SCC: Semantic Context Cascade for Efficient Action Detection

Fabian Caba Heilbron; Wayner Barrios; Victor Escorcia; Bernard Ghanem

doi:10.1109/CVPR.2017.338

SCC: Semantic Context Cascade for Efficient Action Detection

Fabian Caba Heilbron, Wayner Barrios, Victor Escorcia, Bernard Ghanem

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

88 Scopus citations

Abstract

Despite the recent advances in large-scale video analysis, action detection remains as one of the most challenging unsolved problems in computer vision. This snag is in part due to the large volume of data that needs to be analyzed to detect actions in videos. Existing approaches have mitigated the computational cost, but still, these methods lack rich high-level semantics that helps them to localize the actions quickly. In this paper, we introduce a Semantic Cascade Context (SCC) model that aims to detect action in long video sequences. By embracing semantic priors associated with human activities, SCC produces high-quality class-specific action proposals and prune unrelated activities in a cascade fashion. Experimental results in ActivityNet unveils that SCC achieves state-of-the-art performance for action detection while operating at real time.

Original language	English (US)
Title of host publication	2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	3175-3184
Number of pages	10
ISBN (Print)	9781538604571
DOIs	https://doi.org/10.1109/CVPR.2017.338
State	Published - Nov 9 2017

Bibliographical note

KAUST Repository Item: Exported on 2020-10-01
Acknowledgements: Research in this publication was supported by the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research.

Access to Document

10.1109/CVPR.2017.338

Cite this

@inproceedings{2ba262edd3e14136adff7d187bee7b05,

title = "SCC: Semantic Context Cascade for Efficient Action Detection",

abstract = "Despite the recent advances in large-scale video analysis, action detection remains as one of the most challenging unsolved problems in computer vision. This snag is in part due to the large volume of data that needs to be analyzed to detect actions in videos. Existing approaches have mitigated the computational cost, but still, these methods lack rich high-level semantics that helps them to localize the actions quickly. In this paper, we introduce a Semantic Cascade Context (SCC) model that aims to detect action in long video sequences. By embracing semantic priors associated with human activities, SCC produces high-quality class-specific action proposals and prune unrelated activities in a cascade fashion. Experimental results in ActivityNet unveils that SCC achieves state-of-the-art performance for action detection while operating at real time.",

author = "Heilbron, {Fabian Caba} and Wayner Barrios and Victor Escorcia and Bernard Ghanem",

note = "KAUST Repository Item: Exported on 2020-10-01 Acknowledgements: Research in this publication was supported by the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research.",

year = "2017",

month = nov,

day = "9",

doi = "10.1109/CVPR.2017.338",

language = "English (US)",

isbn = "9781538604571",

pages = "3175--3184",

booktitle = "2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

}

TY - GEN

T1 - SCC: Semantic Context Cascade for Efficient Action Detection

AU - Heilbron, Fabian Caba

AU - Barrios, Wayner

AU - Escorcia, Victor

AU - Ghanem, Bernard

N1 - KAUST Repository Item: Exported on 2020-10-01 Acknowledgements: Research in this publication was supported by the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research.

PY - 2017/11/9

Y1 - 2017/11/9

N2 - Despite the recent advances in large-scale video analysis, action detection remains as one of the most challenging unsolved problems in computer vision. This snag is in part due to the large volume of data that needs to be analyzed to detect actions in videos. Existing approaches have mitigated the computational cost, but still, these methods lack rich high-level semantics that helps them to localize the actions quickly. In this paper, we introduce a Semantic Cascade Context (SCC) model that aims to detect action in long video sequences. By embracing semantic priors associated with human activities, SCC produces high-quality class-specific action proposals and prune unrelated activities in a cascade fashion. Experimental results in ActivityNet unveils that SCC achieves state-of-the-art performance for action detection while operating at real time.

AB - Despite the recent advances in large-scale video analysis, action detection remains as one of the most challenging unsolved problems in computer vision. This snag is in part due to the large volume of data that needs to be analyzed to detect actions in videos. Existing approaches have mitigated the computational cost, but still, these methods lack rich high-level semantics that helps them to localize the actions quickly. In this paper, we introduce a Semantic Cascade Context (SCC) model that aims to detect action in long video sequences. By embracing semantic priors associated with human activities, SCC produces high-quality class-specific action proposals and prune unrelated activities in a cascade fashion. Experimental results in ActivityNet unveils that SCC achieves state-of-the-art performance for action detection while operating at real time.

UR - http://hdl.handle.net/10754/626983

UR - http://ieeexplore.ieee.org/document/8099821/

UR - http://www.scopus.com/inward/record.url?scp=85030220370&partnerID=8YFLogxK

U2 - 10.1109/CVPR.2017.338

DO - 10.1109/CVPR.2017.338

M3 - Conference contribution

SN - 9781538604571

SP - 3175

EP - 3184

BT - 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

PB - Institute of Electrical and Electronics Engineers (IEEE)

ER -

SCC: Semantic Context Cascade for Efficient Action Detection

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this