I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
1
5
,
No
.
2
,
A
p
r
il 2
0
2
6
,
p
p
.
1
6
0
5
~
1
6
1
2
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
2
.
p
p
1
6
0
5
-
1
6
1
2
1605
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Deep learni
ng
ens
embles
f
o
r
lung
c
a
ncer detectio
n
in
t
ho
ra
cic
CT sca
ns lev
erag
ing
genera
tive a
dv
ersa
ria
l net
wo
rk
t
echno
lo
g
y
B
inee
s
h M
o
o
zhi
pp
ura
t
h
1,
2
,
J
a
y
a
pa
nd
ia
n Na
t
a
ra
j
a
n
1
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
C
H
R
I
S
T
U
n
i
v
e
r
si
t
y
,
B
e
n
g
a
l
u
r
u
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
A
r
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
D
a
t
a
S
c
i
e
n
c
e
,
J
y
o
t
h
i
En
g
i
n
e
e
r
i
n
g
C
o
l
l
e
g
e
,
Th
r
i
ssu
r
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l 8
,
2
0
2
5
R
ev
is
ed
J
an
2
6
,
2
0
2
6
Acc
ep
ted
Feb
6
,
2
0
2
6
Eff
e
c
ti
v
e
trea
tme
n
t
o
f
l
u
n
g
c
a
n
c
e
r
d
e
p
e
n
d
s
o
n
e
a
rly
a
n
d
a
c
c
u
ra
te
d
e
tec
ti
o
n
,
wh
ich
c
o
n
t
in
u
e
s
to
b
e
a
m
a
jo
r
c
a
u
se
o
f
c
a
n
c
e
r
-
re
l
a
ted
fa
talit
ies
g
lo
b
a
ll
y
.
Co
n
v
e
n
ti
o
n
a
l
d
iag
n
o
stic
tec
h
n
i
q
u
e
s
a
re
u
se
fu
l,
b
u
t
th
e
ir
e
ffica
c
y
in
h
a
n
d
li
n
g
larg
e
a
m
o
u
n
ts o
f
th
o
ra
c
ic
c
o
m
p
u
t
e
d
to
m
o
g
ra
p
h
y
(
CT
)
sc
a
n
d
a
ta i
s l
imited
b
y
th
e
ir
ti
m
e
-
c
o
n
s
u
m
in
g
n
a
tu
re
a
n
d
su
sc
e
p
ti
b
il
it
y
to
h
u
m
a
n
e
rro
r.
T
h
e
re
se
a
rc
h
h
e
re
su
g
g
e
sts
a
n
e
w
d
e
e
p
le
a
rn
in
g
m
o
d
e
l
t
h
a
t
i
n
teg
ra
tes
g
e
n
e
ra
ti
v
e
a
d
v
e
rsa
rial
n
e
two
r
k
s
(G
AN
s)
fo
r
d
a
ta
imp
r
o
v
e
m
e
n
t
wi
th
a
so
p
h
isti
c
a
ted
e
n
se
m
b
le
a
p
p
r
o
a
c
h
t
o
c
las
sifica
ti
o
n
.
G
AN
s
a
re
e
m
p
lo
y
e
d
t
o
g
e
n
e
ra
te
re
a
li
stic
sy
n
th
e
ti
c
CT
ima
g
e
s,
a
d
d
re
ss
in
g
th
e
c
h
a
ll
e
n
g
e
s
o
f
li
m
it
e
d
d
a
tas
e
ts.
Th
e
b
a
c
k
b
o
n
e
o
f
th
e
p
r
o
p
o
se
d
a
p
p
ro
a
c
h
is
a
c
o
n
se
n
su
s
-
g
u
i
d
e
d
a
d
a
p
ti
v
e
b
len
d
in
g
(CG
AB)
e
n
se
m
b
le
m
o
d
e
l
th
a
t
lea
rn
s
to
d
y
n
a
m
ica
ll
y
c
o
m
b
in
e
t
h
e
p
re
d
ictio
n
s
o
f
t
h
re
e
to
p
-
p
e
rfo
rm
i
n
g
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
two
rk
s
(CNN
s)
:
Re
sN
e
t
-
1
5
2
,
De
n
se
Ne
t
-
1
6
9
,
a
n
d
Eff
icie
n
tNe
t
-
B7
.
T
h
e
CG
A
B
m
o
d
e
l
imp
ro
v
e
s
p
re
d
ictio
n
a
c
c
u
ra
c
y
th
r
o
u
g
h
m
o
d
e
l
c
o
n
tri
b
u
ti
o
n
we
i
g
h
ti
n
g
b
a
se
d
o
n
c
o
n
fi
d
e
n
c
e
s
c
o
re
s
a
n
d
in
ter
-
m
o
d
e
l
c
o
n
se
n
su
s,
wh
il
e
a
c
o
n
fli
c
t
-
re
so
lv
in
g
a
u
x
il
iary
d
e
c
isio
n
m
o
d
e
l
is
u
se
d
.
Th
e
a
p
p
r
o
a
c
h
wa
s
tes
ted
u
sin
g
th
e
l
u
n
g
ima
g
e
d
a
tab
a
se
c
o
n
so
rti
u
m
a
n
d
th
e
ima
g
e
d
a
tab
a
se
re
so
u
rc
e
in
it
iati
v
e
(
LIDC
-
IDRI
)
d
a
tas
e
t
wit
h
a
d
e
tec
ti
o
n
ra
te
o
f
9
7
.
