Beyond Top-Grasps Through Scene Completion

Current end-to-end grasp planning methods propose grasps in the order of seconds that attain high grasp success rates on a diverse set of objects, but often by constraining the workspace to top-grasps. In this work, we present a method that allows end-to-end top-grasp planning methods to generate fu...

Full description

Saved in:
Bibliographic Details
Main Authors: Lundell, Jens, Verdoja, Francesco, Kyrki, Ville
Format: Conference Proceeding
Language:English
Subjects:
Cameras
Grasping
Pipelines
Planning
Robot vision systems
Shape
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page 551
container_issue
container_start_page 545
container_title
container_volume
creator Lundell, Jens
Verdoja, Francesco
Kyrki, Ville
description Current end-to-end grasp planning methods propose grasps in the order of seconds that attain high grasp success rates on a diverse set of objects, but often by constraining the workspace to top-grasps. In this work, we present a method that allows end-to-end top-grasp planning methods to generate full six-degree-of-freedom grasps using a single RGBD view as input. This is achieved by estimating the complete shape of the object to be grasped, then simulating different viewpoints of the object, passing the simulated viewpoints to an end-to-end grasp generation method, and finally executing the overall best grasp. The method was experimentally validated on a Franka Emika Panda by comparing 429 grasps generated by the state-of-the-art Fully Convolutional Grasp Quality CNN, both on simulated and real camera images. The results show statistically significant improvements in terms of grasp success rate when using simulated images over real camera images, especially when the real camera viewpoint is angled. Code and video are available at https://irobotics.aalto.fi/beyond-topgrasps-through-scene-completion/.
doi_str_mv 10.1109/ICRA40945.2020.9197320
format conference_proceeding
fullrecord <record><control><sourceid>ieee_CHZPO</sourceid><recordid>TN_cdi_ieee_primary_9197320</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>9197320</ieee_id><sourcerecordid>9197320</sourcerecordid><originalsourceid>FETCH-LOGICAL-i203t-7570fd50f72c14438865c27005e7860debc0eb12b48c6132cb499ba09e44232f3</originalsourceid><addsrcrecordid>eNotz0FLwzAUwPEoCG7TTyBIz0Lqy0vSlxxn0TkYCFrB22jSV1fZmtLOw769B3f6337wF-JeQa4U-Md1-b404I3NERByrzxphAsxV4ROkfaWLsUMLZEER1_XYj5NPwCgdVHMxMMTn1LfZFUa5Gqsp2HKqt2Yfr932UfknrMyHYY9H7vU34irtt5PfHvuQny-PFflq9y8rdblciM7BH2UZAnaxkJLGJUx2rnCRiQAy-QKaDhE4KAwGBcLpTEG432owbMxqLHVC3H373bMvB3G7lCPp-35S_8BxB1ArQ</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>Beyond Top-Grasps Through Scene Completion</title><source>IEEE Xplore All Conference Series</source><creator>Lundell, Jens ; Verdoja, Francesco ; Kyrki, Ville</creator><creatorcontrib>Lundell, Jens ; Verdoja, Francesco ; Kyrki, Ville</creatorcontrib><description>Current end-to-end grasp planning methods propose grasps in the order of seconds that attain high grasp success rates on a diverse set of objects, but often by constraining the workspace to top-grasps. In this work, we present a method that allows end-to-end top-grasp planning methods to generate full six-degree-of-freedom grasps using a single RGBD view as input. This is achieved by estimating the complete shape of the object to be grasped, then simulating different viewpoints of the object, passing the simulated viewpoints to an end-to-end grasp generation method, and finally executing the overall best grasp. The method was experimentally validated on a Franka Emika Panda by comparing 429 grasps generated by the state-of-the-art Fully Convolutional Grasp Quality CNN, both on simulated and real camera images. The results show statistically significant improvements in terms of grasp success rate when using simulated images over real camera images, especially when the real camera viewpoint is angled. Code and video are available at https://irobotics.aalto.fi/beyond-topgrasps-through-scene-completion/.</description><identifier>EISSN: 2577-087X</identifier><identifier>EISBN: 1728173957</identifier><identifier>EISBN: 9781728173955</identifier><identifier>DOI: 10.1109/ICRA40945.2020.9197320</identifier><language>eng</language><publisher>IEEE</publisher><subject>Cameras ; Grasping ; Pipelines ; Planning ; Robot vision systems ; Shape</subject><ispartof>2020 IEEE International Conference on Robotics and Automation (ICRA), 2020, p.545-551</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/9197320$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,780,784,789,790,23930,23931,25140,27925,54555,54932</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/9197320$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Lundell, Jens</creatorcontrib><creatorcontrib>Verdoja, Francesco</creatorcontrib><creatorcontrib>Kyrki, Ville</creatorcontrib><title>Beyond Top-Grasps Through Scene Completion</title><title>2020 IEEE International Conference on Robotics and Automation (ICRA)</title><addtitle>ICRA</addtitle><description>Current end-to-end grasp planning methods propose grasps in the order of seconds that attain high grasp success rates on a diverse set of objects, but often by constraining the workspace to top-grasps. In this work, we present a method that allows end-to-end top-grasp planning methods to generate full six-degree-of-freedom grasps using a single RGBD view as input. This is achieved by estimating the complete shape of the object to be grasped, then simulating different viewpoints of the object, passing the simulated viewpoints to an end-to-end grasp generation method, and finally executing the overall best grasp. The method was experimentally validated on a Franka Emika Panda by comparing 429 grasps generated by the state-of-the-art Fully Convolutional Grasp Quality CNN, both on simulated and real camera images. The results show statistically significant improvements in terms of grasp success rate when using simulated images over real camera images, especially when the real camera viewpoint is angled. Code and video are available at https://irobotics.aalto.fi/beyond-topgrasps-through-scene-completion/.