Publication detail

Detection of Alcohol in Speech Signal Using LF Model

SIGMUND, M. PROKEŠ, A. ZELINKA, P.

Original Title

Detection of Alcohol in Speech Signal Using LF Model

English Title

Detection of Alcohol in Speech Signal Using LF Model

Type

conference paper

Language

en

Original Abstract

Speech changes resulting from alcohol intoxication were investigated in recorded clean speech using the Liljencrant-Fant model of glottal pulses. For use in our experiments, a new collection of intoxicated speech at two levels of intoxication was created. All vowels were tested individually. In person independent experiments, analysis of vowel excitation appears to be a useful technique to provide evidence of alcohol consumption of over 1 per mille of blood-alcohol concentration.

English abstract

Speech changes resulting from alcohol intoxication were investigated in recorded clean speech using the Liljencrant-Fant model of glottal pulses. For use in our experiments, a new collection of intoxicated speech at two levels of intoxication was created. All vowels were tested individually. In person independent experiments, analysis of vowel excitation appears to be a useful technique to provide evidence of alcohol consumption of over 1 per mille of blood-alcohol concentration.

Keywords

Speech signal, LF model

RIV year

2010

Released

15.02.2010

Publisher

IASTED

Location

Calgary

ISBN

978-0-88986-818-2

Book

Proceedings of IASTED International Multi-Conference on Applied Informatics, part Artificial Intelligence and Applications

Pages from

193

Pages to

196

Pages count

4

Documents

BibTex


@inproceedings{BUT33823,
  author="Milan {Sigmund} and Aleš {Prokeš} and Petr {Zelinka}",
  title="Detection of Alcohol in Speech Signal Using LF Model",
  annote="Speech changes resulting from alcohol intoxication were investigated in recorded clean speech using the Liljencrant-Fant model of glottal pulses. For use in our experiments, a new collection of intoxicated speech at two levels of intoxication was created. All vowels were tested individually. In person independent experiments, analysis of vowel excitation appears to be a useful technique to provide evidence of alcohol consumption of over 1 per mille of blood-alcohol concentration.",
  address="IASTED",
  booktitle="Proceedings of IASTED International Multi-Conference on Applied Informatics, part Artificial Intelligence and Applications",
  chapter="33823",
  howpublished="print",
  institution="IASTED",
  year="2010",
  month="february",
  pages="193--196",
  publisher="IASTED",
  type="conference paper"
}