I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
1
2
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
2
,
p
p
.
946
~
956
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
1
2
i
1
.
pp
9
4
6
-
956
946
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Cla
ss
ificatio
n of
t
hree patho
lo
g
ica
l
v
o
ices ba
sed o
n specific
features gro
ups
u
sing
suppo
rt
v
ec
t
o
r
m
a
chin
e
M
un
ee
ra
Alt
a
y
eb
1
,
A
m
a
ni A
l
-
G
hra
iba
h
2
1
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s
a
n
d
C
o
m
mu
n
i
c
a
t
i
o
n
s
En
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
,
Al
-
A
h
l
i
y
y
a
A
mm
a
n
U
n
i
v
e
r
si
t
y
,
A
mm
a
n
,
Jo
r
d
a
n
2
D
e
p
a
r
t
me
n
t
o
f
M
e
d
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
,
Al
-
A
h
l
i
y
y
a
A
mm
a
n
U
n
i
v
e
r
si
t
y
,
A
mm
a
n
,
Jo
r
d
a
n
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Feb
1
0
,
2
0
2
1
R
ev
i
s
ed
A
u
g
3
,
2
0
2
1
A
cc
ep
ted
A
u
g
2
0
,
2
0
2
1
De
ter
m
in
in
g
a
n
d
c
las
sify
in
g
p
a
th
o
l
o
g
ica
l
h
u
m
a
n
so
u
n
d
s
a
re
stil
l
a
n
in
tere
stin
g
a
re
a
o
f
re
se
a
rc
h
in
t
h
e
f
ield
o
f
sp
e
e
c
h
p
r
o
c
e
ss
in
g
.
T
h
is
p
a
p
e
r
e
x
p
lo
re
s
d
if
fe
re
n
t
m
e
th
o
d
s
o
f
v
o
ice
f
e
a
tu
re
s
e
x
tra
c
ti
o
n
,
n
a
m
e
l
y
:
M
e
l
f
re
q
u
e
n
c
y
c
e
p
stra
l
c
o
e
ff
icie
n
ts
(M
F
CCs
),
z
e
ro
-
c
ro
ss
in
g
ra
te
(
ZCR)
a
n
d
d
isc
re
te
w
a
v
e
let
tran
s
f
o
rm
(D
WT
).
A
c
o
m
p
a
riso
n
is
m
a
d
e
b
e
t
we
e
n
th
e
se
m
e
th
o
d
s
in
o
r
d
e
r
to
id
e
n
ti
f
y
th
e
ir
a
b
il
it
y
in
c
las
si
fy
in
g
a
n
y
in
p
u
t
so
u
n
d
a
s
a
n
o
rm
a
l
o
r
p
a
th
o
l
o
g
ica
l
v
o
ice
s
u
sin
g
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
(S
V
M
).
F
irstl
y
,
th
e
v
o
ic
e
sig
n
a
l
is
p
ro
c
e
ss
e
d
a
n
d
f
il
tere
d
,
th
e
n
v
o
c
a
l
f
e
a
tu
re
s
a
r
e
e
x
trac
ted
u
sin
g
th
e
p
ro
p
o
se
d
m
e
th
o
d
s
a
n
d
f
in
a
ll
y
si
x
g
ro
u
p
s
o
f
f
e
a
tu
re
s
a
r
e
u
se
d
to
c
las
si
fy
th
e
v
o
ice
d
a
ta
a
s
h
e
a
l
th
y
,
h
y
p
e
r
k
in
e
ti
c
d
y
sp
h
o
n
ia,
h
y
p
o
k
in
e
ti
c
d
y
sp
h
o
n
ia,
o
r
re
f
lu
x
lar
y
n
g
it
is
u
sin
g
se
p
a
ra
te
c
las
si
f
ica
ti
o
n
p
ro
c
e
ss
e
s.
T
h
e
c
las
si
f
ica
ti
o
n
re
su
lt
s
re
a
c
h
1
0
0
%
a
c
c
u
ra
c
y
u
sin
g
th
e
M
F
CC
a
n
d
k
u
rt
o
sis
f
e
a
tu
re
g
ro
u
p
.
W
h
il
e
t
h
e
o
t
h
e
r
c
las
sif
i
c
a
ti
o
n
a
c
c
u
ra
c
ies
ra
n
g
e
b
e
t
w
e
e
n
~
6
0
%
to
~
9
7
%
.
T
h
e
W
a
v
e
let
f
e
a
tu
re
s
p
ro
v
id
e
v
e
ry
g
o
o
d
c
las
sif
ica
ti
o
n
re
su
lt
s
in
c
o
m
p
a
riso
n
w
it
h
o
th
e
r
c
o
m
m
o
n
v
o
ice
f
e
a
tu
re
s
li
k
e
M
F
CC
a
n
d
ZCR
f
e
a
tu
re
s.
T
h
is p
a
p
e
r
a
im
s to
im
p
r
o
v
e
th
e
d
iag
n
o
sis
o
f
v
o
ice
d
iso
rd
e
rs w
it
h
o
u
t
th
e
n
e
e
d
f
o
r
su
rg
ica
l
in
terv
e
n
ti
o
n
s
a
n
d
e
n
d
o
sc
o
p
ic
p
ro
c
e
d
u
r
e
s
w
h
ich
c
o
n
su
m
e
s
ti
m
e
a
n
d
b
u
rd
e
n
th
e
p
a
ti
e
n
ts.
A
lso
,
th
e
c
o
m
p
a
riso
n
b
e
tw
e
e
n
th
e
p
ro
p
o
se
d
f
e
a
tu
re
e
x
trac
ti
o
n
m
e
th
o
d
s
o
f
f
e
r
s
a
g
o
o
d
re
f
e
r
e
n
c
e
fo
r
f
u
rth
e
r
re
se
a
rc
h
e
s in
th
e
v
o
ice
c
las
sif
ic
a
ti
o
n
a
re
a
.
K
ey
w
o
r
d
s
:
Dis
cr
ete
w
av
ele
t tr
an
s
f
o
r
m
Me
l f
r
eq
u
e
n
c
y
ce
p
s
tr
al
-
co
ef
f
icie
n
t
s
Su
p
p
o
r
t v
ec
to
r
m
ac
h
i
n
e
Vo
ice
d
is
o
r
d
er
s
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
:
Mu
n
ee
r
a
A
lta
y
eb
Dep
ar
t
m
en
t o
f
E
lectr
o
n
ics a
n
d
C
o
m
m
u
n
icat
io
n
s
E
n
g
in
ee
r
i
n
g
,
Facu
lt
y
o
f
E
n
g
i
n
ee
r
in
g
,
Al
-
Ah
li
y
y
a
Am
m
a
n
Un
i
v
er
s
it
y
Al
-
Sar
o
,
A
l
-
Salt,
Am
m
an
,
J
o
r
d
an
E
m
ail:
m
.
alta
y
eb
@
a
m
m
an
u
.
e
d
u
.
j
o
1.
I
NT
RO
D
UCT
I
O
N
Sp
ee
ch
i
s
co
n
s
id
er
ed
as
o
n
e
o
f
t
h
e
m
o
s
t
i
m
p
o
r
tan
t
m
ea
n
s
o
f
co
m
m
u
n
icatio
n
a
m
o
n
g
h
u
m
a
n
s
.
T
h
er
ef
o
r
e,
w
h
e
n
a
n
y
d
ef
ec
t
o
cc
u
r
s
i
n
t
h
e
s
p
ee
ch
s
y
s
te
m
,
t
h
is
co
n
s
id
er
ed
as
a
n
i
m
p
ed
i
m
e
n
t
in
co
m
m
u
n
icatio
n
a
m
o
n
g
p
eo
p
le.
Dif
f
ic
u
lt
y
i
n
s
p
ee
ch
m
a
y
ar
is
e
d
u
e
to
i
m
b
a
lan
ce
in
th
e
s
p
ee
ch
o
r
au
d
ito
r
y
s
y
s
te
m
[
1
]
.
Ma
n
y
r
esear
c
h
er
s
i
n
liter
atu
r
e
h
av
e
s
t
u
d
ied
s
p
ee
ch
d
is
o
r
d
er
s
an
d
v
o
ca
l
p
ath
o
lo
g
y
b
y
an
a
l
y
zin
g
a
n
d
class
i
f
y
in
g
s
a
m
p
les
o
f
p
atie
n
t
’
s
v
o
ice
s
.
T
h
e
p
u
r
p
o
s
e
w
a
s
to
h
elp
p
atie
n
t
s
w
it
h
p
ath
o
lo
g
ic
al
p
r
o
b
lem
s
a
n
d
to
m
o
n
ito
r
t
h
e
p
r
o
g
r
ess
o
f
t
h
e
v
o
ca
l
t
h
er
ap
y
p
ath
w
a
y
a
n
d
to
m
i
n
i
m
ize
t
h
e
u
s
e
o
f
tr
a
d
itio
n
al
d
iag
n
o
s
tic
p
ath
o
lo
g
ies
o
f
v
o
ca
l
p
ath
o
lo
g
y
.
R
esear
ch
er
s
d
ev
e
lo
p
ed
m
a
n
y
d
i
ag
n
o
s
i
s
m
et
h
o
d
s
f
o
r
o
b
s
er
v
atio
n
s
o
f
v
o
ca
l
f
o
ld
s
b
y
m
ea
n
s
o
f
lap
ar
o
s
co
p
ic
to
o
ls
.
Ho
w
e
v
er
,
t
h
ese
tec
h
n
iq
u
es
ar
e
r
is
k
y
,
ti
m
e
co
n
s
u
m
i
n
g
,
d
i
s
co
m
f
o
r
t
an
d
r
eq
u
ir
e
ex
p
en
s
i
v
e
r
eso
u
r
ce
s
[
2
]
,
[
3
]
.
Fro
m
th
e
in
An
k
ış
h
an
w
o
r
k
[
4
]
,
a
n
e
w
ap
p
r
o
ac
h
f
o
r
d
etec
tio
n
o
f
p
ath
o
lo
g
ical
v
o
ice
d
is
o
r
d
er
s
w
a
s
d
ev
elo
p
ed
w
i
th
m
in
i
m
u
m
p
ar
a
m
eter
s
.
T
h
e
p
r
e
p
r
o
ce
s
s
in
g
s
tep
h
as
b
ee
n
ca
r
r
ied
o
u
t
s
i
n
ce
th
e
r
ec
o
r
d
in
g
o
f
t
h
e
s
o
u
n
d
d
ata.
T
h
e
s
o
u
n
d
d
ata
i
s
r
e
-
m
o
d
eled
w
it
h
th
e
ca
lc
u
lated
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
C
la
s
s
i
fica
tio
n
o
f th
r
ee
p
a
th
o
lo
g
ica
l v
o
ices b
a
s
ed
o
n
…
(
Mu
n
ee
r
a
A
lta
ye
b
)
947
co
ef
f
icie
n
t
s
u
s
i
n
g
t
w
o
d
i
f
f
e
r
en
t
m
o
d
els:
lin
ea
r
p
r
ed
icti
v
e
co
d
i
n
g
(
L
P
C
)
an
d
M
el
f
r
eq
u
en
c
y
ce
p
s
tr
al
co
ef
f
icie
n
t
s
(
MF
C
C
)
.
T
h
e
n
,
t
h
e
r
ec
o
r
d
ed
s
p
ee
ch
i
s
d
iv
id
ed
in
to
t
w
o
d
i
f
f
er
en
t
s
i
g
n
al
t
y
p
es
,
clea
n
a
n
d
r
esid
u
al
s
ig
n
al
s
.
T
h
e
s
ig
n
als
t
h
at
m
o
d
eled
f
r
o
m
th
e
co
e
f
f
ic
ien
t
s
a
r
e
ca
lled
clea
n
d
ata
an
d
ar
e
r
e
m
o
v
ed
f
r
o
m
th
e
r
ec
o
r
d
e
d
d
ata
to
g
en
er
ate
th
e
r
esid
u
al
d
ata.
T
h
e
m
o
d
eled
s
ig
n
als
w
er
e
f
ir
s
t
s
ep
ar
ated
in
to
s
u
b
-
f
r
a
m
e
s
,
th
e
n
t
h
e
ch
ar
ac
ter
is
tic
s
(
f
ea
t
u
r
es)
o
f
ea
ch
f
r
a
m
e
w
er
e
e
x
tr
ac
ted
,
th
e
f
ea
tu
r
es
th
a
t
ar
e
ex
tr
ac
ted
in
th
eir
s
tu
d
y
ar
e:
j
itter
,
s
h
i
m
m
er
,
s
k
e
w
n
e
s
s
,
k
u
r
to
s
is
,
en
tr
o
p
y
,
an
d
lar
g
est
l
y
ap
u
n
o
v
e
x
p
o
n
en
ts
(
L
L
E
s
)
.
Fi
n
a
ll
y
,
a
p
ath
o
lo
g
ical
class
i
f
icatio
n
w
as
m
ad
e
d
ep
en
d
in
g
o
n
t
h
ese
f
ea
t
u
r
es,
w
h
e
r
e
3
0
%
o
f
th
e
d
ata
w
er
e
r
an
d
o
m
l
y
s
e
lecte
d
as
test
i
n
g
d
ata
an
d
7
0
%
w
er
e
r
a
n
d
o
m
l
y
s
e
lecte
d
as
tr
ain
i
n
g
d
ata.
T
h
e
class
i
f
icat
io
n
p
er
f
o
r
m
an
ce
i
n
t
h
eir
w
o
r
k
w
a
s
v
er
y
g
o
o
d
.
Usi
n
g
t
h
e
te
n
f
ea
t
u
r
es,
t
h
e
ac
c
u
r
ac
y
r
ate
o
f
tr
ain
i
n
g
d
ata
w
as
1
0
0
%,
h
o
w
ev
er
,
th
e
e
s
ti
m
ated
ac
cu
r
ac
y
f
o
r
test
in
g
d
ata
w
as
9
9
.
5
6
% [
4
]
.
A
d
etec
tio
n
o
f
p
ath
o
lo
g
ical
v
o
ices
w
as
al
s
o
d
ev
elo
p
ed
b
y
Fan
g
[
5
]
,
u
s
in
g
ce
p
s
tr
u
m
v
ec
to
r
s
an
d
a
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
.
T
h
is
s
tu
d
y
r
etr
o
s
p
ec
tiv
el
y
co
llec
ted
6
0
n
o
r
m
al
v
o
ice
s
a
m
p
les
a
n
d
4
0
2
p
ath
o
lo
g
ical
v
o
ice
s
a
m
p
les
o
f
8
co
m
m
o
n
clin
ical
v
o
ice
d
is
o
r
d
er
s
in
a
v
o
ice
clin
ic
o
f
a
ter
tiar
y
teac
h
in
g
h
o
s
p
ital.
T
h
e
y
ex
tr
ac
ted
MFC
C
s
f
r
o
m
3
-
s
ec
o
n
d
s
a
m
p
l
e
s
o
f
a
s
u
s
tai
n
ed
v
o
w
el.
T
h
e
p
er
f
o
r
m
an
ce
s
o
f
th
r
e
e
m
ac
h
i
n
e
lear
n
i
n
g
alg
o
r
ith
m
s
,
n
a
m
e
l
y
,
d
ee
p
n
e
u
r
al
n
et
w
o
r
k
(
DNN)
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e
(
SVM)
,
an
d
Gau
s
s
ian
m
i
x
tu
r
e
m
o
d
el
(
GM
M)
,
w
er
e
ev
al
u
ate
d
b
ased
o
n
a
f
i
v
e
f
o
ld
cr
o
s
s
-
v
a
lid
atio
n
.
C
o
llect
iv
e
ca
s
es
f
r
o
m
t
h
e
v
o
ice
d
is
o
r
d
er
d
atab
ase
o
f
Ma
s
s
ac
h
u
s
ett
s
E
y
e
an
d
E
ar
I
n
f
ir
m
ar
y
(
ME
E
I
)
w
er
e
u
s
ed
to
v
er
if
y
th
e
p
er
f
o
r
m
an
ce
o
f
t
h
e
class
i
f
icatio
n
m
ec
h
an
is
m
s
.
T
h
e
ex
p
er
i
m
en
tal
r
esu
lt
s
d
e
m
o
n
s
tr
ated
th
at
DNN
o
u
tp
er
f
o
r
m
s
GM
M
an
d
SVM.
I
ts
ac
cu
r
ac
y
i
n
d
etec
tin
g
v
o
ic
e
p
ath
o
lo
g
ie
s
r
ea
ch
ed
9
4
.
2
6
%
an
d
9
0
.
5
2
%
in
m
ale
a
n
d
f
e
m
ale
s
u
b
j
ec
ts
,
b
ased
o
n
th
r
ee
r
ep
r
esen
tativ
e
M
FC
C
f
ea
t
u
r
es.
W
h
en
ap
p
lied
to
th
e
ME
E
I
d
atab
ase
f
o
r
v
alid
atio
n
,
th
e
DNN
also
ac
h
iev
ed
a
h
i
g
h
er
ac
cu
r
ac
y
(
9
9
.
3
2
%)
th
an
th
e
o
t
h
er
t
w
o
class
i
f
icat
io
n
al
g
o
r
ith
m
s
.
T
h
e
y
co
n
c
l
u
d
ed
th
at
s
tack
i
n
g
s
e
v
er
al
la
y
er
s
o
f
n
eu
r
o
n
s
w
it
h
o
p
ti
m
ized
w
ei
g
h
ts
,
th
e
p
r
o
p
o
s
ed
DNN
alg
o
r
ith
m
ca
n
f
u
ll
y
u
tili
ze
t
h
e
ac
o
u
s
tic
f
ea
tu
r
e
s
an
d
ef
f
icie
n
tl
y
d
if
f
er
en
t
iate
b
et
w
ee
n
n
o
r
m
a
l a
n
d
p
ath
o
lo
g
ical
v
o
ice
s
a
m
p
l
es [
5
]
.
P
an
ek
et
a
l.
[
6
]
cr
ea
ted
an
a
co
u
s
tic
a
n
a
l
y
s
is
as
s
es
s
m
e
n
t
i
n
d
etec
tin
g
f
o
u
r
m
aj
o
r
s
p
ee
c
h
d
is
ea
s
es
:
ex
ce
s
s
iv
e
d
y
s
f
u
n
ctio
n
,
d
y
s
f
u
n
ctio
n
,
lar
y
n
g
iti
s
,
v
o
ca
l
co
r
d
p
a
r
al
y
s
i
s
.
A
t
t
h
e
b
eg
i
n
n
i
n
g
,
2
8
a
co
u
s
tic
p
ar
a
m
eter
s
w
er
e
e
v
al
u
ated
b
y
e
x
a
m
i
n
ati
o
n
.
T
h
e
an
al
y
s
i
s
o
f
th
e
s
p
e
ec
h
s
i
g
n
al
w
a
s
p
er
f
o
r
m
ed
b
y
e
x
t
r
ac
tin
g
m
a
n
y
f
ea
t
u
r
es,
n
a
m
el
y
:
f
u
n
d
a
m
e
n
tal
f
r
eq
u
e
n
c
y
,
j
itter
a
n
d
s
h
i
m
m
er
co
ef
f
icien
t
s
,
e
n
er
g
y
,
ze
r
o
th
,
f
ir
s
t,
s
ec
o
n
d
,
t
h
ir
d
-
o
r
d
er
m
o
m
en
t,
k
u
r
to
s
i
s
,
p
o
w
e
r
f
ac
to
r
,
1
,
2
a
n
d
3
-
f
o
r
m
a
n
t
a
m
p
lit
u
d
e,
1
,
2
an
d
3
-
f
o
r
m
a
n
t
f
r
eq
u
en
c
y
,
m
ax
i
m
u
m
an
d
m
i
n
i
m
u
m
v
al
u
es
o
f
th
e
s
ig
n
a
l
an
d
1
0
MFC
C
s
.
T
h
e
cla
s
s
i
f
icatio
n
co
n
s
i
s
ted
o
f
r
es
u
lt
s
f
r
o
m
t
h
e
a
n
al
y
s
i
s
f
o
r
ea
ch
p
atie
n
t.
I
t
w
a
s
a
n
al
y
z
ed
u
s
i
n
g
th
r
ee
m
et
h
o
d
s
:
P
r
in
c
ip
al
co
m
p
o
n
en
t
an
al
y
s
is
(
P
C
A
)
,
k
er
n
el
p
r
i
n
cip
al
co
m
p
o
n
e
n
t
an
a
l
y
s
is
(
KP
C
A
)
an
d
an
a
u
to
-
a
s
s
o
ciati
v
e
n
e
u
r
al
n
et
w
o
r
k
(
N
L
P
C
A
)
.
T
en
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
w
a
s
u
s
ed
,
w
h
er
e
th
e
d
ata
w
a
s
d
iv
id
ed
in
to
1
0
s
u
b
s
et
s
;
1
0
%
o
f
th
e
d
ata
w
a
s
u
s
ed
as
a
test
i
n
g
s
et,
an
d
th
e
r
e
m
ain
in
g
9
0
%
w
as
r
ep
r
esen
ti
n
g
as
a
tr
ai
n
i
n
g
s
et.
T
h
e
an
a
l
y
s
is
w
as
co
m
p
leted
i
n
d
iv
id
u
all
y
f
o
r
ea
ch
v
o
w
e
l
a
t
d
if
f
er
e
n
t
in
to
n
atio
n
s
,
s
ep
ar
atel
y
f
o
r
m
e
n
an
d
w
o
m
en
f
o
r
ea
ch
p
ath
o
lo
g
y
a
n
d
ea
ch
v
o
w
el
at
a
d
if
f
er
en
t
p
itch
.
T
h
e
aim
o
f
th
e
ir
r
esear
ch
w
a
s
to
p
er
f
o
r
m
a
class
if
ica
tio
n
th
a
t
ca
n
d
is
t
in
g
u
is
h
b
et
w
ee
n
h
ea
lt
h
y
a
n
d
p
ath
o
lo
g
ical
v
o
ice
s
[
6
]
.
T
h
e
n
o
v
elt
y
o
f
t
h
i
s
w
o
r
k
lie
s
in
e
x
t
r
ac
ti
n
g
n
e
w
f
ea
t
u
r
es
f
r
o
m
h
ea
lt
h
y
v
o
ices
a
n
d
t
h
r
ee
d
if
f
er
e
n
t
p
ath
o
lo
g
ical
v
o
ice
s
a
m
p
le
s
f
o
llo
w
ed
b
y
s
e
v
er
al
cla
s
s
i
f
ica
tio
n
p
r
o
ce
s
s
es
to
cla
s
s
i
f
y
t
h
e
v
o
ice
s
a
m
p
le
s
a
s
h
ea
lt
h
y
o
r
p
ath
o
lo
g
y
v
o
ices
u
s
i
n
g
s
p
ec
i
f
ic
f
ea
t
u
r
e
g
r
o
u
p
s
w
h
ic
h
co
n
tai
n
a
co
m
b
i
n
atio
n
o
f
t
h
e
e
x
tr
ac
ted
f
ea
t
u
r
es.
T
h
e
ex
tr
ac
ted
f
ea
tu
r
es
ar
e
m
ai
n
l
y
t
h
r
ee
d
if
f
er
en
t
MFC
C
f
ea
t
u
r
e
g
r
o
u
p
s
an
d
w
a
v
elet
f
ea
t
u
r
es
g
r
o
u
p
.
A
l
s
o
,
th
e
d
is
cr
ete
w
a
v
elet
tr
a
n
s
f
o
r
m
(
DW
T
)
m
et
h
o
d
is
u
n
i
q
u
e
an
d
h
a
s
n
o
t
b
ee
n
u
s
ed
b
e
f
o
r
e
in
t
h
e
m
ea
n
o
f
v
o
ice
d
is
o
r
d
er
class
if
icat
io
n
.
