Model-based pose estimation by consensus

Anne Jorstad, Philippe Burlina, I. Jeng Wang, Dennis Lucarelli, Daniel DeMenthon

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

Abstract

We present a system for determining a consensus estimate of the pose of an object, as seen from multiple cameras in a distributed network. The cameras are pointed towards a 3D object defined by a configuration of points, which are assumed to be visible and detected in all camera images. The cameras are given a model defining the 3D configuration of these object points, but do not know which image point corresponds to which object point. Each camera estimates the pose of the object, then iteratively exchanges information with its neighbors to arrive at a common decision of the pose over the network. We consider eight variations of the consensus algorithm, and find that each converges to a more accurate result than do the individual cameras alone on average. The method exchanging 3D world coordinates penalized to agree with the input model provides the most accurate results. If bandwidth is limited, performing consensus over rotations and translations requires cameras to exchange only the six values specifying the six degrees of freedom of the object pose, and performing consensus in SE(3) using the Karcher mean is generally the best choice. We show further that interleaving pose calculation with the consensus iterations improves the final result when the image noise is large.

Original languageEnglish (US)
Title of host publicationISSNIP 2008 - Proceedings of the 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing
Pages569-574
Number of pages6
DOIs
StatePublished - Dec 1 2008
Externally publishedYes
Event2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing, ISSNIP 2008 - Sydney, NSW, Australia
Duration: Dec 15 2008Dec 18 2008

Other

Other2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing, ISSNIP 2008
CountryAustralia
CitySydney, NSW
Period12/15/0812/18/08

Fingerprint

Cameras
Bandwidth

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems
  • Electrical and Electronic Engineering

Cite this

Jorstad, A., Burlina, P., Wang, I. J., Lucarelli, D., & DeMenthon, D. (2008). Model-based pose estimation by consensus. In ISSNIP 2008 - Proceedings of the 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing (pp. 569-574). [4762050] https://doi.org/10.1109/ISSNIP.2008.4762050

Model-based pose estimation by consensus. / Jorstad, Anne; Burlina, Philippe; Wang, I. Jeng; Lucarelli, Dennis; DeMenthon, Daniel.

ISSNIP 2008 - Proceedings of the 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing. 2008. p. 569-574 4762050.

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

Jorstad, A, Burlina, P, Wang, IJ, Lucarelli, D & DeMenthon, D 2008, Model-based pose estimation by consensus. in ISSNIP 2008 - Proceedings of the 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing., 4762050, pp. 569-574, 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing, ISSNIP 2008, Sydney, NSW, Australia, 12/15/08. https://doi.org/10.1109/ISSNIP.2008.4762050
Jorstad A, Burlina P, Wang IJ, Lucarelli D, DeMenthon D. Model-based pose estimation by consensus. In ISSNIP 2008 - Proceedings of the 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing. 2008. p. 569-574. 4762050 https://doi.org/10.1109/ISSNIP.2008.4762050
Jorstad, Anne ; Burlina, Philippe ; Wang, I. Jeng ; Lucarelli, Dennis ; DeMenthon, Daniel. / Model-based pose estimation by consensus. ISSNIP 2008 - Proceedings of the 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing. 2008. pp. 569-574
@inproceedings{cbcb9e9ebf5b4462a54b78b4d05b0d32,
title = "Model-based pose estimation by consensus",
abstract = "We present a system for determining a consensus estimate of the pose of an object, as seen from multiple cameras in a distributed network. The cameras are pointed towards a 3D object defined by a configuration of points, which are assumed to be visible and detected in all camera images. The cameras are given a model defining the 3D configuration of these object points, but do not know which image point corresponds to which object point. Each camera estimates the pose of the object, then iteratively exchanges information with its neighbors to arrive at a common decision of the pose over the network. We consider eight variations of the consensus algorithm, and find that each converges to a more accurate result than do the individual cameras alone on average. The method exchanging 3D world coordinates penalized to agree with the input model provides the most accurate results. If bandwidth is limited, performing consensus over rotations and translations requires cameras to exchange only the six values specifying the six degrees of freedom of the object pose, and performing consensus in SE(3) using the Karcher mean is generally the best choice. We show further that interleaving pose calculation with the consensus iterations improves the final result when the image noise is large.",
author = "Anne Jorstad and Philippe Burlina and Wang, {I. Jeng} and Dennis Lucarelli and Daniel DeMenthon",
year = "2008",
month = "12",
day = "1",
doi = "10.1109/ISSNIP.2008.4762050",
language = "English (US)",
isbn = "9781424429578",
pages = "569--574",
booktitle = "ISSNIP 2008 - Proceedings of the 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing",

}

TY - GEN

T1 - Model-based pose estimation by consensus

AU - Jorstad, Anne

AU - Burlina, Philippe

AU - Wang, I. Jeng

AU - Lucarelli, Dennis

AU - DeMenthon, Daniel

PY - 2008/12/1

Y1 - 2008/12/1

N2 - We present a system for determining a consensus estimate of the pose of an object, as seen from multiple cameras in a distributed network. The cameras are pointed towards a 3D object defined by a configuration of points, which are assumed to be visible and detected in all camera images. The cameras are given a model defining the 3D configuration of these object points, but do not know which image point corresponds to which object point. Each camera estimates the pose of the object, then iteratively exchanges information with its neighbors to arrive at a common decision of the pose over the network. We consider eight variations of the consensus algorithm, and find that each converges to a more accurate result than do the individual cameras alone on average. The method exchanging 3D world coordinates penalized to agree with the input model provides the most accurate results. If bandwidth is limited, performing consensus over rotations and translations requires cameras to exchange only the six values specifying the six degrees of freedom of the object pose, and performing consensus in SE(3) using the Karcher mean is generally the best choice. We show further that interleaving pose calculation with the consensus iterations improves the final result when the image noise is large.

AB - We present a system for determining a consensus estimate of the pose of an object, as seen from multiple cameras in a distributed network. The cameras are pointed towards a 3D object defined by a configuration of points, which are assumed to be visible and detected in all camera images. The cameras are given a model defining the 3D configuration of these object points, but do not know which image point corresponds to which object point. Each camera estimates the pose of the object, then iteratively exchanges information with its neighbors to arrive at a common decision of the pose over the network. We consider eight variations of the consensus algorithm, and find that each converges to a more accurate result than do the individual cameras alone on average. The method exchanging 3D world coordinates penalized to agree with the input model provides the most accurate results. If bandwidth is limited, performing consensus over rotations and translations requires cameras to exchange only the six values specifying the six degrees of freedom of the object pose, and performing consensus in SE(3) using the Karcher mean is generally the best choice. We show further that interleaving pose calculation with the consensus iterations improves the final result when the image noise is large.

UR - http://www.scopus.com/inward/record.url?scp=63149109540&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=63149109540&partnerID=8YFLogxK

U2 - 10.1109/ISSNIP.2008.4762050

DO - 10.1109/ISSNIP.2008.4762050

M3 - Conference contribution

AN - SCOPUS:63149109540

SN - 9781424429578

SP - 569

EP - 574

BT - ISSNIP 2008 - Proceedings of the 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing

ER -