Novel data fusion approach for drowsiness detection

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

In this research, a novel method employing the fusion technique of data in and feature out for determining the wake to sleep transition and vice versa is described. Inputs of the process are the Electrooculogram (EOG) and Head Nodding (HN) signals. The method involves a direct fusion of two pairs of signals consisting of the combination of EOG and head nodding signals to generate three new sets of data in the form of Lissajous-like plots. These Lissajous-like plots are further analyzed to extract the feature vector to represent the two states of awake and asleep. In turn, the extracted feature set is used to determine the state of alertness and/or sleepiness. The effectiveness of the derived feature vector from the fusion approach is evaluated and compared against the non-fusion based method. Results show that the fusion based approach fared better than the non-fusion method. This has implications for the application of the developed method in determining driver drowsiness.

Original languageEnglish
Pages (from-to)48-55
Number of pages8
JournalInformation Technology Journal
Volume7
Issue number1
Publication statusPublished - 2008

Fingerprint

Data fusion
Sleep

Keywords

  • Data fusion
  • Electrooculogram
  • Lissajous-like plots

ASJC Scopus subject areas

  • Computer Science (miscellaneous)

Cite this

Novel data fusion approach for drowsiness detection. / Hussain, Aini; Bais, Badariah; Abdul Samad, Salina; Farshad Hendi, S.

In: Information Technology Journal, Vol. 7, No. 1, 2008, p. 48-55.

Research output: Contribution to journalArticle

@article{87adb8c809d440988c35411ff0b38387,
title = "Novel data fusion approach for drowsiness detection",
abstract = "In this research, a novel method employing the fusion technique of data in and feature out for determining the wake to sleep transition and vice versa is described. Inputs of the process are the Electrooculogram (EOG) and Head Nodding (HN) signals. The method involves a direct fusion of two pairs of signals consisting of the combination of EOG and head nodding signals to generate three new sets of data in the form of Lissajous-like plots. These Lissajous-like plots are further analyzed to extract the feature vector to represent the two states of awake and asleep. In turn, the extracted feature set is used to determine the state of alertness and/or sleepiness. The effectiveness of the derived feature vector from the fusion approach is evaluated and compared against the non-fusion based method. Results show that the fusion based approach fared better than the non-fusion method. This has implications for the application of the developed method in determining driver drowsiness.",
keywords = "Data fusion, Electrooculogram, Lissajous-like plots",
author = "Aini Hussain and Badariah Bais and {Abdul Samad}, Salina and {Farshad Hendi}, S.",
year = "2008",
language = "English",
volume = "7",
pages = "48--55",
journal = "Information Technology Journal",
issn = "1812-5638",
publisher = "Asian Network for Scientific Information",
number = "1",

}

TY - JOUR

T1 - Novel data fusion approach for drowsiness detection

AU - Hussain, Aini

AU - Bais, Badariah

AU - Abdul Samad, Salina

AU - Farshad Hendi, S.

PY - 2008

Y1 - 2008

N2 - In this research, a novel method employing the fusion technique of data in and feature out for determining the wake to sleep transition and vice versa is described. Inputs of the process are the Electrooculogram (EOG) and Head Nodding (HN) signals. The method involves a direct fusion of two pairs of signals consisting of the combination of EOG and head nodding signals to generate three new sets of data in the form of Lissajous-like plots. These Lissajous-like plots are further analyzed to extract the feature vector to represent the two states of awake and asleep. In turn, the extracted feature set is used to determine the state of alertness and/or sleepiness. The effectiveness of the derived feature vector from the fusion approach is evaluated and compared against the non-fusion based method. Results show that the fusion based approach fared better than the non-fusion method. This has implications for the application of the developed method in determining driver drowsiness.

AB - In this research, a novel method employing the fusion technique of data in and feature out for determining the wake to sleep transition and vice versa is described. Inputs of the process are the Electrooculogram (EOG) and Head Nodding (HN) signals. The method involves a direct fusion of two pairs of signals consisting of the combination of EOG and head nodding signals to generate three new sets of data in the form of Lissajous-like plots. These Lissajous-like plots are further analyzed to extract the feature vector to represent the two states of awake and asleep. In turn, the extracted feature set is used to determine the state of alertness and/or sleepiness. The effectiveness of the derived feature vector from the fusion approach is evaluated and compared against the non-fusion based method. Results show that the fusion based approach fared better than the non-fusion method. This has implications for the application of the developed method in determining driver drowsiness.

KW - Data fusion

KW - Electrooculogram

KW - Lissajous-like plots

UR - http://www.scopus.com/inward/record.url?scp=41549155127&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=41549155127&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:41549155127

VL - 7

SP - 48

EP - 55

JO - Information Technology Journal

JF - Information Technology Journal

SN - 1812-5638

IS - 1

ER -