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
T
h
e
m
ai
n
s
tep
s
o
f
t
h
is
w
o
r
k
b
eg
an
w
it
h
ex
tr
ac
tin
g
f
ea
tu
r
es
f
r
o
m
th
e
v
o
ice
d
ata,
th
en
u
s
e
th
e
f
ea
t
u
r
e
g
r
o
u
p
s
to
b
u
ild
an
au
to
m
ated
s
y
s
te
m
u
s
i
n
g
SVM
w
h
ich
cla
s
s
if
y
th
e
i
n
p
u
t
d
ata
as
n
o
r
m
al
o
r
p
ath
o
lo
g
y
v
o
ices.
T
h
e
f
ea
tu
r
e
s
o
f
th
e
v
o
ice
s
a
m
p
les
w
er
e
ex
tr
ac
ted
u
s
i
n
g
:
MFC
C
s
,
ze
r
o
-
cr
o
s
s
in
g
r
ate
(
Z
C
R
)
a
n
d
d
is
cr
ete
w
a
v
elet
tr
a
n
s
f
o
r
m
(
DW
T
)
in
ad
d
itio
n
to
o
th
er
s
tati
s
tical
f
ea
tu
r
es
w
h
ic
h
ar
e:
s
k
e
w
n
ess
,
k
u
r
to
s
is
,
an
d
en
tr
o
p
y
.
Fig
u
r
e
1
s
h
o
w
s
t
h
e
b
lo
ck
d
iag
r
a
m
o
f
th
e
p
r
o
p
o
s
ed
w
o
r
k
.
T
h
is
p
ap
er
is
o
r
g
a
n
ized
as
f
o
llo
w
s
:
s
ec
tio
n
1
co
n
tai
n
s
a
b
r
ief
d
escr
ip
tio
n
o
f
t
h
e
d
ata
u
s
ed
in
t
h
i
s
w
o
r
k
;
s
ec
tio
n
2
p
r
esen
ts
th
e
m
e
th
o
d
o
lo
g
y
o
f
t
h
is
w
o
r
k
.
Sectio
n
3
p
r
esen
ts
th
e
r
es
u
lt
s
o
f
class
i
f
icatio
n
.
S
ec
tio
n
4
co
n
tai
n
s
a
co
n
cl
u
s
io
n
o
f
all
t
h
e
w
o
r
k
f
ea
tu
r
ed
in
t
h
is
p
ap
er
.
2
.
1
.
Da
t
a
ba
s
e
T
h
e
r
eg
is
ter
ed
d
atab
ase
w
h
ic
h
w
a
s
u
s
ed
in
t
h
is
s
tu
d
y
co
n
t
ain
s
v
o
ice
s
o
f
h
ea
lt
h
y
a
n
d
p
ath
o
lo
g
ical
p
eo
p
le
b
esid
es
m
u
c
h
in
f
o
r
m
a
tio
n
ab
o
u
t
ea
ch
p
atie
n
t.
T
h
e
v
o
ice
s
i
g
n
a
l
ac
q
u
is
itio
n
s
w
er
e
p
er
f
o
r
m
ed
i
n
t
h
e
Ho
s
p
ital
Un
i
v
er
s
it
y
o
f
Nap
les
“
Fed
er
ico
I
I
”,
at
th
e
m
ed
ic
al
r
o
o
m
o
f
th
e
“
I
n
s
tit
u
te
o
f
Hig
h
-
P
er
f
o
r
m
an
c
e
C
o
m
p
u
tin
g
a
n
d
Net
w
o
r
k
i
n
g
”
[
7
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
1
2
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
2
:
9
4
6
-
956
948
T
h
e
r
ec
o
r
d
in
g
s
w
er
e
m
ad
e
u
s
in
g
a
m
o
b
ile
m
icr
o
p
h
o
n
e
w
h
i
ch
w
as
h
eld
at
ab
o
u
t
2
0
cm
a
w
a
y
f
r
o
m
th
e
p
atien
t.
E
ac
h
s
i
g
n
a
l
co
n
s
is
ts
o
f
a
r
ec
o
r
d
in
g
o
f
v
o
ca
li
za
tio
n
o
f
t
h
e
v
o
w
el
\
\
a
\
\
f
i
v
e
s
ec
o
n
d
s
in
le
n
g
t
h
w
it
h
o
u
t
an
y
i
n
ter
r
u
p
tio
n
o
f
o
th
er
s
o
u
n
d
s
.
T
h
e
r
ec
o
r
d
in
g
s
ig
n
als
w
er
e
s
a
m
p
led
at
8
0
0
0
Hz
an
d
th
eir
r
eso
lu
tio
n
w
a
s
3
2
-
b
its
.
All
s
a
m
p
le
s
w
er
e
r
ec
o
r
d
ed
in
less
th
an
3
0
d
B
o
f
b
ac
k
g
r
o
u
n
d
n
o
is
e
a
n
d
r
o
o
m
h
u
m
id
it
y
w
as
g
r
ea
ter
th
a
n
3
0
-
40
%.
E
ac
h
r
ec
o
r
d
in
g
w
a
s
f
ilter
ed
to
r
em
o
v
e
a
n
y
n
o
is
e
ac
cid
e
n
tall
y
ad
d
ed
d
u
r
in
g
t
h
e
ac
q
u
is
itio
n
[
8
]
.
T
ab
le
1
s
h
o
w
s
th
e
d
etails o
f
t
h
e
n
u
m
b
er
o
f
v
o
ice
s
a
m
p
les o
f
ea
c
h
ca
s
e
u
s
e
d
i
n
th
i
s
w
o
r
k
.
Fig
u
r
e
1
.
B
lo
ck
d
iag
r
a
m
o
f
t
h
e
p
r
o
p
o
s
ed
w
o
r
k
T
ab
le
1
.
P
atien
t d
atab
ase
P
a
t
i
e
n
t
’
s s
t
a
t
u
s
F
e
mal
e
M
a
l
e
T
o
t
a
l
H
e
a
l
t
h
y
33
21
54
H
y
p
e
r
k
i
n
e
t
i
c
d
y
sp
h
o
n
i
a
29
19
48
H
y
p
o
k
i
n
e
t
i
c
d
y
sp
h
o
n
i
a
40
22
62
R
e
f
l
u
x
l
a
r
y
n
g
i
t
i
s
18
20
38
2
.
2
.
F
e
a
t
ure
ex
t
ra
ct
io
n
Ma
n
y
m
eth
o
d
s
w
er
e
u
s
ed
i
n
l
iter
atu
r
e
to
e
x
tr
ac
t
f
ea
t
u
r
es
f
r
o
m
h
u
m
a
n
v
o
ices.
T
h
is
p
ap
er
p
r
o
p
o
s
es
th
r
ee
d
if
f
er
e
n
t
m
e
th
o
d
s
b
esid
e
th
e
s
tatis
t
ical
an
al
y
s
i
s
to
ex
tr
ac
t
s
i
x
f
ea
t
u
r
e
g
r
o
u
p
s
.
T
h
e
f
ir
s
t
t
w
o
m
et
h
o
d
s
MFC
C
s
an
d
Z
C
R
ar
e
ap
p
lied
in
th
e
ti
m
e
an
d
f
r
eq
u
e
n
c
y
d
o
m
ai
n
s
,
al
s
o
o
th
er
f
ea
t
u
r
es
ar
e
ex
tr
ac
ted
u
s
i
n
g
DW
T
.
I
n
th
e
f
ir
s
t
s
tep
,
th
e
d
at
a
w
er
e
f
il
ter
ed
u
s
in
g
p
r
o
p
er
f
il
ter
th
e
n
b
lo
ck
ed
i
n
to
3
8
f
r
a
m
e
s
,
th
e
n
d
elta,
d
elta
-
d
elta
MFC
C
,
Z
C
R
a
n
d
o
th
er
s
tatis
t
ical
f
ea
tu
r
es
w
er
e
ex
tr
ac
ted
f
r
o
m
ea
ch
f
r
a
m
e
f
o
r
all
v
o
ice
s
a
m
p
les.
DW
T
w
a
s
ca
r
r
ied
o
n
th
is
w
o
r
k
b
ase
d
o
n
5
-
lev
el
d
ec
o
m
p
o
s
itio
n
o
f
th
e
v
o
ice
s
i
g
n
a
ls
to
ex
tr
ac
t
5
f
ea
t
u
r
es:
o
n
e
f
r
o
m
ea
ch
lev
el.
Af
ter
p
r
ep
ar
in
g
t
h
e
f
ea
tu
r
e
s
,
class
i
f
icat
io
n
p
r
o
ce
s
s
es
w
er
e
p
er
f
o
r
m
ed
u
s
i
n
g
SVM
b
ased
o
n
s
ix
f
ea
t
u
r
e
g
r
o
u
p
s
an
d
b
ased
o
n
a
co
m
b
i
n
atio
n
o
f
th
e
m
o
s
t
s
i
g
n
i
f
ica
n
t
f
ea
tu
r
e
s
.
I
n
th
e
f
o
ll
o
w
i
n
g
s
ec
tio
n
s
,
an
ex
p
lan
atio
n
o
f
ea
ch
f
ea
t
u
r
e
ex
tr
ac
tio
n
m
et
h
o
d
is
p
r
o
p
o
s
ed
f
o
llo
w
ed
b
y
d
etails
o
f
th
e
f
ea
tu
r
es
g
r
o
u
p
s
th
at
w
er
e
u
s
ed
at
ea
ch
clas
s
i
f
icatio
n
p
r
o
ce
s
s
.
2
.
2
.
1
.
M
el
f
re
qu
enc
y
ce
ps
t
ra
l c
o
ef
f
icient
s
(
M
F
CC)
MFC
C
s
ar
e
t
y
p
e
s
o
f
ce
p
s
tr
al
r
ep
r
esen
tatio
n
o
f
t
h
e
s
i
g
n
al,
wh
er
e
th
e
f
r
eq
u
e
n
c
y
b
an
d
s
ar
e
d
is
tr
ib
u
ted
d
ep
en
d
in
g
o
n
t
h
e
Mel
-
s
ca
le.
T
h
e
MFC
C
s
ar
e
b
asicall
y
i
n
c
lu
d
e
w
i
n
d
o
w
i
n
g
t
h
e
s
i
g
n
al,
a
p
p
ly
i
n
g
t
h
e
d
is
cr
ete
Fo
u
r
ier
tr
an
s
f
o
r
m
a
tio
n
(
D
FT
)
,
f
o
llo
w
ed
b
y
Me
l
f
il
ter
b
an
k
s
,
tak
i
n
g
th
e
lo
g
ar
it
h
m
o
f
all
f
il
ter
b
an
k
e
n
er
g
ie
s
,
th
en
ap
p
l
y
in
g
t
h
e
d
is
cr
ete
co
s
in
e
tr
an
s
f
o
r
m
atio
n
(
D
C
T
)
.
T
h
e
s
tep
s
i
n
v
o
l
v
ed
in
t
h
e
MF
C
C
f
ea
t
u
r
es
ex
tr
ac
tio
n
ar
e
s
u
m
m
ar
ized
i
n
Fi
g
u
r
e
2
[
9
]
.
2
.
2
.
2
.
P
re
-
em
ph
a
s
i
s
P
r
e
-
e
m
p
h
as
is
r
ef
er
s
to
f
ilter
i
n
g
w
h
ic
h
e
m
p
h
as
izes
t
h
e
h
i
g
h
e
r
f
r
eq
u
en
cie
s
,
th
e
m
o
d
eli
n
g
o
f
th
e
s
o
u
n
d
s
ig
n
al
s
an
d
ex
tr
ac
tin
g
o
f
t
h
e
f
ea
t
u
r
es
f
r
o
m
t
h
e
m
o
d
e
led
d
ata
w
h
ich
ar
e
all
p
er
f
o
r
m
ed
af
ter
s
o
m
e
p
r
ep
r
o
ce
s
s
in
g
s
tep
s
.
T
h
e
s
p
ee
ch
s
i
g
n
a
l
S
(
n
)
is
s
e
n
t to
a
h
i
g
h
-
p
as
s
f
i
lter
g
i
v
en
b
y
(
1
)
,
̌
(
)
=
(
)
−
(
−
1
)
(
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
C
la
s
s
i
fica
tio
n
o
f th
r
ee
p
a
th
o
lo
g
ica
l v
o
ices b
a
s
ed
o
n
…
(
Mu
n
ee
r
a
A
lta
ye
b
)
949
w
h
er
e,
S(n
)
is
i
n
p
u
t
s
ig
n
al,
β
is
a
co
n
s
ta
n
t
an
d
it
is
ar
o
u
n
d
0
.
9
7
in
th
is
w
o
r
k
,
an
d
S
̌
(
n
)
is
th
e
s
ig
n
al
af
ter
f
iltra
tio
n
.
Fig
u
r
e
2
.
MFC
C
f
ea
t
u
r
es e
x
tr
ac
tio
n
[
1
0
]
2
.
2
.
3
.
F
ra
m
i
ng
a
nd
w
ind
o
w
ing
T
h
e
s
p
ee
ch
s
ig
n
al
i
s
ti
m
e
-
v
a
r
y
in
g
o
r
n
o
n
-
s
ta
tio
n
ar
y
s
i
g
n
a
l,
th
er
ef
o
r
e
s
p
ee
ch
a
n
al
y
s
i
s
i
s
al
w
a
y
s
p
er
f
o
r
m
ed
b
y
b
r
o
k
e
n
t
h
e
s
i
g
n
al
in
to
p
o
s
s
ib
l
y
o
v
er
lap
p
in
g
f
r
a
m
es,
s
o
th
at
t
h
e
s
p
ee
c
h
s
i
g
n
a
l is co
n
s
tan
t [
1
1
]
.
I
n
th
is
s
tep
,
t
h
e
co
n
ti
n
u
o
u
s
1
D
v
o
ice
s
ig
n
al
s
ar
e
b
lo
c
k
ed
in
to
3
8
f
r
a
m
es
o
f
N=
2
0
0
0
s
a
m
p
le
s
,
w
it
h
n
e
x
t
f
r
a
m
e
s
s
ep
ar
ated
b
y
L
=5
1
2
s
a
m
p
les,
t
h
e
ad
j
ac
en
t
f
r
a
m
e
s
ar
e
o
v
er
lap
p
ed
b
y
N
-
L
s
a
m
p
le
s
ar
o
u
n
d
7
4
%.
W
in
d
o
w
i
n
g
i
s
d
o
n
e
to
en
h
an
ce
th
e
h
ar
m
o
n
ic
s
an
d
s
m
o
o
t
h
t
h
e
ed
g
e
s
at
th
e
b
eg
in
n
i
n
g
an
d
e
n
d
in
g
p
o
in
ts
o
f
t
h
e
f
r
a
m
e
.
Ma
in
l
y
h
a
m
m
i
n
g
w
i
n
d
o
w
r
ep
r
ese
n
ted
b
y
w
(
n
)
m
u
l
tip
lies
b
y
th
e
i
n
p
u
t
s
i
g
n
al
r
ep
r
esen
ted
x
(
n
)
.
T
h
e
h
a
m
m
i
n
g
w
i
n
d
o
w
a
m
p
lit
u
d
e
is
s
h
o
w
n
i
n
Fi
g
u
r
e
3
,
th
e
o
u
tp
u
t si
g
n
al
r
ep
r
esen
te
d
b
y
(
2
)
:
(
)
=
(
)
(
−
)
,
=
0
,
.
.
.
,
−
1
(
2
)
w
h
er
e
K
is
th
e
n
u
m
b
er
o
f
f
r
am
es
a
n
d
is
th
e
n
u
m
b
er
o
f
s
a
m
p
les
b
y
w
h
ic
h
th
e
w
i
n
d
o
w
is
s
h
i
f
ted
in
o
r
d
er
to
y
ield
th
e
i
-
th
f
r
a
m
e
t
h
en
ta
k
i
n
g
th
e
D
FT
o
f
th
e
r
esu
lt
in
g
s
i
g
n
al
(
)
[
9
]
.
Fig
u
r
e
3
.
Ha
m
m
in
g
w
i
n
d
o
w
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
1
2
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
2
:
9
4
6
-
956
950
2
.
2
.
4
.
M
el
-
s
pect
ru
m
T
h
e
r
esu
lti
n
g
s
p
ec
tr
u
m
o
f
Fo
u
r
ier
tr
an
s
f
o
r
m
ed
s
ig
n
al
is
g
i
v
e
n
as
i
n
p
u
t
to
a
Me
l
-
s
ca
le
f
ilter
b
an
k
t
h
at
co
n
s
is
ts
o
f
2
4
f
ilter
s
as
s
h
o
wn
i
n
Fig
u
r
e
4
.
A
Me
l
i
s
a
u
n
i
t
o
f
m
ea
s
u
r
e
b
ased
o
n
t
h
e
h
u
m
an
ea
r
s
p
er
ce
i
v
ed
f
r
eq
u
en
c
y
w
h
er
e
h
u
m
a
n
ea
r
s
ar
e
n
o
t
s
e
n
s
i
tiv
e
en
o
u
g
h
to
d
etec
t
s
o
u
n
d
s
b
elo
w
1
0
0
0
Hz
w
h
e
n
f
r
eq
u
e
n
c
y
w
ar
p
in
g
p
r
o
ce
s
s
o
cc
u
r
s
,
th
e
c
o
ef
f
icie
n
t
s
o
f
ea
c
h
s
h
o
r
t
ti
m
e
Fo
u
r
ier
tr
an
s
f
o
r
m
ed
(
ST
FT
)
a
r
e
m
u
l
tip
lied
b
y
t
h
e
co
r
r
esp
o
n
d
in
g
f
ilter
g
ai
n
.
A
p
o
p
u
lar
f
o
r
m
u
la
to
co
n
v
er
t
f
in
h
er
tz
in
to
is
g
i
v
en
i
n
(
3
)
[
1
2
]
-
[
1
4
]
.
f
m
e
l
=
2595
l
o
g
10
(
1
+
700
)
(
3
)
Fig
u
r
e
4.
Mel
-
s
ca
le
f
ilter
b
an
k
T
h
e
DC
T
ap
p
lied
to
th
e
tr
an
s
f
o
r
m
ed
Me
l
f
r
eq
u
e
n
c
y
co
e
f
f
ic
i
en
ts
p
r
o
d
u
ce
d
a
s
et
o
f
M
FC
C
ca
m
p
estra
l
co
ef
f
icie
n
t
s
.
T
h
e
ce
p
s
tr
al
co
e
f
f
icien
ts
ar
e
u
s
u
al
l
y
r
ef
er
r
ed
to
as
s
tatio
n
ar
y
f
ea
t
u
r
es,
t
h
e
ex
tr
a
i
n
f
o
r
m
atio
n
ab
o
u
t
th
e
d
y
n
a
m
ic
f
ea
tu
r
es
o
f
th
e
s
i
g
n
al
is
o
b
tai
n
ed
b
y
co
m
p
u
tin
g
f
ir
s
t
d
er
iv
at
iv
e
o
f
ce
p
s
t
r
al
co
ef
f
icie
n
t
s
an
d
it is
ca
lled
d
elta
co
ef
f
ic
ien
t
s
,
t
h
e
s
ec
o
n
d
o
r
d
er
d
er
iv
ativ
e
is
c
alled
d
elta
-
d
elta
co
ef
f
icie
n
ts
[
1
5
]
.
Fig
u
r
e
5
(
a)
-
(
d
)
clea
r
l
y
s
h
o
w
s
t
h
e
d
if
f
er
e
n
ce
s
in
th
e
v
al
u
es
o
f
MF
C
C
co
e
f
f
icien
ts
f
o
r
t
h
e
h
ea
lt
h
y
s
i
g
n
al
co
m
p
ar
ed
w
ith
,
h
y
p
er
k
in
et
ic
p
ath
o
lo
g
y
s
i
g
n
al
,
h
y
p
o
k
in
et
ic
p
ath
o
lo
g
y
s
i
g
n
al,
an
d
r
ef
lu
x
p
at
h
o
lo
g
y
s
i
g
n
al.
T
h
ese
o
b
v
io
u
s
d
if
f
er
e
n
ce
s
i
n
t
h
e
MF
C
C
co
ef
f
icien
ts
w
i
ll
b
e
v
er
y
h
e
lp
f
u
l
in
t
h
e
cla
s
s
i
f
icatio
n
p
r
o
ce
s
s
,
w
h
er
e
th
e
x
-
a
x
is
r
ep
r
esen
ts
t
h
e
n
u
m
b
er
o
f
MF
C
C
s
ex
tr
ac
ted
f
r
o
m
t
h
e
i
n
p
u
t
s
ig
n
al
a
n
d
th
e
y
-
ax
is
r
ep
r
esen
ts
th
e
f
ea
t
u
r
e
v
alu
e
s
f
o
r
ea
ch
f
r
a
m
e.
I
n
th
i
s
w
o
r
k
,
1
2
M
FC
C
co
e
f
f
icien
ts
w
er
e
u
s
e
d
w
h
ich
ar
e
f
r
o
m
t
h
e
2
n
d
to
t
h
e
1
3
th
co
ef
f
ic
ien
t
s
an
d
th
e
r
est
w
er
e
d
is
ca
r
d
ed
.
T
h
e
lo
w
er
o
r
d
er
co
ef
f
icien
ts
co
n
tain
m
o
s
t
o
f
t
h
e
in
f
o
r
m
at
i
o
n
ab
o
u
t
th
e
o
v
er
all
s
p
ec
tr
al
s
h
ap
e
ac
co
r
d
in
g
to
t
h
e
f
ea
t
u
r
e
v
al
u
es
s
h
o
w
n
in
Fi
g
u
r
e
5
as
w
e
ca
n
o
b
s
er
v
e
t
h
e
d
i
f
f
er
en
ce
i
n
MF
C
C
s
f
o
r
th
e
f
o
u
r
ca
s
es.
2
.
2
.
5
.
Da
t
a
clus
t
er
ing
us
ing
K
-
m
ea
ns
K
-
m
ea
n
s
cl
u
s
ter
i
n
g
is
p
o
p
u
la
r
in
s
i
g
n
al
p
r
o
ce
s
s
i
n
g
f
ield
,
it
ca
n
b
e
u
s
ed
to
clu
s
ter
t
h
e
ex
tr
ac
ted
f
ea
t
u
r
es
f
r
o
m
s
p
ee
ch
s
i
g
n
als,
an
d
it
is
ap
p
lied
to
r
elativ
el
y
l
ar
g
e
s
ets
o
f
d
ata.
K
-
m
ea
n
s
s
e
p
ar
ate
th
e
d
ata
in
to
s
p
h
er
ical
cl
u
s
ter
s
b
y
s
elec
tin
g
‘
k
’
n
u
m
b
er
o
f
d
i
s
ti
n
ct
cl
u
s
ter
s
t
h
e
n
f
i
n
d
in
g
a
s
et
o
f
c
lu
s
ter
ce
n
ter
s
.
T
h
e
co
m
m
o
n
E
u
clid
ea
n
d
is
tan
ce
b
et
w
ee
n
th
e
p
o
in
t
an
d
ea
ch
clu
s
ter
ce
n
ter
is
co
m
p
u
ted
an
d
th
e
s
u
m
o
f
er
r
o
r
(
SS
E
)
is
th
e
n
ca
lcu
lated
u
s
in
g
(
4
)
[
1
6
]
,
SSE
=
∑
∑
2
∈
=
1
(
,
)
(
4
)
w
h
er
e,
x
is
a
d
ata
p
o
in
t
in
clu
s
ter
,
K
is
th
e
n
u
m
b
er
o
f
cl
u
s
t
er
s
,
co
r
r
esp
o
n
d
s
to
th
e
ce
n
ter
(
m
ea
n
)
o
f
th
e
clu
s
ter
an
d
d
is
t
is
t
h
e
E
u
clid
e
an
d
is
tan
ce
.
