elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Privacy Policy | Contact | Deutsch
Fontsize: [-] Text [+]

"What's This?" - Learning to Segment Unknown Objects from Manipulation Sequences

Boerdijk, Wout and Sundermeyer, Martin and Durner, Maximilian and Triebel, Rudolph (2021) "What's This?" - Learning to Segment Unknown Objects from Manipulation Sequences. In: 2021 IEEE International Conference on Robotics and Automation, ICRA 2021. IEEE Robotics and Automation Society. 2021 IEEE International Conference on Robotics and Automation, ICRA 2021, 31 May - 05 June 2021, Xi'an, China / online (hybrid). ISBN 978-172819077-8. ISSN 1050-4729.

[img] PDF
4MB

Abstract

We present a novel framework for self-supervised grasped object segmentation with a robotic manipulator. Our method successively learns an agnostic foreground segmentation followed by a distinction between manipulator and object solely by observing the motion between consecutive RGB frames. In contrast to previous approaches, we propose a single, end-toend trainable architecture which jointly incorporates motion cues and semantic knowledge. Furthermore, while the motion of the manipulator and the object are substantial cues for our algorithm, we present means to robustly deal with distraction objects moving in the background, as well as with completely static scenes. Our method neither depends on any visual registration of a kinematic robot or 3D object models, nor on precise hand eye calibration or any additional sensor data. By extensive experimental evaluation we demonstrate the superiority of our framework and provide detailed insights on its capability of dealing with the aforementioned extreme cases of motion. We also show that training a semantic segmentation network with the automatically labeled data achieves results on par with manually annotated training data. Code and pretrained model are available at https://github.com/DLR-RM/DistinctNet.

Item URL in elib:https://elib.dlr.de/146603/
Document Type:Conference or Workshop Item (Speech)
Title:"What's This?" - Learning to Segment Unknown Objects from Manipulation Sequences
Authors:
AuthorsInstitution or Email of AuthorsAuthor's ORCID iD
Boerdijk, WoutWout.Boerdijk (at) dlr.dehttps://orcid.org/0000-0003-0789-5970
Sundermeyer, MartinMartin.Sundermeyer (at) dlr.dehttps://orcid.org/0000-0003-0587-9643
Durner, Maximilianmaximilian.durner (at) dlr.dehttps://orcid.org/0000-0001-8885-5334
Triebel, RudolphRudolph.Triebel (at) dlr.dehttps://orcid.org/0000-0002-7975-036X
Date:2021
Journal or Publication Title:2021 IEEE International Conference on Robotics and Automation, ICRA 2021
Refereed publication:Yes
Open Access:Yes
Gold Open Access:No
In SCOPUS:Yes
In ISI Web of Science:No
Publisher:IEEE Robotics and Automation Society
ISSN:1050-4729
ISBN:978-172819077-8
Status:Published
Keywords:Deep Learning for Visual Perception, Semantic Segmentation, Computer Vision for Automation
Event Title:2021 IEEE International Conference on Robotics and Automation, ICRA 2021
Event Location:Xi'an, China / online (hybrid)
Event Type:international Conference
Event Dates:31 May - 05 June 2021
HGF - Research field:Aeronautics, Space and Transport
HGF - Program:Space
HGF - Program Themes:Robotics
DLR - Research area:Raumfahrt
DLR - Program:R RO - Robotics
DLR - Research theme (Project):R - Multisensory World Modelling (RM) [RO]
Location: Oberpfaffenhofen
Institutes and Institutions:Institute of Robotics and Mechatronics (since 2013) > Perception and Cognition
Deposited By: Boerdijk, Wout
Deposited On:08 Dec 2021 14:22
Last Modified:25 Mar 2022 10:33

Repository Staff Only: item control page

Browse
Search
Help & Contact
Information
electronic library is running on EPrints 3.3.12
Copyright © 2008-2017 German Aerospace Center (DLR). All rights reserved.