I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
15
,
No
.
4
,
A
u
g
u
s
t
20
25
,
p
p
.
3
7
3
7
~
3
7
4
7
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
4
.
pp
3
7
3
7
-
3
7
4
7
3737
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Integ
ra
ting time
-
f
requency
features
wit
h
deep
learni
ng
f
o
r
l
ung
so
und clas
sifica
ti
o
n
Su
Yua
n Cha
ng
1,
2
,
M
a
rni A
z
ira
M
a
rk
o
m
1,
3
,
Z
hi Sh
eng
Cho
o
ng
1,
4
,
Arni
M
un
ira
M
a
rk
o
m
5
,
L
a
t
if
a
h M
un
ira
h K
a
m
a
rudd
in
1,
3
,
E
rdy
Su
lin
o
M
o
hd
M
us
lim
T
a
n
3,
6
1
F
a
c
u
l
t
y
o
f
El
e
c
t
r
o
n
i
c
E
n
g
i
n
e
e
r
i
n
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
s
i
t
i
M
a
l
a
y
si
a
P
e
r
l
i
s,
K
a
m
p
u
s
P
a
u
h
P
u
t
r
a
,
A
r
a
u
P
e
r
l
i
s,
M
a
l
a
y
si
a
2
A
d
o
r
e
S
y
s
P
t
e
.
L
t
d
.
,
P
e
n
a
n
g
,
M
a
l
a
y
si
a
3
C
e
n
t
e
r
o
f
E
x
c
e
l
l
e
n
c
e
f
o
r
A
d
v
a
n
c
e
d
S
e
n
s
o
r
T
e
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
s
i
t
i
M
a
l
a
y
si
a
P
e
r
l
i
s,
K
a
m
p
u
s
P
a
u
h
P
u
t
r
a
,
A
r
a
u
P
e
r
l
i
s,
M
a
l
a
y
s
i
a
4
S
k
y
e
C
h
i
p
S
d
n
B
h
d
.
,
P
e
n
a
n
g
,
M
a
l
a
y
s
i
a
5
S
c
h
o
o
l
o
f
E
l
e
c
t
r
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
,
U
n
i
v
e
r
si
t
i
Te
k
n
o
l
o
g
i
M
A
R
A
,
S
e
l
a
n
g
o
r
,
M
a
l
a
y
si
a
6
F
a
c
u
l
t
y
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
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
si
t
i
M
a
l
a
y
si
a
P
e
r
l
i
s,
K
a
mp
u
s P
a
u
h
P
u
t
r
a
,
A
r
a
u
P
e
r
l
i
s
,
M
a
l
a
y
si
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Au
g
5
,
2
0
2
4
R
ev
is
ed
Ap
r
1
5
,
2
0
2
5
Acc
ep
ted
Ma
y
2
4
,
2
0
2
5
De
e
p
lea
rn
in
g
h
a
s
tran
sfo
rm
e
d
m
e
d
ica
l
d
iag
n
o
stics
,
e
sp
e
c
ially
i
n
a
n
a
ly
z
i
n
g
lu
n
g
so
u
n
d
s
t
o
a
ss
e
ss
re
sp
irat
o
ry
c
o
n
d
it
io
n
s.
Trad
it
i
o
n
a
l
m
e
t
h
o
d
s
l
ik
e
c
o
m
p
u
ted
t
o
m
o
g
ra
p
h
y
sc
a
n
s
(CT
sc
a
n
s
)
a
n
d
X
-
ra
y
s
a
re
imp
ra
c
ti
c
a
l
in
re
so
u
rc
e
-
li
m
it
e
d
se
tt
i
n
g
s
d
u
e
to
ra
d
iatio
n
e
x
p
o
su
re
a
n
d
ti
m
e
c
o
n
su
m
p
ti
o
n
,
wh
il
e
c
o
n
v
e
n
ti
o
n
a
l
ste
t
h
o
sc
o
p
e
s
o
ften
lea
d
t
o
m
is
d
iag
n
o
sis
d
u
e
to
su
b
jec
ti
v
e
in
terp
re
tatio
n
a
n
d
e
n
v
iro
n
m
e
n
tal
n
o
ise
.
T
h
is
stu
d
y
e
v
a
l
u
a
tes
d
e
e
p
lea
rn
in
g
m
o
d
e
ls
f
o
r
l
u
n
g
so
u
n
d
c
las
sifica
ti
o
n
u
si
n
g
t
h
e
I
n
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
Bio
m
e
d
ica
l
He
a
lt
h
In
f
o
rm
a
ti
c
s
2
0
1
7
d
a
tas
e
t,
c
o
m
p
risin
g
9
2
0
a
n
n
o
tate
d
sa
m
p
les
fro
m
1
2
6
su
b
jec
ts.
P
re
-
p
ro
c
e
ss
in
g
i
n
c
lu
d
e
s
d
o
wn
sa
m
p
li
n
g
,
se
g
m
e
n
tatio
n
,
n
o
rm
a
li
z
a
ti
o
n
,
a
n
d
a
u
d
i
o
c
li
p
p
i
n
g
,
with
fe
a
tu
re
e
x
trac
ti
o
n
tec
h
n
iq
u
e
s
li
k
e
sp
e
c
tro
g
ra
m
a
n
d
M
e
l
-
fre
q
u
e
n
c
y
c
e
p
stra
l
c
o
e
fficie
n
ts
(M
F
CC).
Th
e
a
d
o
p
te
d
a
u
to
m
a
ti
c
lu
n
g
so
u
n
d
d
iag
n
o
sis
n
e
tw
o
rk
(A
S
LD
-
Ne
t)
m
o
d
e
l
with
tr
ip
le
fe
a
tu
re
in
p
u
t
(ti
m
e
d
o
m
a
in
,
sp
e
c
tro
g
ra
m
,
a
n
d
M
F
CC)
a
c
h
iev
e
d
t
h
e
h
ig
h
e
st
a
c
c
u
ra
c
y
a
t
9
7
.
2
5
%
,
fo
ll
o
we
d
b
y
t
h
e
d
u
a
l
fe
a
tu
re
m
o
d
e
l
(sp
e
c
tro
g
ra
m
a
n
d
M
F
CC)
a
t
9
5
.
6
5
%
.
S
in
g
le
-
in
p
u
t
m
o
d
e
ls
wit
h
sp
e
c
tro
g
ra
m
a
n
d
M
F
CC
p
e
rfo
rm
e
d
we
ll
,
w
h
il
e
th
e
ti
m
e
d
o
m
a
in
in
p
u
t
a
l
o
n
e
h
a
d
th
e
lo
we
st ac
c
u
ra
c
y
.
K
ey
w
o
r
d
s
:
Dee
p
lear
n
in
g
L
u
n
g
s
o
u
n
d
L
u
n
g
s
o
u
n
d
class
if
icatio
n
Me
d
ical
s
ig
n
al
p
r
o
ce
s
s
in
g
T
im
e
-
f
r
eq
u
en
cy
f
ea
tu
r
e
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
:
Ma
r
n
i A
zir
a
Ma
r
k
o
m
B
i
o
m
e
d
i
ca
l
E
l
ec
t
r
o
n
i
c
E
n
g
i
n
e
er
i
n
g
P
r
o
g
r
a
m
m
e
,
D
e
p
a
r
t
m
e
n
t
o
f
E
l
e
c
t
r
o
n
i
c
E
n
g
i
n
e
e
r
i
n
g
,
F
a
c
u
lt
y
o
f
E
l
e
c
t
r
o
n
i
c
E
n
g
i
n
e
e
r
i
n
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
s
it
i
M
al
a
y
s
i
a
P
e
r
l
is
Kam
p
u
s
Pau
h
Pu
tr
a
,
0
2
6
0
0
Ar
au
Per
lis
,
Ma
lay
s
ia
E
m
ail: m
ar
n
i@
u
n
im
ap
.
ed
u
.
m
y
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
n
ec
ess
ity
f
o
r
en
h
a
n
ce
d
d
iag
n
o
s
tics
is
u
n
d
er
s
co
r
ed
b
y
th
e
f
ac
t
th
at
lu
n
g
d
is
ea
s
es,
wh
ich
ar
e
f
u
r
th
er
e
x
ac
er
b
ate
d
b
y
C
OVI
D
-
1
9
,
co
n
tin
u
e
to
b
e
th
e
p
r
im
ar
y
ca
u
s
e
o
f
d
ea
th
o
n
a
g
lo
b
al
s
ca
le
[
1
]
,
[
2
]
.
Ov
er
1
4
%
o
f
f
atalities
wo
r
ld
wid
e
ar
e
ca
u
s
ed
b
y
p
u
lm
o
n
ar
y
d
is
o
r
d
er
s
,
in
clu
d
in
g
asth
m
a,
ch
r
o
n
ic
o
b
s
tr
u
ctiv
e
p
u
lm
o
n
a
r
y
d
is
ea
s
e
(
C
OPD)
,
p
n
eu
m
o
n
ia,
lu
n
g
ca
n
ce
r
,
an
d
tu
b
er
cu
lo
s
is
[
3
]
.
Alth
o
u
g
h
tr
ad
it
io
n
al
to
o
ls
lik
e
C
T
s
ca
n
s
an
d
X
-
r
ay
s
ar
e
u
s
ef
u
l,
th
ey
ar
e
o
f
ten
tim
e
-
c
o
n
s
u
m
in
g
,
e
n
tail
r
ad
iatio
n
e
x
p
o
s
u
r
e,
an
d
m
ay
b
e
in
ac
ce
s
s
ib
le
in
r
eso
u
r
ce
-
lim
ited
s
ettin
g
s
[
4
]
.
C
o
n
ce
r
n
s
r
e
g
ar
d
in
g
d
iag
n
o
s
tic
ac
cu
r
ac
y
ar
e
r
aised
b
y
th
e
im
p
r
ec
is
e
r
esu
lts
th
at
co
n
v
en
tio
n
al
s
teth
o
s
co
p
es
f
r
eq
u
en
tly
p
r
o
d
u
c
e
as
a
r
esu
lt
o
f
s
u
b
jectiv
e
in
ter
p
r
etatio
n
an
d
en
v
ir
o
n
m
en
tal
n
o
is
e
[
5
]
.
Fo
r
im
p
r
o
v
ed
d
ia
g
n
o
s
tic
ac
cu
r
ac
y
,
d
ig
ital
s
teth
o
s
co
p
es
with
d
ee
p
lear
n
in
g
alg
o
r
ith
m
s
co
n
v
er
t
au
d
io
s
ig
n
als
to
d
ig
ital
f
o
r
m
at,
to
p
r
o
v
id
e
n
o
n
-
in
v
asiv
e,
r
ea
l
-
tim
e
m
o
n
ito
r
in
g
[
6
]
,
[
7
]
.
B
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
4
,
Au
g
u
s
t
20
25
:
3
7
3
7
-
3747
3738
p
r
ec
is
ely
ex
tr
ac
tin
g
an
d
c
ateg
o
r
izin
g
p
u
lm
o
n
ar
y
s
o
u
n
d
s
,
th
ey
en
ab
le
r
em
o
te
p
atien
t
m
o
n
ito
r
in
g
an
d
m
e
d
ical
ed
u
ca
tio
n
[
7
]
.
T
h
is
s
tu
d
y
ev
al
u
ates th
e
e
f
f
icac
y
in
th
e
class
i
f
icatio
n
o
f
lu
n
g
s
o
u
n
d
s
u
s
in
g
d
ee
p
lear
n
in
g
with
a
v
ar
iety
o
f
f
ea
tu
r
e
i
n
p
u
ts
,
with
th
e
o
b
jectiv
e
o
f
en
h
a
n
cin
g
d
ia
g
n
o
s
tic
ac
cu
r
ac
y
.
Pre
cise
id
en
tific
atio
n
o
f
l
u
n
g
s
o
u
n
d
s
is
ess
en
tial
f
o
r
d
is
ea
s
e
d
iag
n
o
s
is
an
d
p
atien
t
ca
r
e.
Ho
wev
er
,
co
n
v
en
tio
n
al
s
teth
o
s
co
p
es
f
r
eq
u
en
tly
p
r
o
d
u
ce
in
d
is
tin
ct
o
u
tco
m
es
d
u
e
to
am
b
ien
t
n
o
is
e
an
d
s
u
b
jectiv
e
in
ter
p
r
etatio
n
.
T
h
e
o
b
jectiv
e
o
f
th
is
wo
r
k
is
to
im
p
r
o
v
e
t
h
e
an
aly
s
is
o
f
lu
n
g
s
o
u
n
d
s
b
y
em
p
l
o
y
in
g
p
r
e
-
p
r
o
ce
s
s
in
g
tech
n
iq
u
es,
ex
tr
ac
t
in
g
r
elev
an
t
f
ea
tu
r
es,
u
til
izin
g
d
ee
p
lear
n
in
g
f
o
r
class
if
icatio
n
o
f
n
o
r
m
al
an
d
ab
n
o
r
m
al
lu
n
g
s
o
u
n
d
s
b
y
em
p
lo
y
in
g
d
is
tin
ct
f
ea
tu
r
e
in
p
u
ts
.
T
h
e
r
aw
d
ata
o
f
lu
n
g
s
o
u
n
d
s
ex
h
ib
i
t
ir
r
eg
u
lar
ities
,
wh
ich
p
o
s
es
a
ch
allen
g
e
f
o
r
d
i
r
ec
t
class
if
icatio
n
.
Pre
-
p
r
o
ce
s
s
in
g
g
u
a
r
a
n
tees
co
n
s
is
ten
cy
,
fa
cilitatin
g
p
r
ec
is
e
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
class
if
icatio
n
.
I
n
o
r
d
er
to
en
h
a
n
ce
th
e
class
if
icat
io
n
ef
f
icien
cy
,
it
is
n
ec
ess
ar
y
to
o
p
tim
ize
th
e
f
e
atu
r
e
ex
tr
ac
tio
n
m
eth
o
d
s
d
u
e
to
th
e
in
tr
icate
n
at
u
r
e
o
f
r
esp
ir
ato
r
y
s
o
u
n
d
s
.
C
las
s
if
y
in
g
a
wid
e
r
a
n
g
e
o
f
r
esp
ir
ato
r
y
ab
n
o
r
m
alities
r
em
a
in
s
ch
allen
g
in
g
d
u
e
to
th
e
q
u
a
lity
o
f
th
e
d
ata
an
d
th
e
im
p
o
r
tan
ce
o
f
th
e
f
ea
t
u
r
es,
d
esp
ite
th
e
ex
is
ten
ce
o
f
n
u
m
er
o
u
s
d
ee
p
lear
n
in
g
m
o
d
els.
T
h
is
wo
r
k
in
v
esti
g
ates th
e
ef
f
ec
tiv
en
ess
o
f
v
ar
io
u
s
ty
p
es o
f
f
ea
tu
r
es u
tili
ze
d
in
th
e
class
if
ica
tio
n
o
f
lu
n
g
s
o
u
n
d
s
b
y
u
s
in
g
d
ee
p
lear
n
i
n
g
.
T
h
o
r
o
u
g
h
ass
ess
m
en
t
o
f
th
ese
m
o
d
els
is
cr
u
cial
in
o
r
d
e
r
to
id
en
tify
th
e
m
o
s
t
ef
f
icien
t
f
ea
tu
r
e
s
ets f
o
r
id
en
tify
in
g
r
esp
ir
ato
r
y
p
r
o
b
lem
s
a
n
d
in
c
o
r
p
o
r
atin
g
t
h
em
in
to
clin
ical
p
r
ac
tice.
T
h
is
p
ap
er
s
ig
n
if
ican
tly
co
n
tr
i
b
u
tes
to
lu
n
g
d
is
ea
s
e
d
iag
n
o
s
i
s
b
y
p
r
o
v
i
d
in
g
an
a
d
v
an
ce
d
m
eth
o
d
f
o
r
ca
p
tu
r
in
g
an
d
an
aly
zin
g
lu
n
g
s
o
u
n
d
s
u
s
in
g
d
ee
p
lear
n
in
g
an
d
d
ig
ital
s
teth
o
s
co
p
es.
I
t
in
tr
o
d
u
ce
s
a
r
o
b
u
s
t
p
r
e
-
p
r
o
ce
s
s
in
g
f
r
am
ewo
r
k
to
s
tan
d
ar
d
ize
lu
n
g
s
o
u
n
d
d
ata
an
d
co
m
p
ar
es f
ea
tu
r
e
ex
tr
ac
tio
n
m
eth
o
d
s
,
in
clu
d
i
n
g
Mel
-
f
r
eq
u
e
n
cy
ce
p
s
tr
al
co
ef
f
i
cien
t
(
MFC
C
)
,
s
p
ec
tr
o
g
r
am
s
,
an
d
th
eir
co
m
b
i
n
atio
n
s
.
A
k
ey
co
n
tr
ib
u
tio
n
is
th
e
d
ev
elo
p
m
e
n
t
o
f
th
e
a
u
to
m
atic
lu
n
g
s
o
u
n
d
d
iag
n
o
s
is
n
etwo
r
k
(
AL
SD
-
Net)
,
wh
ich
co
m
b
i
n
es
two
-
d
im
en
s
io
n
al
an
d
o
n
e
-
d
im
en
s
io
n
al
co
n
v
o
lu
t
io
n
al
lay
er
s
to
ca
p
tu
r
e
in
tr
icat
e
lu
n
g
s
o
u
n
d
p
atter
n
s
.
C
o
m
p
r
eh
en
s
iv
e
ev
alu
atio
n
h
ig
h
lig
h
ts
th
e
s
u
p
er
io
r
ity
o
f
i
n
teg
r
ated
f
ea
tu
r
es,
with
th
e
tr
i
p
le
f
ea
tu
r
e
m
o
d
el
ac
h
iev
i
n
g
t
h
e
h
ig
h
est
ac
cu
r
ac
y
.
T
h
ese
f
in
d
in
g
s
d
em
o
n
s
tr
ate
h
o
w
d
ee
p
lear
n
in
g
-
b
ased
d
ig
i
t
al
s
teth
o
s
co
p
es
ca
n
en
h
an
ce
n
o
n
-
in
v
asiv
e,
r
ea
l
-
tim
e
m
o
n
ito
r
in
g
an
d
r
em
o
te
p
atien
t c
ar
e,
esp
ec
ially
in
r
eso
u
r
ce
-
lim
ited
s
ettin
g
s
.
2.
RE
L
AT
E
D
P
RE
VIOU
S
ST
UDIE
S
E
f
f
ec
tiv
e
p
r
e
-
p
r
o
ce
s
s
in
g
is
cr
u
cial
f
o
r
ac
cu
r
ately
class
if
y
in
g
lu
n
g
s
o
u
n
d
s
u
s
in
g
d
ee
p
lear
n
in
g
tech
n
iq
u
es.
R
esam
p
lin
g
,
as
d
o
n
e
b
y
[
8
]
,
en
s
u
r
es
u
n
if
o
r
m
ity
ac
r
o
s
s
d
atasets
b
y
ch
o
o
s
in
g
a
s
am
p
le
r
ate
o
f
4
4
.
1
k
Hz.
I
n
[
9
]
,
th
e
a
u
d
io
f
r
eq
u
en
c
y
was
a
d
ju
s
ted
to
2
2
k
Hz,
wh
er
ea
s
[
1
0
]
r
esam
p
le
d
all
r
ec
o
r
d
i
n
g
s
at
4
,
0
0
0
Hz,
c
o
n
s
id
er
in
g
th
at
th
e
p
r
im
ar
y
s
ig
n
al
o
f
in
ter
est is
m
o
s
tly
b
elo
w
2
,
0
0
0
Hz.
Fil
ter
in
g
m
eth
o
d
s
,
s
u
ch
as
Gau
s
s
ian
B
u
tter
wo
r
th
f
ilter
s
u
s
ed
b
y
[
1
1
]
,
s
e
co
n
d
-
o
r
d
er
B
u
tter
wo
r
th
b
a
n
d
p
ass
f
ilter
s
u
t
ilized
b
y
[
1
2
]
,
an
d
wav
elet
d
en
o
is
in
g
with
h
ig
h
-
p
ass
f
ilter
in
g
em
p
l
o
y
ed
b
y
[
1
3
]
,
s
u
cc
ess
f
u
lly
r
e
m
o
v
e
u
n
wa
n
ted
n
o
is
e
an
d
r
etai
n
im
p
o
r
tan
t
s
ig
n
als.
Fo
r
au
d
io
clip
p
in
g
tech
n
iq
u
es,
s
tu
d
y
[
1
0
]
f
it
cy
cles
in
to
d
u
r
atio
n
s
eg
m
en
ts
o
f
2
.
7
s
,
en
s
u
r
in
g
d
ata
co
n
s
is
ten
cy
an
d
co
m
p
atib
ilit
y
.
Data
a
u
g
m
en
t
atio
n
m
eth
o
d
s
,
s
u
ch
as
tim
e
s
tr
etch
in
g
an
d
v
o
ca
l
tr
ac
t le
n
g
th
p
e
r
tu
r
b
atio
n
(
VT
L
P),
ar
e
em
p
lo
y
ed
b
y
[
9
]
to
i
n
c
r
ea
s
e
d
ata
q
u
ality
an
d
d
iv
e
r
s
ity
.
Mel
-
s
p
ec
tr
o
g
r
am
s
an
d
s
p
ec
t
r
o
g
r
am
s
o
b
tai
n
ed
u
s
in
g
s
h
o
r
t
-
t
im
e
Fo
u
r
ier
tr
an
s
f
o
r
m
(
S
T
FT)
ar
e
ess
en
tial
f
o
r
ex
am
in
in
g
lu
n
g
s
o
u
n
d
s
.
Me
l
-
s
p
ec
tr
o
g
r
am
s
,
as
em
p
lo
y
ed
b
y
[
1
3
]
a
n
d
[
1
4
]
,
co
n
v
er
t
f
r
eq
u
e
n
cy
co
m
p
o
n
en
ts
to
th
e
Me
l
s
ca
le,
p
r
o
v
id
i
n
g
p
e
r
ce
p
tu
ally
s
ig
n
if
ican
t
in
f
o
r
m
atio
n
f
o
r
s
o
u
n
d
an
al
y
s
is
.
STFT
s
p
ec
tr
o
g
r
am
s
,
as
d
em
o
n
s
tr
ated
b
y
[
9
]
a
n
d
[
1
0
]
,
en
a
b
le
th
e
s
im
u
ltan
e
o
u
s
co
llectio
n
o
f
tem
p
o
r
al
an
d
f
r
eq
u
e
n
cy
f
e
atu
r
es,
wh
ich
is
cr
u
cial
f
o
r
ac
c
u
r
ately
class
if
y
in
g
r
esp
ir
ato
r
y
cy
cles.
Sp
e
ctr
o
g
r
am
clip
p
in
g
im
p
r
o
v
es
p
r
o
ce
s
s
in
g
ef
f
icien
cy
b
y
em
p
h
asizin
g
im
p
o
r
tan
t
s
o
u
n
d
f
ea
tu
r
es.
Me
l
-
f
r
eq
u
en
c
y
ce
p
s
tr
al
co
ef
f
icien
ts
(
MFC
C
s
)
,
o
b
t
ain
ed
u
s
in
g
FF
T
an
d
d
is
cr
ete
c
o
s
in
e
tr
an
s
f
o
r
m
(
DC
T
)
,
ac
c
u
r
ately
ch
ar
ac
ter
ize
s
p
ec
tr
u
m
attr
ib
u
tes.
R
esear
ch
er
s
s
u
ch
as
[
1
5
]
a
n
d
[
1
6
]
,
h
av
e
u
tili
ze
d
MFC
C
s
to
im
p
r
o
v
e
class
if
icatio
n
ac
cu
r
ac
y
b
y
r
ed
u
ci
n
g
f
ea
t
u
r
e
co
r
r
elatio
n
an
d
em
u
latin
g
h
u
m
a
n
au
d
ito
r
y
p
er
ce
p
tio
n
.
T
h
ese
tech
n
i
q
u
es
en
h
an
ce
t
h
e
q
u
ality
an
d
co
m
p
r
eh
en
s
ib
ilit
y
o
f
f
ea
tu
r
e
r
ep
r
esen
tatio
n
s
v
ital
f
o
r
th
e
a
n
aly
s
is
an
d
ca
te
g
o
r
iz
atio
n
o
f
lu
n
g
s
o
u
n
d
s
.
Dee
p
lear
n
in
g
alg
o
r
ith
m
s
h
av
e
d
em
o
n
s
tr
ated
en
c
o
u
r
a
g
in
g
o
u
tco
m
es
in
ca
teg
o
r
izin
g
lu
n
g
s
o
u
n
d
s
.
R
esear
ch
er
s
[
1
4
]
ac
h
iev
ed
9
4
%
ac
cu
r
ac
y
u
s
in
g
a
VGGish
-
s
tack
ed
b
id
ir
ec
tio
n
al
g
ated
r
ec
u
r
r
en
t u
n
it
(
B
iGR
U)
m
o
d
el,
f
o
cu
s
in
g
o
n
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e.
I
n
[
1
1
]
,
th
e
AL
SD
-
Net,
a
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
(
C
NN)
,
ac
h
iev
ed
9
4
.
2
4
% a
cc
u
r
ac
y
.
An
o
t
h
er
s
tu
d
y
[
1
2
]
in
teg
r
ated
a
C
NN
with
b
est d
is
cr
ep
an
cy
f
o
r
est
(
B
DF)
,
attain
in
g
r
em
a
r
k
ab
le
p
er
f
o
r
m
an
ce
with
9
9
.
9
4
%
ac
cu
r
ac
y
an
d
im
p
r
ess
iv
e
p
r
ec
is
io
n
,
s
p
e
cif
icity
,
s
en
s
iti
v
ity
,
an
d
F1
-
s
co
r
e
m
etr
ics.
T
r
a
d
itio
n
al
m
ac
h
in
e
lear
n
in
g
m
eth
o
d
s
h
av
e
also
b
ee
n
s
u
cc
ess
f
u
l:
[
8
]
ac
h
iev
e
d
9
9
%
ac
cu
r
ac
y
with
f
in
e
Gau
s
s
ian
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
S
VM
)
,
an
d
[
1
5
]
o
b
tai
n
ed
9
7
.
4
5
%
ac
cu
r
ac
y
with
g
r
ad
ien
t
b
o
o
s
tin
g
.
Hy
b
r
id
m
o
d
els,
lik
e
th
e
C
NN
bi
-
d
ir
ec
tio
n
al
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
B
DL
STM
)
u
s
ed
b
y
[
1
7
]
,
ac
h
iev
e
d
9
8
.
2
6
%
ac
cu
r
ac
y
,
e
f
f
ec
tiv
ely
i
n
teg
r
atin
g
tem
p
o
r
al
an
d
s
p
atial
v
ar
ia
b
les.
T
h
is
r
esear
c
h
h
ig
h
lig
h
ts
s
ig
n
if
ican
t
ad
v
an
c
em
en
ts
in
AI
alg
o
r
ith
m
s
f
o
r
th
e
ac
cu
r
ate
class
if
icatio
n
an
d
d
iag
n
o
s
is
o
f
r
esp
ir
ato
r
y
d
is
o
r
d
er
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
-
8
7
0
8
I
n
teg
r
a
tin
g
time
-
fr
eq
u
en
cy
fea
tu
r
es w
ith
d
ee
p
lea
r
n
in
g
fo
r
lu
n
g
s
o
u
n
d
cla
s
s
ifica
tio
n
…
(
S
u
Yu
a
n
C
h
a
n
g
)
3739
3.
M
E
T
H
O
D
T
h
is
s
tu
d
y
ad
o
p
ts
a
s
tr
u
ctu
r
ed
a
p
p
r
o
ac
h
t
o
d
ev
elo
p
a
d
ee
p
lear
n
i
n
g
m
o
d
el
f
o
r
l
u
n
g
s
o
u
n
d
class
if
icatio
n
.
I
t
in
v
o
lv
es
m
u
ltip
le
s
tag
es,
b
eg
in
n
in
g
with
s
i
g
n
al
ac
q
u
is
itio
n
an
d
f
o
llo
wed
b
y
p
r
e
-
p
r
o
ce
s
s
in
g
,
f
ea
tu
r
e
e
x
tr
ac
tio
n
,
m
o
d
el
d
ev
elo
p
m
en
t,
tr
ain
in
g
,
an
d
ev
al
u
atio
n
.
E
ac
h
s
tag
e
p
lay
s
a
cr
iti
ca
l
r
o
le
in
en
s
u
r
i
n
g
th
e
ac
cu
r
ac
y
a
n
d
r
eliab
ilit
y
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
tem
,
an
d
th
ey
ar
e
d
etailed
in
t
h
e
f
o
llo
win
g
s
u
b
s
ec
tio
n
s
.
3
.
1
.
L
un
g
s
o
un
d si
g
na
l a
cqu
is
it
io
n
T
h
e
d
ataset
f
r
o
m
th
e
I
n
ter
n
atio
n
al
C
o
n
f
er
e
n
ce
o
n
B
io
m
ed
ic
al
an
d
Hea
lth
I
n
f
o
r
m
atics
(
I
C
HB
I
)
2
0
1
7
C
h
allen
g
e
is
u
tili
ze
d
i
n
th
is
s
tu
d
y
.
I
t
is
a
s
cien
tific
c
h
a
llen
g
e
th
at
to
o
k
p
lace
in
2
0
1
7
,
w
h
ich
o
f
f
er
s
a
r
esp
ir
ato
r
y
d
ata
b
ase
an
d
an
o
f
f
icial
s
co
r
e
s
y
s
tem
.
A
to
tal
o
f
5
.
5
h
o
u
r
s
o
f
r
ec
o
r
d
i
n
g
s
wi
th
a
co
llectio
n
o
f
9
2
0
an
n
o
tated
a
u
d
io
s
am
p
les
f
r
o
m
1
2
6
s
u
b
jects
ar
e
in
clu
d
ed
i
n
th
is
d
atab
ase
[
1
8
]
.
W
ith
a
to
t
al
o
f
6
,
8
9
8
cy
cles,
th
ese
cy
cles
ar
e
f
u
r
t
h
er
ca
te
g
o
r
ized
i
n
to
3
,
6
4
2
n
o
r
m
al
c
y
cles,
1
,
8
6
4
co
n
tain
in
g
c
r
ac
k
les,
8
8
6
co
n
tain
in
g
wh
ee
ze
s
,
an
d
5
0
6
co
n
tain
in
g
b
o
th
cr
ac
k
les an
d
wh
ee
ze
s
.
T
a
b
le
1
s
h
o
ws th
e
lu
n
g
s
o
u
n
d
ch
ar
ac
ter
is
tics
.
T
ab
le
1
.
L
u
n
g
s
o
u
n
d
c
h
ar
ac
ter
is
tics
Lu
n
g
so
u
n
d
C
a
u
se
A
c
o
u
st
i
c
c
h
a
r
a
c
t
e
r
i
st
i
c
A
sso
c
i
a
t
e
d
d
i
se
a
se
N
o
r
mal
S
mo
o
t
h
a
n
d
c
o
n
t
i
n
u
o
u
s
a
i
r
f
l
o
w
t
h
r
o
u
g
h
t
h
e
r
e
sp
i
r
a
t
o
r
y
p
a
ss
a
g
e
s wi
t
h
o
u
t
a
n
y
a
b
n
o
r
ma
l
d
i
sr
u
p
t
i
o
n
s
.
F
r
e
q
u
e
n
c
y
r
a
n
g
e
:
60
–
6
0
0
H
z
-
C
r
a
c
k
l
e
Ex
p
l
o
si
v
e
o
p
e
n
i
n
g
o
f
sm
a
l
l
a
i
r
w
a
y
s
o
r
t
h
e
a
l
v
e
o
l
i
,
a
i
r
b
u
b
b
l
e
i
n
l
a
r
g
e
r
a
i
r
w
a
y
s.
N
o
n
-
m
u
si
c
a
l
a
n
d
e
x
p
l
o
si
v
e
i
n
n
a
t
u
r
e
F
r
e
q
u
e
n
c
y
r
a
n
g
e
:
3
5
0
–
6
5
0
H
z
D
u
r
a
t
i
o
n
:
<
2
0
ms
I
n
t
e
r
st
i
t
i
a
l
l
u
n
g
f
i
b
r
o
si
s
P
n
e
u
mo
n
i
a
C
O
P
D
B
r
o
n
c
h
i
e
c
t
a
si
s
A
st
h
m
a
W
h
e
e
z
e
A
i
r
f
l
o
w
l
i
m
i
t
a
t
i
o
n
a
n
d
a
i
r
w
a
y
n
a
r
r
o
w
i
n
g
.
F
r
e
q
u
e
n
c
y
r
a
n
g
e
:
1
0
0
–
1
,
0
0
0
H
z
D
u
r
a
t
i
o
n
:
>
1
0
0
ms
C
O
P
D
A
st
h
m
a
Tu
m
o
r
b
l
o
c
k
i
n
g
a
i
r
w
a
y
3
.
2
.
P
re
-
pro
ce
s
s
ing
T
h
is
s
tu
d
y
u
tili
ze
s
v
ar
io
u
s
p
r
e
-
p
r
o
ce
s
s
in
g
tec
h
n
iq
u
es
t
o
p
r
ep
ar
e
lu
n
g
s
o
u
n
d
d
ata
f
o
r
cla
s
s
if
icatio
n
.
I
n
itially
,
d
o
wn
s
am
p
lin
g
is
e
m
p
lo
y
ed
to
d
ec
r
ea
s
e
th
e
s
am
p
lin
g
r
ate
to
4
,
0
0
0
Hz
in
o
r
d
er
to
s
im
p
lify
th
e
p
r
o
ce
s
s
in
g
,
wh
ile
co
n
tin
u
in
g
to
a
d
h
er
e
to
t
h
e
g
u
id
elin
e
s
g
iv
en
b
y
th
e
N
y
q
u
is
t
th
eo
r
em
[
1
0
]
,
[
1
5
]
.
Seg
m
en
tatio
n
is
th
e
p
r
o
ce
s
s
o
f
d
iv
id
in
g
r
ec
o
r
d
in
g
s
in
to
r
esp
ir
ato
r
y
cy
cles
u
s
in
g
an
n
o
t
ated
s
tar
t
an
d
s
to
p
tim
es.
T
h
is
r
esu
lts
in
d
is
tin
ct
ca
t
eg
o
r
ies
s
u
ch
as
n
o
r
m
al,
cr
a
ck
les,
wh
ee
ze
s
,
an
d
co
m
b
in
ati
o
n
o
f
b
o
th
c
r
ac
k
les
an
d
wh
ee
ze
s
.
T
h
e
au
d
i
o
clip
p
in
g
p
r
o
ce
s
s
en
s
u
r
es
th
at
all
cy
cle
d
u
r
atio
n
s
ar
e
s
tan
d
ar
d
ize
d
to
2
.
7
s
ec
o
n
d
s
b
y
eith
er
cr
o
p
p
in
g
ex
ce
s
s
d
u
r
at
io
n
f
r
o
m
l
o
n
g
e
r
cy
cles
o
r
em
p
lo
y
in
g
ze
r
o
p
a
d
d
in
g
to
s
h
o
r
ter
o
n
es
[
1
0
]
.
Am
p
litu
d
e
n
o
r
m
aliza
tio
n
e
n
s
u
r
es
co
n
s
is
ten
cy
b
y
r
escalin
g
al
l
s
ig
n
als
to
th
e
r
a
n
g
e
o
f
(
-
1
,
1
)
in
o
r
d
er
to
r
ed
u
ce
v
ar
iatio
n
s
ca
u
s
ed
b
y
n
o
is
e
an
d
p
h
y
s
io
lo
g
ical
d
if
f
er
en
ce
s
am
o
n
g
p
atien
ts
[
1
1
]
.
T
h
e
p
u
r
p
o
s
e
o
f
th
ese
p
r
e
-
p
r
o
ce
s
s
in
g
m
eth
o
d
s
is
to
im
p
r
o
v
e
th
e
co
n
s
is
ten
cy
an
d
ex
ce
ll
en
ce
o
f
lu
n
g
s
o
u
n
d
d
ata,
m
ak
i
n
g
it
m
o
r
e
s
u
itab
le
f
o
r
s
u
b
s
eq
u
e
n
t d
ee
p
lear
n
in
g
c
lass
if
icatio
n
m
o
d
els.
3
.
3
.
F
e
a
t
ure
ex
t
r
a
ct
io
n
E
x
tr
ac
tin
g
f
ea
tu
r
es
f
r
o
m
lu
n
g
s
o
u
n
d
au
d
io
r
ec
o
r
d
in
g
s
is
cr
u
c
ial
f
o
r
g
en
e
r
atin
g
m
o
r
e
m
a
n
ag
ea
b
le
an
d
u
s
ef
u
l
in
f
o
r
m
atio
n
.
T
h
e
p
r
o
ce
s
s
in
v
o
lv
es
th
e
ca
r
ef
u
l
s
elec
tio
n
o
f
s
ig
n
if
ican
t
f
ea
tu
r
es,
s
tr
e
am
lin
in
g
t
h
e
d
ata
,
an
d
id
e
n
tify
in
g
p
atter
n
s
ass
o
ciate
d
with
r
esp
i
r
ato
r
y
ab
n
o
r
m
a
liti
es.
T
h
e
s
elec
ted
f
ea
tu
r
es
f
o
r
th
is
s
tu
d
y
c
o
n
s
is
t
o
f
MFC
C
an
d
s
p
ec
tr
o
g
r
am
s
.
Fig
u
r
e
1
illu
s
tr
ates
th
r
ee
s
ig
n
al
f
ea
tu
r
es
p
lo
ts
th
at
s
h
o
w
th
e
r
aw
lu
n
g
s
o
u
n
d
i
n
tim
e
d
o
m
ain
in
Fig
u
r
e
1
(
a)
,
th
e
MFC
C
f
ea
tu
r
e
in
Fig
u
r
e
1
(
b
)
an
d
th
e
s
p
ec
tr
o
g
r
am
in
Fig
u
r
e
1
(
c)
.
3
.
3
.
1
.
M
el
-
f
re
qu
ency
ce
ps
t
ra
l c
o
ef
f
icient
T
h
e
p
r
o
ce
s
s
o
f
ex
tr
ac
tin
g
MF
C
C
b
eg
in
s
b
y
ap
p
ly
in
g
th
e
F
FT
o
n
s
ig
n
als
th
at
h
av
e
b
ee
n
win
d
o
wed
.
A
Me
l
-
s
ca
le
f
ilter
b
an
k
,
s
p
e
cif
ically
a
tr
ian
g
le
b
a
n
d
p
ass
f
ilter
b
an
k
,
is
em
p
lo
y
e
d
to
co
n
v
er
t
th
e
lin
ea
r
f
r
eq
u
e
n
cy
s
p
ec
tr
u
m
i
n
to
t
h
e
Me
l
-
f
r
eq
u
e
n
cy
s
ca
le,
r
ep
licati
n
g
th
e
au
d
ito
r
y
p
er
ce
p
tio
n
o
f
h
u
m
an
s
.
T
h
e
f
o
r
m
u
la
em
p
lo
y
ed
as
(
1
)
[
1
9
]
:
=
2596
∗
(
1
+
700
)
(
1
)
wh
er
e
r
ep
r
esen
ts
f
r
eq
u
en
cy
in
Me
l
s
ca
le
an
d
r
ep
r
esen
ts
f
r
eq
u
e
n
cy
in
li
n
ea
r
s
ca
le.
Ap
p
ly
in
g
a
lo
g
ar
ith
m
ic
tr
a
n
s
f
o
r
m
atio
n
af
ter
f
ilter
in
g
d
ec
r
ea
s
es
th
e
im
p
ac
t
o
f
c
h
an
g
es
i
n
am
p
litu
d
e.
Af
ter
war
d
s
,
th
e
lo
g
ar
ith
m
ic
-
s
ca
le
s
ig
n
al
is
ap
p
lied
to
th
e
DC
T
in
o
r
d
er
to
ca
lcu
late
th
e
MFC
C
s
.
T
h
ese
co
ef
f
icien
ts
q
u
an
tify
th
e
am
p
litu
d
e
o
f
th
e
s
p
ec
tr
u
m
in
th
e
tim
e
d
o
m
ain
.
T
h
is
m
eth
o
d
im
p
r
o
v
es
th
e
r
ep
r
esen
tatio
n
o
f
ess
en
tial
ac
o
u
s
tic
f
ea
tu
r
es f
o
r
class
if
ica
tio
n
task
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
4
,
Au
g
u
s
t
20
25
:
3
7
3
7
-
3747
3740
(
a)
(
b
)
(
c)
Fig
u
r
e
1
.
T
h
e
lu
n
g
s
o
u
n
d
s
ig
n
al,
(
a)
th
e
p
r
e
-
p
r
o
ce
s
s
ed
s
ig
n
al
in
tim
e
d
o
m
ai
n
,
(
b
)
MFC
C
f
ea
tu
r
e,
an
d
(
c)
s
p
ec
tr
o
g
r
am
f
ea
tu
r
e
3
.
3
.
2
.
Sp
ec
t
ro
g
ra
m
T
im
e
-
v
ar
y
in
g
f
r
eq
u
en
cy
c
o
m
p
o
n
en
ts
o
f
au
d
io
s
ig
n
als
ca
n
b
e
id
en
tifie
d
b
y
s
p
ec
tr
o
g
r
am
s
,
wh
ich
ar
e
g
en
er
ated
wh
en
th
e
STFT
t
r
an
s
f
o
r
m
s
win
d
o
wed
p
o
r
tio
n
s
o
f
t
h
e
s
ig
n
als
in
t
o
th
e
f
r
e
q
u
en
cy
[
2
0
]
.
I
t
is
ca
lcu
lated
b
y
(
2
)
:
(
,
)
=
∑
(
)
(
−
)
−
∞
=
−
∞
(
2
)
T
h
e
d
is
cr
ete
-
tim
e
s
ig
n
al
is
r
e
p
r
esen
ted
b
y
(
)
,
t
h
e
win
d
o
w
f
u
n
ctio
n
,
wh
ich
is
u
s
u
ally
Ga
u
s
s
ian
o
r
Han
n
i
n
g
,
is
r
ep
r
esen
ted
b
y
(
)
,
th
e
ti
m
e
in
d
e
x
is
,
an
d
th
e
an
g
u
la
r
f
r
e
q
u
en
c
y
is
in
d
icate
d
b
y
.
T
h
e
s
p
ec
tr
o
g
r
am
is
a
cr
u
cial
to
o
l
f
o
r
a
n
aly
zin
g
c
o
m
p
licated
,
tim
e
-
v
ar
y
in
g
s
ig
n
a
ls
lik
e
r
esp
ir
ato
r
y
cy
cles
b
ec
au
s
e
it
r
ec
o
r
d
s
h
o
w
th
e
f
r
eq
u
en
cy
co
n
ten
t
ch
a
n
g
es
o
v
er
tim
e
b
y
m
o
v
in
g
th
e
win
d
o
w
ac
r
o
s
s
th
e
s
ig
n
al
an
d
c
o
m
p
u
tin
g
th
e
STFT
at
ea
ch
p
lace
.
3
.
4
.
Dee
p
lea
rning
cla
s
s
if
ica
t
io
n
Usi
n
g
s
eq
u
en
tial
C
NN
ar
ch
ite
ctu
r
e
d
esig
n
ed
f
o
r
b
o
th
au
d
i
o
an
d
v
is
u
al
d
ata
p
r
o
ce
s
s
in
g
,
n
a
m
ely
,
th
e
ad
o
p
ted
AL
SD
-
Net
m
o
d
el
o
f
f
er
s
an
ad
v
an
ce
d
m
eth
o
d
f
o
r
au
to
m
ated
lu
n
g
s
o
u
n
d
d
iag
n
o
s
tic
s
.
T
h
is
m
o
d
el
in
v
o
lv
es
two
-
d
im
en
s
io
n
al
co
n
v
o
lu
tio
n
al
lay
er
s
to
ex
tr
ac
t
f
ea
tu
r
es
f
r
o
m
p
ictu
r
e
r
ep
r
esen
tatio
n
s
an
d
o
n
e
-
d
im
en
s
io
n
al
co
n
v
o
lu
tio
n
al
lay
er
s
to
an
aly
ze
au
d
i
o
r
ec
o
r
d
in
g
s
.
Sev
en
co
n
v
o
l
u
tio
n
al
lay
er
s
m
ak
e
u
p
th
e
m
o
d
el
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
-
8
7
0
8
I
n
teg
r
a
tin
g
time
-
fr
eq
u
en
cy
fea
tu
r
es w
ith
d
ee
p
lea
r
n
in
g
fo
r
lu
n
g
s
o
u
n
d
cla
s
s
ifica
tio
n
…
(
S
u
Yu
a
n
C
h
a
n
g
)
3741
ar
ch
itectu
r
e.
T
h
e
f
ir
s
t
f
o
u
r
lay
er
s
ea
ch
h
a
v
e
1
6
f
ilter
s
an
d
r
e
ctif
ied
lin
ea
r
u
n
it
(
R
eL
U)
ac
ti
v
atio
n
.
T
o
im
p
r
o
v
e
tr
ain
in
g
ef
f
icie
n
cy
an
d
s
tab
ilit
y
,
m
ax
-
p
o
o
lin
g
with
a
p
o
o
l
s
i
ze
o
f
2
an
d
b
atch
n
o
r
m
aliza
ti
o
n
ar
e
a
p
p
lied
a
f
ter
ea
ch
co
n
v
o
lu
tio
n
al
lay
e
r
.
W
ith
3
2
f
ilter
s
ap
iece
,
s
u
b
s
eq
u
e
n
t
co
n
v
o
lu
tio
n
al
lay
e
r
s
p
r
eser
v
e
r
eliab
le
f
ea
tu
r
e
ex
tr
ac
tio
n
at
v
ar
io
u
s
ab
s
tr
ac
tio
n
lev
els.
I
n
o
r
d
er
to
av
o
id
o
v
er
f
itti
n
g
d
u
r
i
n
g
m
o
d
el
tr
ain
in
g
,
d
r
o
p
o
u
t
lay
er
s
ar
e
ca
r
ef
u
lly
p
lace
d
af
ter
th
e
co
n
v
o
lu
tio
n
al
an
d
d
e
n
s
e
lay
er
s
,
with
r
ates
o
f
0
.
2
an
d
0
.
4
,
r
esp
ec
tiv
ely
.
I
n
o
r
d
er
to
g
en
er
ate
p
r
o
b
ab
ilis
tic
p
r
ed
icti
o
n
s
f
o
r
th
e
4
ty
p
es
o
f
lu
n
g
s
s
o
u
n
d
s
t
h
at
th
e
m
o
d
el
is
tar
g
etin
g
,
th
e
f
in
al
lay
er
u
s
es
a
d
en
s
e
lay
e
r
with
So
f
t
Ma
x
ac
tiv
atio
n
[
1
1
]
.
T
h
is
ar
c
h
itectu
r
e
h
ig
h
lig
h
ts
th
e
m
o
d
el'
s
ab
ilit
y
to
id
en
tify
co
m
p
lex
p
atter
n
s
in
lu
n
g
s
o
u
n
d
d
ata.
T
a
b
le
2
an
d
Fig
u
r
e
2
s
h
o
w
th
e
h
y
p
er
p
ar
am
eter
s
s
et
a
n
d
th
e
a
r
ch
itectu
r
e
m
o
d
el
f
o
r
AL
SD
-
Net,
r
esp
ec
tiv
ely
.
T
ab
le
2
.
Hy
p
er
p
ar
a
m
eter
s
f
o
r
AL
SD
-
Net
m
o
d
el
H
y
p
e
r
p
a
r
a
me
t
e
r
V
a
l
u
e
Lo
ss f
u
n
c
t
i
o
n
C
a
t
e
g
o
r
i
c
a
l
c
r
o
ss
-
e
n
t
r
o
p
y
B
a
t
c
h
si
z
e
32
Ep
o
c
h
s
50
Ea
r
l
y
st
o
p
p
i
n
g
p
a
t
i
e
n
c
e
10
O
p
t
i
mi
z
e
r
A
d
a
m
Le
a
r
n
i
n
g
r
a
t
e
1
×
10
-
5
A
c
t
i
v
a
t
i
o
n
f
u
n
c
t
i
o
n
R
e
LU
Fig
u
r
e
2
.
Stru
ctu
r
e
o
f
AL
SD
-
Net
m
o
d
el
3
.
5
.
P
er
f
o
r
m
a
nce
e
v
a
lua
t
io
n
T
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
AL
SD
-
Net
m
o
d
el
with
d
is
tin
ct
s
et
s
o
f
f
ea
tu
r
e
in
p
u
t
is
ev
alu
ate
d
b
y
u
s
in
g
im
p
o
r
tan
t
m
etr
ics
an
d
g
r
ap
h
ical
to
o
ls
s
u
ch
as
lear
n
in
g
cu
r
v
es
an
d
p
er
f
o
r
m
an
ce
ev
al
u
atio
n
m
etr
ics.
T
h
e
f
ir
s
t
ev
alu
atio
n
is
th
e
lear
n
i
n
g
c
u
r
v
es
o
f
t
h
e
d
ee
p
lear
n
i
n
g
tr
ain
in
g
an
d
v
alid
ati
o
n
p
r
o
ce
s
s
,
w
h
ich
illu
s
tr
ate
h
o
w
well
a
m
o
d
el
p
r
ed
icts
p
er
f
o
r
m
an
ce
o
v
er
tim
e
as
a
f
u
n
ctio
n
o
f
tr
ain
in
g
ef
f
o
r
t,
ar
e
cr
u
cia
l
v
is
u
al
aid
s
in
d
ee
p
lear
n
in
g
.
T
h
ese
c
u
r
v
es
p
r
im
ar
ily
s
h
o
w
th
e
m
o
d
el'
s
ac
cu
r
ac
y
o
r
lo
s
s
b
ased
o
n
th
e
n
u
m
b
e
r
o
f
tr
ain
i
n
g
e
p
o
ch
s
an
d
b
o
th
tr
ain
in
g
a
n
d
v
alid
at
io
n
d
atasets
.
B
y
an
aly
zin
g
th
ese
cu
r
v
es,
s
ig
n
if
ican
t
d
etails
ab
o
u
t
th
e
m
o
d
el'
s
lear
n
in
g
an
d
g
e
n
er
aliza
tio
n
b
eh
av
io
r
s
ca
n
b
e
id
e
n
tifie
d
[
2
1
]
.
T
h
en
,
o
n
ce
co
m
p
leted
th
e
tr
ain
in
g
an
d
v
alid
atio
n
,
p
er
f
o
r
m
an
ce
e
v
al
u
atio
n
m
etr
ics
th
at
in
v
o
lv
e
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e
,
ar
e
co
n
s
id
er
ed
.
T
h
ese
m
etr
ics
p
r
o
v
id
e
in
s
ig
h
ts
in
to
th
e
s
tr
en
g
th
s
an
d
wea
k
n
ess
es
o
f
d
if
f
er
en
t
f
ea
tu
r
e
s
ets.
W
eig
h
ted
av
er
ag
es
o
f
t
h
ese
m
etr
ics
ar
e
co
m
p
u
ted
t
o
f
air
ly
c
o
m
p
ar
e
m
o
d
els
ac
r
o
s
s
b
alan
c
ed
an
d
im
b
alan
ce
d
d
atasets
.
Acc
o
r
d
in
g
to
[
2
2
]
,
a
s
s
ig
n
in
g
s
m
aller
weig
h
ts
to
cl
ass
es
with
m
o
r
e
in
s
tan
ce
s
an
d
lar
g
er
weig
h
ts
to
m
in
o
r
ity
class
es e
n
s
u
r
es a
b
alan
ce
d
ev
alu
atio
n
,
co
n
s
id
er
in
g
th
e
in
f
lu
en
ce
o
f
in
s
tan
ce
s
f
r
o
m
all
class
es.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
s
ec
tio
n
p
r
esen
ts
th
e
r
esu
lts
an
d
em
p
h
asizes
th
e
ev
alu
ati
o
n
o
f
th
e
p
er
f
o
r
m
an
ce
an
d
ef
f
icien
cy
o
f
th
e
p
r
o
p
o
s
ed
AL
SD
-
Net
m
o
d
el
in
class
if
y
in
g
lu
n
g
s
o
u
n
d
s
.
T
h
e
ass
ess
m
en
t
in
v
o
lv
es
m
u
ltip
le
ty
p
es
o
f
lu
n
g
s
o
u
n
d
f
ea
tu
r
es
to
u
n
d
e
r
s
tan
d
t
h
eir
im
p
ac
t
o
n
t
h
e
m
o
d
el'
s
class
if
icatio
n
ac
cu
r
ac
y
.
T
h
ese
f
e
atu
r
es
in
clu
d
e
tim
e
d
o
m
ain
s
ig
n
als,
s
p
ec
tr
o
g
r
am
r
ep
r
esen
tatio
n
s
,
MFC
C
,
d
u
al
f
ea
tu
r
e
in
p
u
t
(
a
c
o
m
b
in
atio
n
o
f
s
p
ec
tr
o
g
r
am
an
d
MFC
C
)
,
an
d
tr
ip
le
f
ea
tu
r
e
in
p
u
t
(
a
c
o
m
b
in
atio
n
o
f
tim
e
d
o
m
ain
,
s
p
ec
tr
o
g
r
am
,
a
n
d
MFC
C
)
.
T
h
e
a
n
aly
s
is
aim
s
to
id
en
tif
y
wh
ic
h
f
ea
t
u
r
e
s
et
o
r
c
o
m
b
in
atio
n
p
r
o
v
i
d
es
th
e
m
o
s
t
ef
f
ec
tiv
e
in
p
u
t
f
o
r
e
n
h
a
n
cin
g
t
h
e
m
o
d
el’
s
class
if
icatio
n
ca
p
ab
ilit
ies.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
4
,
Au
g
u
s
t
20
25
:
3
7
3
7
-
3747
3742
4
.
1
.
T
he
a
do
pte
d ALS
D
-
Net
t
ra
ini
ng
perf
o
rm
a
nce
T
h
e
an
aly
s
is
f
o
cu
s
es
o
n
ac
cu
r
ac
y
an
d
lo
s
s
cu
r
v
es
d
er
iv
ed
f
r
o
m
tr
ain
in
g
an
d
v
alid
atio
n
p
h
ases
as
s
h
o
wn
in
T
ab
le
3
.
R
ap
id
g
ain
s
in
ac
cu
r
ac
y
ar
e
o
b
s
er
v
ed
in
t
h
e
tim
e
d
o
m
ain
in
p
u
t,
wh
ich
s
tab
ilizes
at
r
o
u
g
h
ly
0
.
9
0
f
o
r
tr
ain
in
g
a
n
d
0
.
8
5
f
o
r
v
alid
atio
n
.
T
h
e
ass
o
ciate
d
lo
s
s
v
alu
es
al
s
o
s
h
o
w
co
n
v
er
g
e
n
ce
b
elo
w
0
.
4
an
d
0
.
5
,
r
esp
ec
tiv
el
y
.
T
ab
le
3
.
T
h
e
tr
ain
in
g
p
er
f
o
r
m
an
ce
F
e
a
t
u
r
e
A
c
c
u
r
a
c
y
c
u
r
v
e
Lo
ss
c
u
r
v
e
Ti
me
d
o
m
a
i
n
S
p
e
c
t
r
o
g
r
a
m
M
F
C
C
S
p
e
c
t
r
o
g
r
a
m
a
n
d
M
F
C
C
Ti
me
d
o
m
a
i
n
,
sp
e
c
t
r
o
g
r
a
m
a
n
d
M
F
C
C
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
-
8
7
0
8
I
n
teg
r
a
tin
g
time
-
fr
eq
u
en
cy
fea
tu
r
es w
ith
d
ee
p
lea
r
n
in
g
fo
r
lu
n
g
s
o
u
n
d
cla
s
s
ifica
tio
n
…
(
S
u
Yu
a
n
C
h
a
n
g
)
3743
T
h
e
s
p
ec
tr
o
g
r
am
in
p
u
t
also
s
h
o
ws
a
f
aster
ac
cu
r
ac
y
in
cr
ea
s
e,
s
tab
ilizin
g
at
v
alu
es
clo
s
e
to
0
.
9
5
f
o
r
tr
ain
in
g
an
d
0
.
9
3
f
o
r
v
alid
atio
n
,
with
ju
s
t sm
all
lo
s
s
es b
etwe
en
0
.
1
a
n
d
0
.
1
8
.
C
o
m
p
ar
a
b
le
p
atter
n
s
ca
n
b
e
s
ee
n
in
th
e
MFC
C
in
p
u
t,
with
lo
s
s
v
alu
es
b
etwe
en
0
.
1
an
d
0
.
2
an
d
ac
cu
r
ac
y
v
alu
es
b
etwe
en
0
.
9
5
an
d
0
.
9
3
.
T
h
e
in
teg
r
atio
n
o
f
MFC
C
an
d
s
p
ec
tr
o
g
r
am
in
p
u
ts
r
esu
lts
in
h
ig
h
ac
cu
r
ac
y
,
ap
p
r
o
x
im
ately
0
.
9
4
f
o
r
v
alid
atio
n
an
d
alm
o
s
t
0
.
9
5
f
o
r
tr
ain
in
g
,
with
lo
s
s
es
d
ec
r
ea
s
in
g
to
ab
o
u
t
0
.
2
0
an
d
0
.
2
5
f
o
r
e
ac
h
s
itu
atio
n
.
Sp
ec
if
ically
,
wh
en
tim
e
d
o
m
ain
,
s
p
ec
tr
o
g
r
a
m
,
a
n
d
MFC
C
in
p
u
ts
ar
e
co
m
b
in
ed
,
tr
ain
in
g
an
d
v
ali
d
atio
n
p
ea
k
ac
cu
r
ac
ies
ar
e
ap
p
r
o
x
im
ately
0
.
9
8
,
wh
ile
lo
s
s
v
alu
es a
r
e
clo
s
e
to
0
.
1
0
.
4
.
2
.
T
he
a
do
pte
d ALS
D
-
Net
t
esting
p
er
f
o
rm
a
nce
E
ac
h
f
ea
t
u
r
e
s
et
was
ass
ess
ed
u
s
in
g
m
etr
ics
s
u
c
h
as
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
a
n
d
F1
-
s
co
r
e
f
o
r
class
if
y
in
g
n
o
r
m
al
lu
n
g
s
o
u
n
d
s
,
cr
ac
k
les,
wh
ee
ze
s
,
an
d
co
m
b
in
atio
n
o
f
cr
ac
k
les
a
n
d
w
h
ee
ze
s
,
as
s
h
o
wn
in
T
ab
le
4
.
T
h
e
r
esu
lts
s
h
o
w
s
ig
n
if
ican
t
d
if
f
e
r
e
n
ce
s
in
p
e
r
f
o
r
m
a
n
ce
b
etwe
en
v
a
r
io
u
s
f
ea
tu
r
e
in
p
u
ts
.
Fo
r
ex
am
p
le,
th
e
tim
e
d
o
m
ain
f
ea
tu
r
e
d
em
o
n
s
tr
ated
o
u
ts
tan
d
in
g
p
r
ec
is
io
n
a
n
d
r
e
ca
ll
m
etr
ics,
r
ea
ch
in
g
its
m
ax
im
u
m
ac
cu
r
ac
y
o
f
9
5
.
6
9
%
in
ca
teg
o
r
izin
g
n
o
r
m
al
lu
n
g
s
o
u
n
d
s
.
B
y
c
o
m
p
ar
is
o
n
,
th
e
s
p
ec
tr
o
g
r
a
m
in
p
u
t
d
em
o
n
s
tr
ated
ex
ce
p
tio
n
al
p
er
f
o
r
m
an
ce
in
id
e
n
tify
in
g
c
r
ac
k
l
es,
with
a
9
9
.
4
4
%
ac
cu
r
ac
y
r
a
te,
as
well
as
g
o
o
d
p
r
ec
is
io
n
an
d
r
ec
all
r
atin
g
s
f
o
r
th
is
class
.
I
n
a
s
im
ilar
v
ein
,
th
e
MFC
C
f
ea
tu
r
e
s
et
p
er
f
o
r
m
ed
a
d
m
ir
ab
l
y
,
ac
h
iev
in
g
a
9
7
.
2
0
% c
lass
if
icat
io
n
ac
cu
r
a
c
y
f
o
r
cr
ac
k
les.
T
ab
le
4
.
Per
f
o
r
m
an
ce
ev
alu
ati
o
n
m
etr
ics o
f
t
h
e
ad
o
p
ted
AL
SD
-
Net
b
ased
o
n
m
u
ltip
le
f
ea
t
u
r
es
C
l
a
s
s
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
S
c
o
r
e
(
%)
Ti
me
d
o
m
a
i
n
f
e
a
t
u
r
e
N
o
r
mal
9
5
.
6
9
8
8
.
7
8
9
5
.
7
0
9
2
.
1
1
C
r
a
c
k
l
e
s
9
4
.
9
7
8
9
.
4
7
9
4
.
9
7
9
2
.
1
4
W
h
e
e
z
e
s
4
4
.
7
3
8
5
.
0
0
4
4
.
7
4
5
8
.
6
2
B
o
t
h
(
C
r
a
c
k
l
e
s
+
W
h
e
e
z
e
s)
8
4
.
1
2
8
9
.
8
3
8
4
.
1
3
8
6
.
8
9
O
v
e
r
a
l
l
A
c
c
u
r
a
c
y
8
8
.
8
4
S
p
e
c
t
r
o
g
r
a
m
N
o
r
mal
9
2
.
4
7
9
8
.
0
1
9
2
.
4
7
9
5
.
1
6
C
r
a
c
k
l
e
s
9
9
.
4
4
9
5
.
1
5
9
9
.
4
4
9
7
.
2
7
W
h
e
e
z
e
s
9
0
.
7
9
7
6
.
6
7
9
0
.
7
9
8
3
.
1
3
B
o
t
h
(
C
r
a
c
k
l
e
s
+
W
h
e
e
z
e
s)
9
6
.
8
3
9
8
.
3
9
9
6
.
8
3
9
7
.
6
O
v
e
r
a
l
l
A
c
c
u
r
a
c
y
9
4
.
4
9
M
F
C
C
N
o
r
mal
9
2
.
4
7
9
7
.
4
5
9
2
.
4
7
9
4
.
9
C
r
a
c
k
l
e
s
9
7
.
2
0
9
6
.
6
7
9
7
.
2
1
9
6
.
9
4
W
h
e
e
z
e
s
9
0
.
7
8
7
1
.
1
3
9
0
.
7
9
7
9
.
7
7
B
o
t
h
(
C
r
a
c
k
l
e
s
+
W
h
e
e
z
e
s)
9
2
.
0
6
9
6
.
6
7
9
2
.
0
6
9
4
.
3
1
O
v
e
r
a
l
l
A
c
c
u
r
a
c
y
9
3
.
4
8
S
p
e
c
t
r
o
g
r
a
m
+
M
F
C
C
N
o
r
mal
9
5
.
8
4
9
5
.
4
2
9
6
.
7
3
9
6
.
0
7
C
r
a
c
k
l
e
s
9
9
.
1
7
9
8
.
8
9
9
8
.
8
9
9
8
.
8
9
W
h
e
e
z
e
s
9
7
.
9
2
8
3
.
3
3
8
5
.
5
3
8
4
.
4
2
B
o
t
h
(
C
r
a
c
k
l
e
s
+
W
h
e
e
z
e
s)
9
9
.
5
8
9
3
.
9
4
9
8
.
4
1
9
6
.
1
2
O
v
e
r
a
l
l
A
c
c
u
r
a
c
y
9
5
.
6
5
Ti
me
d
o
m
a
i
n
+
S
p
e
c
t
r
o
g
r
a
m
+
M
F
C
C
N
o
r
mal
9
7
.
6
7
9
9
.
1
7
9
6
.
5
1
9
7
.
8
2
C
r
a
c
k
l
e
s
9
9
.
5
7
9
8
.
9
0
9
9
.
4
4
9
9
.
1
7
W
h
e
e
z
e
s
9
7
.
6
7
8
5
.
7
1
9
4
.
7
4
9
0
.
0
0
B
o
t
h
(
C
r
a
c
k
l
e
s
+
W
h
e
e
z
e
s)
9
9
.
4
1
9
6
.
8
3
9
6
.
8
3
9
6
.
8
3
O
v
e
r
a
l
l
A
c
c
u
r
a
c
y
9
7
.
2
5
T
h
e
ac
cu
r
ac
y
was
im
p
r
o
v
ed
o
v
er
all
b
y
co
m
b
i
n
in
g
s
p
ec
tr
o
g
r
am
an
d
MFC
C
in
p
u
ts
,
esp
ec
i
ally
wh
en
id
en
tify
in
g
co
m
b
in
atio
n
s
o
f
cr
ac
k
les
an
d
wh
ee
ze
s
,
wh
ich
r
esu
lted
in
a
9
9
.
5
8
%
ac
cu
r
ac
y
r
ate.
T
h
e
co
m
b
in
atio
n
o
f
s
p
ec
tr
o
g
r
am
,
MFC
C
,
an
d
tim
e
d
o
m
ain
ch
ar
ac
ter
is
tics
p
r
o
d
u
ce
d
th
e
b
est
o
v
er
all
ac
cu
r
ac
y
o
f
9
7
.
2
5
%,
d
em
o
n
s
tr
atin
g
r
em
ar
k
ab
le
p
er
f
o
r
m
a
n
ce
in
th
e
clas
s
if
icatio
n
o
f
wh
ee
ze
s
an
d
c
r
a
ck
les.
T
h
ese
r
esu
lts
h
ig
h
lig
h
t
th
e
u
s
ef
u
ln
ess
o
f
f
e
atu
r
e
in
teg
r
atio
n
in
im
p
r
o
v
in
g
th
e
lu
n
g
s
o
u
n
d
class
if
icati
o
n
ac
c
u
r
ac
y
o
f
th
e
ad
o
p
ted
AL
SD
-
Net,
p
r
o
v
id
i
n
g
im
p
o
r
tan
t
n
ew
in
f
o
r
m
atio
n
f
o
r
th
e
d
ev
elo
p
m
en
t
o
f
r
eliab
le
d
iag
n
o
s
tic
in
s
tr
u
m
en
ts
in
r
esp
ir
ato
r
y
h
ea
l
th
ca
r
e
ap
p
licatio
n
s
.
T
ab
le
4
also
co
m
p
ar
es
th
e
p
e
r
f
o
r
m
a
n
ce
o
f
d
if
f
er
en
t
f
ea
tu
r
e
in
p
u
t
s
ets
f
o
r
lu
n
g
s
o
u
n
d
cla
s
s
if
icatio
n
.
T
h
e
tim
e
d
o
m
ain
i
n
p
u
t
m
o
d
el
s
h
o
ws
th
e
lo
west
p
er
f
o
r
m
an
ce
,
with
an
ac
cu
r
ac
y
o
f
8
8
.
8
4
%,
p
r
ec
is
io
n
o
f
8
8
.
6
4
%
,
r
ec
all
o
f
8
8
.
8
4
%,
an
d
an
F1
-
s
co
r
e
o
f
8
7
.
9
1
%.
T
h
i
s
in
d
icate
s
its
lim
ited
ab
ilit
y
t
o
ca
p
tu
r
e
co
m
p
le
x
p
atter
n
s
,
m
ak
in
g
it th
e
least r
e
liab
le
ap
p
r
o
ac
h
.
I
n
co
n
tr
ast,
th
e
tr
i
p
le
f
ea
tu
r
e
in
p
u
t
m
o
d
el,
wh
ic
h
co
m
b
in
es
tim
e
d
o
m
ain
,
s
p
ec
tr
o
g
r
am
,
a
n
d
MFC
C
f
ea
tu
r
es,
a
ch
iev
es
th
e
h
ig
h
est p
er
f
o
r
m
an
ce
.
I
t
r
ec
o
r
d
s
a
n
ac
c
u
r
ac
y
o
f
9
7
.
2
5
%,
a
p
r
ec
is
io
n
o
f
9
7
.
4
0
%,
r
ec
all
o
f
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
4
,
Au
g
u
s
t
20
25
:
3
7
3
7
-
3747
3744
9
7
.
1
0
%,
an
d
an
F1
-
s
co
r
e
o
f
9
7
.
2
2
%.
T
h
is
m
o
d
el'
s
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
is
d
u
e
to
th
e
in
teg
r
atio
n
o
f
v
ar
ie
d
f
ea
tu
r
es,
wh
ich
en
h
an
ce
s
its
p
r
ed
ictio
n
ac
cu
r
ac
y
an
d
r
eliab
il
ity
.
T
h
e
s
p
ec
tr
o
g
r
am
an
d
MFC
C
in
p
u
t
m
o
d
els
also
p
er
f
o
r
m
we
ll
b
u
t
n
o
t
as
h
ig
h
ly
as
th
e
tr
ip
le
f
ea
tu
r
e
m
o
d
el.
T
h
e
s
p
ec
tr
o
g
r
am
m
o
d
el
ac
h
iev
es
an
ac
cu
r
ac
y
o
f
9
4
.
4
9
%,
p
r
ec
is
io
n
o
f
9
4
.
9
5
%,
r
e
ca
ll
o
f
9
4
.
4
9
%,
an
d
an
F1
-
s
co
r
e
o
f
9
4
.
6
1
%.
Similar
ly
,
th
e
MFC
C
m
o
d
el
r
ec
o
r
d
s
an
ac
cu
r
ac
y
o
f
9
3
.
