A human action classifier from 4-D data (3-D+TIME): Based on an invariant body shape descriptor and hidden markov models

Massimiliano Pierobon, Marco Marcon, Augusto Sarti, Stefano Tubaro

Research output: Contribution to conferencePaper

1 Citation (Scopus)

Abstract

Many human action definitions have been provided in the field of human computer interaction studies. These distinctions could be considered merely semantical as human actions are all carried out performing sequences of body postures. In this paper we propose a human action classifier based on volumetric reconstructed sequences (4-D data) acquired from a multi-viewpoint camera system. In order to design the most general action classifier possible, we concentrate our attention in extracting only posture-dependent information from volumetric frames and in performing action distinction only on the basis of the sequence of body postures carried out in the scene. An Invariant Shape Descriptor (ISD) is used in order to properly describe the body shape and its dynamic changes during an action execution. The ISD data is then analyzed in order to extract suitable features able to meaningfully represent a human action independently from body position, orientation, size and proportions. The action classification is performed using a supervised recognizer based on the Hidden Markov Models (HMM) theory. Experimental results, evaluated using an extensive action sequence dataset and applying different training conditions to the HMM-based classifier, confirm the reliability of the proposed approach.

Original languageEnglish (US)
Pages406-413
Number of pages8
StatePublished - Dec 1 2007
EventSIGMAP 2007 - International Conference on Signal Processing and Multimedia Applications - Barcelona, Spain
Duration: Jul 28 2007Jul 31 2007

Other

OtherSIGMAP 2007 - International Conference on Signal Processing and Multimedia Applications
CountrySpain
CityBarcelona
Period7/28/077/31/07

Fingerprint

Hidden Markov models
Classifiers
Human computer interaction
Cameras

Keywords

  • Action classification
  • Action recognition
  • Computer vision
  • Gesture classification
  • Gesture recognition
  • Human machine interaction
  • Human motion analysis
  • Multiple view volumetric reconstruction
  • Video surveillance
  • Voxel based representation of human body

ASJC Scopus subject areas

  • Computer Science Applications
  • Signal Processing
  • Electrical and Electronic Engineering
  • Media Technology

Cite this

Pierobon, M., Marcon, M., Sarti, A., & Tubaro, S. (2007). A human action classifier from 4-D data (3-D+TIME): Based on an invariant body shape descriptor and hidden markov models. 406-413. Paper presented at SIGMAP 2007 - International Conference on Signal Processing and Multimedia Applications, Barcelona, Spain.

A human action classifier from 4-D data (3-D+TIME) : Based on an invariant body shape descriptor and hidden markov models. / Pierobon, Massimiliano; Marcon, Marco; Sarti, Augusto; Tubaro, Stefano.

2007. 406-413 Paper presented at SIGMAP 2007 - International Conference on Signal Processing and Multimedia Applications, Barcelona, Spain.

