A new likelihood function for stereo matching: how to achieve invariance to unknown texture, gains and offsets?

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

1 Citation (Scopus)

Abstract

We introduce a new likelihood function for window-based stereo matching. This likelihood can cope with unknown textures, uncertain gain factors, uncertain offsets, and correlated noise. The method can be finetuned to the uncertainty ranges of the gains and offsets, rather than a full, blunt normalization as in NCC (normalized cross correlation). The likelihood is based on a sound probabilistic model. As such it can be directly used within a probabilistic framework. We demonstrate this by embedding the likelihood in a HMM (hidden Markov model) formulation of the 3D reconstruction problem, and applying this to a test scene. We compare the reconstruction results with the results when the similarity measure is the NCC, and we show that our likelihood fits better within the probabilistic frame for stereo matching than NCC.
Original languageEnglish
Title of host publicationInternational Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2009)
PublisherINSTICC PRESS
Pages603-608
Number of pages6
ISBN (Print)978-989-8111-74-6
DOIs
Publication statusPublished - Feb 2009
EventInternational Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2009 - Lisboa, Portugal
Duration: 5 Feb 20098 Feb 2009

Conference

ConferenceInternational Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2009
Abbreviated titleVISIGRAPP
CountryPortugal
CityLisboa
Period5/02/098/02/09

Fingerprint

Hidden Markov models
Invariance
Textures
Acoustic waves
Statistical Models
Uncertainty

Keywords

  • NCC
  • HMM
  • SCS-Safety
  • Probabilistic framework
  • Likelihood
  • Stereo reconstruction

Cite this

Damjanovic, S., van der Heijden, F., & Spreeuwers, L. J. (2009). A new likelihood function for stereo matching: how to achieve invariance to unknown texture, gains and offsets? In International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2009) (pp. 603-608). INSTICC PRESS. https://doi.org/10.5220/0001793606030608
Damjanovic, Sanja ; van der Heijden, Ferdinand ; Spreeuwers, Luuk J. / A new likelihood function for stereo matching: how to achieve invariance to unknown texture, gains and offsets?. International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2009). INSTICC PRESS, 2009. pp. 603-608
@inproceedings{bcc761b63ca64cdf84356bc29d88e817,
title = "A new likelihood function for stereo matching: how to achieve invariance to unknown texture, gains and offsets?",
abstract = "We introduce a new likelihood function for window-based stereo matching. This likelihood can cope with unknown textures, uncertain gain factors, uncertain offsets, and correlated noise. The method can be finetuned to the uncertainty ranges of the gains and offsets, rather than a full, blunt normalization as in NCC (normalized cross correlation). The likelihood is based on a sound probabilistic model. As such it can be directly used within a probabilistic framework. We demonstrate this by embedding the likelihood in a HMM (hidden Markov model) formulation of the 3D reconstruction problem, and applying this to a test scene. We compare the reconstruction results with the results when the similarity measure is the NCC, and we show that our likelihood fits better within the probabilistic frame for stereo matching than NCC.",
keywords = "NCC, HMM, SCS-Safety, Probabilistic framework, Likelihood, Stereo reconstruction",
author = "Sanja Damjanovic and {van der Heijden}, Ferdinand and Spreeuwers, {Luuk J.}",
year = "2009",
month = "2",
doi = "10.5220/0001793606030608",
language = "English",
isbn = "978-989-8111-74-6",
pages = "603--608",
booktitle = "International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2009)",
publisher = "INSTICC PRESS",

}

Damjanovic, S, van der Heijden, F & Spreeuwers, LJ 2009, A new likelihood function for stereo matching: how to achieve invariance to unknown texture, gains and offsets? in International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2009). INSTICC PRESS, pp. 603-608, International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2009, Lisboa, Portugal, 5/02/09. https://doi.org/10.5220/0001793606030608

A new likelihood function for stereo matching: how to achieve invariance to unknown texture, gains and offsets? / Damjanovic, Sanja; van der Heijden, Ferdinand; Spreeuwers, Luuk J.

International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2009). INSTICC PRESS, 2009. p. 603-608.

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

TY - GEN

T1 - A new likelihood function for stereo matching: how to achieve invariance to unknown texture, gains and offsets?

AU - Damjanovic, Sanja

AU - van der Heijden, Ferdinand

AU - Spreeuwers, Luuk J.

PY - 2009/2

Y1 - 2009/2

N2 - We introduce a new likelihood function for window-based stereo matching. This likelihood can cope with unknown textures, uncertain gain factors, uncertain offsets, and correlated noise. The method can be finetuned to the uncertainty ranges of the gains and offsets, rather than a full, blunt normalization as in NCC (normalized cross correlation). The likelihood is based on a sound probabilistic model. As such it can be directly used within a probabilistic framework. We demonstrate this by embedding the likelihood in a HMM (hidden Markov model) formulation of the 3D reconstruction problem, and applying this to a test scene. We compare the reconstruction results with the results when the similarity measure is the NCC, and we show that our likelihood fits better within the probabilistic frame for stereo matching than NCC.

AB - We introduce a new likelihood function for window-based stereo matching. This likelihood can cope with unknown textures, uncertain gain factors, uncertain offsets, and correlated noise. The method can be finetuned to the uncertainty ranges of the gains and offsets, rather than a full, blunt normalization as in NCC (normalized cross correlation). The likelihood is based on a sound probabilistic model. As such it can be directly used within a probabilistic framework. We demonstrate this by embedding the likelihood in a HMM (hidden Markov model) formulation of the 3D reconstruction problem, and applying this to a test scene. We compare the reconstruction results with the results when the similarity measure is the NCC, and we show that our likelihood fits better within the probabilistic frame for stereo matching than NCC.

KW - NCC

KW - HMM

KW - SCS-Safety

KW - Probabilistic framework

KW - Likelihood

KW - Stereo reconstruction

U2 - 10.5220/0001793606030608

DO - 10.5220/0001793606030608

M3 - Conference contribution

SN - 978-989-8111-74-6

SP - 603

EP - 608

BT - International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2009)

PB - INSTICC PRESS

ER -

Damjanovic S, van der Heijden F, Spreeuwers LJ. A new likelihood function for stereo matching: how to achieve invariance to unknown texture, gains and offsets? In International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2009). INSTICC PRESS. 2009. p. 603-608 https://doi.org/10.5220/0001793606030608