TP-AE: temporally primed 6D object pose tracking with auto-encoders

Linfang Zheng; Ales Leonardis; Tze Ho Elden Tse; Nora Horanyi; Hua Chen; Wei Zhang; Hyung Jin Chang

doi:10.1109/ICRA46639.2022.9811890

TP-AE: temporally primed 6D object pose tracking with auto-encoders

Linfang Zheng, Ales Leonardis, Tze Ho Elden Tse, Nora Horanyi, Hua Chen, Wei Zhang, Hyung Jin Chang

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

524 Downloads (Pure)

Abstract

Fast and accurate tracking of an object’s motion is one of the key functionalities of a robotic system for achieving reliable interaction with the environment. This paper focuses on the instance-level six-dimensional (6D) pose tracking problem with a symmetric and textureless object under occlusion. We propose a Temporally Primed 6D pose tracking framework with Auto-Encoders (TP-AE) to tackle the pose tracking problem. The framework consists of a prediction step and a temporally primed pose estimation step. The prediction step aims to quickly and efficiently generate a guess on the object’s real-time pose based on historical information about the target object’s motion. Once the prior prediction is obtained, the temporally primed pose estimation step embeds the prior pose into the RGB-D input, and leverages auto-encoders to reconstruct the target object with higher quality under occlusion, thus improving the framework’s performance. Extensive experiments show that the proposed 6D pose tracking method can accurately estimate the 6D pose of a symmetric and textureless object under occlusion, and significantly outperforms the state-of-the-art on T-LESS dataset while running in real-time at 26 FPS.

Original language	English
Title of host publication	2022 IEEE International Conference on Robotics and Automation (ICRA)
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	10616-10623
Number of pages	8
ISBN (Electronic)	9781728196817
ISBN (Print)	9781728196824 (PoD)
DOIs	https://doi.org/10.1109/ICRA46639.2022.9811890
Publication status	Published - 12 Jul 2022
Event	39th IEEE International Conference on Robotics and Automation, ICRA 2022 - Philadelphia, United States Duration: 23 May 2022 → 27 May 2022

Publication series

Name	Proceedings - IEEE International Conference on Robotics and Automation
Publisher	IEEE
ISSN (Print)	1050-4729
ISSN (Electronic)	2577-087X

Conference

Conference	39th IEEE International Conference on Robotics and Automation, ICRA 2022
Country/Territory	United States
City	Philadelphia
Period	23/05/22 → 27/05/22

Bibliographical note

Funding Information:
This work is supported in part by the National Natural Science Foundation of China under Grants 62073159 and 62003155, the Shenzhen Science and Technology Program under Grant JCYJ20200109141601708, the Science, Technology and Innovation Commission of Shenzhen Municipality under grant ZDSYS20200811143601004, and the Institute of Information and communications Technology Planning and evaluation (IITP) grant funded by the Korea government (MSIT) (2021-0-00537, Visual common sense through self-supervised learning for restoration of invisible parts in images). (Corresponding author: Wei Zhang).

Publisher Copyright:
© 2022 IEEE.

Keywords

Target tracking
Automation
Pose estimation
Real-time systems
Reliability
Robots

Access to Document

10.1109/ICRA46639.2022.9811890Licence: None: All rights reserved

ZhengL2022TPAEAccepted author manuscript, 15.8 MB

Cite this

Zheng, L., Leonardis, A., Tse, T. H. E., Horanyi, N., Chen, H., Zhang, W., & Chang, H. J. (2022). TP-AE: temporally primed 6D object pose tracking with auto-encoders. In 2022 IEEE International Conference on Robotics and Automation (ICRA) (pp. 10616-10623). Article 9811890 (Proceedings - IEEE International Conference on Robotics and Automation). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/ICRA46639.2022.9811890

