HTK is a portable software toolkit for developing systems using continuous density hidden Markov models developed by the Cambridge University Speech Group. This paper describes speech recognition experiments using HTK based systems for the DARPA Resource Management (RM) task. In particular good performance is obtained using a tied-state triphone based multiple mixture approach. This system was used in the nal DARPA RM evaluation (September 1992) and was found to perform at a similar level to the main DARPA systems, and yet be eecient in terms of the total of parameters and computational load. The results for that system are given along with some recent experiments that investigated the use of male-female modelling in a tied-state HMM system.
[1]
Steve J. Young,et al.
The use of state tying in continuous speech recognition
,
1993,
EUROSPEECH.
[2]
John Makhoul,et al.
Context-dependent modeling for acoustic-phonetic recognition of continuous speech
,
1985,
ICASSP '85. IEEE International Conference on Acoustics, Speech, and Signal Processing.
[3]
Chin-Hui Lee,et al.
Bayesian learning for hidden Markov model with Gaussian mixture state observation densities
,
1991,
Speech Commun..
[4]
Steve Young,et al.
Benchmark DARPA RM results using the HTK portable HMM toolkit
,
1992
.