A second-order HMM for high performance word and phoneme-based continuous speech recognition.

Jean-Franois Mari, Dominique Fohr, Jean-Claude Junqua

Browse the full ICASSP paper archive.