@inproceedings{8fff41ce9c0841879f81f56d244f265b,

title = "TP-AE: temporally primed 6D object pose tracking with auto-encoders",

abstract = "Fast and accurate tracking of an object{\textquoteright}s motion is one of the key functionalities of a robotic system for achieving reliable interaction with the environment. This paper focuses on the instance-level six-dimensional (6D) pose tracking problem with a symmetric and textureless object under occlusion. We propose a Temporally Primed 6D pose tracking framework with Auto-Encoders (TP-AE) to tackle the pose tracking problem. The framework consists of a prediction step and a temporally primed pose estimation step. The prediction step aims to quickly and efficiently generate a guess on the object{\textquoteright}s real-time pose based on historical information about the target object{\textquoteright}s motion. Once the prior prediction is obtained, the temporally primed pose estimation step embeds the prior pose into the RGB-D input, and leverages auto-encoders to reconstruct the target object with higher quality under occlusion, thus improving the framework{\textquoteright}s performance. Extensive experiments show that the proposed 6D pose tracking method can accurately estimate the 6D pose of a symmetric and textureless object under occlusion, and significantly outperforms the state-of-the-art on T-LESS dataset while running in real-time at 26 FPS.",

keywords = "Target tracking, Automation, Pose estimation, Real-time systems, Reliability, Robots",

author = "Linfang Zheng and Ales Leonardis and Tse, {Tze Ho Elden} and Nora Horanyi and Hua Chen and Wei Zhang and Chang, {Hyung Jin}",

note = "Funding Information: This work is supported in part by the National Natural Science Foundation of China under Grants 62073159 and 62003155, the Shenzhen Science and Technology Program under Grant JCYJ20200109141601708, the Science, Technology and Innovation Commission of Shenzhen Municipality under grant ZDSYS20200811143601004, and the Institute of Information and communications Technology Planning and evaluation (IITP) grant funded by the Korea government (MSIT) (2021-0-00537, Visual common sense through self-supervised learning for restoration of invisible parts in images). (Corresponding author: Wei Zhang). Publisher Copyright: {\textcopyright} 2022 IEEE.; 39th IEEE International Conference on Robotics and Automation, ICRA 2022 ; Conference date: 23-05-2022 Through 27-05-2022",

year = "2022",

month = jul,

day = "12",

doi = "10.1109/ICRA46639.2022.9811890",

language = "English",

isbn = "9781728196824 (PoD)",

series = "Proceedings - IEEE International Conference on Robotics and Automation",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

pages = "10616--10623",

booktitle = "2022 IEEE International Conference on Robotics and Automation (ICRA)",

}

Zheng, L, Leonardis, A, Tse, THE, Horanyi, N, Chen, H, Zhang, W & Chang, HJ 2022, TP-AE: temporally primed 6D object pose tracking with auto-encoders. in 2022 IEEE International Conference on Robotics and Automation (ICRA)., 9811890, Proceedings - IEEE International Conference on Robotics and Automation, Institute of Electrical and Electronics Engineers (IEEE), pp. 10616-10623, 39th IEEE International Conference on Robotics and Automation, ICRA 2022, Philadelphia, United States, 23/05/22. https://doi.org/10.1109/ICRA46639.2022.9811890

TP-AE: temporally primed 6D object pose tracking with auto-encoders. / Zheng, Linfang; Leonardis, Ales; Tse, Tze Ho Elden et al.
2022 IEEE International Conference on Robotics and Automation (ICRA). Institute of Electrical and Electronics Engineers (IEEE), 2022. p. 10616-10623 9811890 (Proceedings - IEEE International Conference on Robotics and Automation).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - TP-AE

T2 - 39th IEEE International Conference on Robotics and Automation, ICRA 2022

AU - Zheng, Linfang

AU - Leonardis, Ales

AU - Tse, Tze Ho Elden

AU - Horanyi, Nora

AU - Chen, Hua

AU - Zhang, Wei

AU - Chang, Hyung Jin

N1 - Funding Information: This work is supported in part by the National Natural Science Foundation of China under Grants 62073159 and 62003155, the Shenzhen Science and Technology Program under Grant JCYJ20200109141601708, the Science, Technology and Innovation Commission of Shenzhen Municipality under grant ZDSYS20200811143601004, and the Institute of Information and communications Technology Planning and evaluation (IITP) grant funded by the Korea government (MSIT) (2021-0-00537, Visual common sense through self-supervised learning for restoration of invisible parts in images). (Corresponding author: Wei Zhang). Publisher Copyright: © 2022 IEEE.

