ABSTRACT
The first item, of course, is an interesting one-particularly
the biometrics of lip reading movement. More interesting
is the integration of this modality with the other two-face
and voice. The advantage of this system is that if one
modality is not working properly, the other two modalities
will compensate for the errors of the first. What this means
is if one modularity is disturbed (e.g., a noisy environment
drowning out the voice), the other two modalities still lead
to an accurate identification.