Conversion of Body Conducted Unvoiced Speech(Murmur) to Normal Speech Using Hidden Markov Model (HMM)

 
 
 
  • Abstract
  • Keywords
  • References
  • PDF
  • Abstract


    The main purpose of this paper is Conversion of  non- audible murmured voice into the normal speech using Hidden Markov Model(HMM).This non audible murmur voice NAM is a one type of murmured voice which can be delivered by a NAM microphone which is attached behind the speaker’s ear. The Hidden Markov Models(HMMs) are stochastic models of statistical learning .These are very useful in speech recognition .The point of the paper is to collect as much as data from the device and convert it into audible and clear data signal that can be used for further sensory based applications. Hence, having an insight of how to convert the NAM to speech and then to whisper has a lot of benefits while keeping in mind the disadvantages of such conversion. Since, NAM is minute details of a communication between one’s own self it is highly recommended to the data in as much as discrete format as necessary since a speech signal can have various frequencies over a portion of the signal, big data approach is recommended.

     


  • Keywords


    Non audible murmur, NAM microphone, Hidden Markov Model, Speech Signals.

  • References


      [1] BDenby, TSchultz, KHonda, THueber, JMGilbert, and JS.Brumberg, “Silent speech interfaces,” Speech Communication., vol52, no4,pp270–287, 2010.

      [2] T.Rajeshkumar and G.R.Suresh “Examination of militants utiliz-ing NAM microphone and wireless handset for murmured speech in a view of concealed markov model.” IIRJET In 2017.

      [3] S.-CJou, TSchultz, and AWaibel, “Adaptation for soft whisper recognition using a throat microphone,” in ProcINTER-SPEECH, Jeju Island, Korea, Sep2004, pp1493–1496.

      [4] TSchultz and MWand, “ Modeling coarticulation in EMG-based continuous speech recognition,” Speech Communication., vol52, no4, pp.341–353, 2010.

      [5] THueber, E.-LBenaroya, GChollet, BDenby, GDreyfus, and M.Stone, “Development of a silent speech interface driven by ultrasound and optical images of the tongue and lips,” Speech Communication., vol52, no4, pp288–300, 2010.

      [6] ASubramanya, ZZhang, ZLiu, and AAcero,“Multisensory pro-cessing for speech enhancement and magnitude-normalized spectra for Speech modeling,”SpeechCommun.,vol.50,no.3,pp.228–243,2008.

      [7] Y.Nakajima,H.Kashioka,N.Cambell,andK.Shika no.”Non-audible murmur (NAM) recognition,” IEICE TransInfSyst., volE89-d, no1, pp1-8,2006

      [8] Panikos Heracleous,Tomomi Kaino,Hiroshi Saruwatari and Kiyohiro Shika-no”Unvoiced speech recognition using Tissue conductive acoustic sensor” In 2006

      [9] T.Toda,K..Nakamura,T.Nagai,T.Kanio,Y.Nakajima, and KShi-kano, “Technologies for processing body-conducted speech detect-ed with non-audible murmur microphone,” in proc.INTERSPEECH, Brighton, U.K., Sep2009, pp,632-635.

      [10] Yamato Ohtani and Tomkoi “non parallel preparing for nu-merous to numerous Eigen voice change” IEEEIn 2010.


 

View

Download

Article ID: 15723
 
DOI: 10.14419/ijet.v7i2.32.15723




Copyright © 2012-2015 Science Publishing Corporation Inc. All rights reserved.