PY - 2022/7/12

Y1 - 2022/7/12

N2 - Fast and accurate tracking of an object’s motion is one of the key functionalities of a robotic system for achieving reliable interaction with the environment. This paper focuses on the instance-level six-dimensional (6D) pose tracking problem with a symmetric and textureless object under occlusion. We propose a Temporally Primed 6D pose tracking framework with Auto-Encoders (TP-AE) to tackle the pose tracking problem. The framework consists of a prediction step and a temporally primed pose estimation step. The prediction step aims to quickly and efficiently generate a guess on the object’s real-time pose based on historical information about the target object’s motion. Once the prior prediction is obtained, the temporally primed pose estimation step embeds the prior pose into the RGB-D input, and leverages auto-encoders to reconstruct the target object with higher quality under occlusion, thus improving the framework’s performance. Extensive experiments show that the proposed 6D pose tracking method can accurately estimate the 6D pose of a symmetric and textureless object under occlusion, and significantly outperforms the state-of-the-art on T-LESS dataset while running in real-time at 26 FPS.

AB - Fast and accurate tracking of an object’s motion is one of the key functionalities of a robotic system for achieving reliable interaction with the environment. This paper focuses on the instance-level six-dimensional (6D) pose tracking problem with a symmetric and textureless object under occlusion. We propose a Temporally Primed 6D pose tracking framework with Auto-Encoders (TP-AE) to tackle the pose tracking problem. The framework consists of a prediction step and a temporally primed pose estimation step. The prediction step aims to quickly and efficiently generate a guess on the object’s real-time pose based on historical information about the target object’s motion. Once the prior prediction is obtained, the temporally primed pose estimation step embeds the prior pose into the RGB-D input, and leverages auto-encoders to reconstruct the target object with higher quality under occlusion, thus improving the framework’s performance. Extensive experiments show that the proposed 6D pose tracking method can accurately estimate the 6D pose of a symmetric and textureless object under occlusion, and significantly outperforms the state-of-the-art on T-LESS dataset while running in real-time at 26 FPS.

KW - Target tracking

KW - Automation

KW - Pose estimation

KW - Real-time systems

KW - Reliability

KW - Robots

UR - https://www.icra2022.org/

UR - http://www.scopus.com/inward/record.url?scp=85136335236&partnerID=8YFLogxK

U2 - 10.1109/ICRA46639.2022.9811890

DO - 10.1109/ICRA46639.2022.9811890

M3 - Conference contribution

SN - 9781728196824 (PoD)

T3 - Proceedings - IEEE International Conference on Robotics and Automation

SP - 10616

EP - 10623

BT - 2022 IEEE International Conference on Robotics and Automation (ICRA)

PB - Institute of Electrical and Electronics Engineers (IEEE)

Y2 - 23 May 2022 through 27 May 2022

ER -

Zheng L, Leonardis A, Tse THE, Horanyi N, Chen H, Zhang W et al. TP-AE: temporally primed 6D object pose tracking with auto-encoders. In 2022 IEEE International Conference on Robotics and Automation (ICRA). Institute of Electrical and Electronics Engineers (IEEE). 2022. p. 10616-10623. 9811890. (Proceedings - IEEE International Conference on Robotics and Automation). doi: 10.1109/ICRA46639.2022.9811890

TP-AE: temporally primed 6D object pose tracking with auto-encoders

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this