File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: Referring Image Segmentation via Recurrent Refinement Networks

TitleReferring Image Segmentation via Recurrent Refinement Networks
Authors
Issue Date2018
Citation
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2018, p. 5745-5753 How to Cite?
Abstract© 2018 IEEE. We address the problem of image segmentation from natural language descriptions. Existing deep learning-based methods encode image representations based on the output of the last convolutional layer. One general issue is that the resulting image representation lacks multi-scale semantics, which are key components in advanced segmentation systems. In this paper, we utilize the feature pyramids inherently existing in convolutional neural networks to capture the semantics at different scales. To produce suitable information flow through the path of feature hierarchy, we propose Recurrent Refinement Network (RRN) that takes pyramidal features as input to refine the segmentation mask progressively. Experimental results on four available datasets show that our approach outperforms multiple baselines and state-of-the-art1.
Persistent Identifierhttp://hdl.handle.net/10722/281969
ISSN

 

DC FieldValueLanguage
dc.contributor.authorLi, Ruiyu-
dc.contributor.authorLi, Kaican-
dc.contributor.authorKuo, Yi Chun-
dc.contributor.authorShu, Michelle-
dc.contributor.authorQi, Xiaojuan-
dc.contributor.authorShen, Xiaoyong-
dc.contributor.authorJia, Jiaya-
dc.date.accessioned2020-04-09T09:19:16Z-
dc.date.available2020-04-09T09:19:16Z-
dc.date.issued2018-
dc.identifier.citationProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2018, p. 5745-5753-
dc.identifier.issn1063-6919-
dc.identifier.urihttp://hdl.handle.net/10722/281969-
dc.description.abstract© 2018 IEEE. We address the problem of image segmentation from natural language descriptions. Existing deep learning-based methods encode image representations based on the output of the last convolutional layer. One general issue is that the resulting image representation lacks multi-scale semantics, which are key components in advanced segmentation systems. In this paper, we utilize the feature pyramids inherently existing in convolutional neural networks to capture the semantics at different scales. To produce suitable information flow through the path of feature hierarchy, we propose Recurrent Refinement Network (RRN) that takes pyramidal features as input to refine the segmentation mask progressively. Experimental results on four available datasets show that our approach outperforms multiple baselines and state-of-the-art1.-
dc.languageeng-
dc.relation.ispartofProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition-
dc.titleReferring Image Segmentation via Recurrent Refinement Networks-
dc.typeConference_Paper-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1109/CVPR.2018.00602-
dc.identifier.scopuseid_2-s2.0-85062859929-
dc.identifier.spage5745-
dc.identifier.epage5753-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats