<s>
Audio	B-General_Concept
visual	I-General_Concept
speech	I-General_Concept
recognition	I-General_Concept
(	O
AVSR	B-General_Concept
)	O
is	O
a	O
technique	O
that	O
uses	O
image	B-Algorithm
processing	I-Algorithm
capabilities	O
in	O
lip	O
reading	O
to	O
aid	O
speech	B-Application
recognition	I-Application
systems	O
in	O
recognizing	O
undeterministic	O
phones	O
or	O
giving	O
preponderance	O
among	O
near	O
probability	O
decisions	O
.	O
</s>
<s>
Each	O
system	O
of	O
lip	O
reading	O
and	O
speech	B-Application
recognition	I-Application
works	O
separately	O
,	O
then	O
their	O
results	O
are	O
mixed	O
at	O
the	O
stage	O
of	O
feature	O
fusion	O
.	O
</s>
