Ego-Motion Classification for Body-Worn Videos

Zhaoyi MENG; Javier SÁNCHEZ; Jean-Michel MOREL; Andrea L. BERTOZZI; P. Jeffrey BRANTINGHAM

doi:10.1007/978-3-319-91274-5_10

Ego-Motion Classification for Body-Worn Videos

Zhaoyi MENG
, Javier SÁNCHEZ
, Jean-Michel MOREL
, Andrea L. BERTOZZI^*
, P. Jeffrey BRANTINGHAM

^*Corresponding author for this work

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Research › peer-review

6 Citations (Scopus)

Abstract

Portable cameras record dynamic first-person video footage and these videos contain information on the motion of the individual to whom the camera is mounted, defined as ego. We address the task of discovering ego-motion from the video itself, without other external calibration information. We investigate the use of similarity transformations between successive video frames to extract signals reflecting ego-motions and their frequencies. We use novel graph-based unsupervised and semi-supervised learning algorithms to segment the video frames into different ego-motion categories. Our results show very accurate results on both choreographed test videos and ego-motion videos provided by the Los Angeles Police Department.

Original language	English
Title of host publication	Imaging, Vision and Learning Based on Optimization and PDEs
Editors	Xue-Cheng TAI, Egil BAE, Marius LYSAKER
Publisher	Springer, Cham
Pages	221-239
Number of pages	19
ISBN (Electronic)	9783319912745
ISBN (Print)	9783319912738
DOIs	https://doi.org/10.1007/978-3-319-91274-5_10
Publication status	Published - 2018
Externally published	Yes
Event	International conference on Imaging, Vision and Learning Based on Optimization and PDEs, IVLOPDE 2016 - Bergen, Norway Duration: 29 Aug 2016 → 2 Sept 2016

Publication series

Name	Mathematics and Visualization
ISSN (Print)	1612-3786
ISSN (Electronic)	2197-666X

Conference

Conference	International conference on Imaging, Vision and Learning Based on Optimization and PDEs, IVLOPDE 2016
Country/Territory	Norway
City	Bergen
Period	29/08/16 → 2/09/16

Bibliographical note

Publisher Copyright:
© Springer International Publishing AG 2018.

Funding

The work was supported by the ONR grant N00014-16-1-2119, NSF grant DMS-1737770, NSF grant DMS-1417674, FUI project Plein Phare by BPI-France and NIJ Grant 2014-R2-CX-0101.

Access to Document

10.1007/978-3-319-91274-5_10

Cite this

@inproceedings{50c32724d60f42e6ac40819897eec09b,

title = "Ego-Motion Classification for Body-Worn Videos",

abstract = "Portable cameras record dynamic first-person video footage and these videos contain information on the motion of the individual to whom the camera is mounted, defined as ego. We address the task of discovering ego-motion from the video itself, without other external calibration information. We investigate the use of similarity transformations between successive video frames to extract signals reflecting ego-motions and their frequencies. We use novel graph-based unsupervised and semi-supervised learning algorithms to segment the video frames into different ego-motion categories. Our results show very accurate results on both choreographed test videos and ego-motion videos provided by the Los Angeles Police Department.",

author = "Zhaoyi MENG and Javier S{\'A}NCHEZ and Jean-Michel MOREL and BERTOZZI, \{Andrea L.\} and BRANTINGHAM, \{P. Jeffrey\}",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2018.; International conference on Imaging, Vision and Learning Based on Optimization and PDEs, IVLOPDE 2016 ; Conference date: 29-08-2016 Through 02-09-2016",

year = "2018",

doi = "10.1007/978-3-319-91274-5\_10",

language = "English",

isbn = "9783319912738",

series = "Mathematics and Visualization",

publisher = "Springer, Cham",

pages = "221--239",

editor = "Xue-Cheng TAI and Egil BAE and Marius LYSAKER",

booktitle = "Imaging, Vision and Learning Based on Optimization and PDEs",

}

MENG, Z, SÁNCHEZ, J, MOREL, J-M, BERTOZZI, AL & BRANTINGHAM, PJ 2018, Ego-Motion Classification for Body-Worn Videos. in X-C TAI, E BAE & M LYSAKER (eds), Imaging, Vision and Learning Based on Optimization and PDEs. Mathematics and Visualization, Springer, Cham, pp. 221-239, International conference on Imaging, Vision and Learning Based on Optimization and PDEs, IVLOPDE 2016, Bergen, Norway, 29/08/16. https://doi.org/10.1007/978-3-319-91274-5_10

Ego-Motion Classification for Body-Worn Videos. / MENG, Zhaoyi; SÁNCHEZ, Javier; MOREL, Jean-Michel et al.
Imaging, Vision and Learning Based on Optimization and PDEs. ed. / Xue-Cheng TAI; Egil BAE; Marius LYSAKER. Springer, Cham, 2018. p. 221-239 (Mathematics and Visualization).

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Research › peer-review

TY - GEN

T1 - Ego-Motion Classification for Body-Worn Videos

AU - MENG, Zhaoyi

AU - SÁNCHEZ, Javier

AU - MOREL, Jean-Michel

AU - BERTOZZI, Andrea L.

AU - BRANTINGHAM, P. Jeffrey

PY - 2018

Y1 - 2018

N2 - Portable cameras record dynamic first-person video footage and these videos contain information on the motion of the individual to whom the camera is mounted, defined as ego. We address the task of discovering ego-motion from the video itself, without other external calibration information. We investigate the use of similarity transformations between successive video frames to extract signals reflecting ego-motions and their frequencies. We use novel graph-based unsupervised and semi-supervised learning algorithms to segment the video frames into different ego-motion categories. Our results show very accurate results on both choreographed test videos and ego-motion videos provided by the Los Angeles Police Department.

AB - Portable cameras record dynamic first-person video footage and these videos contain information on the motion of the individual to whom the camera is mounted, defined as ego. We address the task of discovering ego-motion from the video itself, without other external calibration information. We investigate the use of similarity transformations between successive video frames to extract signals reflecting ego-motions and their frequencies. We use novel graph-based unsupervised and semi-supervised learning algorithms to segment the video frames into different ego-motion categories. Our results show very accurate results on both choreographed test videos and ego-motion videos provided by the Los Angeles Police Department.

UR - https://www.scopus.com/pages/publications/85057477586

U2 - 10.1007/978-3-319-91274-5_10

DO - 10.1007/978-3-319-91274-5_10

M3 - Conference paper (refereed)

AN - SCOPUS:85057477586

SN - 9783319912738

T3 - Mathematics and Visualization

SP - 221

EP - 239

BT - Imaging, Vision and Learning Based on Optimization and PDEs

A2 - TAI, Xue-Cheng

A2 - BAE, Egil

A2 - LYSAKER, Marius

PB - Springer, Cham

T2 - International conference on Imaging, Vision and Learning Based on Optimization and PDEs, IVLOPDE 2016

Y2 - 29 August 2016 through 2 September 2016

ER -

Ego-Motion Classification for Body-Worn Videos

Abstract

Publication series

Conference

Bibliographical note

Funding

Access to Document

Other files and links

Fingerprint

Cite this