DCT-based video features for audio-visual speech recognition.

Martin Heckmann, Kristian Kroschel, Christophe Savariaux, Frdric Berthommier

Browse the full Interspeech paper archive.