DBN based multi-stream models for audio-visual speech recognition.

John N. Gowdy, Amarnag Subramanya, Chris D. Bartels, Jeff A. Bilmes

Browse the full ICASSP paper archive.