• DocumentCode
    3194293
  • Title

    A Valid Frequency Range Analysis of Throat Signal for Voice Command System

  • Author

    Jung, Young-Giu ; Han, Mun-sung ; Chung, Kyung-Yong ; Lee, Sang Jo

  • Author_Institution
    Multimodal & Human Interaction Lab., YM-NaeulTech, Incheon, South Korea
  • fYear
    2011
  • fDate
    26-29 April 2011
  • Firstpage
    1
  • Lastpage
    5
  • Abstract
    In this paper, a valid frequency range of throat signal analyses is presented. Compare to speech signals, the throat signals have many differences in the amount of information. In this context, targeting on throat signals to find valid frequency range is one of the most important analyses for the development of new breed of feature extraction algorithm. In this respect, using analysis similarity of throat signals feature distribution and correlation analysis of throat signals in the frequency domain, the frequency range can be found with effective and quick response time to be applied to a voice command recognition system using neck-microphone. As a result, the proposed frequency range is from 200Hz (center frequency) to 2001Hz (center frequency). For the neck-microphone voice command system using the proposed frequency range, a significant reduction in time complexity of 25% can be achieved without any loss in recognition rate.
  • Keywords
    computational complexity; feature extraction; microphones; speech recognition; wearable computers; correlation analysis; feature extraction algorithm; neck microphone; throat signal analyses; throat signals feature distribution; time complexity; valid frequency range analysis; voice command recognition system; Band pass filters; Correlation; Feature extraction; Finite impulse response filter; Indexes; Microphones; Speech recognition;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Information Science and Applications (ICISA), 2011 International Conference on
  • Conference_Location
    Jeju Island
  • Print_ISBN
    978-1-4244-9222-0
  • Electronic_ISBN
    978-1-4244-9223-7
  • Type

    conf

  • DOI
    10.1109/ICISA.2011.5772383
  • Filename
    5772383