MAP disparity estimation using hidden markov trees

Eric T. Psota, Jedrzej Kowalczuk, Mateusz Mittek, Lance C Perez

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

25 Citations (Scopus)

Abstract

A new method is introduced for stereo matching that operates on minimum spanning trees (MSTs) generated from the images. Disparity maps are represented as a collection of hidden states on MSTs, and each MST is modeled as a hidden Markov tree. An efficient recursive message-passing scheme designed to operate on hidden Markov trees, known as the upward-downward algorithm, is used to compute the maximum a posteriori (MAP) disparity estimate at each pixel. The messages processed by the upward-downward algorithm involve two types of probabilities: the probability of a pixel having a particular disparity given a set of per-pixel matching costs, and the probability of a disparity transition between a pair of connected pixels given their similarity. The distributions of these probabilities are modeled from a collection of images with ground truth disparities. Performance evaluation using the Middlebury stereo benchmark version 3 demonstrates that the proposed method ranks second and third in terms of overall accuracy when evaluated on the training and test image sets, respectively.

Original languageEnglish (US)
Title of host publicationProceedings - 2015 IEEE International Conference on Computer Vision, ICCV 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages2219-2227
Number of pages9
Volume11-18-December-2015
ISBN (Electronic)9781467383912
DOIs
StatePublished - Feb 17 2016
Event15th IEEE International Conference on Computer Vision, ICCV 2015 - Santiago, Chile
Duration: Dec 11 2015Dec 18 2015

Other

Other15th IEEE International Conference on Computer Vision, ICCV 2015
CountryChile
CitySantiago
Period12/11/1512/18/15

Fingerprint

Pixels
Message passing
Costs

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition

Cite this

Psota, E. T., Kowalczuk, J., Mittek, M., & Perez, L. C. (2016). MAP disparity estimation using hidden markov trees. In Proceedings - 2015 IEEE International Conference on Computer Vision, ICCV 2015 (Vol. 11-18-December-2015, pp. 2219-2227). [7410613] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICCV.2015.256

MAP disparity estimation using hidden markov trees. / Psota, Eric T.; Kowalczuk, Jedrzej; Mittek, Mateusz; Perez, Lance C.

Proceedings - 2015 IEEE International Conference on Computer Vision, ICCV 2015. Vol. 11-18-December-2015 Institute of Electrical and Electronics Engineers Inc., 2016. p. 2219-2227 7410613.

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

Psota, ET, Kowalczuk, J, Mittek, M & Perez, LC 2016, MAP disparity estimation using hidden markov trees. in Proceedings - 2015 IEEE International Conference on Computer Vision, ICCV 2015. vol. 11-18-December-2015, 7410613, Institute of Electrical and Electronics Engineers Inc., pp. 2219-2227, 15th IEEE International Conference on Computer Vision, ICCV 2015, Santiago, Chile, 12/11/15. https://doi.org/10.1109/ICCV.2015.256
Psota ET, Kowalczuk J, Mittek M, Perez LC. MAP disparity estimation using hidden markov trees. In Proceedings - 2015 IEEE International Conference on Computer Vision, ICCV 2015. Vol. 11-18-December-2015. Institute of Electrical and Electronics Engineers Inc. 2016. p. 2219-2227. 7410613 https://doi.org/10.1109/ICCV.2015.256
Psota, Eric T. ; Kowalczuk, Jedrzej ; Mittek, Mateusz ; Perez, Lance C. / MAP disparity estimation using hidden markov trees. Proceedings - 2015 IEEE International Conference on Computer Vision, ICCV 2015. Vol. 11-18-December-2015 Institute of Electrical and Electronics Engineers Inc., 2016. pp. 2219-2227
@inproceedings{938a7107f5d4491d814c2fe16af4a5d0,
title = "MAP disparity estimation using hidden markov trees",
abstract = "A new method is introduced for stereo matching that operates on minimum spanning trees (MSTs) generated from the images. Disparity maps are represented as a collection of hidden states on MSTs, and each MST is modeled as a hidden Markov tree. An efficient recursive message-passing scheme designed to operate on hidden Markov trees, known as the upward-downward algorithm, is used to compute the maximum a posteriori (MAP) disparity estimate at each pixel. The messages processed by the upward-downward algorithm involve two types of probabilities: the probability of a pixel having a particular disparity given a set of per-pixel matching costs, and the probability of a disparity transition between a pair of connected pixels given their similarity. The distributions of these probabilities are modeled from a collection of images with ground truth disparities. Performance evaluation using the Middlebury stereo benchmark version 3 demonstrates that the proposed method ranks second and third in terms of overall accuracy when evaluated on the training and test image sets, respectively.",
author = "Psota, {Eric T.} and Jedrzej Kowalczuk and Mateusz Mittek and Perez, {Lance C}",
year = "2016",
month = "2",
day = "17",
doi = "10.1109/ICCV.2015.256",
language = "English (US)",
volume = "11-18-December-2015",
pages = "2219--2227",
booktitle = "Proceedings - 2015 IEEE International Conference on Computer Vision, ICCV 2015",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - MAP disparity estimation using hidden markov trees

AU - Psota, Eric T.

AU - Kowalczuk, Jedrzej

AU - Mittek, Mateusz

AU - Perez, Lance C

PY - 2016/2/17

Y1 - 2016/2/17

N2 - A new method is introduced for stereo matching that operates on minimum spanning trees (MSTs) generated from the images. Disparity maps are represented as a collection of hidden states on MSTs, and each MST is modeled as a hidden Markov tree. An efficient recursive message-passing scheme designed to operate on hidden Markov trees, known as the upward-downward algorithm, is used to compute the maximum a posteriori (MAP) disparity estimate at each pixel. The messages processed by the upward-downward algorithm involve two types of probabilities: the probability of a pixel having a particular disparity given a set of per-pixel matching costs, and the probability of a disparity transition between a pair of connected pixels given their similarity. The distributions of these probabilities are modeled from a collection of images with ground truth disparities. Performance evaluation using the Middlebury stereo benchmark version 3 demonstrates that the proposed method ranks second and third in terms of overall accuracy when evaluated on the training and test image sets, respectively.

AB - A new method is introduced for stereo matching that operates on minimum spanning trees (MSTs) generated from the images. Disparity maps are represented as a collection of hidden states on MSTs, and each MST is modeled as a hidden Markov tree. An efficient recursive message-passing scheme designed to operate on hidden Markov trees, known as the upward-downward algorithm, is used to compute the maximum a posteriori (MAP) disparity estimate at each pixel. The messages processed by the upward-downward algorithm involve two types of probabilities: the probability of a pixel having a particular disparity given a set of per-pixel matching costs, and the probability of a disparity transition between a pair of connected pixels given their similarity. The distributions of these probabilities are modeled from a collection of images with ground truth disparities. Performance evaluation using the Middlebury stereo benchmark version 3 demonstrates that the proposed method ranks second and third in terms of overall accuracy when evaluated on the training and test image sets, respectively.

UR - http://www.scopus.com/inward/record.url?scp=84973860966&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84973860966&partnerID=8YFLogxK

U2 - 10.1109/ICCV.2015.256

DO - 10.1109/ICCV.2015.256

M3 - Conference contribution

VL - 11-18-December-2015

SP - 2219

EP - 2227

BT - Proceedings - 2015 IEEE International Conference on Computer Vision, ICCV 2015

PB - Institute of Electrical and Electronics Engineers Inc.

ER -