Comparing audio- and a-posteriori-probability-based stream confidence measures for audio-visual speech recognition.

Martin Heckmann, Thorsten Wild, Frdric Berthommier, Kristian Kroschel

Browse the full Interspeech paper archive.