I
A
E
S
I
n
t
e
r
n
at
io
n
al
Jou
r
n
al
of
A
r
t
if
ic
ia
l
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
V
ol
.
14
, N
o.
4
,
A
ugus
t
20
25
, pp.
3192
~
3200
I
S
S
N
:
2252
-
8938
,
D
O
I
:
10.11591/
ij
a
i.
v
14
.i
4
.pp
3192
-
3200
3192
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ij
ai
.
ia
e
s
c
or
e
.c
om
A
d
e
e
p
l
e
ar
n
i
n
g
-
b
ase
d
f
r
am
e
w
or
k
f
o
r
au
t
om
at
i
c
d
e
t
e
c
t
i
on
of
C
O
V
ID
-
19
u
si
n
g
c
h
e
st
X
-
r
ay an
d
C
T
-
sc
an
i
m
age
s
S
iv
an
agi
r
e
d
d
y K
al
li
1
, B
u
k
k
a N
ar
e
n
d
r
a K
u
m
ar
2
, S
aggu
r
t
h
i
Jagad
e
e
s
h
1
,
K
u
s
h
agar
i
C
h
an
d
r
am
ou
li
R
avi
K
u
m
ar
2
1
D
e
pa
r
t
m
e
nt
of
E
l
e
c
t
r
oni
c
s
a
nd C
om
m
uni
c
a
t
i
on E
ngi
ne
e
r
i
ng
, S
r
i
de
vi
W
om
e
n’
s
E
ngi
ne
e
r
i
ng C
ol
l
e
ge
, H
yde
r
a
ba
d, I
ndi
a
2
D
e
pa
r
t
m
e
nt
of
C
om
put
e
r
S
c
i
e
nc
e
a
nd E
ngi
ne
e
r
i
ng
, S
r
i
de
vi
W
om
e
n’
s
E
ngi
ne
e
r
i
ng C
ol
l
e
ge
, H
yde
r
a
ba
d, I
ndi
a
A
r
t
ic
le
I
n
f
o
A
B
S
T
R
A
C
T
A
r
ti
c
le
h
is
to
r
y
:
R
e
c
e
iv
e
d
A
ug
7, 2024
R
e
vi
s
e
d
A
pr
14, 2025
A
c
c
e
pt
e
d
J
un 8, 2025
COVID
-
19
has
profoundly
impacted
global
public
h
ealth,
underscori
ng
the
need
for
rapid
detection
methods.
Radiography
and
radiologic
im
aging,
especiall
y
chest
X
-
rays,
enable
swift
diagnosis
of
infected
individual
s.
This
study
delves
into
leveraging
machine
learning
to
identify
COVID
-
1
9
from
X
-
ray
images.
By
gathering
a
dataset
of
9
,
000
chest
X
-
rays
and
CT
scans
from
public
resources,
meticulously
vetted
by
board
-
licensed
radiologists
to
confirm
COVID
-
19
presence,
the
research
sets
a
robust
foun
dation.
However,
further
validation
is
essential
expanding
datasets
to
enco
mpass
enough
COVID
-
19
cases
enhances
convolut
ional
neural
network
(
CNN
)
accuracy.
Among
various
machine
learning
techniqu
es,
deep
learning
excels
in
identifying
distinct
patterns
on
imaging
characteristics
discernible
i
n
chest
radiographs
of
COVID
-
19
patients.
Yet,
extensive
validation
across
diverse
datasets
and
clinical
trials
is
crucial
to
ensure
the
robustnes
s
and
generalizability
of
these
models.
The
conversation
extends
into
complexi
ties,
includi
ng
ethical
considerat
ions
around
patient
priva
cy
and
integrating
intelligent
tech
into
clinical
workflows.
Collaborating
closel
y
with
healthcare
professionals
ensures
this
technology
complements
the
establi
shed
diagnost
ic
approach.
Despite
the
potenti
al
to
detect
CO
VID
-
19
using
chest
X
-
ray
imaging
findings,
thorough
research
and
validation,
alongsi
de
ethical
deliberati
ons,
are
vital
before
implem
enting
it
in
the
healthcare
field.
The
results
show
that
the
proposed
model
ac
hieved
classifi
cation
accuracy
and
F1
score
of
96%
and
98%,
respectively,
for
the
X
-
ray images.
K
e
y
w
o
r
d
s
:
C
onvolut
io
na
l
ne
ur
a
l
ne
twor
k
C
O
V
I
D
-
19
C
T
i
m
a
ge
s
D
e
e
p l
e
a
r
ni
ng
D
e
ns
e
N
e
t
-
121
This is an
open
acce
ss artic
le unde
r the
CC BY
-
SA
license.
C
or
r
e
s
pon
di
n
g A
u
th
or
:
S
iv
a
na
gi
r
e
ddy Ka
ll
i
D
e
pa
r
tm
e
nt
of
E
le
c
tr
oni
c
s
a
nd C
om
m
uni
c
a
ti
on E
ngi
ne
e
r
in
g
, S
r
id
e
vi
W
om
e
n’
s
E
ngi
ne
e
r
in
g C
ol
le
ge
H
yde
r
a
ba
d, T
e
la
ng
a
na
, I
ndi
a
E
m
a
il
:
s
iv
a
na
gi
r
e
ddyka
ll
i@gma
il
.c
om
1.
I
N
T
R
O
D
U
C
T
I
O
N
L
ung
di
s
e
a
s
e
s
s
p
a
n
a
va
s
t
a
r
r
a
y
of
c
ondi
ti
ons
,
im
pa
c
ti
ng
m
il
li
ons
gl
oba
ll
y.
B
r
oa
dl
y
c
a
te
gor
iz
e
d,
th
e
y
a
f
f
e
c
t
lu
ng
ti
s
s
ue
s
,
bl
ood
f
lo
w
,
a
nd
a
i
r
w
a
ys
.
D
is
or
de
r
s
s
uc
h
a
s
a
s
th
m
a
,
c
hr
oni
c
obs
tr
uc
ti
ve
pul
m
ona
r
y
di
s
e
a
s
e
(
C
O
P
D
)
,
a
nd
pne
um
oni
a
pr
im
a
r
il
y
obs
tr
uc
t
a
ir
w
a
ys
,
im
pa
ir
in
g
th
e
body’
s
oxyge
n
a
bs
or
pt
io
n
a
nd
c
a
r
bon
di
oxi
de
e
xpul
s
io
n.
B
lo
od
f
lo
w
c
ondi
ti
ons
li
ke
pul
m
ona
r
y
e
m
bol
is
m
a
nd
pul
m
ona
r
y
hype
r
te
ns
io
n
im
pe
de
oxyge
n
tr
a
ns
por
t
due
to
a
r
te
r
ia
l
c
lo
tt
in
g
in
th
e
lu
ngs
.
A
ddi
ti
ona
ll
y,
di
s
e
a
s
e
s
s
u
c
h
a
s
s
a
r
c
oi
dos
i
s
a
nd
pul
m
ona
r
y
f
ib
r
os
is
f
ur
th
e
r
c
om
pl
ic
a
te
r
e
s
pi
r
a
to
r
y
he
a
lt
h.
T
he
C
O
V
I
D
-
19
pa
nde
m
ic
,
in
s
ti
ga
te
d
by
th
e
S
A
R
S
-
C
oV
-
2
vi
r
us
,
ha
s
he
ig
ht
e
ne
d
gl
oba
l
c
onc
e
r
ns
a
bout
r
e
s
pi
r
a
t
or
y
he
a
lt
h.
C
om
m
on
s
ym
pt
om
s
in
c
lu
de
d
r
y
c
ough,
f
a
ti
gue
,
r
e
s
pi
r
a
to
r
y
di
s
tr
e
s
s
r
a
ngi
ng
f
r
om
m
il
d
to
s
e
ve
r
e
,
lo
s
s
of
ta
s
te
or
s
m
e
ll
,
a
nd
f
e
ve
r
.
T
he
vi
r
us
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
de
e
p l
e
ar
ni
ng
-
bas
e
d f
r
am
e
w
or
k
f
or
aut
om
at
ic
de
te
c
ti
on C
O
V
I
D
-
19 us
in
g c
he
s
t
…
(
Si
v
anagir
e
dd
y
K
al
li
)
3193
s
pr
e
a
ds
th
r
ough
r
e
s
pi
r
a
to
r
y
dr
opl
e
ts
e
xpe
ll
e
d
w
he
n
in
f
e
c
t
e
d
in
di
vi
dua
ls
s
n
e
e
z
e
,
c
ough,
or
ta
lk
,
pos
in
g
s
ig
ni
f
ic
a
nt
r
is
ks
to
vul
ne
r
a
bl
e
popula
ti
ons
,
in
c
lu
di
ng
th
os
e
w
it
h
unde
r
ly
in
g
he
a
lt
h
c
ondi
ti
ons
li
ke
di
a
be
te
s
,
c
hr
oni
c
r
e
s
pi
r
a
to
r
y
di
s
e
a
s
e
s
,
c
a
r
di
ova
s
c
ul
a
r
di
s
or
de
r
s
,
a
nd
c
a
nc
e
r
.
O
ld
e
r
a
dul
ts
w
it
h
m
ul
ti
pl
e
h
e
a
lt
h
is
s
ue
s
a
r
e
pa
r
ti
c
ul
a
r
ly
s
us
c
e
pt
ib
le
t
o s
e
ve
r
e
i
ll
ne
s
s
f
r
om
C
O
V
I
D
-
19.
D
ia
gnos
ti
c
s
tr
a
te
gi
e
s
f
or
C
O
V
I
D
-
19
e
nc
om
pa
s
s
a
nt
ig
e
n
te
s
ts
to
de
te
c
t
c
ur
r
e
nt
in
f
e
c
ti
on
a
nd
a
nt
ib
ody
te
s
ts
to
a
s
c
e
r
ta
in
pa
s
t
e
xpos
ur
e
.
W
id
e
ly
us
e
d
P
C
R
te
s
ts
de
t
e
c
t
vi
r
a
l
R
N
A
f
r
om
na
s
opha
r
ynge
a
l
s
w
a
bs
w
it
h
hi
gh
a
c
c
ur
a
c
y.
H
ow
e
ve
r
,
gl
oba
l
c
ont
a
in
m
e
nt
e
f
f
or
ts
f
a
c
e
c
ha
ll
e
nge
s
due
to
li
m
it
e
d
te
s
ti
ng
c
a
pa
c
it
ie
s
,
une
ve
n
r
e
s
our
c
e
di
s
tr
ib
ut
io
n,
a
nd
th
e
a
b
s
e
nc
e
o
f
uni
ve
r
s
a
ll
y
a
va
il
a
bl
e
v
a
c
c
in
e
s
or
s
p
e
c
if
ic
tr
e
a
tm
e
nt
s
.
I
n
di
a
gno
s
in
g
C
O
V
I
D
-
19,
m
e
di
c
a
l
im
a
gi
ng
t
e
c
hnol
ogi
e
s
pl
a
y
a
c
r
uc
i
a
l
r
ol
e
,
e
s
pe
c
ia
ll
y
in
r
a
pi
dl
y
a
s
s
e
s
s
in
g
lu
ng
a
bnor
m
a
li
ti
e
s
.
C
T
s
c
a
ns
a
nd
X
-
r
a
ys
c
a
pt
ur
e
d
e
ta
il
e
d
c
he
s
t
im
a
ge
s
,
w
it
h
C
T
s
c
a
ns
of
f
e
r
in
g
th
r
e
e
-
di
m
e
ns
io
na
l
vi
e
w
s
th
a
t
e
nha
nc
e
vi
s
ua
li
z
a
ti
on
of
in
t
e
r
na
l
or
ga
ns
a
nd
pr
e
c
i
s
e
di
s
e
a
s
e
lo
c
a
ti
on
id
e
nt
if
ic
a
ti
on.
I
n
c
ont
r
a
s
t,
X
-
r
a
ys
pr
ovi
de
two
-
di
m
e
ns
io
na
l
im
a
ge
s
s
ui
ta
bl
e
f
or
e
xa
m
in
in
g
de
n
s
e
ti
s
s
ue
s
but
la
c
k
de
pt
h
pe
r
c
e
pt
io
n.
D
e
s
pi
te
th
e
be
ne
f
it
s
of
C
T
s
c
a
ns
,
a
c
c
e
s
s
to
hi
gh
-
qua
li
ty
im
a
gi
ng
e
qui
pm
e
nt
va
r
ie
s
gl
oba
ll
y.
R
e
s
e
a
r
c
h
e
r
s
in
c
r
e
a
s
in
gl
y
f
oc
us
on
us
in
g
c
he
s
t
C
T
s
c
a
ns
to
di
a
gnos
e
C
O
V
I
D
-
19
due
to
th
e
ir
s
upe
r
io
r
im
a
gi
ng
c
a
pa
bi
li
ti
e
s
.
H
ow
e
ve
r
,
a
c
c
ur
a
te
in
te
r
pr
e
ta
ti
on
r
e
qui
r
e
s
s
ki
ll
e
d
r
a
di
ol
ogi
s
t
s
to
pr
om
pt
ly
a
nd
a
c
c
ur
a
te
ly
i
de
nt
if
y C
O
V
I
D
-
19
-
r
e
la
te
d a
bnor
m
a
li
ti
e
s
.
T
o
s
tr
e
a
m
li
ne
C
O
V
I
D
-
19
di
a
gnos
is
a
nd
r
e
duc
e
hum
a
n
in
te
r
ve
nt
io
n,
r
e
s
e
a
r
c
he
r
s
de
ve
lo
p
a
ut
om
a
te
d
di
a
gnos
ti
c
m
ode
ls
.
O
ne
in
nova
ti
ve
a
ppr
oa
c
h
in
vol
ve
s
a
th
r
e
e
-
s
ta
ge
m
e
th
odol
ogy
in
c
or
por
a
ti
ng
w
a
ve
le
t
-
e
nha
nc
e
d
da
ta
a
ugm
e
nt
a
ti
on,
di
s
e
a
s
e
de
te
c
ti
on,
a
nd
a
nom
a
ly
lo
c
a
li
z
a
ti
on.
T
hi
s
a
ppr
oa
c
h
c
om
pe
ns
a
te
s
f
or
li
m
it
e
d
C
O
V
I
D
-
19
C
T
im
a
ge
s
by
pr
e
-
p
r
oc
e
s
s
in
g
th
e
m
us
in
g
s
ta
ti
ona
r
y
w
a
ve
le
ts
to
e
nha
nc
e
f
e
a
tu
r
e
s
a
nd
m
it
ig
a
te
ove
r
-
f
it
ti
ng.
S
ubs
e
que
nt
ly
,
e
a
c
h
im
a
ge
unde
r
goe
s
tr
a
ns
f
or
m
a
ti
ons
li
ke
s
he
a
r
in
g,
r
ot
a
ti
on,
a
nd
tr
a
ns
la
ti
on
f
or
e
f
f
e
c
ti
ve
da
ta
s
e
t
a
ugm
e
nt
a
ti
on.
I
n
th
e
s
e
c
ond
s
ta
ge
,
tr
a
ns
f
e
r
le
a
r
ni
ng
te
c
hni
que
s
c
la
s
s
if
y
C
T
s
c
a
ns
in
to
C
O
V
I
D
a
nd
non
-
C
O
V
I
D
c
a
t
e
gor
ie
s
.
M
ode
l
s
li
ke
R
e
s
N
e
t5
0,
w
it
h
a
ddi
ti
ona
l
c
onvolut
io
na
l
la
ye
r
s
f
or
e
nha
nc
e
d
f
e
a
tu
r
e
e
xt
r
a
c
ti
on,
opt
im
iz
e
c
la
s
s
if
ic
a
ti
on
a
c
c
u
r
a
c
y.
T
he
be
s
t
-
pe
r
f
or
m
in
g
m
ode
l
is
s
e
le
c
te
d
ba
s
e
d
on
c
om
pa
r
is
ons
w
it
h
be
n
c
hm
a
r
k
tr
a
ns
f
e
r
le
a
r
ni
ng
m
ode
ls
,
e
n
s
ur
in
g
r
obus
t
pe
r
f
or
m
a
nc
e
in
di
s
ti
ngui
s
hi
ng
C
O
V
I
D
-
19
-
r
e
la
te
d
a
nom
a
li
e
s
f
r
om
nor
m
a
l
lu
ng
im
a
ge
s
.
I
n
th
e
f
in
a
l
s
ta
ge
,
th
e
s
e
le
c
te
d
m
ode
l’
s
f
e
a
tu
r
e
m
a
p
s
a
nd
a
c
ti
va
ti
on
la
ye
r
s
de
te
c
t
a
nom
a
li
e
s
w
it
hi
n
c
he
s
t
C
T
im
a
ge
s
of
C
O
V
I
D
-
19
-
pos
it
iv
e
pa
ti
e
nt
s
,
f
a
c
il
it
a
ti
ng
e
a
r
ly
a
nd
a
c
c
ur
a
te
c
li
ni
c
a
l
in
te
r
ve
nt
io
n.
T
hi
s
r
e
s
e
a
r
c
h
a
im
s
to
e
va
lu
a
t
e
a
nd
c
om
p
a
r
e
tr
a
ns
f
e
r
le
a
r
ni
ng
m
ode
ls
'
p
e
r
f
or
m
a
nc
e
in
de
te
c
ti
ng
C
O
V
I
D
-
19
f
r
om
C
T
s
c
a
n
s
u
s
in
g
a
li
m
it
e
d
da
ta
s
e
t.
A
ddi
ti
ona
ll
y,
th
e
s
tu
dy
e
xpl
or
e
s
in
nova
ti
ve
da
ta
a
ugm
e
n
ta
ti
on
te
c
hni
que
s
to
e
nha
nc
e
f
e
a
tu
r
e
m
a
p
in
te
r
pr
e
ta
bi
li
ty
i
n de
e
pe
r
ne
ur
a
l
ne
twor
k l
a
ye
r
s
, i
m
pr
ovi
ng
di
a
gnos
ti
c
a
c
c
ur
a
c
y a
nd r
e
li
a
bi
li
ty
.
A
dva
nc
e
m
e
nt
s
in
m
e
di
c
a
l
im
a
gi
ng
a
nd
a
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
hol
d
pr
om
is
e
f
or
e
nha
nc
in
g
C
O
V
I
D
-
19
di
a
gnos
is
a
nd
m
a
na
ge
m
e
nt
.
B
y
le
ve
r
a
gi
ng
a
ut
om
a
te
d
di
a
gnos
ti
c
m
ode
ls
a
nd
in
nova
ti
ve
im
a
ge
pr
oc
e
s
s
in
g
te
c
hni
que
s
,
r
e
s
e
a
r
c
he
r
s
a
im
to
ove
r
c
om
e
e
xi
s
ti
ng
c
ha
ll
e
nge
s
,
e
xp
e
di
te
C
O
V
I
D
-
19
in
f
e
c
ti
on
id
e
nt
if
ic
a
ti
on
f
r
om
c
he
s
t
C
T
s
c
a
ns
,
s
uppor
t
ti
m
e
ly
he
a
lt
hc
a
r
e
i
nt
e
r
ve
nt
io
ns
, a
nd mi
ti
ga
te
t
he
pa
nde
m
ic
'
s
i
m
pa
c
t
on globa
l
he
a
lt
h
s
ys
te
m
s
.
2.
L
I
T
E
R
A
T
U
R
E
S
U
R
V
E
Y
T
he
C
O
V
I
D
-
19
pa
nde
m
ic
,
c
a
us
e
d
by
th
e
S
A
R
S
-
C
oV
-
2
vi
r
us
,
ha
s
s
tr
a
in
e
d
he
a
lt
hc
a
r
e
s
ys
te
m
s
w
or
ld
w
id
e
,
ne
c
e
s
s
it
a
ti
ng
e
f
f
ic
ie
nt
a
nd
a
c
c
ur
a
te
di
a
gnos
ti
c
to
ol
s
.
E
a
r
ly
a
nd
a
c
c
ur
a
te
di
a
gnos
i
s
is
c
r
uc
ia
l
f
or
e
f
f
e
c
ti
ve
tr
e
a
tm
e
nt
a
nd
c
ont
a
in
m
e
nt
.
W
hi
le
R
T
-
P
C
R
te
s
ts
a
r
e
th
e
s
ta
nd
a
r
d,
im
a
gi
ng
te
c
hni
que
s
li
ke
c
h
e
s
t
X
-
r
a
ys
a
nd
C
T
s
c
a
ns
ha
ve
pr
ove
n
in
va
lu
a
bl
e
in
di
a
gnos
in
g
a
nd
m
oni
to
r
in
g
lu
ng
a
bnor
m
a
li
ti
e
s
a
s
s
oc
ia
te
d
w
it
h
C
O
V
I
D
-
19.
R
e
c
e
nt
a
dva
nc
e
m
e
nt
s
in
de
e
p
l
e
a
r
ni
ng
of
f
e
r
pr
om
is
in
g
s
ol
ut
io
ns
f
or
a
ut
om
a
ti
ng
a
nd
e
nha
nc
in
g
th
e
s
e
di
a
gnos
ti
c
pr
oc
e
s
s
e
s
[
1]
.
C
he
s
t
X
-
r
a
ys
pr
ovi
d
e
a
qui
c
k
a
nd
a
c
c
e
s
s
ib
le
m
e
th
od
f
or
de
te
c
ti
ng
lu
ng
a
bnor
m
a
li
ti
e
s
,
a
lt
hough
th
e
y
of
f
e
r
li
m
it
e
d
de
pt
h
pe
r
c
e
pt
io
n.
C
T
s
c
a
n
s
,
on
th
e
ot
h
e
r
ha
nd,
pr
ovi
de
de
ta
il
e
d
th
r
e
e
-
di
m
e
ns
io
na
l
im
a
ge
s
,
a
ll
ow
in
g
f
or
a
m
or
e
c
o
m
pr
e
he
ns
iv
e
a
s
s
e
s
s
m
e
nt
of
lu
ng
c
ondi
ti
ons
.
S
tu
di
e
s
ha
ve
s
how
n
th
a
t
C
T
s
c
a
ns
a
r
e
m
or
e
s
e
n
s
it
iv
e
in
de
te
c
ti
ng
C
O
V
I
D
-
19
-
r
e
la
te
d
lu
ng
a
bnor
m
a
li
ti
e
s
c
om
pa
r
e
d
to
c
he
s
t
X
-
r
a
ys
[
2]
.
C
O
V
I
D
-
19
e
xhi
bi
ts
di
s
ti
nc
t
im
a
gi
ng
f
e
a
tu
r
e
s
on
c
he
s
t
X
-
r
a
ys
a
nd
C
T
s
c
a
ns
,
in
c
lu
di
ng
gr
ound
-
gl
a
s
s
opa
c
it
ie
s
,
c
ons
ol
id
a
ti
on,
a
nd
bi
la
te
r
a
l
lu
ng
in
vol
ve
m
e
nt
.
I
de
nt
if
yi
ng
th
e
s
e
f
e
a
tu
r
e
s
a
c
c
ur
a
te
ly
r
e
qui
r
e
s
s
ki
ll
e
d
r
a
di
ol
ogi
s
ts
,
w
hi
c
h c
a
n
be
a
bot
tl
e
ne
c
k
in
a
r
e
a
s
w
it
h
li
m
it
e
d
m
e
di
c
a
l
e
xpe
r
ti
s
e
[
3]
.
C
onvolut
io
na
l
ne
ur
a
l
ne
twor
ks
(
C
N
N
s
)
ha
v
e
r
e
vol
ut
io
ni
z
e
d
im
a
ge
a
na
ly
s
i
s
by
a
ut
om
a
ti
ng
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
a
nd
c
la
s
s
if
ic
a
ti
on
pr
oc
e
s
s
e
s
.
T
h
e
s
e
n
e
twor
ks
ha
v
e
s
how
n
r
e
m
a
r
ka
bl
e
s
uc
c
e
s
s
in
m
e
di
c
a
l
im
a
gi
ng
ta
s
k
s
,
in
c
lu
di
ng
di
s
e
a
s
e
de
te
c
ti
on
a
nd
a
nom
a
ly
lo
c
a
li
z
a
ti
on
[
4]
.
T
r
a
ns
f
e
r
le
a
r
ni
ng
in
vol
ve
s
pr
e
-
t
r
a
in
in
g
a
ne
ur
a
l
ne
twor
k
on
a
la
r
ge
da
ta
s
e
t
a
nd
f
in
e
-
tu
ni
ng
it
on
a
s
pe
c
if
ic
ta
s
k.
T
hi
s
a
ppr
oa
c
h
is
pa
r
ti
c
ul
a
r
ly
us
e
f
ul
in
m
e
di
c
a
l
im
a
gi
ng, whe
r
e
l
a
be
le
d da
ta
i
s
of
te
n s
c
a
r
c
e
.
M
o
de
ls
li
k
e
R
e
s
N
e
t,
V
G
G
,
a
n
d
d
e
n
s
e
l
y
c
on
ne
c
t
e
d
c
o
nv
ol
u
ti
o
na
l
ne
tw
or
k
s
(
D
e
n
s
e
N
e
t
)
h
a
v
e
b
e
e
n
e
f
f
e
c
t
iv
e
ly
e
m
p
lo
y
e
d
i
n
tr
a
n
s
f
e
r
l
e
a
r
ni
ng
f
or
C
O
V
I
D
-
19
d
e
t
e
c
ti
on
[
5]
.
S
e
v
e
r
a
l
s
tu
d
ie
s
h
a
v
e
e
xpl
or
e
d
th
e
u
s
e
of
de
e
p
l
e
a
r
ni
n
g
m
od
e
l
s
f
or
a
ut
om
a
ti
c
C
O
V
I
D
-
19
d
e
t
e
c
ti
on
f
r
om
c
he
s
t
X
-
r
a
y
s
.
A
p
os
to
l
op
oul
os
a
nd
M
p
e
s
i
a
n
a
[
6]
de
ve
lo
pe
d
a
C
N
N
-
ba
s
e
d
m
ode
l
a
c
hi
e
vi
ng
hi
gh
a
c
c
ur
a
c
y
in
c
la
s
s
if
yi
ng
C
O
V
I
D
-
19
c
a
s
e
s
. T
he
y
u
s
e
d
a
da
ta
s
e
t
of
c
he
s
t
X
-
r
a
ys
in
c
lu
di
ng
C
O
V
I
D
-
19,
pne
um
oni
a
,
a
nd
he
a
lt
hy c
a
s
e
s
,
de
m
ons
tr
a
ti
ng
th
e
pot
e
nt
ia
l
of
C
N
N
s
in
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
4
,
A
ugus
t
2025
:
3192
-
3200
3194
di
s
ti
ng
ui
s
hi
ng
b
e
tw
e
e
n
th
e
s
e
c
on
di
ti
on
s
.
O
z
tu
r
k
e
t
al
.
[
7]
i
nt
r
oduc
e
d
a
de
e
p
l
e
a
r
ni
n
g
m
od
e
l
c
a
pa
bl
e
of
de
te
c
ti
ng
C
O
V
I
D
-
19
f
r
om
c
h
e
s
t
X
-
r
a
y
s
w
it
h
a
n
a
c
c
ur
a
c
y
of
98.08%
.
T
he
ir
a
p
pr
oa
c
h
e
m
pl
o
ye
d
a
D
a
r
kN
e
t
m
ode
l,
hi
g
hl
ig
ht
in
g
t
he
e
f
f
e
c
ti
ve
ne
s
s
of
de
e
p
le
a
r
ni
ng
a
r
c
hi
te
c
t
ur
e
s
in
m
e
di
c
a
l
im
a
ge
a
na
ly
s
i
s
.
C
T
s
c
a
n
s
of
f
e
r
hi
ghe
r
s
e
n
s
it
iv
it
y
a
nd
s
p
e
c
if
i
c
it
y
in
de
t
e
c
ti
n
g
C
O
V
I
D
-
19
-
r
e
l
a
te
d
lu
ng
a
bnor
m
a
li
ti
e
s
.
W
a
ng
e
t
al
.
[
8]
pr
opo
s
e
d
a
de
e
p
le
a
r
ni
ng
m
ode
l
us
in
g
a
da
t
a
s
e
t
of
c
he
s
t
C
T
im
a
g
e
s
t
o
de
te
c
t
C
O
V
I
D
-
19
w
it
h
a
n
a
c
c
ur
a
c
y
of
82.
9%
.
T
he
ir
m
od
e
l
ut
il
i
z
e
d
a
c
o
m
bi
n
a
ti
on
of
C
N
N
a
nd
lo
ng
s
hor
t
-
te
r
m
m
e
m
or
y
(
L
S
T
M
)
ne
t
w
or
k
s
to
c
a
pt
ur
e
s
p
a
ti
a
l
a
nd
te
m
p
or
a
l
f
e
a
t
ur
e
s
f
r
om
C
T
s
c
a
ns
.
J
a
i
s
w
a
l
e
t
al
.
[
9]
e
m
p
lo
ye
d
a
tr
a
n
s
f
e
r
le
a
r
ni
ng
a
ppr
o
a
c
h
u
s
in
g
th
e
V
G
G
16
m
ode
l,
a
c
h
ie
vi
ng
a
n
a
c
c
ur
a
c
y
of
91%
i
n
c
la
s
s
if
y
in
g
C
O
V
I
D
-
19
f
r
om
c
he
s
t
C
T
im
a
ge
s
.
T
h
e
ir
s
tu
d
y
de
m
on
s
tr
a
t
e
d
t
he
im
por
t
a
nc
e
of
da
t
a
a
ugm
e
nt
a
ti
on
a
nd
pr
e
pr
oc
e
s
s
in
g
in
e
nha
nc
in
g
m
o
de
l
pe
r
f
or
m
a
nc
e
.
C
om
pa
r
a
ti
v
e
s
tu
di
e
s
ha
ve
s
how
n
th
a
t
c
om
bi
ni
ng
c
h
e
s
t
X
-
r
a
y
s
a
nd
C
T
s
c
a
n
s
c
a
n
im
pr
o
ve
di
a
gno
s
ti
c
a
c
c
ur
a
c
y.
F
or
in
s
ta
n
c
e
,
S
ong
e
t
al
.
[
10]
de
v
e
lo
p
e
d
a
hybr
id
m
od
e
l
i
nt
e
g
r
a
ti
ng
f
e
a
tu
r
e
s
f
r
om
bot
h
im
a
gi
ng
m
oda
li
ti
e
s
,
a
c
hi
e
vi
ng
s
u
pe
r
io
r
pe
r
f
or
m
a
nc
e
c
om
pa
r
e
d
t
o
m
od
e
l
s
u
s
in
g
a
s
i
ngl
e
m
oda
li
t
y.
T
h
e
ir
r
e
s
e
a
r
c
h
e
m
p
ha
s
iz
e
s
th
e
pot
e
nt
i
a
l
of
m
ul
ti
-
m
od
a
l
a
ppr
oa
c
he
s
in
m
e
di
c
a
l
im
a
g
in
g.
A
s
t
udy
by
Y
a
n
a
nd
L
im
in
g
[
11]
e
va
lu
a
te
d
c
h
e
s
t
X
-
r
a
ys
'
pe
r
f
or
m
a
nc
e
in
di
a
gno
s
in
g
C
O
V
I
D
-
19
a
nd
a
s
s
e
s
s
e
d
r
a
di
ol
ogi
s
t
in
te
r
pr
e
ta
ti
ons
'
a
c
c
ur
a
c
y.
T
he
uni
que
C
T
c
ha
r
a
c
te
r
is
ti
c
s
of
C
O
V
I
D
-
19
pr
ovi
de
d
c
r
uc
ia
l
in
s
ig
ht
s
f
or
he
a
lt
hc
a
r
e
pr
of
e
s
s
io
na
ls
on
di
s
ti
ngui
s
hi
ng
C
O
V
I
D
-
19 f
r
om
ot
he
r
v
ir
a
l
pne
um
oni
a
s
ba
s
e
d on im
a
gi
ng f
in
d
in
gs
. T
he
e
m
e
r
ge
nc
e
of
t
he
nove
l
c
or
ona
vi
r
us
,
S
A
R
S
-
C
oV
-
2,
unde
r
s
c
or
e
d
th
e
im
por
ta
nc
e
of
a
c
c
ur
a
te
a
nd
ti
m
e
ly
di
a
gnos
ti
c
im
a
gi
ng
in
m
a
na
gi
ng
gl
ob
a
l
he
a
lt
h
c
r
is
is
.
T
r
a
ns
f
e
r
le
a
r
ni
ng
ha
s
pr
ove
n
e
f
f
e
c
ti
ve
in
a
ddr
e
s
s
in
g
th
e
a
nom
a
ly
de
te
c
ti
on
c
ha
ll
e
nge
in
s
m
a
ll
m
e
di
c
a
l
im
a
ge
da
ta
s
e
ts
,
d
e
m
ons
tr
a
ti
ng
pr
om
is
in
g
r
e
s
ul
ts
in
di
s
ti
ngui
s
hi
ng
C
O
V
I
D
-
19
c
a
s
e
s
f
r
om
ot
he
r
r
e
s
pi
r
a
to
r
y
c
ondi
ti
ons
ba
s
e
d
on
im
a
ge
f
e
a
tu
r
e
s
.
O
ne
of
m
a
in
c
ha
ll
e
nge
s
in
de
ve
lo
pi
ng
r
obus
t
de
e
p
le
a
r
ni
ng
m
ode
ls
f
or
C
O
V
I
D
-
19
de
te
c
ti
on
is
s
c
a
r
c
it
y
a
nd
qua
li
ty
of
a
n
not
a
te
d
da
ta
s
e
ts
.
L
a
r
ge
-
s
c
a
le
,
di
ve
r
s
e
da
ta
s
e
ts
a
r
e
c
r
uc
ia
l
f
or
t
r
a
in
in
g m
ode
ls
t
ha
t
ge
ne
r
a
li
z
e
w
e
ll
a
c
r
os
s
di
f
f
e
r
e
nt
popula
ti
ons
a
nd i
m
a
gi
ng de
vi
c
e
s
[
12]
.
T
h
e
w
e
ll
-
kn
ow
n D
e
e
p C
N
N
e
s
ta
bl
i
s
h
e
s
th
e
c
r
it
i
c
a
l
b
e
n
c
hm
a
r
k
s
by
W
a
n
g
e
t
a
l
.
[
8]
,
a
nd
M
in
e
t
al
.
[
13]
pr
opos
e
d
ne
twor
k
in
ne
twor
k
,
e
nha
n
c
in
g
C
N
N
s
w
it
h
m
ic
r
o
m
ul
ti
la
ye
r
pe
r
c
e
pt
r
on
(
M
L
P
)
f
or
im
pr
ove
d
f
e
a
tu
r
e
a
bs
tr
a
c
ti
on
a
nd
c
la
s
s
if
ic
a
ti
on
a
c
c
ur
a
c
y
.
N
i
s
hi
ur
a
e
t
al
.
[
14]
a
na
ly
z
e
d
C
O
V
I
D
-
19
s
e
r
ia
l
in
te
r
va
ls
,
pr
ovi
di
ng
e
s
s
e
nt
ia
l
da
ta
f
or
tr
a
ns
m
is
s
io
n
m
ode
li
ng
a
nd
e
a
r
ly
out
br
e
a
k
r
e
s
pons
e
s
tr
a
te
gi
e
s
.
I
m
a
ge
s
di
r
e
c
tl
y
obt
a
in
e
d
f
r
om
pa
ti
e
nt
s
s
uf
f
e
r
in
g
w
it
h
s
e
ve
r
e
C
O
V
I
D
-
19
o
r
p
n
e
um
oni
a
a
r
e
us
e
d
in
th
is
s
tu
dy
[
15]
–
[
19]
.
T
he
l
a
c
k
of
C
T
s
c
a
n
s
w
it
h
th
e
la
be
l
"
da
ta
"
in
r
a
di
ol
ogy
[
20]
.
A
ddi
ti
ona
ll
y,
th
e
pr
e
tr
a
in
e
d
C
N
N
m
ode
l
a
nd
t
e
xt
ur
e
de
s
c
r
ip
to
r
s
[
21]
.
T
he
goa
l
of
th
e
I
m
a
gi
ng
C
O
V
I
D
-
19
A
I
i
ni
ti
a
ti
ve
in
E
ur
ope
[
22]
,
[
23
]
a
nd
th
e
r
a
di
ol
ogi
c
a
l
s
oc
ie
ty
of
N
or
th
A
m
e
r
ic
a
(
R
S
N
A
)
[
24]
,
[
25]
is
to
m
a
ke
da
ta
e
a
s
il
y
a
c
c
e
s
s
ib
le
to
th
e
ge
ne
r
a
l
publ
ic
.
W
it
h
th
e
he
lp
of
th
e
s
e
da
ta
,
di
f
f
e
r
e
nt
f
e
a
tu
r
e
s
f
r
om
di
f
f
e
r
e
nt
c
a
te
go
r
ie
s
c
a
n
im
pr
ove
in
te
r
c
la
s
s
va
r
ia
nc
e
,
w
hi
c
h
im
pr
ove
s
de
e
p
le
a
r
ni
ng
pe
r
f
or
m
a
nc
e
.
T
h
e
m
ode
l
w
il
l
ove
r
f
it
a
nd
yi
e
ld
c
onc
lu
s
io
n
s
th
a
t
a
r
e
onl
y
w
e
a
kl
y
ge
ne
r
a
li
z
e
d
due
to
a
pa
uc
it
y
of
da
ta
[
26]
,
[
27]
.
T
he
r
e
f
or
e
,
it
h
a
s
be
e
n
de
m
on
s
tr
a
te
d
th
a
t
d
a
ta
a
ugm
e
nt
a
ti
on
w
or
ks
w
e
ll
f
or
tr
a
in
in
g
di
s
c
r
im
in
a
ti
ve
de
e
p
le
a
r
ni
ng
m
ode
l
s
.
F
li
ppi
ng,
r
ot
a
ti
ng,
c
ol
or
ji
tt
e
r
in
g,
r
a
ndom
c
r
oppi
ng,
e
la
s
ti
c
di
s
to
r
ti
ons
,
a
nd s
ynt
he
ti
c
d
a
ta
s
ynt
he
s
i
s
us
in
g
ge
ne
r
a
ti
ve
a
dve
r
s
a
r
ia
l
ne
twor
ks
(
G
A
N
s
)
a
r
e
a
f
e
w
e
xa
m
pl
e
s
of
da
t
a
a
ugm
e
nt
a
ti
on
te
c
hni
que
s
[
28]
,
[
29]
.
S
e
ve
r
a
l
vi
s
ua
l
tr
a
it
s
of
th
e
m
e
di
c
a
l
phot
os
in
I
m
a
ge
N
e
t
e
xhi
bi
t
s
tr
ong
in
te
r
c
la
s
s
s
im
il
a
r
it
y
[
30]
,
[
31]
.
A
s
a
r
e
s
ul
t,
c
onve
nt
io
na
l
a
ugm
e
nt
a
ti
on
te
c
hni
que
s
th
a
t
ju
s
t
m
a
ke
m
in
or
i
m
a
ge
a
dj
us
tm
e
nt
s
a
r
e
l
e
s
s
s
uc
c
e
s
s
f
ul
[
32]
.
3.
P
R
O
P
O
S
E
D
M
E
T
H
O
D
I
n
m
e
di
c
a
l
im
a
gi
ng,
de
e
p
le
a
r
ni
ng
te
c
hni
que
s
ha
ve
e
m
e
r
ge
d
a
s
pow
e
r
f
ul
to
ol
s
f
or
a
ut
om
a
ti
ng
di
a
gnos
ti
c
s
a
nd
im
pr
ovi
ng
di
s
e
a
s
e
de
te
c
ti
on.
A
ut
om
a
te
d
C
O
V
I
D
-
19
di
a
gnos
is
us
in
g
m
e
di
c
a
l
im
a
gi
ng
ha
s
be
e
n
e
xpl
or
e
d
us
in
g
da
ta
s
e
ts
c
om
pr
is
in
g
c
he
s
t
X
-
r
a
y
im
a
g
e
s
f
r
om
pa
ti
e
nt
s
w
it
h
b
a
c
te
r
ia
l
pne
um
oni
a
,
c
onf
ir
m
e
d
C
O
V
I
D
-
19
c
a
s
e
s
,
a
nd
uni
nf
e
c
te
d
c
ont
r
ol
s
.
D
e
e
p
ne
ur
a
l
ne
twor
k
a
r
c
hi
te
c
tu
r
e
s
ha
ve
be
e
n
in
ve
s
ti
ga
te
d t
o e
nha
nc
e
m
e
di
c
a
l
im
a
ge
c
l
a
s
s
if
ic
a
ti
on a
c
c
ur
a
c
y.
D
e
ns
e
N
e
t
is
a
de
e
p
le
a
r
ni
ng
a
r
c
hi
te
c
tu
r
e
th
a
t
w
a
s
de
ve
lo
pe
d
t
o
s
ol
ve
s
om
e
of
th
e
s
hor
tc
om
in
gs
of
C
N
N
s
.
T
h
e
s
e
pr
obl
e
m
s
in
c
lu
de
va
ni
s
hi
ng
gr
a
di
e
nt
s
,
in
f
or
m
a
ti
on
lo
s
s
,
a
nd
c
ha
ll
e
nge
s
in
tr
a
in
in
g
e
xt
r
e
m
e
ly
de
e
p
ne
twor
k.
T
he
a
bi
li
ty
of
th
e
D
e
n
s
e
N
e
t
de
s
ig
n
to
e
f
f
e
c
ti
ve
ly
le
a
r
n
f
r
om
da
ta
,
e
li
m
in
a
te
di
f
f
ic
ul
ti
e
s
w
it
h
va
ni
s
hi
ng
gr
a
di
e
nt
s
,
a
nd
a
c
hi
e
ve
gr
e
a
t
pe
r
f
or
m
a
nc
e
w
it
h
r
e
la
ti
ve
ly
f
e
w
e
r
pa
r
a
m
e
te
r
s
th
a
n
ot
he
r
a
r
c
hi
te
c
tu
r
e
s
ha
s
c
ont
r
ib
ut
e
d
to
th
e
r
is
e
in
popula
r
it
y
of
th
is
pa
r
ti
c
ul
a
r
a
r
c
hi
te
c
tu
r
e
.
T
he
num
be
r
"
121"
in
D
e
ns
e
N
e
t1
21
r
e
f
e
r
s
to
th
e
to
ta
l
num
be
r
of
la
ye
r
s
,
w
hi
c
h
in
c
lu
de
s
a
ll
f
ul
ly
c
onne
c
te
d,
c
onvolut
io
na
l,
pool
in
g,
a
nd
ba
tc
h
nor
m
a
li
z
a
ti
on
la
ye
r
s
.
D
e
ns
e
N
e
t1
21,
a
C
N
N
a
r
c
hi
te
c
tu
r
e
,
ha
s
s
e
ve
r
a
l
not
a
bl
e
be
ne
f
it
s
in
di
ve
r
s
e
c
om
put
e
r
vi
s
io
n
a
nd
im
a
ge
a
na
ly
s
is
a
ppl
ic
a
ti
ons
.
O
ne
not
a
bl
e
be
ne
f
it
is
in
it
s
e
xt
e
ns
iv
e
c
onne
c
ti
on
ne
twor
k,
w
hi
c
h
f
a
c
il
it
a
te
s
th
e
r
e
us
e
of
f
e
a
tu
r
e
s
a
nd
th
e
e
f
f
ic
ie
nt
tr
a
ns
m
i
s
s
io
n
of
in
f
or
m
a
ti
on
a
c
r
os
s
di
f
f
e
r
e
nt
la
ye
r
s
.
I
n C
N
N
s
,
th
e
pr
oc
e
s
s
of
c
om
bi
ni
ng f
e
a
tu
r
e
m
a
ps
i
s
t
ypi
c
a
ll
y pe
r
f
or
m
e
d i
n a
s
e
que
nt
ia
l
m
a
nne
r
. H
ow
e
v
e
r
, D
e
ns
e
N
e
t1
21
in
tr
oduc
e
s
a
nove
l
a
ppr
oa
c
h
w
he
r
e
e
a
c
h
la
ye
r
e
s
t
a
bl
is
he
s
di
r
e
c
t
c
onne
c
ti
ons
w
it
h
a
ll
f
ol
lo
w
in
g
la
ye
r
s
.
T
he
hi
gh
le
ve
l
of
c
onne
c
te
dne
s
s
in
s
id
e
th
e
ne
twor
k
f
a
c
il
it
a
te
s
th
e
tr
a
ns
m
is
s
io
n
of
gr
a
di
e
nt
s
,
m
it
ig
a
te
s
th
e
li
ke
li
hood
of
di
s
a
ppe
a
r
in
g
gr
a
di
e
nt
s
,
a
nd
a
ugm
e
nt
s
th
e
ne
two
r
k'
s
to
ta
l
c
a
pa
c
it
y
f
or
le
a
r
ni
ng.
C
on
s
e
que
nt
ly
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
de
e
p l
e
ar
ni
ng
-
bas
e
d f
r
am
e
w
or
k
f
or
aut
om
at
ic
de
te
c
ti
on C
O
V
I
D
-
19 us
in
g c
he
s
t
…
(
Si
v
anagir
e
dd
y
K
al
li
)
3195
D
e
ns
e
N
e
t1
21
h
a
s
a
te
nde
nc
y
to
n
e
e
d
a
r
e
du
c
e
d
num
be
r
of
pa
r
a
m
e
te
r
s
in
c
om
pa
r
is
on
to
ot
he
r
de
e
p
de
s
ig
ns
,
m
a
ki
ng
it
c
om
put
a
ti
ona
ll
y
e
f
f
e
c
ti
ve
a
nd
m
or
e
a
m
e
na
bl
e
to
tr
a
in
in
g,
e
s
pe
c
i
a
ll
y
w
he
n
c
onf
r
ont
e
d
w
it
h
li
tt
le
da
ta
. T
he
c
om
pa
c
t
na
tu
r
e
of
t
hi
s
f
e
a
tu
r
e
m
a
ke
s
i
t
e
s
pe
c
ia
ll
y a
dva
nt
a
ge
ous
i
n s
it
ua
ti
ons
w
he
n t
he
r
e
a
r
e
l
im
it
e
d
pr
oc
e
s
s
in
g
r
e
s
our
c
e
s
a
va
il
a
bl
e
.
O
ne
f
ur
th
e
r
be
ne
f
it
of
D
e
ns
e
N
e
t1
21
li
e
s
in
it
s
r
e
m
a
r
ka
bl
e
e
f
f
ic
a
c
y
in
th
e
pr
oc
e
s
s
e
s
of
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
a
nd
r
e
pr
e
s
e
nt
a
ti
on
le
a
r
ni
ng.
T
he
ne
twor
k'
s
ti
ght
in
te
r
c
onne
c
ti
ons
f
a
c
il
it
a
te
th
e
e
xt
r
a
c
ti
on
of
s
ubt
le
a
nd
hi
ghl
y
de
ta
il
e
d
in
f
or
m
a
ti
on
f
r
om
pi
c
tu
r
e
s
,
r
e
nde
r
in
g
it
pa
r
ti
c
ul
a
r
ly
a
dva
nt
a
ge
ous
f
or
jo
bs
de
m
a
ndi
ng
m
e
ti
c
ul
ous
a
na
ly
s
is
,
s
uc
h
a
s
im
a
ge
c
a
te
gor
iz
a
ti
on,
obj
e
c
t
r
e
c
ogni
ti
on,
a
nd
m
e
di
c
a
l
im
a
ge
a
na
ly
s
is
.
P
r
opos
e
d de
e
p l
e
a
r
ni
ng D
e
ns
e
N
e
t
a
r
c
hi
te
c
tu
r
e
a
s
s
ho
w
n i
n F
ig
ur
e
1
.
I
n
a
ddi
ti
on,
th
e
in
c
lu
s
io
n
of
de
ns
e
s
ki
p
c
onn
e
c
ti
ons
e
n
a
bl
e
s
t
he
e
s
ta
bl
is
hm
e
nt
of
s
ki
p
c
onne
c
ti
ons
a
c
r
os
s
va
r
io
us
le
ve
ls
w
it
hi
n
th
e
ne
twor
k,
he
nc
e
a
s
s
is
ti
ng
in
th
e
r
e
te
nt
io
n
of
lo
w
-
le
ve
l
c
ha
r
a
c
te
r
is
ti
c
s
a
nd
c
ont
e
xt
ua
l
in
f
or
m
a
ti
on
ove
r
th
e
w
hol
e
of
th
e
ne
twor
k.
T
hi
s
a
ppr
oa
c
h
pr
ove
s
to
be
ve
r
y
be
ne
f
ic
ia
l
in
th
e
c
ont
e
xt
of
ta
s
ks
s
uc
h a
s
s
e
m
a
nt
ic
s
e
gm
e
nt
a
ti
on,
w
he
r
e
th
e
a
va
i
la
bi
li
ty
of
pi
xe
l
-
le
ve
l
in
f
o
r
m
a
ti
on
is
of
ut
m
os
t
im
por
ta
nc
e
.
I
n
s
e
ve
r
a
l
c
om
put
e
r
vi
s
io
n
a
ppl
ic
a
ti
on
s
,
th
e
pop
ul
a
r
it
y
a
nd
s
uc
c
e
s
s
of
D
e
n
s
e
N
e
t1
21
m
a
y
b
e
a
tt
r
ib
ut
e
d
to
it
s
e
f
f
ic
ie
nt
pa
r
a
m
e
te
r
ut
il
iz
a
ti
on
a
nd
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
c
a
pa
bi
li
ti
e
s
.
D
e
n
s
e
N
e
t1
21
i
s
a
n e
xa
m
pl
e
of
a
C
N
N
de
s
ig
n
th
a
t
ha
s
be
e
n
ga
in
in
g
tr
a
c
ti
on
in
th
e
w
or
ld
of
m
e
di
c
a
l
im
a
gi
ng,
na
m
e
ly
in
th
e
id
e
nt
if
ic
a
ti
on
of
C
O
V
I
D
-
19
ut
il
iz
in
g
c
he
s
t
X
-
r
a
y
pi
c
tu
r
e
s
.
T
hi
s
is
one
of
th
e
m
os
t
im
por
ta
nt
a
ppl
ic
a
ti
ons
of
th
is
ki
nd
of
ne
twor
k.
T
he
s
ig
ni
f
ic
a
n
c
e
of
D
e
ns
e
N
e
t1
21
in
th
e
d
e
te
c
ti
on
of
C
O
V
I
D
-
19
us
in
g
c
he
s
t
X
-
r
a
y
im
a
ge
s
li
e
s
in
th
e
f
a
c
t
th
a
t
it
is
a
bl
e
to
e
f
f
e
c
ti
ve
ly
le
a
r
n
a
nd
a
na
ly
z
e
f
e
a
tu
r
e
s
f
r
om
m
e
di
c
a
l
im
a
ge
s
,
pr
ovi
de
s
ta
te
-
of
-
th
e
-
a
r
t
pe
r
f
or
m
a
nc
e
,
a
nd
e
ns
ur
e
da
ta
e
f
f
ic
ie
nc
y
a
nd
in
te
r
pr
e
ta
bi
li
ty
,
a
ll
of
w
hi
c
h
a
r
e
e
s
s
e
nt
ia
l
f
or
a
c
c
ur
a
te
a
nd
tr
us
twor
th
y di
s
e
a
s
e
di
a
gno
s
is
.
F
ig
ur
e
1. P
r
opos
e
d de
e
p l
e
a
r
ni
ng D
e
ns
e
N
e
t
a
r
c
hi
te
c
tu
r
e
4.
E
X
P
E
R
I
M
E
N
T
A
L
R
E
S
U
L
T
S
I
n
th
is
s
e
c
ti
on,
th
e
r
e
s
ul
ts
of
th
e
s
ugge
s
te
d
m
e
th
od
a
r
e
a
na
l
yz
e
d
a
nd
e
xa
m
in
e
d.
K
a
ggl
e
w
a
s
th
e
s
our
c
e
of
th
e
da
ta
s
e
t
th
a
t
w
a
s
u
s
e
d.
T
h
e
da
ta
s
e
t
w
a
s
pr
oc
e
s
s
e
d
us
in
g
th
e
a
ppr
oa
c
h
th
a
t
w
a
s
pr
opos
e
d.
T
he
da
ta
s
e
t
i
s
s
tr
uc
tu
r
e
d
w
it
h
th
r
e
e
m
a
in
f
ol
de
r
s
c
a
ll
e
d
tr
a
in
,
te
s
t,
a
nd
va
l,
a
nd
w
it
hi
n
th
o
s
e
th
r
e
e
m
a
in
f
ol
de
r
s
a
r
e
s
ubf
ol
de
r
s
la
be
le
d
"
P
ne
um
oni
a
"
a
nd
"
N
or
m
a
l.
"
T
he
r
e
a
r
e
5,86
3
X
-
R
a
y
pi
c
tu
r
e
s
in
th
e
J
P
E
G
f
or
m
a
t,
a
nd
th
e
y
a
r
e
s
e
pa
r
a
te
d
in
to
two
c
a
te
gor
ie
s
:
pne
um
oni
a
a
nd
nor
m
a
l.
C
he
s
t
X
-
r
a
y
pi
c
tu
r
e
s
w
e
r
e
c
ho
s
e
n
f
r
om
r
e
tr
os
pe
c
ti
ve
c
ohor
ts
of
pe
di
a
tr
ic
pa
ti
e
nt
s
a
ge
d one
t
o f
iv
e
ye
a
r
s
ol
d a
t
th
e
G
ua
ngz
hou W
om
e
n a
nd C
hi
ld
r
e
n'
s
M
e
di
c
a
l
C
e
nt
e
r
in
G
ua
ngz
hou.
T
he
phot
os
w
e
r
e
ta
ke
n
in
a
n
a
nt
e
r
io
r
-
pos
te
r
io
r
or
ie
nt
a
ti
on.
I
n
th
e
c
our
s
e
of
pr
ovi
di
ng
nor
m
a
l
c
li
ni
c
a
l
tr
e
a
tm
e
nt
f
or
th
e
pa
ti
e
nt
s
,
a
ll
c
h
e
s
t
X
-
r
a
y
im
a
gi
ng
w
a
s
c
onduc
te
d.
B
e
f
or
e
doi
ng
th
e
a
na
ly
s
is
of
c
he
s
t
x
-
r
a
y
pi
c
tu
r
e
s
,
e
v
e
r
y
c
he
s
t
r
a
di
ogr
a
ph
w
a
s
f
ir
s
t
s
ubj
e
c
t
e
d
to
a
s
c
r
e
e
ni
ng
f
or
qua
li
ty
c
ont
r
ol
.
T
hi
s
in
c
lu
de
d
de
le
ti
ng
a
ny
s
c
a
ns
th
a
t
w
e
r
e
of
poor
qua
li
ty
or
c
oul
d
not
be
r
e
a
d.
A
f
te
r
th
a
t,
th
e
di
a
gno
s
e
s
f
or
th
e
phot
os
w
e
r
e
s
c
or
e
d
by
two
hi
ghl
y
qua
li
f
ie
d
doc
to
r
s
be
f
o
r
e
be
in
g
gi
ve
n
th
e
gr
e
e
n
li
ght
f
or
us
e
in
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
4
,
A
ugus
t
2025
:
3192
-
3200
3196
tr
a
in
in
g
of
th
e
A
I
s
ys
t
e
m
.
I
n
or
de
r
to
ta
ke
in
to
c
ons
id
e
r
a
ti
on,
th
e
pos
s
ib
il
it
y
of
gr
a
di
ng
m
is
ta
ke
s
,
th
e
a
s
s
e
s
s
m
e
nt
s
e
t
w
a
s
a
ls
o r
e
vi
e
w
e
d by a
t
hi
r
d s
pe
c
ia
li
s
t.
T
he
R
O
C
c
ur
ve
is
c
r
e
a
te
d
by
c
om
p
a
r
in
g
th
e
tr
ue
po
s
it
iv
e
r
a
te
(
T
P
R
)
to
th
e
f
a
ls
e
pos
it
iv
e
r
a
te
(
F
P
R
)
,
w
hi
c
h
is
pl
ot
te
d
a
ga
in
s
t
e
a
c
h
ot
he
r
on
th
e
x
-
a
nd
y
-
a
xe
s
,
r
e
s
pe
c
ti
ve
ly
.
E
ve
r
y
poi
nt
on
th
e
R
O
C
c
ur
ve
r
e
pr
e
s
e
nt
s
a
di
f
f
e
r
e
nt
c
la
s
s
if
ic
a
ti
on
th
r
e
s
hol
d
th
a
t
is
a
ppl
ie
d
t
o
th
e
pr
oba
bi
li
ti
e
s
th
a
t
a
r
e
ge
ne
r
a
te
d
by
th
e
m
ode
l.
W
e
a
r
e
a
bl
e
to
r
e
gul
a
te
th
e
ba
la
nc
e
be
twe
e
n
s
e
ns
it
iv
it
y
a
nd
s
pe
c
if
ic
it
y
by
m
odi
f
yi
ng
th
e
th
r
e
s
hol
d,
a
nd t
hi
s
m
a
y be
done
i
n a
c
c
or
da
nc
e
w
it
h t
he
ne
e
ds
of
t
he
i
s
s
ue
.
T
he
F
ig
ur
e
2
is
a
gr
a
phi
c
a
l
il
lu
s
tr
a
ti
on
of
how
w
e
ll
a
bi
na
r
y
c
la
s
s
if
ie
r
di
f
f
e
r
e
nt
ia
te
s
be
twe
e
n
th
e
two
c
a
te
gor
ie
s
in
que
s
ti
on.
A
c
l
a
s
s
if
ie
r
th
a
t
a
c
hi
e
v
e
s
hi
gh
s
e
ns
it
iv
it
y
(
T
P
R
)
w
hi
le
s
im
ul
ta
ne
ous
ly
r
e
ta
in
in
g
a
lo
w
FPR
is
c
ons
id
e
r
e
d
to
ha
ve
gr
e
a
te
r
pe
r
f
o
r
m
a
nc
e
.
T
hi
s
is
s
how
n
by
a
c
ur
ve
th
a
t
is
c
lo
s
e
r
to
th
e
to
p
-
le
f
t
c
or
ne
r
.
I
t
is
c
om
m
on
pr
a
c
ti
c
e
to
ut
il
iz
e
th
e
a
r
e
a
unde
r
th
e
R
O
C
c
ur
ve
(
A
U
C
-
R
O
C
)
a
s
a
s
um
m
a
r
y
in
di
c
a
to
r
f
or
th
e
ove
r
a
ll
pe
r
f
or
m
a
nc
e
of
th
e
m
od
e
l.
A
gr
e
a
te
r
c
a
p
a
c
it
y
f
or
c
a
t
e
g
or
iz
a
ti
on
is
in
di
c
a
te
d
by
a
n
A
U
C
-
R
O
C
v
a
lu
e
th
a
t
is
c
lo
s
e
r
to
1.0.
O
n
th
e
R
O
C
gr
a
ph,
th
e
di
a
gona
l
li
ne
a
t
45
de
gr
e
e
s
de
pi
c
ts
r
a
ndom
gue
s
s
in
g.
T
hi
s
i
s
th
e
c
a
s
e
w
he
n
th
e
T
P
R
a
nd
th
e
FPR
a
r
e
id
e
nt
ic
a
l
to
one
a
not
he
r
.
A
c
la
s
s
if
ie
r
th
a
t
f
a
ll
s
be
lo
w
th
is
di
a
gona
l
is
in
f
e
r
io
r
to
r
a
ndom
gue
s
s
in
g
in
te
r
m
s
of
it
s
p
r
e
di
c
ti
ve
pow
e
r
.
I
n
c
ont
r
a
s
t,
a
c
la
s
s
if
ie
r
th
a
t
is
lo
c
a
te
d
a
bove
th
e
di
a
gona
l
ha
s
s
om
e
d
e
gr
e
e
of
pr
e
di
c
ti
ve
pow
e
r
. T
he
di
s
c
r
im
in
a
ti
ng powe
r
of
t
he
m
ode
l
in
c
r
e
a
s
e
s
i
n pr
opor
ti
on
to
th
e
s
te
e
pne
s
s
of
th
e
c
ur
ve
'
s
a
s
c
e
nt
to
w
a
r
d
th
e
uppe
r
le
f
t
c
or
ne
r
.
T
he
c
onf
us
io
n
m
a
tr
ix
is
a
c
om
m
onl
y
us
e
d
ta
bul
a
r
r
e
pr
e
s
e
nt
a
ti
on t
ha
t
is
ut
il
iz
e
d t
o e
va
lu
a
te
t
h
e
e
f
f
ic
a
c
y of
a
c
la
s
s
if
ic
a
ti
on mode
l.
T
he
a
s
s
e
s
s
m
e
nt
of
th
e
de
gr
e
e
to
w
hi
c
h
a
m
ode
l'
s
pr
e
di
c
ti
ons
c
o
r
r
e
s
pond
w
it
h
th
e
obs
e
r
ve
d
r
e
s
ul
ts
i
s
a
n
im
por
ta
nt
a
nd
us
e
f
ul
te
c
hni
que
,
pa
r
ti
c
ul
a
r
ly
in
s
c
e
na
r
io
s
w
he
r
e
th
e
r
e
is
a
n
une
qua
l
di
s
tr
ib
ut
io
n
of
c
la
s
s
e
s
.
I
n
F
ig
ur
e
3,
s
how
s
th
e
m
a
tr
ix
pr
e
s
e
nt
e
d
s
e
e
m
s
to
pe
r
ta
in
to
a
bi
na
r
y
c
la
s
s
if
ic
a
ti
on
ta
s
k,
w
he
r
e
by
th
e
c
la
s
s
e
s
a
r
e
de
not
e
d
a
s
"
N
or
m
a
l"
a
nd
"
C
O
V
I
D
."
T
he
r
ow
i
n
que
s
ti
on pe
r
ta
in
s
t
o s
it
ua
ti
ons
i
n w
hi
c
h t
he
a
c
tu
a
l
c
la
s
s
i
s
c
la
s
s
if
ie
d
a
s
"
N
or
m
a
l.
"
I
n
th
e
pr
e
s
e
nt
s
c
e
na
r
io
,
th
e
m
od
e
l
ha
s
a
c
c
ur
a
te
ly
c
la
s
s
if
ie
d
265
c
a
s
e
s
a
s
"
N
or
m
a
l,
"
but
it
ha
s
m
a
de
e
r
r
one
ous
pr
e
di
c
ti
ons
by
c
la
s
s
if
yi
ng
4
in
s
ta
nc
e
s
a
s
"
C
O
V
I
D
"
w
he
n
th
e
y
w
e
r
e
r
e
a
ll
y
"
N
or
m
a
l.
"
T
he
te
r
m
"
A
c
tu
a
l
C
O
V
I
D
"
r
e
f
e
r
s
to
c
a
s
e
s
in
w
hi
c
h
th
e
tr
ue
c
la
s
s
is
la
be
le
d
a
s
"
C
O
V
I
D
."
T
he
m
ode
l
ha
s
a
c
c
ur
a
te
ly
c
la
s
s
if
ie
d 67 c
a
s
e
s
a
s
"
C
O
V
I
D
,"
but
i
t
ha
s
m
a
de
e
r
r
one
ous
pr
e
di
c
ti
ons
i
n 3 i
ns
ta
nc
e
s
by
c
la
s
s
if
yi
ng
th
e
m
a
s
"
N
or
m
a
l"
w
he
n
th
e
y
w
e
r
e
in
f
a
c
t
"
C
O
V
I
D
."
T
he
c
ol
um
n
la
be
le
d
"
P
r
e
di
c
te
d
n
or
m
a
l
"
de
not
e
s
c
a
s
e
s
th
a
t
ha
ve
be
e
n
c
la
s
s
if
ie
d
a
s
"
N
or
m
a
l"
by
th
e
m
ode
l.
A
m
ong
th
e
oc
c
ur
r
e
nc
e
s
c
la
s
s
e
d
a
s
"
N
or
m
a
l,
"
a
to
ta
l
of
265
w
e
r
e
a
c
c
ur
a
te
ly
la
be
le
d
a
s
s
uc
h,
w
hi
le
3
in
s
ta
nc
e
s
w
e
r
e
e
r
r
one
ous
ly
c
la
s
s
if
ie
d
a
s
"
N
or
m
a
l"
w
he
n
th
e
y
w
e
r
e
r
e
a
ll
y
c
a
s
e
s
of
"
C
O
V
I
D
."
T
he
c
ol
um
n
la
be
le
d
"
P
r
e
di
c
te
d
C
O
V
I
D
"
de
not
e
s
th
e
oc
c
ur
r
e
nc
e
s
th
a
t
th
e
m
ode
l
ha
s
id
e
nt
if
ie
d
a
nd
c
la
s
s
if
ie
d
a
s
"
C
O
V
I
D
."
A
m
ong
th
e
oc
c
ur
r
e
nc
e
s
th
a
t
w
e
r
e
c
a
te
gor
iz
e
d
a
s
"
C
O
V
I
D
,"
a
to
ta
l
of
67
in
s
ta
nc
e
s
w
e
r
e
a
c
c
ur
a
t
e
ly
la
be
le
d
a
s
"
C
O
V
I
D
,"
w
he
r
e
a
s
4
in
s
ta
nc
e
s
w
e
r
e
e
r
r
one
ous
ly
c
la
s
s
if
ie
d
a
s
"
C
O
V
I
D
"
w
he
n
th
e
y
w
e
r
e
r
e
a
ll
y
e
xa
m
pl
e
s
of
"
N
or
m
a
l.
"
T
he
num
be
r
o
f
in
s
ta
nc
e
s
a
c
c
ur
a
te
ly
c
la
s
s
if
ie
d
a
s
"
C
O
V
I
D
"
is
67,
w
hi
c
h
a
r
e
r
e
f
e
r
r
e
d
to
a
s
tr
ue
pos
it
iv
e
s
(
T
P
)
.
T
r
ue
ne
ga
ti
ve
s
(
T
N
)
r
e
f
e
r
to
in
s
ta
nc
e
s
th
a
t
ha
ve
be
e
n
a
c
c
ur
a
te
ly
f
or
e
c
a
s
te
d
a
s
"
N
or
m
a
l.
"
I
n
th
is
s
pe
c
if
ic
c
a
s
e
,
th
e
r
e
a
r
e
265
in
s
ta
nc
e
s
th
a
t
h
a
ve
be
e
n
pr
ope
r
ly
c
la
s
s
if
ie
d
a
s
"
N
or
m
a
l.
"
F
a
ls
e
pos
it
iv
e
s
(
FP
)
r
e
f
e
r
to
in
s
ta
nc
e
s
th
a
t
a
r
e
in
c
or
r
e
c
tl
y
f
or
e
c
a
s
te
d
a
s
"
C
O
V
I
D
"
w
he
n
th
e
y
a
r
e
r
e
a
ll
y
c
la
s
s
i
f
ie
d
a
s
"
N
or
m
a
l.
"
I
n
th
is
pa
r
ti
c
ul
a
r
c
a
s
e
,
th
e
r
e
a
r
e
f
our
in
s
ta
nc
e
s
th
a
t
f
a
ll
unde
r
th
is
c
a
t
e
gor
y.
F
a
ls
e
ne
ga
ti
ve
s
(
FN
)
r
e
f
e
r
to
in
s
ta
nc
e
s
in
w
hi
c
h
th
e
pr
e
di
c
ti
on
is
c
la
s
s
if
ie
d
a
s
"
N
or
m
a
l,
"
w
he
n
in
r
e
a
li
ty
,
it
s
houl
d
ha
ve
be
e
n
c
la
s
s
if
ie
d
a
s
"
C
O
V
I
D
."
I
n
th
is
pa
r
ti
c
ul
a
r
c
a
s
e
,
th
e
r
e
w
e
r
e
t
hr
e
e
i
ns
ta
nc
e
s
of
FN
.
F
ig
ur
e
2. R
O
C
c
ur
ve
F
ig
ur
e
3. C
onf
us
io
n m
a
tr
ix
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
de
e
p l
e
ar
ni
ng
-
bas
e
d f
r
am
e
w
or
k
f
or
aut
om
at
ic
de
te
c
ti
on C
O
V
I
D
-
19 us
in
g c
he
s
t
…
(
Si
v
anagir
e
dd
y
K
al
li
)
3197
T
he
pr
e
c
is
io
n
-
r
e
c
a
ll
c
ur
ve
r
e
pr
e
s
e
nt
e
d i
n
F
ig
ur
e
4
th
a
t
vi
s
ua
li
z
e
s
t
he
t
r
a
de
-
of
f
be
twe
e
n a
c
c
ur
a
c
y a
nd
r
e
c
a
ll
f
or
va
r
io
us
c
a
te
gor
iz
a
ti
on
th
r
e
s
hol
ds
.
I
t
is
a
ls
o
known
a
s
th
e
pr
e
c
i
s
io
n
-
r
e
c
a
ll
c
ur
ve
.
P
r
e
di
c
ti
ons
a
r
e
pr
oduc
e
d
us
in
g
a
c
onf
id
e
nc
e
s
c
or
e
or
a
pr
oba
bi
li
ty
va
lu
e
in
m
a
ny
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
.
A
th
r
e
s
hol
d
is
th
e
n
us
e
d
to
de
c
id
e
w
he
th
e
r
a
pr
e
di
c
ti
on
is
de
e
m
e
d
pos
it
iv
e
o
r
ne
ga
ti
ve
.
T
he
c
ur
ve
is
pr
oduc
e
d
by
m
a
ki
n
g
s
m
a
ll
,
in
c
r
e
m
e
nt
a
l
c
ha
nge
s
to
a
pr
e
de
te
r
m
in
e
d
th
r
e
s
hol
d
th
a
t
is
us
e
d
f
or
id
e
nt
if
yi
ng
c
a
s
e
s
.
W
he
n
th
e
th
r
e
s
hol
d
is
a
dj
u
s
te
d,
th
e
num
be
r
of
TP
,
FP
,
TN
,
a
nd
FN
a
ls
o
c
ha
nge
s
.
T
hi
s
,
in
tu
r
n,
h
a
s
a
n
im
pa
c
t
on
th
e
a
c
c
ur
a
c
y
a
nd
r
e
c
a
ll
of
th
e
te
s
t.
T
he
c
la
s
s
if
ie
r
ha
s
a
te
nde
nc
y
t
o
pr
oduc
e
f
e
w
e
r
pos
it
iv
e
pr
e
di
c
ti
ons
w
he
n
th
e
th
r
e
s
hol
d
is
s
e
t
e
xt
r
e
m
e
ly
hi
gh,
w
hi
c
h
le
a
ds
to
hi
gh
a
c
c
ur
a
c
y
b
ut
pe
r
ha
ps
r
e
duc
e
d
r
e
c
a
ll
.
W
he
n
th
e
th
r
e
s
hol
d
is
lo
w
,
on
th
e
ot
he
r
ha
nd,
m
or
e
c
a
s
e
s
a
r
e
pr
oj
e
c
te
d
a
s
pos
it
iv
e
,
w
hi
c
h
in
c
r
e
a
s
e
s
r
e
c
a
ll
but
m
a
y
r
e
s
ul
t
in
de
c
r
e
a
s
e
d
a
c
c
ur
a
c
y.
F
ig
ur
e
5
is
F
1 s
c
or
e
e
vol
ut
io
n pl
ot
w
hi
c
h i
s
a
s
ta
ti
s
ti
c
t
ha
t
c
om
bi
ne
s
a
c
c
ur
a
c
y
a
nd r
e
c
a
ll
i
nt
o a
s
in
gl
e
num
be
r
.
T
hi
s
va
lu
e
m
a
y
be
us
e
d
to
e
v
a
lu
a
te
pe
r
f
or
m
a
nc
e
.
I
t
is
e
s
pe
c
ia
ll
y
he
lp
f
ul
in
s
it
ua
ti
ons
in
w
hi
c
h
you
w
is
h t
o s
tr
ik
e
a
c
om
pr
om
is
e
be
twe
e
n t
he
c
om
p
e
ti
ng va
lu
e
s
of
a
c
c
ur
a
c
y a
nd r
e
c
a
ll
. T
h
e
f
or
m
ul
a
t
ha
t
is
us
e
d t
o
de
te
r
m
in
e
t
he
F
1 s
c
or
e
i
s
a
s
f
ol
lo
w
s
:
1
=
2
×
(
×
)
/
(
+
)
T
he
m
ul
ti
pl
e
it
e
r
a
ti
ons
or
di
f
f
e
r
e
nt
im
pl
e
m
e
nt
a
ti
ons
of
th
e
c
a
te
gor
iz
a
ti
on
m
ode
l
a
r
e
s
how
n
a
lo
ng
th
e
x
-
a
xi
s
of
th
e
F
ig
ur
e
5
.
E
a
c
h
it
e
r
a
ti
on
i
s
a
s
e
p
a
r
a
te
e
f
f
or
t
to
e
nha
n
c
e
t
he
m
ode
l'
s
f
unc
ti
ona
li
ty
in
a
va
r
ie
ty
of
w
a
ys
,
a
nd
th
e
s
e
im
pr
ove
m
e
nt
s
a
r
e
r
e
pr
e
s
e
nt
e
d
a
s
di
s
c
r
e
te
a
tt
e
m
pt
s
.
T
he
F
1
s
c
or
e
th
a
t
th
e
m
ode
l
m
a
na
g
e
d
to
a
tt
a
in
dur
in
g
e
a
c
h
it
e
r
a
ti
on
is
s
how
n
a
lo
ng
th
e
y
-
a
xi
s
.
T
he
F
1
s
c
or
e
is
in
te
nde
d
to
be
a
m
e
a
s
ur
e
of
th
e
de
gr
e
e
to
w
hi
c
h
th
e
m
ode
l
is
a
bl
e
to
s
tr
ik
e
a
ba
la
nc
e
be
twe
e
n
r
e
c
a
ll
a
nd
a
c
c
ur
a
c
y.
W
he
n
c
om
pa
r
e
d
to
pr
e
vi
ous
s
c
or
e
s
,
a
hi
ghe
r
F
1
s
c
or
e
im
pl
ie
s
im
pr
ove
d
ove
r
a
ll
pe
r
f
or
m
a
nc
e
.
T
he
pe
r
f
or
m
a
nc
e
m
e
tr
ic
s
of
th
e
pr
opos
e
d
m
ode
l
is
r
e
por
te
d i
n T
a
bl
e
1.
F
ig
ur
e
4. P
r
e
c
is
io
n vs
r
e
c
a
ll
c
ur
ve
F
ig
ur
e
5. F
1 s
c
or
e
e
vol
ut
io
n pl
ot
T
a
bl
e
1. P
e
r
f
or
m
a
nc
e
m
e
tr
ic
s
of
t
he
pr
opos
e
d m
ode
l
L
a
be
l
P
r
e
c
e
s
s
i
on
R
e
c
a
l
l
F1
s
c
or
e
N
or
m
a
l
0.98
0.97
0.98
C
O
V
I
D
0.88
0.94
0.91
A
c
c
ur
a
c
y
0.96
T
he
pe
r
f
or
m
a
nc
e
m
e
a
s
ur
e
s
th
a
t
ha
ve
b
e
e
n
pr
e
s
e
nt
e
d
i
n
T
a
bl
e
1,
th
e
e
f
f
ic
i
e
nc
y
of
th
e
m
od
e
l
w
i
th
r
e
ga
r
d
to
t
he
s
e
im
por
t
a
nt
m
e
tr
ic
s
is
s
how
n
f
or
t
w
o
s
e
pa
r
a
te
c
la
s
s
e
s
,
n
a
m
e
ly
"
N
or
m
a
l"
a
nd
"
C
O
V
I
D
."
A
n
a
c
c
ur
a
c
y
v
a
lu
e
of
0.
96
i
s
s
how
n
by
th
e
m
ode
l
w
he
n
a
p
pl
ie
d
to
t
he
"
N
or
m
a
l"
c
l
a
s
s
.
T
hi
s
s
how
s
th
a
t
a
r
e
la
ti
ve
ly
hi
gh
p
e
r
c
e
nt
a
g
e
of
c
or
r
e
c
t
po
s
it
iv
e
pr
e
di
c
ti
ons
w
e
r
e
m
a
de
i
n
c
om
p
a
r
is
o
n
to
t
he
t
ot
a
l
num
be
r
of
po
s
it
iv
e
pr
e
di
c
ti
ons
m
a
d
e
.
T
he
f
a
c
t
th
a
t
th
e
m
ode
l
h
a
s
a
r
e
c
a
ll
v
a
lu
e
of
0.97
in
di
c
a
t
e
s
t
ha
t
it
i
s
c
a
pa
bl
e
of
id
e
nt
if
yi
n
g
a
s
ig
ni
f
i
c
a
nt
pr
opor
ti
on
of
oc
c
ur
r
e
n
c
e
s
t
ha
t
a
r
e
r
e
a
ll
y
p
os
it
iv
e
.
A
s
c
or
e
of
0.98
f
or
th
e
e
qui
v
a
le
n
t
F
1
in
di
c
a
te
s
a
ba
la
n
c
e
d
pe
r
f
or
m
a
nc
e
i
n
te
r
m
s
of
bot
h
a
c
c
ur
a
c
y
a
n
d
r
e
c
a
ll
.
I
n
a
s
im
il
a
r
w
a
y
a
pr
e
c
is
io
n
v
a
lu
e
of
0
.88
f
or
th
e
"
C
O
V
I
D
"
c
la
s
s
in
di
c
a
t
e
s
a
s
a
ti
s
f
a
c
to
r
y
d
e
gr
e
e
of
a
c
c
ur
a
c
y
in
t
he
id
e
nt
if
ic
a
ti
on
of
a
f
f
ir
m
a
ti
ve
c
a
s
e
s
.
T
h
e
f
a
c
t
th
a
t
th
e
m
od
e
l
w
a
s
s
uc
c
e
s
s
f
ul
in
c
a
pt
ur
i
ng
a
s
i
gni
f
ic
a
nt
pe
r
c
e
nt
a
g
e
of
e
xa
m
pl
e
s
w
he
n
t
he
hyp
ot
he
s
i
s
w
a
s
c
or
r
e
c
t
is
s
ho
w
n
by
t
he
r
e
c
a
ll
va
lu
e
of
0.
94.
T
he
f
a
c
t
th
a
t
t
hi
s
s
pe
c
if
i
c
c
l
a
s
s
h
a
s
a
n
F
1
s
c
or
e
of
0.
91
in
di
c
a
t
e
s
th
a
t
th
e
r
e
i
s
a
he
a
lt
h
y ba
l
a
n
c
e
e
xi
s
ti
ng
be
tw
e
e
n
a
c
c
ur
a
c
y
a
nd r
e
c
a
ll
.
C
om
pa
r
a
ti
v
e
r
e
s
ul
t
s
a
s
s
ho
w
n i
n
T
a
bl
e
2.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
4
,
A
ugus
t
2025
:
3192
-
3200
3198
T
a
bl
e
2. C
om
pa
r
a
ti
ve
r
e
s
ul
ts
M
ode
l
A
c
c
ur
a
c
y (
%
)
A
l
e
xN
e
t
87
V
G
G
16
89
R
e
s
N
e
t
50
94
P
r
opos
e
d D
e
ns
e
N
e
t
m
ode
l
96
5.
C
O
N
C
L
U
S
I
O
N
I
n
c
on
c
lu
s
io
n,
a
c
l
a
s
s
if
ic
a
ti
on
m
od
e
l
w
a
s
d
e
ve
l
ope
d
in
our
s
tu
dy
to
id
e
n
ti
f
y
C
O
V
I
D
-
19
c
a
s
e
s
b
a
s
e
d
on
m
e
di
c
a
l
im
a
ge
a
n
a
ly
s
i
s
.
A
d
a
ta
s
e
t
c
om
pr
is
i
ng
b
ot
h
"
N
or
m
a
l"
a
nd
"
C
O
V
I
D
"
c
a
s
e
s
w
a
s
ut
il
i
z
e
d,
a
nd
m
a
c
hi
n
e
l
e
a
r
ni
n
g
te
c
hni
qu
e
s
w
e
r
e
e
m
p
lo
ye
d
to
tr
a
in
a
n
d
e
v
a
lu
a
te
th
e
m
od
e
l'
s
pe
r
f
or
m
a
nc
e
.
P
r
om
i
s
in
g
r
e
s
u
lt
s
in
te
r
m
s
of
a
c
c
ur
a
c
y,
pr
e
c
i
s
io
n
,
r
e
c
a
ll
,
a
nd
F
1
s
c
or
e
w
e
r
e
e
xhi
b
it
e
d
by
th
e
m
o
de
l.
S
p
e
c
if
i
c
a
ll
y
,
a
n
a
c
c
ur
a
c
y
of
96%
w
a
s
a
c
hi
e
ve
d
f
or
th
e
"
N
or
m
a
l"
c
l
a
s
s
,
in
di
c
a
ti
ng
a
hi
g
h
pe
r
c
e
nt
a
ge
of
c
or
r
e
c
tl
y
pr
e
di
c
t
e
d
ne
ga
ti
v
e
c
a
s
e
s
.
T
he
r
e
c
a
ll
of
97%
s
u
gge
s
te
d
t
ha
t
a
c
tu
a
l
n
e
ga
ti
ve
c
a
s
e
s
w
e
r
e
pr
of
i
c
ie
n
tl
y
id
e
nt
if
ie
d,
a
nd
a
w
e
ll
-
ba
l
a
nc
e
d
pe
r
f
or
m
a
n
c
e
b
e
twe
e
n
a
c
c
ur
a
c
y
a
nd
r
e
c
a
ll
w
a
s
in
di
c
a
t
e
d
by
th
e
F
1
s
c
or
e
of
98%
.
F
or
th
e
"
C
O
V
I
D
"
c
la
s
s
,
a
pr
e
c
i
s
io
n
of
88%
w
a
s
a
t
ta
in
e
d
by
th
e
m
ode
l,
d
e
m
on
s
tr
a
t
in
g
a
s
a
ti
s
f
a
c
to
r
y
le
ve
l
of
a
c
c
ur
a
c
y
in
id
e
nt
if
yi
ng
pos
it
iv
e
c
a
s
e
s
.
T
h
e
r
e
c
a
ll
of
94%
s
ig
ni
f
ie
d
th
e
m
od
e
l'
s
c
a
p
a
bi
li
ty
to
c
a
pt
ur
e
a
s
ig
ni
f
ic
a
nt
pr
op
or
ti
on
of
a
c
tu
a
l
pos
it
iv
e
c
a
s
e
s
, a
n
d t
he
F
1
s
c
or
e
of
91%
c
o
nf
ir
m
e
d
a
he
a
lt
hy
ba
l
a
nc
e
be
t
w
e
e
n a
c
c
ur
a
c
y
a
nd r
e
c
a
ll
f
or
t
hi
s
c
l
a
s
s
.
C
om
pa
r
i
s
on
to
ot
he
r
e
xi
s
ti
ng
m
od
e
ls
,
in
c
lu
di
ng
A
l
e
xN
e
t,
V
G
G
16,
a
nd
R
e
s
N
e
t5
0,
w
a
s
c
ondu
c
te
d
to
e
v
a
lu
a
te
our
m
ode
l'
s
p
e
r
f
or
m
a
n
c
e
. O
ur
pr
opo
s
e
d
D
e
n
s
e
N
e
t
m
ode
l
out
p
e
r
f
or
m
e
d t
h
e
s
e
m
od
e
ls
w
it
h
a
n
a
c
c
ur
a
c
y of
9
6%
,
s
how
c
a
s
in
g
i
ts
e
f
f
e
c
ti
ve
ne
s
s
i
n t
h
e
d
e
te
c
ti
on
of
C
O
V
I
D
-
19
f
r
om
m
e
di
c
a
l
im
a
ge
s
.
A
C
K
N
O
WL
E
D
G
M
E
N
T
S
T
he
a
ut
hor
s
w
i
s
h
to
th
a
nk
s
a
he
a
d
of
ti
m
e
th
e
D
e
pa
r
tm
e
nt
of
E
le
c
tr
oni
c
s
a
nd
C
om
m
uni
c
a
ti
ons
E
ngi
ne
e
r
in
g
a
nd
D
e
pa
r
tm
e
nt
of
C
om
put
e
r
S
c
ie
nc
e
a
nd
E
ngi
n
e
e
r
in
g,
S
r
id
e
vi
W
om
e
n
E
ngi
ne
e
r
in
g
C
ol
le
ge
,
H
yde
r
a
ba
d,
to
a
s
s
i
s
t
th
e
m
in
th
e
ir
t
e
c
hni
c
a
l
a
dvi
c
e
a
nd
r
e
s
e
a
r
c
h
a
m
e
ni
ti
e
s
.
W
e
a
l
s
o
a
c
knowle
dge
,
D
r
.
A
.
N
a
r
m
a
da
, w
ho pr
ovi
de
d us
e
f
ul
c
om
m
e
nt
s
dur
in
g m
ode
l
va
li
da
t
io
n.
F
U
N
D
I
N
G
I
N
F
O
R
M
A
T
I
O
N
T
he
s
tu
dy
is
pa
r
ti
a
ll
y
s
pons
or
e
d
by
th
e
I
nt
e
r
na
l
Q
ua
li
ty
A
s
s
ur
a
nc
e
C
e
ll
(
I
Q
A
C
)
s
c
he
m
e
of
th
e
S
r
id
e
vi
W
om
e
n
E
ngi
ne
e
r
in
g
C
ol
le
ge
in
te
r
m
s
of
in
f
r
a
s
tr
uc
tu
r
e
a
nd
r
e
s
e
a
r
c
h
f
a
c
il
it
ie
s
.
T
he
pr
oduc
e
d
w
or
k
w
a
s
not
f
unde
d by a
ny a
ge
n
c
y.
A
U
T
H
O
R
C
O
N
T
R
I
B
U
T
I
O
N
S
S
T
A
T
E
M
E
N
T
T
hi
s
jo
ur
na
l
us
e
s
th
e
C
ont
r
ib
ut
or
R
ol
e
s
T
a
xonomy
(
C
R
e
di
T
)
to
r
e
c
ogni
z
e
in
di
vi
dua
l
a
ut
hor
c
ont
r
ib
ut
io
ns
, r
e
duc
e
a
ut
hor
s
hi
p di
s
put
e
s
,
a
nd f
a
c
il
it
a
te
c
ol
la
bo
r
a
ti
on.
N
am
e
o
f
A
u
t
h
or
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
S
iv
a
na
gi
r
e
ddy Ka
ll
i
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
B
u
kk
a
N
a
r
e
n
dr
a
K
um
a
r
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
S
a
ggur
th
i
J
a
ga
de
e
s
h
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
K
u
s
h
a
g
a
r
i
C
h
a
ndr
a
m
o
ul
i
R
a
v
i
K
um
a
r
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
C
:
C
onc
e
pt
ua
l
i
z
a
t
i
on
M
:
M
e
t
hodol
ogy
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
da
t
i
on
Fo
:
Fo
r
m
a
l
a
na
l
ys
i
s
I
:
I
nve
s
t
i
ga
t
i
on
R
:
R
e
s
our
c
e
s
D
:
D
a
t
a
C
ur
a
t
i
on
O
:
W
r
i
t
i
ng
-
O
r
i
gi
na
l
D
r
a
f
t
E
:
W
r
i
t
i
ng
-
R
e
vi
e
w
&
E
di
t
i
ng
Vi
:
Vi
s
ua
l
i
z
a
t
i
on
Su
:
Su
pe
r
vi
s
i
on
P
:
P
r
oj
e
c
t
a
dm
i
ni
s
t
r
a
t
i
on
Fu
:
Fu
ndi
ng a
c
qui
s
i
t
i
on
C
O
N
F
L
I
C
T
O
F
I
N
T
E
R
E
S
T
S
T
A
T
E
M
E
N
T
T
he
a
ut
hor
s
in
di
c
a
te
th
a
t
th
e
y
do
not
h
a
ve
a
ny
known
c
om
pe
ti
ng
f
in
a
nc
ia
l
in
te
r
e
s
ts
or
pe
r
s
on
a
l
r
e
la
ti
ons
hi
ps
th
a
t
m
ig
ht
ha
ve
s
e
e
m
e
d
to
a
f
f
e
c
t
th
e
w
or
k
r
e
por
te
d
in
th
is
pa
pe
r
.
T
he
r
e
i
s
no
c
onf
li
c
t
of
in
te
r
e
s
t
de
c
la
r
e
d by a
ut
hor
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
de
e
p l
e
ar
ni
ng
-
bas
e
d f
r
am
e
w
or
k
f
or
aut
om
at
ic
de
te
c
ti
on C
O
V
I
D
-
19 us
in
g c
he
s
t
…
(
Si
v
anagir
e
dd
y
K
al
li
)
3199
I
N
F
O
R
M
E
D
C
O
N
S
E
N
T
A
ll
th
e
pe
opl
e
in
c
lu
de
d
in
th
is
w
or
k
ha
ve
s
ig
ne
d
a
n
in
f
o
r
m
e
d
c
ons
e
nt
w
it
h
us
.
P
a
r
ti
c
ip
a
nt
s
w
e
r
e
m
a
de
t
o unde
r
s
ta
nd w
ha
t
th
e
r
e
s
e
a
r
c
h i
nvol
ve
d a
nd t
he
ir
c
on
s
e
nt
w
a
s
gi
ve
n a
s
p
e
r
t
he
e
th
ic
a
l
r
e
qui
r
e
m
e
nt
s
.
E
T
H
I
C
A
L
A
P
P
R
O
V
A
L
A
ll
th
e
r
e
s
e
a
r
c
h
w
or
ki
ng
w
it
h
hum
a
n
s
ubj
e
c
ts
f
ol
lo
w
e
d
a
ll
th
e
ne
c
e
s
s
a
r
y
na
ti
ona
l
r
e
gul
a
ti
ons
a
nd
in
s
ti
tu
ti
ona
l
pol
ic
ie
s
,
a
s
w
e
ll
a
s
f
ol
lo
w
e
d
th
e
pr
in
c
ip
le
s
of
th
e
H
e
ls
in
ki
D
e
c
la
r
a
ti
on.
T
he
I
ns
ti
tu
ti
ona
l
R
e
vi
e
w
B
oa
r
d
(
I
R
B
)
of
S
r
id
e
vi
W
om
e
n
E
ngi
ne
e
r
in
g
C
ol
le
ge
,
H
yde
r
a
ba
d,
T
e
la
ng
a
na
,
I
ndi
a
.
T
h
e
s
tu
dy
i
s
not
li
nke
d
to
huma
n s
ubj
e
c
ts
or
a
ni
m
a
l
r
e
s
e
a
r
c
h
a
nd t
hus
do
e
s
not
ne
e
d
e
t
hi
c
a
l
s
c
r
ut
in
y.
D
A
T
A
A
V
A
I
L
A
B
I
L
I
T
Y
T
he
da
ta
unde
r
ly
in
g
th
e
f
in
di
ngs
of
th
e
pr
e
s
e
nt
s
tu
dy
c
a
n
be
r
e
que
s
te
d
by
a
ddr
e
s
s
in
g
th
e
c
or
r
e
s
ponding
a
ut
hor
w
it
h
a
r
e
a
s
ona
bl
e
r
e
que
s
t.
T
he
da
ta
th
a
t
w
a
s
ut
il
iz
e
d
in
th
e
r
e
s
e
a
r
c
h
i
s
not
ope
nl
y
a
va
il
a
bl
e
or
c
onf
id
e
nt
ia
l.
T
he
da
ta
th
a
t
w
a
s
ut
il
iz
e
d
in
th
e
r
e
s
e
a
r
c
h
is
not
ope
nl
y
a
va
il
a
bl
e
a
s
pe
r
in
s
ti
tu
ti
ona
l
c
onf
id
e
nt
ia
li
ty
pol
ic
y but c
a
n be
s
ha
r
e
d a
s
pe
r
r
e
a
s
ona
bl
e
r
e
que
s
t
a
ddr
e
s
s
e
d t
o t
he
c
or
r
e
s
ponding a
ut
hor
.
R
E
F
E
R
E
N
C
E
S
[
1]
W
H
O
,
“
C
l
i
ni
c
a
l
m
a
na
ge
m
e
nt
of
s
e
ve
r
e
a
c
ut
e
r
e
s
pi
r
a
t
or
y
i
nf
e
c
t
i
on
w
he
n
nove
l
c
or
ona
vi
r
us
(
2019
-
nC
oV
)
i
nf
e
c
t
i
on
i
s
s
us
pe
c
t
e
d:
i
nt
e
r
i
m
gui
da
nc
e
,”
W
or
l
d H
e
al
t
h O
r
gani
z
at
i
on
, 2020.
[
2]
H
.
S
hi
e
t
al
.
,
“
R
a
di
ol
ogi
c
a
l
f
i
nd
i
ngs
f
r
om
81
pa
t
i
e
nt
s
w
i
t
h
C
O
V
I
D
-
19
pne
u
m
oni
a
i
n
W
uha
n,
C
hi
na
:
a
de
s
c
r
i
pt
i
ve
s
t
udy,”
T
he
L
anc
e
t
I
nf
e
c
t
i
ous
D
i
s
e
as
e
s
, vol
. 20, no. 4, pp. 425
–
434, 2020, doi
:
10.1016/
S
1473
-
3099(
20)
30086
-
4.
[
3]
A
.
W
.
S
a
l
e
hi
,
P
.
B
a
gl
a
t
,
a
nd
G
.
G
upt
a
,
“
R
e
vi
e
w
on
m
a
c
hi
ne
a
nd
de
e
p
l
e
a
r
ni
ng
m
ode
l
s
f
or
t
he
de
t
e
c
t
i
on
a
nd
pr
e
di
c
t
i
on
of
C
or
ona
vi
r
us
,”
M
at
e
r
i
al
s
T
oday
:
P
r
oc
e
e
di
ngs
, vol
. 33, pp. 3896
–
3901, 2020, do
i
:
10.1016/
j
.m
a
t
pr
.2020.06.245.
[
4]
Y
. L
e
C
un,
Y
.
B
e
ngi
o,
a
n
d G
. H
i
n
t
o
n,
“
D
e
e
p
l
e
a
r
n
i
n
g,”
N
at
ur
e
,
v
ol
. 5
21
, n
o.
75
53
, p
p.
43
6
–
4
44
, 2
01
5
,
do
i
:
10
.1
038
/
n
a
t
u
r
e
14
539
.
[
5]
K
.
H
e
,
X
.
Z
ha
ng,
S
.
R
e
n,
a
nd
J
.
S
un,
“
D
e
e
p
r
e
s
i
dua
l
l
e
a
r
ni
ng
f
or
i
m
a
ge
r
e
c
ogni
t
i
on,”
i
n
P
r
oc
e
e
di
ng
s
of
t
he
I
E
E
E
C
om
put
e
r
Soc
i
e
t
y
C
onf
e
r
e
n
c
e
on C
om
put
e
r
V
i
s
i
on and P
at
t
e
r
n R
e
c
ogni
t
i
on
, 2016, pp. 770
–
778
, doi
:
10.1109/
C
V
P
R
.2016.90.
[
6]
I
.
D
.
A
pos
t
ol
opoul
os
a
nd
T
.
A
.
M
pe
s
i
a
n
a
,
“
C
O
V
I
D
-
19:
a
ut
om
a
t
i
c
de
t
e
c
t
i
on
f
r
om
X
-
r
a
y
i
m
a
ge
s
ut
i
l
i
z
i
ng
t
r
a
ns
f
e
r
l
e
a
r
ni
ng
w
i
t
h
c
onvol
ut
i
ona
l
ne
ur
a
l
ne
t
w
or
ks
,”
P
hy
s
i
c
al
and
E
ngi
ne
e
r
i
ng
Sc
i
e
nc
e
s
i
n
M
e
di
c
i
ne
,
vol
.
43,
no.
2,
pp.
635
–
640,
2020,
doi
:
10.1007/
s
13246
-
020
-
00865
-
4.
[
7]
T
.
O
z
t
ur
k,
M
.
T
a
l
o,
E
.
A
.
Y
i
l
di
r
i
m
,
U
.
B
.
B
a
l
ogl
u,
O
.
Y
i
l
di
r
i
m
,
a
nd
U
.
R
a
j
e
n
dr
a
A
c
ha
r
ya
,
“
A
ut
om
a
t
e
d
d
e
t
e
c
t
i
on
of
C
O
V
I
D
-
19
c
a
s
e
s
us
i
ng
de
e
p
ne
ur
a
l
ne
t
w
or
ks
w
i
t
h
X
-
r
a
y
i
m
a
ge
s
,”
C
om
put
e
r
s
i
n
B
i
ol
ogy
and
M
e
di
c
i
ne
,
vol
.
121,
2020,
doi
:
10.1016/
j
.c
om
pbi
om
e
d.2020.103792.
[
8]
L
.
W
a
ng,
Z
.
Q
.
L
i
n,
a
nd
A
.
W
ong,
“
C
O
V
I
D
-
N
e
t
:
a
t
a
i
l
o
r
e
d
de
e
p
c
onvol
ut
i
ona
l
ne
ur
a
l
ne
t
w
or
k
de
s
i
gn
f
or
de
t
e
c
t
i
on
of
C
O
V
I
D
-
19
c
a
s
e
s
f
r
om
c
he
s
t
X
-
r
a
y i
m
a
ge
s
,”
Sc
i
e
nt
i
f
i
c
R
e
por
t
s
, vol
. 10, no. 1, N
ov. 2020, d
oi
:
10.1038/
s
41598
-
020
-
76550
-
z.
[
9]
A
.
J
a
i
s
w
a
l
,
N
.
G
i
a
nc
ha
nda
ni
,
D
.
S
i
ngh,
V
.
K
um
a
r
,
a
nd
M
.
K
a
ur
,
“
C
l
a
s
s
i
f
i
c
a
t
i
on
of
t
he
C
O
V
I
D
-
19
i
nf
e
c
t
e
d
pa
t
i
e
nt
s
us
i
ng
D
e
ns
e
N
e
t
201
ba
s
e
d
de
e
p
t
r
a
ns
f
e
r
l
e
a
r
ni
ng,”
J
our
nal
of
B
i
om
ol
e
c
ul
a
r
St
r
uc
t
u
r
e
and
D
y
nam
i
c
s
,
vol
.
39,
no.
15,
pp.
5682
–
5689
,
2021, doi
:
10.1080/
07391102.2020.1788642.
[
10]
Y
.
S
on
g
e
t
a
l
.
,
“
D
e
e
p
l
e
a
r
n
i
ng
e
n
a
b
l
e
s
a
c
c
u
r
a
t
e
d
i
a
g
nos
i
s
o
f
n
ov
e
l
c
or
o
na
vi
r
us
(
C
O
V
I
D
-
1
9)
w
i
t
h
C
T
i
m
a
ge
s
,”
I
E
E
E
/
A
C
M
T
r
ans
ac
t
i
o
n
s
o
n
C
om
pu
t
at
i
on
a
l
B
i
ol
og
y
a
nd
B
i
o
i
nf
or
m
a
t
i
c
s
,
vo
l
.
1
8,
n
o
.
6,
pp
.
2
77
5
–
2
78
0,
20
21,
do
i
:
1
0.
11
0
9/
T
C
B
B
.2
0
21
.3
0
65
36
1
.
[
11]
L
.
Y
a
n
a
nd
X
.
L
i
m
i
ng,
“
C
or
ona
vi
r
us
di
s
e
a
s
e
2019
(
C
O
V
I
D
-
19)
:
r
ol
e
of
c
he
s
t
C
T
i
n
di
a
gnos
i
s
a
nd
m
a
n
a
ge
m
e
nt
,”
A
m
e
r
i
c
an
J
our
nal
of
R
oe
nt
ge
nol
ogy
, vol
. 214, no. 6, pp. 1280
–
1286, 2020.
[
12]
P
.
Z
hou
e
t
al
.
,
“
A
pne
um
oni
a
out
br
e
a
k
a
s
s
oc
i
a
t
e
d
w
i
t
h
a
ne
w
c
or
ona
vi
r
us
of
pr
oba
bl
e
ba
t
or
i
gi
n,”
N
at
u
r
e
,
vol
.
579,
no.
7798
,
pp. 270
–
273, 2020
, doi
:
10.1038/
s
41586
-
020
-
2012
-
7
.
[
13]
L
.
M
i
n,
C
.
Q
i
a
ng,
a
nd
Y
.
S
hui
c
he
ng,
“
N
e
t
w
or
k
i
n
ne
t
w
or
k,”
i
n
2nd I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on
L
e
ar
ni
ng
R
e
pr
e
s
e
nt
at
i
on
s
,
I
C
L
R
2014
-
C
onf
e
r
e
nc
e
T
r
a
c
k
P
r
oc
e
e
di
ng
s
, 2014.
[
14]
H
. N
i
s
hi
ur
a
, N
.
M
.
L
i
nt
on, a
nd
A
.
R
. A
khm
e
t
z
ha
nov,
“
S
e
r
i
a
l
i
nt
e
r
va
l
of
nove
l
c
or
ona
vi
r
us
(
C
O
V
I
D
-
19)
i
nf
e
c
t
i
ons
,”
I
nt
e
r
nat
i
onal
J
our
nal
of
I
nf
e
c
t
i
ous
D
i
s
e
a
s
e
s
, vol
. 93, pp. 284
–
286, 2020, doi
:
10.1016/
j
.i
j
i
d.2
020.02.060.
[
15]
M
.
T
ur
kogl
u,
“
C
O
V
I
D
e
t
e
c
t
i
oN
e
t
:
C
O
V
I
D
-
19
d
i
a
gnos
i
s
s
ys
t
e
m
ba
s
e
d
on
X
-
r
a
y
i
m
a
ge
s
us
i
ng
f
e
a
t
ur
e
s
s
e
l
e
c
t
e
d
f
r
om
p
r
e
-
l
e
a
r
ne
d
de
e
p f
e
a
t
ur
e
s
e
ns
e
m
bl
e
,”
A
ppl
i
e
d I
nt
e
l
l
i
ge
nc
e
, vol
. 51, no. 3, pp. 1213
–
1226, 2021.
[
16]
A
.
A
.
A
r
da
ka
ni
,
A
.
R
.
K
a
na
f
i
,
U
.
R
.
A
c
ha
r
ya
,
N
.
K
h
a
de
m
,
a
nd
A
.
M
oha
m
m
a
di
,
“
A
ppl
i
c
a
t
i
on
of
de
e
p
l
e
a
r
ni
ng
t
e
c
hni
que
t
o
m
a
na
ge
C
O
V
I
D
-
19
i
n
r
out
i
ne
c
l
i
ni
c
a
l
pr
a
c
t
i
c
e
us
i
ng
C
T
i
m
a
ge
s
:
r
e
s
ul
t
s
of
10
c
onvol
ut
i
ona
l
ne
ur
a
l
ne
t
w
or
ks
,”
C
om
put
e
r
s
i
n
B
i
ol
ogy
and M
e
di
c
i
ne
, vol
. 121, 2020, doi
:
10.1016/
j
.c
om
pbi
om
e
d.2020.10379
5.
