In
d
o
n
e
si
a
n
J
o
u
r
n
a
l
o
f
E
l
e
c
t
r
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
Vo
l
.
10
, N
o
.
2
,
Ma
y
201
8
, p
p
.
554~
561
IS
S
N
:
2502
-
4752
,
DOI
:
1
0
.
1
1
5
9
1
/
ij
e
e
c
s
.
v10.
i
2
.p
p
554
-
561

554
Jou
r
n
al
h
om
e
page
:
ht
t
p:
/
/
i
ae
s
co
r
e.
co
m
/
j
o
u
r
n
a
l
s
/
i
n
d
ex.
p
h
p
/
i
j
eecs
Spe
e
c
h
Em
o
t
i
o
n
R
e
c
o
g
n
i
t
i
o
n
U
s
i
n
g
D
e
e
p
F
e
e
d
f
o
r
w
a
r
d
Ne
u
r
a
l
Ne
t
w
o
r
k
Mu
h
a
m
m
a
d
F
a
h
r
e
z
a
A
l
g
h
i
f
a
r
i
1
, T
e
d
d
y
S
u
r
y
a
G
u
n
a
w
a
n
*
2
, M
ir
a
K
a
r
t
iw
i
3
1
,2
De
p
a
r
t
me
n
t
o
f
E
l
e
c
t
r
i
c
a
l
a
n
d
C
o
mp
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
,
Ku
l
l
i
y
y
a
h
o
f
E
n
g
i
n
e
e
r
i
n
g
, M
a
la
y
s
ia
3
De
p
a
r
t
me
n
t
o
f
I
n
f
o
r
m
at
i
o
n
S
y
st
em
s,
K
u
l
l
i
y
y
ah
o
f
I
C
T
,
In
t
e
rn
a
t
i
o
n
a
l
Is
l
a
m
i
c
U
n
i
v
e
rs
i
t
y
M
a
l
a
y
s
i
a
, M
a
la
y
s
ia
Ar
t
i
c
l
e
I
n
f
o
AB
S
T
RACT
Ar
t
i
c
l
e
h
i
s
t
o
r
y
:
Re
c
e
i
v
e
d
No
v
26
,
201
7
Re
v
i
s
e
d
Ja
n
2
3
, 2
0
1
8
Ac
c
e
p
t
e
d
Fe
b
2
1
, 2
0
1
8
Sp
e
e
c
h
e
m
o
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
(
SER
)
i
s
c
u
r
r
e
n
t
l
y
a
r
e
s
e
a
r
c
h
h
o
t
s
p
o
t
d
u
e
t
o
i
t
s
ch
al
l
en
g
i
n
g
n
at
u
r
e
b
u
t
b
o
u
n
t
i
f
u
l
f
u
t
u
r
e
p
r
o
sp
ect
s.
T
h
e
o
b
j
ect
i
v
e
o
f
t
h
i
s
re
s
e
a
rc
h
i
s
t
o
u
t
i
l
i
z
e
D
e
e
p
N
e
u
ra
l
N
e
t
w
o
rk
s
(D
N
N
s
)
t
o
re
c
o
g
n
i
z
e
h
u
m
a
n
sp
eech
em
o
t
i
o
n
.
F
i
r
st
,
t
h
e
ch
o
sen
sp
eech
feat
u
r
e
M
el
-
fr
e
q
u
e
n
c
y
c
e
p
st
r
a
l
co
ef
f
i
ci
en
t
(
M
F
C
C
)
w
er
e
ex
t
r
act
ed
f
r
o
m
r
aw
au
d
i
o
d
at
a.
S
eco
n
d
,
t
h
e
sp
eech
fe
a
t
u
r
e
s
e
x
t
r
a
c
t
e
d
w
e
r
e
fe
d
i
n
t
o
t
h
e
D
N
N
t
o
t
r
a
i
n
t
h
e
n
e
t
w
o
r
k
.
T
h
e
t
r
a
i
n
e
d
ne
t
w
or
k
w
a
s
t
he
n
t
e
st
ed
o
n
t
o
a
set
o
f
l
ab
el
l
ed
em
o
t
i
o
n
sp
eech
au
d
i
o
an
d
t
h
e
re
c
o
g
n
i
t
i
o
n
ra
t
e
w
a
s
e
v
a
l
u
a
t
e
d
.
B
a
s
e
d
o
n
t
h
e
a
c
c
u
ra
c
y
ra
t
e
t
h
e
M
F
C
C
,
num
be
r
of
ne
ur
ons
a
nd
l
a
ye
r
s
a
r
e
a
dj
us
t
e
d
f
or
opt
i
m
i
z
a
t
i
on.
M
or
e
ove
r
,
a
cu
st
o
m
-
ma
d
e
d
a
t
a
b
a
s
e
i
s
i
n
t
r
o
d
u
c
e
d
a
n
d
v
a
l
i
d
a
t
e
d
u
s
i
n
g
t
h
e
ne
t
w
or
k
opt
i
m
i
z
e
d.
Th
e
o
p
t
i
m
u
m
c
o
n
f
i
g
u
r
a
t
i
o
n
f
o
r
SER
i
s
1
3
M
FC
C
,
1
2
n
e
u
r
o
n
s
a
n
d
2
l
a
ye
r
s
f
or
3
e
m
ot
i
ons
a
nd
25
M
F
C
C
,
21
ne
ur
ons
a
nd
4
l
a
ye
r
s
f
or
4
em
o
t
i
o
n
s,
ach
i
ev
i
n
g
a
t
o
t
al
r
eco
g
n
i
t
i
o
n
r
at
e
o
f
9
6
.
3
%
f
o
r
3
em
o
t
i
o
n
s
an
d
97.
1%
f
or
4
e
m
ot
i
ons
.
Ke
y
wo
r
d
s
:
De
e
p
n
e
u
r
a
l
n
e
t
wo
r
k
Me
l
-
fr
e
q
u
e
n
c
y
c
e
p
st
r
a
l
c
o
e
ffi
c
i
e
n
t
s
(M
F
C
C
)
Sp
e
e
c
h
e
m
o
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
(
SER
)
Co
p
y
r
i
g
h
t
©
2
0
1
8
In
s
t
i
t
u
t
e
o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
r
i
n
g
a
n
d
S
c
i
e
n
c
e
.
Al
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
Co
r
r
e
s
p
o
n
d
i
n
g
Au
t
h
o
r
:
Te
d
d
y
S
u
r
y
a
G
u
n
a
w
a
n
,
De
p
a
r
t
m
e
n
t
o
f
E
l
e
c
t
r
i
c
a
l
a
n
d
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
,
Ku
l
l
i
y
y
a
h
o
f
E
n
g
i
n
e
e
r
i
n
g
, M
a
la
y
s
ia
.
Em
a
i
l
:
ts
g
u
n
a
w
a
n
@
iiu
m
.e
d
u
.m
y
1.
IN
T
R
O
D
U
C
T
IO
N
Sp
e
e
c
h
Em
o
t
i
o
n
R
e
c
o
g
n
i
t
i
o
n
(S
E
R
)
c
a
n
b
e
d
e
f
i
n
e
d
a
s
t
h
e
i
d
e
n
t
i
f
i
c
a
t
i
o
n
o
f
t
h
e
e
m
o
t
i
o
n
a
l
s
t
a
t
e
o
f
t
h
e
sp
e
a
k
e
r
f
r
o
m
h
i
s
o
r
h
e
r
sp
e
e
c
h
si
g
n
a
l
[1
]
.
S
E
R
is
o
n
e
o
f
th
e
to
p
ic
s
in
s
p
e
e
c
h
p
r
o
c
e
s
s
in
g
th
a
t
h
a
s
b
e
e
n
co
n
t
i
n
u
o
u
s
l
y
r
es
ear
ch
ed
f
o
r
d
ecad
es
,
t
h
e
s
i
m
p
l
es
t
at
t
em
p
t
s
d
at
es
b
ack
f
r
o
m
t
h
e
l
at
e
f
i
f
t
i
es
[2
]
.
I
n
to
d
a
y
’
s
wo
r
l
d
,
S
E
R
h
a
s
s
h
o
wn
t
o
b
e
q
u
i
t
e
a
r
e
s
e
a
r
c
h
h
o
t
s
p
o
t
,
a
s
i
n
d
i
c
a
t
e
d
b
y
t
h
e
g
r
o
wt
h
o
f
p
u
b
l
i
c
a
t
i
o
n
p
a
p
e
r
s
in
e
a
c
h
y
e
a
r
.
Th
e
a
p
p
l
i
c
a
t
i
o
n
o
f
S
ER
c
a
n
b
e
t
a
r
g
e
t
e
d
t
o
s
e
v
e
r
a
l
se
c
t
o
r
s.
I
n
b
a
n
k
i
n
g
,
a
n
a
u
t
o
c
a
l
l
e
r
e
q
u
i
p
p
e
d
w
i
t
h
SE
R
m
a
y
a
s
s
i
s
t
i
n
d
e
t
e
c
t
i
n
g
t
h
e
e
m
o
t
i
o
n
o
f
t
h
e
c
u
s
t
o
m
e
r
,
g
e
n
e
r
a
t
i
n
g
c
u
s
t
o
m
r
e
s
p
o
n
s
e
s
b
a
s
e
d
o
n
t
h
e
r
e
s
u
l
t
.
I
n
ed
u
cat
i
o
n
,
an
e
-
le
a
r
n
in
g
p
o
r
ta
l
w
ith
S
E
R
c
a
n
d
e
te
c
t
th
e
e
m
o
tio
n
s
o
f
th
e
u
s
e
r
s
u
c
h
a
s
f
r
u
s
tr
a
tio
n
an
d
s
t
r
es
s
,
de
t
e
r
m
i
ni
ng
w
he
t
he
r
t
he
s
t
udyi
ng
i
s
c
onduc
i
ve
or
not
a
nd
gi
ve
a
ppr
opr
i
a
t
e
c
ount
e
r
m
e
a
s
ur
e
s
.
Y
e
t
a
not
he
r
ap
p
l
i
cat
i
o
n
i
s
i
n
t
r
an
s
p
o
r
t
at
i
o
n
,
w
h
er
e
i
n
t
h
e
n
ear
-
fu
t
u
r
e
t
h
a
t
v
e
h
i
c
l
e
s
a
r
e
c
a
p
a
b
l
e
o
f
a
u
t
o
-
dr
i
vi
ng,
t
he
s
ys
t
e
m
can
t
ak
e
o
v
er
t
h
e
s
t
eer
i
n
g
w
he
e
l
i
n
t
he
c
a
s
e
w
he
r
e
a
n
unhe
a
l
t
hy
a
m
ount
of
e
m
ot
i
on
i
s
de
t
e
c
t
e
d
fr
o
m
t
h
e
d
r
i
v
e
r
.
A
t
y
p
i
c
a
l
s
p
e
e
c
h
e
m
o
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
i
s
i
l
l
u
s
t
r
a
t
e
d
i
n
F
i
g
u
r
e
1
.
T
h
e
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
m
a
r
k
s
t
h
e
s
t
a
r
t
of
a
S
E
R
s
ys
t
e
m
.
T
hi
s
i
nc
l
ude
s
s
e
l
e
c
t
i
ng
t
he
f
e
a
t
ur
e
s
a
ppr
opr
i
a
t
e
f
o
r
e
m
o
t
i
o
n
re
c
o
g
n
i
t
i
o
n
.
N
e
x
t
t
h
e
s
e
f
e
a
t
u
re
s
ar
e
p
r
o
ces
s
ed
b
y
a
cl
as
s
i
f
i
er
.
T
h
es
e
cl
as
s
i
f
i
er
s
ar
e
t
r
ai
n
ed
b
y
r
ef
er
r
i
n
g
t
o
an
em
o
t
i
o
n
d
at
ab
as
e.
N
ex
t
,
t
h
e
sy
st
e
m
w
i
l
l
b
e
p
u
t
i
n
t
o
t
e
st
i
n
g
b
y
c
r
o
ssc
h
e
c
k
i
n
g
w
i
t
h
t
h
e
sa
m
e
d
a
t
a
b
a
se
.
T
h
e
p
r
o
c
e
sse
d
d
a
t
a
o
b
t
a
i
n
e
d
w
i
l
l
be
th
e
d
e
te
r
m
in
a
n
t o
f
th
e
d
e
c
is
io
n
, ty
p
ic
a
lly
in
te
r
m
s
o
f
a
c
c
u
r
a
c
y
a
n
d
p
r
o
c
e
s
s
in
g
tim
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
o
n
e
s
i
a
n
J
E
l
e
c
E
n
g
&
C
o
m
p
S
c
i
IS
S
N
:
2502
-
4752

Spe
e
c
h
E
m
ot
i
on
R
e
c
ogni
t
i
on
U
s
i
ng
D
e
e
p
F
e
e
df
or
w
ar
d
Ne
u
r
a
l
Ne
t
wo
r
k
(
Mu
h
a
m
m
a
d
F
a
h
r
e
z
a
A
l
g
h
i
f
a
r
i
)
555
Fi
g
u
r
e
1
.
A
T
y
p
i
c
a
l
Sp
e
e
c
h
E
m
o
t
i
o
n
R
e
c
o
g
n
i
t
i
o
n
A
l
g
o
r
i
t
h
m
In
[3
]
,
w
e
c
o
n
d
u
c
te
d
a
c
o
m
p
r
e
h
e
n
s
iv
e
lite
r
a
tu
r
e
r
e
v
ie
w
o
n
S
E
R
in
w
h
ic
h
o
u
r
p
r
o
p
o
s
e
d
s
y
s
te
m
is
ba
s
e
d
on
t
ha
t
c
r
i
t
i
c
a
l
r
e
vi
e
w
.
C
ur
r
e
nt
S
E
R
i
s
s
t
i
l
l
i
n
de
ve
l
opm
e
nt
s
t
a
ge
.
A
l
t
hough
it
h
a
s
b
e
e
n
c
o
n
tin
u
o
u
s
ly
re
s
e
a
rc
h
e
d
i
n
t
h
e
p
a
s
t
d
e
c
a
d
e
,
t
h
e
i
m
p
l
e
m
e
n
t
a
t
i
o
n
i
n
o
u
r
d
a
i
l
y
l
i
f
e
a
n
d
p
ra
c
t
i
c
a
l
u
s
e
s
i
s
s
t
i
l
l
v
e
ry
l
i
m
i
t
e
d
.
Ma
n
y
t
h
i
n
g
s
a
r
e
l
e
f
t
t
o
b
e
d
e
s
i
r
e
d
,
i
n
t
e
r
m
s
o
f
a
c
c
u
r
a
c
y
,
p
r
o
c
e
s
s
i
n
g
t
i
m
e
a
n
d
p
r
a
c
t
i
c
a
l
i
t
y
.
I
n
t
h
i
s
s
t
u
d
y
,
w
e
im
p
r
o
v
e
d
t
he
a
c
c
ur
a
c
y
pr
oc
e
s
s
i
ng
t
i
m
e
by
opt
i
m
i
z
i
ng
va
r
i
ous
de
e
p
ne
ur
a
l
ne
t
w
or
k
c
onf
i
gur
a
t
i
ons
.
2.
PR
O
PO
S
E
D
SP
E
E
C
H
E
M
O
T
I
O
N
R
E
C
O
G
N
I
T
I
O
N
SY
ST
E
M
Th
e
S
ER
f
l
o
w
c
o
n
d
u
c
t
e
d
i
n
t
h
i
s
r
e
s
e
a
r
c
h
i
s
s
h
o
w
n
i
n
F
i
g
u
r
e
2
.
F
o
r
t
h
e
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
,
t
h
i
s
s
t
u
d
y
ha
s
c
hos
e
n
t
he
M
e
l
-
fr
e
q
u
e
n
c
y
c
e
p
s
t
r
a
l
c
o
e
ffi
c
i
e
n
t
s
(
M
F
C
C
)
d
u
e
t
o
t
h
e
i
t
s
n
a
t
u
r
e
t
o
b
e
t
u
n
e
d
i
n
a
s
c
a
l
e
t
h
a
t
i
s
su
i
t
a
b
l
e
f
o
r
t
h
e
h
u
m
a
n
e
a
r
[4
]
,
b
e
s
t
s
u
ite
d
f
o
r
N
-
wa
y
c
l
a
s
s
i
f
i
e
r
s
[2
]
,
a
n
d
is
o
n
e
o
f
th
e
m
o
s
t
p
o
p
u
la
r
f
e
a
tu
r
e
to
be
e
xt
r
a
c
t
e
d
i
n
S
E
R
,
s
uc
h
a
s
i
n
[5
]
an
d
[6
]
.
F
o
r
th
e
c
la
s
s
if
ie
r
,
th
e
c
h
o
s
e
n
a
lg
o
r
ith
m
is
th
e
d
e
e
p
n
e
u
r
a
l
ne
t
w
or
k,
a
br
a
nc
h
of
t
he
a
r
t
i
f
i
c
i
a
l
ne
ur
a
l
ne
t
w
or
k.
I
t
e
m
pl
oys
l
e
s
s
pa
r
a
m
e
t
e
r
s
,
hi
ghe
r
pe
r
f
or
m
a
nc
e
c
om
pa
r
e
d
to
V
Q
m
o
d
e
l,
h
a
s
v
e
r
y
h
ig
h
p
o
te
n
tia
l
g
iv
e
n
m
o
r
e
h
id
d
e
n
la
y
e
r
s
[7
]
.
N
e
u
r
a
l
n
e
tw
o
r
k
s
h
a
s
g
a
in
e
d
p
o
p
u
la
r
ity
in
SE
R
s
y
s
t
e
m
s
,
s
u
c
h
a
s
c
o
n
d
u
c
t
e
d
b
y
[8
]
an
d
[9
]
.
Fi
g
u
r
e
2
.
Pr
o
p
o
s
e
d
Sp
e
e
c
h
E
m
o
t
i
o
n
R
e
c
o
g
n
i
t
i
o
n
Sy
s
t
e
m
[3
]
2.
1
Me
l
-
Fr
e
q
u
e
n
c
y
C
e
p
s
t
r
a
l
C
o
e
f
f
i
c
i
e
n
t
s
(
M
FC
C
s
)
Fe
a
t
u
r
e
E
x
t
r
a
c
t
i
o
n
MF
C
C
s
u
s
e
a
n
o
n
-
lin
e
a
r
f
r
e
q
u
e
n
c
y
s
c
a
le
,
i.e
.
m
e
l
s
c
a
le
,
b
a
s
e
d
o
n
th
e
a
u
d
ito
r
y
p
e
r
c
e
p
tio
n
.
A
me
l
is
a
uni
t
of
m
e
a
s
ur
e
of
pe
r
c
e
i
ve
d
pi
t
c
h
or
f
r
e
que
nc
y
of
a
t
one
.
E
q.
(
1)
c
a
n
be
us
e
d
t
o
c
onve
r
t
f
r
e
que
nc
y
s
c
a
l
e
t
o
me
l
s
c
a
l
e
.



=
177
ln
1
+



(1
)
W
he
r
e



is
th
e
f
r
e
q
u
e
n
c
y
in
m
e
ls
a
n
d


is
th
e
n
o
r
m
a
l
f
r
e
q
u
e
n
c
y
in
H
z
.
M
F
C
C
s
a
r
e
o
f
te
n
c
a
lc
u
la
te
d
us
i
ng
a
f
i
l
t
e
r
ba
nk
of

fi
l
t
e
r
s
,
i
n
w
h
i
c
h
e
a
c
h
fi
l
t
e
r
h
a
s
a
t
r
i
a
n
g
u
l
a
r
s
h
a
p
e
a
n
d
i
s
s
p
a
c
e
d
u
n
i
fo
r
m
l
y
on
t
he
m
e
l
sc
a
l
e
a
s
sh
o
w
n
i
n
E
q
u
a
t
i
o
n
(2
).



=
0

<


−
1















−
1
<

≤

















<

≤


+
1
0

>


+
1
(2
)
Evaluation Warning : The document was created with Spire.PDF for Python.

IS
S
N
:
2502
-
4752
In
d
o
n
e
s
i
a
n
J
E
l
e
c
E
n
g
&
C
o
m
p
S
c
i
,
Vo
l
.
10
, N
o
.
2
,
Ma
y
2
0
1
8
:
554
–
561
556
wh
e
r
e

=
0
,
1
,
⋯
,

−
1
. T
h
e
lo
g
-
en
er
g
y
m
el
s
p
ect
r
u
m
i
s
t
h
en
cal
cu
l
at
ed
as
f
o
l
l
o
w
s
:


=
ln













=
0
,
1
,
⋯
,

−
1
(3
)
wh
e
r
e


is
th
e
d
is
c
r
e
te
F
o
u
r
ie
r
tr
a
n
s
f
o
r
m
(
D
F
T
)
o
f
a
s
p
e
e
c
h
in
p
u
t


.
Al
t
h
o
u
g
h
t
r
a
d
i
t
i
o
n
a
l
c
e
p
s
t
r
u
m
u
s
e
s
i
n
v
e
r
s
e
d
i
s
c
r
e
t
e
F
o
u
r
i
e
r
t
r
a
n
s
f
o
r
m
(
I
DF
T
)
,
m
e
l
f
r
e
q
u
e
n
c
y
cep
s
t
r
u
m
i
s
n
o
r
m
al
l
y
i
m
p
l
em
en
t
ed
u
s
i
n
g
d
i
s
cr
et
e
co
s
i
n
e
t
r
an
s
f
o
r
m
(
D
C
T
)
s
i
n
ce
[
]
m
S
is
e
v
e
n
a
s
s
h
o
w
n
in
E
q
.
(4
),
a
s
f
o
l
l
o
w
s
:


=


co
s

+











=
0
,
1
,
⋯
,

−
1
(4
)
Ty
p
i
c
a
l
l
y
,
t
h
e
n
u
m
b
e
r
o
f
f
i
l
t
e
r
s
M
ra
n
g
e
s
f
ro
m
2
0
t
o
4
0
,
a
n
d
t
h
e
n
u
m
b
e
r
o
f
k
e
p
t
c
o
e
f
f
i
c
i
e
n
t
s
i
s
1
3
.
So
m
e
r
e
s
e
a
r
c
h
r
e
p
o
r
t
e
d
t
h
a
t
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
s
p
e
e
ch
r
eco
g
n
i
t
i
o
n
an
d
s
p
eak
er
i
d
en
t
i
f
i
cat
i
o
n
s
y
s
t
em
s
re
a
c
h
e
d
p
e
a
k
w
i
t
h
3
2
-
35
f
i
l
t
e
r
s
[1
0
]
.
2.
2
De
e
p
Ne
u
r
a
l
Ne
t
w
o
r
k
s
(
DNNs
)
Cl
a
s
s
i
f
i
e
r
Th
e
m
a
i
n
p
r
i
n
c
i
p
l
e
o
f
D
N
N
i
s
t
o
u
t
i
l
i
z
e
l
o
w
e
r
l
e
v
e
l
f
e
a
t
u
r
e
s
l
e
a
r
n
i
n
g
t
o
u
p
d
a
t
e
t
h
e
l
e
a
r
n
i
n
g
o
f
h
i
g
h
e
r
fe
a
t
u
r
e
s
.
T
h
e
r
e
a
r
e
m
a
n
y
a
v
a
i
l
a
b
l
e
de
e
p
ar
c
hi
t
e
c
t
ur
e
,
s
u
c
h
a
s
n
e
u
r
a
l
n
e
tw
o
r
k
s
w
ith
m
a
n
y
h
id
d
e
n
la
y
e
r
s
an
d
/
o
r
m
an
y
h
i
d
d
en
v
ar
i
ab
l
es
,
co
n
v
o
l
u
t
i
o
n
al
n
eu
r
al
ne
t
w
or
ks
,
r
e
c
ur
r
e
nt
ne
ur
a
l
ne
t
w
or
ks
,
a
nd
de
e
p
be
l
i
e
f
ne
t
w
or
k
[1
1
]
.
I
n
th
is
r
e
s
e
a
r
c
h
,
w
e
u
s
e
d
d
e
e
p
le
a
r
n
in
g
u
s
in
g
f
e
e
d
f
o
r
w
a
r
d
n
e
u
r
a
l
n
e
tw
o
r
k
a
r
c
h
ite
c
tu
r
e
s
w
ith
mu
l
t
i
l
a
y
e
r
s
h
i
d
d
e
n
l
a
y
e
r
s
w
i
t
h
ma
n
y
h
i
d
d
e
n
v
a
r
i
a
b
l
e
s
.
F
i
g
u
r
e
3
i
l
l
u
s
t
r
a
t
e
s
t
h
e
d
e
e
p
f
e
e
d
f
o
r
w
a
r
d
n
e
u
r
a
l
ne
t
w
or
k
s
t
r
uc
t
ur
e
.
Fi
g
u
r
e
3
.
D
e
e
p
Fe
e
d
f
o
r
w
a
r
d
N
e
u
r
a
l
N
e
t
w
o
r
k
St
r
u
c
t
u
r
e
2.
3
Spe
e
c
h
Em
o
t
i
o
n
D
a
t
a
b
a
s
e
a
n
d
R
e
c
o
r
d
i
n
g
o
f
N
e
w
S
p
e
e
c
h
S
i
g
n
a
l
s
D
a
t
a
b
a
s
e
f
o
r
Te
s
t
i
n
g
Th
e
d
a
t
a
b
a
s
e
u
s
e
d
f
o
r
t
r
a
i
n
i
n
g
i
n
t
h
i
s
r
e
s
e
a
r
c
h
i
s
t
h
e
o
p
e
n
-
so
u
r
c
e
B
e
r
l
i
n
D
a
ta
b
a
s
e
o
f
E
m
o
tio
n
a
l
Sp
e
e
c
h
(
E
m
o
-
DB
)
[1
2
]
,
w
h
ic
h
c
o
n
ta
in
s
5
3
5
e
m
o
tio
n
a
l
a
u
d
io
f
ile
s
in
(
.w
a
v
)
f
o
r
m
a
t
f
r
o
m
1
0
d
if
f
e
r
e
n
t
s
p
e
a
k
e
r
s
(5
m
a
l
e
5
f
e
m
a
l
e
s
).
E
a
c
h
s
p
e
a
k
e
r
u
t
t
e
rs
1
0
G
e
rm
a
n
s
e
n
t
e
n
c
e
s
u
n
d
e
r
7
d
i
f
f
e
re
n
t
em
o
t
i
o
n
s
(a
n
g
e
r,
b
o
re
d
o
m
,
di
s
gus
t
,
f
e
a
r
,
ha
ppi
ne
s
s
,
sa
d
n
e
ss,
n
e
u
t
r
a
l
)
.
F
o
r
t
h
i
s
r
e
se
a
r
c
h
,
o
n
l
y
f
o
u
r
e
m
o
t
i
o
n
s
a
r
e
se
l
e
c
t
e
d
,
i
n
c
l
u
d
i
n
g
a
n
g
e
r
,
ha
ppi
ne
s
s
,
s
a
dne
s
s
a
nd
ne
ut
r
a
l
.
As
a
d
d
i
t
i
o
n
a
l
c
o
n
t
r
i
b
u
t
i
o
n
t
o
t
h
e
S
E
R
r
e
s
e
a
r
c
h
f
i
e
l
d
,
t
h
i
s
r
e
s
e
a
r
c
h
h
a
s
a
t
t
e
m
p
t
e
d
t
o
c
r
e
a
t
e
a
n
e
w
em
o
t
i
o
n
al
d
at
as
et
.
I
n
s
p
i
r
ed
b
y
t
h
e
d
a
ta
b
a
s
e
s
c
r
e
a
te
d
in
[1
3
]
an
d
[1
4
]
,
th
e
c
u
s
to
m
d
a
ta
s
e
t
c
o
n
ta
in
s
e
lic
ite
d
En
g
l
i
s
h
e
m
o
t
i
o
n
a
l
d
a
t
a
f
r
o
m
9
s
p
e
a
k
e
r
s
i
n
4
d
i
f
f
e
r
e
n
t
e
m
o
t
i
o
n
s
,
h
a
p
p
y
,
a
n
g
r
y
,
s
a
d
a
n
d
n
e
u
t
r
a
l
.
F
o
r
e
a
c
h
em
o
t
i
o
n
,
t
h
e
s
p
eak
er
i
s
p
r
o
m
p
t
ed
t
o
act
o
u
t
5
l
i
n
es
,
f
o
r
an
i
n
i
t
ia
l to
ta
l o
f
1
8
0
e
m
o
tio
n
a
l lin
e
s
.
Th
e
r
e
c
o
r
d
i
n
g
m
e
d
i
u
m
u
s
e
d
i
s
a
c
o
n
v
e
n
t
i
o
n
a
l
p
h
o
n
e
’
s
r
e
c
o
r
d
e
r
a
n
d
a
u
d
i
o
t
r
a
n
s
m
i
t
t
e
d
t
h
r
o
u
g
h
Wh
a
t
s
A
p
p
v
o
i
c
e
m
e
s
s
a
g
i
n
g
i
n
a
n
o
i
s
y
e
n
v
i
r
o
n
m
e
n
t
.
T
h
e
p
r
i
m
a
r
y
r
e
a
s
o
n
i
s
t
o
s
i
m
u
l
a
t
e
t
h
e
c
o
n
v
e
r
s
a
t
i
o
n
i
n
a
re
a
l
-
lif
e
e
n
v
ir
o
n
m
e
n
t.
T
h
e
r
eco
r
d
i
n
g
i
s
n
o
r
m
al
l
y
p
er
f
o
r
m
ed
i
n
a
r
eco
r
d
i
n
g
s
t
u
d
i
o
w
h
er
e
t
h
er
e
i
s
m
i
n
i
m
al
noi
s
e
a
nd
i
nt
e
r
f
e
r
e
nc
e
.
B
ut
i
n
r
e
a
l
i
t
y,
c
onve
r
s
a
t
i
ons
a
r
e
he
l
d
e
ve
r
yw
he
r
e
,
t
he
r
e
f
or
e
s
a
m
pl
e
s
t
ha
t
ha
ve
noi
s
e
s
ar
e
m
o
r
e
v
al
u
ab
l
e
t
o
t
r
ai
n
t
h
e
n
et
w
o
r
k
.
Th
e
s
e
c
o
n
d
r
e
a
s
o
n
i
s
b
e
c
a
u
se
o
f
t
h
e
c
o
m
p
r
e
ssi
o
n
f
a
c
t
o
r
.
I
n
c
o
n
v
e
n
t
i
o
n
a
l
e
m
o
t
i
o
n
a
l
d
a
t
a
b
a
se
s,
a
u
d
i
o
qua
l
i
t
y
i
s
one
of
t
he
pr
e
f
e
r
r
e
d
f
a
c
t
or
s
s
o
t
ha
t
f
e
a
t
ur
e
e
xt
r
a
c
t
i
on
i
s
a
c
hi
e
ve
d
w
i
t
hout
m
uc
h
di
f
f
i
c
ul
t
i
e
s
.
O
n
t
he
Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
o
n
e
s
i
a
n
J
E
l
e
c
E
n
g
&
C
o
m
p
S
c
i
IS
S
N
:
2502
-
4752

Spe
e
c
h
E
m
ot
i
on
R
e
c
ogni
t
i
on
U
s
i
ng
D
e
e
p
F
e
e
df
or
w
ar
d
Ne
u
r
a
l
Ne
t
wo
r
k
(
Mu
h
a
m
m
a
d
F
a
h
r
e
z
a
A
l
g
h
i
f
a
r
i
)
557
ot
he
r
ha
nd,
i
n
m
ode
r
n
V
oi
c
e
-
ove
r
-
In
t
e
rn
e
t
-
Pr
o
t
o
c
o
l
(
V
o
I
P)
,
a
u
d
i
o
c
o
m
p
r
es
s
i
o
n
i
s
o
n
e
o
f
t
h
e
m
aj
o
r
co
n
s
i
d
er
at
i
o
n
s
t
o
en
s
u
r
e
t
h
at
t
h
e
ch
an
n
el
can
s
i
m
u
l
at
e
‘
r
eal
-
tim
e
’
c
o
n
v
e
r
s
a
tio
n
s
w
h
ile
s
e
n
d
in
g
m
in
im
u
m
d
a
ta
ove
r
t
he
c
onne
c
t
i
on.
To
e
n
s
u
r
e
t
h
a
t
t
h
e
q
u
a
l
i
t
y
o
f
t
h
e
e
m
o
t
i
o
n
a
l
a
u
d
i
o
i
s
a
c
c
e
p
t
a
b
l
e
,
e
a
c
h
a
u
d
i
o
f
i
l
e
s
a
r
e
c
o
d
e
d
ap
p
r
o
p
r
i
at
el
y
an
d
r
an
d
o
m
l
y
s
el
ect
ed
.
F
o
r
each
au
d
i
o
f
i
l
e,
t
h
e
v
al
i
d
at
o
r
w
i
l
l
g
u
es
s
w
h
at
em
o
t
i
o
n
t
h
at
i
s
co
n
v
ey
ed
.
F
o
r
each
co
r
r
ect
an
s
w
er
t
h
e
au
d
i
o
f
i
l
e
w
i
l
l
b
e
m
o
v
ed
t
o
a
v
al
i
d
at
ed
p
o
o
l
.
F
o
r
v
al
i
d
at
i
o
n
i
n
t
eg
r
i
t
y
,
th
e
p
r
o
c
e
s
s
w
a
s
c
o
n
d
u
c
te
d
in
a
n
e
a
r
s
ile
n
c
e
la
b
, e
qui
ppe
d
w
i
t
h
a
noi
s
e
-
can
cel
l
i
n
g
h
ead
p
h
o
n
e.
F
r
o
m
1
8
0
v
o
i
ce
lin
e
s
, 1
4
8
a
r
e
d
is
tin
g
u
is
h
a
b
le
a
n
d
c
o
n
s
id
e
r
e
d
to
b
e
a
d
d
e
d
to
o
u
r
S
E
R
d
a
ta
b
a
s
e
.
3.
RE
S
UL
T
S
AND
DI
S
CUS
S
I
O
N
Th
e
m
a
i
n
f
o
c
u
s
o
f
t
h
i
s
s
e
c
t
i
o
n
i
s
t
o
i
n
v
e
s
t
i
g
a
t
e
t
h
e
i
d
e
a
l
n
e
t
w
o
r
k
c
o
n
f
i
g
u
r
a
t
i
o
n
b
y
a
d
j
u
s
t
i
n
g
th
e
num
be
r
of
M
F
C
C
,
ne
ur
ons
i
n
e
a
c
h
l
a
ye
r
,
a
s
w
e
l
l
a
s
num
be
r
of
l
a
ye
r
s
.
F
or
t
he
f
i
r
s
t
s
t
a
ge
,
3
e
m
ot
i
ons
a
r
e
t
a
ke
n
in
to
c
o
n
s
id
e
r
a
tio
n
–
ha
ppy,
a
ngr
y,
a
nd
s
a
d.
A
f
t
e
r
w
a
r
ds
,
a
n
a
ddi
t
i
ona
l
e
m
ot
i
on,
ne
ut
r
a
l
i
s
a
dde
d
t
o
t
he
s
ys
t
e
m
,
re
t
ra
i
n
e
d
a
n
d
t
e
s
t
e
d
.
T
h
e
f
i
rs
t
in
v
e
s
tig
a
tio
n
w
a
s
th
e
e
x
tr
a
c
tio
n
o
f
M
F
C
C
f
r
o
m
d
a
ta
b
a
s
e
in
te
r
m
s
o
f
p
r
o
c
e
s
s
in
g
tim
e
.
F
o
r
e
a
c
h
c
o
e
f
f
ic
ie
n
t
f
r
o
m
1
to
1
0
0
,
th
e
p
r
o
c
e
s
s
in
g
tim
e
is
r
e
c
o
r
d
e
d
a
n
d
r
e
p
e
a
te
d
f
o
r
5
tim
e
s
,
th
e
n
th
e
av
er
ag
e
wa
s
c
a
l
c
u
l
a
t
e
d
a
s
s
h
o
wn
i
n
F
i
g
u
r
e
4.
Fi
g
u
r
e
4
.
Ex
p
e
r
i
m
e
n
t
s
o
n
N
u
m
b
e
r
o
f
M
F
C
C
a
n
d
I
t
s
P
r
o
c
e
s
s
i
n
g
Ti
m
e
Fr
o
m
t
h
e
r
e
s
u
l
t
s
,
i
t
c
a
n
b
e
o
b
s
e
r
v
e
d
t
h
a
t
t
h
e
M
FC
C
e
x
t
r
a
c
t
i
o
n
p
r
o
c
e
s
s
i
s
l
i
g
h
t
w
e
i
g
h
t
a
c
r
o
s
s
n
u
m
b
e
r
of
c
oe
f
f
i
c
i
e
nt
s
w
i
t
h
s
m
a
l
l
va
r
i
a
nc
e
a
nd
s
t
a
nda
r
d
de
vi
a
t
i
on
of
0.
265
a
nd
0.
51,
r
e
s
pe
c
t
i
ve
l
y.
E
ve
n
goi
n
g
up
t
o
100
M
F
C
C
,
t
he
pr
oc
e
s
s
i
ng
t
i
m
e
i
s
s
t
i
l
l
c
ons
i
de
r
a
bl
y
l
ow
a
t
7.
756
m
s
.
H
a
vi
ng
a
ne
gl
i
gi
bl
e
pr
oc
e
s
s
i
ng
t
i
m
e
me
a
n
s
t
h
a
t
p
r
o
p
o
s
e
d
s
y
s
t
e
m
d
o
n
o
t
h
a
v
e
c
o
n
s
t
r
a
i
n
e
d
o
n
t
h
e
n
u
mb
e
r
o
f
c
o
e
f
f
i
c
i
e
n
t
s
w
h
i
c
h
e
n
a
b
l
e
s
t
h
e
f
o
c
u
s
t
o
be
on
ot
he
r
pa
r
a
m
e
t
e
r
s
.
Th
e
n
e
x
t
s
t
ep
i
s
i
n
v
es
t
i
g
at
i
n
g
t
h
e
b
es
t
p
er
f
o
r
m
an
ce
o
f
n
eu
r
al
n
et
w
o
r
k
co
n
f
i
g
u
r
at
i
o
n
,
i
n
w
h
i
ch
t
h
e
fo
l
l
o
w
i
n
g
M
a
t
l
a
b
’
s
fu
n
c
t
i
o
n
w
e
r
e
u
s
e
d
,
i
n
c
l
u
d
i
n
g
fe
e
d
fo
r
w
a
r
d
n
e
t
(
)
,
p
a
t
t
e
r
n
n
e
t
(
)
,
fi
t
n
e
t
(),
a
n
d
cas
cad
ed
f
eed
f
o
r
w
ar
d
n
et
(
)
.
T
h
e
d
e
f
a
u
lt
n
u
m
b
e
r
o
f
n
e
u
r
o
n
s
is
1
0
w
ith
a
s
in
g
l
e
h
i
d
d
en
l
ay
er
.
T
h
e
t
r
ai
n
i
n
g
,
va
l
i
da
t
i
on
a
nd
t
e
s
t
i
ng
r
a
t
i
o
i
s
70%
:
15%
:
15%
,
r
e
s
pe
c
t
i
ve
l
y.
T
he
r
e
s
ul
t
s
i
s
s
how
n
i
n
T
a
bl
e
1.
Ta
b
l
e
1
.
R
e
s
u
l
t
s
o
f
V
a
r
i
o
u
s
N
e
u
r
a
l
N
e
t
w
o
r
k
S
t
r
u
c
t
u
r
e
s
Ma
t
l
a
b
F
u
n
c
t
i
o
n
Va
r
i
a
n
c
e
St
a
n
d
a
r
d
D
e
v
i
a
t
i
o
n
Be
s
t
P
e
r
f
o
r
m
a
n
c
e
Wo
r
s
t
P
e
r
f
o
r
m
a
n
c
e
feed
fo
r
w
ar
d
n
et
(
)
0.
00050
0.
02231
0.
96648
0.
84575
pa
t
t
e
r
nne
t
(
)
0.
00315
0.
05610
0.
92778
0.
63111
fi
t
n
et
(
)
0.
003574
0.
059782
0.
95778
0.
64444
cascad
efo
r
w
ar
d
n
et
(
)
0.
003816
0.
006177
0.
96333
0.
63778
Fr
o
m
o
b
s
e
r
v
a
t
i
o
n
,
f
e
e
d
f
o
r
w
a
r
d
n
e
t
(
)
a
l
g
o
r
i
t
h
m
p
r
o
v
i
d
e
s
t
h
e
be
s
t
ove
r
a
l
l
r
e
l
i
a
bl
e
pe
r
f
or
m
a
nc
e
co
m
p
ar
ed
t
o
i
t
s
co
u
n
t
er
p
ar
t
s
,
b
o
o
s
t
i
n
g
a
r
eco
g
n
i
t
i
o
n
r
at
e
o
f
0
.
9
6
6
at
1
3
M
F
C
C
w
h
i
l
e
m
ai
n
t
ai
n
i
n
g
l
es
s
va
r
i
a
nc
e
a
nd
s
t
a
nda
r
d
de
vi
a
t
i
on.
T
hi
s
a
l
gor
i
t
hm
i
s
c
hos
e
n
t
o
be
t
he
f
ounda
t
i
on
of
de
e
p
ne
ur
a
l
ne
t
w
or
k.
T
he
ne
xt
s
t
a
ge
of
opt
i
m
i
z
i
ng
i
s
f
i
ndi
ng
t
he
i
de
a
l
num
be
r
of
ne
ur
ons
i
n
a
l
a
ye
r
.
F
or
t
hi
s
pr
oc
e
s
s
,
t
he
M
F
C
C
i
s
ke
pt
co
n
s
t
an
t
at
1
3
an
d
t
h
e
n
u
m
b
er
o
f
n
eu
r
o
n
s
i
s
i
n
cr
eas
ed
.
0
2
4
6
8
10
12
1
9
17
25
33
41
49
57
65
73
81
89
97
Pro
ces
s
i
n
g
T
i
me
(
ms
)

N
u
m
b
e
r

o
f

MF
C
C

Evaluation Warning : The document was created with Spire.PDF for Python.

IS
S
N
:
2502
-
4752
In
d
o
n
e
s
i
a
n
J
E
l
e
c
E
n
g
&
C
o
m
p
S
c
i
,
Vo
l
.
10
, N
o
.
2
,
Ma
y
2
0
1
8
:
554
–
561
558
Fi
g
u
r
e
5
.
E
x
p
e
r
i
m
e
n
t
s
o
n
N
u
m
b
e
r
o
f
N
e
u
r
o
n
i
n
t
h
e
H
i
d
d
e
n
L
a
y
e
r
As
s
h
o
wn
i
n
F
i
g
u
r
e
5,
t
he
ove
r
a
l
l
r
es
u
l
t
s
i
m
p
r
o
v
ed
i
n
t
er
m
s
o
f
r
eco
g
n
i
t
i
o
n
r
at
e,
p
eak
i
n
g
at
n
u
m
b
er
of
ne
ur
on
of
12.
F
i
na
l
l
y,
t
he
num
be
r
of
hi
dde
n
l
ay
er
s
w
i
l
l
b
e
v
ar
i
ed
b
et
w
een
1
an
d
4.
U
s
i
ng
t
he
s
a
m
e
num
be
r
MF
C
C
a
n
d
n
e
u
r
o
n
s
,
l
a
y
e
r
s
a
r
e
a
d
d
e
d
i
n
c
r
e
m
e
n
t
a
l
l
y
b
y
1
u
n
t
i
l
t
h
e
r
e
s
u
l
t
s
c
a
n
n
o
l
o
n
g
e
r
i
m
p
ro
v
e
.
T
h
e
i
n
c
re
a
s
e
in
la
y
e
r
s
s
h
o
u
ld
s
ig
n
if
ic
a
n
tly
im
p
a
c
t
th
e
p
r
o
c
e
s
s
in
g
tim
e
,
th
e
r
e
f
o
r
e
it
s
h
o
u
ld
n
o
w
b
e
ta
k
e
n
in
to
a
c
c
o
u
n
t.
T
h
e
fi
n
a
l
r
e
s
u
l
t
s
a
r
e
s
h
o
w
n
i
n
T
a
b
l
e
2
.
Ta
b
l
e
2
.
Ex
p
e
r
i
m
e
n
t
s
o
n
N
u
m
b
e
r
o
f
H
i
d
d
e
n
La
y
e
r
s
Nu
mb
e
r
o
f
L
a
y
e
r
Be
s
t
P
e
r
f
o
r
m
a
n
c
e
Pr
o
c
e
s
s
i
n
g
Ti
m
e
(
s
)
1
0.
960
3.
919
2
0.
972
4.
128
3
0.
968
4.
608
4
0.
969
4.
104
Fo
r
t
h
r
e
e
e
m
o
t
i
o
n
s
e
v
a
l
u
a
t
e
d
,
i
t
h
a
s
b
e
e
n
f
o
u
n
d
t
h
a
t
t
h
e
o
p
t
i
m
u
m
c
o
n
f
i
g
u
r
a
t
i
o
n
i
s
1
3
M
FC
C
,
1
2
ne
ur
ons
i
n
2
hi
dde
n
l
a
ye
r
s
.
T
he
ne
xt
s
t
e
p
i
s
de
t
e
r
m
i
ni
ng
w
he
t
he
r
t
hi
s
co
n
f
i
g
u
r
at
i
o
n
i
s
s
u
i
t
ab
l
e
f
o
r
al
l
M
F
C
C
ba
s
e
d
S
E
R
s
ys
t
e
m
s
.
T
he
ne
ut
r
a
l
e
m
ot
i
on
i
s
a
dde
d
i
nt
o
t
he
t
r
a
i
ni
ng
pool
.
T
he
f
i
r
s
t
r
e
s
ul
t
i
s
obt
a
i
ne
d
us
i
ng
t
he
fo
r
m
e
r
n
e
t
w
o
r
k
c
o
n
fi
g
u
r
a
t
i
o
n
o
f
1
3
M
F
C
C
,
1
2
n
e
u
r
o
n
s
i
n
2
h
i
d
d
e
n
l
a
y
e
r
s
r
e
s
u
l
t
i
n
g
a
b
e
s
t
c
a
s
e
r
e
c
o
g
n
i
t
i
o
n
r
at
e
o
f
9
6
.
6
7
%
w
i
t
h
p
r
o
ces
s
i
n
g
t
i
m
e
o
f
4
.
9
9
7
s
.
Th
e
s
t
e
p
s
o
f
o
p
t
i
m
i
z
a
t
i
o
n
a
r
e
t
h
e
n
r
e
p
e
a
t
e
d
–
opt
i
m
i
z
e
num
be
r
of
M
F
C
C
,
ne
ur
on,
a
nd
l
a
ye
r
,
re
p
e
a
t
i
n
g
t
ra
i
n
i
n
g
a
n
d
t
e
s
t
i
n
g
5
t
i
m
e
s
e
a
c
h
t
o
e
n
s
u
re
a
c
c
u
ra
c
y
.
A
f
t
e
r
o
p
t
i
m
i
z
a
t
i
o
n
,
t
h
e
i
d
e
a
l
c
o
n
f
i
g
u
ra
t
i
o
n
i
s
25
M
F
C
C
,
21
ne
ur
ons
,
a
nd
4
hi
dde
n
l
a
ye
r
s
f
or
4
e
m
ot
i
ons
a
c
hi
e
vi
ng
a
r
e
c
ogni
t
i
on
r
a
t
e
of
97.
1%
i
n
12.
013s
.
Th
e
c
o
n
f
u
s
i
o
n
m
a
t
r
i
x
a
n
d
n
e
t
w
o
r
k
c
o
n
f
i
g
u
r
a
t
i
o
n
a
r
e
d
i
s
p
l
a
y
e
d
i
n
F
i
g
u
r
e
6
a
n
d
7
.
Fi
g
u
r
e
6
.
O
p
t
i
m
i
z
e
d
N
e
t
w
o
r
k
C
o
n
f
u
s
i
o
n
M
a
t
r
i
x
0,5
0,7
0,9
1,1
1
4
7
10
13
16
19
22
25
28
Re
c
o
g
n
ti
o
n

Rate

Nu
mb
er
o
f

Neu
ro
n
s

Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
o
n
e
s
i
a
n
J
E
l
e
c
E
n
g
&
C
o
m
p
S
c
i
IS
S
N
:
2502
-
4752

Spe
e
c
h
E
m
ot
i
on
R
e
c
ogni
t
i
on
U
s
i
ng
D
e
e
p
F
e
e
df
or
w
ar
d
Ne
u
r
a
l
Ne
t
wo
r
k
(
Mu
h
a
m
m
a
d
F
a
h
r
e
z
a
A
l
g
h
i
f
a
r
i
)
559
Fi
g
u
r
e
7
.
O
p
t
i
m
i
z
e
d
D
e
e
p
Fe
e
d
f
o
w
a
r
d
N
e
u
r
a
l
N
e
t
w
o
r
k
f
o
r
4
E
m
o
t
i
o
n
s
On
c
e
t
h
e
i
d
e
a
l
n
e
t
wo
r
k
c
o
n
f
i
g
u
r
a
t
i
o
n
o
f
d
e
e
p
f
e
e
d
f
o
r
wa
r
d
n
e
u
r
a
l
n
e
t
wo
r
k
i
s
f
i
n
a
l
i
z
e
d
,
t
h
e
f
i
n
a
l
s
t
e
p
of
t
he
s
t
udy
i
s
va
l
i
da
t
i
on
of
t
he
c
us
t
om
e
m
ot
i
ona
l
da
t
a
ba
s
e
.
W
e
us
e
d
t
he
s
a
m
e
t
r
a
i
ne
d
ne
t
w
or
k
c
onf
i
gur
a
t
i
on
but
now
w
e
us
e
d
c
us
t
om
da
t
a
ba
s
e
.
N
ot
e
t
ha
t
,
a
s
e
xpl
a
i
ne
d
i
n
t
he
pr
e
vi
ous
s
e
c
t
i
on,
t
he
c
us
t
om
da
t
a
ba
s
e
w
a
s
re
c
o
rd
e
d
i
n
t
h
e
re
a
l
l
i
f
e
s
e
t
t
i
n
g
w
i
t
h
e
n
v
i
ro
n
m
e
n
t
a
l
n
o
i
s
e
.
A
s
s
h
o
w
n
i
n
F
i
g
u
re
8
(a
),
t
h
e
re
c
o
g
n
i
t
i
o
n
ra
t
e
i
s
ra
t
h
e
r
l
o
w
a
t
2
3
.
3
%
.
A
s
t
h
e
re
s
u
l
t
s
w
e
re
ra
t
h
er
u
n
s
at
i
s
f
y
i
n
g
at
2
3
.
3
%
r
eco
g
n
i
t
i
o
n
r
at
e,
a
n
ew
n
et
w
o
r
k
i
s
cr
eat
ed
.
T
h
i
s
n
et
w
o
r
k
u
s
es
t
h
e
cu
s
t
o
m
em
o
t
i
o
n
al
d
at
ab
as
e’
s
o
w
n
au
d
i
o
d
at
a
f
o
r
t
r
ai
n
i
n
g
an
d
t
es
t
i
n
g
.
T
h
e
re
s
u
l
t
s
i
n
F
i
g
u
re
8(
b)
ha
ve
s
how
n
t
o
be
m
or
e
pr
om
i
s
i
ng
a
t
70.
8%
r
e
c
ogni
t
i
on
r
a
t
e
.
Fr
o
m
t
h
e
r
es
u
l
t
s
o
b
t
ai
n
ed
,
o
n
e
can
o
b
s
er
v
e
t
h
at
t
h
e
n
ew
l
y
cal
i
b
r
at
ed
n
et
w
o
r
k
f
o
r
4
em
o
t
i
o
n
s
pe
r
f
or
m
e
d
be
t
t
e
r
(
97.
1%
)
t
ha
n
t
he
f
or
m
e
r
one
t
r
a
i
ne
d
f
or
3
e
m
ot
i
ons
(
96.
67%
)
,
a
t
a
t
r
a
de
of
f
of
a
si
g
n
i
f
i
c
a
n
t
l
y
h
i
g
h
e
r
p
r
o
c
e
ssi
n
g
t
i
m
e
.
T
h
e
r
e
f
o
r
e
,
i
f
t
h
e
o
b
j
e
c
t
i
v
e
i
s
t
o
m
a
x
i
mi
z
e
r
e
c
o
g
n
i
t
i
o
n
r
a
t
e
f
o
r
e
a
c
h
uni
que
num
be
r
of
e
m
ot
i
ons
,
i
t
i
s
r
e
c
om
m
e
nde
d
t
o
r
e
c
a
l
i
br
a
t
e
t
he
i
de
a
l
num
be
r
of
M
F
C
C
,
ne
ur
on
a
nd
l
a
ye
r
.
Wi
t
h
t
h
a
t
s
a
i
d
,
t
h
e
d
i
f
f
e
r
e
n
c
e
i
n
r
e
c
o
g
n
i
t
i
o
n
r
a
t
e
i
s
a
m
e
r
e
d
i
f
f
e
r
e
n
c
e
o
f
1
%
.
T
h
i
s
i
n
d
i
c
a
t
e
s
t
h
a
t
t
h
e
o
p
t
i
m
i
z
e
d
ne
t
w
or
k
co
n
f
i
g
u
r
at
i
o
n
can
b
e
u
s
ed
f
o
r
o
t
h
er
n
u
m
b
er
o
f
em
o
t
i
o
n
s
w
i
t
h
a
cer
t
ai
n
d
eg
r
ee
o
f
t
o
l
er
at
i
o
n
o
f
er
r
o
r
.
Th
e
r
e
s
u
l
t
s
o
b
t
a
i
n
e
d
i
n
t
h
e
o
p
t
i
m
i
z
a
t
i
o
n
o
f
n
e
t
w
o
r
k
h
a
s
s
h
o
w
n
t
o
h
a
v
e
a
s
i
g
n
i
f
i
c
a
n
t
i
m
p
r
o
v
e
m
e
n
t
co
m
p
ar
ed
t
o
t
h
e
S
E
R
r
es
ear
ch
co
u
n
t
er
p
ar
t
s
.
T
h
e
b
en
ch
m
ar
k
s
t
udy
c
onduc
t
e
d
by
[1
5
]
wi
t
h
a
n
u
n
m
e
n
t
i
o
n
e
d
da
t
a
ba
s
e
ha
s
a
c
hi
e
ve
d
a
92.
3%
pe
r
f
or
m
a
nc
e
w
hi
l
e
t
he
s
t
udy
by
[1
6
]
us
i
ng
t
he
s
a
m
e
B
e
r
l
i
n
E
m
ot
i
on
D
a
t
a
ba
s
e
ach
i
ev
ed
6
5
%
r
eco
g
n
i
t
i
o
n
.
C
o
m
p
ar
at
i
v
el
y
,
t
h
e
r
eco
g
n
i
t
i
o
n
r
at
e
a
ch
i
ev
ed
u
s
i
n
g
t
h
e
s
y
s
t
em
p
r
o
p
o
s
ed
i
n
t
h
i
s
st
u
d
y
i
s
9
7
.
1
%
.
A
f
a
c
t
o
r
t
o
c
o
n
s
i
d
e
r
i
s
t
h
e
v
a
r
i
a
t
i
o
n
o
f
p
a
r
a
m
e
t
e
r
s
u
s
e
d
.
I
n
t
h
i
s
s
t
u
d
y
,
M
F
C
C
i
s
t
h
e
s
o
l
e
i
n
p
u
t
wh
i
l
e
th
e
s
tu
d
y
c
o
n
d
u
c
te
d
b
y
[1
5
]
us
e
s
M
F
C
C
s
,
pe
r
cep
t
u
al
l
i
n
ear
p
r
ed
i
ct
i
v
e
(
P
L
P
s
)
,
an
d
F
i
l
t
er
b
an
k
s
(
F
B
A
N
K
s
)
.
By
t
h
e
o
r
y
,
a
d
d
i
n
g
m
o
r
e
p
a
r
a
m
e
t
e
r
s
s
h
o
u
l
d
i
m
p
r
o
v
e
t
h
e
r
e
c
o
g
n
i
t
i
o
n
r
a
t
e
a
t
t
h
e
c
o
s
t
o
f
p
r
o
c
e
s
s
i
n
g
c
o
m
p
l
e
x
i
t
y
an
d
t
i
m
e.
I
n
o
t
h
er
w
o
r
d
s
,
ach
i
ev
i
n
g
a
h
i
g
h
r
eco
g
n
i
t
i
o
n
r
at
e
u
s
i
n
g
l
es
s
p
ar
am
et
er
s
i
s
d
es
ir
a
b
le
.
An
o
t
h
e
r
f
a
c
t
o
r
t
h
a
t
s
h
o
u
l
d
b
e
u
n
d
e
r
s
t
o
o
d
t
h
a
t
t
h
e
n
u
m
b
e
r
o
f
e
m
o
t
i
o
n
a
n
a
l
y
z
e
d
.
I
n
t
h
i
s
s
t
u
d
y
,
o
n
l
y
3
an
d
4
em
o
t
i
o
n
s
w
er
e
an
al
y
zed
w
h
i
l
e
t
h
ei
r
s
t
u
d
y
at
t
em
p
t
s
t
o
p
r
o
ces
s
6
-
7
e
m
ot
i
ons
.
T
he
or
e
t
i
c
a
l
l
y,
i
nc
r
e
a
s
i
ng
th
e
n
u
m
b
e
r
o
f
e
m
o
tio
n
s
to
b
e
r
e
c
o
g
n
iz
e
d
w
ill
d
e
te
r
io
r
a
te
th
e
a
c
c
u
r
a
c
y
r
a
te
.
H
e
n
c
e
a
c
h
ie
v
in
g
h
ig
h
r
e
c
o
g
n
itio
n
ra
t
e
f
o
r
m
u
l
t
i
p
l
e
e
m
o
t
i
o
n
s
i
s
c
o
m
m
e
n
d
a
b
l
e
.
Wi
t
h
t
h
a
t
s
a
i
d
,
o
u
r
r
e
s
u
l
t
u
s
i
n
g
l
i
m
i
t
e
d
e
m
o
t
i
o
n
s
i
s
a
c
c
e
p
t
a
b
l
e
a
s
th
e
p
u
r
p
o
s
e
o
f
th
e
n
e
tw
o
r
k
is
to
b
e
la
te
r
e
m
p
lo
y
e
d
in
r
e
a
l
-
lif
e
s
itu
a
tio
n
.
T
h
e
4
e
m
ot
i
ons
a
na
l
yz
e
d
i
n
t
hi
s
s
t
udy
ar
e
t
h
e
p
r
i
m
ar
i
l
y
em
o
t
i
o
n
s
t
h
at
h
av
e
p
r
act
i
cal
ap
p
l
i
cat
i
o
n
s
.
M
ax
i
m
i
zi
n
g
t
h
e
r
eco
g
n
i
t
i
o
n
r
at
e
o
f
t
h
es
e
em
o
t
i
o
n
s
ar
e
o
f
a
h
i
g
h
er
p
r
i
o
r
i
t
y
.
(a
)
W
i
t
h
o
u
t
N
e
t
w
o
rk
R
e
t
ra
i
n
i
n
g
(b
)
W
i
t
h
N
e
t
w
o
rk
R
e
t
ra
i
n
i
n
g
Fi
g
u
r
e
8
.
Pe
r
f
o
r
m
a
n
c
e
o
f
t
h
e
SE
R
Sy
st
e
m
o
n
t
h
e
C
u
st
o
m
D
a
t
a
b
a
se
Evaluation Warning : The document was created with Spire.PDF for Python.

