T
E
L
K
O
MN
I
K
A
T
el
eco
m
m
u
n
i
ca
t
i
o
n
,
C
o
m
p
u
t
i
n
g
,
E
l
ect
ro
n
i
cs
a
n
d
C
o
n
t
ro
l
Vo
l
.
1
9
, N
o
.
5
,
O
ct
o
b
er
202
1
, p
p
.
1581
~
1587
I
SSN
:
1693
-
6930,
a
c
c
r
e
di
t
e
d F
i
r
s
t
G
r
a
de
by K
e
m
e
nr
i
s
t
e
kdi
kt
i
,
D
e
c
r
e
e
N
o:
21/
E
/
K
P
T
/
2018
D
O
I
:
10.
12928/
T
E
L
K
O
M
N
I
K
A
.
v1
9
i
5
.
18535
1581
Jou
r
n
al
h
om
e
page
:
ht
t
p:
/
/
j
our
nal
.
uad
.
ac
.
i
d/
i
nde
x
.
php/
T
E
L
K
O
M
N
I
K
A
I
m
pro
v
em
ent
o
n
KN
N
us
i
ng
g
ene
t
i
c a
l
g
o
ri
t
hm
a
nd
co
m
bi
n
ed
f
ea
t
ur
e
ex
t
ra
ct
i
o
n
t
o
id
e
n
t
if
y
CO
VI
D
-
1
9
s
uf
f
e
re
rs
ba
s
e
d o
n
CT
s
c
an
i
m
age
R
ad
i
t
yo A
d
i
N
u
gr
oh
o
,
A
r
i
e
S
ap
t
a N
u
gr
ah
a
,
Ay
l
wi
n
A
l
R
as
yi
d
,
F
e
n
n
y Wi
n
d
a R
ah
ayu
De
pa
r
tm
e
n
t of
C
om
pu
te
r
S
c
ie
nc
e
,
F
a
c
u
lt
y
of
M
a
the
m
a
ti
c
s
a
nd
Na
t
ur
a
l
S
c
ie
nc
e
s
,
L
a
m
bu
ng
M
a
ng
kur
a
t
U
ni
ve
r
s
it
y,
Ban
j
arma
s
i
n
,
I
nd
one
sia
A
rt
i
cl
e I
n
f
o
AB
S
T
RACT
A
r
tic
le
h
is
to
r
y
:
R
ecei
v
ed
N
ov 13
,
2020
R
ev
i
s
ed
J
ul
19
,
2021
A
ccep
t
ed
A
ug 5
,
2021
C
or
on
a
v
ir
u
s d
ise
a
se
2
01
9 (
CO
V
ID
-
19
)
ha
s
spr
e
a
d t
hr
o
u
gho
ut
th
e
w
or
l
d.
The
de
te
c
ti
on of
th
is d
ise
a
se
is u
sua
ll
y c
a
r
r
ie
d o
ut u
si
ng the
r
e
ve
r
se
tr
a
nsc
r
ip
ta
se
pol
ym
e
r
a
se
c
ha
i
n r
e
a
c
ti
on
(
RT
-
P
C
R
)
s
wa
b
te
s
t.
Ho
we
ve
r
,
lim
ite
d r
e
so
ur
c
e
s
be
c
a
m
e
a
n ob
sta
c
le
t
o c
a
r
r
yi
ng o
ut
the
m
a
s
si
ve
te
s
t.
T
o so
lve
t
hi
s pr
o
ble
m
,
c
om
p
ute
r
iz
e
d t
om
o
gr
a
p
hy
(
CT
)
sc
a
n im
a
ge
s a
r
e
u
se
d a
s one
of
t
he
s
ol
ut
io
ns
to de
te
c
t the
suf
f
e
r
e
r
.
Thi
s te
c
hn
iq
ue
ha
s be
e
n u
se
d b
y r
e
se
a
r
c
he
r
s
but
m
o
st
ly
usi
ng c
la
s
sif
ie
r
s
tha
t r
e
q
uir
e
d h
ig
h r
e
s
our
c
e
s,
suc
h a
s
c
on
vo
lu
ti
ona
l ne
ur
a
l
ne
t
wor
k
(
CN
N
)
.
I
n th
is
st
ud
y,
we
pr
op
ose
d a
wa
y t
o c
la
s
sif
y the
C
T
s
c
an
im
a
ge
s by u
si
ng t
he
m
or
e
e
f
f
ic
ie
nt c
la
s
sif
ie
r
,
k
-
ne
a
r
e
s
t ne
i
gh
bor
s
(
KN
N)
,
fo
r
im
a
ge
s tha
t a
r
e
p
r
oc
e
s
se
d u
si
ng a
c
om
bi
na
t
io
n of
the
se
f
e
a
tur
e
e
x
tr
a
c
t
io
n
m
e
th
od
s,
Ha
r
a
lic
k,
his
to
gr
a
m
,
a
nd loc
a
l bina
r
y pa
t
t
e
r
n
(L
BP
)
.
G
e
n
e
t
i
c
a
lg
or
i
thm
i
s a
l
so
use
d f
or
f
e
a
tur
e
se
le
c
ti
on.
T
he
r
e
s
ult
s s
ho
we
d
tha
t t
he
pr
o
po
se
d m
e
th
od wa
s a
ble
t
o im
pr
ove
KN
N
pe
r
f
or
m
a
nc
e
,
wit
h the
be
st
a
c
c
ur
a
c
y of
93.
3
0% f
or
the
c
om
b
ina
ti
on o
f
Ha
r
a
l
ic
k a
n
d
loc
a
l b
ina
r
y pa
t
te
r
n
f
e
a
t
u
r
e
e
x
tr
a
c
t
io
n,
a
nd the
be
st
a
r
e
a
u
n
d
e
r
t
h
e
c
u
r
v
e
(
AUC
)
f
or
the
c
om
b
ina
ti
on of
Ha
r
a
li
c
k,
hi
st
og
r
a
m
,
a
nd
loc
a
l
b
ina
r
y p
a
tte
r
n w
it
h a
va
l
ue
of
0.
94
8.
The
be
s
t a
c
c
ur
a
c
y of
o
ur
m
ode
ls a
l
so o
ut
pe
r
f
or
m
s C
NN by a
4.
3
%
marg
i
n
.
Ke
y
wo
r
d
s
:
G
e
n
e
tic
a
lg
o
r
ith
m
H
ar
al
i
ck
H
i
s
t
ogr
a
m
k
-
n
ear
es
t
ne
i
ghbour
L
o
cal
b
i
n
ar
y
p
at
t
er
n
T
his
is
a
n
o
pe
n
ac
c
e
s
s
ar
tic
le
u
nde
r
the
CC
B
Y
-
SA
lic
e
n
se
.
C
or
r
e
s
pon
di
n
g A
u
t
h
or
:
R
a
di
t
yo A
di
N
ugr
oho
D
ep
ar
t
em
en
t
o
f
C
o
m
p
u
t
er
S
ci
en
ce
,
F
acu
l
t
y
o
f
M
at
h
em
at
i
cs
an
d
N
at
u
r
al
S
ci
en
ces
L
a
m
bung M
a
ngkur
a
t
U
ni
ve
r
s
i
t
y
A.
Ya
n
i
S
t.
Km
.
3
6
,
UL
M
C
a
m
pus
B
an
j
ar
b
ar
u
,
S
o
ut
h
K
a
l
i
m
a
nt
a
n 70714
,
I
ndone
s
i
a
E
ma
il: r
a
d
ity
o
.
a
d
i@
u
lm
.
a
c
.
id
1.
I
NT
RO
DUC
T
I
O
N
R
e
c
e
nt
l
y,
I
ndone
s
i
a
i
s
hi
t
by t
he
C
or
ona
vi
r
us
di
s
e
a
s
e
2019
(
C
OVI
D
-
19
)
pa
nde
m
i
c
c
a
us
e
d by t
he
S
ev
er
e A
cu
t
e R
es
p
i
r
at
o
r
y
S
y
n
d
r
o
m
e
C
or
ona
vi
r
u
s
-
2
(
S
A
RS
-
Co
V
-
2)
.
S
i
nc
e
i
t
w
a
s
f
i
r
s
t
a
nnounc
e
d by
t
he
gove
r
nm
e
nt
i
n M
a
r
c
h 2020,
t
hi
s
vi
r
us
ha
s
c
ont
i
nue
d t
o s
pr
e
a
d t
o va
r
i
ous
pr
ovi
nc
e
s
i
n I
ndone
s
i
a
a
nd ha
s
i
nf
e
c
t
e
d hundr
e
ds
of
t
hous
a
nds
of
pe
opl
e
.
S
out
h K
a
l
i
m
a
nt
a
n,
a
pr
ovi
nc
e
i
n I
ndone
s
i
a
,
i
s o
n
e o
f
t
h
e ar
eas
w
i
t
h
t
he
hi
ghe
s
t
i
nf
e
c
t
i
on
r
a
t
e
s
i
n I
ndone
s
i
a
.
