Learning slow features for behavior analysis

Lazaros Zafeiriou, Mihalis A. Nicolaou, Stefanos Zafeiriou, Symeon Nikitids, Maja Pantic

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

17 Citations (Scopus)
23 Downloads (Pure)

Abstract

A recently introduced latent feature learning technique for time varying dynamic phenomena analysis is the socalled Slow Feature Analysis (SFA). SFA is a deterministic component analysis technique for multi-dimensional sequences that by minimizing the variance of the first order time derivative approximation of the input signal finds uncorrelated projections that extract slowly-varying features ordered by their temporal consistency and constancy. In this paper, we propose a number of extensions in both the deterministic and the probabilistic SFA optimization frameworks. In articular, we derive a novel deterministic SFA algorithm that is able to identify linear projections that extract the common slowest varying features of two or more sequences. In addition, we propose an Expectation Maximization (EM) algorithm to perform inference in a probabilistic ormulation of SFA and similarly extend it in order to handle two and more time varying data sequences. Moreover, we demonstrate that the probabilistic SFA (EMSFA) algorithm that discovers the common slowest varying latent space of multiple sequences can be combined with dynamic time warping chniques for robust sequence timealignment. The proposed SFA algorithms were applied for facial behavior analysis demonstrating their usefulness and appropriateness for this task.
Original languageUndefined
Title of host publicationProceedings of the IEEE International Conference on Computer Vision, ICCV 2013
Place of PublicationUSA
PublisherIEEE Computer Society
Pages2840-2847
Number of pages8
ISBN (Print)1550-5499
DOIs
Publication statusPublished - 1 Dec 2013
EventIEEE International Conference on Computer Vision 2013 - Sydney Conference Centre, Sydney, Australia
Duration: 1 Dec 20138 Dec 2013
http://www.pamitc.org/iccv13/

Publication series

NameIEEE International Conference on Computer Vision
PublisherIEEE Computer Society
ISSN (Print)1550-5499

Conference

ConferenceIEEE International Conference on Computer Vision 2013
Abbreviated titleICCV 2013
CountryAustralia
CitySydney
Period1/12/138/12/13
Internet address

Keywords

  • HMI-HF: Human Factors
  • EC Grant Agreement nr.: FP7/288235
  • METIS-302865
  • EWI-24261
  • IR-89697
  • EC Grant Agreement nr.: FP7/2007-2013

Cite this

Zafeiriou, L., Nicolaou, M. A., Zafeiriou, S., Nikitids, S., & Pantic, M. (2013). Learning slow features for behavior analysis. In Proceedings of the IEEE International Conference on Computer Vision, ICCV 2013 (pp. 2840-2847). (IEEE International Conference on Computer Vision). USA: IEEE Computer Society. https://doi.org/10.1109/ICCV.2013.353
Zafeiriou, Lazaros ; Nicolaou, Mihalis A. ; Zafeiriou, Stefanos ; Nikitids, Symeon ; Pantic, Maja. / Learning slow features for behavior analysis. Proceedings of the IEEE International Conference on Computer Vision, ICCV 2013. USA : IEEE Computer Society, 2013. pp. 2840-2847 (IEEE International Conference on Computer Vision).
@inproceedings{2915458ccdfd457e94f77be560f1acbf,
title = "Learning slow features for behavior analysis",
abstract = "A recently introduced latent feature learning technique for time varying dynamic phenomena analysis is the socalled Slow Feature Analysis (SFA). SFA is a deterministic component analysis technique for multi-dimensional sequences that by minimizing the variance of the first order time derivative approximation of the input signal finds uncorrelated projections that extract slowly-varying features ordered by their temporal consistency and constancy. In this paper, we propose a number of extensions in both the deterministic and the probabilistic SFA optimization frameworks. In articular, we derive a novel deterministic SFA algorithm that is able to identify linear projections that extract the common slowest varying features of two or more sequences. In addition, we propose an Expectation Maximization (EM) algorithm to perform inference in a probabilistic ormulation of SFA and similarly extend it in order to handle two and more time varying data sequences. Moreover, we demonstrate that the probabilistic SFA (EMSFA) algorithm that discovers the common slowest varying latent space of multiple sequences can be combined with dynamic time warping chniques for robust sequence timealignment. The proposed SFA algorithms were applied for facial behavior analysis demonstrating their usefulness and appropriateness for this task.",
keywords = "HMI-HF: Human Factors, EC Grant Agreement nr.: FP7/288235, METIS-302865, EWI-24261, IR-89697, EC Grant Agreement nr.: FP7/2007-2013",
author = "Lazaros Zafeiriou and Nicolaou, {Mihalis A.} and Stefanos Zafeiriou and Symeon Nikitids and Maja Pantic",
note = "eemcs-eprint-24261",
year = "2013",
month = "12",
day = "1",
doi = "10.1109/ICCV.2013.353",
language = "Undefined",
isbn = "1550-5499",
series = "IEEE International Conference on Computer Vision",
publisher = "IEEE Computer Society",
pages = "2840--2847",
booktitle = "Proceedings of the IEEE International Conference on Computer Vision, ICCV 2013",
address = "United States",

