File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Weakly-Supervised Learning of Category-Specific 3D Object Shapes

TitleWeakly-Supervised Learning of Category-Specific 3D Object Shapes
Authors
Keywords3D shape reconstruction
common object segmentation
viewpoint estimation
Issue Date2021
Citation
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, v. 43, n. 4, p. 1423-1437 How to Cite?
AbstractCategory-specific 3D object shape models have greatly boosted the recent advances in object detection, recognition and segmentation. However, even the most advanced approach for learning 3D object shapes still requires heavy manual annotations on large-scale 2D images. Such annotations include object categories, object keypoints, and figure-ground segmentation for the instances in each image. In particular, annotating figure-ground segmentation is unbearably labor-intensive and time-consuming. To address this problem, this paper devotes to learn category-specific 3D shape models under weak supervision, where only object categories and keypoints are required to be manually annotated on the training 2D images. By exploring the underlying relationship between two tasks: object segmentation and category-specific 3D shape reconstruction, we propose a novel weakly-supervised learning framework to jointly address these two tasks and combine them to boost the final performance of the learned 3D shape models. Moreover, learning without using figure-ground segmentation leads to ambiguous solutions. To this end, we develop the confidence weighting schemes in the viewpoint estimation and 3D shape learning procedure. These schemes effectively reduce the confusion caused by the noisy data and thus increase the chances for recovering more reliable 3D object shapes. Comprehensive experiments on the challenging PASCAL VOC benchmark show that our framework achieves comparable performance with the state-of-the-art methods that use expensive manual segmentation-level annotations. In addition, our experiments also demonstrate that our 3D shape models improve object segmentation performance.
Persistent Identifierhttp://hdl.handle.net/10722/321920
ISSN
2023 Impact Factor: 20.8
2023 SCImago Journal Rankings: 6.158
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorHan, Junwei-
dc.contributor.authorYang, Yang-
dc.contributor.authorZhang, Dingwen-
dc.contributor.authorHuang, Dong-
dc.contributor.authorXu, Dong-
dc.contributor.authorDe La Torre, Fernando-
dc.date.accessioned2022-11-03T02:22:22Z-
dc.date.available2022-11-03T02:22:22Z-
dc.date.issued2021-
dc.identifier.citationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, v. 43, n. 4, p. 1423-1437-
dc.identifier.issn0162-8828-
dc.identifier.urihttp://hdl.handle.net/10722/321920-
dc.description.abstractCategory-specific 3D object shape models have greatly boosted the recent advances in object detection, recognition and segmentation. However, even the most advanced approach for learning 3D object shapes still requires heavy manual annotations on large-scale 2D images. Such annotations include object categories, object keypoints, and figure-ground segmentation for the instances in each image. In particular, annotating figure-ground segmentation is unbearably labor-intensive and time-consuming. To address this problem, this paper devotes to learn category-specific 3D shape models under weak supervision, where only object categories and keypoints are required to be manually annotated on the training 2D images. By exploring the underlying relationship between two tasks: object segmentation and category-specific 3D shape reconstruction, we propose a novel weakly-supervised learning framework to jointly address these two tasks and combine them to boost the final performance of the learned 3D shape models. Moreover, learning without using figure-ground segmentation leads to ambiguous solutions. To this end, we develop the confidence weighting schemes in the viewpoint estimation and 3D shape learning procedure. These schemes effectively reduce the confusion caused by the noisy data and thus increase the chances for recovering more reliable 3D object shapes. Comprehensive experiments on the challenging PASCAL VOC benchmark show that our framework achieves comparable performance with the state-of-the-art methods that use expensive manual segmentation-level annotations. In addition, our experiments also demonstrate that our 3D shape models improve object segmentation performance.-
dc.languageeng-
dc.relation.ispartofIEEE Transactions on Pattern Analysis and Machine Intelligence-
dc.subject3D shape reconstruction-
dc.subjectcommon object segmentation-
dc.subjectviewpoint estimation-
dc.titleWeakly-Supervised Learning of Category-Specific 3D Object Shapes-
dc.typeArticle-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1109/TPAMI.2019.2949562-
dc.identifier.pmid31670664-
dc.identifier.scopuseid_2-s2.0-85099725616-
dc.identifier.volume43-
dc.identifier.issue4-
dc.identifier.spage1423-
dc.identifier.epage1437-
dc.identifier.eissn1939-3539-
dc.identifier.isiWOS:000629017400001-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats