Using the multi-stream approach for continuous audio-visual speech recognition: experiments on the M2VTS database.

Stphane Dupont, Juergen Luettin

Browse the full Interspeech paper archive.