4
8
%,
p
r
ec
is
io
n
o
f
9
4
.
2
8
%,
r
ec
all
o
f
9
3
.
4
8
%,
an
d
an
F1
-
s
co
r
e
o
f
9
3
.
7
1
%.
T
h
ese
m
o
d
els
o
u
tp
er
f
o
r
m
th
e
tim
e
d
o
m
ain
m
o
d
el
s
ig
n
if
ican
tly
,
h
ig
h
lig
h
tin
g
th
e
ad
v
an
tag
e
o
f
ca
p
tu
r
in
g
f
r
eq
u
en
cy
d
o
m
ain
in
f
o
r
m
atio
n
.
T
h
e
d
u
al
f
ea
t
u
r
e
in
p
u
t
m
o
d
el,
co
m
b
in
in
g
s
p
ec
tr
o
g
r
am
a
n
d
MFC
C
f
ea
tu
r
es,
s
h
o
ws
i
m
p
r
ess
iv
e
p
er
f
o
r
m
an
ce
with
a
n
ac
cu
r
ac
y
o
f
9
5
.
6
5
%,
p
r
ec
is
io
n
o
f
9
4
.
8
5
%,
r
ec
all
o
f
9
6
.
2
1
%,
a
n
d
a
n
F
1
-
s
co
r
e
o
f
9
5
.
5
2
%.
T
h
is
m
o
d
el
o
u
tp
er
f
o
r
m
s
th
e
in
d
iv
id
u
al
s
p
ec
tr
o
g
r
am
an
d
MF
C
C
m
o
d
els
an
d
clo
s
ely
m
atc
h
es
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
tr
ip
le
f
ea
tu
r
e
m
o
d
el.
I
n
teg
r
atin
g
d
if
f
e
r
en
t
f
ea
tu
r
e
ty
p
es
allo
ws
f
o
r
th
e
u
tili
za
tio
n
o
f
co
m
p
lem
en
tar
y
in
f
o
r
m
atio
n
,
r
esu
ltin
g
i
n
im
p
r
o
v
ed
o
v
er
all
p
e
r
f
o
r
m
an
ce
.
4
.
3
.
O
v
er
a
ll r
esu
lt
s
T
ab
le
5
co
m
p
ar
es
th
e
p
er
f
o
r
m
an
ce
o
f
d
if
f
er
e
n
t
f
e
atu
r
e
in
p
u
t
s
ets
f
o
r
lu
n
g
s
o
u
n
d
class
if
icatio
n
.
T
h
e
tim
e
d
o
m
ain
in
p
u
t
m
o
d
el
s
h
o
ws
th
e
lo
west
p
er
f
o
r
m
an
ce
,
with
an
ac
cu
r
ac
y
o
f
8
8
.
8
4
%,
p
r
ec
is
io
n
o
f
8
8
.
6
4
%,
r
ec
all
o
f
8
8
.
8
4
%,
an
d
an
F1
-
s
co
r
e
o
f
8
7
.
9
1
%.
T
h
is
in
d
icate
s
its
lim
ited
ab
ilit
y
to
ca
p
t
u
r
e
co
m
p
le
x
p
a
tter
n
s
,
m
ak
in
g
it th
e
least r
eliab
le
ap
p
r
o
ac
h
.
I
n
co
n
tr
ast,
th
e
tr
i
p
le
f
ea
tu
r
e
in
p
u
t
m
o
d
el,
wh
ic
h
co
m
b
in
es
tim
e
d
o
m
ain
,
s
p
ec
tr
o
g
r
am
,
a
n
d
MFC
C
f
ea
tu
r
es,
ac
h
iev
es
th
e
h
ig
h
est p
er
f
o
r
m
an
ce
.
I
t
r
ec
o
r
d
s
a
n
ac
c
u
r
ac
y
o
f
9
7
.
2
5
%,
a
p
r
ec
is
io
n
o
f
9
7
.
4
0
%,
r
ec
all
o
f
97.
1
0
%,
an
d
an
F1
-
s
co
r
e
o
f
9
7
.
2
2
%.
T
h
is
m
o
d
el'
s
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
is
d
u
e
to
th
e
in
teg
r
atio
n
o
f
v
ar
ie
d
f
ea
tu
r
es,
wh
ich
en
h
an
ce
s
its
p
r
ed
ictio
n
ac
cu
r
ac
y
an
d
r
eliab
il
ity
.
T
h
e
s
p
ec
tr
o
g
r
am
an
d
MFC
C
in
p
u
t
m
o
d
els
also
p
er
f
o
r
m
we
ll
b
u
t
n
o
t
as
h
ig
h
ly
as
th
e
tr
ip
le
f
ea
tu
r
e
m
o
d
el.
T
h
e
s
p
ec
tr
o
g
r
am
m
o
d
el
ac
h
iev
es
an
ac
cu
r
ac
y
o
f
9
4
.
4
9
%,
p
r
ec
is
io
n
o
f
9
4
.
9
5
%,
r
e
ca
ll
o
f
9
4
.
4
9
%,
an
d
an
F1
-
s
co
r
e
o
f
9
4
.
6
1
%.
Similar
ly
,
th
e
MFC
C
m
o
d
el
r
ec
o
r
d
s
an
ac
cu
r
ac
y
o
f
9
3
.
4
8
%,
p
r
ec
is
io
n
o
f
9
4
.
2
8
%,
r
ec
all
o
f
9
3
.
4
8
%,
an
d
an
F1
-
s
co
r
e
o
f
9
3
.
7
1
%
.
T
h
ese
m
o
d
els
o
u
tp
er
f
o
r
m
th
e
tim
e
d
o
m
ain
m
o
d
el
s
ig
n
if
ican
tly
,
h
ig
h
lig
h
tin
g
th
e
ad
v
an
tag
e
o
f
ca
p
tu
r
in
g
f
r
eq
u
en
cy
d
o
m
ain
in
f
o
r
m
atio
n
.
T
h
e
d
u
al
f
ea
t
u
r
e
in
p
u
t
m
o
d
el,
co
m
b
in
in
g
s
p
ec
tr
o
g
r
am
a
n
d
MFC
C
f
ea
tu
r
es,
s
h
o
ws
i
m
p
r
ess
iv
e
p
er
f
o
r
m
an
ce
with
a
n
ac
cu
r
ac
y
o
f
9
5
.
6
5
%,
p
r
ec
is
io
n
o
f
9
4
.
8
5
%,
r
ec
all
o
f
9
6
.
2
1
%,
a
n
d
a
n
F
1
-
s
co
r
e
o
f
9
5
.
5
2
%.
T
h
is
m
o
d
el
o
u
tp
er
f
o
r
m
s
th
e
in
d
iv
id
u
al
s
p
ec
tr
o
g
r
am
an
d
MF
C
C
m
o
d
els
an
d
clo
s
ely
m
atch
es
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
tr
ip
le
f
ea
tu
r
e
m
o
d
el.
I
n
teg
r
atin
g
d
if
f
e
r
en
t
f
ea
tu
r
e
ty
p
es
allo
ws
f
o
r
th
e
u
ti
lizatio
n
o
f
co
m
p
lem
en
tar
y
in
f
o
r
m
atio
n
,
r
esu
ltin
g
i
n
im
p
r
o
v
ed
o
v
er
all
p
e
r
f
o
r
m
an
ce
.
T
h
e
tr
ip
le
f
ea
t
u
r
e
m
o
d
el
in
co
r
p
o
r
atin
g
tim
e
d
o
m
ain
,
s
p
ec
tr
o
g
r
am
,
an
d
MFC
C
em
er
g
ed
as
th
e
to
p
p
er
f
o
r
m
er
with
an
ac
c
u
r
ac
y
o
f
9
7
.
2
5
%
in
th
is
s
tu
d
y
th
at
ass
es
s
es
v
ar
io
u
s
f
ea
tu
r
e
in
p
u
ts
f
o
r
lu
n
g
s
o
u
n
d
class
if
icatio
n
.
T
h
is
m
o
d
el
s
h
o
ws
s
tr
o
n
g
p
r
ec
is
io
n
an
d
r
ec
all
m
etr
ics
an
d
p
er
f
o
r
m
s
well
in
all
ca
teg
o
r
ies
(
'
No
r
m
al'
,
'C
r
ac
k
les
'
,
'W
h
ee
ze
s
'
,
an
d
co
m
b
in
atio
n
s
)
.
T
h
e
t
im
e
d
o
m
ain
m
o
d
el,
o
n
th
e
o
th
er
h
an
d
,
h
as
th
e
lo
west
ac
cu
r
ac
y
o
f
a
n
y
m
o
d
e
l,
wh
ich
is
8
8
.
8
4
%,
an
d
f
in
d
s
it
d
if
f
icu
lt
to
d
is
tin
g
u
is
h
b
et
wee
n
v
ar
io
u
s
lu
n
g
so
u
n
d
ab
n
o
r
m
alities
d
u
e
t
o
its
lim
ited
ab
ilit
y
to
ca
p
tu
r
e
in
tr
ic
ate
f
r
eq
u
e
n
cy
an
d
tem
p
o
r
al
f
lu
ctu
atio
n
s
.
I
n
co
m
p
a
r
is
o
n
,
s
in
g
le
-
in
p
u
t
m
o
d
els
s
u
ch
as
s
p
ec
tr
o
g
r
am
(
9
4
.
4
9
%
ac
cu
r
ac
y
)
a
n
d
MFC
C
(
9
3
.
4
8
%
ac
cu
r
ac
y
)
p
er
f
o
r
m
b
etter
in
ce
r
tain
ar
ea
s
.
Fo
r
e
x
am
p
le,
s
p
ec
tr
o
g
r
am
is
b
etter
at
ca
p
t
u
r
in
g
f
r
eq
u
en
c
y
d
ata
o
v
e
r
tim
e,
wh
ile
MFC
C
m
ak
es
u
s
e
o
f
s
p
ec
tr
al
f
ea
tu
r
es,
b
u
t
it
m
ig
h
t
b
e
less
s
u
cc
es
s
f
u
l
at
ca
p
tu
r
in
g
s
u
b
tle
tim
e
-
f
r
eq
u
e
n
cy
r
eso
lu
tio
n
s
.
T
h
ese
r
esu
lts
ar
e
co
n
f
ir
m
ed
b
y
ea
r
l
ier
r
esear
ch
,
in
clu
d
in
g
wo
r
k
s
b
y
[
2
3
]
an
d
[
2
4
]
,
wh
ich
d
em
o
n
s
tr
ate
h
o
w
well
s
p
ec
tr
o
g
r
a
m
p
er
f
o
r
m
s
in
c
o
m
p
ar
is
o
n
to
MFC
C
in
task
s
s
im
i
l
ar
to
class
if
icatio
n
.
T
h
e
in
teg
r
atio
n
o
f
m
u
ltip
le
f
e
atu
r
es
in
th
e
d
u
al
f
ea
tu
r
e
m
o
d
el
(
co
m
b
in
i
n
g
s
p
ec
t
r
o
g
r
am
a
n
d
MFC
C
)
ac
h
iev
es
an
ac
cu
r
ac
y
o
f
9
5
.
6
5
%,
s
u
r
p
ass
in
g
s
in
g
le
-
f
ea
tu
r
e
m
o
d
els
b
y
lev
er
a
g
in
g
th
e
ir
co
m
p
lem
en
tar
y
s
tr
en
g
th
s
.
T
h
is
m
eth
o
d
is
co
n
s
is
ten
t
with
r
esear
ch
b
y
[
2
5
]
,
wh
ich
s
h
o
wed
th
at
m
er
g
in
g
s
p
ec
tr
o
g
r
am
a
n
d
MFC
C
ch
ar
ac
ter
is
tics
in
cr
ea
s
ed
ac
cu
r
ac
y
wh
en
c
o
m
p
ar
e
d
t
o
u
tili
zin
g
th
em
s
ep
a
r
ately
.
Al
l
th
in
g
s
co
n
s
id
er
ed
,
th
e
m
o
d
el'
s
ca
p
ac
ity
to
p
r
ec
is
ely
ca
teg
o
r
ize
in
tr
icate
lu
n
g
s
o
u
n
d
p
atter
n
s
is
im
p
r
o
v
e
d
b
y
th
e
in
teg
r
atio
n
o
f
tim
e
d
o
m
ain
,
s
p
ec
tr
o
g
r
am
,
a
n
d
MFC
C
f
ea
tu
r
es,
wh
ic
h
m
ak
es
it
a
p
r
o
m
i
s
in
g
d
e
v
elo
p
m
en
t
f
o
r
r
esp
ir
ato
r
y
h
ea
lth
ca
r
e
ap
p
licatio
n
s
.
T
ab
le
5
.
C
o
m
p
a
r
is
o
n
o
f
weig
h
ted
av
er
ag
e
ac
r
o
s
s
d
is
tin
ct
f
ea
tu
r
e
in
p
u
t sets
F
e
a
t
u
r
e
i
n
p
u
t
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
S
c
o
r
e
(
%)
Ti
me
d
o
m
a
i
n
8
8
.
8
4
8
8
.
6
4
8
8
.
8
4
8
7
.
9
1
S
p
e
c
t
r
o
g
r
a
m
9
4
.
4
9
9
4
.
9
5
9
4
.
4
9
9
4
.
6
1
M
F
C
C
9
3
.
4
8
9
4
.
2
8
9
3
.
4
8
9
3
.
7
1
D
u
a
l
f
e
a
t
u
r
e
i
n
p
u
t
9
5
.
6
5
9
4
.
8
5
9
6
.
2
1
9
5
.
5
2
Tr
i
p
l
e
f
e
a
t
u
r
e
i
n
p
u
t
9
7
.
2
5
9
7
.
4
0
9
7
.
1
0
9
7
.
2
2
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
-
8
7
0
8
I
n
teg
r
a
tin
g
time
-
fr
eq
u
en
cy
fea
tu
r
es w
ith
d
ee
p
lea
r
n
in
g
fo
r
lu
n
g
s
o
u
n
d
cla
s
s
ifica
tio
n
…
(
S
u
Yu
a
n
C
h
a
n
g
)
3745
5.