O
ne
of
t
he
f
a
c
t
or
s
t
ha
t
c
a
us
e
d t
he
hi
gh
num
be
r
of
pa
t
i
e
nt
s
w
a
s
t
he
de
l
a
y
i
n
t
he
i
de
nt
i
f
i
c
a
t
i
on
pr
oc
e
s
s
of
t
he
r
ev
er
s
e t
r
an
s
cr
i
p
t
as
e p
o
l
y
m
er
as
e ch
ai
n
r
eact
i
o
n
(
R
T
-
P
CR
)
s
w
a
b
te
s
t d
u
e
to
th
e
la
r
g
e
nu
m
be
r
of
s
pe
c
i
m
e
ns
t
ha
t
ha
d t
o be
e
xa
m
i
ne
d
by t
he
l
a
bor
a
t
o
r
y.
T
hi
s
m
a
ke
s
t
he
t
e
s
t
r
e
s
ul
t
s
know
n 14
da
ys
a
f
t
e
r
t
he
t
e
s
t
i
s
car
r
i
ed
o
u
t
.
P
C
R
is
a
s
a
mp
le
te
s
t b
y
ta
k
in
g
s
a
mp
le
s
f
r
o
m p
la
c
e
s
w
h
e
r
e
th
e
v
ir
u
s
is
mo
s
t lik
e
ly
to
b
e
pr
e
s
e
nt
,
s
uc
h a
s
t
he
ba
c
k of
t
he
nos
e
or
m
out
h
or
de
e
p
i
n
t
he
l
ungs
[1
].
T
h
e
P
C
R
t
es
t
w
as
al
s
o
d
ec
l
ar
ed
b
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1693
-
6930
T
E
L
KOM
NI
KA
T
el
eco
m
m
u
n
C
om
put
E
l
C
ont
r
o
l
,
Vo
l
.
1
9
, N
o
.
5
,
O
ct
o
b
er
2021
:
1581
-
1587
1582
W
or
l
d H
e
a
l
t
h O
r
ga
ni
z
a
t
i
on
(
WH
O
)
a
s
t
he
gol
d
e
n s
t
a
nda
r
d f
or
de
t
e
c
t
i
ng t
he
pr
e
s
e
nc
e
of
C
O
V
I
D
-
19 i
n
hum
a
ns
.
A
l
t
hough know
n
f
or
i
t
s
e
f
f
e
c
t
i
ve
ne
s
s
,
P
C
R
t
e
s
t
i
ng i
s
not
t
he
onl
y
w
a
y.
T
he
co
m
p
u
t
er
i
zed
t
om
ogr
a
phy
(
CT
)
s
can
i
s
m
o
r
e
accu
r
at
e t
h
an
t
h
e
P
C
R
s
w
ab
t
es
t
i
n
ear
l
y
d
et
ect
i
o
n
o
f
C
O
V
I
D
-
19
[2
]
.
M
a
ny
r
es
ear
ch
er
s
h
av
e i
d
en
t
i
f
i
ed
s
u
ffe
re
rs
o
f
C
O
V
ID
-
19 t
hr
ough
C
T
s
can
i
m
ag
es
s
u
ch
as
[
3]
-
[
5]
.
T
he
y
us
e
t
he
c
onvol
ut
i
ona
l
ne
ur
a
l
ne
t
w
or
k
(
C
NN)
m
e
t
hod
t
o c
l
a
s
s
i
f
y pos
i
t
i
ve
a
nd ne
ga
t
i
ve
C
O
V
I
D
-
19 pa
t
i
e
nt
s
w
i
t
h a
n
accu
r
acy
r
at
e o
f
m
o
r
e t
h
an
9
0
%
.
C
N
N
i
s
a
t
ype
of
ne
ur
a
l
ne
t
w
o
r
k f
o
r
p
r
oc
e
s
s
i
ng da
t
a
t
ha
t
ha
s
a
ne
t
w
or
k
-
l
i
ke
t
opol
ogy
[
6]
.
C
NN i
s
w
i
de
l
y us
e
d i
n c
om
put
e
r
vi
s
i
on,
a
s
i
s
done
by [
7
]
-
[
10]
.
D
e
s
pi
t
e
ha
vi
ng va
r
i
ous
a
dva
nt
a
ge
s
,
C
N
N
i
s
a
m
e
t
hod
t
ha
t
r
e
qui
r
e
s
e
nor
m
ous
c
om
put
a
t
i
ona
l
r
e
s
our
c
e
s
[
11
]
.
H
o
w
ev
er
,
i
n
m
ach
i
n
e l
ear
n
i
n
g
th
e
r
e
a
r
e
s
till
ma
n
y
o
th
e
r
c
l
a
s
s
i
f
i
c
a
t
i
on a
l
gor
i
t
hm
s
t
ha
t
c
a
n be
us
e
d w
i
t
h
l
ow
r
e
s
our
c
e
s
,
one
of
w
hi
c
h i
s
k
-
ne
a
r
e
s
t
ne
i
ghbor
s
(
K
NN)
.
T
h
e
KNN
a
l
gor
i
t
hm
w
a
s
f
o
r
m
ul
a
t
e
d by pe
r
f
or
m
i
ng a
non
-
p
ar
am
et
r
i
c m
et
h
o
d
f
o
r
p
at
t
er
n
c
la
s
s
if
ic
a
tio
n
[
12]
.
KNN
al
s
o
s
t
at
ed
as
a
s
i
m
p
l
e b
u
t
ef
f
ect
i
v
e al
g
o
r
i
t
h
m
f
o
r
s
ev
er
al
cas
es
[
13]
.
T
h
e s
u
cces
s
o
f
t
h
e
KNN
a
l
gor
i
t
h
m
de
pe
nds
on
s
e
l
e
c
t
i
ng t
he
c
o
r
r
e
c
t
k va
l
ue
.
I
n
t
hi
s
s
t
udy,
w
e
us
e
d t
he
KNN
to
id
e
n
tif
y
s
u
ffe
re
rs
o
f
C
O
V
ID
-
19 ba
s
e
d on
C
T
s
can
i
m
ag
es
.
T
h
e
ima
g
e
s
w
a
s
c
ol
l
e
c
t
e
d
f
r
om
T
ongj
i
H
os
pi
t
a
l
i
n
W
uha
n,
C
hi
na
[
4]
.
B
ef
o
r
e t
h
e d
at
a m
i
n
i
n
g
p
r
o
ces
s
i
s
car
r
i
ed
o
u
t
,
t
h
e
obt
a
i
ne
d
C
T
s
can
i
m
ag
e
i
s
ex
t
r
act
ed
b
as
ed
o
n
te
x
tu
r
e
to
o
b
ta
in
its
c
h
a
r
a
c
te
r
is
tic
v
a
lu
e
s
.
F
e
a
tu
r
e
e
x
tr
a
c
tio
n
in
ima
g
e
s
is
d
iv
id
e
d
in
to
s
e
v
e
r
a
l c
a
te
g
o
r
ie
s
,
n
am
el
y
,
c
ol
or
,
t
e
xt
ur
e
,
a
nd s
ha
pe
[
14]
.
T
e
x
t
ur
e
-
ba
s
e
d f
e
a
t
ur
e
e
xt
r
a
c
t
i
on i
s
know
n t
o ha
ve
a
dva
nt
ag
es
,
n
am
el
y
it h
a
s
lo
w
c
o
mp
u
ta
tio
n
a
l
c
o
mp
le
x
ity
a
n
d
is
e
a
s
y
t
o
imp
le
me
n
t
[
1
5]
.
T
he
f
e
a
t
ur
e
e
xt
r
a
c
t
i
on m
e
t
hods
us
e
d
i
n
t
h
i
s
s
t
u
d
y
ar
e H
ar
al
i
ck
,
l
o
cal
b
i
n
ar
y
p
at
t
er
n
(
LB
P
)
,
a
nd 32
-
bi
n
hi
s
t
ogr
a
m
.
O
ne
of
t
he
c
ha
l
l
e
nge
s
i
n t
hi
s
s
t
udy i
s
t
ha
t
t
he
f
e
a
t
ur
e
e
xt
r
a
c
t
i
on r
e
s
ul
t
s
i
n e
a
c
h m
e
t
hod ha
ve
a
l
a
r
ge
n
u
m
b
er
o
f
f
eat
u
r
es
.
T
h
i
s
can
cau
s
e a
c
ur
s
e
o
f
di
m
e
ns
i
ona
l
i
t
y
(
C
o
D
)
w
h
ic
h
le
a
d
s
to
a
h
ig
h
time
c
o
mp
le
x
ity
pr
obl
e
m
[
1
6]
.
C
o
D
m
ay
al
s
o
d
ecr
eas
e t
h
e accu
r
acy
g
en
er
at
ed
b
y
t
h
e
al
g
o
r
i
t
h
m
.
T
o
o
v
er
co
m
e t
h
i
s
w
eak
n
es
s
,
S
ay
ed
e
t a
l.
[
1
7]
s
u
g
g
es
t
ed
t
h
e u
s
e o
f
f
eat
u
r
e s
el
ect
i
o
n
.
F
eat
u
r
e s
el
ect
i
o
n
i
s
a m
et
h
o
d
f
o
r
s
el
ect
i
n
g
t
he
m
os
t
r
e
l
e
va
nt
f
e
a
t
ur
e
s
f
r
om
a
da
t
a
s
e
t
.
R
e
duc
i
ng t
he
da
t
a
di
m
e
ns
i
on w
oul
d a
l
s
o r
e
s
ul
t
i
n pe
r
f
or
m
a
nc
e
i
m
p
r
o
v
em
en
t
i
n
m
an
y
cas
es
.
O
n
e t
y
p
e o
f
f
eat
u
r
e
s
el
ect
i
o
n
i
s
w
r
ap
p
er
[
1
8]
.
T
he
w
r
a
ppe
r
us
e
s
m
a
c
hi
ne
l
e
a
r
ni
ng
t
o
r
un
t
hr
ough
a
l
l
pos
s
i
bl
e
f
e
a
t
ur
e
c
om
bi
na
t
i
ons
,
t
he
n s
e
l
e
c
t
s
t
he
c
o
m
bi
na
t
i
on t
ha
t
pr
oduc
e
s
t
he
be
s
t
pe
r
f
or
m
a
nc
e
.
T
h
e w
r
ap
p
er
m
e
t
hod de
t
e
r
m
i
ne
s
t
he
be
s
t
f
e
a
t
ur
e
c
om
bi
na
t
i
o
n by
c
om
pa
r
i
ng
t
he
e
va
l
u
a
tio
n
c
r
ite
r
ia
d
e
te
r
mi
n
e
d
f
r
o
m
va
r
i
ous
f
e
a
t
ur
e
c
om
bi
na
t
i
ons
,
t
he
n
f
r
om
t
he
c
om
pa
r
i
s
on r
e
s
ul
t
s
s
e
l
e
c
t
t
he
f
e
a
t
ur
e
c
om
bi
na
t
i
on
t
ha
t
ha
s
t
he
m
os
t
opt
i
m
a
l
r
e
s
ul
t
s
.
O
ne
a
l
gor
i
t
hm
t
ha
t
c
a
n be
us
e
d t
o pe
r
f
o
r
m
w
r
a
ppe
r
-
b
as
ed
f
eat
u
r
e s
el
ect
i
o
n
i
s
t
h
e
g
en
et
i
c al
g
o
ri
t
hm
(
GA)
,
a
s
ha
s
be
e
n done
by
[
19
]
-
[
21]
.
I
n t
hi
s
s
t
udy,
w
e
pr
opos
e
d a
m
e
t
hod
t
o i
m
p
r
ove
KN
N
c
la
s
s
if
ic
a
tio
n
in
th
e
c
o
mp
u
te
r
v
is
io
n
f
ie
ld
.
T
h
e
ge
ne
t
i
c
a
l
gor
i
t
hm
w
a
s u
se
d
a
s
t
he
f
e
a
tu
r
e
s
e
le
c
tio
n
in
c
la
s
s
if
y
in
g
C
O
V
I
D
-
19
s
uf
f
e
r
e
r
s
t
hr
ough
C
T
S
c
a
n
i
m
ag
es
ex
t
r
ac
t
e
d us
i
ng
t
h
e
H
ar
al
i
ck
m
et
h
o
d
,
3
2
-
bi
n
hi
s
t
ogr
a
m
,
a
nd l
oc
a
l
bi
na
r
y
pa
t
t
e
r
n
.
A
f
te
r
g
e
ttin
g
th
e
c
l
a
s
s
i
f
i
c
a
t
i
on r
e
s
ul
t
s
of
t
he
pr
opos
e
d m
e
t
hod,
w
e
c
om
pa
r
e
d t
he
m
by t
he
be
s
t
r
e
s
ul
t
s
i
n t
he
p
r
e
vi
ous
w
or
k
[
4]
.
T
he
y us
e
d t
he
C
N
N
D
e
ns
e
N
e
t
-
169 a
nd R
e
s
N
e
t
-
5
0
ar
ch
i
t
ec
t
ur
e
s
.
B
e
f
o
r
e
e
nt
e
r
i
ng
t
he
C
N
N
pr
oc
e
s
s
,
t
he
y di
d
pr
e
-
p
r
o
ces
s
st
e
p
by r
e
s
i
z
i
ng t
he
i
m
a
ge
t
o
480x4
80.
A
n i
m
a
ge
s
e
gm
e
nt
a
t
i
on pr
oc
e
s
s
i
s
a
l
s
o c
a
r
r
i
e
d out
t
o
i
m
pr
ove
a
c
c
ur
a
c
y.
M
e
a
nw
hi
l
e
,
i
n our
m
e
t
hod,
w
e
pr
oc
e
s
s
t
he
i
m
a
ge
di
r
e
c
t
l
y
i
nt
o t
he
f
e
a
t
ur
e
e
xt
r
a
c
t
i
on,
w
i
t
hout
r
e
s
i
z
e
a
nd s
e
gm
e
nt
a
t
i
on.
2.
RE
S
E
ARCH
M
ETH
O
D
O
u
r
r
es
ear
ch
w
as
car
r
i
ed
o
u
t
as
i
n
F
i
g
u
r
e
1
.
I
n
t
hi
s
s
t
udy,
t
he
C
oe
l
ho
[
2
2]
l
i
b
ra
ry
f
or
P
yt
hon
w
a
s
us
e
d t
o pe
r
f
or
m
f
e
a
t
ur
e
ex
t
r
act
i
o
n
.
M
ean
w
h
i
l
e,
t
o
p
er
f
o
r
m
f
eat
u
r
e s
el
ect
i
o
n
an
d
cl
as
s
i
f
i
cat
i
o
n
,
t
h
e
R
a
pi
dM
i
ne
r
[
23]
so
f
t
w
a
r
e
i
s
u
se
d
.
2.
1
.
D
at
a
s
e
t
T
h
e
d
a
ta
s
e
t u
s
e
d
in
th
is
s
tu
d
y
is
th
e
C
T
s
can
d
at
as
et
co
m
p
i
l
ed
b
y
[
4]
.
T
h
er
e
ar
e 7
4
6
g
r
ay
s
cal
e
i
m
a
ge
s
c
ons
i
s
t
i
ng of
349
C
T
s
c
a
ns
o
f
C
OVI
D
-
19
pa
t
i
e
nt
s
a
nd 397 non
-
C
OVI
D
-
19 pa
t
i
e
nt
s
.
T
h
e i
m
ag
es
ar
e
v
ar
y
i
n *
.
j
pg
a
nd *
.
png
f
or
m
a
t
s
.
2.
2
.
F
ea
t
u
re
ex
t
ra
ct
i
o
n
B
ef
o
r
e en
t
er
i
n
g
t
h
e
cl
as
s
i
f
i
cat
i
o
n
s
t
ag
e,
t
h
e
f
eat
u
r
es
o
f
t
h
e d
o
w
n
l
o
ad
ed
d
at
as
et
ar
e
ex
t
r
act
ed
u
s
i
n
g
H
a
r
a
l
i
c
k m
e
t
hod,
32
-
bi
n
hi
s
t
ogr
a
m
,
a
nd
l
o
cal
b
i
n
ar
y
p
at
t
er
n
.
A
t
t
h
i
s
s
t
ag
e each
i
m
ag
e i
s
co
n
v
er
t
ed
i
n
t
o
a
num
be
r
o
f
ma
tr
ix
.
T
h
e f
i
r
s
t
f
eat
u
r
e ex
t
r
act
i
o
n
i
s
H
a
r
a
l
i
c
k.
T
hi
s
f
e
a
t
ur
e
c
ont
a
i
ns
i
nf
or
m
a
t
i
on
a
bout
t
he
i
nt
e
ns
i
t
y of
t
he
i
m
a
ge
i
n pi
xe
l
s
w
i
t
h c
e
r
t
a
i
n pos
i
t
i
ons
i
n r
e
l
a
t
i
on t
o e
a
c
h ot
he
r
oc
c
ur
r
i
ng s
i
m
ul
t
a
ne
o
us
l
y
[
24]
.
T
h
is
me
th
o
d
c
a
lc
u
la
te
s
its
f
e
a
tu
r
e
v
a
lu
e
f
r
o
m
8
an
g
l
es
,
n
am
el
y
0
,
45,
9
0,
135
,
180,
225,
270
,
315,
3
60.
