Human gesture recognition using a simplified dynamic Bayesian network

Myung Cheol Roh, Seong Whan Lee

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

In video-based human gesture recognition, it is very important to combine useful features and analyze the dynamic structure thereof as efficiently as possible. In this paper, we proposed a dynamic Bayesian network model that is a simplified model of dynamics at the level of hidden variables and employs observation windows of observation time slices for robust modeling and handling of noise and other variabilities. The proposed Simplified dynamic Bayesian network (DBN) was tested on a gesture database and an American sign language database. According to the experiments, the proposed DBN outperformed other methods: Conditional Random Fields (CRFs), conventional Bayesian Networks (BNs), DBNs, and Hidden Markov Models (HMMs). The proposed DBN achieved 98 % recognition accuracy in gesture recognition and 94.6 % in ASL recognition whereas the HMM and the CRF did 80 and 86 % in gesture recognition and 75.4 and 85.4 % in ASL (American Sign Language) recognition, respectively.

Original languageEnglish
Pages (from-to)557-568
Number of pages12
JournalMultimedia Systems
Volume21
Issue number6
DOIs
Publication statusPublished - 2014 Oct 9

Fingerprint

Gesture recognition
Bayesian networks
Hidden Markov models

Keywords

  • Dynamic Bayesian network
  • Gesture recognition
  • Sign language recognition

ASJC Scopus subject areas

  • Media Technology
  • Hardware and Architecture
  • Information Systems
  • Software
  • Computer Networks and Communications

Cite this

Human gesture recognition using a simplified dynamic Bayesian network. / Roh, Myung Cheol; Lee, Seong Whan.

In: Multimedia Systems, Vol. 21, No. 6, 09.10.2014, p. 557-568.

Research output: Contribution to journalArticle

@article{bb45027002924aa392458a41ed38a44f,
title = "Human gesture recognition using a simplified dynamic Bayesian network",
abstract = "In video-based human gesture recognition, it is very important to combine useful features and analyze the dynamic structure thereof as efficiently as possible. In this paper, we proposed a dynamic Bayesian network model that is a simplified model of dynamics at the level of hidden variables and employs observation windows of observation time slices for robust modeling and handling of noise and other variabilities. The proposed Simplified dynamic Bayesian network (DBN) was tested on a gesture database and an American sign language database. According to the experiments, the proposed DBN outperformed other methods: Conditional Random Fields (CRFs), conventional Bayesian Networks (BNs), DBNs, and Hidden Markov Models (HMMs). The proposed DBN achieved 98 {\%} recognition accuracy in gesture recognition and 94.6 {\%} in ASL recognition whereas the HMM and the CRF did 80 and 86 {\%} in gesture recognition and 75.4 and 85.4 {\%} in ASL (American Sign Language) recognition, respectively.",
keywords = "Dynamic Bayesian network, Gesture recognition, Sign language recognition",
author = "Roh, {Myung Cheol} and Lee, {Seong Whan}",
year = "2014",
month = "10",
day = "9",
doi = "10.1007/s00530-014-0414-9",
language = "English",
volume = "21",
pages = "557--568",
journal = "Multimedia Systems",
issn = "0942-4962",
publisher = "Springer Verlag",
number = "6",

}

TY - JOUR

T1 - Human gesture recognition using a simplified dynamic Bayesian network

AU - Roh, Myung Cheol

AU - Lee, Seong Whan

PY - 2014/10/9

Y1 - 2014/10/9

N2 - In video-based human gesture recognition, it is very important to combine useful features and analyze the dynamic structure thereof as efficiently as possible. In this paper, we proposed a dynamic Bayesian network model that is a simplified model of dynamics at the level of hidden variables and employs observation windows of observation time slices for robust modeling and handling of noise and other variabilities. The proposed Simplified dynamic Bayesian network (DBN) was tested on a gesture database and an American sign language database. According to the experiments, the proposed DBN outperformed other methods: Conditional Random Fields (CRFs), conventional Bayesian Networks (BNs), DBNs, and Hidden Markov Models (HMMs). The proposed DBN achieved 98 % recognition accuracy in gesture recognition and 94.6 % in ASL recognition whereas the HMM and the CRF did 80 and 86 % in gesture recognition and 75.4 and 85.4 % in ASL (American Sign Language) recognition, respectively.

AB - In video-based human gesture recognition, it is very important to combine useful features and analyze the dynamic structure thereof as efficiently as possible. In this paper, we proposed a dynamic Bayesian network model that is a simplified model of dynamics at the level of hidden variables and employs observation windows of observation time slices for robust modeling and handling of noise and other variabilities. The proposed Simplified dynamic Bayesian network (DBN) was tested on a gesture database and an American sign language database. According to the experiments, the proposed DBN outperformed other methods: Conditional Random Fields (CRFs), conventional Bayesian Networks (BNs), DBNs, and Hidden Markov Models (HMMs). The proposed DBN achieved 98 % recognition accuracy in gesture recognition and 94.6 % in ASL recognition whereas the HMM and the CRF did 80 and 86 % in gesture recognition and 75.4 and 85.4 % in ASL (American Sign Language) recognition, respectively.

KW - Dynamic Bayesian network

KW - Gesture recognition

KW - Sign language recognition

UR - http://www.scopus.com/inward/record.url?scp=84942193195&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84942193195&partnerID=8YFLogxK

U2 - 10.1007/s00530-014-0414-9

DO - 10.1007/s00530-014-0414-9

M3 - Article

AN - SCOPUS:84942193195

VL - 21

SP - 557

EP - 568

JO - Multimedia Systems

JF - Multimedia Systems

SN - 0942-4962

IS - 6

ER -