CO
NCLU
SI
O
N
I
n
co
n
cl
u
s
io
n
,
th
is
s
tu
d
y
s
y
s
te
m
atica
lly
ev
alu
ated
th
e
p
er
f
o
r
m
an
ce
o
f
d
if
f
er
e
n
t
f
ea
tu
r
e
i
n
p
u
ts
with
in
th
e
ad
o
p
te
d
AL
SD
-
Net
f
o
r
class
if
y
in
g
lu
n
g
s
o
u
n
d
s
.
T
h
e
tr
ip
le
f
ea
tu
r
e
m
o
d
el,
c
o
m
b
i
n
in
g
tim
e
d
o
m
ain
,
s
p
ec
tr
o
g
r
am
,
an
d
MFC
C
,
p
er
f
o
r
m
ed
th
e
b
est
with
9
7
.
2
5
%
ac
cu
r
ac
y
.
T
h
e
tim
e
d
o
m
ai
n
m
o
d
el,
o
n
th
e
o
t
h
e
r
h
an
d
,
s
tr
u
g
g
led
with
t
h
e
in
tr
ic
ate
f
r
eq
u
e
n
cy
an
d
tim
in
g
p
att
er
n
s
in
lu
n
g
s
o
u
n
d
s
,
y
ield
in
g
th
e
lo
west
ac
cu
r
ac
y
o
f
8
8
.
8
4
%.
W
h
ile
MFC
C
u
s
ed
s
p
ec
tr
u
m
c
h
ar
ac
ter
is
tics
b
u
t
s
tr
u
g
g
led
with
d
etailed
ti
m
in
g
,
s
p
ec
tr
o
g
r
am
(
9
4
.
4
9
%
ac
cu
r
ac
y
)
an
d
MFC
C
(
9
3
.
4
8
%
ac
cu
r
ac
y
)
wer
e
two
ex
am
p
les
o
f
s
in
g
l
e
-
f
ea
tu
r
e
m
o
d
els
th
at
d
em
o
n
s
tr
ated
s
tr
en
g
th
s
in
p
a
r
ticu
lar
ar
ea
s
.
Sp
ec
tr
o
g
r
am
wa
s
p
ar
ticu
lar
ly
g
o
o
d
at
ca
p
tu
r
in
g
tim
e
-
f
r
eq
u
en
cy
d
y
n
am
ics.
Fu
r
th
er
m
o
r
e
,
th
e
r
esear
ch
h
i
g
h
lig
h
ts
th
e
n
o
tewo
r
th
y
b
en
ef
it
o
f
co
m
b
in
in
g
s
ev
er
al
f
e
atu
r
es,
as
d
em
o
n
s
tr
ated
b
y
th
e
d
u
al
f
ea
tu
r
e
m
o
d
el
(
c
o
m
b
in
i
n
g
s
p
ec
tr
o
g
r
am
an
d
MFC
C
)
th
at
attain
s
a
9
5
.
6
5
%
ac
cu
r
ac
y
r
ate,
u
tili
zin
g
co
m
p
lem
en
tar
y
ad
v
an
tag
es
to
im
p
r
o
v
e
class
if
icatio
n
ca
p
ab
ilit
y
.
T
h
e
b
est
f
e
atu
r
e
co
m
b
i
n
atio
n
s
f
o
r
en
h
an
cin
g
l
u
n
g
s
o
u
n
d
clas
s
if
icatio
n
ac
cu
r
ac
y
ar
e
s
h
o
w
n
b
y
th
ese
s
tu
d
ies,
wh
ich
ad
v
a
n
ce
th
e
f
ield
.
T
h
e
y
ar
e
co
n
s
is
ten
t
with
ea
r
lier
s
tu
d
ies
s
h
o
win
g
s
p
ec
tr
o
g
r
am
s
to
b
e
m
o
r
e
ef
f
ec
tiv
e
th
an
M
FC
C
in
co
m
p
ar
a
b
le
task
s
,
co
n
f
ir
m
in
g
th
e
s
tr
ateg
y
o
f
u
s
in
g
f
ea
tu
r
e
r
ich
n
ess
to
im
p
r
o
v
e
d
iag
n
o
s
tic
p
r
ec
is
io
n
in
r
esp
ir
ato
r
y
ca
r
e.
T
o
fu
r
th
er
im
p
r
o
v
e
a
n
d
v
alid
ate
th
ese
f
in
d
in
g
s
f
o
r
wid
er
clin
i
ca
l
ap
p
licatio
n
s
,
f
u
tu
r
e
ef
f
o
r
t
s
co
u
ld
in
v
esti
g
ate
h
y
b
r
id
f
ea
tu
r
e
co
m
b
in
atio
n
s
an
d
lar
g
er
d
atasets
,
u
ltima
tely
en
h
a
n
cin
g
th
e
ca
p
ab
ilit
ies
o
f
au
t
o
m
ated
l
u
n
g
s
o
u
n
d
an
al
y
s
is
s
y
s
tem
s
.
ACK
NO
WL
E
DG
M
E
N
T
S
A
u
th
o
r
s
h
av
e
n
o
ac
k
n
o
wled
g
m
en
ts
to
d
ec
lar
e.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
C
h
an
g
Su
Yu
an
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Ma
r
n
i A
zir
a
Ma
r
k
o
m
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
C
h
o
o
n
g
Z
h
i Sh
en
g
✓
✓
✓
✓
✓
✓
Ar
n
i M
u
n
ir
a
Ma
r
k
o
m
✓
✓
✓
✓
L
atif
ah
Mu
n
ir
a
h
Kam
ar
u
d
in
✓
✓
✓
E
r
d
y
Su
lin
o
Mo
h
d
Mu
s
lim
T
an
✓
✓
✓
C
:
C
o
n
ce
p
tu
aliza
tio
n
M
:
M
eth
o
d
o
lo
g
y
So
:
So
f
twar
e
Va
:
Va
lid
atio
n
Fo
:
Fo
r
m
al
an
aly
s
is
I
:
I
n
v
esti
g
atio
n
R
:
R
eso
u
r
ce
s
D
:
D
ata
C
u
r
atio
n
O
:
W
r
itin
g
-
O
r
ig
in
al
Dr
af
t
E
:
W
r
itin
g
-
R
ev
iew
&
E
d
itin
g
Vi
:
Vi
s
u
aliza
tio
n
Su
:
Su
p
er
v
is
io
n
P
:
P
r
o
ject
ad
m
in
is
tr
atio
n
Fu
:
Fu
n
d
in
g
ac
q
u
is
itio
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
I
NF
O
RM
E
D
CO
NS
E
N
T
No
in
f
o
r
m
ed
c
o
n
s
en
t
was
r
e
q
u
ir
ed
,
as
th
e
d
ata
u
s
ed
in
t
h
is
s
tu
d
y
wer
e
o
b
tain
ed
f
r
o
m
p
u
b
licly
av
ailab
le
d
atab
ases
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
4
,
Au
g
u
s
t
20
25
:
3
7
3
7
-
3747
3746
E
T
H
I
CAL AP
P
RO
V
AL
No
eth
ical
ap
p
r
o
v
al
was r
eq
u
i
r
ed
f
o
r
th
is
s
tu
d
y
.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
i
n
g
s
o
f
th
is
s
tu
d
y
ar
e
o
p
en
ly
a
v
ailab
le
in
th
e
I
C
B
HI
R
esp
ir
ato
r
y
So
u
n
d
Data
b
ase
at
h
ttp
s
://d
o
i.o
r
g
/1
0
.
1
0
8
8
/1
3
6
1
-
6
5
7
9
/ab
0
3
ea
,
as d
e
s
cr
ib
ed
in
R
o
ch
a
et
a
l.
(
2
0
1
9
)
[
1
8
]
.
RE
F
E
R
E
NC
E
S
[
1
]
W
.
G
u
a
n
e
t
a
l
.
,
“
C
l
i
n
i
c
a
l
c
h
a
r
a
c
t
e
r
i
st
i
c
s
o
f
C
o
r
o
n
a
v
i
r
u
s
D
i
s
e
a
se
2
0
1
9
i
n
C
h
i
n
a
,
”
N
e
w
E
n
g
l
a
n
d
J
o
u
r
n
a
l
o
f
M
e
d
i
c
i
n
e
,
v
o
l
.
3
8
2
,
n
o
.
1
8
,
p
p
.
1
7
0
8
–
1
7
2
0
,
A
p
r
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
5
6
/
N
EJ
M
o
a
2
0
0
2
0
3
2
.
[
2
]
C
.
H
u
a
n
g
e
t
a
l
.
,
“
C
l
i
n
i
c
a
l
f
e
a
t
u
r
e
s
o
f
p
a
t
i
e
n
t
s
i
n
f
e
c
t
e
d
w
i
t
h
2
0
1
9
n
o
v
e
l
c
o
r
o
n
a
v
i
r
u
s
i
n
W
u
h
a
n
,
C
h
i
n
a
,
”
T
h
e
L
a
n
c
e
t
,
v
o
l
.
3
9
5
,
n
o
.
1
0
2
2
3
,
p
p
.
4
9
7
–
5
0
6
,
F
e
b
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
S
0
1
4
0
-
6
7
3
6
(
2
0
)
3
0
1
8
3
-
5.
[
3
]
D
.
C
h
a
mb
e
r
l
a
i
n
,
J.
M
o
f
o
r
,
R
.
F
l
e
t
c
h
e
r
,
a
n
d
R
.
K
o
d
g
u
l
e
,
“
M
o
b
i
l
e
st
e
t
h
o
s
c
o
p
e
a
n
d
si
g
n
a
l
p
r
o
c
e
ss
i
n
g
a
l
g
o
r
i
t
h
ms
f
o
r
p
u
l
m
o
n
a
r
y
scree
n
i
n
g
a
n
d
d
i
a
g
n
o
st
i
c
s
,
”
i
n
2
0
1
5
I
EEE
G
l
o
b
a
l
H
u
m
a
n
i
t
a
r
i
a
n
T
e
c
h
n
o
l
o
g
y
C
o
n
f
e
r
e
n
c
e
(
G
H
T
C
)
,
O
c
t
.
2
0
1
5
,
p
p
.
3
8
5
–
3
9
2
,
d
o
i
:
1
0
.
1
1
0
9
/
G
H
TC
.
2
0
1
5
.
7
3
4
4
0
0
1
.
[
4
]
H
.
C
h
o
i
e
t
a
l
.
,
“
E
x
t
e
n
s
i
o
n
o
f
C
o
r
o
n
a
v
i
r
u
s
D
i
s
e
a
s
e
2
0
1
9
o
n
c
h
e
s
t
C
T
a
n
d
i
m
p
l
i
c
a
t
i
o
n
s
f
o
r
c
h
e
s
t
r
a
d
i
o
g
r
a
p
h
i
c
i
n
t
e
r
p
r
e
t
a
t
i
o
n
,
”
Ra
d
i
o
l
o
g
y
:
C
a
r
d
i
o
t
h
o
r
a
c
i
c
I
m
a
g
i
n
g
,
v
o
l
.
2
,
n
o
.
2
,
p
.
e
2
0
0
1
0
7
,
A
p
r
.
2
0
2
0
,
d
o
i
:
1
0
.
1
1
4
8
/
r
y
c
t
.
2
0
2
0
2
0
0
1
0
7
.
[
5
]
Y.
-
C
.
W
u
e
t
a
l
.
,
“
D
e
v
e
l
o
p
me
n
t
o
f
a
n
El
e
c
t
r
o
n
i
c
S
t
e
t
h
o
s
c
o
p
e
a
n
d
a
c
l
a
ssi
f
i
c
a
t
i
o
n
a
l
g
o
r
i
t
h
m
f
o
r
c
a
r
d
i
o
p
u
l
m
o
n
a
r
y
s
o
u
n
d
s,”
S
e
n
s
o
rs
,
v
o
l
.
2
2
,
n
o
.
1
1
,
p
.
4
2
6
3
,
J
u
n
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
2
1
1
4
2
6
3
.
[
6
]
J.
J.
S
e
a
h
,
J
.
Zh
a
o
,
D
.
Y
.
W
a
n
g
,
a
n
d
H
.
P
.
L
e
e
,
“
R
e
v
i
e
w
o
n
t
h
e
a
d
v
a
n
c
e
me
n
t
s
o
f
s
t
e
t
h
o
sc
o
p
e
t
y
p
e
s
i
n
c
h
e
st
a
u
s
c
u
l
t
a
t
i
o
n
,
”
D
i
a
g
n
o
s
t
i
c
s
,
v
o
l
.
1
3
,
n
o
.
9
,
p
.
1
5
4
5
,
A
p
r
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
d
i
a
g
n
o
s
t
i
c
s
1
3
0
9
1
5
4
5
.
[
7
]
C
.
K
o
n
i
n
g
a
n
d
A
.
L
o
c
k
,
“
A
s
y
st
e
ma
t
i
c
r
e
v
i
e
w
a
n
d
u
t
i
l
i
z
a
t
i
o
n
s
t
u
d
y
o
f
d
i
g
i
t
a
l
s
t
e
t
h
o
sc
o
p
e
s
f
o
r
c
a
r
d
i
o
p
u
l
mo
n
a
r
y
a
ss
e
ssm
e
n
t
s,
”
J
o
u
rn
a
l
o
f
Me
d
i
c
a
l
Re
se
a
r
c
h
a
n
d
I
n
n
o
v
a
t
i
o
n
,
v
o
l
.
5
,
n
o
.
2
,
p
p
.
4
–
1
4
,
F
e
b
.
2
0
2
2
,
d
o
i
:
1
0
.
2
5
2
5
9
/
JM
R
I
_
2
_
2
0
2
1
.
[
8
]
B
.
A
b
e
r
a
T
e
ssem
a
,
H
.
N
e
m
o
mss
a
,
a
n
d
G
.
