Human action recognition using shape and CLG-motion flow from multi-view image sequences

Mohiuddin Ahmad, Seong Whan Lee

Research output: Contribution to journalArticle

108 Citations (Scopus)

Abstract

In this paper, we present a method for human action recognition from multi-view image sequences that uses the combined motion and shape flow information with variability consideration. A combined local-global (CLG) optic flow is used to extract motion flow feature and invariant moments with flow deviations are used to extract the global shape flow feature from the image sequences. In our approach, human action is represented as a set of multidimensional CLG optic flow and shape flow feature vectors in the spatial-temporal action boundary. Actions are modeled by using a set of multidimensional HMMs for multiple views using the combined features, which enforce robust view-invariant operation. We recognize different human actions in daily life successfully in the indoor and outdoor environment using the maximum likelihood estimation approach. The results suggest robustness of the proposed method with respect to multiple views action recognition, scale and phase variations, and invariant analysis of silhouettes.

Original languageEnglish
Pages (from-to)2237-2252
Number of pages16
JournalPattern Recognition
Volume41
Issue number7
DOIs
Publication statusPublished - 2008 Jul 1

Fingerprint

Optics
Maximum likelihood estimation

Keywords

  • Action matrix
  • Action recognition
  • Combined local-global (CLG) optic flow
  • Invariant Zernike moments
  • Multi-view image sequence
  • Multidimensional hidden Markov model (MDHMM)

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Human action recognition using shape and CLG-motion flow from multi-view image sequences. / Ahmad, Mohiuddin; Lee, Seong Whan.

In: Pattern Recognition, Vol. 41, No. 7, 01.07.2008, p. 2237-2252.

Research output: Contribution to journalArticle

@article{a0d91cbc940647d8bbb195aad5b7dade,
title = "Human action recognition using shape and CLG-motion flow from multi-view image sequences",
abstract = "In this paper, we present a method for human action recognition from multi-view image sequences that uses the combined motion and shape flow information with variability consideration. A combined local-global (CLG) optic flow is used to extract motion flow feature and invariant moments with flow deviations are used to extract the global shape flow feature from the image sequences. In our approach, human action is represented as a set of multidimensional CLG optic flow and shape flow feature vectors in the spatial-temporal action boundary. Actions are modeled by using a set of multidimensional HMMs for multiple views using the combined features, which enforce robust view-invariant operation. We recognize different human actions in daily life successfully in the indoor and outdoor environment using the maximum likelihood estimation approach. The results suggest robustness of the proposed method with respect to multiple views action recognition, scale and phase variations, and invariant analysis of silhouettes.",
keywords = "Action matrix, Action recognition, Combined local-global (CLG) optic flow, Invariant Zernike moments, Multi-view image sequence, Multidimensional hidden Markov model (MDHMM)",
author = "Mohiuddin Ahmad and Lee, {Seong Whan}",
year = "2008",
month = "7",
day = "1",
doi = "10.1016/j.patcog.2007.12.008",
language = "English",
volume = "41",
pages = "2237--2252",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",
number = "7",

}

TY - JOUR

T1 - Human action recognition using shape and CLG-motion flow from multi-view image sequences

AU - Ahmad, Mohiuddin

AU - Lee, Seong Whan

PY - 2008/7/1

Y1 - 2008/7/1

N2 - In this paper, we present a method for human action recognition from multi-view image sequences that uses the combined motion and shape flow information with variability consideration. A combined local-global (CLG) optic flow is used to extract motion flow feature and invariant moments with flow deviations are used to extract the global shape flow feature from the image sequences. In our approach, human action is represented as a set of multidimensional CLG optic flow and shape flow feature vectors in the spatial-temporal action boundary. Actions are modeled by using a set of multidimensional HMMs for multiple views using the combined features, which enforce robust view-invariant operation. We recognize different human actions in daily life successfully in the indoor and outdoor environment using the maximum likelihood estimation approach. The results suggest robustness of the proposed method with respect to multiple views action recognition, scale and phase variations, and invariant analysis of silhouettes.

AB - In this paper, we present a method for human action recognition from multi-view image sequences that uses the combined motion and shape flow information with variability consideration. A combined local-global (CLG) optic flow is used to extract motion flow feature and invariant moments with flow deviations are used to extract the global shape flow feature from the image sequences. In our approach, human action is represented as a set of multidimensional CLG optic flow and shape flow feature vectors in the spatial-temporal action boundary. Actions are modeled by using a set of multidimensional HMMs for multiple views using the combined features, which enforce robust view-invariant operation. We recognize different human actions in daily life successfully in the indoor and outdoor environment using the maximum likelihood estimation approach. The results suggest robustness of the proposed method with respect to multiple views action recognition, scale and phase variations, and invariant analysis of silhouettes.

KW - Action matrix

KW - Action recognition

KW - Combined local-global (CLG) optic flow

KW - Invariant Zernike moments

KW - Multi-view image sequence

KW - Multidimensional hidden Markov model (MDHMM)

UR - http://www.scopus.com/inward/record.url?scp=41549136442&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=41549136442&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2007.12.008

DO - 10.1016/j.patcog.2007.12.008

M3 - Article

AN - SCOPUS:41549136442

VL - 41

SP - 2237

EP - 2252

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

IS - 7

ER -