E
a
c
h
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
KA
T
el
eco
m
m
u
n
C
om
put
E
l
C
ont
r
o
l
I
m
pr
ov
e
m
e
nt
on K
N
N
us
i
ng ge
ne
t
i
c
al
gor
i
t
hm
and c
om
bi
ne
d f
e
at
ur
e
e
x
t
r
ac
t
i
on
…
(
R
adi
t
y
o
A
di
N
ugr
oho
)
1583
a
ngl
e
pr
oduc
e
s
14 f
e
a
t
ur
e
s
us
i
ng t
he
f
o
r
m
ul
a
i
n
T
a
bl
e
1.
T
hus
,
e
a
c
h i
m
a
ge
w
i
l
l
pr
oduc
e
8
x14 f
e
a
t
ur
e
s
.
T
he
n,
s
i
m
p
l
i
f
i
cat
i
o
n
i
s
car
r
i
ed
o
u
t
b
y
cal
cu
l
at
i
n
g
t
h
e av
e
r
ag
e v
al
u
e o
f
each
co
r
n
e
r
f
eat
u
r
e
i
n
each
i
m
ag
e
,
s
o
t
h
at
t
h
e
n
u
m
b
er
o
f
f
eat
u
r
es
f
o
r
each
i
m
ag
e
b
eco
m
es
1
4
f
ea
t
u
r
es
.
T
he
f
or
m
ul
a
i
n T
a
b
e
l
1
i
s
f
or
m
e
d
f
r
om
t
he
g
r
a
y
-
t
one
s
p
at
i
al
-
de
p
e
nde
nc
y m
a
t
r
i
x of
a
n
i
m
a
ge
a
nd de
f
i
ne
d by
[
2
4]
a
s f
o
l
l
o
w
s:
-
p
(
i,
j)
:
(
i,
j)
th
e
n
tr
y
in
th
e
ma
tr
ic
e
;
-
p
x
(
i)
:
t
h
e en
t
r
y
i
n
t
h
e m
ar
g
i
n
al
-
pr
oba
bi
l
i
t
y m
a
t
r
i
x
o
bt
a
i
ne
d by s
um
m
i
ng t
he
r
ow
s
of
p
(
i,
j)
;
-
Ng
:
t
he
num
be
r
o
f
di
s
t
i
nc
t
gr
a
y l
e
ve
l
s
i
n
t
he
qua
nt
i
z
e
d i
m
a
ge
.
T
he
ne
xt
e
xt
r
a
c
t
i
on m
e
t
hod i
s
hi
s
t
ogr
a
m
.
A
n i
m
a
ge
hi
s
t
ogr
a
m
i
s
t
he
i
nt
e
ns
i
t
y of
t
he
num
be
r
of
pi
xe
l
s
w
hi
c
h i
s
f
or
m
e
d
i
n
gr
a
phi
c
a
l
f
or
m
a
t
.
T
he
va
l
ue
s
f
o
r
m
e
d i
n
e
a
c
h i
m
a
ge
a
r
e
gr
oupe
d
i
nt
o
32 va
l
ue
r
an
g
es
.
S
o
,
i
n
t
h
i
s
s
t
ep
3
2
f
eat
u
r
es
ar
e
g
en
er
at
ed
.
T
h
e
th
ir
d
is
l
oc
a
l
bi
na
r
y pa
t
t
e
r
n
.
T
h
is
me
th
o
d
is
a
s
imp
le
but
ve
r
y e
f
f
i
c
i
e
nt
t
e
xt
u
r
e
ope
r
a
t
or
t
ha
t
l
ab
el
s
i
m
ag
e
p
i
x
e
l
s
by l
i
m
i
t
i
n
g
t
he
e
n
vi
r
o
nm
e
nt
of
e
a
c
h
p
i
x
e
l
a
n
d
c
ons
i
d
e
r
s
t
he
r
e
s
u
l
t
t
o be
a
bi
na
r
y
nu
m
b
e
r
[
2
5]
.
T
he
n,
t
he
l
a
be
l
hi
s
t
ogr
a
m
c
a
n be
us
e
d a
s
a
t
e
xt
ur
e
de
s
c
r
i
pt
or
.
Th
i
s
m
e
t
hod pr
oduc
e
d
25 f
e
a
t
u
r
e
s
.
F
i
gur
e
1.
P
r
opos
e
d m
e
t
hod
a
bs
t
r
a
c
t
i
on de
s
i
gn
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1693
-
6930
T
E
L
KOM
NI
KA
T
el
eco
m
m
u
n
C
om
put
E
l
C
ont
r
o
l
,
Vo
l
.
1
9
, N
o
.
5
,
O
ct
o
b
er
2021
:
1581
-
1587
1584
T
ab
l
e 1
.
H
ar
al
i
ck
’
s
F
eat
u
r
e
an
d
i
t
s
f
o
r
mu
la
No
F
eat
u
r
es
F
o
r
mu
la
1
A
ngul
a
r
S
e
c
ond M
om
e
nt
Σ
Σ
(
,
)
2
2
C
ont
r
a
s
t
Σ
=
0
−
1
2
Σ
=
1
Σ
=
1
(
,
)
,
|
−
|
=
3
C
o
r
r
e
la
tio
n
Σ
Σ
(
)
(
,
)
−
4
S
u
m
o
f
S
q
u
ar
es
:
V
ar
i
an
ce
Σ
Σ
(
−
)
2
(
,
)
5
I
n
v
er
s
e D
i
f
f
er
en
ce M
o
m
en
t
Σ
Σ
1
1
+
(
−
)
2
(
,
)
6
S
um
A
ve
r
a
ge
Σ
=
2
2
+
(
)
7
S
u
m
V
ar
i
an
ce
Σ
=
2
2
(
−
8
)
2
+
(
)
8
S
um
E
nt
r
opy
−
Σ
=
2
2
+
(
)
l
og
+
(
)
=
8
9
E
nt
r
opy
−
Σ
Σ
(
,
)
l
og
(
(
,
)
)
10
D
i
f
f
er
en
ce V
ar
i
an
ce
Σ
=
0
−
1
2
−
(
)
11
D
i
ffe
re
n
c
e
E
n
t
ro
p
y
−
Σ
=
0
−
1
−
(
)
l
og
{
−
(
)
}
12
I
nf
o. M
e
a
s
ur
e
of
C
ol
l
e
c
t
i
on 1
H
X
Y
−
H
X
Y
1
ma
x
{
,
}
13
I
nf
o. M
e
a
s
ur
e
of
C
ol
l
e
c
t
i
on 2
(
1
−
e
xp
[
−
2
(
2
−
)
]
)
1
2
14
M
a
x. C
or
r
e
l
a
t
i
on C
oe
f
f
i
c
i
e
nt
T
he
s
quar
e
r
oot
of
t
he
s
e
c
ond l
a
r
ge
s
t
e
i
ge
nv
al
ue
of
Q
, w
h
e
r
e
(
,
)
=
Σ
(
,
)
(
,
)
(
)
(
)
2.
3
.
G
en
era
t
e
n
ew
d
a
t
a
s
et
T
hi
s
s
t
a
ge
i
s
t
he
f
or
m
a
t
i
on of
a
ne
w
da
t
a
s
e
t
by
c
o
m
bi
ni
ng t
he
f
e
a
t
ur
e
s
f
or
m
e
d i
n 2.
2
.
A
t
t
hi
s
s
t
a
ge
,
7
n
ew
d
at
as
et
s
ar
e
g
en
er
at
ed
w
h
i
ch
ar
e d
es
cr
i
b
ed
i
n
T
ab
l
e 2
.
E
ve
r
y
da
t
a
s
e
t
ha
s
di
f
f
e
r
e
nt
di
m
e
ns
i
on de
p
e
nds
on
its
f
e
a
tu
r
e
e
x
tr
a
c
tio
n
me
th
o
d
.
T
ab
l
e 2
.
D
et
ai
l
o
f
n
ew
d
at
as
et
s
D
at
as
et
N
um
of
F
eat
u
r
e
Ha
r
14
F
o
r
me
d
f
r
o
m H
a
r
a
lic
k
e
x
tr
a
c
tio
n
H
is
t3
2
32
F
o
r
me
d
f
r
o
m H
is
to
g
r
a
m e
x
tr
a
c
tio
n
LB
P
25
F
o
r
me
d
f
r
o
m
L
o
c
a
l B
in
a
r
y
P
a
tte
r
n
e
x
tr
a
c
tio
n
H
a
r
+
H
is
t3
2
46
C
om
bi
na
t
i
on of
H
a
r
a
l
i
c
k &
H
i
s
t
ogr
a
m
32bi
n
H
a
r
+
LB
P
39
C
o
mb
in
a
tio
n
o
f
H
a
r
a
lic
k
& L
o
c
a
l
B
in
a
r
y
P
a
tte
r
n
H
is
t3
2
+
L
B
P
57
C
om
bi
na
t
i
on of
H
i
s
t
ogr
a
m
32bi
n &
L
oc
a
l
B
i
na
r
y P
a
t
t
e
r
n
H
a
r
+H
i
st
3
2
+L
B
P
71
C
om
bi
na
t
i
on of
H
a
r
a
l
i
c
k, H
i
s
t
ogr
a
m
32bi
n, &
L
oc
a
l
B
i
na
r
y P
a
t
t
e
r
n
2.
4
.
C
l
as
s
i
f
i
c
at
i
on
an
d
c
r
os
s
val
i
d
at
i
on
A
t th
is
s
ta
g
e
,
th
e
d
a
ta
s
e
t f
o
r
me
d
in
T
a
b
le
2
is
c
l
a
s
s
if
ie
d
u
s
in
g
th
e
KNN
a
l
gor
i
t
hm
a
nd va
l
i
da
t
e
d
us
i
ng 10
-
f
ol
d c
r
os
s
va
l
i
da
t
i
on
.
T
h
e
KNN
cl
as
s
i
f
i
c
at
i
o
n
i
s
car
r
i
ed
o
u
t
w
i
t
h
a v
al
u
e o
f
k
=2
t
o
k
=1
7
.
T
h
e v
al
u
e
of
k=
1 w
a
s
not
i
nc
l
ude
d be
c
a
us
e
of
t
he
hi
gh
va
r
i
a
n
ce
[
26]
.
2.
4.
1
.
C
la
s
s
if
ic
a
t
io
n
w
i
t
h
o
u
t
f
ea
t
u
re
s
el
ect
i
o
n
(
KN
N
On
ly
)
T
h
i
s
cl
as
s
i
f
i
cat
i
o
n
i
n
v
o
l
v
es
al
l
t
h
e
f
eat
u
r
es
t
h
at
ar
e f
o
r
m
ed
f
r
o
m
T
ab
l
e 2
.
H
er
e,
w
e d
o
n
o
t
s
el
ect
t
h
e
f
eat
u
r
es
y
et
.
L
at
er
,
t
h
e accu
r
acy
o
f
t
h
e
KNN
cl
as
s
i
f
i
er
w
i
l
l
b
e co
m
p
ar
ed
t
o
t
h
e
accu
r
acy
o
f
GA+
KNN
.
2.
4.
2
.
C
la
s
s
if
ic
a
t
io
n
u
s
in
g
g
en
et
i
c
a
l
g
o
ri
t
h
m
f
ea
t
u
re s
el
ect
i
o
n
(
G
A+
K
NN
)
E
ach
d
at
as
et
i
n
T
ab
l
e 2
i
s
cr
eat
ed
a n
ew
d
at
a s
u
b
s
et
co
n
t
ai
n
i
n
g
o
n
l
y
t
h
e f
eat
u
r
es
s
el
ect
ed
b
y
t
h
e
ge
ne
t
i
c
a
l
gor
i
t
hm
.
T
hi
s
a
l
go
r
i
t
hm
w
or
ks
a
s
f
ol
l
ow
s
[
27]
:
i
)
S
t
e
p 1:
I
n
i
t
i
a
l
i
z
e
r
a
ndom
i
ndi
v
i
dua
l
pop
ul
a
t
i
ons
;
ii)
S
t
e
p
2:
A
s
s
i
gn f
i
t
ne
s
s
va
l
ue
s
f
or
e
a
c
h i
nd
i
vi
du
a
l
i
n t
he
popul
a
t
i
on
; iii
)
S
t
e
p
3:
M
a
ke
i
ndi
vi
dua
l
s
e
l
e
c
t
i
ons
on t
he
popul
a
t
i
on t
o
c
r
e
a
t
e
ne
w
ge
ne
r
a
t
i
on
; iv
)
S
t
ep
4
:
P
er
f
or
m
c
r
os
s
ove
r
s
on t
he
s
e
l
e
c
t
e
d i
ndi
vi
dua
l
s
;
v)
St
e
p
5:
P
e
r
f
or
m
m
ut
a
t
i
ons
t
o
a
voi
d
s
i
m
i
l
a
r
i
t
y
i
n t
he
ge
ne
r
a
t
i
on
of
r
e
s
ul
t
s
c
r
os
s
ove
r
a
nd pa
r
e
nt
po
pul
a
t
i
on
;
vi
)
S
t
ep
6
:
R
ep
eat
s
t
ep
2
-
5
u
n
til
th
e
s
to
p
c
r
ite
r
ia
a
r
e
me
t.
2.
5
.
E
val
u
at
i
o
n
A
t
t
h
i
s
s
t
ag
e,
t
h
e
p
er
f
o
r
m
an
ce o
f
t
h
e
KNN
al
g
o
r
i
t
h
m
i
s
ev
al
u
at
ed
b
as
ed
o
n
i
t
s
accu
r
acy
an
d
a
r
ea
unde
r
t
he
c
ur
ve
(
AUC
)
.
T
h
e h
i
g
h
er
t
h
e accu
r
acy
v
al
u
e,
t
h
e b
et
t
er
t
h
e p
er
f
o
r
m
an
ce o
f
t
h
e
m
o
d
el
.
T
hi
s
r
ul
e
a
l
s
o a
ppl
i
e
d t
o t
he
A
U
C
va
l
ue
.
2.
6
.
S
i
g
n
i
f
i
ca
n
ce
t
e
st
At
t
h
i
s st
a
g
e
,
w
e
u
s
e t
h
e t
-
te
s
t me
th
o
d
.
T
hi
s
m
e
t
hod w
a
s
a
ppl
i
e
d
t
o
t
es
t
t
h
e
s
i
g
n
i
f
i
can
ce o
f
each
o
f
t
he
be
s
t
va
l
ue
s
pr
oduc
e
d by
KNN
a
nd
GA
+
KNN
f
o
r
each
d
at
as
et
i
n
T
ab
l
e 2
.
Wi
t
h
al
p
h
a v
al
u
e=0
.
0
5
,
m
ean
s
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
KA
T
el
eco
m
m
u
n
C
om
put
E
l
C
ont
r
o
l
I
m
pr
ov
e
m
e
nt
on K
N
N
us
i
ng ge
ne
t
i
c
al
gor
i
t
hm
and c
om
bi
ne
d f
e
at
ur
e
e
x
t
r
ac
t
i
on
…
(
R
adi
t
y
o
A
di
N
ugr
oho
)
1585
t
h
at
t
h
e s
i
g
n
i
f
i
can
ce v
al
u
e o
f
KNN
a
n
d
GA+
KNN
i
s
l
e
s
s
t
ha
n 0.
05 (
p
-
v
al
u
e
<
α)
i
n
d
i
cat
es
t
h
e
t
w
o
m
o
d
el
s
can
be
s
a
i
d t
o be
s
i
gni
f
i
c
a
nt
l
y
di
f
f
e
r
e
nt
.
3.
RE
S
UL
T
AND ANAL
YS
I
S
A
t
t
h
i
s
s
t
ag
e,
t
h
e b
es
t
accu
r
acy
f
o
r
t
h
e k
-
N
N
m
o
de
l
i
s
c
om
pa
r
e
d w
i
t
h t
he
be
s
t
f
or
t
he
G
A
+
KNN
m
ode
l
.
T
he
n
,
t
o
s
how
t
h
at
t
h
e b
es
t
accu
r
acy
o
f
t
h
e t
w
o
m
o
d
el
s
h
as
a s
t
at
i
s
t
i
cal
l
y
s
i
g
n
i
f
i
can
t
d
i
f
f
er
en
ce,
a
di
f
f
e
r
e
nt
t
e
s
t
i
s
pe
r
f
o
r
m
e
d
us
i
ng t
he
t
-
t
es
t
.
T
h
e
t
es
t
r
es
u
l
t
s
can
b
e
s
een
i
n
T
ab
l
e 3
.
F
r
o
m
t
h
e
t
es
t
,
w
e
can
s
ee
t
ha
t
,
a
l
t
hough
not
a
l
l
p
r
oduc
e
s
i
gni
f
i
c
a
nt
di
f
f
e
r
e
nc
e
s
,
t
he
r
e
s
ul
t
s
obt
a
i
ne
d a
r
e
t
he
p
r
opos
e
d m
e
t
hod
(
GA+
KNN
)
out
pe
r
f
or
m
i
ng
KNN
.
G
en
et
i
c
a
l
gor
i
t
hm
ha
s
be
e
n
s
how
n t
o
i
m
p
r
ove
K
NN
cl
as
s
i
f
i
cat
i
o
n
accu
r
acy
i
n
i
m
ag
es
ex
t
r
act
ed
w
i
t
h
H
ar
al
i
ck
,
L
B
P
,
an
d
H
ar
+
LB
P
.
