• DocumentCode
    1020823
  • Title

    Speaker adaptation via VQ prototype modification

  • Author

    Rtischev, D. ; Nahamoo, David ; Picheny, Michael

  • Author_Institution
    IBM Thomas J. Watson Res. Center, Yorktown Heights, NY, USA
  • Volume
    2
  • Issue
    1
  • fYear
    1994
  • Firstpage
    94
  • Lastpage
    97
  • Abstract
    A statistical technique for vector quantizer (VQ) prototype adaptation, based on tied-mixture continuous-parameter HMM´s, is derived and evaluated on the basis of experimental evidence. Performance on difficult adaptation tasks indicates that VQ-prototype adaptation via tied-mixture HMM´s constitutes a useful mechanism for speaker adaptation, particularly when there are substantial channel differences or when there is a large mismatch between reference and target speaker characteristics.
  • Keywords
    decoding; hidden Markov models; speech coding; speech recognition; vector quantisation; VQ prototype modification; decoder; hidden Markov model; recognition performance; speaker adaptation; speaker characteristics mismatch; speech recognition system; statistical technique; tied-mixture continuous-parameter HMM; vector quantizer; Clustering algorithms; Decoding; Hidden Markov models; Iterative algorithms; Labeling; Loudspeakers; Maximum likelihood estimation; Modems; Prototypes; Speech recognition;
  • fLanguage
    English
  • Journal_Title
    Speech and Audio Processing, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    1063-6676
  • Type

    jour

  • DOI
    10.1109/89.260340
  • Filename
    260340