Research output: Contribution to conferencePaper

Pierobon, M, Marcon, M, Sarti, A & Tubaro, S 2007, 'A human action classifier from 4-D data (3-D+TIME): Based on an invariant body shape descriptor and hidden markov models', Paper presented at SIGMAP 2007 - International Conference on Signal Processing and Multimedia Applications, Barcelona, Spain, 7/28/07 - 7/31/07 pp. 406-413.
Pierobon M, Marcon M, Sarti A, Tubaro S. A human action classifier from 4-D data (3-D+TIME): Based on an invariant body shape descriptor and hidden markov models. 2007. Paper presented at SIGMAP 2007 - International Conference on Signal Processing and Multimedia Applications, Barcelona, Spain.
Pierobon, Massimiliano ; Marcon, Marco ; Sarti, Augusto ; Tubaro, Stefano. / A human action classifier from 4-D data (3-D+TIME) : Based on an invariant body shape descriptor and hidden markov models. Paper presented at SIGMAP 2007 - International Conference on Signal Processing and Multimedia Applications, Barcelona, Spain.8 p.
@conference{71c846de290f4287ac78feffe6000867,
title = "A human action classifier from 4-D data (3-D+TIME): Based on an invariant body shape descriptor and hidden markov models",
abstract = "Many human action definitions have been provided in the field of human computer interaction studies. These distinctions could be considered merely semantical as human actions are all carried out performing sequences of body postures. In this paper we propose a human action classifier based on volumetric reconstructed sequences (4-D data) acquired from a multi-viewpoint camera system. In order to design the most general action classifier possible, we concentrate our attention in extracting only posture-dependent information from volumetric frames and in performing action distinction only on the basis of the sequence of body postures carried out in the scene. An Invariant Shape Descriptor (ISD) is used in order to properly describe the body shape and its dynamic changes during an action execution. The ISD data is then analyzed in order to extract suitable features able to meaningfully represent a human action independently from body position, orientation, size and proportions. The action classification is performed using a supervised recognizer based on the Hidden Markov Models (HMM) theory. Experimental results, evaluated using an extensive action sequence dataset and applying different training conditions to the HMM-based classifier, confirm the reliability of the proposed approach.",
keywords = "Action classification, Action recognition, Computer vision, Gesture classification, Gesture recognition, Human machine interaction, Human motion analysis, Multiple view volumetric reconstruction, Video surveillance, Voxel based representation of human body",
author = "Massimiliano Pierobon and Marco Marcon and Augusto Sarti and Stefano Tubaro",
year = "2007",
month = "12",
day = "1",
language = "English (US)",
pages = "406--413",
note = "SIGMAP 2007 - International Conference on Signal Processing and Multimedia Applications ; Conference date: 28-07-2007 Through 31-07-2007",

}

TY - CONF

T1 - A human action classifier from 4-D data (3-D+TIME)

T2 - Based on an invariant body shape descriptor and hidden markov models

AU - Pierobon, Massimiliano

AU - Marcon, Marco

AU - Sarti, Augusto

AU - Tubaro, Stefano

PY - 2007/12/1

Y1 - 2007/12/1

N2 - Many human action definitions have been provided in the field of human computer interaction studies. These distinctions could be considered merely semantical as human actions are all carried out performing sequences of body postures. In this paper we propose a human action classifier based on volumetric reconstructed sequences (4-D data) acquired from a multi-viewpoint camera system. In order to design the most general action classifier possible, we concentrate our attention in extracting only posture-dependent information from volumetric frames and in performing action distinction only on the basis of the sequence of body postures carried out in the scene. An Invariant Shape Descriptor (ISD) is used in order to properly describe the body shape and its dynamic changes during an action execution. The ISD data is then analyzed in order to extract suitable features able to meaningfully represent a human action independently from body position, orientation, size and proportions. The action classification is performed using a supervised recognizer based on the Hidden Markov Models (HMM) theory. Experimental results, evaluated using an extensive action sequence dataset and applying different training conditions to the HMM-based classifier, confirm the reliability of the proposed approach.

AB - Many human action definitions have been provided in the field of human computer interaction studies. These distinctions could be considered merely semantical as human actions are all carried out performing sequences of body postures. In this paper we propose a human action classifier based on volumetric reconstructed sequences (4-D data) acquired from a multi-viewpoint camera system. In order to design the most general action classifier possible, we concentrate our attention in extracting only posture-dependent information from volumetric frames and in performing action distinction only on the basis of the sequence of body postures carried out in the scene. An Invariant Shape Descriptor (ISD) is used in order to properly describe the body shape and its dynamic changes during an action execution. The ISD data is then analyzed in order to extract suitable features able to meaningfully represent a human action independently from body position, orientation, size and proportions. The action classification is performed using a supervised recognizer based on the Hidden Markov Models (HMM) theory. Experimental results, evaluated using an extensive action sequence dataset and applying different training conditions to the HMM-based classifier, confirm the reliability of the proposed approach.

KW - Action classification

KW - Action recognition

KW - Computer vision

KW - Gesture classification

KW - Gesture recognition

KW - Human machine interaction

KW - Human motion analysis

KW - Multiple view volumetric reconstruction

KW - Video surveillance

KW - Voxel based representation of human body

UR - http://www.scopus.com/inward/record.url?scp=70349443696&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70349443696&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:70349443696

SP - 406

EP - 413

ER -