File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: Command-driven Articulated Object Understanding and Manipulation

TitleCommand-driven Articulated Object Understanding and Manipulation
Authors
Issue Date22-Aug-2023
Abstract

We present Cart, a new approach towards articulatedobject manipulations by human commands. Beyond the existing work that focuses on inferring articulation structures, we further support manipulating articulated shapes to align them subject to simple command templates. The key of Cart is to utilize the prediction of object structures to connect visual observations with user commands for effective manipulations. It is achieved by encoding command messages for motion prediction and a test-time adaptation to adjust the amount of movement from only command supervision. For a rich variety of object categories, Cart can accurately manipulate object shapes and outperform the state-of-the-art approaches in understanding the inherent articulation structures. Also, it can well generalize to unseen object categories and real-world objects. We hope Cart could open new directions for instructing machines to operate articulated objects. Code is available at https://github.com/dvlab-research/Cart.


Persistent Identifierhttp://hdl.handle.net/10722/333839

 

DC FieldValueLanguage
dc.contributor.authorChu, Ruihang-
dc.contributor.authorLiu, Zhengzhe-
dc.contributor.authorYe, Xiaoqing-
dc.contributor.authorTan, Xiao-
dc.contributor.authorQi, Xiaojuan-
dc.contributor.authorFu, Chi-Wing-
dc.contributor.authorJia, Jiaya-
dc.date.accessioned2023-10-06T08:39:30Z-
dc.date.available2023-10-06T08:39:30Z-
dc.date.issued2023-08-22-
dc.identifier.urihttp://hdl.handle.net/10722/333839-
dc.description.abstract<p>We present Cart, a new approach towards articulatedobject manipulations by human commands. Beyond the existing work that focuses on inferring articulation structures, we further support manipulating articulated shapes to align them subject to simple command templates. The key of Cart is to utilize the prediction of object structures to connect visual observations with user commands for effective manipulations. It is achieved by encoding command messages for motion prediction and a test-time adaptation to adjust the amount of movement from only command supervision. For a rich variety of object categories, Cart can accurately manipulate object shapes and outperform the state-of-the-art approaches in understanding the inherent articulation structures. Also, it can well generalize to unseen object categories and real-world objects. We hope Cart could open new directions for instructing machines to operate articulated objects. Code is available at https://github.com/dvlab-research/Cart.<br></p>-
dc.languageeng-
dc.relation.ispartof2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (17/06/2023-24/06/2023, Vancouver, BC, Canada)-
dc.titleCommand-driven Articulated Object Understanding and Manipulation-
dc.typeConference_Paper-
dc.identifier.doi10.1109/CVPR52729.2023.00851-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats