Multistream Articulatory Feature-Based Models for Visual Speech Recognition | IEEE Journals & Magazine | IEEE Xplore