File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: Unsupervised moving object detection via contextual information separation

TitleUnsupervised moving object detection via contextual information separation
Authors
KeywordsGrouping and Shape
Representation Learning
Scene Analysis and Understanding
Segmentation
Statistical Learning
Issue Date2019
Citation
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2019, v. 2019-June, p. 879-888 How to Cite?
AbstractWe propose an adversarial contextual model for detecting moving objects in images. A deep neural network is trained to predict the optical flow in a region using information from everywhere else but that region (context), while another network attempts to make such context as uninformative as possible. The result is a model where hypotheses naturally compete with no need for explicit regularization or hyper-parameter tuning. Although our method requires no supervision whatsoever, it outperforms several methods that are pre-trained on large annotated datasets. Our model can be thought of as a generalization of classical variational generative region-based segmentation, but in a way that avoids explicit regularization or solution of partial differential equations at run-time.
Persistent Identifierhttp://hdl.handle.net/10722/325447
ISSN
2023 SCImago Journal Rankings: 10.331
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorYang, Yanchao-
dc.contributor.authorLoquercio, Antonio-
dc.contributor.authorScaramuzza, Davide-
dc.contributor.authorSoatto, Stefano-
dc.date.accessioned2023-02-27T07:33:24Z-
dc.date.available2023-02-27T07:33:24Z-
dc.date.issued2019-
dc.identifier.citationProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2019, v. 2019-June, p. 879-888-
dc.identifier.issn1063-6919-
dc.identifier.urihttp://hdl.handle.net/10722/325447-
dc.description.abstractWe propose an adversarial contextual model for detecting moving objects in images. A deep neural network is trained to predict the optical flow in a region using information from everywhere else but that region (context), while another network attempts to make such context as uninformative as possible. The result is a model where hypotheses naturally compete with no need for explicit regularization or hyper-parameter tuning. Although our method requires no supervision whatsoever, it outperforms several methods that are pre-trained on large annotated datasets. Our model can be thought of as a generalization of classical variational generative region-based segmentation, but in a way that avoids explicit regularization or solution of partial differential equations at run-time.-
dc.languageeng-
dc.relation.ispartofProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition-
dc.subjectGrouping and Shape-
dc.subjectRepresentation Learning-
dc.subjectScene Analysis and Understanding-
dc.subjectSegmentation-
dc.subjectStatistical Learning-
dc.titleUnsupervised moving object detection via contextual information separation-
dc.typeConference_Paper-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1109/CVPR.2019.00097-
dc.identifier.scopuseid_2-s2.0-85073466198-
dc.identifier.volume2019-June-
dc.identifier.spage879-
dc.identifier.epage888-
dc.identifier.isiWOS:000529484001003-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats