Noisy speech recognition by using output combination of discrete-mixture HMMs and continuous-mixture HMMs.

Tetsuo Kosaka, You Saito, Masaharu Kato

Browse the full Interspeech paper archive.