</description><subject>Cameras</subject><subject>Grasping</subject><subject>Pipelines</subject><subject>Planning</subject><subject>Robot vision systems</subject><subject>Shape</subject><issn>2577-087X</issn><isbn>1728173957</isbn><isbn>9781728173955</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2020</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><recordid>eNotz0FLwzAUwPEoCG7TTyBIz0Lqy0vSlxxn0TkYCFrB22jSV1fZmtLOw769B3f6337wF-JeQa4U-Md1-b404I3NERByrzxphAsxV4ROkfaWLsUMLZEER1_XYj5NPwCgdVHMxMMTn1LfZFUa5Gqsp2HKqt2Yfr932UfknrMyHYY9H7vU34irtt5PfHvuQny-PFflq9y8rdblciM7BH2UZAnaxkJLGJUx2rnCRiQAy-QKaDhE4KAwGBcLpTEG432owbMxqLHVC3H373bMvB3G7lCPp-35S_8BxB1ArQ</recordid><startdate>202005</startdate><enddate>202005</enddate><creator>Lundell, Jens</creator><creator>Verdoja, Francesco</creator><creator>Kyrki, Ville</creator><general>IEEE</general><scope>6IE</scope><scope>6IH</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIO</scope></search><sort><creationdate>202005</creationdate><title>Beyond Top-Grasps Through Scene Completion</title><author>Lundell, Jens ; Verdoja, Francesco ; Kyrki, Ville</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i203t-7570fd50f72c14438865c27005e7860debc0eb12b48c6132cb499ba09e44232f3</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Cameras</topic><topic>Grasping</topic><topic>Pipelines</topic><topic>Planning</topic><topic>Robot vision systems</topic><topic>Shape</topic><toplevel>online_resources</toplevel><creatorcontrib>Lundell, Jens</creatorcontrib><creatorcontrib>Verdoja, Francesco</creatorcontrib><creatorcontrib>Kyrki, Ville</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan (POP) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE</collection><collection>IEEE Proceedings Order Plans (POP) 1998-present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Lundell, Jens</au><au>Verdoja, Francesco</au><au>Kyrki, Ville</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>Beyond Top-Grasps Through Scene Completion</atitle><btitle>2020 IEEE International Conference on Robotics and Automation (ICRA)</btitle><stitle>ICRA</stitle><date>2020-05</date><risdate>2020</risdate><spage>545</spage><epage>551</epage><pages>545-551</pages><eissn>2577-087X</eissn><eisbn>1728173957</eisbn><eisbn>9781728173955</eisbn><abstract>Current end-to-end grasp planning methods propose grasps in the order of seconds that attain high grasp success rates on a diverse set of objects, but often by constraining the workspace to top-grasps. In this work, we present a method that allows end-to-end top-grasp planning methods to generate full six-degree-of-freedom grasps using a single RGBD view as input. This is achieved by estimating the complete shape of the object to be grasped, then simulating different viewpoints of the object, passing the simulated viewpoints to an end-to-end grasp generation method, and finally executing the overall best grasp. The method was experimentally validated on a Franka Emika Panda by comparing 429 grasps generated by the state-of-the-art Fully Convolutional Grasp Quality CNN, both on simulated and real camera images. The results show statistically significant improvements in terms of grasp success rate when using simulated images over real camera images, especially when the real camera viewpoint is angled. Code and video are available at https://irobotics.aalto.fi/beyond-topgrasps-through-scene-completion/.</abstract><pub>IEEE</pub><doi>10.1109/ICRA40945.2020.9197320</doi><tpages>7</tpages></addata></record>
fulltext fulltext_linktorsrc
identifier EISSN: 2577-087X
ispartof 2020 IEEE International Conference on Robotics and Automation (ICRA), 2020, p.545-551
issn 2577-087X
language eng
recordid cdi_ieee_primary_9197320
source IEEE Xplore All Conference Series
subjects Cameras
Grasping
Pipelines
Planning
Robot vision systems
Shape
title Beyond Top-Grasps Through Scene Completion
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-01T16%3A48%3A40IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_CHZPO&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=Beyond%20Top-Grasps%20Through%20Scene%20Completion&rft.btitle=2020%20IEEE%20International%20Conference%20on%20Robotics%20and%20Automation%20(ICRA)&rft.au=Lundell,%20Jens&rft.date=2020-05&rft.spage=545&rft.epage=551&rft.pages=545-551&rft.eissn=2577-087X&rft_id=info:doi/10.1109/ICRA40945.2020.9197320&rft.eisbn=1728173957&rft.eisbn_list=9781728173955&rft_dat=%3Cieee_CHZPO%3E9197320%3C/ieee_CHZPO%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-i203t-7570fd50f72c14438865c27005e7860debc0eb12b48c6132cb499ba09e44232f3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=9197320&rfr_iscdi=true