Egocentric visual event classification with location-based priors

Sudeep Sundaram; Walterio Mayol-Cuevas

doi:10.1007/978-3-642-17274-8

Egocentric visual event classification with location-based priors

Sudeep Sundaram, Walterio Mayol-Cuevas

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

7 Citations (Scopus)

462 Downloads (Pure)

Abstract

We present a method for visual classification of actions and events captured from an egocentric point of view. The method tackles the challenge of a moving camera by creating deformable graph models for classification of actions. Action models are learned from low resolution, roughly stabilized difference images acquired using a single monocular camera. In parallel, raw images from the camera are used to estimate the user’s location using a visual Simultaneous Localization and Mapping (SLAM) system. Action-location priors, learned using a labelled set of locations, further aid action classification and bring events into context. We present results on a dataset collected within a cluttered environment, consisting of routine manipulations performed on objects without tags.

Translated title of the contribution	Egocentric Visual Event Classification with Location-Based Priors
Original language	English
Title of host publication	International Symposium on Visual Computing
Publisher	Springer
Pages	596-605
Number of pages	10
ISBN (Electronic)	9783642172748
ISBN (Print)	9783642172731
DOIs	https://doi.org/10.1007/978-3-642-17274-8
Publication status	Published - 10 Nov 2010
Event	6th International Symposium on Visual Computing - Las Vegas, United States Duration: 10 Nov 2010 → …

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer Berlin Heidelberg
Volume	6454
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	6th International Symposium on Visual Computing
Country/Territory	United States
City	Las Vegas
Period	10/11/10 → …

Bibliographical note

Other page information: -
Conference Proceedings/Title of Journal: International Symposium on Visual Computing
Other identifier: 2001256
Rose publication type: Conference contribution

Additional information: Author's own post-print version of a paper published by Springer in volume 6454 of the Lecture Notes in Computer Science series. The original publication is available at www.springerlink.com

Sponsorship: The authors are deeply grateful to the British Council for the PhD studentship granted to SS,
and to the EUFP7 COGNITO project for partially funding WMC

ISSN: 0302-9743 (print), 1611-3349 (online)

Keywords

event recognition
SLAM
wearable computing
simultaneous localization and mapping

Access to Document

10.1007/978-3-642-17274-8

SundaramMayol ISVC10Accepted author manuscript, 1.46 MB

Handle.net

Persistent link

Cite this

@inproceedings{0cbca16f2c114d7d9e8cc52690c4c2c0,

title = "Egocentric visual event classification with location-based priors",

abstract = "We present a method for visual classification of actions and events captured from an egocentric point of view. The method tackles the challenge of a moving camera by creating deformable graph models for classification of actions. Action models are learned from low resolution, roughly stabilized difference images acquired using a single monocular camera. In parallel, raw images from the camera are used to estimate the user{\textquoteright}s location using a visual Simultaneous Localization and Mapping (SLAM) system. Action-location priors, learned using a labelled set of locations, further aid action classification and bring events into context. We present results on a dataset collected within a cluttered environment, consisting of routine manipulations performed on objects without tags.",

keywords = "event recognition, SLAM, wearable computing, simultaneous localization and mapping",

author = "Sudeep Sundaram and Walterio Mayol-Cuevas",

note = "Other page information: - Conference Proceedings/Title of Journal: International Symposium on Visual Computing Other identifier: 2001256 Rose publication type: Conference contribution Additional information: Author's own post-print version of a paper published by Springer in volume 6454 of the Lecture Notes in Computer Science series. The original publication is available at www.springerlink.com Sponsorship: The authors are deeply grateful to the British Council for the PhD studentship granted to SS, and to the EUFP7 COGNITO project for partially funding WMC ISSN: 0302-9743 (print), 1611-3349 (online); 6th International Symposium on Visual Computing ; Conference date: 10-11-2010",

year = "2010",

month = nov,

day = "10",

doi = "10.1007/978-3-642-17274-8",

language = "English",

isbn = "9783642172731",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "596--605",

booktitle = "International Symposium on Visual Computing",

}

TY - GEN

T1 - Egocentric visual event classification with location-based priors

AU - Sundaram, Sudeep

AU - Mayol-Cuevas, Walterio

N1 - Other page information: - Conference Proceedings/Title of Journal: International Symposium on Visual Computing Other identifier: 2001256 Rose publication type: Conference contribution Additional information: Author's own post-print version of a paper published by Springer in volume 6454 of the Lecture Notes in Computer Science series. The original publication is available at www.springerlink.com Sponsorship: The authors are deeply grateful to the British Council for the PhD studentship granted to SS, and to the EUFP7 COGNITO project for partially funding WMC ISSN: 0302-9743 (print), 1611-3349 (online)

PY - 2010/11/10

Y1 - 2010/11/10

N2 - We present a method for visual classification of actions and events captured from an egocentric point of view. The method tackles the challenge of a moving camera by creating deformable graph models for classification of actions. Action models are learned from low resolution, roughly stabilized difference images acquired using a single monocular camera. In parallel, raw images from the camera are used to estimate the user’s location using a visual Simultaneous Localization and Mapping (SLAM) system. Action-location priors, learned using a labelled set of locations, further aid action classification and bring events into context. We present results on a dataset collected within a cluttered environment, consisting of routine manipulations performed on objects without tags.

AB - We present a method for visual classification of actions and events captured from an egocentric point of view. The method tackles the challenge of a moving camera by creating deformable graph models for classification of actions. Action models are learned from low resolution, roughly stabilized difference images acquired using a single monocular camera. In parallel, raw images from the camera are used to estimate the user’s location using a visual Simultaneous Localization and Mapping (SLAM) system. Action-location priors, learned using a labelled set of locations, further aid action classification and bring events into context. We present results on a dataset collected within a cluttered environment, consisting of routine manipulations performed on objects without tags.

KW - event recognition

KW - SLAM

KW - wearable computing

KW - simultaneous localization and mapping

U2 - 10.1007/978-3-642-17274-8

DO - 10.1007/978-3-642-17274-8

M3 - Conference Contribution (Conference Proceeding)

SN - 9783642172731

T3 - Lecture Notes in Computer Science

SP - 596

EP - 605

BT - International Symposium on Visual Computing

PB - Springer

T2 - 6th International Symposium on Visual Computing

Y2 - 10 November 2010

ER -

Egocentric visual event classification with location-based priors

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Handle.net

Fingerprint

Cite this