Enhancement and optimisation of a speech recognition front end based on hidden Markov models

A method for performance evaluation of the acousticphonetic front end of a continuous speech recognition system, using the entropy of its output, is described. Results are given for a front end based on phonemic hidden Markov models, with various optional enhancements which have been optimised using the entropy criterion.

[1]  R. Moore,et al.  Explicit modelling of state occupancy in hidden Markov models for automatic speech recognition , 1985, ICASSP '85. IEEE International Conference on Acoustics, Speech, and Signal Processing.

[2]  Yasuo Ariki,et al.  Hierarchical phoneme discrimination by hidden Markov modelling using cepstrum and formant information , 1989, International Conference on Acoustics, Speech, and Signal Processing,.

[3]  L. R. Rabiner,et al.  On the application of vector quantization and hidden Markov models to speaker-independent, isolated word recognition , 1983, The Bell System Technical Journal.