PointPoseNet: Point Pose Network for Robust 6D Object Pose Estimation

Wei Chen; Jinming Duan; Hector Basevi; Hyung Jin Chang; Ales Leonardis

doi:10.1109/WACV45572.2020.9093272

PointPoseNet: Point Pose Network for Robust 6D Object Pose Estimation

Wei Chen^*, Jinming Duan, Hector Basevi, Hyung Jin Chang, Ales Leonardis

^*Corresponding author for this work

Computer Science

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

4 Citations (Scopus)

31 Downloads (Pure)

Abstract

In this paper, we propose a novel pipeline to estimate 6D object pose from RGB-D images of known objects present in complex scenes. The pipeline directly operates on raw point clouds extracted from RGB-D scans. Specifically, our method takes the point cloud as input and regresses the point-wise unit vectors pointing to the 3D keypoints. We then use these vectors to generate keypoint hypotheses from which the 6D object pose hypotheses are computed. Finally, we select the best 6D object pose from the hypotheses based on a proposed scoring mechanism with geometry constraints. Extensive experiments show that the proposed method is robust against the variety in object shape and appearance as well as occlusions between objects, and that our method outperforms the state-of-the-art methods on the LINEMOD and Occlusion LINEMOD datasets.

Original language	English
Title of host publication	2020 IEEE Winter Conference on Applications of Computer Vision (WACV)
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	2813-2822
Number of pages	10
ISBN (Electronic)	9781728165530, 9781728165523 (USB)
ISBN (Print)	9781728165547 (PoD)
DOIs	https://doi.org/10.1109/WACV45572.2020.9093272
Publication status	Published - 14 May 2020
Event	2020 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2020 - Snowmass Village, United States Duration: 1 Mar 2020 → 5 Mar 2020

Publication series

Name	IEEE Workshop on Applications of Computer Vision
Publisher	IEEE
ISSN (Print)	2472-6737
ISSN (Electronic)	2642-9381

Conference

Conference	2020 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2020
Country/Territory	United States
City	Snowmass Village
Period	1/03/20 → 5/03/20

Bibliographical note

Publisher Copyright:
© 2020 IEEE.

Keywords

Three-dimensional displays
Pose estimation
Two dimensional displays
Geometry
Feature extraction
Robustness
Pipelines

ASJC Scopus subject areas

Computer Science Applications
Computer Vision and Pattern Recognition

Access to Document

10.1109/WACV45572.2020.9093272

ChenW2020PointPoseNet
W. Chen, J. Duan, H. Basevi, H. J. Chang and A. Leonardis, "PointPoseNet: Point Pose Network for Robust 6D Object Pose Estimation," 2020 IEEE Winter Conference on Applications of Computer Vision (WACV), Snowmass, CO, USA, 2020, pp. 2813-2822, doi: 10.1109/WACV45572.2020.9093272. © 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Accepted author manuscript, 8.13 MBLicence: Other (please specify with Rights Statement)

Understanding scenes and events through joint parsing, cognitive reasoning and lifelong learning (Oxford lead)
Leonardis, A. (Principal Investigator)
Engineering & Physical Science Research Council
1/01/16 → 28/02/22
Project: Research Councils

Cite this

Chen, W., Duan, J., Basevi, H., Chang, H. J., & Leonardis, A. (2020). PointPoseNet: Point Pose Network for Robust 6D Object Pose Estimation. In 2020 IEEE Winter Conference on Applications of Computer Vision (WACV) (pp. 2813-2822). Article 9093272 (IEEE Workshop on Applications of Computer Vision). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/WACV45572.2020.9093272

@inproceedings{d825386cdb5c4004932087b62635f7c5,

title = "PointPoseNet: Point Pose Network for Robust 6D Object Pose Estimation",

abstract = "In this paper, we propose a novel pipeline to estimate 6D object pose from RGB-D images of known objects present in complex scenes. The pipeline directly operates on raw point clouds extracted from RGB-D scans. Specifically, our method takes the point cloud as input and regresses the point-wise unit vectors pointing to the 3D keypoints. We then use these vectors to generate keypoint hypotheses from which the 6D object pose hypotheses are computed. Finally, we select the best 6D object pose from the hypotheses based on a proposed scoring mechanism with geometry constraints. Extensive experiments show that the proposed method is robust against the variety in object shape and appearance as well as occlusions between objects, and that our method outperforms the state-of-the-art methods on the LINEMOD and Occlusion LINEMOD datasets.",

