I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
23
,
No
.
2
,
A
u
g
u
s
t
2
0
2
1
,
p
p
.
1
2
1
2
~
121
8
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
23
.i
2
.
pp
1
2
1
2
-
1
2
1
8
1212
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Ara
bic spea
ker
re
co
g
nition us
ing
HM
M
J
a
bb
a
r
S.
H
us
s
ein
1
,
Ab
du
lk
a
dh
im
A.
Sa
lm
a
n
2
,
T
hm
er
R
.
Sa
ee
d
3
1
Ka
rb
a
la Un
iv
e
rsity
,
c
o
ll
e
g
e
o
f
E
n
g
i
n
e
e
rin
g
,
Ka
rb
a
la,
Ira
q
2
Tec
h
n
ica
l
In
stit
u
te
o
f
Ka
rb
a
la,
A
l
-
F
u
ra
t
Al
-
Aw
sa
t
Tec
h
n
ica
l
U
n
iv
e
rsity
,
Ka
rb
a
la,
Ira
q
3
De
p
a
rtme
n
t
o
f
El
e
c
tri
c
a
l
E
n
g
i
n
e
e
rin
g
,
U
n
iv
e
rsit
y
o
f
Tec
h
n
o
lo
g
y
,
Ba
g
h
d
a
d
,
Ira
q
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
an
1
2
,
2
0
2
1
R
ev
is
ed
J
u
n
3
0
,
2
0
2
1
Acc
ep
ted
J
u
l 1
3
,
2
0
2
1
In
th
is
p
a
p
e
r,
a
n
e
w
su
g
g
e
ste
d
sy
ste
m
fo
r
sp
e
a
k
e
r
re
c
o
g
n
i
ti
o
n
b
y
u
sin
g
h
id
d
e
n
m
a
rk
o
v
m
o
d
e
l
(
HHM
)
a
l
g
o
rit
h
m
.
M
a
n
y
re
se
a
rc
h
e
s
h
a
v
e
b
e
e
n
writt
e
n
in
th
is
s
u
b
jec
t,
e
sp
e
c
ially
b
y
H
M
M
.
Ara
b
ic
lan
g
u
a
g
e
is
o
n
e
o
f
th
e
d
iffi
c
u
lt
lan
g
u
a
g
e
s
a
n
d
th
e
wo
rk
wit
h
it
is
v
e
ry
li
tt
le,
a
lso
,
th
e
wo
rk
h
a
s
b
e
e
n
d
o
n
e
fo
r
tex
t
d
e
p
e
n
d
e
n
t
s
y
ste
m
wh
e
re
HMM
is
v
e
ry
e
ffe
c
ti
v
e
a
n
d
t
h
e
a
lg
o
rit
h
m
train
e
d
a
t
th
e
wo
r
d
lev
e
l.
O
n
e
th
e
p
ro
b
lem
s
in
s
u
c
h
s
y
ste
m
s
is
th
e
n
o
ise
,
s
o
we
tak
e
it
in
c
o
n
sid
e
ra
ti
o
n
b
y
a
d
d
in
g
a
d
d
it
iv
e
w
h
it
e
g
a
u
ss
ian
n
o
ise
(
AWG
N
)
to
th
e
sp
e
e
c
h
sig
n
a
ls
to
se
e
it
s
e
ffe
c
t.
He
re
,
we
u
se
d
HMM
with
n
e
w
a
lg
o
rit
h
m
wit
h
o
n
e
sta
te,
w
h
e
re
t
wo
o
f
t
h
e
se
c
o
m
p
o
n
e
n
ts,
i.
e
.
(π
a
n
d
A)
a
re
re
m
o
v
e
d
.
Th
is
g
i
v
e
e
x
trem
e
ly
a
c
c
e
lera
tes
th
e
train
in
g
a
n
d
tes
ti
n
g
sta
g
e
s
o
f
re
c
o
g
n
it
i
o
n
s
p
e
e
d
s
with
lo
we
st
m
e
m
o
ry
u
sa
g
e
,
a
s
se
e
n
in
th
e
wo
rk
.
T
h
e
re
su
lt
s
sh
o
w
a
n
e
x
c
e
ll
e
n
t
o
u
tco
m
e
.
1
0
0
%
re
c
o
g
n
it
i
o
n
ra
te
fo
r
t
h
e
tes
ted
d
a
ta,
a
b
o
u
t
9
1
.
6
%
re
c
o
g
n
it
io
n
ra
te wit
h
AWG
N n
o
ise
.
K
ey
w
o
r
d
s
:
Ar
ab
ic
lan
g
u
a
g
e
Hid
d
en
m
ar
k
o
v
m
o
d
el
Sp
ea
k
er
r
ec
o
g
n
itio
n
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
J
ab
b
ar
S.
Hu
s
s
ein
Dep
ar
tm
en
t o
f
Pro
s
th
etic
&
O
r
th
etic
E
n
g
in
ee
r
i
n
g
Kar
b
ala
Un
iv
er
s
ity
,
co
lleg
e
o
f
E
n
g
in
ee
r
i
n
g
Kar
b
ala,
I
r
aq
jab
b
a
r
.
s
alm
an
@
u
o
k
er
b
ala.
ed
u
.
iq
1.
I
NT
RO
D
UCT
I
O
N
T
h
ese
d
ay
s
,
th
e
u
p
h
ea
v
al
in
th
e
h
ar
d
war
e
in
n
o
v
atio
n
g
iv
es
a
wid
e
r
eg
io
n
to
t
h
e
s
p
ec
ialis
ts
f
o
r
tak
in
g
ca
r
e
o
f
c
o
m
p
lex
is
s
u
es,
f
o
r
e
x
am
p
le,
s
p
ea
k
er
r
ec
o
g
n
itio
n
(
SR
)
wi
th
n
o
is
y
en
v
ir
o
n
m
en
t.
T
h
e
s
ig
n
if
ican
ce
o
f
SR
ca
n
b
e
s
ee
n
th
r
o
u
g
h
its
ap
p
licatio
n
s
in
s
ec
u
r
ity
a
n
d
r
ec
o
n
n
aiss
an
ce
f
r
am
ewo
r
k
s
.
I
n
th
e
wr
itin
g
,
d
i
f
f
er
en
t
p
r
o
ce
d
u
r
es
f
o
r
SR
h
av
e
b
ee
n
illu
s
tr
ated
,
h
id
d
en
m
a
r
k
o
v
m
o
d
el
(
HM
M)
[
1
]
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
[
2
]
,
lin
ea
r
d
is
cr
im
in
an
t
an
aly
s
is
[
3
]
,
in
d
e
p
en
d
e
n
t
co
m
p
o
n
en
t
an
aly
s
is
[
4
]
,
p
r
in
cip
al
co
m
p
o
n
en
t
an
aly
s
is
[
5
]
,
q
u
an
tizatio
n
o
f
m
el
-
f
r
e
q
u
en
c
y
ce
p
s
tr
al
co
ef
f
icien
ts
[
6
]
an
d
ar
tific
ial
n
eu
r
al
n
etwo
r
k
[
7
]
.
I
n
s
p
ite
o
f
d
if
f
er
e
n
t
p
r
o
ce
d
u
r
es
wh
ich
n
ee
d
to
r
etr
ain
th
e
f
r
am
ewo
r
k
if
th
e
r
e
s
h
o
u
ld
b
e
a
n
o
cc
u
r
r
en
ce
o
f
r
ef
r
esh
in
g
th
e
d
atab
ase
,
th
e
HM
M
ca
n
b
e
u
tili
ze
d
s
o
th
at
ea
ch
m
o
d
el
is
in
d
e
p
e
n
d
en
tly
p
r
e
p
ar
ed
.
I
n
d
i
f
f
er
en
t
wo
r
d
s
,
ad
d
in
g
o
r
ex
p
ellin
g
an
y
in
d
iv
i
d
u
al
t
o
/f
r
o
m
th
e
f
r
a
m
ewo
r
k
ca
n
b
e
ef
f
ec
tiv
ely
p
er
f
o
r
m
e
d
with
o
u
t
t
h
e
n
ee
d
to
r
etr
ain
th
e
f
r
am
ewo
r
k
.
T
h
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
a
n
d
r
e
c
o
g
n
i
t
i
o
n
o
f
t
h
e
A
r
a
b
i
c
l
a
n
g
u
e
w
o
r
d
s
i
s
t
h
e
m
o
t
i
v
a
t
i
n
g
t
o
p
i
c
i
n
t
h
e
a
p
p
l
i
c
a
t
i
o
n
s
o
f
t
h
e
A
r
a
b
i
c
c
o
m
p
u
t
e
r
i
n
t
e
r
f
a
c
e
.
T
h
e
c
o
m
p
u
t
e
r
i
n
t
e
r
f
a
c
e
i
s
a
s
i
g
n
i
f
i
c
a
n
t
m
e
a
n
s
i
n
t
h
e
i
n
t
e
l
l
i
g
e
n
t
s
t
r
u
c
t
u
r
e
s
a
n
d
t
h
e
t
e
c
h
n
o
l
o
g
i
e
s
.
T
h
e
L
i
n
g
u
i
s
t
i
c
r
e
c
o
g
n
i
t
i
o
n
i
s
t
a
l
k
i
n
g
r
e
c
o
g
n
i
t
i
o
n
,
a
n
d
i
t
i
s
c
h
a
r
a
c
t
e
r
i
z
e
d
s
u
c
h
a
s
t
h
e
m
e
t
h
o
d
t
o
v
a
r
y
i
n
g
o
v
e
r
a
c
o
u
s
t
i
c
d
i
s
c
o
u
r
s
e
s
i
g
n
a
l
s
t
o
i
t
s
c
o
n
n
e
c
t
i
n
g
s
e
t
o
f
w
o
r
d
s
o
r
o
t
h
e
r
l
a
n
g
u
a
g
e
u
n
i
t
s
[
8
]
-
[
12]
.
Sp
ea
k
er
r
ec
o
g
n
itio
n
is
a
m
u
lti
-
d
is
cip
lin
ar
y
in
n
o
v
atio
n
wh
ic
h
u
tili
ze
s
th
e
v
o
ca
l
f
ea
tu
r
es
o
f
s
p
ea
k
er
s
to
in
f
er
d
ata
ab
o
u
t
th
eir
ch
ar
a
c
ter
s
.
I
t
is
a
p
ar
t
o
f
b
io
m
etr
ics
th
at
m
ig
h
t
b
e
u
tili
ze
d
f
o
r
d
is
tin
g
u
is
h
in
g
p
r
o
o
f
,
ch
ec
k
,
an
d
r
ec
o
g
n
itio
n
o
f
in
d
i
v
id
u
al
s
p
ea
k
er
s
,
with
th
e
ca
p
a
city
o
f
d
etec
tio
n
,
tr
ac
k
in
g
,
an
d
s
eg
m
en
tatio
n
b
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
A
r
a
b
ic
s
p
ea
ke
r
r
ec
o
g
n
itio
n
u
s
in
g
HMM
(
Ja
b
b
a
r
S
.
Hu
s
s
ein
)
1213
ex
ten
s
io
n
.
Sp
ea
k
er
r
ec
o
g
n
itio
n
an
d
s
p
ea
k
er
c
h
ec
k
s
tr
u
ctu
r
e
a
b
ig
g
er
co
n
tr
o
l
o
f
s
p
ea
k
er
class
if
icatio
n
[
1
3
]
.
Sp
ea
k
er
r
ec
o
g
n
itio
n
attem
p
ts
to
f
ig
u
r
e
o
u
t
wh
ich
s
p
ea
k
er
p
r
o
d
u
ce
d
a
d
is
co
u
r
s
e
s
ig
n
al
th
o
u
g
h
s
p
ea
k
er
ch
ec
k
af
f
ir
m
s
if
th
e
p
a
r
t
o
f
th
e
d
is
c
o
u
r
s
e
h
as
a
p
lace
with
t
h
e
p
er
s
o
n
wh
o
alleg
atio
n
it.
I
t
o
u
g
h
t
to
b
e
n
o
ticed
th
at
th
er
e
ar
e
two
s
o
r
ts
o
f
s
p
ea
k
er
r
ec
o
g
n
itio
n
,
wh
ich
a
r
e;
tex
t
i
n
d
ep
en
d
en
t
an
d
tex
t
d
e
p
en
d
e
n
t
[
1
4
]
.
T
h
is
p
ap
er
will
an
y
way
c
o
n
ce
n
t
r
ate
o
n
t
ex
t
d
ep
e
n
d
en
t
s
p
ea
k
er
r
ec
o
g
n
itio
n
.
Pre
s
en
t
co
n
ten
t
tex
t
d
ep
en
d
en
t
p
r
o
d
u
ce
s
s
en
s
ib
le
o
u
tco
m
es,
y
et
at
th
e
s
am
e
tim
e
d
o
n
o
t h
av
e
t
h
e
f
u
n
d
am
en
tal
ex
ec
u
tio
n
o
n
th
e
o
f
f
c
h
an
ce
th
at
th
ey
ar
e
to
b
e
u
tili
ze
d
b
y
th
e
o
v
er
all
p
o
p
u
latio
n
(
f
o
r
e
x
am
p
le
liv
e
test
in
g
)
.
So
as
to
les
s
en
th
e
co
m
p
licated
n
atu
r
e
o
f
SR
f
r
am
ewo
r
k
th
a
t
u
tili
ze
s
HM
M,
a
f
ew
p
r
o
ce
d
u
r
es
h
av
e
b
ee
n
attem
p
ted
,
wh
e
r
e
th
e
m
o
s
t
tr
an
s
ce
n
d
en
t
s
tr
ateg
y
is
th
e
d
ec
r
ea
s
e
o
f
s
p
ea
k
er
f
ile
s
ize
u
tili
zin
g
o
n
e
o
f
th
e
tr
an
s
f
o
r
m
atio
n
s
tr
ateg
ies,
f
o
r
ex
am
p
le,
d
is
cr
ete
wav
elet
t
r
an
s
f
o
r
m
atio
n
(
DW
T
)
[
1
]
an
d
d
is
cr
ete
co
s
in
e
tr
an
s
f
o
r
m
atio
n
[
1
5
]
.
T
h
e
n
a
g
ain
,
th
e
d
o
wn
s
id
e
o
f
ac
c
o
m
p
lis
h
in
g
f
u
r
th
er
d
ec
r
ea
s
e
i
n
th
e
f
r
am
ewo
r
k
'
s
u
n
p
r
e
d
ictab
ilit
y
is
th
e
im
p
r
o
p
er
n
u
m
b
er
o
f
HM
M
s
tates
u
tili
ze
d
[
1
6
]
,
[
1
7
]
,
wh
er
e
th
is
d
is
ad
v
an
tag
e
is
u
n
d
er
s
to
o
d
b
y
u
tili
zin
g
o
n
e
-
s
tate
HM
M.
I
n
d
is
co
v
er
in
g
th
is
to
p
ic,
p
r
im
ar
y
,
a
th
eo
r
y
p
ar
t c
o
v
er
in
g
th
e
c
o
n
ce
p
t
o
f
MH
H
with
o
n
e
s
tate
[
1
7
]
an
d
th
e
m
et
h
o
d
o
f
d
ec
r
ea
s
in
g
th
e
s
ize
o
f
th
e
s
p
o
k
en
wo
r
d
,
d
is
cr
ete
wav
elet
tr
an
s
f
o
r
m
atio
n
(
DW
T
)
.
T
h
en
th
e
Me
th
o
d
o
lo
g
y
o
f
th
e
wo
r
k
with
its
s
tep
s
,
f
in
ally
,
th
e
o
u
tco
m
es
an
d
th
e
co
n
clu
s
io
n
o
f
th
e
s
p
ea
k
e
r
r
ec
o
g
n
itio
n
u
tili
zin
g
t
h
e
o
n
e
-
s
tate
Hid
d
en
.
2.
M
E
T
H
O
DO
L
O
G
Y
T
h
e
wo
r
k
d
o
n
e
th
r
o
u
g
h
t
h
e
f
o
llo
win
g
s
tep
s
;
i
)
r
ec
o
r
d
i
n
g
Ar
ab
ic
wo
r
d
s
;
ii)
p
r
e
-
p
r
o
ce
s
s
in
g
;
iii
)
f
ea
tu
r
es
ex
tr
ac
tio
n
a
n
d
iv
)
r
ec
o
g
n
i
tio
n
,
with
two
p
h
ases
;
tr
ain
in
g
,
test
i
n
g
,
a
n
d
ex
p
er
im
en
ts
:
2
.
1
.
Da
t
a
s
et
s
Ar
ab
ic
wo
r
d
s
ar
e
r
ec
o
r
d
ed
u
s
in
g
a
m
icr
o
p
h
o
n
e,
with
p
er
s
o
n
s
liv
e
ar
o
u
n
d
u
s
,
an
d
f
r
o
m
lear
n
in
g
p
r
o
g
r
a
m
f
o
r
Ar
ab
ic
lan
g
u
ag
e,
all
th
at
h
av
e
b
ee
n
d
o
n
e
with
r
ea
l
en
v
ir
o
n
m
e
n
ts
,
n
o
t
in
esp
ec
ial
en
v
ir
o
n
m
en
ts
lik
e
in
[
6
]
,
th
en
to
th
e
co
m
p
u
ter
th
r
o
u
g
h
th
e
au
d
i
o
p
o
r
t,
t
h
at
is
ac
co
m
p
lis
h
ed
with
8
0
0
0
Hz
as
s
am
p
lin
g
f
r
eq
u
e
n
cy
an
d
1
6
b
it
r
eso
lu
tio
n
f
o
r
clea
r
r
ec
o
r
d
i
n
g
an
d
s
in
g
le
ch
an
n
el.
T
h
e
r
ec
o
d
in
g
p
r
o
ce
s
s
r
ev
ea
led
th
at
u
s
in
g
th
e
m
icr
o
p
h
o
n
e
r
esu
lts
in
g
o
o
d
q
u
ality
o
u
tp
u
t
s
ig
n
al
s
.
Ho
wev
er
,
it
m
ig
h
t
b
e
a
d
if
f
ic
u
lt
p
r
o
ce
s
s
,
d
u
e
to
th
e
n
o
is
e
ef
f
ec
t a
s
well
as u
n
s
tab
le
d
is
tan
ce
b
etwe
en
th
e
s
p
ea
k
er
s
an
d
th
e
m
icr
o
p
h
o
n
e
.
2
.
2
.
P
re
-
pro
ce
s
s
ing
Af
ter
co
n
v
er
tin
g
th
e
au
d
io
s
i
g
n
al
to
d
ig
itized
f
o
r
m
,
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
s
tag
e
s
tar
ts
,
th
e
o
r
ig
in
al
s
ig
n
al
co
n
s
is
ts
o
f
two
p
ar
ts
,
i.e
.
,
in
f
o
r
m
atio
n
p
ar
t
an
d
s
ilen
t
p
ar
t
with
8
0
0
0
d
o
u
b
le
s
am
p
les.
I
t
s
h
o
u
ld
b
e
m
en
tio
n
ed
t
h
at
s
ilen
t
p
ar
t
m
u
s
t
b
e
r
em
o
v
ed
th
at
g
iv
e
a
s
ig
n
al
ab
o
u
t
4
0
0
0
d
o
a
b
l
e,
s
u
ch
as
in
[
1
8
]
.
No
r
m
aliza
tio
n
p
a
r
ts
n
ec
ess
ar
y
f
o
r
m
ak
in
g
th
e
s
ig
n
al
s
m
o
o
th
er
f
o
r
n
ex
t
o
p
e
r
atio
n
s
.
Pre
-
em
p
h
asis
p
ar
t
am
en
d
s
th
e
lo
s
s
o
f
h
ig
h
er
f
r
eq
u
e
n
cies
th
at
h
av
e
b
ee
n
lo
s
t
th
r
o
u
g
h
th
e
p
r
o
p
ag
atio
n
a
n
d
r
ad
iatio
n
f
o
r
m
v
o
ice
s
o
u
r
ce
to
th
e
m
icr
o
p
h
o
n
e
,
im
p
r
o
v
in
g
ef
f
icien
cy
f
o
r
th
e
n
ex
t
s
tag
es,
t
h
e
f
r
am
i
n
g
a
n
d
win
d
o
w
in
g
a
r
e
ac
co
m
p
lis
h
ed
.
As
th
e
h
u
m
an
s
p
ee
c
h
s
ig
n
al
is
v
ar
y
in
g
s
lo
wly
in
tim
e,
it
is
n
o
r
m
ally
d
iv
id
ed
in
t
o
f
r
am
es,
wh
ich
ar
e
o
v
er
la
p
p
in
g
with
ea
ch
o
th
er
.
W
h
ile
win
d
o
win
g
p
r
o
ce
s
s
in
clu
d
es
d
iv
id
in
g
th
e
f
r
am
es
with
a
win
d
o
w,
s
u
ch
a
Ham
m
in
g
,
s
u
ch
p
r
o
ce
s
s
d
ec
r
ea
s
e
s
th
e
ef
f
ec
ts
o
f
d
is
co
n
tin
u
ity
t
h
at
is
p
r
o
d
u
ce
d
b
y
f
r
am
i
n
g
p
r
o
ce
s
s
.
Fin
ally
r
esizin
g
o
f
th
e
s
p
o
k
en
f
ile
b
y
u
s
in
g
d
is
c
r
ete
wav
elet
tr
an
s
f
o
r
m
(
DW
T
)
,
h
en
ce
,
with
1
s
t
lev
el
o
f
DW
T
we
g
et
ab
o
u
t
1
0
0
0
d
o
u
b
le
s
am
p
les,
wh
ile
with
m
o
r
e
lev
els,
th
e
d
ata
will lo
s
e
th
e
m
ea
n
p
a
r
t o
f
it.
2
.
3
.
F
e
a
t
ures e
x
t
ra
c
t
io
n
I
n
th
e
wh
o
le
wo
r
k
,
f
ea
t
u
r
e
ex
tr
ac
tio
n
an
d
r
ec
o
g
n
itio
n
wer
e
im
p
lem
en
ted
in
MA
T
L
AB
2
0
1
7
b
s
o
f
twar
e.
E
ac
h
s
p
ee
ch
s
ig
n
al
co
r
r
esp
o
n
d
in
g
to
an
y
wo
r
d
is
p
u
t
in
a
s
p
ec
if
ic
f
ile.
Ma
n
y
s
p
ee
ch
f
ea
tu
r
es
h
a
v
e
b
ee
n
s
tu
d
ie
d
,
f
o
r
co
n
s
id
er
i
n
g
th
e
s
p
o
k
en
A
r
ab
ic
wo
r
d
s
as
a
u
d
io
s
ig
n
al,
an
d
f
r
o
m
th
at
a
n
a
u
d
io
f
ea
t
u
r
es
ca
n
b
e
ex
tr
ac
ted
an
d
b
r
o
ad
l
y
class
if
ied
b
ased
o
n
th
eir
s
em
an
tic
in
ter
p
r
etatio
n
as
p
er
ce
p
t
u
al
an
d
p
h
y
s
ical
f
ea
t
u
r
es.
Mo
r
eo
v
er
,
s
tatis
tical
f
ea
tu
r
es
in
clu
d
in
g
,
m
ea
n
v
alu
e,
r
o
o
t
m
ea
n
s
q
u
ar
e
(
R
MS
)
,
s
tan
d
ar
d
d
ev
iatio
n
,
m
ed
ian
v
alu
e,
c
o
v
ar
ian
ce
,
v
ar
ian
ce
v
a
lu
e,
m
a
x
im
u
m
v
alu
e
an
d
m
i
n
im
u
m
v
alu
e.
I
n
th
is
wo
r
k
,
th
e
s
tatis
t
ical
f
ea
tu
r
es
(
m
ea
n
v
alu
e
an
d
c
o
v
ar
ian
ce
)
ar
e
th
e
d
ep
en
d
ed
f
ea
tu
r
es
b
ec
au
s
e
th
e
s
tatis
tical
f
ea
tu
r
es
r
e
p
r
esen
t
th
e
c
o
r
e
o
f
th
e
s
ig
n
al
an
d
r
ed
u
ce
t
h
e
r
eq
u
ir
ed
s
ize
an
d
th
e
p
r
o
ce
s
s
in
g
ti
m
e.
2
.
4
.
H
idd
en
m
a
r
k
o
v
mo
del
HM
M
[
1
9
]
is
a
s
to
ch
asti
c
s
y
s
tem
u
s
ed
to
f
o
r
esee
a
f
u
tu
r
e
o
cc
asio
n
s
d
ep
en
d
e
n
t
o
n
a
p
r
e
v
io
u
s
d
ata.
T
h
e
s
y
s
tem
in
clu
d
es
an
ass
o
r
tm
en
t
o
f
s
tates,
wh
er
e
ju
s
t
th
e
y
ield
s
o
f
th
e
s
tates
ca
n
b
e
v
iewe
d
an
d
all
th
e
ch
an
g
es a
m
o
n
g
th
e
s
tates a
r
e
u
n
k
n
o
wn
.
HM
M
ca
n
b
e
g
r
o
u
p
e
d
in
to
two
class
es a
s
in
d
icate
d
b
y
th
e
k
n
o
wled
g
e
o
f
th
e
y
ield
s
:
d
is
cr
ete
HM
M
a
n
d
co
n
tin
u
es
HM
M
[
2
0
]
,
Fig
u
r
e
1
s
h
o
ws
th
e
s
tate
d
iag
r
am
3
-
s
tate
lef
t
-
to
-
r
ig
h
t
HM
M
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
23
,
No
.
2
,
Au
g
u
s
t 2
0
2
1
:
1
2
1
2
-
1
2
1
8
1214
Fig
u
r
e
1
.
State
d
iag
r
am
o
f
3
-
s
t
ate
lef
t
-
to
-
r
ig
h
t H
MM
Dis
cr
ete
HM
M,
th
is
ty
p
e
m
an
ag
es
d
is
c
r
ete
co
d
es
th
at
ar
e
tr
an
s
m
itted
f
r
o
m
th
e
s
tates
an
d
th
e
s
y
s
tem
is
d
em
o
n
s
tr
ated
b
y
th
e
th
r
ee
b
o
u
n
d
ar
ies
(
π,
A,
B
)
.
C
o
n
tin
u
o
u
s
HM
M,
T
h
e
ex
p
r
e
s
s
io
n
"c
o
n
tin
u
o
u
s
"
in
d
icate
s
th
e
id
ea
o
f
th
e
y
ield
d
en
s
ities
o
f
th
e
m
ask
ed
s
tates.
L
ik
e
a
Gau
s
s
ian
ca
p
ac
ity
,
th
e
s
e
y
ield
s
tr
ac
k
th
e
p
r
o
b
a
b
ilit
y
d
en
s
ity
f
u
n
ctio
n
(
PDF),
wh
er
e
it
i
s
a
s
y
m
m
etr
ic
b
en
d
f
r
am
in
g
a
f
o
r
m
r
esem
b
l
es
a
ch
im
e.
PD
F
o
f
th
e
p
er
ce
p
tio
n
v
ec
to
r
O
is
d
eter
m
in
ed
b
y
th
e
ac
c
o
m
p
an
y
in
g
co
n
d
itio
n
[
2
0
]
:
(
)
=
∑
√
2
2
e
xp
[
(
−
)
2
2
]
=
1
(
1
)
W
h
er
e
;
wn
,
σ
n
an
d
µn
ar
e
,
in
d
iv
id
u
ally
,
th
e
weig
h
t,
s
tan
d
ar
d
d
ev
iatio
n
an
d
m
ea
n
o
f
th
e
n
th
Gau
s
s
ian
b
len
d
.
I
t
is
im
p
o
r
tan
t
th
at
th
e
co
v
ar
ian
ce
(
∑)
o
f
a
v
ec
to
r
is
eq
u
iv
alen
t
to
th
e
s
q
u
ar
e
o
f
th
e
s
tan
d
ar
d
d
ev
iatio
n
an
d
th
u
s
,
th
e
co
n
tin
u
o
u
s
HM
M
is
r
ep
r
esen
ted
as i
n
th
e
ass
o
ciate
d
tu
p
le:
=
(
,
,
,
Σ
)
(
2
)
T
h
e
f
o
llo
win
g
p
o
i
n
ts
g
iv
e
an
o
v
er
v
iew
o
f
its
co
n
s
tr
u
ctio
n
: s
y
m
b
o
ls
N:
States
n
u
m
b
er
in
ea
ch
s
y
s
t
em
.
M:
C
o
d
e
n
u
m
b
e
r
in
th
e
y
ield
s
.
π:
T
h
e
f
u
n
d
am
en
tal
s
tate
p
r
o
b
ab
ilit
y
p
ar
am
eter
o
f
s
ize
N
×
1
.
A:
T
h
e
ch
an
g
e
p
r
o
b
ab
ilit
y
f
r
a
m
ewo
r
k
o
f
s
ize
N
×
N.
B
: T
h
e
r
elea
s
e
p
r
o
b
ab
ilit
y
f
r
a
m
ewo
r
k
o
f
s
ize
N
×
M.
T
h
e
co
n
tr
ast b
etwe
en
th
e
c
o
n
ti
n
u
o
u
s
an
d
d
is
cr
ete
HM
Ms,
co
n
ce
r
n
in
g
th
e
HM
M
b
o
u
n
d
ar
ie
s
,
is
in
th
e
d
is
ch
ar
g
e
b
o
u
n
d
ar
y
,
wh
er
e
in
co
n
tin
u
o
u
s
HM
M;
it
is
in
d
icate
d
b
y
th
e
c
o
v
ar
ian
ce
an
d
m
ea
n
r
ath
er
th
a
n
d
is
cr
ete
co
d
es.
2
.
5
.
Rec
o
g
nitio
n
R
ec
o
g
n
itio
n
h
as two
p
a
r
ts
:
tr
ain
in
g
an
d
test
in
g
,
2
.
5
.
1
.
T
ra
ini
ng
Fo
r
ev
er
y
s
p
o
k
en
wo
r
d
,
an
ar
r
ay
is
cr
ea
ted
b
y
lin
k
in
g
all
th
e
s
eq
u
en
ce
s
g
o
t
f
r
o
m
th
e
tr
a
in
in
g
wo
r
d
as
clar
if
ied
ea
r
lier
.
W
h
en
t
h
e
ar
r
ay
is
f
r
am
e
d
,
it
is
d
eliv
e
r
e
d
to
th
e
HM
M
f
o
r
t
r
ain
in
g
.
H
MM
u
tili
ze
d
in
th
e
p
r
o
p
o
s
ed
wo
r
k
is
a
u
n
iq
u
e
s
y
s
tem
th
at
co
n
tain
s
j
u
s
t
o
n
e
s
tate
with
co
n
tin
u
o
u
s
y
ield
d
en
s
ities
.
Neith
er
s
tar
tin
g
v
ec
to
r
π
n
o
r
tr
an
s
f
o
r
m
atio
n
m
atr
ix
A,
o
cc
u
r
s
in
o
n
e
-
s
tate
s
y
s
tem
an
d
,
f
o
r
t
h
is
s
itu
atio
n
,
th
ey
ar
e
eq
u
i
v
alen
t
t
o
o
n
e.
I
n
th
is
m
an
n
er
,
th
e
s
y
s
tem
is
co
m
m
o
n
ly
f
o
u
n
d
ed
o
n
th
e
µ
an
d
∑ o
f
th
e
p
e
r
ce
p
tio
n
v
ec
to
r
s
,
as sh
o
wn
in
F
ig
u
r
e
2
.
T
h
e
B
au
m
-
W
elch
ca
lcu
latio
n
[
1
0
]
with
a
o
n
e
iter
at
io
n
is
u
tili
ze
d
to
tr
ain
th
e
s
y
s
t
em
o
f
ev
er
y
wo
r
d
.
J
u
s
t
o
n
e
Gau
s
s
ian
m
ix
tu
r
e
is
u
tili
ze
d
an
d
th
e
PDFs
ar
e
d
eter
m
in
ed
as
in
(
1
)
,
wh
er
e
(
)
=
[
1
,
2
,
3
,
.
.
.
,
]
.
Fig
u
r
e
2
s
h
o
ws th
e
s
tate
ch
ar
t o
f
th
e
C
OSM
.
Fig
u
r
e
2
.
State
d
iag
r
am
o
f
th
e
C
OSM
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
A
r
a
b
ic
s
p
ea
ke
r
r
ec
o
g
n
itio
n
u
s
in
g
HMM
(
Ja
b
b
a
r
S
.
Hu
s
s
ein
)
1215
2
.
5
.
2
.
T
esting
All
s
p
o
k
en
wo
r
d
s
th
at
ar
e
n
o
t
u
tili
ze
d
in
tr
ain
in
g
o
f
t
h
e
HM
M
tr
ac
k
th
e
co
r
r
esp
o
n
d
in
g
ab
o
v
e
p
o
in
ts
f
o
r
test
in
g
,
wh
e
r
e
ea
ch
w
o
r
d
i
s
in
d
ep
en
d
e
n
tly
tr
ea
ted
.
T
h
e
µ
an
d
∑
o
f
th
e
p
e
r
ce
p
tio
n
v
ec
to
r
s
ar
e
d
eter
m
in
e
d
an
d
th
e
Viter
b
i
ca
lcu
latio
n
[
1
0
]
is
u
tili
ze
d
to
f
in
d
th
eir
p
r
o
b
a
b
ilit
ies
b
y
all
PDFs
th
at
ar
e
g
o
tten
in
th
e
tr
ai
n
in
g
p
r
o
ce
d
u
r
e.
Su
b
s
eq
u
en
tly
,
th
e
in
d
ex
o
f
th
e
m
o
s
t
ex
tr
e
m
e
p
r
o
b
a
b
ilit
y
m
ig
h
t
b
e
u
tili
ze
d
to
d
is
tin
g
u
is
h
th
e
u
n
k
n
o
wn
wo
r
d
.
2
.
5
.
3
.
Ca
s
e
s
t
ud
y
T
h
e
ex
p
er
im
en
ts
ar
e
p
e
r
f
o
r
m
e
d
o
n
th
e
wo
r
k
d
ata
b
ases
:
wh
er
e
f
o
r
f
iv
e
p
er
s
o
n
s
,
o
n
e
h
u
n
d
r
e
d
p
atter
n
s
f
o
r
ea
ch
o
n
e,
7
0
wo
r
d
s
f
o
r
tr
ai
n
in
g
,
a
n
d
3
0
wo
r
d
s
f
o
r
test
in
g
.
W
ith
HM
M,
th
e
ex
p
er
im
e
n
ts
s
h
o
w
th
at
th
e
tech
n
iq
u
e
o
f
u
s
in
g
th
e
m
ea
n
v
alu
e
an
d
co
n
f
e
r
en
ce
ar
e
th
e
f
in
e
o
n
e.
So
,
th
is
m
eth
o
d
is
ex
am
in
ed
u
s
in
g
c
o
n
tin
u
es
HM
M,
an
d
th
e
f
o
llo
win
g
s
p
ec
if
icatio
n
s
ar
e
em
p
lo
y
ed
:
1.
Pre
p
r
o
ce
s
s
in
g
: Fo
r
th
e
wo
r
d
s
d
atab
ase: 1
s
t stag
e
o
f
DW
T
p
r
o
d
u
ce
s
v
ec
to
r
o
f
s
ize
(
2
0
0
2
×
1
)
2.
7
5
%
o
v
er
la
p
p
ed
Ham
m
in
g
wi
n
d
o
w
o
f
len
g
th
n
=1
0
0
3.
Featu
r
e
ex
tr
ac
tio
n
= [
]
4.
T
r
ain
in
g
Af
ter
th
e
in
f
o
r
m
atio
n
g
ath
er
in
g
,
we
tr
ied
o
u
r
lear
n
in
g
ca
lcu
l
atio
n
as tak
es a
f
ter
:
•
R
an
d
o
m
ly
p
ick
7
0
•
T
est o
n
th
e
r
est o
f
t
h
e
3
0
•
R
ep
ea
t stag
es 1
an
d
2
o
r
d
in
ar
i
ly
W
h
er
e
s
tag
e
(
c)
is
ad
d
ed
t
o
d
i
m
in
is
h
th
e
v
ar
iety
f
r
o
m
th
e
d
e
cisi
o
n
o
f
th
e
p
r
ep
ar
atio
n
s
et.
3.
RE
SU
L
T
S
T
h
e
r
esu
lts
s
h
o
wn
in
T
ab
le
1
,
is
f
o
u
n
d
ed
f
o
r
f
iv
e
p
e
r
s
o
n
s
ea
ch
o
n
e
h
as
1
0
0
p
atter
s
(
wo
r
d
s
)
,
7
0
o
n
e
f
o
r
tr
ain
i
n
g
a
n
d
3
0
p
atter
n
s
f
o
r
test
.
W
ith
Fig
u
r
e
3
,
we
tak
e
th
e
p
atter
n
s
f
o
r
o
n
e
p
e
r
s
o
n
,
a
n
d
also
b
e
g
an
wit
h
7
0
o
n
e
f
o
r
tr
ai
n
in
g
a
n
d
3
0
p
a
tter
n
s
f
o
r
test
,
th
e
n
in
s
tep
o
f
f
iv
e
p
atter
s
we
r
ed
u
ce
d
th
e
t
r
ain
in
g
p
atter
s
an
d
in
cr
ea
s
ed
th
e
test
o
n
e,
o
u
r
g
o
al
to
s
ee
th
e
ef
f
ec
t
o
f
th
e
n
u
m
b
er
o
f
p
atter
s
o
n
th
e
r
ec
o
g
n
i
tio
n
r
ate
an
d
HM
M
alg
o
r
ith
m
,
as sh
o
wn
in
Fig
u
r
e
4
,
th
e
r
ec
o
g
n
itio
n
r
ate
d
ec
r
ea
s
ed
with
d
ec
r
ea
s
in
g
th
e
tr
ain
in
g
p
atter
s
an
d
th
at
is
a
n
atu
r
al
r
esu
lt with
s
u
ch
al
g
o
r
ith
m
.
T
ab
le1
.
HM
M
r
ec
o
g
n
itio
n
r
ate
R
e
c
o
g
n
i
t
i
o
n
r
a
t
e
%
Te
st
w
o
r
d
s
Tr
a
i
n
i
n
g
w
o
r
d
s
S
p
e
a
k
e
r
s
1
0
0
30
70
1
1
0
0
30
70
2
1
0
0
30
70
3
1
0
0
30
70
4
1
0
0
30
70
5
T
o
s
im
u
late
th
e
ef
f
ec
ts
o
f
e
r
r
o
r
o
r
n
o
is
e
o
n
t
h
e
p
er
f
o
r
m
a
n
c
e
o
f
th
e
r
ec
o
g
n
itio
n
s
y
s
tem
,
a
n
ad
d
itiv
e
wh
ite
g
au
s
s
ian
n
o
is
e
(
AW
GN
)
was
ad
d
e
d
to
th
e
wo
r
d
s
p
att
er
n
s
,
tr
ain
in
g
an
d
test
o
n
es,
b
ec
au
s
e
s
u
ch
a
n
o
is
e
ca
v
er
all
th
e
s
p
ec
tr
u
m
,
t
h
e
r
esu
lts
s
h
o
w
g
o
o
d
o
u
tc
om
es,
as sh
o
wn
in
T
a
b
le
2
.
W
h
i
l
e
F
i
g
u
r
e
2
s
h
o
w
t
h
e
e
f
f
e
c
t
o
f
a
d
d
i
t
i
v
e
n
o
i
s
e
f
o
r
o
n
e
p
e
r
s
o
n
r
e
c
o
g
n
i
t
i
o
n
.
W
i
t
h
l
e
s
s
n
o
i
s
e
l
e
v
e
l
s
,
o
n
e
c
a
n
g
e
t
b
e
t
t
e
r
r
e
s
u
l
t
s
,
s
u
c
h
a
s
i
n
[
2
1
]
-
[
2
3
]
.
T
ab
le
2
.
HM
M
r
ec
o
g
n
itio
n
r
at
e
with
ad
d
itiv
e
n
o
is
e
R
e
c
o
g
n
i
t
i
o
n
r
a
t
e
%
Te
st
w
o
r
d
s
Tr
a
i
n
i
n
g
w
o
r
d
s
S
p
e
a
k
e
r
s
9
1
.
6
30
70
1
8
3
.
3
30
70
2
9
1
.
6
30
70
3
8
3
.
3
30
70
4
8
3
.
3
30
70
5
Fo
r
co
m
p
r
is
in
g
with
o
th
e
r
tech
n
o
lo
g
ies,
lik
e
n
e
u
r
al
n
etwo
r
k
an
d
o
r
d
i
n
ar
y
HM
M
with
m
o
r
e
th
an
o
n
e
s
tate)
,
an
d
as
s
h
o
wn
in
T
ab
le
3
,
HM
M
with
o
n
e,
two
an
d
th
r
ee
s
tate
ar
e
s
h
o
w
n
,
o
n
e
ca
n
n
o
te
f
r
o
m
th
e
r
esu
lts
,
th
at
th
e
o
n
e
s
tate
HM
M
h
as
b
etter
o
u
tp
u
t
th
an
t
h
e
o
th
e
r
s
,
an
d
th
at
also
h
a
v
e
b
ee
n
p
u
b
lis
h
ed
as
in
[
2
4
]
.
W
h
ile
th
e
co
m
p
ar
is
o
n
with
NN,
lik
e
m
u
lti
-
lay
er
f
ee
d
f
o
r
war
d
n
eu
r
al
n
etwo
r
k
(
ML
FF
NN)
,
an
d
a
s
s
h
o
wn
in
T
ab
le
4
,
s
till
th
e
HM
M
h
as b
etter
r
esu
lts
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
23
,
No
.
2
,
Au
g
u
s
t 2
0
2
1
:
1
2
1
2
-
1
2
1
8
1216
Fig
u
r
e
3
.
R
ec
o
g
n
itio
n
r
ate
f
o
r
o
n
e
p
er
s
o
n
with
v
ar
ia
b
les tr
ain
in
g
an
d
test
wo
r
d
s
Fig
u
r
e
4
.
R
ec
o
g
n
itio
n
r
ate
f
o
r
o
n
e
p
er
s
o
n
with
v
ar
ia
b
les tr
ain
in
g
an
d
test
wo
r
d
s
an
d
ad
d
iti
v
e
n
o
is
e
T
ab
le
3
.
C
o
m
p
a
r
is
o
n
o
n
e
s
tate
HM
M
with
o
u
t state
with
,
two
an
d
th
r
ee
s
tate
HM
M
R
e
c
o
g
n
i
t
i
o
n
r
a
t
e
%
t
h
r
e
e
s
t
a
t
e
s
R
e
c
o
g
n
i
t
i
o
n
r
a
t
e
%
t
w
o
s
t
a
t
e
s
R
e
c
o
g
n
i
t
i
o
n
r
a
t
e
%
o
n
e
s
t
a
t
e
Te
st
w
o
r
d
s
Tr
a
i
n
i
n
g
w
o
r
d
s
S
p
e
a
k
e
r
s
7
6
.
6
6
8
6
.
6
6
1
0
0
30
70
1
7
6
.
6
6
8
6
.
6
6
1
0
0
30
70
2
7
6
.
6
6
8
6
.
6
6
1
0
0
30
70
3
7
6
.
6
6
8
6
.
6
6
1
0
0
30
70
4
7
6
.
6
6
8
6
.
6
6
1
0
0
30
70
5
T
ab
le
4
.
C
o
m
p
a
r
is
o
n
HM
M
with
o
u
t state
with
ML
FF
NN
R
e
c
o
g
n
i
t
i
o
n
r
a
t
e
%
M
LFF
N
N
[
2
5
]
R
e
c
o
g
n
i
t
i
o
n
r
a
t
e
%
o
n
e
s
t
a
t
e
H
M
M
Te
st
w
o
r
d
s
Tr
a
i
n
i
n
g
w
o
r
d
s
S
p
e
a
k
e
r
s
90
1
0
0
30
70
1
90
1
0
0
30
70
2
90
1
0
0
30
70
3
90
1
0
0
30
70
4
90
1
0
0
30
70
5
4.
CO
NCLU
SI
O
NS
Sp
ea
k
er
r
ec
o
g
n
itio
n
is
th
e
u
s
e
o
f
a
m
ac
h
in
e
t
o
r
ec
o
g
n
ize
a
p
er
s
o
n
f
r
o
m
a
s
p
o
k
en
p
h
r
ase
.
Sp
ea
k
er
-
r
ec
o
g
n
itio
n
s
y
s
tem
s
ca
n
b
e
u
s
ed
to
id
en
tify
a
p
a
r
ticu
lar
p
er
s
o
n
o
r
to
v
er
if
y
a
p
er
s
o
n
’
s
claim
ed
id
en
tity
.
Sp
ee
ch
p
r
o
ce
s
s
in
g
,
s
p
ee
c
h
p
r
o
d
u
ctio
n
,
a
n
d
f
ea
tu
r
es
an
d
p
atter
n
m
atc
h
in
g
f
o
r
s
p
ea
k
e
r
r
ec
o
g
n
itio
n
wer
e
in
tr
o
d
u
ce
d
.
A
u
n
iq
u
e
tech
n
iq
u
e
is
estab
lis
h
ed
f
o
r
r
ec
o
g
n
izin
g
s
p
o
k
e
n
wo
r
d
b
y
m
ea
n
s
o
f
c
o
n
tin
u
o
u
s
o
n
e
-
s
tate
s
y
s
tem
in
co
m
b
in
atio
n
with
a
DW
T
.
Dis
s
im
ilar
to
o
th
er
m
et
h
o
d
s
th
at
d
ep
e
n
d
o
n
all
p
ar
a
m
eter
s
o
f
HM
M,
th
e
s
u
g
g
e
s
ted
wo
r
k
r
em
o
v
es
two
o
f
th
ese
co
m
p
o
n
e
n
ts
,
i.e
.
π
a
n
d
A,
an
d
th
e
r
ec
o
g
n
itio
n
b
e
d
e
ter
m
in
ed
b
y
s
im
p
ly
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
A
r
a
b
ic
s
p
ea
ke
r
r
ec
o
g
n
itio
n
u
s
in
g
HMM
(
Ja
b
b
a
r
S
.
Hu
s
s
ein
)
1217
th
e
PDF o
f
o
n
e
Gau
s
s
ian
m
ix
tu
r
e
elem
en
t.
E
n
v
ir
o
n
m
en
t n
o
is
e
is
ac
co
r
d
in
g
ly
d
etac
h
e
d
f
r
o
m
th
e
wo
r
d
s
v
ia
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
an
d
th
e
DW
T
.
T
h
e
1
s
t
lev
el
o
f
DW
T
was
ap
p
lied
to
th
e
s
p
o
k
e
n
wo
r
d
o
f
th
e
wo
r
d
s
d
atab
ases
,
wh
ich
in
tu
r
n
d
ec
r
ea
s
e
th
e
s
p
o
k
en
wo
r
d
s
ize,
wh
er
e
c
o
n
s
tr
u
ctin
g
f
ea
tu
r
e
v
ec
to
r
s
b
y
DW
T
is
a
v
er
y
h
o
p
ef
u
l
m
eth
o
d
f
o
r
th
e
task
o
f
s
p
o
k
e
n
wo
r
d
.
Als
o
,
th
e
u
tili
zin
g
o
f
o
n
e
s
tate
ex
tr
em
ely
ac
ce
ler
at
es
th
e
t
r
ain
in
g
a
n
d
test
in
g
s
tag
es
o
f
r
ec
o
g
n
itio
n
s
p
ee
d
s
with
lo
west
m
em
o
r
y
u
s
ag
e.
T
h
e
ex
p
e
r
im
en
tal
o
u
tco
m
es
d
is
p
lay
th
at
th
e
p
r
ec
is
io
n
o
f
th
e
s
u
g
g
ested
w
o
r
k
is
ar
o
u
n
d
1
0
0
%
in
s
p
ite
o
f
a
d
d
itiv
e
n
o
is
e,
th
at
af
f
ec
t
th
e
s
p
o
k
e
n
wo
r
d
r
ec
o
g
n
itio
n
b
u
t
with
m
in
im
u
m
ef
f
ec
t
.
T
h
e
in
f
lu
en
ce
o
f
th
e
q
u
an
tity
o
f
s
tate
d
ep
en
d
s
o
n
th
e
d
ata
s
co
p
e.
Fo
r
m
in
o
r
d
ata,
s
m
all
q
u
a
n
tity
o
f
s
tates
h
as
g
r
ea
ter
r
ec
o
g
n
itio
n
p
er
ce
n
tag
e.
Fo
r
b
ig
g
er
d
ata,
th
e
q
u
an
tity
o
f
s
tates
h
as v
er
y
m
i
n
o
r
r
esu
lt o
n
r
ec
o
g
n
itio
n
r
ate
RE
F
E
R
E
NC
E
S
[1
]
E.
Ab
b
a
s
a
n
d
H.
F
a
r
h
a
n
,
“
F
a
c
e
re
c
o
g
n
it
i
o
n
u
si
n
g
DWT
wit
h
HM
M
,
”
E
n
g
i
n
e
e
rin
g
&
T
e
c
h
n
o
l
o
g
y
J
o
u
rn
a
l
,
v
o
l.
3
0
,
n
o
.
1
,
p
p
.
1
4
2
-
1
5
4
,
2
0
1
2
.
[2
]
Z.
Li
a
n
d
X.
Tan
g
,
“
Us
in
g
s
u
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
s
to
e
n
h
a
n
c
e
t
h
e
p
e
rfo
rm
a
n
c
e
o
f
Ba
y
e
sia
n
fa
c
e
re
c
o
g
n
it
i
o
n
,
”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
I
n
f
o
rm
a
ti
o
n
F
o
re
n
sic
s
a
n
d
S
e
c
u
r
it
y
,
v
o
l.
2
,
n
o
.
2
,
p
p
.
1
7
4
-
1
8
0
,
2
0
0
7
,
d
o
i:
1
0
.
1
1
0
9
/T
IF
S
.
2
0
0
7
.
8
9
7
2
4
7
.
[3
]
J.
Lu
,
K.
N.
P
lata
n
i
o
ti
s
,
a
n
d
A
.
N.
Ve
n
e
tsa
n
o
p
o
u
lo
s
,
“
F
a
c
e
re
c
o
g
n
i
ti
o
n
u
si
n
g
LDA
-
b
a
se
d
a
lg
o
rit
h
m
s
,
”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Ne
u
ra
l
Ne
two
rk
s
,
v
o
l.
1
4
,
n
o
.
1
,
p
p
.
1
9
5
-
2
0
0
,
2
0
0
3
,
d
o
i:
1
0
.
1
1
0
9
/
TNN.2
0
0
2
.
8
0
6
6
4
7
.
[4
]
M
.
S
.
Ba
rt
lett,
J
.
R.
M
o
v
e
l
lan
,
a
n
d
T
.
J.
S
e
j
n
o
ws
k
i,
“
F
a
c
e
re
c
o
g
n
i
t
io
n
b
y
in
d
e
p
e
n
d
e
n
t
c
o
m
p
o
n
e
n
t
a
n
a
ly
sis
,
”
I
EE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Ne
u
ra
l
Ne
two
rk
s
,
v
o
l.
13
,
n
o
.
6
,
p
p
.
1
4
5
0
-
1
4
6
4
,
2
0
0
2
,
d
o
i:
1
0
.
1
1
0
9
/
TNN.2
0
0
2
.
8
0
4
2
8
7
.
[5
]
J.
Ya
n
g
,
D.
Zh
a
n
g
,
A.
F
.
F
ra
n
g
i
,
a
n
d
J
.
Ya
n
g
,
“
Two
-
d
ime
n
si
o
n
a
l
P
CA:
a
n
e
w
a
p
p
ro
a
c
h
to
a
p
p
e
a
r
a
n
c
e
b
a
se
d
fa
c
e
re
p
re
se
n
tatio
n
a
n
d
re
c
o
g
n
it
io
n
,
”
IEE
E
T
ra
n
s
a
c
ti
o
n
s
o
n
P
a
tt
e
rn
An
a
lys
is
a
n
d
M
a
c
h
i
n
e
In
telli
g
e
n
c
e
,
v
o
l.
2
6
,
n
o
.
1
,
p
p
.
1
3
1
-
1
3
7
,
2
0
0
4
,
d
o
i:
1
0
.
1
1
0
9
/T
P
AMI.
2
0
0
4
.
1
2
6
1
0
9
7
.
[6
]
V
.
N
.
K
.
R
.
De
v
a
n
a
a
n
d
T
.
Ra
je
sh
,
“
A
Hi
g
h
Bit
-
Ra
te
S
p
e
e
c
h
Re
c
o
g
n
i
ti
o
n
S
y
ste
m
t
h
ro
u
g
h
Q
u
a
n
ti
z
a
ti
o
n
o
f
M
e
l
-
F
re
q
u
e
n
c
y
Ce
p
stra
l
C
o
e
fficie
n
t
s
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
E
lec
trica
l,
El
e
c
tro
n
ics
a
n
d
C
o
m
p
u
ter
S
y
ste
ms
(IJ
EE
CS
)
,
v
ol
.
2
,
n
o
.
8
-
9
,
2
0
1
4
.
[7
]
M
.
Ow
a
y
jan
,
R.
Ac
h
k
a
r
,
a
n
d
M
.
Isk
a
n
d
a
r,
“
F
a
c
e
d
e
tec
ti
o
n
w
it
h
e
x
p
re
ss
io
n
re
c
o
g
n
i
ti
o
n
u
sin
g
a
rt
ifi
c
ial
n
e
u
ra
l
n
e
two
rk
s
,
”
2
0
1
6
3
rd
M
id
d
le
Ea
st
Co
n
fer
e
n
c
e
o
n
Bi
o
me
d
ic
a
l
E
n
g
i
n
e
e
rin
g
(
M
ECB
M
E)
,
Be
iru
t
,
6
-
7
Oc
t.
2
0
1
6
,
p
p
.
1
1
5
-
1
1
9
,
2
0
1
6
,
d
o
i:
1
0
.
1
1
0
9
/
M
ECBM
E.
2
0
1
6
.
7
7
4
5
4
2
1
.
[8
]
Kh
.
M
.
O
.
Na
h
a
r,
Na
h
a
r,
M
.
El
sh
a
fe
i,
W.
G
,
Al
-
Kh
a
ti
b
,
a
n
d
H
.
Al
-
M
u
h
tas
e
b
,
“
S
tat
isti
c
a
l
An
a
l
y
sis
o
f
Ara
b
ic
P
h
o
n
e
m
e
s
fo
r
C
o
n
t
in
u
o
u
s
Ara
b
ic
S
p
e
e
c
h
Re
c
o
g
n
it
io
n
,
”
I
n
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
C
o
mp
u
ter
a
n
d
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
,
v
o
l.
0
1
,
n
o
.
0
2
,
No
v
.
2
0
1
2
.
[9
]
J.
S
.
Hu
ss
e
in
,
A.
H.
Ali
,
a
n
d
Th
.
R.
S
a
e
e
d
,
“
Im
p
ro
v
e
th
e
Re
c
o
g
n
i
ti
o
n
o
f
Ara
b
ic
S
ig
n
Lan
g
u
a
g
e
s
Ba
se
d
o
n
S
tatisti
c
a
l
F
e
a
tu
re
s,”
Ira
q
i
J
o
u
rn
a
l
o
f
Co
m
p
u
ter
s,
Co
mm
u
n
ica
t
io
n
s,
Co
n
tro
l
&
S
y
ste
ms
En
g
in
e
e
rin
g
(
IJ
CCCE)
,
v
o
l.
1
8
,
n
o
.
3
,
p
p
.
2
6
-
32
,
2
0
1
8
.
[1
0
]
Th
.
R
.
S
a
e
e
d
,
J.
S
.
H
u
ss
e
in
,
a
n
d
A.
H.
Ali
,
“
Clas
sifica
ti
o
n
imp
r
o
v
e
m
e
n
t
o
f
s
p
o
k
e
n
a
ra
b
ic
lan
g
u
a
g
e
b
a
se
d
o
n
ra
d
ial
b
a
sis
fu
n
c
ti
o
n
,
”
I
n
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
mp
u
ter
En
g
i
n
e
e
rin
g
(IJ
ECE
)
,
v
o
l.
9
,
n
o
.
1
,
p
p
.
4
0
2
-
4
0
8
,
F
e
b
.
2
0
1
9
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jec
e
.
v
9
i1
.
p
p
4
0
2
-
4
0
8
.
[1
1
]
A.
Hu
ss
e
in
,
S
.
Wata
n
a
b
e
,
a
n
d
A.
Alia,
“
Ara
b
ic
S
p
e
e
c
h
Re
c
o
g
n
i
ti
o
n
b
y
E
n
d
-
to
-
En
d
,
M
o
d
u
lar S
y
ste
m
s
a
n
d
Hu
m
a
n
,
”
J
o
u
rn
a
l
o
f
Co
m
p
u
ter
S
p
e
e
c
h
a
n
d
L
a
n
g
u
a
g
e
,
a
rXi
v
p
re
p
rin
t
a
rXiv
:
2
1
0
1
.
0
8
4
5
4
,
2
0
2
1
.
[1
2
]
M
.
A.
A
h
m
a
d
a
n
d
R.
M
.
El
Aw
a
d
y
,
“
P
h
o
n
e
ti
c
Re
c
o
g
n
it
i
o
n
o
f
A
ra
b
ic
Alp
h
a
b
e
t
letters
u
si
n
g
Ne
u
ra
l
Ne
two
rk
s,”
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
E
lec
tric &
Co
mp
u
ter
S
c
ien
c
e
s
,
v
o
l.
1
1
,
n
o
.
1
,
p
p
.
4
4
-
4
9
,
2
0
1
1
.
[1
3
]
R.
M
.
Ha
n
ifa1
,
Kh
.
Isa
a
n
d
S
h
.
M
o
h
a
m
a
d
,
“
S
p
e
a
k
e
r
e
th
n
ic
i
d
e
n
ti
fica
ti
o
n
f
o
r
c
o
n
t
in
u
o
u
s
sp
e
e
c
h
i
n
M
a
lay
lan
g
u
a
g
e
u
sin
g
p
it
c
h
a
n
d
M
F
CC
,
”
In
d
o
n
e
sia
n
J
o
u
r
n
a
l
o
f
E
lec
trica
l
E
n
g
i
n
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
(IJ
EE
CS
)
,
v
o
l.
1
9
,
n
o
.
1
,
p
p
.
2
0
7
-
2
1
4
,
2
0
2
0
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
1
9
.
i
1
.
p
p
2
0
7
-
2
1
4
.
[1
4
]
M
.
T
.
S
.
A
l
-
K
a
l
t
a
k
c
h
i
,
H
.
A
.
A
.
T
a
h
a
,
M
.
A
.
S
h
e
h
a
b
,
a
n
d
M
.
A.
M
.
A
b
d
u
l
l
a
h
,
“
C
o
m
p
a
r
i
s
o
n
o
f
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
a
n
d
n
o
r
m
a
l
i
z
a
t
i
o
n
m
e
t
h
o
d
s
f
o
r
s
p
e
a
k
e
r
r
e
c
o
g
n
i
t
i
o
n
u
s
i
n
g
g
r
i
d
-
a
u
d
i
o
v
i
s
u
a
l
d
a
t
a
b
a
s
e
,
”
I
n
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
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
I
J
E
E
C
S
)
, v
o
l
.
1
8
,
n
o
.
2
,
p
p
.
7
8
2
-
7
8
9
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s
.
v
1
8
.
i
2
.
p
p
7
8
2
-
789
.
[1
5
]
K.
S
in
g
h
,
M
.
Zav
e
ri,
a
n
d
M
.
R
a
g
h
u
wa
n
s
h
i,
“
Re
c
o
g
n
izin
g
fa
c
e
s
u
n
d
e
r
v
a
ry
i
n
g
p
o
se
s
with
t
h
re
e
sta
tes
h
id
d
e
n
M
a
rk
o
v
m
o
d
e
l
,
”
in
Pro
c
.
o
f
IEE
E
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
Co
mp
u
ter
S
c
ien
c
e
a
n
d
Au
t
o
ma
t
i
o
n
En
g
in
e
e
rin
g
(CS
AE
)
,
Zh
a
n
g
ji
a
ji
e
,
C
h
in
a
,
v
o
l.
2
,
p
p
.
3
5
9
-
3
6
3
,
2
0
1
2
,
d
o
i:
1
0
.
1
1
0
9
/CS
AE.
2
0
1
2
.
6
2
7
2
7
9
2
.
[1
6
]
H.
F
a
rh
a
n
,
M
.
Al
-
M
u
ifraje
,
a
n
d
T.
S
a
e
e
d
,
“
Us
in
g
o
n
l
y
tw
o
sta
tes
o
f
d
isc
re
te
HMM
fo
r
h
i
g
h
-
sp
e
e
d
fa
c
e
re
c
o
g
n
it
i
o
n
,
”
in
2
0
1
6
Al
-
S
a
d
e
q
I
n
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
M
u
l
t
id
isc
ip
li
n
a
ry
in
I
T
a
n
d
Co
mm
u
n
i
c
a
ti
o
n
S
c
ien
c
e
a
n
d
Ap
p
li
c
a
ti
o
n
s (A
IC
-
M
IT
CS
A)
,
Ba
g
h
d
a
d
,
Ira
q
,
2
0
1
6
,
p
p
.
1
-
5
,
d
o
i
:
1
0
.
1
1
0
9
/AIC
-
M
ITCS
A
.
2
0
1
6
.
7
7
5
9
9
3
9
.
[1
7
]
H.
F
a
rh
a
n
,
M
.
Al
-
M
u
ifraje
,
a
n
d
T.
S
a
e
e
d
,
“
A
No
v
e
l
F
a
c
e
Re
c
o
g
n
it
io
n
M
e
th
o
d
b
a
se
d
o
n
On
e
S
tate
o
f
Disc
re
te
Hid
d
e
n
M
a
r
k
o
v
M
o
d
e
l
,
”
2
0
1
7
A
n
n
u
a
l
Co
n
fer
e
n
c
e
o
n
Ne
w
T
re
n
d
s
in
fo
rm
a
ti
o
n
&
Co
mm
u
n
ica
t
io
n
s
T
e
c
h
n
o
lo
g
y
Ap
p
li
c
a
ti
o
n
s (NT
IC
T
)
,
Ba
g
h
d
a
d
,
2
0
1
7
,
p
p
.
2
5
2
-
2
5
7
,
d
o
i:
1
0
.
1
1
0
9
/
NTICT.
2
0
1
7
.
7
9
7
6
1
5
2
.
[1
8
]
R.
M
.
Ha
n
ifa,
e
t
a
l.
,
“
Vo
ice
d
a
n
d
u
n
v
o
ice
d
se
p
a
ra
ti
o
n
in
m
a
lay
sp
e
e
c
h
u
sin
g
z
e
ro
c
ro
ss
in
g
ra
te
a
n
d
e
n
e
rg
y
,
”
In
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
i
n
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
(IJ
EE
CS
)
,
v
o
l.
1
6
,
n
o
.
2
,
p
p
.
7
7
5
-
7
8
0
,
2
0
1
9
,
d
o
i:
1
0
.
1
1
5
9
1
/
ij
e
e
c
s.v
1
6
.
i2
.
p
p
7
7
5
-
7
8
0
.
[1
9
]
L.
R.
Ra
b
in
e
r,
“
A
t
u
to
rial
o
n
h
id
d
e
n
M
a
r
k
o
v
m
o
d
e
ls
a
n
d
se
lec
ted
a
p
p
l
ica
ti
o
n
s
in
sp
e
e
c
h
re
c
o
g
n
it
i
o
n
,
”
in
Pro
c
.
o
f
th
e
IEE
E
,
v
o
l.
7
7
,
n
o
.
2
,
1
9
8
9
,
p
p
.
2
5
7
-
2
8
6
,
d
o
i:
1
0
.
1
1
0
9
/5
.
1
8
6
2
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
23
,
No
.
2
,
Au
g
u
s
t 2
0
2
1
:
1
2
1
2
-
1
2
1
8
1218
[2
0
]
L.
S
h
i,
I.
A
h
m
a
d
,
Y.
He
,
a
n
d
K.
Ch
a
n
g
,
”
Hi
d
d
e
n
M
a
rk
o
v
M
o
d
e
l
b
a
se
d
Dro
n
e
S
o
u
n
d
Re
c
o
g
n
it
i
o
n
u
si
n
g
M
F
CC
Tec
h
n
iq
u
e
in
P
ra
c
ti
c
a
l
No
isy
En
v
ir
o
n
m
e
n
ts
,
”
J
o
u
rn
a
l
o
f
Co
mm
u
n
ica
ti
o
n
s
a
n
d
Ne
two
rk
s
,
v
o
l
.
2
0
,
no
.
5
,
p
p
.
5
0
9
-
5
1
8
,
2
0
1
8
,
d
o
i:
1
0
.
1
1
0
9
/J
CN.2
0
1
8
.
0
0
0
0
7
5
.
[2
1
]
M
u
k
h
e
rjee
a
n
d
A.
S
e
n
g
u
p
ta,
“
Es
ti
m
a
ti
n
g
t
h
e
p
r
o
b
a
b
il
it
y
d
e
n
sity
f
u
n
c
ti
o
n
o
f
a
n
o
n
sta
ti
o
n
a
ry
n
o
n
-
Ga
u
ss
ian
n
o
ise
,
”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
In
d
u
stri
a
l
El
e
c
tro
n
ics
,
v
o
l
.
5
7
,
n
o
.
4
,
p
p
.
1
4
2
9
-
1
4
3
5
,
2
0
1
0
,
d
o
i:
1
0
.
1
1
0
9
/T
IE
.
2
0
0
9
.
2
0
3
9
4
5
1
.
[2
2
]
M
.
T.
S
.
Al
-
Ka
lt
a
k
c
h
i,
R
.
R.
O
.
Al
-
Nim
a
,
M
.
A.
M
.
Ab
d
u
ll
a
h
,
a
n
d
H
.
N.
Ab
d
u
ll
a
h
,
“
Th
o
ro
u
g
h
e
v
a
lu
a
ti
o
n
o
f
TIM
I
T
d
a
tab
a
se
sp
e
a
k
e
r
id
e
n
ti
fica
ti
o
n
p
e
rfo
rm
a
n
c
e
u
n
d
e
r
n
o
ise
wit
h
a
n
d
with
o
u
t
t
h
e
G
.
7
1
2
ty
p
e
h
a
n
d
se
t,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
S
p
e
e
c
h
T
e
c
h
n
o
l
o
g
y
,
v
ol
.
2
2
,
n
o
.
3
,
p
p
.
8
5
1
-
8
6
3
,
2
0
1
9
,
d
o
i:
1
0
.
1
0
0
7
/s1
0
7
7
2
-
0
1
9
-
0
9
6
3
0
-
9
.
[2
3
]
O
.
A
.
N
o
o
r
,
“
R
o
b
u
s
t
s
p
e
a
k
e
r
v
e
r
i
f
i
c
a
t
i
o
n
i
n
b
a
n
d
-
l
o
c
a
l
i
z
e
d
n
o
i
s
e
c
o
n
d
i
t
i
o
n
s
,
”
I
n
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
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
I
J
E
E
C
S
)
,
v
o
l
.
