The Multi-Stream automatic speech recognition approach was investigated in this work as a framework for Audio-Visual data fusion and speech recognition. This method presents many potential advantages for such a task. It particularly allows for synchronous ...
IDIAP1997