keywords = "Three-dimensional displays, Pose estimation, Two dimensional displays, Geometry, Feature extraction, Robustness, Pipelines",

author = "Wei Chen and Jinming Duan and Hector Basevi and Chang, {Hyung Jin} and Ales Leonardis",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 2020 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2020 ; Conference date: 01-03-2020 Through 05-03-2020",

year = "2020",

month = may,

day = "14",

doi = "10.1109/WACV45572.2020.9093272",

language = "English",

isbn = "9781728165547 (PoD)",

series = "IEEE Workshop on Applications of Computer Vision",

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

pages = "2813--2822",

booktitle = "2020 IEEE Winter Conference on Applications of Computer Vision (WACV)",

}

Chen, W, Duan, J , Basevi, H , Chang, HJ & Leonardis, A 2020, PointPoseNet: Point Pose Network for Robust 6D Object Pose Estimation. in 2020 IEEE Winter Conference on Applications of Computer Vision (WACV)., 9093272, IEEE Workshop on Applications of Computer Vision, Institute of Electrical and Electronics Engineers (IEEE), pp. 2813-2822, 2020 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2020, Snowmass Village, United States, 1/03/20. https://doi.org/10.1109/WACV45572.2020.9093272

PointPoseNet: Point Pose Network for Robust 6D Object Pose Estimation. / Chen, Wei; Duan, Jinming ; Basevi, Hector et al.
2020 IEEE Winter Conference on Applications of Computer Vision (WACV). Institute of Electrical and Electronics Engineers (IEEE), 2020. p. 2813-2822 9093272 (IEEE Workshop on Applications of Computer Vision).

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

TY - GEN

T1 - PointPoseNet

T2 - 2020 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2020

AU - Chen, Wei

AU - Duan, Jinming

AU - Basevi, Hector

AU - Chang, Hyung Jin

AU - Leonardis, Ales

PY - 2020/5/14

Y1 - 2020/5/14

N2 - In this paper, we propose a novel pipeline to estimate 6D object pose from RGB-D images of known objects present in complex scenes. The pipeline directly operates on raw point clouds extracted from RGB-D scans. Specifically, our method takes the point cloud as input and regresses the point-wise unit vectors pointing to the 3D keypoints. We then use these vectors to generate keypoint hypotheses from which the 6D object pose hypotheses are computed. Finally, we select the best 6D object pose from the hypotheses based on a proposed scoring mechanism with geometry constraints. Extensive experiments show that the proposed method is robust against the variety in object shape and appearance as well as occlusions between objects, and that our method outperforms the state-of-the-art methods on the LINEMOD and Occlusion LINEMOD datasets.

AB - In this paper, we propose a novel pipeline to estimate 6D object pose from RGB-D images of known objects present in complex scenes. The pipeline directly operates on raw point clouds extracted from RGB-D scans. Specifically, our method takes the point cloud as input and regresses the point-wise unit vectors pointing to the 3D keypoints. We then use these vectors to generate keypoint hypotheses from which the 6D object pose hypotheses are computed. Finally, we select the best 6D object pose from the hypotheses based on a proposed scoring mechanism with geometry constraints. Extensive experiments show that the proposed method is robust against the variety in object shape and appearance as well as occlusions between objects, and that our method outperforms the state-of-the-art methods on the LINEMOD and Occlusion LINEMOD datasets.

KW - Three-dimensional displays

KW - Pose estimation

KW - Two dimensional displays

KW - Geometry

KW - Feature extraction

KW - Robustness

KW - Pipelines

UR - http://www.scopus.com/inward/record.url?scp=85085495782&partnerID=8YFLogxK

U2 - 10.1109/WACV45572.2020.9093272

DO - 10.1109/WACV45572.2020.9093272

M3 - Conference contribution

AN - SCOPUS:85085495782

SN - 9781728165547 (PoD)

T3 - IEEE Workshop on Applications of Computer Vision

SP - 2813

EP - 2822

BT - 2020 IEEE Winter Conference on Applications of Computer Vision (WACV)

PB - Institute of Electrical and Electronics Engineers (IEEE)

Y2 - 1 March 2020 through 5 March 2020

ER -

PointPoseNet: Point Pose Network for Robust 6D Object Pose Estimation

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Projects

Understanding scenes and events through joint parsing, cognitive reasoning and lifelong learning (Oxford lead)

Cite this