}

Zafeiriou, L, Nicolaou, MA, Zafeiriou, S, Nikitids, S & Pantic, M 2013, Learning slow features for behavior analysis. in Proceedings of the IEEE International Conference on Computer Vision, ICCV 2013. IEEE International Conference on Computer Vision, IEEE Computer Society, USA, pp. 2840-2847, IEEE International Conference on Computer Vision 2013, Sydney, Australia, 1/12/13. https://doi.org/10.1109/ICCV.2013.353

Learning slow features for behavior analysis. / Zafeiriou, Lazaros; Nicolaou, Mihalis A.; Zafeiriou, Stefanos; Nikitids, Symeon; Pantic, Maja.

Proceedings of the IEEE International Conference on Computer Vision, ICCV 2013. USA : IEEE Computer Society, 2013. p. 2840-2847 (IEEE International Conference on Computer Vision).

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Learning slow features for behavior analysis

AU - Zafeiriou, Lazaros

AU - Nicolaou, Mihalis A.

AU - Zafeiriou, Stefanos

AU - Nikitids, Symeon

AU - Pantic, Maja

N1 - eemcs-eprint-24261

PY - 2013/12/1

Y1 - 2013/12/1

N2 - A recently introduced latent feature learning technique for time varying dynamic phenomena analysis is the socalled Slow Feature Analysis (SFA). SFA is a deterministic component analysis technique for multi-dimensional sequences that by minimizing the variance of the first order time derivative approximation of the input signal finds uncorrelated projections that extract slowly-varying features ordered by their temporal consistency and constancy. In this paper, we propose a number of extensions in both the deterministic and the probabilistic SFA optimization frameworks. In articular, we derive a novel deterministic SFA algorithm that is able to identify linear projections that extract the common slowest varying features of two or more sequences. In addition, we propose an Expectation Maximization (EM) algorithm to perform inference in a probabilistic ormulation of SFA and similarly extend it in order to handle two and more time varying data sequences. Moreover, we demonstrate that the probabilistic SFA (EMSFA) algorithm that discovers the common slowest varying latent space of multiple sequences can be combined with dynamic time warping chniques for robust sequence timealignment. The proposed SFA algorithms were applied for facial behavior analysis demonstrating their usefulness and appropriateness for this task.

AB - A recently introduced latent feature learning technique for time varying dynamic phenomena analysis is the socalled Slow Feature Analysis (SFA). SFA is a deterministic component analysis technique for multi-dimensional sequences that by minimizing the variance of the first order time derivative approximation of the input signal finds uncorrelated projections that extract slowly-varying features ordered by their temporal consistency and constancy. In this paper, we propose a number of extensions in both the deterministic and the probabilistic SFA optimization frameworks. In articular, we derive a novel deterministic SFA algorithm that is able to identify linear projections that extract the common slowest varying features of two or more sequences. In addition, we propose an Expectation Maximization (EM) algorithm to perform inference in a probabilistic ormulation of SFA and similarly extend it in order to handle two and more time varying data sequences. Moreover, we demonstrate that the probabilistic SFA (EMSFA) algorithm that discovers the common slowest varying latent space of multiple sequences can be combined with dynamic time warping chniques for robust sequence timealignment. The proposed SFA algorithms were applied for facial behavior analysis demonstrating their usefulness and appropriateness for this task.

KW - HMI-HF: Human Factors

KW - EC Grant Agreement nr.: FP7/288235

KW - METIS-302865

KW - EWI-24261

KW - IR-89697

KW - EC Grant Agreement nr.: FP7/2007-2013

U2 - 10.1109/ICCV.2013.353

DO - 10.1109/ICCV.2013.353

M3 - Conference contribution

SN - 1550-5499

T3 - IEEE International Conference on Computer Vision

SP - 2840

EP - 2847

BT - Proceedings of the IEEE International Conference on Computer Vision, ICCV 2013

PB - IEEE Computer Society

CY - USA

ER -

Zafeiriou L, Nicolaou MA, Zafeiriou S, Nikitids S, Pantic M. Learning slow features for behavior analysis. In Proceedings of the IEEE International Conference on Computer Vision, ICCV 2013. USA: IEEE Computer Society. 2013. p. 2840-2847. (IEEE International Conference on Computer Vision). https://doi.org/10.1109/ICCV.2013.353