Her
e,
th
e
MF
C
C
s
co
ef
f
icie
n
ts
w
er
e
co
m
p
u
ted
f
o
r
ab
o
u
t
3
8
f
r
am
e
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
C
la
s
s
i
fica
tio
n
o
f th
r
ee
p
a
th
o
lo
g
ica
l v
o
ices b
a
s
ed
o
n
…
(
Mu
n
ee
r
a
A
lta
ye
b
)
951
o
f
v
o
ice
s
a
m
p
les
w
it
h
1
2
co
ef
f
icie
n
ts
r
elate
d
to
ea
ch
f
r
a
m
e.
Vec
to
r
q
u
an
tizatio
n
b
ased
o
n
K
-
m
ea
n
s
clu
s
ter
in
g
w
a
s
d
o
n
e
w
i
th
r
esp
ec
t
to
t
h
e
c
lu
s
ter
in
d
e
x
to
r
ed
u
ce
t
h
e
s
ize
o
f
t
h
e
f
ea
tu
r
e
v
ec
to
r
f
o
r
ea
ch
v
o
ice
s
i
g
n
al.
T
h
u
s
,
v
ec
to
r
o
f
1
2
f
ea
tu
r
es
w
er
e
ex
tr
ac
ted
f
r
o
m
ea
ch
v
o
ice
s
i
g
n
al
a
n
d
u
s
ed
i
n
f
u
r
th
er
cla
s
s
i
f
icatio
n
p
r
o
ce
s
s
es.
(
a)
(
b
)
(
c)
(
d
)
Fig
u
r
e
5
.
MFC
C
s
ex
tr
ac
ted
f
r
o
m
t
h
e
s
ig
n
als
;
(
a)
h
ea
lth
y
s
ig
n
al,
(
b
)
h
y
p
er
k
i
n
etic
p
ath
o
lo
g
y
s
i
g
n
a
l,
(
c)
h
y
p
o
k
i
n
etic
p
ath
o
lo
g
y
s
i
g
n
al,
(
d
)
r
ef
lu
x
p
ath
o
lo
g
y
s
i
g
n
al
2
.
3
.
St
a
t
is
t
ica
l f
ea
t
ure
ex
t
ra
ct
io
n
I
n
th
i
s
s
tu
d
y
,
th
e
f
i
lter
ed
s
ig
n
a
ls
f
r
o
m
t
h
e
p
r
ep
r
o
ce
s
s
in
g
s
tep
w
er
e
s
ep
ar
ated
i
n
to
f
r
a
m
es,
wh
ich
w
er
e
th
en
u
s
ed
to
ca
lcu
late
th
e
2
4
s
t
atis
tical
f
ea
t
u
r
es:
6
-
s
k
e
w
n
e
s
s
,
6
-
k
u
r
to
s
is
,
6
-
e
n
tr
o
p
y
a
n
d
6
-
Z
C
R
.
2
.
3
.
1
.
Sk
ew
nes
s
Sk
e
w
n
es
s
is
a
m
ea
s
u
r
e
o
f
t
h
e
d
is
to
r
tio
n
as
y
m
m
etr
y
o
f
t
h
e
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
an
d
f
o
r
an
y
s
i
g
n
al
s
̂
(
t
)
an
d
it is
d
ef
i
n
ed
as th
e
s
tan
d
a
r
d
ized
th
ir
d
m
o
m
en
t o
f
th
i
s
s
i
g
n
al
s
w
h
ich
i
s
g
iv
e
n
b
y
(
5
)
[
4
]
:
Sk
e
w
[
S
̂
(
t
)
]
=
E
⌈
S
̂
−
μ
σ
⌉
3
(
5)
w
h
er
e
,
E
r
ep
r
esen
t
s
t
h
e
e
x
p
ec
ted
o
p
er
ato
r
,
μ
is
t
h
e
m
ea
n
o
f
t
h
e
s
ig
n
al
a
n
d
σ
is
t
h
e
s
tan
d
ar
d
d
ev
iatio
n
o
f
t
h
e
s
ig
n
al.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
1
2
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
2
:
9
4
6
-
956
952
2
.
3
.
2
.
K
urt
o
s
is
T
h
e
k
u
r
to
s
i
s
o
f
t
h
e
s
i
g
n
al
i
s
d
ef
i
n
ed
as
t
h
e
s
tan
d
ar
d
ized
f
o
u
r
th
m
o
m
e
n
t
o
f
t
h
e
s
ig
n
als
S
̂
(
t
)
an
d
it
is
a
m
ea
s
u
r
e
o
f
th
e
co
m
b
i
n
ed
w
ei
g
h
t
o
f
a
d
is
tr
ib
u
t
io
n
'
s
tails
r
elativ
e
to
th
e
ce
n
ter
o
f
t
h
e
d
is
tr
ib
u
tio
n
,
a
n
d
it
i
s
g
iv
e
n
b
y
(
6
)
[
4
]
:
Sk
e
w
[
S
̂
(
t
)
]
=
E
[
(
S
̂
−
μ
)
4
]
(
E
[
(
S
̂
−
μ
)
2
]
)
2
(
6
)
w
h
er
e
,
E
r
ep
r
esen
ts
th
e
e
x
p
ec
t
ed
o
p
er
at
o
r
,
μ
is
th
e
m
ea
n
o
f
t
h
e
s
ig
n
al.
2
.
3
.
3
.
E
ntr
o
py
E
n
tr
o
p
y
i
s
th
e
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
o
f
th
e
s
i
g
n
al
S
̂
(
t
)
,
o
r
is
t
h
e
av
er
ag
e
le
v
el
o
f
in
f
o
r
m
ati
o
n
o
r
u
n
ce
r
tai
n
t
y
.
T
h
e
r
elatio
n
al
p
r
o
b
ab
ilit
y
o
f
t
h
e
ev
e
n
t
s
S
̂
i
(
w
h
er
e
i
=1
,
2,
3
,
…,
k
)
ca
lled
th
e
s
el
f
-
p
r
o
b
a
b
ilit
y
h
(
p
i
)
an
d
it
is
d
ef
i
n
ed
in
(
7
)
.
Th
e
E
n
tr
o
p
y
(
H)
is
d
ef
i
n
ed
in
(
8
)
an
d
it
is
th
e
w
e
ig
h
ts
o
f
K
n
u
m
b
er
s
o
f
s
el
f
-
in
f
o
r
m
atio
n
v
al
u
es [
1
7
]
.
ℎ
(
)
=
2
(
1
)
(
7
)
H
=
−
∑
2
(
1
)
=
1
(
8
)
2
.
3
.
4
.
Z
er
o
-
cr
o
s
s
ing
r
a
t
e
(
Z
CR)
T
h
e
ze
r
o
-
cr
o
s
s
in
g
r
ate
(
Z
C
R
)
o
f
an
au
d
io
f
r
a
m
e
is
t
h
e
r
ate
o
f
s
ig
n
c
h
an
g
es
o
f
th
e
s
ig
n
al
d
u
r
in
g
th
e
f
r
a
m
e.
T
h
e
Z
C
R
i
s
d
ef
i
n
ed
in
th
e
(
9
)
[
1
8
]
,
Z
(
i
)
=
1
2
∑
|
[
̂
(
)
]
−
=
1
[
̂
(
−
1
)
]
|
(
9
)
w
h
er
e
,
is
th
e
len
g
t
h
o
f
th
e
f
r
a
m
e
an
d
(
·
)
is
th
e
s
ig
n
f
u
n
ctio
n
.
Usu
all
y
,
Z
C
R
is
u
s
ed
to
s
ep
ar
ate
s
ig
n
al
as
v
o
iced
a
n
d
u
n
v
o
iced
,
b
u
t
h
er
e
th
is
m
eth
o
d
i
s
u
s
ed
to
e
x
tr
ac
t
f
ea
t
u
r
es
w
h
ic
h
w
ill
h
elp
t
o
class
i
f
y
t
h
e
v
o
ice
s
ig
n
al
a
s
n
o
r
m
al
o
r
ab
n
o
r
m
a
l
v
o
ices
ac
co
r
d
in
g
to
t
h
e
v
o
ice
s
i
g
n
al
n
at
u
r
e.
Z
C
R
ca
n
b
e
in
ter
p
r
eted
as
a
m
ea
s
u
r
e
o
f
t
h
e
n
o
is
i
n
e
s
s
o
f
a
s
ig
n
a
l
; it
u
s
u
all
y
r
et
u
r
n
s
h
i
g
h
er
v
alu
e
s
in
t
h
e
ca
s
e
o
f
a
n
o
is
y
s
ig
n
a
l [
1
8
]
.
2
.
4
.
Dis
cr
et
e
w
a
v
ele
t
t
ra
ns
f
o
r
m
(
D
WT
)
T
h
e
v
o
ice
s
ig
n
al
is
d
ec
o
m
p
o
s
ed
in
to
N
lev
els
u
s
in
g
DW
T
,
w
h
er
e
N
m
u
s
t
b
e
a
s
tr
ictly
p
o
s
iti
v
e
in
te
g
er
ch
o
s
en
to
b
e
f
iv
e
lev
e
ls
in
th
i
s
p
ap
er
.
I
n
th
e
f
ir
s
t
s
tep
o
f
th
e
DW
T
-
b
ased
an
al
y
s
i
s
,
th
e
DW
T
o
f
th
e
v
o
ice
s
i
g
n
a
l
(
)
p
r
o
d
u
ce
s
t
w
o
s
e
ts
o
f
co
ef
f
ic
ien
t
s
:
ap
p
r
o
x
i
m
a
t
io
n
co
ef
f
icie
n
ts
c
A
1
,
a
n
d
d
etail
co
ef
f
icien
ts
cD1
.
T
h
ese
v
ec
to
r
s
ar
e
o
b
tain
ed
b
y
co
n
v
o
lv
in
g
th
e
s
i
g
n
al
s
w
it
h
th
e
lo
w
-
p
a
s
s
f
ilter
L
o
_
D
f
o
r
ap
p
r
o
x
i
m
atio
n
,
an
d
w
it
h
t
h
e
h
i
g
h
-
p
ass
f
ilter
Hi_
D
f
o
r
d
etail,
f
o
llo
w
ed
b
y
d
y
ad
ic
d
ec
i
m
atio
n
(
d
o
w
n
s
a
m
p
lin
g
)
a
s
s
h
o
w
n
i
n
th
e
b
lo
ck
d
iag
r
a
m
i
n
Fi
g
u
r
e
6
,
w
h
er
e
t
h
e
len
g
t
h
o
f
ea
ch
f
ilt
er
is
eq
u
a
l
to
2
N
[
1
9
]
,
[
2
0
]
.
T
h
e
n
ex
t
s
tep
s
p
lit
s
th
e
ap
p
r
o
x
i
m
atio
n
co
e
f
f
icie
n
t
s
cA
1
i
n
to
t
w
o
p
ar
ts
f
o
llo
w
i
n
g
s
a
m
e
s
c
h
e
m
e
i
n
t
h
e
f
ir
s
t
s
te
p
b
y
r
ep
lacin
g
s
b
y
cA
1
,
an
d
h
en
ce
p
r
o
d
u
cin
g
cA
2
an
d
cD
2
; a
n
d
s
tep
s
co
n
tin
u
e
as
s
u
c
h
N
ti
m
es.
Fo
llo
w
i
n
g
th
e
s
e
s
tep
s
,
th
e
w
a
v
elet
d
ec
o
m
p
o
s
itio
n
o
f
th
e
v
o
ice
s
i
g
n
al
s
(
t)
(
an
al
y
ze
d
at
le
v
el
N
=
5
)
r
esu
lts
in
th
e
s
tr
u
c
tu
r
e:
[
c
A
5
,
cD5
,
.
.
.
,
cD1
]
as
s
h
o
w
n
in
F
ig
u
r
e
7
.