T
h
e
b
es
t
o
v
er
al
l
accu
r
acy
r
es
u
l
t
s
w
er
e
ach
i
ev
ed
b
y
G
A
+
KN
N
on
t
he
H
ar
al
i
ck
+L
B
P
f
eat
u
r
e ex
t
r
act
i
o
n
.
A
l
s
o
,
t
h
e
b
es
t
A
U
C
v
al
u
e i
s
g
en
er
at
ed
b
y
G
A
+
KNN
o
n
H
ar
al
i
c
k
+3
2
-
bi
n
hi
s
t
ogr
a
m
+L
B
P
d
at
as
et
.
T
o
d
et
er
m
i
n
e t
h
e e
f
f
e
ct
i
v
en
es
s
o
f
t
h
e m
o
d
el
,
w
e co
m
p
ar
e
t
h
e b
es
t
accu
r
acy
GA+
KNN
w
i
t
h t
he
C
N
N
m
ode
l
pr
oduc
e
d
by
[
4]
a
s
i
n
T
a
b
l
e 4
.
Y
an
g
m
o
d
el
ex
cel
s
i
n
t
h
e A
U
C
s
co
r
e,
w
h
i
l
e
t
he
our
p
r
opos
e
d m
ode
l
i
s
s
upe
r
i
or
i
n
t
e
r
m
s
of
a
c
c
ur
a
c
y.
T
ab
l
e 3
.
R
es
u
l
t
s
o
f
t
h
e b
es
t
accu
r
acy
an
d
i
t
s
t
-
t
e
st
No
D
at
as
et
B
es
t
A
ccu
r
a
cy
K
N
N
B
e
s
t
Ac
c
u
r
a
c
y
G
A+
KNN
B
es
t
A
ccu
r
a
cy
t
-
T
e
s
t
(
α
=
0
.
05)
k
A
ccu
r
acy
AUC
k
A
ccu
r
acy
AUC
1
Ha
r
15
79.10%
0.835
13
83.70%
0.896
S
ig
n
if
ic
a
n
t
2
H
is
t3
2
2
90.90%
0.936
2
91.80%
0.935
N
o
t S
ig
n
if
ic
a
n
t
3
LB
P
4
81.00%
0.859
2
89.00%
0.901
S
ig
n
if
ic
a
n
t
4
H
a
r
+
H
is
t3
2
2
90.62%
0.933
2
92.23%
0.937
N
o
t S
ig
n
if
ic
a
n
t
5
H
a
r
+
LB
P
2
84.60%
0.868
2
93.30%
0.937
S
ig
n
if
ic
a
n
t
6
H
is
t3
2
+
L
B
P
2
91.55%
0.942
2
92.63%
0.94
N
o
t S
ig
n
if
ic
a
n
t
7
H
a
r
+H
i
st
3
2
+L
B
P
2
91.16%
0.926
2
92.76%
0.948
N
o
t S
ig
n
if
ic
a
n
t
T
ab
l
e 4
.
G
A
+
KNN
c
om
pa
r
i
s
on a
ga
i
ns
t
C
N
N
Y
a
ng’
s
m
ode
l
Y
a
ng,
e
t a
l
[4
]
P
r
opos
e
d M
ode
l
(
G
A
+
K
N
N
)
A
ccu
r
acy
AUC
A
ccu
r
acy
AUC
89%
0.98
93.30%
0.937
4.
CO
NCL
US
I
O
N
T
he
r
e
s
e
a
r
c
h pr
ove
d
t
ha
t
t
he
m
ode
l
bui
l
t
us
i
ng
t
he
g
e
n
e
tic
a
lg
o
r
ith
m
(
GA+
KNN
)
c
om
bi
ne
d
w
i
t
h
H
ar
al
i
ck
an
d
l
o
cal
bi
na
r
y pa
t
t
e
r
n
w
a
s
a
bl
e
t
o i
m
pr
ove
t
he
pe
r
f
or
m
a
nc
e
of
t
he
KNN
o
n
ly
c
la
s
s
if
ic
a
tio
n
a
l
gor
i
t
hm
a
nd p
r
oduc
e
t
he
be
s
t
a
c
c
ur
a
c
y w
i
t
h a
v
a
l
ue
of
93
.
30%
a
nd A
U
C
of
0.
937
.
T
he
m
a
c
hi
ne
l
e
a
r
ni
ng
m
ode
l
pr
oduc
e
d i
s
a
l
s
o a
bl
e
t
o pr
ov
i
d
e
e
xc
e
l
l
e
nt
r
e
s
ul
t
s
by out
pe
r
f
or
m
s
t
he
C
N
N
Y
a
ng m
ode
l
,
w
h
i
c
h w
a
s
f
o
r
m
ed
b
y
t
h
e i
d
en
t
i
cal
d
at
as
et
.
R
EF
ER
EN
C
ES
[1
]
J.
Ha
da
ya
,
M
.
S
c
hum
m
a
n
d
E.
H.
L
ivi
ng
st
on,
"
Te
s
ti
n
g
I
nd
iv
id
ua
l
s f
or
C
or
o
na
v
ir
u
s D
ise
a
se
2
01
9 (
C
O
VI
D
-
19
)
,
"
J
am
a,
vo
l.
32
3,
no.
1
9,
p.
19
81,
2
02
0,
do
i:
1
0.
1
00
1/
ja
m
a
.
202
0.
53
88.
[2
]
T.
Ai,
e
t al
.
,
"
C
or
r
e
la
ti
on
of
C
he
s
t C
T a
n
d R
T
-
P
C
R
Te
sti
ng
f
or
C
or
o
na
v
ir
u
s
Di
se
a
se
20
19
(
C
O
VI
D
-
19)
in
C
hi
na
:
A R
e
por
t of
1
01
4 C
a
se
s,
"
Rad
io
lo
gy
,
vo
l.
29
6,
no.
2,
2
0
20,
do
i:
10.
1
14
8/
r
a
d
io
l.
20
20
20
06
42.
[3
]
D.
S
in
gh,
V.
K
um
a
r
,
V
a
i
sha
li a
nd M
.
Ka
ur
,
"
C
la
ss
if
ic
a
ti
on of
C
OVI
D
-
1
9 pa
tie
nt
s f
r
om
c
he
st C
T im
a
ge
s u
s
in
g
m
ult
i
-
ob
je
c
t
ive
dif
f
e
r
e
nt
ia
l
e
vo
lu
ti
on
–
b
a
se
d c
o
nv
ol
ut
ion
a
l
ne
ur
a
l ne
tw
or
k
s,
"
Eu
ro
pe
a
n J
ou
rn
al
of C
li
ni
c
a
l
Mic
ro
bi
ol
ogy
&
I
n
fe
c
t
io
n Di
se
a
se
s,
v
ol.
3
9,
no.
7,
p
p.
1
397
-
13
89
,
20
20,
d
oi
:
10.
10
07
/s1
00
96
-
02
0
-
0
39
01
-
z
.
[4
]
X.
Ya
ng,
X.
He
,
J.
Z
ha
o,
Y.
Z
ha
ng,
S
.
Z
ha
ng a
nd P
.
Xie
,
"
C
OVI
D
-
CT
-
Da
ta
se
t: A C
T S
c
a
n Da
ta
se
t a
b
ou
t
C
OVI
D
-
19
,
"
arX
iv
e
-
pr
in
ts,
20
20
.
[
On
li
ne
]
.
Av
a
i
la
b
le
:
htt
ps
:/
/a
r
xi
v.
or
g
/a
b
s/
20
03.
13
86
5
[5
]
I
.
Kha
n,
J.
L
.
S
ha
h
a
n
d M
.
B
ha
t,
"
C
or
oNe
t:
A
de
e
p
ne
u
r
a
l ne
tw
or
k
f
or
de
te
c
ti
on
a
n
d d
ia
g
no
si
s
of
C
O
VI
D
-
1
9
f
ro
m
c
h
e
s
t
x
-
r
a
y im
a
ge
s,
"
C
omp
ute
r Me
t
ho
ds a
n
d Pro
gr
am
s in B
io
me
d
ic
i
ne
,
p.
105
58
1,
2020,
doi
:
1
0.
10
16
/j.
c
m
pb.
20
20.
1
05
58
1.
[6
]
I
.
G
oodf
e
l
lo
w,
Y.
B
e
ng
io a
n
d A.
C
o
ur
v
il
le
,
“
De
e
p L
e
a
r
nin
g,
”
C
a
m
br
id
ge
: M
A
: T
he
M
I
T P
r
e
s
s,
20
16
.
[7
]
T.
S
ha
n
th
i,
S
.
R
.
S
a
nd A.
R
a
ju,
"
Au
tom
a
tic
d
ia
g
n
osi
s of
sk
in d
ise
a
se
s
us
in
g c
o
nv
ol
ut
io
n ne
ur
a
l ne
tw
o
r
k,
"
Mic
ro
pr
oc
e
s
so
rs
a
nd Mic
ro
s
y
st
e
ms
,
vo
l.