1
3
,
n
o
.
2
,
p
p
.
4
9
9
-
5
0
6
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s
.
v
1
3
.
i
2
.
p
p
4
9
9
-
506
.
[2
4
]
M
.
Alsu
laim
a
n
,
Y
.
Al
o
taib
i
,
M
.
G
h
u
lam
,
M
.
A.
Be
n
c
h
e
rif
,
a
n
d
A
.
M
a
h
m
o
u
d
,
“
Ara
b
ic
S
p
e
a
k
e
r
Re
c
o
g
n
it
i
o
n
:
Ba
b
y
lo
n
Lev
a
n
ti
n
e
S
u
b
se
t
Ca
se
S
tu
d
y
,
”
J
o
u
r
n
a
l
o
f
C
o
mp
u
ter
S
c
ien
c
e
,
v
o
l
.
6
,
n
o
.
4
,
p
p
.
381
-
3
8
5
,
2
0
1
0
,
d
o
i:
1
0
.
3
8
4
4
/j
c
ss
p
.
2
0
1
0
.
3
8
1
.
3
8
5
.
[2
5
]
D
.
M
e
n
g
istu
a
n
d
D.
M
.
Ale
m
a
y
e
h
u
,
“
S
p
e
e
c
h
P
ro
c
e
ss
in
g
fo
r
Tex
t
In
d
e
p
e
n
d
e
n
t
Am
h
a
ric
Lan
g
u
a
g
e
Dia
lec
t
Re
c
o
g
n
it
i
o
n
,
”
In
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
i
n
e
e
rin
g
a
n
d
Co
mp
u
ter
S
c
ien
c
e
(IJ
EE
CS
)
,
v
o
l.
5
,
n
o
.
1
,
p
p
.
1
1
5
-
1
2
2
,
2
0
1
7
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
5
.
i
1
.
p
p
1
1
5
-
1
2
2
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
J
a
b
b
a
r
S
a
l
m
a
n
H
u
ss
e
in
wa
s
b
o
rn
i
n
Ba
g
d
a
d
Ira
q
in
1
9
7
4
.
He
re
c
e
iv
e
d
t
h
e
B.
S
.
,
M
.
S
.
,
a
n
d
P
h
.
D.
d
e
g
re
e
s
i
n
El
e
c
tro
n
ic
E
n
g
in
e
e
rin
g
fr
o
m
Un
i
v
e
rsity
o
f
Tec
h
n
o
l
o
g
y
in
1
9
9
8
,
2
0
0
1
a
n
d
2
0
1
9
re
sp
e
c
ti
v
e
ly
.
F
r
o
m
2
0
0
0
to
2
0
0
7
h
e
wo
r
k
e
d
a
t
Un
i
v
e
rsity
o
f
Tec
h
n
o
lo
g
y
/
El
e
c
tri
c
a
l
En
g
i
n
e
e
rin
g
D
e
p
a
rtme
n
t
a
s
lec
tu
re
r.
S
i
n
c
e
2
0
0
7
,
h
e
h
a
s
b
e
e
n
wo
rk
i
n
g
a
s
a
lec
tu
re
r
in
Un
iv
e
rsity
o
f
Ke
rb
a
la
/
C
o
ll
a
g
e
o
f
E
n
g
i
n
e
e
rin
g
.
He
wo
rk
e
d
a
lso
in
m
a
n
y
c
o
m
p
a
n
ies
a
s
a
Bro
a
d
c
a
stin
g
e
n
g
i
n
e
e
r.
Du
rin
g
t
h
is
p
e
rio
d
m
a
n
y
re
se
a
rc
h
we
re
p
u
b
li
sh
e
d
i
n
in
ter
n
a
ti
o
n
a
l
a
n
d
lo
c
a
l
c
o
n
fe
re
n
c
e
s
a
n
d
j
o
u
r
n
a
ls
in
th
e
fiel
d
o
f
S
p
e
e
c
h
Re
c
o
g
n
i
ti
o
n
s,
P
ro
st
h
e
sis
Co
n
tr
o
l,
An
ten
n
a
De
sig
n
a
n
d
M
o
b
il
e
Ra
d
i
a
ti
o
n
P
ro
tec
ti
o
n
S
y
ste
m
s.
Dr
.
Abd
u
l
k
a
d
h
im
A.
S
a
lm
a
n
w
a
s
b
o
r
n
in
Na
jaf,
Ira
q
i
n
1
9
6
4
.
H
e
re
c
e
iv
e
d
t
h
e
B.
S
.
,
M
.
S
.
a
n
d
P
h
.
D.
d
e
g
re
e
s
in
El
e
c
tro
n
ic
En
g
i
n
e
e
rin
g
fr
o
m
Un
iv
e
rsity
o
f
Tec
h
n
o
l
o
g
y
in
1
9
8
7
,
2
0
0
2
a
n
d
2
0
1
8
re
sp
e
c
ti
v
e
l
y
.
F
r
o
m
1
9
8
7
to
1
9
9
3
,
h
e
wo
r
k
e
d
a
t
t
h
e
Ato
m
ic
En
e
rg
y
Org
a
n
iza
ti
o
n
i
n
th
e
field
o
f
d
e
sig
n
i
n
g
e
lec
tro
n
ic
c
ircu
it
s
fo
r
g
e
n
e
ra
ti
n
g
a
n
d
m
e
a
su
rin
g
h
i
g
h
-
s
p
e
e
d
c
u
rre
n
t
p
u
lse
s
a
s
we
ll
a
s
in
th
e
u
se
o
f
d
i
g
it
a
l
a
n
d
l
o
g
ica
l
c
ircu
it
s
i
n
t
h
e
d
e
sig
n
s
o
f
c
o
n
tro
l
sy
ste
m
s.
S
in
c
e
1
9
9
4
,
h
e
h
a
s
b
e
e
n
wo
rk
in
g
a
s
a
lec
tu
re
r
a
t
Tec
h
n
ica
l
In
stit
u
te
o
f
Ka
rb
a
la
a
t
Al
-
F
u
ra
t
Al
-
Aw
sa
t
Tec
h
n
ica
l
Un
i
v
e
rsity
.
Du
ri
n
g
th
is
p
e
ri
o
d
m
a
n
y
re
se
a
rc
h
e
s
we
re
p
u
b
l
ish
e
d
in
in
tern
a
ti
o
n
a
l
a
n
d
lo
c
a
l
c
o
n
fe
re
n
c
e
s
a
n
d
j
o
u
r
n
a
ls
i
n
t
h
e
fiel
d
o
f
fil
ter
d
e
sig
n
,
sw
it
c
h
in
g
c
a
p
a
c
it
o
r
tec
h
n
iq
u
e
,
c
h
a
n
n
e
l
m
o
d
e
l,
p
o
we
r
li
n
e
c
o
m
m
u
n
ica
ti
o
n
s
(
P
LC),
O
F
DM
m
o
d
u
latio
n
,
a
u
to
m
a
ti
o
n
u
si
n
g
F
P
G
A an
d
M
icr
o
c
o
n
tr
o
ll
e
r.
Th
a
m
ir
Ra
shed
S
a
e
e
d
wa
s
Bo
r
n
i
n
Ba
g
h
d
a
d
,
Ira
q
,
o
n
F
e
b
ru
a
ry
1
0
,
1
9
6
5
.
He
re
c
e
iv
e
d
t
h
e
B.
S
c
.
De
g
re
e
fro
m
m
il
it
a
ry
e
n
g
i
n
e
e
rin
g
c
o
ll
e
g
e
i
n
Ba
g
h
d
a
d
i
n
1
9
8
7
,
th
e
M
.
S
c
.
De
g
re
e
fro
m
m
il
it
a
ry
e
n
g
i
n
e
e
rin
g
c
o
ll
e
g
e
i
n
B
a
g
h
d
a
d
in
1
9
9
4
a
n
d
P
h
.
D.
d
e
g
re
e
fro
m
AL
-
Ra
sh
e
d
c
o
ll
e
g
e
o
f
e
n
g
in
e
e
rin
g
a
n
d
S
e
c
in
e
c
i
n
Ba
g
h
d
a
d
2
0
0
3
.
F
ro
m
1
9
9
4
to
2
0
0
3
,
h
e
wo
r
k
e
d
wit
h
m
il
it
a
ry
e
n
g
in
e
e
rin
g
c
o
ll
e
g
e
i
n
Ba
g
h
d
a
d
a
s
a
m
e
m
b
e
r
o
f
tea
c
h
i
n
g
sta
ff.
F
ro
m
2
0
0
3
ti
ll
n
o
w,
h
e
wo
rk
e
d
wit
h
th
e
Un
i
v
e
rsity
o
f
Tec
h
n
o
l
o
g
y
in
Ba
g
h
d
a
d
a
s
a
m
e
m
b
e
r
o
f
tea
c
h
in
g
sta
ff.
Cu
rre
n
tl
y
,
h
e
is
th
e
As
st.
P
ro
fe
ss
o
r
o
f
e
lec
tri
c
a
l
e
n
g
in
e
e
rin
g
a
t
u
n
i
v
e
rsity
o
f
Tec
h
n
o
l
o
g
y
a
n
d
a
h
e
a
d
o
f
ra
d
a
r
re
se
a
rc
h
g
ro
u
p
in
th
e
El
e
c
tri
c
a
l
En
g
.
De
p
t.
His
m
a
jo
r
i
n
tere
sts
a
re
in
d
i
g
it
a
l
sig
n
a
l
p
ro
c
e
ss
in
g
,
d
ig
it
a
l
c
ircu
it
d
e
sig
n
f
o
r
DS
P
b
a
se
d
o
n
F
P
G
A,
se
n
so
r
n
e
tw
o
rk
a
n
d
P
a
tt
e
rn
Re
c
o
g
n
it
i
o
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.