Emotional expression classification using time-series kernels

A. Lőrincz, Laszlo Attila Jeni, Zoltan Szabo, Jeffrey F. Cohn, Takeo Kanade

Research output: Chapter in Book/Report/Conference proceedingConference contribution

32 Citations (Scopus)

Abstract

Estimation of facial expressions, as spatio-temporal processes, can take advantage of kernel methods if one considers facial landmark positions and their motion in 3D space. We applied support vector classification with kernels derived from dynamic time-warping similarity measures. We achieved over 99% accuracy-measured by area under ROC curve-using only the 'motion pattern' of the PCA compressed representation of the marker point vector, the so-called shape parameters. Beyond the classification of full motion patterns, several expressions were recognized with over 90% accuracy in as few as 5-6 frames from their onset, about 200 milliseconds.

Original languageEnglish
Title of host publicationIEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops
Pages889-895
Number of pages7
DOIs
Publication statusPublished - 2013
Event2013 IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2013 - Portland, OR, United States
Duration: Jun 23 2013Jun 28 2013

Other

Other2013 IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2013
CountryUnited States
CityPortland, OR
Period6/23/136/28/13

Fingerprint

Time series

Keywords

  • 3d shape
  • dynamic time warping kernel
  • emotional expression classification
  • global alignment kernel
  • time-series

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Electrical and Electronic Engineering

Cite this

Lőrincz, A., Jeni, L. A., Szabo, Z., Cohn, J. F., & Kanade, T. (2013). Emotional expression classification using time-series kernels. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (pp. 889-895). [6595976] https://doi.org/10.1109/CVPRW.2013.131

Emotional expression classification using time-series kernels. / Lőrincz, A.; Jeni, Laszlo Attila; Szabo, Zoltan; Cohn, Jeffrey F.; Kanade, Takeo.

IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops. 2013. p. 889-895 6595976.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Lőrincz, A, Jeni, LA, Szabo, Z, Cohn, JF & Kanade, T 2013, Emotional expression classification using time-series kernels. in IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops., 6595976, pp. 889-895, 2013 IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2013, Portland, OR, United States, 6/23/13. https://doi.org/10.1109/CVPRW.2013.131
Lőrincz A, Jeni LA, Szabo Z, Cohn JF, Kanade T. Emotional expression classification using time-series kernels. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops. 2013. p. 889-895. 6595976 https://doi.org/10.1109/CVPRW.2013.131
Lőrincz, A. ; Jeni, Laszlo Attila ; Szabo, Zoltan ; Cohn, Jeffrey F. ; Kanade, Takeo. / Emotional expression classification using time-series kernels. IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops. 2013. pp. 889-895
@inproceedings{7c33bef0a71b4cd5b3bd2d4a26424064,
title = "Emotional expression classification using time-series kernels",
abstract = "Estimation of facial expressions, as spatio-temporal processes, can take advantage of kernel methods if one considers facial landmark positions and their motion in 3D space. We applied support vector classification with kernels derived from dynamic time-warping similarity measures. We achieved over 99{\%} accuracy-measured by area under ROC curve-using only the 'motion pattern' of the PCA compressed representation of the marker point vector, the so-called shape parameters. Beyond the classification of full motion patterns, several expressions were recognized with over 90{\%} accuracy in as few as 5-6 frames from their onset, about 200 milliseconds.",
keywords = "3d shape, dynamic time warping kernel, emotional expression classification, global alignment kernel, time-series",
author = "A. Lőrincz and Jeni, {Laszlo Attila} and Zoltan Szabo and Cohn, {Jeffrey F.} and Takeo Kanade",
year = "2013",
doi = "10.1109/CVPRW.2013.131",
language = "English",
isbn = "9780769549903",
pages = "889--895",
booktitle = "IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops",

}

TY - GEN

T1 - Emotional expression classification using time-series kernels

AU - Lőrincz, A.

AU - Jeni, Laszlo Attila

AU - Szabo, Zoltan

AU - Cohn, Jeffrey F.

AU - Kanade, Takeo

PY - 2013

Y1 - 2013

N2 - Estimation of facial expressions, as spatio-temporal processes, can take advantage of kernel methods if one considers facial landmark positions and their motion in 3D space. We applied support vector classification with kernels derived from dynamic time-warping similarity measures. We achieved over 99% accuracy-measured by area under ROC curve-using only the 'motion pattern' of the PCA compressed representation of the marker point vector, the so-called shape parameters. Beyond the classification of full motion patterns, several expressions were recognized with over 90% accuracy in as few as 5-6 frames from their onset, about 200 milliseconds.

AB - Estimation of facial expressions, as spatio-temporal processes, can take advantage of kernel methods if one considers facial landmark positions and their motion in 3D space. We applied support vector classification with kernels derived from dynamic time-warping similarity measures. We achieved over 99% accuracy-measured by area under ROC curve-using only the 'motion pattern' of the PCA compressed representation of the marker point vector, the so-called shape parameters. Beyond the classification of full motion patterns, several expressions were recognized with over 90% accuracy in as few as 5-6 frames from their onset, about 200 milliseconds.

KW - 3d shape

KW - dynamic time warping kernel

KW - emotional expression classification

KW - global alignment kernel

KW - time-series

UR - http://www.scopus.com/inward/record.url?scp=84884918508&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84884918508&partnerID=8YFLogxK

U2 - 10.1109/CVPRW.2013.131

DO - 10.1109/CVPRW.2013.131

M3 - Conference contribution

AN - SCOPUS:84884918508

SN - 9780769549903

SP - 889

EP - 895

BT - IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops

ER -