A two-channel training algorithm for hidden Markov model to identify visual speech elements.

Say Wei Foo, Yong Lian, Liang Dong

Browse the full ISCAS paper archive.