Skip to main content
Publication

Selecting relevant visual features for speechreading