A fast match for continuous speech recognition using allophonic models.

Lalit R. Bahl, Peter V. de Souza, P. S. Gopalakrishnan, David Nahamoo, Michael A. Picheny

Browse the full ICASSP paper archive.