Fig
u
r
e
6
.
T
h
e
f
ir
s
t step
o
f
DW
T
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
C
la
s
s
i
fica
tio
n
o
f th
r
ee
p
a
th
o
lo
g
ica
l v
o
ices b
a
s
ed
o
n
…
(
Mu
n
ee
r
a
A
lta
ye
b
)
953
Fig
u
r
e
7
.
T
h
e
g
en
er
al
s
tr
u
ct
u
r
e
o
f
DW
T
o
f
5
lev
els
3.
CL
AS
SI
F
I
CAT
I
O
N
RE
SU
L
T
S
Su
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
es
(
SV
Ms)
ar
e
s
tate
-
of
-
t
h
e
-
ar
t
class
i
f
ier
s
,
SVM
ta
k
es
a
s
et
o
f
i
n
p
u
t
d
ata
an
d
p
r
ed
icts
,
f
o
r
ea
ch
g
iv
e
n
i
n
p
u
t,
w
h
ich
o
f
t
w
o
p
o
s
s
ib
le
cl
ass
es
f
o
r
m
s
t
h
e
o
u
tp
u
t,
ac
co
r
d
in
g
to
t
h
e
SVM
m
et
h
o
d
o
lo
g
y
,
a
k
er
n
el
f
u
n
ctio
n
is
u
s
ed
in
o
r
d
er
t
o
m
ap
th
e
f
ea
tu
r
e
v
ec
to
r
s
to
th
e
‘
k
er
n
el
s
p
ac
e’
.
I
n
th
is
w
o
r
k
,
10
-
f
o
ld
c
r
o
s
s
-
v
al
id
atio
n
o
n
t
h
e
tr
ain
i
n
g
d
ata
w
er
e
u
s
ed
to
c
r
ea
te
th
e
m
o
d
el.
1
0
s
a
m
p
le
s
f
r
o
m
ea
c
h
ca
s
e
ar
e
k
ep
t
f
o
r
test
in
g
w
h
ile
th
e
r
e
m
ai
n
in
g
s
a
m
p
les
w
er
e
u
s
ed
in
tr
ain
i
n
g
.
T
h
e
tar
g
et
v
ar
iab
le
co
r
r
esp
o
n
d
s
to
a
d
ec
is
io
n
th
a
t in
p
u
t d
ata
x
b
elo
n
g
s
to
n
o
r
m
al
v
o
ice
(
c
lass
0
)
o
r
ab
n
o
r
m
al
v
o
ice
(
class
1
)
[
2
1
]
-
[
2
3
]
.
T
h
e
class
i
f
icatio
n
p
r
o
ce
s
s
w
a
s
p
er
f
o
r
m
ed
s
ix
ti
m
es
u
s
in
g
th
e
f
o
llo
w
i
n
g
f
ea
t
u
r
e
g
r
o
u
p
s
:
Gr
o
u
p
1
in
cl
u
d
es
d
elta
-
MF
C
C
f
ea
t
u
r
es
alo
n
e,
g
r
o
u
p
2
h
as
d
elta
-
d
elt
a
MFC
C
a
n
d
Ku
r
to
s
i
s
f
ea
t
u
r
es,
g
r
o
u
p
3
i
n
clu
d
e
s
d
elta
-
d
el
ta
MF
C
C
a
n
d
s
k
e
w
n
ess
f
ea
tu
r
es,
w
h
ile
g
r
o
u
p
4
h
as
d
elta
-
d
elta
MF
C
C
a
n
d
Z
C
R
f
ea
t
u
r
es
to
g
e
th
er
.
A
l
s
o
,
g
r
o
u
p
5
h
as
th
e
d
elta
-
d
e
lta
MFC
C
p
lu
s
t
h
e
en
tr
o
p
y
f
ea
tu
r
es.
T
h
e
f
i
v
e
DW
T
f
ea
tu
r
es
ar
e
u
s
ed
as
g
r
o
u
p
6
in
t
h
e
la
s
t
cla
s
s
i
f
icatio
n
p
r
o
ce
s
s
.
Sect
io
n
6
p
r
o
v
id
es
d
et
ails
ab
o
u
t
th
e
f
ea
t
u
r
e
g
r
o
u
p
s
an
d
t
h
e
r
es
u
lta
n
t
class
i
f
icatio
n
ac
cu
r
ac
ies [
2
1
]
,
[
2
2
]
.
3
.
1
.
P
er
f
o
rm
a
nce
ev
a
lua
t
io
n
T
h
e
to
tal
n
u
m
b
er
s
o
f
s
p
ee
ch
s
a
m
p
les
u
s
ed
in
th
is
w
o
r
k
ar
e
2
0
2
f
o
r
th
e
e
v
alu
a
tio
n
p
u
r
p
o
s
e
o
f
w
h
ic
h
1
4
8
ar
e
p
ath
o
lo
g
y
w
h
ile
5
4
a
r
e
n
o
r
m
al
v
o
ices.
T
h
e
ter
m
s
u
s
ed
in
th
e
co
n
f
u
s
io
n
m
atr
i
x
a
s
s
h
o
w
n
in
T
ab
le
2
ca
n
b
r
ief
l
y
b
e
d
escr
ib
ed
as:
tr
u
e
p
o
s
itiv
e
(
TP
)
:
tr
u
e
d
ec
is
iv
e
s
y
s
te
m
clas
s
i
f
ied
as
tr
u
e
;
tr
u
e
n
e
g
ati
v
e
(
TN
)
:
f
alse
ev
e
n
t
d
etec
ted
as
f
al
s
e
;
f
al
s
e
p
o
s
iti
v
e
(
FP
)
:
t
h
e
e
v
e
n
t
i
s
f
alse
an
d
d
is
c
r
i
m
in
ated
as
tr
u
e
;
a
n
d
f
alse
n
eg
at
iv
e
(
FN
)
: tr
u
e
ev
e
n
t c
las
s
if
ied
as
f
alse [
2
4
]
-
[
2
6
]
.
T
ab
le
2
.
C
o
n
f
u
s
io
n
m
atr
i
x
C
o
n
f
u
s
i
o
n
ma
t
r
i
x
N
o
r
mal
P
a
t
h
o
l
o
g
y
N
o
r
mal
TP
FP
P
a
t
h
o
l
o
g
y
FN
TN
A
l
s
o
,
ac
cu
r
ac
y
(
AC
)
is
d
ef
i
n
e
d
as th
e
p
r
o
b
ab
ilit
y
t
h
at
t
h
e
cl
ass
i
f
icatio
n
b
y
th
e
s
y
s
te
m
is
c
o
r
r
ec
t a
n
d
it is
g
i
v
en
b
y
(
1
0
)
[
2
0
]
:
=
+
(
+
+
+
)
∗
100
(
1
0
)
T
h
e
s
en
s
it
iv
i
t
y
(
tr
u
e
p
o
s
itiv
e
r
ate
(
T
P
R
)
)
an
d
s
p
ec
if
icit
y
(
tr
u
e
n
e
g
ati
v
e
r
ate
(
T
NR
)
)
ar
e
a
ls
o
ca
lcu
lated
f
r
o
m
th
e
co
n
f
u
s
io
n
m
atr
i
x
u
s
in
g
(
11
)
,
an
d
(
12
)
r
esp
ec
tiv
el
y
[
2
0
]
:
T
P
R
=
TP
TP
+
FN
(
1
1
)
T
NR
=
TN
TP
+
FN
(
1
2
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
1
2
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
2
:
9
4
6
-
956
954
3
.
2
.
Cla
s
s
if
ica
t
io
n us
ing
delt
a
,
delt
a
-
delt
a
M
F
CC,
Z
CR
a
nd
o
t
her
s
t
a
t
is
t
ica
l f
ea
t
ures
Her
e,
d
elta
-
MFC
C
,
d
elta
-
d
elt
a
MFC
C
,
Z
C
R
a
n
d
o
th
er
s
ta
tis
tical
f
ea
t
u
r
es
ar
e
u
s
ed
to
cr
ea
te
f
iv
e
f
ea
t
u
r
e
g
r
o
u
p
s
w
h
ich
ar
e
n
a
m
ed
b
y
F1
:
Delta
M
FC
C
f
ea
t
u
r
es,
F2
:
R
e
lated
to
d
elta
-
d
elta
MFC
C
an
d
k
u
r
to
s
is
f
ea
t
u
r
es,
F3
:
Delta
-
d
elta
MF
C
C
a
n
d
s
k
e
w
n
ess
f
ea
tu
r
es,
F4
:
Delta
-
d
elta
MF
C
C
w
it
h
Z
C
R
f
ea
t
u
r
es,
w
h
ile
F5
:
Delta
-
d
elta
MFC
C
p
l
u
s
e
n
t
r
o
p
y
f
ea
t
u
r
es.
W
h
er
e
t
h
e
n
u
m
b
er
o
f
n
o
r
m
a
l
d
ata=
5
4
,
h
y
p
er
k
in
et
ic=
6
4
,
h
y
p
o
k
in
etic=
4
5
,
r
ef
l
u
x
=3
8
s
am
p
les.
I
n
ea
c
h
ca
s
e
1
0
s
a
m
p
l
es
ar
e
k
ep
t
f
o
r
test
i
n
g
an
d
th
e
r
em
a
in
i
n
g
s
a
m
p
le
s
ar
e
u
s
ed
in
tr
ai
n
i
n
g
.
T
ab
les 3
,
4
an
d
5
s
h
o
w
th
e
r
e
s
u
l
t o
f
t
h
e
class
i
f
icatio
n
p
r
o
ce
s
s
in
c
lu
d
i
n
g
tr
ain
in
g
-
,
test
i
n
g
-
ac
cu
r
ac
y
,
T
NR
an
d
T
P
R
u
s
in
g
ea
c
h
f
ea
t
u
r
e
g
r
o
u
p
an
d
r
ep
ea
ted
f
o
r
th
e
th
r
ee
d
if
f
er
en
t
p
ath
o
lo
g
ical
ca
s
es
v
er
s
u
s
t
h
e
n
o
r
m
al
v
o
ice
s
i
g
n
al
.
T
h
e
v
o
ice
clas
s
i
f
icatio
n
s
r
es
u
l
ts
ar
e
s
h
o
w
n
i
n
T
ab
les
3
,
4
an
d
5
w
h
er
e
t
h
e
co
n
f
u
s
io
n
m
atr
i
x
i
s
u
s
ed
to
en
v
i
s
io
n
th
e
p
er
f
o
r
m
an
ce
.
T
h
e
m
a
x
i
m
u
m
r
es
u
lts
o
f
t
h
e
m
o
d
el
-
,
te
s
t
-
ac
cu
r
ac
y
,
T
NR
a
n
d
T
P
R
w
er
e
f
o
u
n
d
u
s
i
n
g
th
e
f
ea
t
u
r
es
g
r
o
u
p
3
(
F3
)
,
as
th
e
y
r
ea
ch
ed
1
0
0
%
in
t
h
e
class
i
f
icatio
n
o
f
all
p
at
h
o
lo
g
ical
ca
s
es.
An
o
th
er
g
o
o
d
ac
cu
r
ac
y
w
as
f
o
u
n
d
w
h
e
n
cla
s
s
i
f
y
in
g
th
e
d
ata
as
H
y
p
e
r
k
in
et
ic
o
r
n
o
r
m
al
v
o
ices
u
s
in
g
f
ea
t
u
r
es
g
r
o
u
p
4
(
F4
)
in
T
ab
le
4
,
also
class
if
y
i
n
g
th
e
d
ata
a
s
H
y
p
o
k
i
n
etic
o
r
n
o
r
m
al
u
s
i
n
g
f
ea
tu
r
e
s
g
r
o
u
p
1
(
F1
)
as
s
ee
n
i
n
T
ab
le
4
.
