3D object pose estimation using viewpoint generative learning

Dissaphong Thachasongtham, Takumi Yoshida, François De Sorbier, Hideo Saito

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

11 Citations (Scopus)

Abstract

Conventional local features such as SIFT or SURF are robust to scale and rotation changes but sensitive to large perspective change. Because perspective change always occurs when 3D object moves, using these features to estimate the pose of a 3D object is a challenging task. In this paper, we extend one of our previous works on viewpoint generative learning to 3D objects. Given a model of a textured object, we virtually generate several patterns of the model from different viewpoints and select stable keypoints from those patterns. Then our system learns a collection of feature descriptors from the stable keypoints. Finally, we are able to estimate the pose of a 3D object by using these robust features. In our experimental results, we demonstrate that our system is robust against large viewpoint change and even under partial occlusion.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages512-521
Number of pages10
Volume7944 LNCS
DOIs
Publication statusPublished - 2013
Event18th Scandinavian Conference on Image Analysis, SCIA 2013 - Espoo, Finland
Duration: 2013 Jun 172013 Jun 20

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7944 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other18th Scandinavian Conference on Image Analysis, SCIA 2013
CountryFinland
CityEspoo
Period13/6/1713/6/20

Fingerprint

Pose Estimation
Scale Invariant Feature Transform
Local Features
Occlusion
Estimate
Descriptors
Learning
Object
Partial
Experimental Results
Model
Demonstrate

Keywords

  • generative learning
  • pose estimation
  • stable keypoint

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Thachasongtham, D., Yoshida, T., De Sorbier, F., & Saito, H. (2013). 3D object pose estimation using viewpoint generative learning. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7944 LNCS, pp. 512-521). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7944 LNCS). https://doi.org/10.1007/978-3-642-38886-6_48

3D object pose estimation using viewpoint generative learning. / Thachasongtham, Dissaphong; Yoshida, Takumi; De Sorbier, François; Saito, Hideo.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7944 LNCS 2013. p. 512-521 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7944 LNCS).

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

Thachasongtham, D, Yoshida, T, De Sorbier, F & Saito, H 2013, 3D object pose estimation using viewpoint generative learning. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 7944 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7944 LNCS, pp. 512-521, 18th Scandinavian Conference on Image Analysis, SCIA 2013, Espoo, Finland, 13/6/17. https://doi.org/10.1007/978-3-642-38886-6_48
Thachasongtham D, Yoshida T, De Sorbier F, Saito H. 3D object pose estimation using viewpoint generative learning. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7944 LNCS. 2013. p. 512-521. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-38886-6_48
Thachasongtham, Dissaphong ; Yoshida, Takumi ; De Sorbier, François ; Saito, Hideo. / 3D object pose estimation using viewpoint generative learning. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7944 LNCS 2013. pp. 512-521 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{decaae2196b9421fa9422f35ca6650e5,
title = "3D object pose estimation using viewpoint generative learning",
abstract = "Conventional local features such as SIFT or SURF are robust to scale and rotation changes but sensitive to large perspective change. Because perspective change always occurs when 3D object moves, using these features to estimate the pose of a 3D object is a challenging task. In this paper, we extend one of our previous works on viewpoint generative learning to 3D objects. Given a model of a textured object, we virtually generate several patterns of the model from different viewpoints and select stable keypoints from those patterns. Then our system learns a collection of feature descriptors from the stable keypoints. Finally, we are able to estimate the pose of a 3D object by using these robust features. In our experimental results, we demonstrate that our system is robust against large viewpoint change and even under partial occlusion.",
keywords = "generative learning, pose estimation, stable keypoint",
author = "Dissaphong Thachasongtham and Takumi Yoshida and {De Sorbier}, Fran{\cc}ois and Hideo Saito",
year = "2013",
doi = "10.1007/978-3-642-38886-6_48",
language = "English",
isbn = "9783642388859",
volume = "7944 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "512--521",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - 3D object pose estimation using viewpoint generative learning

AU - Thachasongtham, Dissaphong

AU - Yoshida, Takumi

AU - De Sorbier, François

AU - Saito, Hideo

PY - 2013

Y1 - 2013

N2 - Conventional local features such as SIFT or SURF are robust to scale and rotation changes but sensitive to large perspective change. Because perspective change always occurs when 3D object moves, using these features to estimate the pose of a 3D object is a challenging task. In this paper, we extend one of our previous works on viewpoint generative learning to 3D objects. Given a model of a textured object, we virtually generate several patterns of the model from different viewpoints and select stable keypoints from those patterns. Then our system learns a collection of feature descriptors from the stable keypoints. Finally, we are able to estimate the pose of a 3D object by using these robust features. In our experimental results, we demonstrate that our system is robust against large viewpoint change and even under partial occlusion.

AB - Conventional local features such as SIFT or SURF are robust to scale and rotation changes but sensitive to large perspective change. Because perspective change always occurs when 3D object moves, using these features to estimate the pose of a 3D object is a challenging task. In this paper, we extend one of our previous works on viewpoint generative learning to 3D objects. Given a model of a textured object, we virtually generate several patterns of the model from different viewpoints and select stable keypoints from those patterns. Then our system learns a collection of feature descriptors from the stable keypoints. Finally, we are able to estimate the pose of a 3D object by using these robust features. In our experimental results, we demonstrate that our system is robust against large viewpoint change and even under partial occlusion.

KW - generative learning

KW - pose estimation

KW - stable keypoint

UR - http://www.scopus.com/inward/record.url?scp=84884486752&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84884486752&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-38886-6_48

DO - 10.1007/978-3-642-38886-6_48

M3 - Conference contribution

SN - 9783642388859

VL - 7944 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 512

EP - 521

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -