Title :
Automatic word recognition based on second-order hidden Markov models
Author :
Mari, Jean-Francois ; Haton, Jean-Paul ; Kriouile, Abdelaziz
Author_Institution :
CRIN, Vandoeuvre les Nancy, France
fDate :
1/1/1997 12:00:00 AM
Abstract :
We propose an extension of the Viterbi algorithm that makes second-order hidden Markov models computationally efficient. A comparative study between first-order (HMM1s) and second-order Markov models (HMM2s) is carried out. Experimental results show that HMM2s provide a better state occupancy modeling and, alone, have performances comparable with HMM1s plus postprocessing
Keywords :
hidden Markov models; maximum likelihood estimation; speech processing; speech recognition; Viterbi algorithm; automatic word recognition; computationally efficient HMM; first-order Markov models; performance; postprocessing; second-order hidden Markov models; state occupancy modeling; Computational complexity; Gaussian distribution; Hidden Markov models; Random processes; Shape; Solid modeling; Speech processing; State estimation; Stochastic processes; Viterbi algorithm;
Journal_Title :
Speech and Audio Processing, IEEE Transactions on
