음성인식을 위한 은닉마코프모형 연구

  • 손건태 (부산대학교 통계학과) ;
  • 정상화 (부산대학교 컴퓨터공학과) ;
  • 박민욱 (부산대학교 컴퓨터공학과)
  • Published : 1998.04.01

Abstract

음성자동인식을 위한 통계적 방법으로 은닉마코프모형이 널리 사용되고 있다. 이산형 은닉마코프모형보다 인식률이 우수한 연속형 은닉마코프모형을 고려하였으며, 인식을 위한 비터비(Viterbi) 알고리즘을 병렬화시켜 인식속도를 빠르게 하는 인식 알고리즘을 제안하였다. 제안된 방법으로 실험을 통하여 인식률과 인식속도 개선률(speed-up)을 살펴보았다.

Keywords

References

  1. Inequalities v.3 An inequality and associated maximization technique in statistical estimation for probabilistic functions of Markov chains Baum, L.E.
  2. Journal of Royal Statistician Society v.39 no.1 Maximum likelihood from incomplete data via the EM algorithm Dempster, A.P.;Laird, N.M.;Rubin, D.B.
  3. Technometrics v.33 Hidden Markov models for speech recognition Juang, B.H.;Rabiner, L.R.
  4. IEEE transactions on information theory v.IT-13 Error bounds for convolutional codes and an asymptotically optimal decoding algorithm Viterbi, A.J.
  5. Annals of Mathematical Statistics v.37 Statistical inference for probabilistic functions of finite state Markov chains Baum, L.E.;Petrie, T.
  6. AT&T Technical Journal v.64 Maximum likelihood estimation for mixture mutivariate stochastic observations of Markov chain Juang, B.H.
  7. IEEE Transactions on Acoustics, Speech, Signal processing v.ASSP-23 The dragon system - an overview Baker, J.K.
  8. Annals of Mathematical Statistics v.41 A maximization technique occurring in the statistical analysis of probabilistic functions of Markov chains Baum, L.E.;Petrie, T.;Soules, G.;Weiss, N.
  9. IEEE ASSP Magazine An introduction to hidden Markov models Rabiner, L.R.;Juang, B.H.
  10. Thesis, Alparon report Parallel Implementation of Hidden Markov Models on the nCUBE2 Huijsen, G.
  11. Korean journal of computational and applied mathematics v.4 Estimation in a mixture normal distribution Sohn, K.T.;Baik, J.S.
  12. IEEE Transactions on Information theory v.IT-21 Decoding for channels with insertions, deletions and substitutions with applications to speech recognition Bahl, L.R.;Jelinek, F.
  13. IEEE Transactions on Pattern Analysis and Machine Intelligence v.PAMI-5 A maximum likelihood approach to continous speech recognition Bahl, L.R.;Jelinek, F.;Mercer, R.L.
  14. Digital signal Processing-87 Speaker independent isolated word recognition based on continous hidden Markov models using mutidimensional spherically invariant functions Euler, S.;Wolf, D.
  15. Fundamentals of Speech Recognition Rabiner, L.R.;Juang, B.H.
  16. IEEE Transactions on Information Theory v.IT-21 Design of a linguisitc statistical decoder for the recognition of continous speech Jelinek, F.;Bahl, L.R.;Mercer, R.L.