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:0033284918

SP - 660

EP - 665

ER -