Bayesian Optimization of 3D Feature Parameters for 6D Pose Estimation

Frederik Hagelskjær*, Norbert Krüger, Anders Glent Buch

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

36 Downloads (Pure)

Abstract

6D pose estimation using local features has shown state-of-the-art performance for object recognition and pose estimation from 3D data in a number of benchmarks. However, this method requires extensive knowledge and elaborate parameter tuning to obtain optimal performances. In this paper, we propose an optimization method able to determine feature parameters automatically, providing improved point matches to a robust pose estimation algorithm. Using labeled data, our method measures the performance of the current parameter setting using a scoring function based on both true and false positive detections. Combined with a Bayesian optimization strategy, we achieve automatic tuning using few labeled examples. Experiments were performed on two recent RGB-D benchmark datasets. The results show significant improvements by tuning an existing algorithm, with state-of-art performance.
Original languageEnglish
Title of host publicationProceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications
EditorsAndreas Kerren, Christophe Hurter, Jose Braz
Volume5: VISAPP
PublisherSCITEPRESS Digital Library
Publication date2019
Pages135-142
ISBN (Electronic)978-989-758-354-4
DOIs
Publication statusPublished - 2019
Event14th International Conference on Computer Vision Theory and Applications - Prague, Czech Republic
Duration: 25. Feb 201927. Feb 2019

Conference

Conference14th International Conference on Computer Vision Theory and Applications
CountryCzech Republic
CityPrague
Period25/02/201927/02/2019

Fingerprint

Tuning
Object recognition
Experiments

Keywords

  • Machine Learning Technologies for Vision
  • Shape Representation and Matching
  • Features Extraction
  • Object Detection and Localization
  • Bayesian Optimization
  • Object Detection
  • Pose Estimation
  • Machine Learning
  • Optimization
  • Feature Matching

Cite this

Hagelskjær, F., Krüger, N., & Buch, A. G. (2019). Bayesian Optimization of 3D Feature Parameters for 6D Pose Estimation. In A. Kerren, C. Hurter, & J. Braz (Eds.), Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (Vol. 5: VISAPP, pp. 135-142). SCITEPRESS Digital Library. https://doi.org/10.5220/0007568801350142
Hagelskjær, Frederik ; Krüger, Norbert ; Buch, Anders Glent. / Bayesian Optimization of 3D Feature Parameters for 6D Pose Estimation. Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. editor / Andreas Kerren ; Christophe Hurter ; Jose Braz. Vol. 5: VISAPP SCITEPRESS Digital Library, 2019. pp. 135-142
@inproceedings{834769e0764a4540a2067835ada86507,
title = "Bayesian Optimization of 3D Feature Parameters for 6D Pose Estimation",
abstract = "6D pose estimation using local features has shown state-of-the-art performance for object recognition and pose estimation from 3D data in a number of benchmarks. However, this method requires extensive knowledge and elaborate parameter tuning to obtain optimal performances. In this paper, we propose an optimization method able to determine feature parameters automatically, providing improved point matches to a robust pose estimation algorithm. Using labeled data, our method measures the performance of the current parameter setting using a scoring function based on both true and false positive detections. Combined with a Bayesian optimization strategy, we achieve automatic tuning using few labeled examples. Experiments were performed on two recent RGB-D benchmark datasets. The results show significant improvements by tuning an existing algorithm, with state-of-art performance.",
keywords = "Machine Learning Technologies for Vision, Shape Representation and Matching, Features Extraction, Object Detection and Localization, Bayesian Optimization, Object Detection, Pose Estimation, Machine Learning, Optimization, Feature Matching",
author = "Frederik Hagelskj{\ae}r and Norbert Kr{\"u}ger and Buch, {Anders Glent}",
year = "2019",
doi = "10.5220/0007568801350142",
language = "English",
volume = "5: VISAPP",
pages = "135--142",
editor = "Andreas Kerren and Christophe Hurter and Jose Braz",
booktitle = "Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications",
publisher = "SCITEPRESS Digital Library",

}

Hagelskjær, F, Krüger, N & Buch, AG 2019, Bayesian Optimization of 3D Feature Parameters for 6D Pose Estimation. in A Kerren, C Hurter & J Braz (eds), Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. vol. 5: VISAPP, SCITEPRESS Digital Library, pp. 135-142, 14th International Conference on Computer Vision Theory and Applications, Prague, Czech Republic, 25/02/2019. https://doi.org/10.5220/0007568801350142

Bayesian Optimization of 3D Feature Parameters for 6D Pose Estimation. / Hagelskjær, Frederik; Krüger, Norbert; Buch, Anders Glent.

Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. ed. / Andreas Kerren; Christophe Hurter; Jose Braz. Vol. 5: VISAPP SCITEPRESS Digital Library, 2019. p. 135-142.

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

TY - GEN

T1 - Bayesian Optimization of 3D Feature Parameters for 6D Pose Estimation

AU - Hagelskjær, Frederik

AU - Krüger, Norbert

AU - Buch, Anders Glent

PY - 2019

Y1 - 2019

N2 - 6D pose estimation using local features has shown state-of-the-art performance for object recognition and pose estimation from 3D data in a number of benchmarks. However, this method requires extensive knowledge and elaborate parameter tuning to obtain optimal performances. In this paper, we propose an optimization method able to determine feature parameters automatically, providing improved point matches to a robust pose estimation algorithm. Using labeled data, our method measures the performance of the current parameter setting using a scoring function based on both true and false positive detections. Combined with a Bayesian optimization strategy, we achieve automatic tuning using few labeled examples. Experiments were performed on two recent RGB-D benchmark datasets. The results show significant improvements by tuning an existing algorithm, with state-of-art performance.

AB - 6D pose estimation using local features has shown state-of-the-art performance for object recognition and pose estimation from 3D data in a number of benchmarks. However, this method requires extensive knowledge and elaborate parameter tuning to obtain optimal performances. In this paper, we propose an optimization method able to determine feature parameters automatically, providing improved point matches to a robust pose estimation algorithm. Using labeled data, our method measures the performance of the current parameter setting using a scoring function based on both true and false positive detections. Combined with a Bayesian optimization strategy, we achieve automatic tuning using few labeled examples. Experiments were performed on two recent RGB-D benchmark datasets. The results show significant improvements by tuning an existing algorithm, with state-of-art performance.

KW - Machine Learning Technologies for Vision

KW - Shape Representation and Matching

KW - Features Extraction

KW - Object Detection and Localization

KW - Bayesian Optimization

KW - Object Detection

KW - Pose Estimation

KW - Machine Learning

KW - Optimization

KW - Feature Matching

U2 - 10.5220/0007568801350142

DO - 10.5220/0007568801350142

M3 - Article in proceedings

VL - 5: VISAPP

SP - 135

EP - 142

BT - Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

A2 - Kerren, Andreas

A2 - Hurter, Christophe

A2 - Braz, Jose

PB - SCITEPRESS Digital Library

ER -

Hagelskjær F, Krüger N, Buch AG. Bayesian Optimization of 3D Feature Parameters for 6D Pose Estimation. In Kerren A, Hurter C, Braz J, editors, Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. Vol. 5: VISAPP. SCITEPRESS Digital Library. 2019. p. 135-142 https://doi.org/10.5220/0007568801350142