IS
S
N
:
2502
-
4752
In
d
o
n
e
s
i
a
n
J
E
l
e
c
E
n
g
&
C
o
m
p
S
c
i
,
Vo
l
.
10
, N
o
.
2
,
Ma
y
2
0
1
8
:
554
–
561
560
Fo
r
t
h
e
c
u
s
t
o
m
d
a
t
a
b
a
s
e
,
t
h
e
r
e
s
u
l
t
s
w
h
e
n
u
s
i
n
g
t
h
e
o
p
t
i
m
i
z
e
d
n
e
t
w
o
r
k
o
n
l
y
a
c
h
i
e
v
e
d
a
p
e
r
f
o
r
m
a
n
c
e
of
23.
3%
.
T
he
r
e
a
r
e
s
e
ve
r
a
l
hypot
he
s
i
s
t
o
e
xpl
a
i
n
t
he
poor
r
e
s
ul
t
.
T
he
pr
i
m
a
r
y
r
e
a
s
on
pr
opos
e
d
is
d
u
e
to
th
e
la
n
g
u
a
g
e
d
if
f
e
r
e
n
c
e
.
F
o
r
e
x
a
m
p
le
,
th
e
A
r
a
b
ic
la
n
g
u
a
g
e
m
a
y
s
o
u
n
d
m
o
r
e
in
te
n
s
e
th
a
n
s
a
y
,
th
e
tr
a
d
itio
n
a
l
Su
n
d
a
n
e
s
e
l
a
n
g
u
a
g
e
o
f
I
n
d
o
n
e
s
i
a
.
I
n
t
h
e
o
r
i
g
i
n
a
l
n
e
t
w
o
r
k
,
G
e
r
m
a
n
e
m
o
t
i
o
n
a
l
s
e
n
t
e
n
c
e
s
w
e
r
e
u
s
e
d
t
o
t
r
a
i
n
th
e
d
e
e
p
n
e
u
r
a
l
n
e
tw
o
r
k
w
h
ile
th
e
cu
s
t
o
m
d
at
ab
as
e
u
s
es
E
n
g
l
i
s
h
.
T
ak
i
n
g
t
h
i
s
i
n
t
o
acco
u
n
t
,
t
h
e
r
es
u
l
t
s
h
av
e
gr
e
a
t
l
y
i
m
pr
ove
d
w
he
n
t
r
a
i
ni
ng
us
i
ng
t
he
s
a
m
e
da
t
a
s
e
t
,
a
c
hi
e
vi
ng
70.
8%
i
n
be
s
t
c
a
s
e
pe
r
f
or
m
a
nc
e
.
T
hi
s
va
l
i
da
t
e
s
t
he
i
de
a
of
l
a
ngua
ge
be
i
ng
a
f
a
c
t
or
i
n
S
E
R
.
A
s
i
de
f
r
om
t
he
l
a
ngua
ge
di
f
fe
r
e
n
c
e
,
t
h
e
l
o
s
s
y
co
m
p
r
es
s
i
o
n
i
s
an
o
t
h
er
l
i
k
el
y
f
act
o
r
t
o
ex
p
l
ai
n
t
h
e
d
i
f
f
er
en
ce
i
n
p
er
f
o
r
m
an
ce.
4.
CO
NCL
US
I
O
NS
AND
F
UT
URE
W
O
RK
S
Af
t
e
r
e
x
t
e
n
s
i
v
e
e
x
p
e
r
i
m
e
n
t
a
t
i
o
n
,
S
E
R
s
y
s
t
e
m
h
a
s
b
e
e
n
i
m
p
l
e
m
e
n
t
e
d
a
n
d
o
p
t
i
m
i
z
e
d
.
T
h
e
o
p
t
i
m
u
m
co
n
f
i
g
u
r
at
i
o
n
f
o
r
S
E
R
i
s
1
3
M
F
C
C
,
1
2
ne
ur
ons
a
nd
2
l
a
ye
r
s
f
or
3
e
m
ot
i
ons
a
nd
25
M
F
C
C
,
21
ne
ur
ons
a
nd
4
la
y
e
r
s
f
o
r
4
e
m
o
tio
n
s
,
a
c
h
ie
v
in
g
a
to
ta
l
r
e
c
o
g
n
itio
n
r
a
te
o
f
9
6
.3
%
f
o
r
3
e
m
o
tio
n
s
a
n
d
9
7
.1
%
f
o
r
4
e
m
o
tio
n
s
.
I
f
th
e
o
b
j
e
c
tiv
e
is
to
m
a
x
im
iz
e
r
e
c
o
g
n
itio
n
r
a
te
f
o
r
e
a
c
h
u
n
iq
u
e
n
u
m
b
e
r
o
f
e
m
o
ti
ons
,
i
t
i
s
r
e
c
om
m
e
nde
d
t
o
re
c
a
l
i
b
ra
t
e
a
n
d
re
t
ra
i
n
t
h
e
n
e
t
w
o
rk
,
o
t
h
e
rw
i
s
e
t
h
e
f
o
rm
e
r
n
e
t
w
o
rk
i
s
s
t
i
l
l
u
t
i
l
i
z
a
b
l
e
g
i
v
e
n
a
c
e
rt
a
i
n
d
e
g
re
e
o
f
re
c
o
g
n
i
t
i
o
n
e
rro
r.
In
S
E
R
,
l
a
n
g
u
a
g
e
u
s
e
d
i
s
a
f
a
c
t
o
r
t
o
b
e
c
o
n
s
i
d
e
re
d
f
o
r
re
c
o
g
n
i
t
i
o
n
ra
t
e
.
F
u
t
u
re
re
s
e
a
rc
h
in
c
lu
d
e
th
e
i
m
p
r
o
v
em
en
t
o
n
d
eep
l
ear
n
i
n
g
co
n
f
i
g
u
r
at
i
o
n
,
d
i
f
f
er
en
t
d
at
ab
as
e,
an
d
r
eal
l
i
f
e
ap
p
l
i
cat
i
o
n
s
.
ACK
NO
W
L
E
DG
E
M
E
NT
Th
e
a
u
t
h
o
r
s
w
o
u
l
d
l
i
k
e
t
o
e
x
p
r
e
s
s
t
h
e
i
r
g
r
a
t
i
t
u
d
e
t
o
t
h
e
M
a
l
a
y
s
i
a
n
M
i
n
i
s
t
r
y
o
f
H
i
g
h
e
r
Ed
u
c
a
t
i
o
n
(M
O
H
E
),
w
h
i
c
h
h
a
s
p
ro
v
i
d
e
d
f
u
n
d
i
n
g
f
o
r
t
h
e
re
s
e
a
rc
h
t
h
ro
u
g
h
t
h
e
F
u
n
d
a
m
e
n
t
a
l
R
e
s
e
a
rc
h
G
ra
n
t
S
c
h
e
m
e
,
FR
G
S1
5
-
194
-
0435.
RE
F
E
RE
NCE
S
[1
]
A.
J
o
s
h
i
,
R.
K
a
u
r
,
"
A
S
t
u
d
y
o
f
s
p
e
e
c
h
e
m
o
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
m
e
t
h
o
d
s
,
"
In
t
.
J
.
C
o
m
p
u
t
.
S
c
i
.
M
o
b
.
C
o
m
p
u
t
.
(IJ
C
S
M
C
)
,
vol
.
2,
pp.
28
-
31,
2013.
[
2]
M.
E
l
A
y
ad
i
,
M
.
S
.
K
am
el
,
F.
K
a
r
r
a
y
,
"
Su
r
v
e
y
o
n
s
p
e
e
c
h
e
m
o
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
:
Fe
a
t
u
r
e
s
,
c
l
a
s
s
i
f
i
c
a
t
i
o
n
s
c
h
e
m
e
s
,
an
d
d
at
ab
ases,
"
Pa
t
t
e
r
n
Re
c
o
g
n
i
t
i
o
n
, v
o
l. 4
4
, p
p
. 5
7
2
-
587,
2011.
[3
]
T.
S.
G
u
n
a
w
a
n
,
M
.
F
.
A
lg
h
if
a
r
i,
M
.
A
.
M
o
r
s
h
id
i,
M.
K
a
r
t
i
w
i
,
"
A
R
e
v
i
e
w
o
n
S
p
e
e
c
h
E
m
o
t
i
on
R
e
c
ogni
t
i
on
Al
g
o
r
i
t
h
ms
,
"
In
d
o
n
e
s
i
a
n
J
o
u
r
n
a
l
o
f
E
l
e
c
t
r
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
a
n
d
In
f
o
r
m
a
t
i
c
s
, v
o
l. 6
, 2
0
1
8
.
[4
]
A.
Ko
n
a
r
,
A.
C
h
a
k
r
a
b
o
r
t
y
,
Em
o
t
i
o
n
Re
c
o
g
n
i
t
i
o
n
:
A
Pa
t
t
e
r
n
An
a
l
y
s
i
s
Ap
p
r
o
a
c
h
, J
o
h
n
W
ile
y
&
S
o
n
s
, 2
0
1
4
.
[5
]
S.
R
.
B
a
n
d
e
l
a
,
T.
K
.
K
u
m
a
r
,
"S
t
r
e
s
s
e
d
s
p
e
e
c
h
e
m
ot
i
on
r
e
c
ogni
t
i
on
us
i
ng
f
e
at
ur
e
f
us
i
on
of
t
e
age
r
e
ne
r
gy
ope
r
at
or
and
M
F
C
C
,
"
in
2
0
1
7
8
th
I
n
te
r
n
a
tio
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
tin
g
,
C
o
m
m
u
n
ic
a
tio
n
a
n
d
N
e
tw
o
r
k
in
g
T
e
c
h
n
o
lo
g
ie
s
(IC
C
C
N
T
),
p
p
.
1
-
5,
2017.
[6
]
S.
T.
Sa
s
t
e
,
S.
M
.
J
a
g
d
a
l
e
,
"E
m
o
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
f
r
o
m
sp
eech
u
si
n
g
M
F
C
C
a
n
d
D
W
T
f
o
r
secu
ri
t
y
syst
em
,
"
in
2
0
1
7
In
t
e
rn
a
t
i
o
n
a
l
c
o
n
f
e
re
n
c
e
o
f
E
l
e
c
t
ro
n
i
c
s
,
C
o
m
m
u
n
i
c
a
t
i
o
n
a
n
d
A
e
ro
s
p
a
c
e
T
e
c
h
n
o
l
o
g
y
(IC
E
C
A
),
p
p
.
7
0
1
-
704,
2017.
[7
]
E.
G
o
p
i
,
Di
g
i
t
a
l
s
p
e
e
c
h
p
r
o
c
e
s
s
i
n
g
u
s
i
n
g
M
a
t
l
a
b
, S
p
r
in
g
e
r
, 2
0
1
4
.
[8
]
S.
Zh
a
n
g
,
S.
Zh
a
n
g
,
T.
H
u
a
n
g
,
a
n
d
W
.
G
a
o
,
"
Sp
e
e
c
h
Em
o
t
i
o
n
R
e
c
o
g
n
i
t
i
o
n
U
s
i
n
g
D
e
e
p
C
o
n
v
o
l
u
t
i
o
n
a
l
N
e
u
r
a
l
Ne
t
wo
r
k
a
n
d
Di
s
c
r
i
mi
n
a
n
t
T
e
mp
o
r
a
l
P
y
r
a
mi
d
M
a
t
c
h
i
n
g
,
"
IE
E
E
T
r
a
n
s
a
c
t
i
o
n
s
o
n
M
u
l
t
i
m
e
d
i
a
,
v
o
l.
P
P
,
p
p
.
1
-
1,
2017.
[9
]
C.
W
.
H
u
a
n
g
,
S.
S.
N
a
r
a
y
a
n
a
n
,
"D
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
r
e
c
u
r
r
e
n
t
ne
ur
al
ne
t
w
or
k
w
i
t
h
at
t
e
nt
i
on
m
e
c
hani
s
m
f
or
r
obus
t
sp
eech
em
o
t
i
o
n
reco
g
n
i
t
i
o
n
,
"
in
2
0
1
7
I
E
E
E
I
n
te
r
n
a
tio
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
M
u
ltim
e
d
ia
a
n
d
E
x
p
o
(
I
C
M
E
)
,
p
p
.
5
8
3
-
588,
2017.
[1
0
]
V.
T
i
wa
r
i
,
"
M
F
C
C
a
n
d
i
t
s
a
p
p
l
i
c
a
t
i
o
n
s
i
n
s
p
e
a
k
e
r
r
e
c
o
g
n
i
t
i
o
n
,
"
In
t
e
r
n
a
t
i
o
n
a
l
j
o
u
r
n
al
on
e
m
e
r
gi
ng
t
e
c
hnol
ogi
e
s
,
vol
.
1,
pp.
19
-
22,
2010.
[1
1
]
Y.
L
e
C
u
n
,
Y.
B
e
n
g
i
o
,
G.
Hi
n
t
o
n
,
"
De
e
p
l
e
a
r
n
i
n
g
,
"
Na
t
u
r
e
, v
o
l. 5
2
1
, p
p
. 4
3
6
, 2
0
1
5
.
[1
2
]
F.
B
u
r
k
h
a
r
d
t
,
A
.
Pa
e
s
c
h
k
e
,
M
.
R
o
l
f
e
s
,
W
.
F.
Se
n
d
l
m
e
i
e
r
,
a
n
d
B
.
W
e
i
s
s
,
"
A
d
a
t
a
b
a
s
e
o
f
g
e
r
m
a
n
e
m
o
t
i
o
n
a
l
s
p
e
e
c
h
,
"
pp.
1517
-
1520.
[1
3
]
D.
P
r
a
v
e
n
a
,
S
.
Na
n
d
h
a
k
u
ma
r
,
D.
Go
v
i
n
d
,
"
S
i
g
n
i
f
i
c
a
n
c
e
o
f
n
a
t
u
r
a
l
e
l
i
c
i
t
a
t
i
o
n
i
n
d
e
v
e
l
o
p
i
n
g
s
i
mu
l
a
t
e
d
f
u
l
l
b
l
o
wn
sp
eech
em
o
t
i
o
n
d
at
ab
ases,
"
p
p
.
2
6
1
-
265.
[1
4
]
A.
C
h
a
n
d
r
a
n
,
P
.
Du
p
l
e
x
,
G.
Di
v
u
,
De
v
e
l
o
p
me
n
t
o
f
s
p
e
e
c
h
e
mo
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
s
y
s
t
em
u
si
n
g
d
eep
b
el
i
ef
n
et
w
o
rks
i
n
ma
l
a
y
a
l
a
m
l
a
n
g
u
a
g
e
, 2
0
1
7
.
[1
5
]
J.
N
i
u
,
Y
.
Q
i
an
,
K.
Yu
,
"
Ac
o
u
s
t
i
c
e
mo
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
u
s
i
n
g
d
e
e
p
n
e
u
r
a
l
n
e
t
wo
r
k
,
"
p
p
.
1
2
8
-
132,
2014.
[1
6
]
X.
Z
h
o
u
,
J
.
Gu
o
,
R.
Bi
e
,
"D
e
e
p
L
e
a
r
n
i
n
g
B
a
s
e
d
A
f
f
e
c
t
i
v
e
M
o
d
e
l
f
o
r
S
p
e
e
c
h
E
m
o
t
i
o
n
R
e
c
o
g
ni
t
i
on,
"
in
2
0
1
6
I
n
tl
IE
E
E
C
o
n
f
e
re
n
c
e
s
o
n
U
b
i
q
u
i
t
o
u
s
In
t
e
l
l
i
g
e
n
c
e
&
C
o
m
p
u
t
i
n
g
,
A
d
v
a
n
c
e
d
a
n
d
T
ru
s
t
e
d
C
o
m
p
u
t
i
n
g
,
S
c
a
l
a
b
l
e
Co
m
p
u
t
i
n
g
a
n
d
Co
m
m
u
n
i
c
a
t
i
o
n
s
,
Cl
o
u
d
a
n
d
Bi
g
D
a
t
a
Co
m
p
u
t
i
n
g
,
I
n
t
e
r
n
e
t
o
f
P
e
o
p
l
e
,
a
n
d
S
m
a
r
t
W
o
r
l
d
Co
n
g
r
e
s
s
(U
IC
/
A
T
C
/
S
c
a
l
C
o
m
/
C
B
D
C
o
m
/
Io
P
/
S
m
a
rt
W
o
rl
d
),
p
p
.
8
4
1
-
846,
2016.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
o
n
e
s
i
a
n
J
E
l
e
c
E
n
g
&
C
o
m
p
S
c
i
IS
S
N
:
2502
-
4752

