Title :
A new variable frame analysis method for speech recognition
Author :
Le Cerf, Philippe ; Van Compernolle, Dirk
Author_Institution :
ESAT, Katholieke Univ., Leuven, Heverlee, Belgium
Abstract :
Variable frame rate (VFR) analysis is a technique used in speech processing and recognition for discarding frames that are too much alike. The article introduces a new method for VFR. Instead of calculating the distance between frames, the norm of the derivative parameters is used in deciding to retain or to discard a frame, informal inspection of speech spectrograms shows that this new method puts more emphasis on the transient regions of the speech signal. Experimental results with a hidden Markov model (HMM) based system show that the new method outperforms the classical method.<>
Keywords :
hidden Markov models; speech processing; speech recognition; HMM; derivative parameters; experimental results; hidden Markov model; informal inspection; speech processing; speech recognition; speech signal; speech spectrograms; transient regions; variable frame analysis method; Hidden Markov models; Inspection; Neural networks; Signal processing algorithms; Spectral analysis; Spectrogram; Speech analysis; Speech processing; Speech recognition; Steady-state;
Journal_Title :
Signal Processing Letters, IEEE
