Phonetic recognition using hidden Markov models and maximum mutual information training (1988)

by B Merialdo
Venue:Proceedings of ICASSP-88