Fast Temporal Activity Proposals for Efficient Detection of Human Actions in Untrimmed Videos

Handle URI:
http://hdl.handle.net/10754/622892
Title:
Fast Temporal Activity Proposals for Efficient Detection of Human Actions in Untrimmed Videos
Authors:
Heilbron, Fabian Caba; Niebles, Juan Carlos; Ghanem, Bernard ( 0000-0002-5534-587X )
Abstract:
In many large-scale video analysis scenarios, one is interested in localizing and recognizing human activities that occur in short temporal intervals within long untrimmed videos. Current approaches for activity detection still struggle to handle large-scale video collections and the task remains relatively unexplored. This is in part due to the computational complexity of current action recognition approaches and the lack of a method that proposes fewer intervals in the video, where activity processing can be focused. In this paper, we introduce a proposal method that aims to recover temporal segments containing actions in untrimmed videos. Building on techniques for learning sparse dictionaries, we introduce a learning framework to represent and retrieve activity proposals. We demonstrate the capabilities of our method in not only producing high quality proposals but also in its efficiency. Finally, we show the positive impact our method has on recognition performance when it is used for action detection, while running at 10FPS.
KAUST Department:
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division; Electrical Engineering Program; Visual Computing Center (VCC)
Citation:
Heilbron FC, Niebles JC, Ghanem B (2016) Fast Temporal Activity Proposals for Efficient Detection of Human Actions in Untrimmed Videos. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Available: http://dx.doi.org/10.1109/CVPR.2016.211.
Publisher:
Institute of Electrical and Electronics Engineers (IEEE)
Journal:
2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
Issue Date:
13-Dec-2016
DOI:
10.1109/CVPR.2016.211
Type:
Conference Paper
Sponsors:
Research in this publication was supported by the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research, the Stanford AI Lab-Toyota Center for Artificial Intelligence Research, and a Google Faculty Research Award (2015).
Additional Links:
http://ieeexplore.ieee.org/document/7780580/
Appears in Collections:
Conference Papers; Electrical Engineering Program; Visual Computing Center (VCC); Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division

Full metadata record

DC FieldValue Language
dc.contributor.authorHeilbron, Fabian Cabaen
dc.contributor.authorNiebles, Juan Carlosen
dc.contributor.authorGhanem, Bernarden
dc.date.accessioned2017-02-15T08:32:14Z-
dc.date.available2017-02-15T08:32:14Z-
dc.date.issued2016-12-13en
dc.identifier.citationHeilbron FC, Niebles JC, Ghanem B (2016) Fast Temporal Activity Proposals for Efficient Detection of Human Actions in Untrimmed Videos. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Available: http://dx.doi.org/10.1109/CVPR.2016.211.en
dc.identifier.doi10.1109/CVPR.2016.211en
dc.identifier.urihttp://hdl.handle.net/10754/622892-
dc.description.abstractIn many large-scale video analysis scenarios, one is interested in localizing and recognizing human activities that occur in short temporal intervals within long untrimmed videos. Current approaches for activity detection still struggle to handle large-scale video collections and the task remains relatively unexplored. This is in part due to the computational complexity of current action recognition approaches and the lack of a method that proposes fewer intervals in the video, where activity processing can be focused. In this paper, we introduce a proposal method that aims to recover temporal segments containing actions in untrimmed videos. Building on techniques for learning sparse dictionaries, we introduce a learning framework to represent and retrieve activity proposals. We demonstrate the capabilities of our method in not only producing high quality proposals but also in its efficiency. Finally, we show the positive impact our method has on recognition performance when it is used for action detection, while running at 10FPS.en
dc.description.sponsorshipResearch in this publication was supported by the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research, the Stanford AI Lab-Toyota Center for Artificial Intelligence Research, and a Google Faculty Research Award (2015).en
dc.publisherInstitute of Electrical and Electronics Engineers (IEEE)en
dc.relation.urlhttp://ieeexplore.ieee.org/document/7780580/en
dc.subjectfeature extractionen
dc.subjectimage recognitionen
dc.subjectimage representationen
dc.subjectimage retrievalen
dc.subjectDictionariesen
dc.subjectFeature extractionen
dc.subjectImage reconstructionen
dc.subjectProposalsen
dc.subjectTrainingen
dc.subjectVideo sequencesen
dc.subjectVideosen
dc.titleFast Temporal Activity Proposals for Efficient Detection of Human Actions in Untrimmed Videosen
dc.typeConference Paperen
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Divisionen
dc.contributor.departmentElectrical Engineering Programen
dc.contributor.departmentVisual Computing Center (VCC)en
dc.identifier.journal2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)en
dc.contributor.institutionDepartment of Computer Science, Stanford Universityen
dc.contributor.institutionUniversidad del Norte, Colombiaen
kaust.authorHeilbron, Fabian Cabaen
kaust.authorGhanem, Bernarden
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.