Spe
e
c
h
E
m
ot
i
on
R
e
c
ogni
t
i
on
U
s
i
ng
D
e
e
p
F
e
e
df
or
w
ar
d
Ne
u
r
a
l
Ne
t
wo
r
k
(
Mu
h
a
m
m
a
d
F
a
h
r
e
z
a
A
l
g
h
i
f
a
r
i
)
561
BI
O
G
R
A
P
H
I
ES
O
F
A
U
TH
O
R
S
Mu
h
a
m
m
a
d
F
a
h
r
e
z
a
A
l
g
h
i
f
a
r
i
h
as
co
m
p
l
et
ed
h
i
s
B
.
E
n
g
.
(
H
o
n
s)
d
eg
r
ee
i
n
E
l
ect
r
o
n
i
cs:
Co
m
p
u
t
e
r
I
n
f
o
r
m
a
t
i
o
n
E
n
g
i
n
e
e
r
i
n
g
f
r
o
m
I
n
t
e
r
n
a
t
i
o
n
a
l
I
s
l
a
m
i
c
U
n
i
v
e
r
s
i
t
y
M
a
l
a
y
s
i
a
(
I
I
U
M
)
i
n
2018.
H
i
s
r
e
s
e
a
r
c
h
in
te
r
e
s
ts
a
r
e
in
s
ig
n
a
l
p
r
o
c
e
s
s
in
g
,
a
r
tif
ic
ia
l
in
te
llig
e
n
c
e
a
n
d
a
f
f
e
c
tiv
e
co
m
p
u
t
i
n
g
.
H
e
r
ecei
v
ed
a
b
est
F
Y
P
aw
ar
d
f
r
o
m
I
E
E
E
S
i
g
n
al
P
r
o
cessi
n
g
–
Ma
l
a
y
s
i
a
c
h
a
p
t
e
r
.
Cu
r
r
e
n
t
l
y
,
h
e
i
s
w
o
r
k
i
n
g
o
n
t
h
e
a
p
p
l
i
c
a
t
i
o
n
o
f
s
p
e
e
c
h
e
m
o
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
f
o
r
s
u
i
c
i
d
e
p
r
e
v
e
n
t
i
on.
Te
d
d
y
S
u
r
y
a
G
u
n
a
w
a
n
re
c
e
i
v
e
d
h
i
s
B
E
n
g
d
e
g
re
e
i
n
E
l
e
c
t
ri
c
a
l
E
n
g
i
n
e
e
ri
n
g
w
i
t
h
c
u
m
l
a
u
d
e
aw
ar
d
f
r
o
m
I
n
st
i
t
u
t
T
ek
n
o
l
o
g
i
B
an
d
u
n
g
(
I
T
B
)
,
I
n
d
o
n
esi
a
i
n
1
9
9
8
.
H
e
o
b
t
ai
n
ed
h
i
s
M
.
E
n
g
d
eg
r
ee
in
2
0
0
1
f
r
o
m
th
e
S
c
h
o
o
l
o
f
C
o
m
p
u
te
r
E
n
g
in
e
e
r
in
g
a
t
N
a
n
y
a
n
g
T
e
c
h
n
o
lo
g
ic
al
U
n
i
v
er
si
t
y
,
Si
n
g
a
p
o
r
e
,
a
n
d
Ph
D
d
e
g
r
e
e
i
n
2
0
0
7
f
r
o
m
t
h
e
Sc
h
o
o
l
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
r
i
n
g
a
n
d
Te
l
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
,
Th
e
U
n
i
v
e
r
s
i
t
y
o
f
N
e
w
So
u
t
h
W
a
l
e
s
,
A
u
s
t
r
a
l
i
a
.
H
i
s
r
e
s
e
a
r
c
h
i
n
t
e
r
e
s
t
s
a
r
e
i
n
sp
eech
an
d
au
d
i
o
p
r
o
cessi
n
g
,
b
i
o
m
ed
i
cal
si
g
n
al
p
r
o
cessi
n
g
an
d
i
n
s
tr
u
m
e
n
ta
tio
n
,
im
a
g
e
a
n
d
v
id
e
o
pr
oc
e
s
s
i
ng,
a
nd
pa
r
a
l
l
e
l
c
om
put
i
ng.
H
e
i
s
c
ur
r
e
nt
l
y
a
n
I
E
E
E
S
e
ni
or
M
e
m
be
r
(
s
i
nc
e
2012)
,
w
a
s
ch
ai
r
m
an
o
f
I
E
E
E
I
n
st
r
u
m
en
t
at
i
o
n
an
d
M
easu
r
em
en
t
S
o
ci
et
y
–
Ma
l
a
y
s
i
a
S
e
c
t
i
o
n
(
2
0
1
3
a
n
d
2
0
1
4
)
,
As
s
o
c
i
a
t
e
P
r
o
f
e
s
s
o
r
(
s
i
n
c
e
2
0
1
2
)
,
He
a
d
o
f
De
p
a
r
t
me
n
t
(
2
0
1
5
-
2016)
a
t
D
e
pa
r
t
m
e
nt
of
E
l
e
c
t
r
i
c
a
l
an
d
C
o
m
p
u
t
er
E
n
g
i
n
eer
i
n
g
,
an
d
H
ead
o
f
P
r
o
g
r
am
m
e
A
ccr
ed
i
t
at
i
o
n
an
d
Q
u
al
i
t
y
A
ssu
r
an
ce
f
o
r
Fa
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
(
s
i
n
c
e
2
0
1
7
)
,
I
n
t
e
r
n
a
t
i
o
n
a
l
I
s
l
a
m
i
c
U
n
i
v
e
r
s
i
t
y
M
a
l
a
y
s
i
a
.
H
e
i
s
C
h
a
r
t
e
r
e
d
En
g
i
n
e
e
r
(
I
ET,
U
K
)
a
nd
I
ns
i
nyur
P
r
of
e
s
i
ona
l
M
a
dya
(
P
I
I
,
I
ndone
s
i
a
)
s
i
nc
e
2016.
Mi
r
a
K
a
r
t
i
w
i
co
m
p
l
et
ed
h
er
st
u
d
i
es
at
t
h
e
U
n
i
v
er
si
t
y
o
f
W
o
l
l
o
n
g
o
n
g
,
A
u
st
r
al
i
a
r
esu
l
t
i
n
g
i
n
t
h
e
fo
l
l
o
w
i
n
g
d
e
g
r
e
e
s
b
e
i
n
g
c
o
n
fe
r
r
e
d
:
B
a
c
h
e
l
o
r
o
f
C
o
m
m
e
r
c
e
i
n
B
u
si
n
e
ss
I
n
fo
r
m
a
t
i
o
n
S
y
st
e
m
s,
Ma
s
t
e
r
i
n
I
n
f
o
r
m
a
t
i
o
n
S
y
s
t
e
m
s
i
n
2
0
0
1
a
n
d
h
e
r
D
o
c
t
o
r
o
f
P
h
i
l
o
s
o
p
h
y
i
n
2
0
0
9
.
S
h
e
i
s
c
ur
r
e
nt
l
y
a
n
As
s
o
c
i
a
t
e
P
r
o
f
e
s
s
o
r
i
n
De
p
a
r
t
me
n
t
o
f
I
n
f
o
r
ma
t
i
o
n
S
y
s
t
e
ms
,
Ku
l
i
y
y
a
h
o
f
I
n
f
o
r
ma
t
i
o
n
a
n
d
Co
m
m
u
n
i
c
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
I
n
t
e
r
n
a
t
i
o
n
a
l
I
s
l
a
m
i
c
U
n
i
v
e
r
s
i
t
y
M
a
l
a
y
s
i
a
.
H
e
r
r
e
s
e
a
r
c
h
i
n
t
e
r
e
s
t
s
in
c
lu
d
e
e
le
c
tr
o
n
ic
c
o
m
m
e
r
c
e
, d
a
ta
m
in
in
g
, e
-
he
a
l
t
h
a
nd
m
obi
l
e
a
ppl
ic
a
tio
n
s
d
e
v
e
lo
p
m
e
n
t.
Evaluation Warning : The document was created with Spire.PDF for Python.