A new approach to continuous speech recognition based on considerations on human processes of speech perception.

Hiroya Fujisaki, Keikichi Hirose, Hirofumi Udagawa, Noboru Kanedera

Browse the full ICASSP paper archive.