76,
2
02
0,
do
i:
10.
1
01
6/
j.
m
ic
pr
o.
2
02
0.
1
03
07
4.
[8
]
D.
L
iu,
e
t al,
"
C
a
r
dia
c
m
a
gne
tic
r
e
s
ona
nc
e
im
a
ge
se
gm
e
nta
ti
on ba
se
d on c
o
nv
ol
ut
io
na
l ne
ur
a
l ne
tw
or
k,
"
C
omp
ute
r
Me
t
ho
ds a
nd
Pr
og
ra
ms i
n B
io
me
d
ic
i
ne
,
vo
l.
19
7,
p.
10
5
755,
2
02
0,
do
i:
10.
1
01
6/
j.
c
m
p
b.
2
02
0.
10
57
55.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1693
-
6930
T
E
L
KOM
NI
KA
T
el
eco
m
m
u
n
C
om
put
E
l
C
ont
r
o
l
,
Vo
l
.
1
9
, N
o
.
5
,
O
ct
o
b
er
2021
:
1581
-
1587
1586
[9
]
E.
P
é
r
e
z
,
O.
R
e
ye
s a
nd S
.
Ve
n
tur
a
,
"
C
on
vo
lu
ti
ona
l ne
u
r
a
l ne
t
wor
ks f
or
the
a
u
tom
a
tic
dia
gn
os
is of
m
e
la
n
om
a
:
An
e
xte
ns
ive
e
xpe
r
im
e
n
ta
l st
ud
y,
"
Me
d
ic
al
I
m
age
An
aly
si
s,
v
ol.
6
7,
p.
10
18
58,
20
20,
doi
:
1
0.
10
16
/j.
m
e
dia
.
20
20.
1
01
85
8.
[
10]
Z.
-
j.
L
u,
Q.
Qin,
H.
-
y.
S
hi a
n
d H.
H
ua
n
g,
"
S
AR
m
ov
i
ng ta
r
ge
t
im
a
g
in
g ba
se
d o
n c
on
vo
lu
ti
ona
l ne
ur
a
l
ne
t
wo
r
k,
"
Dig
it
al Si
gn
al
P
ro
c
e
s
si
ng,
v
ol.
1
06,
n
o.
1,
20
20,
d
oi
:
10.
101
6/
j.
d
sp.
2
02
0.
1
02
83
2.
[
11]
P
.
M
a
ji a
nd R
.
M
ul
li
ns,
"
On t
he
R
e
d
uc
t
io
n of
C
o
m
put
a
t
io
na
l C
om
ple
xi
ty of
De
e
p C
o
nv
ol
ut
io
na
l N
e
ur
a
l
Ne
t
wor
ks,
"
En
tr
opy
,
v
ol.
2
0,
no.
4,
p.
3
05,
2
01
8,
do
i:
1
0
.
339
0/e
20
04
03
05.
[
12]
E.
F
ix
a
n
d J.
J.
L
.
Ho
dge
s,
"
Di
sc
r
im
ina
tor
y A
na
l
ys
is
.
Non
pa
r
a
m
e
tr
ic
Di
sc
r
im
ina
ti
on
: C
o
ns
is
te
nc
y P
r
ope
r
ti
e
s,
"
I
nte
rn
at
io
na
l Sta
ti
st
ic
a
l Re
v
ie
w
/ Re
v
u
e
I
nte
rn
at
io
n
ale
de
St
at
is
ti
que
,
v
ol.
57,
n
o.
3,
p
p
.
238
-
2
47
,
1989
,
doi
:
1
0.
23
07
/1
40
37
97.
[
13]
H.
W
a
n
g,
I
.
Dü
nt
sc
h,
G
.
G
e
diga
a
nd G
.
G
uo,
"
Ne
a
r
e
st N
e
i
gh
bo
ur
s w
it
ho
ut
k,
"
Adv
anc
e
s i
n S
of
t C
o
mp
ut
in
g
Mo
ni
tor
in
g,
Se
c
u
ri
ty
,
an
d Re
sc
ue
T
e
c
hn
iq
ue
s i
n
Mul
ti
age
nt Sy
ste
ms,
vo
l.
28,
p
p.
1
79
-
18
9,
20
05,
doi
:
1
0.
10
07
/3
-
54
0
-
3
23
70
-
8_
12
.
[
14]
R
.
M
.
Kum
a
r
a
nd K.
S
r
e
e
k
um
a
r
,
"
A S
ur
ve
y o
n I
m
a
ge
F
e
a
tur
e
De
sc
r
i
pt
or
s,
"
I
nte
rn
at
io
na
l J
o
ur
na
l o
f C
om
pu
te
r
Sc
ie
nc
e
an
d I
n
fo
rm
at
io
n T
e
c
h
no
lo
gie
s,
vo
l.
5,
pp.
76
68
-
76
73,
2
01
4.
[
O
nl
ine
]
.
A
va
i
l
a
ble
:
htt
p:
//
ij
c
si
t.
c
om
/d
oc
s
/V
ol
um
e
%
20
5/
vo
l5
is
sue
06
/i
jc
s
it
20
140
50
61
68.
pdf
[
15]
A.
Hum
e
a
u
-
He
ur
tie
r
,
"
Te
x
tur
e
F
e
a
t
ur
e
Extr
a
c
ti
on M
e
t
hod
s: A S
ur
ve
y,
"
in
I
E
E
E
A
c
c
e
s
s
,
vol.
7,
pp.
89
75
-
9
00
0,
201
9,
do
i: 10.
1
10
9/
AC
C
ES
S
.
20
18.
2
89
07
43.
[
16]
F
.
G
.
M
oha
m
m
a
di,
M
.
H.
Am
in
i a
n
d
H.
R
.
Ar
a
b
nia
,
"
Evo
lu
ti
ona
r
y C
om
p
uta
ti
on,
O
pt
im
i
z
a
t
io
n,
a
nd
L
e
a
r
n
in
g
Alg
or
it
hm
s f
or
Da
ta
S
c
ie
nc
e
,
"
in
Opt
im
iz
at
io
n,
L
e
ar
nin
g,
an
d C
on
tro
l fo
r I
nte
rde
pe
n
de
n
t C
om
ple
x
N
e
tw
o
rk
s.
Adv
anc
e
s i
n
I
n
te
l
li
ge
n
t
Sy
s
te
m
s a
nd
C
om
pu
ti
ng
,
v
ol.
1
1
2
3,
C
ha
m
,
S
pr
i
nge
r
I
nte
r
na
t
io
na
l
P
u
bl
is
hi
ng,
2
0
20,
pp.
37
-
5
6,
do
i:
10
.
10
07
/9
78
-
3
-
0
30
-
34
09
4
-
0_
3.
[
17]
C
he
n,
Z
.
F
e
ng
-
Yo
u a
nd
Y.
X
ia
n
-
F
e
ng,
"
H
ybr
id
pa
r
t
ic
le
s
wa
r
m
o
pt
im
iz
a
ti
on
w
it
h s
pir
a
l
-
s
ha
pe
d
m
e
c
ha
ni
sm
f
or
f
e
a
tur
e
se
le
c
t
io
n,
"
Ex
pe
rt Sy
ste
ms w
it
h Ap
pl
ic
a
t
io
ns
,
vo
l
.
128,
pp.
1
40
-
1
56,
2
01
9,
do
i:
1
0.
1
01
6/
j.
e
swa
.
20
19.
0
3.
0
39.
[
18]
G
.
C
ha
ndr
a
she
ka
r
a
nd F
.
S
a
hi
n,
"
A sur
ve
y on f
e
a
t
ur
e
s
e
le
c
t
io
n m
e
th
od
s,
"
C
ompu
te
r
s an
d El
e
c
t
ric
al E
ng
ine
e
ri
ng,
vol.
4
0,
no.
1,
p
p
. 1
6
-
2
8,
20
14,
d
oi
:
10.
10
16
/j.
c
om
pe
le
c
e
ng.
2
01
3.
1
1.
02
4.
[
19]
S
.
S
a
ye
d,
M
.
Na
s
se
f
,
A.
B
a
dr
a
nd I
.
F
a
r
a
g,
"
A
Ne
s
te
d
G
e
ne
tic
Al
gor
it
hm
f
or
f
e
a
tur
e
se
le
c
ti
on
in
hi
gh
-
dim
e
ns
i
ona
l
c
a
nc
e
r
M
ic
r
oa
r
r
a
y da
ta
se
ts,
"
Ex
pe
rt
Sy
ste
ms w
ith Ap
pl
ic
a
ti
on
s,
vo
l.
121,
p
p
.
2
33
-
2
43,
2
019
,
doi
:1
0.
1
01
6/
j.
e
s
wa
.
20
18.
12.
0
22.
[
20]
S
.
Ja
dha
v,
H.
He
a
n
d K.
