Title :
Dynamic Gaussian selection technique for speeding up HMM-based continuous speech recognition
Author :
Cai, Jun ; Bouselmi, Ghazi ; Fohr, Dominique ; Laprie, Yves
Author_Institution :
Groupe Parole, LOPJA-CNRS & INPJA, Vandoeuvre-les-Nancy
fDate :
March 31 2008-April 4 2008
Abstract :
A fast likelihood computation approach called dynamic Gaussian selection (DGS) is proposed for HMM-based continuous speech recognition. DGS approach is a one-pass search technique which generates a dynamic shortlist of Gaussians for each state during the procedure of likelihood computation. The shortlist consists of the Gaussians which make prominent contribution to the likelihood. In principle, DGS is an extension of the technique of partial distance elimination, and it requires almost no additional memory for the storage of Gaussian shortlists. DGS algorithm has been implemented by modifying the likelihood computation module in HTK 3.4 system. Results from experiments on TIMIT and HIWIRE corpora indicate that this approach can speed up the likelihood computation significantly without introducing apparent additional recognition error.
Keywords :
Gaussian processes; hidden Markov models; speech recognition; HMM-based continuous speech recognition; dynamic Gaussian selection technique; error recognition; likelihood computation approach; one-pass search technique; partial distance elimination; Cognitive science; Distortion measurement; Distributed computing; Hidden Markov models; Nearest neighbor searches; Probability distribution; Speech analysis; Speech recognition; Testing; Vocabulary; Gaussian selection; fast likelihood computation; hidden Markov models; speech recognition;
Conference_Titel :
Acoustics, Speech and Signal Processing, 2008. ICASSP 2008. IEEE International Conference on
Conference_Location :
Las Vegas, NV
Print_ISBN :
978-1-4244-1483-3
Electronic_ISBN :
1520-6149
DOI :
10.1109/ICASSP.2008.4518646