Game Development Reference
In-Depth Information
Figure 28. The form of membership functions.
µ
(
k
j
)
i
,
s
(
k
j
)
i
,
(
k
j
)
c
(
,
k
j
)
s
(
k
)
s
,
i
,
j
i
(
,
k
j
)
where
(
,
k
j
)
(
,
k
j
)
expresses the
relevance
r
of the
i
-th element of the
r
=
max{
g
∩
A
}
i
i
i
(
,
k
j
)
input feature vector with respect to class
is
the fuzzified input vector resulting from a
singleton
fuzzification procedure (Klir
& Yuan, 1995).
The various emotion profiles correspond to the fuzzy intersection of several sets
and are implemented through a
A
. Actually
g
=
A
'
(
G
)
=
{
g
1
g
,
,...}
2
−
−
-norm
of the form
t
(
a
,
b
)=
a
·
b
. Similarly the
belief that an observed feature vector corresponds to a particular emotion results
from a fuzzy union of several sets through an
-norm
which is implemented as
u
(
a
,
b
)=max(
a
,
b)
.
Figure 29. Facial expression analysis interface.
Search WWH ::
Custom Search