Comparison of low- and high-level visual features for audio-visual continuous automatic speech recognition.

Petar S. Aleksic, Aggelos K. Katsaggelos

Browse the full ICASSP paper archive.