GOTOH AND SILVERMAN : ANALYSIS OF LPC / DFT FEATURES 105 Nasal - set Confusion

نویسندگان

  • Daniel J. Mashao
  • Yoshihiko Gotoh
  • Harvey F. Silverman
چکیده

| The search for better and more robust performance of speech recognition systems is ongoing. Much of the improvement is likely to come from better acoustic feature analysis. In this letter, the results from a signiicant experiment are reported; these show how a warped-DFT analysis outperforms an LPC-cepstral analysis in a signiicant way, supporting results by other researchers for diierent recognition tasks. An analysis of nasal-letter performance is used to show the development of the warped-DFT feature analysis.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Using MAP estimated parameters to improve HMM speech recognition performance

RECOGNITION PERFORMANCE Yoshihiko Gotoh 1 Michael M. Hochberg 2 Harvey F. Silverman 1 1 LEMS, Division of Engineering, Brown University, Providence, RI 02912 USA 2 Cambridge University Engineering Department, Trumpington Street, Cambridge CB2 1PZ UK ABSTRACT Hidden Markov models (HMMs) have been quite successfully applied to speech recognition tasks, but many unsolved problems still remain. HMM...

متن کامل

Incremental ML estimation of HMM parameters for efficient training

Conventional training of a hidden Markov model (HMM) is performed by an expectation-maximization algorithm using a maximum likelihood (ML) criterion. Recently it was reported that, using an incremental variant of maximum a posteriori estimation, substantial speed improvements could be obtained. The approach requires a prior distribution when the training starts, although it is diicult to nd an ...

متن کامل

Efficient training algorithms for HMMs using incremental estimation

Typically, parameter estimation for a hidden Markov model (HMM) is performed using an expectation-maximization (EM) algorithm with the maximum-likelihood (ML) criterion. The EM algorithm is an iterative scheme which is well-deened and numerically stable, but convergence may require a large number of iterations. For speech recognition systems utilizing large amounts of training material, this re...

متن کامل

Microphone-array speech recognition via incremental map training

For a hidden Markov model (HMM) based speech recognition system it is desirable to combine enhancement of the acoustical signal and statistical representation of model parameters , ensuring both a high quality speech signal and an appropriately trained HMM. In this paper the incre-mental variant of maximum a posteriori (MAP) estimation is used to adjust the parameters of a talker-independent HM...

متن کامل

Gains from diversification on convex combinations: A majorization and stochastic dominance approach

By incorporating both majorization theory and stochastic dominance theory, this paper presents a general theory and a unifying framework for determining the diversification preferences of risk-averse investors and conditions under which they would unanimously judge a particular asset to be superior. In particular, we develop a theory for comparing the preferences of different convex combination...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1995