[
17]
L
. B
r
une
s
e
,
F
. M
e
r
c
a
l
d
o,
A
. R
e
gi
ne
l
l
i
, a
nd A
.
S
a
n
t
on
e
, “
E
xp
l
a
i
na
b
l
e
de
e
p
l
e
a
r
ni
n
g f
or
pu
l
m
ona
r
y d
i
s
e
a
s
e
a
nd
c
o
r
o
na
v
i
r
us
C
O
V
I
D
-
19
de
t
e
c
t
i
on
f
r
om
X
-
r
a
ys
,”
C
o
m
pu
t
e
r
m
e
t
ho
ds
a
nd
pr
o
gr
am
s
i
n b
i
o
m
e
d
i
c
i
ne
,
v
ol
. 1
9
6,
202
0
,
do
i
:
1
0.1
01
6/
j
.c
m
p
b.2
02
0.1
05
608
.
[
18]
Y
.
C
he
n,
Q
.
L
i
u,
a
nd
D
.
G
uo,
“
E
m
e
r
gi
ng
c
or
ona
vi
r
us
e
s
:
G
e
nom
e
s
t
r
uc
t
ur
e
,
r
e
pl
i
c
a
t
i
on,
a
nd
pa
t
hoge
ne
s
i
s
,”
J
our
nal
of
M
e
di
c
a
l
V
i
r
ol
ogy
, vol
. 92, no. 4, pp. 418
–
423, 2020, doi
:
10.1002/
j
m
v.25681.
[
19]
S
.
B
.
S
t
oe
c
kl
i
n
e
t
al
.
,
“
F
i
r
s
t
c
a
s
e
s
of
c
or
ona
vi
r
us
di
s
e
a
s
e
2019
(
C
O
V
I
D
-
19)
i
n
F
r
a
nc
e
:
s
ur
ve
i
l
l
a
nc
e
,
i
nve
s
t
i
ga
t
i
on
s
a
nd
c
ont
r
ol
m
e
a
s
ur
e
s
, J
a
nua
r
y 2020,”
E
ur
os
u
r
v
e
i
l
l
anc
e
, vol
. 25, no. 6, 2020, doi
:
10.2807/
1560
-
7917.E
S
.2020.25.6.2000094.
[
20]
D
.
P
.
F
a
n
e
t
al
.
,
“
I
nf
-
N
e
t
:
a
ut
om
a
t
i
c
C
O
V
I
D
-
19
l
ung
i
nf
e
c
t
i
on
s
e
gm
e
nt
a
t
i
on
f
r
om
C
T
i
m
a
ge
s
,”
I
E
E
E
T
r
ans
ac
t
i
ons
on
M
e
di
c
al
I
m
agi
ng
, vol
. 39, no. 8, pp. 2626
–
2637, 2020, doi
:
10.1109/
T
M
I
.2020.2996645
.
[
21]
R
.
M
.
P
e
r
e
i
r
a
,
D
.
B
e
r
t
ol
i
ni
,
L
.
O
.
T
e
i
xe
i
r
a
,
C
.
N
.
S
i
l
l
a
,
a
nd
Y
.
M
.
G
.
C
os
t
a
,
“
C
O
V
I
D
-
19
i
de
nt
i
f
i
c
a
t
i
on
i
n
c
he
s
t
X
-
r
a
y
i
m
a
ge
s
on
f
l
a
t
a
nd
hi
e
r
a
r
c
hi
c
a
l
c
l
a
s
s
i
f
i
c
a
t
i
on
s
c
e
na
r
i
os
,”
C
om
put
e
r
M
e
t
hods
and
P
r
ogr
am
s
i
n
B
i
om
e
di
c
i
ne
,
vol
.
194,
2020,
doi
:
10.1016/
j
.c
m
pb.2020.105532.
[
22]
E
u
s
o
M
I
I
, “
A
E
u
r
op
e
a
n
i
n
i
t
i
a
t
i
v
e
f
o
r
a
u
t
o
m
a
t
e
d d
i
a
g
n
os
i
s
a
n
d q
ua
n
t
i
t
a
t
i
ve
a
n
a
l
ys
i
s
o
f
C
O
V
I
D
-
1
9 o
n i
m
a
g
i
n
g,
”
E
u
s
o
M
I
I
,
2
02
0
. [
O
n
l
i
ne
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
ps
:
/
/
w
w
w
.e
u
s
o
m
i
i
.
o
r
g
/
a
-
e
u
r
op
e
a
n
-
i
n
i
t
i
a
t
i
v
e
-
f
o
r
-
a
u
t
o
m
a
t
e
d
-
d
i
a
g
n
os
i
s
-
a
n
d
-
q
u
a
n
t
i
t
a
t
i
ve
-
a
na
l
ys
i
s
-
of
-
c
o
v
i
d
-
19
-
on
-
i
m
a
g
i
n
g
/
[
23]
N
.
Z
ha
ng
e
t
al
.
,
“
R
e
c
e
nt
a
dva
nc
e
s
i
n
t
he
de
t
e
c
t
i
on
of
r
e
s
pi
r
a
t
or
y
vi
r
us
i
nf
e
c
t
i
on
i
n
hum
a
ns
,”
J
our
nal
of
M
e
di
c
al
V
i
r
ol
ogy
,
vol
. 92, no. 4, pp. 408
–
417, 2020, doi
:
10.1002/
j
m
v.25674.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
4
,
A
ugus
t
2025
:
3192
-
3200
3200
[
24]
R
S
N
A
,
“
R
S
N
A
a
nnounc
e
s
C
O
V
I
D
-
19
i
m
a
gi
ng
da
t
a
r
e
pos
i
t
or
y,”
R
ad
i
ol
og
i
c
al
Soc
i
e
t
y
of
N
or
t
h
A
m
e
r
i
c
a
,
2021
.
[
O
nl
i
ne
]
.
A
va
i
l
a
bl
e
:
ht
t
ps
:
/
/
w
w
w
.r
s
na
.or
g/
ne
w
s
/
2020/
m
a
r
c
h/
c
ovi
d
-
19
-
i
m
a
gi
ng
-
da
t
a
-
r
e
pos
i
t
or
y
[
25]
V
. M
.
C
or
m
a
n
e
t
al
.
,
“
D
e
t
e
c
t
i
on
of
2019 nove
l
c
or
ona
vi
r
us
(
2019
-
nC
oV
)
by r
e
a
l
-
t
i
m
e
R
T
-
P
C
R
,”
E
ur
os
u
r
v
e
i
l
l
anc
e
, vol
.
25, no. 3
,
2020, doi
:
10.2807/
1560
-
7917.E
S
.2020.25.3.2000045.
[
26]
J
.
W
a
ng
a
nd
L
.
P
e
r
e
z
,
“
T
he
e
ff
e
c
t
i
ve
ne
s
s
of
da
t
a
a
ugm
e
nt
a
t
i
on
i
n
i
m
a
ge
c
l
a
s
s
i
f
i
c
a
t
i
on
us
i
ng
de
e
p
l
e
a
r
ni
ng,”
ar
X
i
v
-
C
om
put
e
r
Sc
i
e
nc
e
, pp. 1
-
8,
2017.
[
27]
P
.
Y
a
da
v,
N
.
M
e
non,
V
.
R
a
vi
,
a
nd
S
.
V
i
s
hva
na
t
ha
n,
“
L
ung
-
G
A
N
s
:
un
s
u
pe
r
vi
s
e
d
r
e
pr
e
s
e
nt
a
t
i
on
l
e
a
r
ni
ng
f
or
l
ung
di
s
e
a
s
e
c
l
a
s
s
i
f
i
c
a
t
i
on
us
i
ng
c
he
s
t
C
T
a
nd
X
-
r
a
y
i
m
a
ge
s
,”
I
E
E
E
T
r
ans
ac
t
i
ons
on
E
ngi
n
e
e
r
i
ng
M
anage
m
e
nt
,
vol
.
70,
no.
8,
pp.
2774
–
2786,
2023, doi
:
10.1109/
T
E
M
.2021.3103334.
[
28]
J
.
D
e
ng
,
W
.
D
ong
,
R
.
S
oc
he
r
,
L
.
J
.
L
i
,
K
.
L
i
,
a
n
d
L
.
F
e
i
-
F
e
i
,
“
I
m
a
ge
N
e
t
:
a
l
a
r
g
e
-
s
c
a
l
e
h
i
e
r
a
r
c
h
i
c
a
l
i
m
a
ge
da
t
a
b
a
s
e
,”
i
n
2
009
I
E
E
E
C
onf
e
r
e
nc
e
o
n C
om
p
ut
e
r
V
i
s
i
on
a
nd
P
at
t
e
r
n
R
e
c
o
gn
i
t
i
on
, C
V
P
R
200
9
,
20
09
, p
p.
248
–
25
5
,
do
i
:
1
0.1
10
9/
C
V
P
R
.20
09
.52
06
848
.
[
29]
A
.
A
r
una
c
ha
l
a
m
,
V
.
R
a
vi
,
V
.
A
c
ha
r
ya
,
a
nd
T
.
D
.
P
ha
m
,
“
T
ow
a
r
d
da
t
a
-
m
o
de
l
-
a
gnos
t
i
c
a
ut
onom
ous
m
a
c
hi
ne
-
ge
ne
r
a
t
e
d
da
t
a
l
a
be
l
i
ng
a
nd
a
nnot
a
t
i
on
pl
a
t
f
or
m
:
C
O
V
I
D
-
19
a
ut
oa
nnot
a
t
i
on
us
e
c
a
s
e
,”
I
E
E
E
T
r
ans
ac
t
i
ons
on
E
ngi
ne
e
r
i
ng
M
anage
m
e
nt
,
vol
.
70,
no. 8, pp. 2695
–
2706, 2023, doi
:
10.1109/
T
E
M
.2021.3094544.
[
30]
A
.
B
e
n
-
C
o
h
e
n
,
E
.
K
l
a
n
g
,
M
.
M
.
A
m
i
t
a
i
,
J
.
G
o
l
d
be
r
g
e
r
,
a
n
d
H
.
G
r
e
e
n
s
p
a
n
,
“
A
n
a
t
o
m
i
c
a
l
da
t
a
a
u
gm
e
n
t
a
t
i
o
n
f
o
r
C
N
N
ba
s
e
d
p
i
xe
l
-
w
i
s
e
c
l
a
s
s
i
f
i
c
a
t
i
on
,
”
i
n
2
0
18
I
E
E
E
1
5
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
Sy
m
po
s
i
um
o
n
B
i
om
e
d
i
c
a
l
I
m
a
g
i
n
g
(
I
S
B
I
2
0
18
)
,
2
0
18
,
p
p
.
1
0
9
6
–
1099
,
d
o
i
:
10.1109/
I
S
B
I
.20
18.83
63762
.
[
31]
D
.
M
ukht
or
ov,
M
.
R
a
khm
onova
,
S
.
M
uk
s
i
m
ova
,
a
nd
Y
.
I
.
C
ho,
“
E
ndos
c
opi
c
i
m
a
ge
c
l
a
s
s
i
f
i
c
a
t
i
on
ba
s
e
d
on
e
xpl
a
i
na
bl
e
de
e
p
l
e
a
r
ni
ng,”
Se
ns
or
s
, vol
. 23, no. 6, 2023, doi
:
10.3390/
s
23063176.
[
32]
C
.
I
.
P
a
ul
e
s
,
H
.
D
.
M
a
r
s
t
on,
a
nd
A
.
S
.
F
a
uc
i
,
“
C
or
ona
vi
r
us
i
nf
e
c
t
i
ons
m
or
e
t
ha
n
j
us
t
t
he
c
om
m
on
c
ol
d,”
J
A
M
A
-
J
our
nal
of
t
h
e
A
m
e
r
i
c
an M
e
di
c
al
A
s
s
oc
i
at
i
on
, vol
. 323, no. 8, pp. 707
–
708, 2020, doi
:
10.1001/
j
a
m
a
.2020.0757.
B
I
O
G
R
A
P
H
I
E
S
O
F
A
U
T
H
O
R
S
S
ivanagireddy
Kalli
is
presently
a
Profes
sor
of
Electronics
and
Communication
Engineering.
He
did
his
Ph.D.
degree
in
Electronics
and
Commu
nication
Engineering
from
JNTU
Hyderabad
in
2019.
He
is
having
more
than
22
years
of
teachi
ng
and
8
years
of
research
experience.
He
has
a
total
of
60
research
publicat
ions
in
internationa
l
journals
and
9
patents.
His
current
research
areas
are
artificial
intell
igence,
machine
learning,
a
nd
deep
learning.
He
is
a
member of
IEEE.
He can be contacted at email:
sivanagireddyka
lli@
gmail.com
.
Bukka
Narendra
Kumar
is
presently
a
Profes
sor
of
Computer
Scienc
e
and
Engineering
and
HoD
.
He
did
his
Ph.D.
degree
in
Computer
Scie
nce
and
Engineering
from
JNTU
Hyderabad
in
2019.
He
is
having
more
than
25
years
of
teachi
ng
and
8
years
of
research
experience.
He
has
a
total
of
20
research
publicat
ions
in
internationa
l
journals
and
4
patents.
His
current
research
areas
are
informat
ion
security,
artifici
al
intell
igence, m
achine
learning,
and
dee
p
learning.
He is a lif
e member
of MISTE.
He can be contacted at email:
bnkphd@
gmail.com
.
Saggurt
hi
Jagadee
sh
is
presently
Profes
sor
of
Electronics
and
Communication
Engineering.
He
did
his
Ph.D.
degree
in
Electronics
and
Commu
nication
Engineering
from
JNTU
Hyderabad
in
2019.
He
is
having
more
th
an
25
years
of
Teachi
ng
and
8
years
of
researc
h
experience.
He
has
a
total
of
40
research
publicat
ions
in
internationa
l
journals
and
9
patents.
His
current
research
areas
are
artificial
intell
igence,
machine
learning,
a
nd
deep
learning.
He
is
a
member of
IEEE.
He can be contacted at email:
jaaga.ssje
c@
gmail.com.
Kushagari
Chandramou
li
Ravi
K
umar
is
presently
Profess
or
of
Computer
Scienc
e
and
Engine
ering
.
He
did
his
Ph.D.
degree
in
Computer
Scie
nce
and
Engineering
from
JNTU
Hyderabad
in
2019.
He
is
having
more
than
29
years
of
teachi
ng
and
8
years
of
research
experience.
He
has
a
total
of
20
research
publicat
ions
in
internationa
l
journals
and
3
patents.
His
current
research
areas
are
artificial
intell
igence,
machine
learning,
a
nd
deep
learning.
He
is
a
member of
MISTE.
He can be contacted at email:
kcravikunar1971@gmail.com
.
Evaluation Warning : The document was created with Spire.PDF for Python.