La
m
e
s
g
i
n
S
i
me
g
n
,
“
A
c
q
u
i
s
i
t
i
o
n
a
n
d
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
Lu
n
g
s
o
u
n
d
s
f
o
r
i
mp
r
o
v
i
n
g
t
h
e
e
f
f
i
c
a
c
y
o
f
a
u
s
c
u
l
t
a
t
i
o
n
d
i
a
g
n
o
si
s
o
f
p
u
l
m
o
n
a
r
y
d
i
se
a
ses,
”
M
e
d
i
c
a
l
D
e
v
i
c
e
s:
Ev
i
d
e
n
c
e
a
n
d
R
e
se
a
rc
h
,
v
o
l
.
V
o
l
u
m
e
1
5
,
p
p
.
8
9
–
1
0
2
,
A
p
r
.
2
0
2
2
,
d
o
i
:
1
0
.
2
1
4
7
/
M
D
ER
.
S
3
6
2
4
0
7
.
[
9
]
H
.
C
h
a
n
a
n
e
a
n
d
M
.
B
a
h
o
u
r
a
,
“
C
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
-
b
a
s
e
d
m
o
d
e
l
f
o
r
L
u
n
g
s
o
u
n
d
s
c
l
a
ssi
f
i
c
a
t
i
o
n
,
”
i
n
2
0
2
1
I
EE
E
I
n
t
e
r
n
a
t
i
o
n
a
l
Mi
d
w
e
s
t
S
y
m
p
o
s
i
u
m
o
n
C
i
rc
u
i
t
s
a
n
d
S
y
st
e
m
s
(
MW
S
C
AS
)
,
A
u
g
.
2
0
2
1
,
p
p
.
5
5
5
–
5
5
8
,
d
o
i
:
1
0
.
1
1
0
9
/
M
W
S
C
A
S
4
7
6
7
2
.
2
0
2
1
.
9
5
3
1
8
8
7
.
[
1
0
]
G
.
P
e
t
me
z
a
s
e
t
a
l
.
,
“
A
u
t
o
m
a
t
e
d
L
u
n
g
so
u
n
d
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
si
n
g
a
h
y
b
r
i
d
C
N
N
-
LSTM
n
e
t
w
o
r
k
a
n
d
f
o
c
a
l
l
o
ss
f
u
n
c
t
i
o
n
,
”
S
e
n
s
o
rs
,
v
o
l
.
2
2
,
n
o
.
3
,
p
.
1
2
3
2
,
F
e
b
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
2
0
3
1
2
3
2
.
[
1
1
]
N
.
B
a
g
h
e
l
,
V
.
N
a
n
g
i
a
,
a
n
d
M
.
K
.
D
u
t
t
a
,
“
A
L
SD
-
N
e
t
:
A
u
t
o
m
a
t
i
c
l
u
n
g
s
o
u
n
d
s
d
i
a
g
n
o
s
i
s
n
e
t
w
o
r
k
f
r
o
m
p
u
l
m
o
n
a
r
y
si
g
n
a
l
s,”
N
e
u
r
a
l
C
o
m
p
u
t
i
n
g
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
3
3
,
n
o
.
2
4
,
p
p
.
1
7
1
0
3
–
1
7
1
1
8
,
D
e
c
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
5
2
1
-
021
-
0
6
3
0
2
-
1.
[
1
2
]
M
.
Z
h
a
n
g
,
M
.
Li
,
L.
G
u
o
,
a
n
d
J.
L
i
u
,
“
A
l
o
w
-
c
o
st
A
I
-
e
mp
o
w
e
r
e
d
s
t
e
t
h
o
s
c
o
p
e
a
n
d
a
l
i
g
h
t
w
e
i
g
h
t
mo
d
e
l
f
o
r
d
e
t
e
c
t
i
n
g
c
a
r
d
i
a
c
a
n
d
r
e
sp
i
r
a
t
o
r
y
d
i
s
e
a
s
e
s
f
r
o
m
l
u
n
g
a
n
d
h
e
a
r
t
a
u
sc
u
l
t
a
t
i
o
n
s
o
u
n
d
s,
”
S
e
n
so
rs
,
v
o
l
.
2
3
,
n
o
.
5
,
p
.
2
5
9
1
,
F
e
b
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
3
0
5
2
5
9
1
.
[
1
3
]
Y
.
S
h
i
,
Y
.
L
i
,
M
.
C
a
i
,
a
n
d
X
.
D
.
Z
h
a
n
g
,
“
A
L
u
n
g
so
u
n
d
c
a
t
e
g
o
r
y
r
e
c
o
g
n
i
t
i
o
n
met
h
o
d
b
a
s
e
d
o
n
w
a
v
e
l
e
t
d
e
c
o
mp
o
si
t
i
o
n
a
n
d
B
P
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
Bi
o
l
o
g
i
c
a
l
S
c
i
e
n
c
e
s
,
v
o
l
.
1
5
,
n
o
.
1
,
p
p
.
1
9
5
–
2
0
7
,
2
0
1
9
,
d
o
i
:
1
0
.
7
1
5
0
/
i
j
b
s
.
2
9
8
6
3
.
[
1
4
]
K
.
N
.
La
l
,
“
A
l
u
n
g
so
u
n
d
r
e
c
o
g
n
i
t
i
o
n
mo
d
e
l
t
o
d
i
a
g
n
o
ses
t
h
e
r
e
s
p
i
r
a
t
o
r
y
d
i
sea
s
e
s
b
y
u
s
i
n
g
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
,
”
Mu
l
t
i
m
e
d
i
a
T
o
o
l
s
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
8
2
,
n
o
.
2
3
,
p
p
.
3
6
6
1
5
–
3
6
6
3
1
,
S
e
p
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
0
4
2
-
0
2
3
-
1
4
7
2
7
-
0.
[
1
5
]
B
.
V
i
d
h
y
a
,
M
.
N
i
k
h
i
l
M
a
d
h
a
v
,
M
.
S
u
r
e
s
h
K
u
mar,
a
n
d
S
.
K
a
l
a
n
a
n
d
i
n
i
,
“
A
I
b
a
sed
d
i
a
g
n
o
s
i
s
o
f
P
n
e
u
mo
n
i
a
,
”
Wi
re
l
e
ss
Pe
rso
n
a
l
C
o
m
m
u
n
i
c
a
t
i
o
n
s
,
v
o
l
.
1
2
6
,
n
o
.
4
,
p
p
.
3
6
7
7
–
3
6
9
2
,
O
c
t
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
2
7
7
-
0
2
2
-
0
9
8
8
5
-
7.
[
1
6
]
J.
A
m
o
se,
P
.
M
a
n
i
me
g
a
l
a
i
,
a
n
d
R
.
P
o
n
S
e
l
c
h
i
y
a
,
“
C
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
a
d
v
e
n
t
i
t
i
o
u
s
L
u
n
g
so
u
n
d
s
:
W
h
e
e
z
e
,
C
r
a
c
k
l
e
u
si
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s,”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
S
y
s
t
e
m
s a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
i
n
E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
1
,
n
o
.
3
,
p
p
.
1
1
4
3
–
1
1
5
2
,
2
0
2
3
.
[
1
7
]
M
.
F
r
a
i
w
a
n
,
L
.
F
r
a
i
w
a
n
,
M
.
A
l
k
h
o
d
a
r
i
,
a
n
d
O
.
H
a
ss
a
n
i
n
,
“
R
e
c
o
g
n
i
t
i
o
n
o
f
p
u
l
mo
n
a
r
y
d
i
se
a
ses
f
r
o
m
l
u
n
g
s
o
u
n
d
s
u
si
n
g
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
a
n
d
l
o
n
g
sh
o
r
t
-
t
e
r
m
m
e
m
o
r
y
,
”
J
o
u
r
n
a
l
o
f
Am
b
i
e
n
t
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
H
u
m
a
n
i
z
e
d
C
o
m
p
u
t
i
n
g
,
v
o
l
.
1
3
,
n
o
.
1
0
,
p
p
.
4
7
5
9
–
4
7
7
1
,
O
c
t
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
2
6
5
2
-
0
2
1
-
0
3
1
8
4
-
y.
[
1
8
]
B
.
M
.
R
o
c
h
a
e
t
a
l
.
,
“
A
n
o
p
e
n
a
c
c
e
ss
d
a
t
a
b
a
se
f
o
r
t
h
e
e
v
a
l
u
a
t
i
o
n
o
f
r
e
sp
i
r
a
t
o
r
y
so
u
n
d
c
l
a
s
si
f
i
c
a
t
i
o
n
a
l
g
o
r
i
t
h
ms,
”
P
h
y
si
o
l
o
g
i
c
a
l
Me
a
su
r
e
m
e
n
t
,
v
o
l
.
4
0
,
n
o
.
3
,
p
.
0
3
5
0
0
1
,
M
a
r
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
8
8
/
1
3
6
1
-
6
5
7
9
/
a
b
0
3
e
a
.
[
1
9
]
B
.
A
c
a
r
d
e
m
i
r
c
i
,
Y
.
K
o
ç
y
i
ğ
i
t
,
D
.
K
i
z
i
l
i
r
ma
k
,
a
n
d
Y
.
H
a
v
l
u
c
u
,
“
A
d
v
e
n
t
i
t
i
o
u
s
a
n
d
n
o
r
mal
r
e
s
p
i
r
a
t
o
r
y
s
o
u
n
d
a
n
a
l
y
s
i
s
w
i
t
h
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
m
e
t
h
o
d
s,
”
C
e
l
a
l
Ba
y
a
r
Ü
n
i
v
e
rsi
t
e
s
i
Fe
n
B
i
l
i
m
l
e
ri
D
e
r
g
i
s
i
,
v
o
l
.
1
8
,
n
o
.
2
,
p
p
.
1
6
9
–
1
8
0
,
Ju
n
.
2
0
2
1
,
d
o
i
:
1
0
.
1
8
4
6
6
/
c
b
a
y
a
r
f
b
e
.
1
0
0
2
9
1
7
.
[
2
0
]
F
.
D
e
m
i
r
,
A
.
S
e
n
g
u
r
,
a
n
d
V
.
B
a
j
a
j
,
“
C
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
b
a
se
d
e
f
f
i
c
i
e
n
t
a
p
p
r
o
a
c
h
f
o
r
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
l
u
n
g
d
i
s
e
a
s
e
s,”
H
e
a
l
t
h
I
n
f
o
rm
a
t
i
o
n
S
c
i
e
n
c
e
a
n
d
S
y
st
e
m
s
,
v
o
l
.
8
,
n
o
.
1
,
p
.
4
,
D
e
c
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
7
5
5
-
0
1
9
-
0
0
9
1
-
3.
[
2
1
]
C
.
P
e
r
l
i
c
h
,
“
Le
a
r
n
i
n
g
c
u
r
v
e
s
i
n
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
,
”
i
n
E
n
c
y
c
l
o
p
e
d
i
a
o
f
Ma
c
h
i
n
e
L
e
a
rn
i
n
g
,
B
o
st
o
n
,
M
A
:
S
p
r
i
n
g
e
r
U
S
,
2
0
1
1
,
p
p
.
5
7
7
–
5
8
0
.
[
2
2
]
S
.
TA
N
,
“
N
e
i
g
h
b
o
r
-
w
e
i
g
h
t
e
d
K
-
n
e
a
r
e
st
n
e
i
g
h
b
o
r
f
o
r
u
n
b
a
l
a
n
c
e
d
t
e
x
t
c
o
r
p
u
s,”
Ex
p
e
r
t
S
y
st
e
m
s
w
i
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
2
8
,
n
o
.
4
,
p
p
.
6
6
7
–
6
7
1
,
M
a
y
2
0
0
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
sw
a
.
2
0
0
4
.
1
2
.
0
2
3
.
[
2
3
]
Z.
W
a
n
g
a
n
d
Z.
S
u
n
,
“
P
e
r
f
o
r
ma
n
c
e
e
v
a
l
u
a
t
i
o
n
o
f
l
u
n
g
s
o
u
n
d
s
c
l
a
ssi
f
i
c
a
t
i
o
n
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
u
n
d
e
r
v
a
r
i
a
b
l
e
p
a
r
a
m
e
t
e
r
s,”
EU
RAS
I
P
J
o
u
rn
a
l
o
n
Ad
v
a
n
c
e
s
i
n
S
i
g
n
a
l
Pr
o
c
e
ss
i
n
g
,
v
o
l
.
2
0
2
4
,
n
o
.
1
,
p
.
5
1
,
A
p
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
1
8
6
/
s
1
3
6
3
4
-
024
-
0
1
1
4
8
-
w.
[
2
4
]
L.
D
.
M
a
n
g
,
F
.
D
.
G
.
M
a
r
t
í
n
e
z
,
D
.
M
.
M
u
ñ
o
z
,
S
.
G
.
G
a
l
á
n
,
a
n
d
R
.
C
o
r
t
i
n
a
,
“
C
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
a
d
v
e
n
t
i
t
i
o
u
s
s
o
u
n
d
s
c
o
mb
i
n
i
n
g
c
o
c
h
l
e
o
g
r
a
m a
n
d
v
i
s
i
o
n
t
r
a
n
sf
o
r
m
e
r
s,
”
S
e
n
so
rs
,
v
o
l
.
2
4
,
n
o
.
2
,
p
.
6
8
2
,
Ja
n
.
2
0
2
4
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
4
0
2
0
6
8
2
.
[
2
5
]
S.
-
Y
.
Ju
n
g
,
C
.
-
H
.
L
i
a
o
,
Y
.
-
S
.
W
u
,
S
.
-
M
.
Y
u
a
n
,
a
n
d
C
.
-
T.
S
u
n
,
“
Ef
f
i
c
i
e
n
t
l
y
c
l
a
ss
i
f
y
i
n
g
l
u
n
g
s
o
u
n
d
s
t
h
r
o
u
g
h
d
e
p
t
h
w
i
s
e
se
p
a
r
a
b
l
e
C
N
N
mo
d
e
l
s
w
i
t
h
f
u
s
e
d
S
TFT
a
n
d
M
F
C
C
f
e
a
t
u
r
e
s,
”
D
i
a
g
n
o
s
t
i
c
s
,
v
o
l
.
1
1
,
n
o
.
4
,
p
.
7
3
2
,
A
p
r
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
d
i
a
g
n
o
s
t
i
c
s1
1
0
4
0
7
3
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.