Skip to main content
Publication

Using entropy as a stream reliability estimate for audio-visual speech recognition