Je
nki
ns,
"
I
nf
or
m
a
ti
on ga
i
n dir
e
c
te
d ge
ne
tic
a
l
gor
it
hm
wr
a
p
pe
r
f
e
a
tur
e
se
le
c
ti
on f
or
c
r
e
di
t
r
a
ti
ng,
"
Ap
pl
ie
d S
of
t C
o
mp
ut
in
g,
vo
l.
69,
p.
5
41
-
55
3,
20
18,
do
i:
10.
1
01
6/
j.
a
s
oc
.
2
01
8.
04.
03
3.
[
21]
R
.
S
.
Wa
ho
no a
nd
N.
S
.
He
r
m
a
n,
"
G
e
ne
tic
F
e
a
t
ur
e
S
e
le
c
t
io
n f
or
S
of
t
w
a
r
e
D
e
f
e
c
t
,
"
A
dv
a
nc
e
d Sc
ie
nc
e
L
e
tt
e
rs,
vol.
2
0,
no.
1,
p
p
.
2
39
-
24
4,
Am
e
r
ic
a
n S
c
ie
n
tif
ic
P
u
bl
is
h
e
r
s
,
20
14,
d
oi
:
10.
11
66
/a
s
l.
20
14.
52
83.
[
22]
L.
P
.
C
oe
lho,
"
M
a
ho
ta
s
: Ope
n s
our
c
e
s
of
t
wa
r
e
f
or
sc
r
ip
t
a
ble
c
om
pu
te
r
vi
si
on,
"
J
ou
rn
al of
O
pe
n
Re
se
a
rc
h So
ftw
are
,
vol.
1,
n
o
. 1
,
p
p
.
1
-
7
,
2
01
3,
do
i:
1
0.
5
33
4/
jor
s.
a
c
.
[
23]
M
ie
r
s
wa
,
M
.
W
ur
st,
R
.
Kl
in
ke
n
be
r
g,
M
.
S
c
ho
lz
a
n
d
T.
E
ule
r
,
"
Y
AL
E: R
a
pi
d P
r
ot
ot
yp
in
g f
or
C
om
p
le
x
Da
ta
M
ini
ng Ta
sk
s,
"
in
Proc
e
e
di
ng
s of t
he
12
th AC
M SI
GK
DD I
nte
rn
at
io
na
l C
on
fe
re
nc
e
on K
now
le
d
ge
Di
sc
ov
e
ry
an
d
Dat
a M
in
in
g
,
P
hi
la
de
lp
hia
,
2
00
6
, p
p
.
93
5
-
9
40
,
do
i:
10.
1
145
/1
15
04
02.
11
50
53
1.
[
24]
R
.
M
.
Ha
r
a
lic
k,
K.
S
ha
nm
u
ga
m
a
nd I
.
Din
ste
in,
"
Te
xt
ur
a
l F
e
a
tur
e
s f
or
I
m
a
ge
C
la
s
sif
ic
a
t
io
n,
"
in
I
EEE
T
r
an
sac
ti
on
s
on Sy
ste
ms,
M
an,
a
nd C
y
be
r
ne
t
ic
s
, v
o
l
. S
M
C
-
3,
n
o.
6,
p
p.
610
-
6
21,
N
ov.
1
97
3,
do
i:
1
0.
1
10
9/
TS
M
C
.
19
73.
4
30
93
14.
[
25]
T.
Aho
ne
n,
A.
Ha
di
d a
nd M
.
P
ie
t
ika
ine
n,
"
F
a
c
e
De
sc
r
ip
ti
on w
it
h L
oc
a
l B
ina
r
y P
a
tte
r
ns
: A
pp
lic
a
ti
on t
o F
a
c
e
R
e
c
og
ni
ti
on,
"
i
n
I
E
EE T
ra
ns
ac
t
io
ns
o
n P
at
te
r
n A
na
ly
si
s a
nd
Mac
hi
ne
I
nte
ll
ige
nc
e
,
v
ol.
28,
n
o.
1
2,
p
p.
2
03
7
-
20
4
1,
De
c
.
20
06,
d
oi
: 10.
11
09
/TP
AM
I
.
20
06.
24
4.
[
26]
T.
Ha
s
tie
,
R
.
Ti
bsh
ir
a
ni a
nd
J.
F
r
ie
dm
a
n,
"
T
he
E
le
m
e
nt
s of
S
ta
t
is
tic
a
l L
e
a
r
ni
ng
Da
ta
M
i
ni
ng,
I
nf
e
r
e
nc
e
,
a
nd
P
r
e
dic
ti
on"
,
Sp
ri
nge
r Se
r
ie
s
in S
ta
ti
st
ic
s
,
v
ol.
2
7,
pp.
8
3
-
85
,
Ne
w Yor
k: S
pr
i
nge
r
,
2
00
9,
do
i:
1
0.
1
00
7/B
F
02
98
58
0
2.
[
27]
R
.
L
e
a
r
di,
"
Appl
ic
a
ti
on of
a
ge
ne
tic
a
lg
or
i
thm
to
f
e
a
tu
r
e
se
le
c
t
io
n u
nde
r
f
ul
l
va
li
da
t
io
n
c
o
nd
it
io
ns a
nd
t
o o
ut
lie
r
de
te
c
ti
on,
"
J
ou
rn
al o
f C
he
mo
me
t
ric
s,
vo
l.
8,
no.
1,
p
p
.
65
-
7
9,
19
94,
d
oi
:
10.
10
02
/c
e
m
.
11
80
08
01
07.
B
I
OGR
A
P
HI
E
S
OF
A
U
T
HOR
S
Rad
it
yo Ad
i
Nu
gr
oh
o
is a
n A
ss
is
ta
n
t P
r
of
e
s
sor
i
n the
De
pa
r
te
m
e
nt of
C
om
pu
te
r
S
c
ie
nc
e
a
t
L
a
m
bun
g M
a
ngk
ur
a
t U
ni
ve
r
s
it
y.
His r
e
se
a
r
c
h inte
r
e
st
s inc
l
ude
S
of
twa
r
e
De
f
e
c
t P
r
e
dic
t
io
n,
a
nd C
om
pu
te
r
V
is
io
n.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
KA
T
el
eco
m
m
u
n
C
om
put
E
l
C
ont
r
o
l
I
m
pr
ov
e
m
e
nt
on K
N
N
us
i
ng ge
ne
t
i
c
al
gor
i
t
hm
and c
om
bi
ne
d f
e
at
ur
e
e
x
t
r
ac
t
i
on
…
(
R
adi
t
y
o
A
di
N
ugr
oho
)
1587
Ar
ie
S
ap
t
a Nu
gr
ah
a
is a
n u
nde
r
gr
a
d
ua
te
s
tu
de
n
t in t
he
De
pa
r
te
m
e
nt of
C
om
pu
te
r
S
c
ie
nc
e
a
t
the
L
a
m
bu
ng M
a
n
gk
ur
a
t
Un
ive
r
si
ty a
n
d wi
ll be
gr
a
d
ua
t
ing
in 2
02
1.
Ar
ie
ha
s a
str
on
g in
te
r
e
s
t
in
t
he
f
ie
ld
of
M
a
c
h
i
n
e
L
e
a
r
nin
g a
n
d
So
ft
w
are
E
ng
ine
e
r
i
ng
.
Aylw
in
A
l R
as
yid
is a
n u
nde
r
gr
a
d
ua
te
s
tu
de
n
t of
the
C
om
pu
te
r
S
c
ie
nc
e
De
pa
r
tm
e
n
t,
F
a
c
u
lt
y
of
M
a
t
he
m
a
t
ic
s a
nd
Na
tur
a
l S
c
ie
nc
e
s,
L
a
m
bun
g M
a
ng
k
ur
a
t
Un
ive
r
si
ty.
Ay
lw
in
ha
s
i
nte
r
e
st
in
S
of
tw
a
r
e
En
gi
ne
e
r
i
ng a
n
d S
ys
te
m
s P
r
ogr
a
m
m
in
g.
B
e
si
de
s t
ha
t,
Ay
lw
in a
ls
o ha
s a
n i
nte
r
e
st
in
UI
/
UX de
si
gn.
F
e
n
n
y W
in
d
a Rah
ayu
i
s a
n unde
r
gr
a
d
ua
te
s
tu
de
n
t of
the
C
om
pu
te
r
S
c
ie
nc
e
De
pa
r
tm
e
n
t,
F
a
c
ul
ty
of
M
a
t
he
m
a
t
ic
s a
nd
Na
t
ur
a
l S
c
ie
n
c
e
s,
L
a
m
bu
n
g M
a
n
gk
ur
a
t
Un
ive
r
si
ty.
F
e
n
ny
ha
s a
n
int
e
r
e
s
t
in
S
of
t
wa
r
e
En
gi
ne
e
r
i
ng
a
nd
S
y
ste
m
s
P
r
o
gr
a
m
m
ing
Evaluation Warning : The document was created with Spire.PDF for Python.