Two features combination with gated recurrent unit for visual speech recognition.

Masaya Iwasaki, Michiko Kubokawa, Takeshi Saitoh

Browse the full MVA paper archive.