Ver
y
g
o
o
d
r
esu
lts
w
e
r
e
p
er
f
o
r
m
ed
u
s
in
g
t
h
e
f
ir
s
t
f
e
atu
r
e
g
r
o
u
p
(
F1
)
an
d
f
i
f
t
h
g
r
o
u
p
(
F5
)
in
th
e
ca
s
e
o
f
clas
s
if
y
i
n
g
t
h
e
d
ata
as
R
e
f
l
u
x
o
r
n
o
r
m
al
d
ata
a
s
s
h
o
w
n
i
o
n
T
ab
le
5
.
Fro
m
t
h
e
tab
le,
it
ca
n
b
e
n
o
ticed
t
h
at
s
o
m
e
o
f
t
h
e
f
ea
tu
r
e
s
ar
e
co
n
s
u
m
in
g
l
o
w
er
ac
c
u
r
ac
y
th
a
n
o
t
h
er
s
,
f
o
r
ex
a
m
p
le
t
h
e
f
ea
t
u
r
es
g
r
o
u
p
4
(
F4
)
g
i
v
e
s
test
ac
cu
r
ac
y
less
t
h
a
n
o
r
eq
u
al
to
5
0
% in
m
o
s
t o
f
t
h
e
ca
s
e
s
.
I
n
o
r
d
er
to
r
ev
ea
l
th
e
b
est
f
ea
tu
r
e
co
m
b
i
n
atio
n
s
an
d
to
o
b
tain
t
h
e
h
i
g
h
est
ac
c
u
r
ac
y
i
n
cla
s
s
if
ica
tio
n
,
T
ab
le
6
s
h
o
w
s
t
h
e
r
esu
lt
u
s
i
n
g
a
co
m
b
in
at
io
n
o
f
t
h
e
b
est
th
r
ee
f
ea
tu
r
e
g
r
o
u
p
s
at
ea
ch
p
ath
o
lo
g
ical
ca
s
e
f
o
u
n
d
in
T
ab
le
3
.
T
h
e
f
ir
s
t
co
lu
m
n
in
T
ab
le
6
s
h
o
w
s
th
e
r
es
u
lts
o
f
class
i
f
y
in
g
t
h
e
d
ata
as
n
o
r
m
al
o
r
h
y
p
er
k
i
n
etic
u
s
i
n
g
d
elta
-
d
elta
M
FC
C
,
s
k
e
w
n
es
s
,
an
d
Z
C
R
f
ea
tu
r
e
s
to
g
eth
er
.
I
t
is
f
o
u
n
d
t
h
at
t
h
e
tes
t
-
,
tr
ain
-
ac
c
u
r
ac
ies,
T
NR
an
d
T
P
R
ar
e
all
1
0
0
%
ac
cu
r
ac
y
.
T
h
e
s
a
m
e
r
es
u
lts
ar
e
f
o
u
n
d
w
h
e
n
cla
s
s
i
f
y
in
g
t
h
e
d
a
ta
as
h
y
p
o
k
i
n
etic
o
r
n
o
r
m
al
u
s
i
n
g
d
elta
-
d
elta
MF
C
C
,
s
k
e
w
n
es
s
an
d
d
elta
-
MF
C
C
to
g
eth
er
.
T
h
e
d
elta
-
d
elta
MFC
C
,
en
tr
o
p
y
an
d
d
elta
-
MF
C
C
f
ea
t
u
r
es
co
m
b
i
n
atio
n
ar
e
u
s
ed
in
t
h
e
t
h
ir
d
p
ath
o
lo
g
ical
ca
s
e
(
R
ef
l
u
x
)
an
d
t
h
e
r
esu
lt
ac
cu
r
ac
ies
w
er
e
v
er
y
g
o
o
d
.
T
ab
le
3
.
Featu
r
e
co
m
b
i
n
atio
n
s
an
d
ac
cu
r
ac
y
o
b
tain
ed
f
o
r
h
y
p
er
k
in
etic
p
at
h
o
lo
g
y
v
s
n
o
r
m
a
l c
ases
N
o
r
mal
F
1
H
y
p
e
r
F
1
N
o
r
mal
F
2
H
y
p
e
r
F
2
N
o
r
mal
F
3
H
y
p
e
r
F
3
N
o
r
mal
F
4
H
y
p
e
r
F
4
N
o
r
mal
F
5
H
y
p
e
r
F
5
T
r
a
i
n
A
c
c
u
r
a
c
y
5
8
%
6
1
%
1
0
0
%
6
7
%
5
5
%
T
e
st
A
c
c
u
r
a
c
y
5
0
%
6
5
%
1
0
0
%
7
0
%
5
0
%
T
N
R
5
7
.
8
9
%
5
8
.
3
3
%
1
0
0
%
6
4
.
2
8
%
0
T
P
R
5
8
.
0
3
%
6
2
.
5
%
1
0
0
%
6
8
.
9
7
%
5
5
%
T
ab
le
4
.
Featu
r
e
co
m
b
i
n
atio
n
s
an
d
ac
cu
r
ac
y
o
b
tain
ed
f
o
r
h
y
p
o
k
in
etic
p
ath
o
lo
g
y
v
s
n
o
r
m
al
ca
s
es
N
o
r
mal
F
1
H
y
p
o
F
1
N
o
r
mal
F
2
H
y
p
o
F
2
N
o
r
mal
F
3
H
y
p
o
F
3
N
o
r
mal
F
4
H
y
p
o
F
4
N
o
r
mal
F
5
H
y
p
o
F
5
T
r
a
i
n
A
c
c
u
r
a
c
y
7
0
.
3
7
%
7
0
.
3
7
%
1
0
0
%
6
4
.
1
9
%
5
8
.
0
2
4
%
T
e
st
A
c
c
u
r
a
c
y
8
0
%
5
5
%
1
0
0
%
5
0
%
4
0
%
T
N
R
7
4
.
4
2
%
6
8
.
4
2
%
1
0
0
%
8
6
.
8
6
5
9
.
0
1
%
T
P
R
6
5
.
7
8
%
7
5
%
1
0
0
%
5
9
.
4
5
%
5
5
%
T
ab
le
5
.
Featu
r
e
co
m
b
i
n
atio
n
s
an
d
ac
cu
r
ac
y
o
b
tain
ed
f
o
r
r
ef
l
u
x
p
at
h
o
lo
g
y
v
s
n
o
r
m
al
ca
s
es
N
o
r
mal
F
1
R
e
f
l
u
x
F
1
N
o
r
mal
F
2
R
e
f
l
u
x
F
2
N
o
r
mal
F
3
R
e
f
l
u
x
F
3
N
o
r
mal
F
4
R
e
f
l
u
x
F
4
N
o
r
mal
F
5
R
e
f
l
u
x
F
5
T
r
a
i
n
A
c
c
u
r
a
c
y
9
8
.
6
5
%
6
6
.
2
2
%
1
0
0
%
6
7
.
5
7
%
7
9
.
7
3
%
T
e
st
A
c
c
u
r
a
c
y
1
0
0
%
5
0
%
1
0
0
%
4
5
%
8
0
%
T
N
R
9
7
.
8
2
%
6
8
.
5
2
%
1
0
0
%
6
9
.
0
9
%
8
1
.
2
5
%
T
P
R
1
0
0
%
6
0
%
1
0
0
%
6
3
.
1
6
%
7
6
.
9
2
%
T
ab
le
6
.
B
est f
ea
tu
r
e
co
m
b
i
n
a
tio
n
s
a
n
d
ac
cu
r
ac
y
o
b
tain
ed
N
o
r
mal
/
H
y
p
e
r
∆∆
M
F
C
C
+
S
k
e
w
n
e
ss+
Z
C
R
N
o
r
mal
/
H
y
p
o
∆∆
M
F
C
C
+
S
k
e
w
n
e
ss+
∆M
F
C
C
N
o
r
mal
/
R
e
f
l
u
x
∆∆
M
F
C
C
+
En
t
r
o
p
y
+
∆M
F
C
C
T
r
a
i
n
A
c
c
u
r
a
c
y
1
0
0
%
1
0
0
%
9
7
.
3
%
T
e
st
A
c
c
u
r
a
c
y
1
0
0
%
1
0
0
%
1
0
0
%
T
N
R
1
0
0
%
1
0
0
%
1
0
0
%
T
P
R
1
0
0
%
1
0
0
%
9
3
.
5
4
%
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
C
la
s
s
i
fica
tio
n
o
f th
r
ee
p
a
th
o
lo
g
ica
l v
o
ices b
a
s
ed
o
n
…
(
Mu
n
ee
r
a
A
lta
ye
b
)
955
3
.
3
.
Cla
s
s
if
ica
t
io
n us
ing
DWT
f
ea
t
ures
Dis
cr
ete
w
a
v
elet
tr
an
s
f
o
r
m
(
DW
T
)
an
aly
s
i
s
is
u
s
ed
to
e
x
tr
ac
t
5
f
ea
t
u
r
es;
o
n
e
en
er
g
y
f
ea
t
u
r
e
i
s
ex
tr
ac
ted
f
r
o
m
ea
c
h
o
n
e
o
f
t
h
e
f
i
v
e
w
a
v
elet
le
v
el
s
.
T
ab
le
7
p
r
esen
ts
th
e
ac
cu
r
ac
y
o
b
t
ain
ed
b
y
W
a
v
ele
t
f
ea
t
u
r
es.
T
h
e
T
ab
le
7
s
h
o
w
s
t
h
at
t
h
e
clas
s
if
icatio
n
tr
ai
n
ac
c
u
r
ac
y
(
m
o
d
el
ac
c
u
r
ac
y
)
ar
e
r
an
g
ed
b
et
w
ee
n
6
0
%
an
d
7
0
%
in
all
p
at
h
o
lo
g
ical
c
ases
,
w
h
ile
te
s
t
ac
c
u
r
ac
ies
ar
e
r
an
g
ed
b
et
w
ee
n
8
0
%
a
n
d
9
0
%
w
h
ich
ar
e
g
o
o
d
r
esu
lt
s
I
in
co
m
p
ar
is
o
n
w
it
h
s
o
m
e
o
f
t
h
e
r
es
u
lts
s
ee
n
i
n
T
ab
le
3
.
T
o
th
e
b
est
o
f
o
u
r
k
n
o
w
led
g
e,
t
h
e
DW
T
f
ea
t
u
r
es
ar
e
n
o
t
w
id
el
y
u
s
ed
in
v
o
ice
clas
s
i
f
icatio
n
f
ield
,
b
u
t
in
th
i
s
w
o
r
k
,
th
e
r
es
u
lt
s
s
h
o
w
t
h
at
th
e
y
ar
e
b
etter
th
an
s
o
m
e
o
f
t
h
e
M
FC
C
a
n
d
Z
C
R
f
ea
t
u
r
es
w
h
ic
h
ar
e
w
id
el
y
u
s
ed
i
n
th
is
ar
ea
.
T
h
e
W
av
e
let
f
ea
t
u
r
es
p
r
ese
n
t
g
o
o
d
r
esu
lts
a
n
d
co
u
ld
b
e
u
s
e
d
in
class
i
f
y
i
n
g
t
h
e
v
o
ice
as
n
o
r
m
al
o
r
p
ath
o
lo
g
ical
v
o
ices
w
it
h
g
o
o
d
ac
cu
r
ac
y
.
T
ab
le
7
.
W
av
elet
f
ea
tu
r
es a
n
d
ac
cu
r
ac
y
o
b
tain
ed
W
a
v
e
l
e
t
f
e
a
t
u
r
e
s
H
y
p
e
r
k
i
n
e
t
i
c
H
y
p
o
k
i
n
e
t
i
c
R
e
f
l
u
x
T
r
a
i
n
A
c
c
u
r
a
c
y
7
0
%
6
0
.
4
9
%
6
7
.
5
7
%
T
e
st
A
c
c
u
r
a
c
y
9
0
%
8
0
%
9
0
%
T
N
R
7
4
.
1
9
%
6
1
.
0
2
%
6
6
.
6
7
%
T
P
R
6
8
.
1
2
%
5
9
.
0
9
%
7
2
.
7
3
%
4.
