Clustering of human actions using invariant body shape descriptor and dynamic time warping

Massimiliano Pierobon, Marco Marcon, Augusto Sarti, Stefano Tubaro

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

16 Citations (Scopus)

Abstract

We propose a human action clustering method based on a 3D representation of the body in terms of volumetric coordinates. Features representing body postures are extracted directly from 3D data, making the system inherently insensitive to viewpoint dependence, motion ambiguities and self-occlusions. An Invariant Shape Descriptor of human body is obtained in order to capture only posture-dependent characteristics, despite possible differences in translation, orientation, scale and body size. Frame-by-frame descriptions, generated from a gesture sequence, are collected together in matrices. Clustering of action matrices is eventually performed, and through a Dynamic Time Warping (while computing the distance metric), we gain independence from possible temporal nonlinear distortions among different instances of the same gesture.

Original languageEnglish (US)
Title of host publicationIEEE Conference on Advanced Video and Signal Based Based Surveillance - Proceedings of AVSS 2005
Pages22-27
Number of pages6
DOIs
StatePublished - Dec 1 2005
EventIEEE Conference on Advanced Video and Signal Based Surveillance, AVSS 2005 - Como, Italy
Duration: Sep 15 2005Sep 16 2005

Publication series

NameIEEE International Conference on Advanced Video and Signal Based Surveillance - Proceedings of AVSS 2005
Volume2005

Other

OtherIEEE Conference on Advanced Video and Signal Based Surveillance, AVSS 2005
CountryItaly
CityComo
Period9/15/059/16/05

Fingerprint

Nonlinear distortion

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Pierobon, M., Marcon, M., Sarti, A., & Tubaro, S. (2005). Clustering of human actions using invariant body shape descriptor and dynamic time warping. In IEEE Conference on Advanced Video and Signal Based Based Surveillance - Proceedings of AVSS 2005 (pp. 22-27). [1577237] (IEEE International Conference on Advanced Video and Signal Based Surveillance - Proceedings of AVSS 2005; Vol. 2005). https://doi.org/10.1109/AVSS.2005.1577237

Clustering of human actions using invariant body shape descriptor and dynamic time warping. / Pierobon, Massimiliano; Marcon, Marco; Sarti, Augusto; Tubaro, Stefano.

IEEE Conference on Advanced Video and Signal Based Based Surveillance - Proceedings of AVSS 2005. 2005. p. 22-27 1577237 (IEEE International Conference on Advanced Video and Signal Based Surveillance - Proceedings of AVSS 2005; Vol. 2005).

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

Pierobon, M, Marcon, M, Sarti, A & Tubaro, S 2005, Clustering of human actions using invariant body shape descriptor and dynamic time warping. in IEEE Conference on Advanced Video and Signal Based Based Surveillance - Proceedings of AVSS 2005., 1577237, IEEE International Conference on Advanced Video and Signal Based Surveillance - Proceedings of AVSS 2005, vol. 2005, pp. 22-27, IEEE Conference on Advanced Video and Signal Based Surveillance, AVSS 2005, Como, Italy, 9/15/05. https://doi.org/10.1109/AVSS.2005.1577237
Pierobon M, Marcon M, Sarti A, Tubaro S. Clustering of human actions using invariant body shape descriptor and dynamic time warping. In IEEE Conference on Advanced Video and Signal Based Based Surveillance - Proceedings of AVSS 2005. 2005. p. 22-27. 1577237. (IEEE International Conference on Advanced Video and Signal Based Surveillance - Proceedings of AVSS 2005). https://doi.org/10.1109/AVSS.2005.1577237
Pierobon, Massimiliano ; Marcon, Marco ; Sarti, Augusto ; Tubaro, Stefano. / Clustering of human actions using invariant body shape descriptor and dynamic time warping. IEEE Conference on Advanced Video and Signal Based Based Surveillance - Proceedings of AVSS 2005. 2005. pp. 22-27 (IEEE International Conference on Advanced Video and Signal Based Surveillance - Proceedings of AVSS 2005).
@inproceedings{64db69a395c1490bb2fd3d30303dd30d,
title = "Clustering of human actions using invariant body shape descriptor and dynamic time warping",
abstract = "We propose a human action clustering method based on a 3D representation of the body in terms of volumetric coordinates. Features representing body postures are extracted directly from 3D data, making the system inherently insensitive to viewpoint dependence, motion ambiguities and self-occlusions. An Invariant Shape Descriptor of human body is obtained in order to capture only posture-dependent characteristics, despite possible differences in translation, orientation, scale and body size. Frame-by-frame descriptions, generated from a gesture sequence, are collected together in matrices. Clustering of action matrices is eventually performed, and through a Dynamic Time Warping (while computing the distance metric), we gain independence from possible temporal nonlinear distortions among different instances of the same gesture.",
author = "Massimiliano Pierobon and Marco Marcon and Augusto Sarti and Stefano Tubaro",
year = "2005",
month = "12",
day = "1",
doi = "10.1109/AVSS.2005.1577237",
language = "English (US)",
isbn = "0780393856",
series = "IEEE International Conference on Advanced Video and Signal Based Surveillance - Proceedings of AVSS 2005",
pages = "22--27",
booktitle = "IEEE Conference on Advanced Video and Signal Based Based Surveillance - Proceedings of AVSS 2005",

}

TY - GEN

T1 - Clustering of human actions using invariant body shape descriptor and dynamic time warping

AU - Pierobon, Massimiliano

AU - Marcon, Marco

AU - Sarti, Augusto

AU - Tubaro, Stefano

PY - 2005/12/1

Y1 - 2005/12/1

N2 - We propose a human action clustering method based on a 3D representation of the body in terms of volumetric coordinates. Features representing body postures are extracted directly from 3D data, making the system inherently insensitive to viewpoint dependence, motion ambiguities and self-occlusions. An Invariant Shape Descriptor of human body is obtained in order to capture only posture-dependent characteristics, despite possible differences in translation, orientation, scale and body size. Frame-by-frame descriptions, generated from a gesture sequence, are collected together in matrices. Clustering of action matrices is eventually performed, and through a Dynamic Time Warping (while computing the distance metric), we gain independence from possible temporal nonlinear distortions among different instances of the same gesture.

AB - We propose a human action clustering method based on a 3D representation of the body in terms of volumetric coordinates. Features representing body postures are extracted directly from 3D data, making the system inherently insensitive to viewpoint dependence, motion ambiguities and self-occlusions. An Invariant Shape Descriptor of human body is obtained in order to capture only posture-dependent characteristics, despite possible differences in translation, orientation, scale and body size. Frame-by-frame descriptions, generated from a gesture sequence, are collected together in matrices. Clustering of action matrices is eventually performed, and through a Dynamic Time Warping (while computing the distance metric), we gain independence from possible temporal nonlinear distortions among different instances of the same gesture.

UR - http://www.scopus.com/inward/record.url?scp=33846992276&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33846992276&partnerID=8YFLogxK

U2 - 10.1109/AVSS.2005.1577237

DO - 10.1109/AVSS.2005.1577237

M3 - Conference contribution

AN - SCOPUS:33846992276

SN - 0780393856

SN - 9780780393851

T3 - IEEE International Conference on Advanced Video and Signal Based Surveillance - Proceedings of AVSS 2005

SP - 22

EP - 27

BT - IEEE Conference on Advanced Video and Signal Based Based Surveillance - Proceedings of AVSS 2005

ER -