DIS
CU
SS
I
O
N
C
las
s
i
f
icatio
n
o
f
p
ath
o
lo
g
ica
l
v
o
ices
u
s
i
n
g
m
ac
h
i
n
e
lear
n
i
n
g
h
a
s
s
i
g
n
i
f
ica
n
t
b
en
e
f
its
f
o
r
p
atien
t
ass
es
s
m
en
t
an
d
i
m
p
r
o
v
e
m
e
n
t
co
m
p
u
ter
-
a
id
ed
s
y
s
te
m
s
,
a
s
t
h
er
e
ar
e
m
a
n
y
p
r
ev
io
u
s
r
esear
c
h
es
in
t
h
is
f
ie
ld
th
a
t
ap
p
ly
v
ar
io
u
s
m
et
h
o
d
s
o
f
f
ea
t
u
r
e
ex
tr
ac
tio
n
a
n
d
clas
s
if
icati
o
n
alg
o
r
it
h
m
s
.
I
n
th
is
p
ap
er
,
t
h
e
p
r
o
p
o
s
ed
m
e
th
o
d
f
o
r
d
etec
tin
g
a
n
d
class
if
y
i
n
g
v
o
ca
l
d
is
o
r
d
er
is
co
m
p
ar
ed
w
ith
t
h
e
m
et
h
o
d
s
f
o
u
n
d
i
n
p
r
ev
io
u
s
s
t
u
d
ies
[
4
]
-
[
6
]
,
[
2
5
]
w
h
ich
s
h
o
w
t
h
at
r
elate
d
v
o
ices
co
u
ld
b
e
class
i
f
ied
i
n
to
n
o
r
m
al/p
at
h
o
lo
g
ical
d
ep
en
d
s
o
n
s
o
u
n
d
s
f
ea
t
u
r
e
s
an
d
th
e
ac
cu
r
ac
y
o
f
t
h
e
class
if
icatio
n
alg
o
r
it
h
m
s
.
O
n
th
e
o
th
er
h
an
d
,
th
e
r
es
u
lt
ac
cu
r
ac
y
d
e
m
o
n
s
tr
ated
in
t
h
i
s
r
esear
ch
is
s
h
o
w
n
to
b
e
s
u
p
er
io
r
to
ea
r
lier
r
esear
ch
es.
T
h
e
a
cc
u
r
ac
y
r
ate
w
a
s
9
9
.
5
6
%
in
[
4
]
,
th
e
ac
cu
r
ac
y
r
ate
w
a
s
9
4
.
2
6
%
in
[
5
]
,
th
e
ac
cu
r
a
c
y
r
ate
w
as
b
et
w
ee
n
9
0
-
100
%
in
[
6
]
,
an
d
t
h
e
ac
c
u
r
ac
y
r
ate
w
a
s
9
7
.
9
%
i
n
[
2
]
.
Ho
w
e
v
er
,
in
th
i
s
s
t
u
d
y
,
th
e
ac
cu
r
ac
y
r
ate
n
o
t
o
n
l
y
i
n
cr
ea
s
ed
to
ar
o
u
n
d
1
0
0
%,
b
u
t
also
th
e
m
et
h
o
d
s
p
r
esen
ted
ar
e
ab
le
to
class
if
y
t
h
e
r
elate
d
v
o
ices
in
to
f
o
u
r
d
if
f
er
en
t
cl
ass
es
(
n
o
r
m
al,
h
y
p
er
k
i
n
etic,
h
y
p
o
k
i
n
etic,
r
ef
l
u
x
)
,
w
h
ic
h
is
i
m
p
o
r
tan
t i
n
v
o
ic
e
d
i
s
ea
s
es d
ia
g
n
o
s
tic
f
ie
ld
.
5.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
ex
p
lo
r
es
an
d
co
m
p
ar
e
s
s
ev
er
al
v
o
ice
f
ea
tu
r
e
s
ex
tr
ac
tio
n
m
eth
o
d
s
w
h
ich
ar
e
u
s
ed
to
class
i
f
y
th
e
v
o
ices
as
n
o
r
m
al
o
r
p
ath
o
lo
g
ical
v
o
ices.
T
h
r
ee
d
if
f
er
en
t
ab
n
o
r
m
al
c
ases
w
er
e
s
t
u
d
ied
:
h
y
p
er
k
in
et
ic,
h
y
p
o
k
i
n
etic,
an
d
r
ef
lu
x
.
T
h
r
ee
d
if
f
er
e
n
t
m
e
th
o
d
s
w
er
e
u
s
ed
to
ex
tr
ac
t
f
ea
tu
r
es
w
h
ic
h
ar
e:
MFC
C
,
Z
C
R
,
DW
T
an
d
a
r
e
lated
s
tati
s
tical
f
ea
t
u
r
e
ar
e
f
o
u
n
d
u
s
in
g
s
k
e
w
n
e
s
s
,
k
u
r
to
s
is
,
an
d
e
n
tr
o
p
y
.
T
h
e
p
u
r
p
o
s
e
o
f
th
is
w
o
r
k
is
to
class
i
f
y
th
e
v
o
ice
d
atase
t
an
d
co
m
p
ar
e
t
h
e
cla
s
s
i
f
icatio
n
r
es
u
lts
u
s
in
g
d
i
f
f
er
en
t
f
ea
t
u
r
e
g
r
o
u
p
s
w
h
er
e
t
h
e
cla
s
s
if
icatio
n
p
r
o
ce
s
s
w
as r
ep
ea
ted
.
T
h
e
class
i
f
icat
io
n
p
r
o
ce
s
s
e
s
w
er
e
al
l d
o
n
e
u
s
in
g
SVM
an
d
th
e
tr
ain
-
,
test
-
ac
cu
r
ac
ies,
T
NR
an
d
T
P
R
ar
e
c
alcu
lated
f
r
o
m
t
h
e
r
esu
lta
n
t
c
o
n
f
u
s
io
n
m
atr
i
x
i
n
ea
ch
ca
s
e.
T
h
e
b
est
class
if
ic
atio
n
r
esu
lt
s
w
er
e
r
ea
ch
ed
u
s
in
g
th
e
f
ea
t
u
r
e
g
r
o
u
p
th
at
i
n
clu
d
e
s
d
elta
-
d
elta
MFC
C
an
d
s
k
e
w
n
e
s
s
f
ea
tu
r
e
s
,
as
it
g
av
e
1
0
0
%
ac
cu
r
ac
y
in
all
ca
s
e
s
.
A
co
m
b
i
n
ati
o
n
o
f
s
o
m
e
o
f
t
h
e
d
elta
-
d
elta
M
FC
C
a
n
d
Z
C
R
f
e
atu
r
es
ar
e
also
g
a
v
e
v
er
y
g
o
o
d
ac
cu
r
ac
y
.
T
h
e
DW
T
f
ea
t
u
r
es
ar
e
n
o
t
co
m
m
o
n
l
y
u
s
ed
i
n
v
o
ice
c
lass
if
ica
tio
n
,
b
u
t
i
n
th
is
p
ap
er
,
t
h
e
r
es
u
lt
s
s
h
o
w
t
h
at
t
h
e
y
ar
e
b
etter
t
h
an
t
h
e
d
elta
-
d
elta
MF
C
C
an
d
Z
C
R
f
ea
tu
r
es
w
h
ich
ar
e
w
id
el
y
u
s
ed
i
n
t
h
is
ar
ea
a
n
d
ca
n
b
e
u
s
ed
to
class
if
y
t
h
e
v
o
ice
as
n
o
r
m
a
l
o
r
p
ath
o
lo
g
ical
w
it
h
g
o
o
d
ac
cu
r
ac
y
.
I
n
th
e
f
u
t
u
r
e
r
esear
ch
,
o
th
er
m
et
h
o
d
s
,
o
r
a
c
o
m
b
i
n
ati
o
n
o
f
class
i
f
icatio
n
m
et
h
o
d
s
th
a
n
S
VM
m
a
y
b
e
u
s
ed
to
en
h
a
n
ce
th
e
r
e
s
u
l
ts
w
h
er
e
lo
w
er
ac
cu
r
ac
ie
s
w
er
e
f
o
u
n
d
also
class
i
f
ica
tio
n
o
f
o
th
er
d
is
ea
s
es t
h
at
ca
u
s
e
te
m
p
o
r
a
r
y
v
o
ca
l i
m
p
air
m
e
n
ts
,
s
u
ch
a
s
C
OVI
D
-
19.
RE
F
E
R
E
NC
E
S
[
1
]
A
.
V
i
sav
e
,
P
.
K
a
c
h
a
r
e
,
A
.
Jey
a
k
u
m
a
r
,
A
.
N
.
C
h
e
e
r
a
n
,
a
n
d
G
.
B
a
c
h
h
e
r
,
“
V
o
c
a
l
f
e
a
t
u
r
e
s
f
o
r
g
l
o
t
t
a
l
p
a
t
h
o
l
o
g
y
d
e
t
e
c
t
i
o
n
u
s
i
n
g
B
P
N
N
,”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
C
o
m
p
u
t
e
r
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
1
8
,
n
o
.
1
7
,
p
p
.
1
-
6
,
2
0
1
5
,
d
o
i
:
1
0
.
5
1
2
0
/
2
0
8
3
4
-
3
5
7
1
.
[
2
]
V
.
S
e
l
l
a
m
a
n
d
J.
Jag
a
d
e
e
san
,
“
C
l
a
ssi
f
i
c
a
t
i
o
n
o
f
n
o
r
mal
a
n
d
p
a
t
h
o
l
o
g
i
c
a
l
v
o
i
c
e
u
si
n
g
S
V
M
a
n
d
R
B
F
N
N
,”
J
o
u
r
n
a
l
o
f
S
i
g
n
a
l
a
n
d
I
n
f
o
rm
a
t
i
o
n
Pr
o
c
e
ssi
n
g
,
v
o
l
.
5
,
n
o
.
1
,
2
0
1
4
,
d
o
i
:
1
0
.
4
2
3
6
/
j
s
i
p
.
2
0
1
4
.
5
1
0
0
1
.
[
3
]
D
.
P
r
a
v
e
n
a
,
S
.
D
h
i
v
y
a
,
a
n
d
A
.
D
.
D
e
v
i
,
“
P
a
t
h
o
l
o
g
i
c
a
l
v
o
i
c
e
r
e
c
o
g
n
i
t
i
o
n
f
o
r
v
o
c
a
l
f
o
l
d
d
i
se
a
se
,”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
C
o
m
p
u
t
e
r
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
4
7
,
n
o
.
1
3
,
p
p
.
3
1
-
3
7
,
2
0
1
2
,
d
o
i
:
1
0
.
5
1
2
0
/
7
2
5
0
-
0
3
1
4
.
[
4
]
H
.
A
n
k
ı
ş
h
a
n
,
“
A
n
e
w
a
p
p
r
o
a
c
h
f
o
r
d
e
t
e
c
t
i
o
n
o
f
p
a
t
h
o
l
o
g
i
c
a
l
v
o
i
c
e
d
i
so
r
d
e
r
s
w
i
t
h
r
e
d
u
c
e
d
p
a
r
a
me
t
e
r
s
,”
E
l
e
c
t
r
i
c
a
,
v
o
l
.
1
8
,
n
o
.
1
,
p
p
.
6
0
-
7
1
,
2
0
1
8
,
d
o
i
:
1
0
.
5
1
5
2
/
i
u
j
e
e
e
.
2
0
1
8
.
1
8
1
0
.
[
5
]
S
.
H
.
F
a
n
g
,
“
D
e
t
e
c
t
i
o
n
o
f
p
a
t
h
o
l
o
g
i
c
a
l
v
o
i
c
e
u
si
n
g
c
e
p
st
r
u
m
v
e
c
t
o
r
s:
A
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
,
”
J
o
u
r
n
a
l
o
f
Vo
i
c
e
,
v
o
l
.
3
3
,
n
o
.
5
,
p
p
.
6
3
4
-
6
4
1
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
v
o
i
c
e
.
2
0
1
8
.
0
2
.
0
0
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.