3
5
%
,
su
r
p
a
ss
in
g
sin
g
le
-
m
o
d
e
l
a
n
d
trad
it
io
n
a
l
e
n
se
m
b
le
m
e
th
o
d
s.
Th
e
c
u
r
re
n
t
w
o
rk
p
r
o
v
i
d
e
s
a
so
li
d
a
n
d
sc
a
lab
le
a
p
p
r
o
a
c
h
t
o
lu
n
g
c
a
n
c
e
r
d
e
tec
ti
o
n
wit
h
b
e
tt
e
r
g
e
n
e
ra
li
z
a
ti
o
n
,
in
c
re
a
se
d
d
iag
n
o
stic c
o
n
siste
n
c
y
,
a
n
d
a
p
p
li
c
a
b
il
i
ty
f
o
r
c
li
n
ica
l
u
se
.
K
ey
w
o
r
d
s
:
Dee
p
lear
n
in
g
E
n
s
em
b
le
lear
n
in
g
Gen
er
ativ
e
ad
v
e
r
s
ar
ial
n
etwo
r
k
s
L
u
n
g
ca
n
ce
r
d
etec
tio
n
Me
d
ical
im
ag
in
g
T
h
o
r
ac
ic
co
m
p
u
ted
to
m
o
g
r
ap
h
y
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
J
ay
ap
an
d
ian
Nata
r
ajan
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
an
d
E
n
g
in
ee
r
in
g
,
C
HR
I
ST
Un
iv
er
s
ity
B
en
g
alu
r
u
,
I
n
d
ia
E
m
ail:
jay
ap
an
d
ian
.
n
@
ch
r
is
tu
n
iv
er
s
ity
.
in
1.
I
NT
RO
D
UCT
I
O
N
Ar
tific
ial
in
tellig
en
ce
(
AI
)
is
tr
an
s
f
o
r
m
in
g
h
ea
lth
ca
r
e
b
y
en
ab
lin
g
m
o
r
e
ac
c
u
r
ate,
ef
f
i
cien
t,
an
d
s
ca
lab
le
d
ec
is
io
n
-
m
ak
in
g
[
1
]
.
AI
s
y
s
tem
s
ar
e
g
iv
in
g
ass
is
tan
ce
in
d
is
ea
s
e
d
iag
n
o
s
is
,
o
u
tco
m
e
p
r
ed
ictio
n
,
a
n
d
tr
ea
tm
en
t o
p
tim
izatio
n
.
AI
im
p
r
o
v
es v
alu
es in
r
eso
u
r
ce
r
estricte
d
en
v
ir
o
n
m
e
n
ts
th
r
o
u
g
h
a
u
to
m
atin
g
r
ep
etitiv
e
task
s
,
p
r
o
ce
s
s
in
g
lar
g
e
-
s
ca
le
d
ata
in
r
ea
l
tim
e,
an
d
im
p
r
o
v
in
g
d
iag
n
o
s
tic
co
n
s
is
ten
cy
[
2
]
.
Of
th
e
m
an
y
AI
tech
n
o
lo
g
ies,
d
ee
p
lear
n
in
g
,
an
d
m
o
r
e
s
p
ec
if
ically
,
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
C
NNs),
is
th
e
b
est
tech
n
o
lo
g
y
f
o
r
e
x
tr
ac
tin
g
h
ier
ar
ch
ical
f
ea
tu
r
es
f
r
o
m
a
v
ar
iet
y
o
f
m
ed
ical
d
iag
n
o
s
tics
in
clu
d
in
g
ca
r
d
i
o
v
ascu
lar
im
ag
in
g
,
b
r
ain
tu
m
o
r
d
etec
tio
n
,
an
d
d
iab
etic
r
etin
o
p
ath
y
[
3
]
,
[
4
]
.
Patien
t
ca
r
e
d
ep
en
d
s
p
r
im
ar
il
y
o
n
th
e
ab
ilit
y
to
d
iag
n
o
s
e
a
p
atien
t'
s
co
n
d
itio
n
in
a
tim
ely
m
an
n
er
.
C
o
n
v
en
tio
n
al
im
ag
i
n
g
tech
n
iq
u
es
f
o
r
d
ia
g
n
o
s
is
ar
e,
h
o
wev
er
,
tim
e
-
co
n
s
u
m
in
g
,
a
r
e
d
ep
en
d
e
n
t
o
n
t
h
e
ex
p
er
tis
e
o
f
th
e
p
er
s
o
n
co
n
d
u
ctin
g
th
e
test
,
an
d
ar
e
s
u
b
ject
to
a
h
ig
h
d
eg
r
ee
o
f
h
u
m
an
v
ar
iab
ilit
y
.
Sy
s
tem
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
2
,
Ap
r
il 2
0
2
6
:
1
6
0
5
-
1
6
1
2
1606
b
ased
o
n
C
NN
tech
n
o
lo
g
y
ad
d
r
ess
th
ese
is
s
u
es
b
y
au
to
m
atica
lly
in
ter
p
r
etin
g
an
d
d
iag
n
o
s
in
g
im
ag
es
an
d
ar
e
th
er
ef
o
r
e
ess
en
tial to
r
a
d
io
lo
g
y
an
d
p
ath
o
lo
g
y
f
o
r
s
eg
m
e
n
tatio
n
,
class
if
icatio
n
,
an
d
p
atter
n
r
ec
o
g
n
itio
n
[
5
]
.
L
u
n
g
c
an
ce
r
is
t
h
e
m
o
s
t
co
m
m
o
n
ca
u
s
e
o
f
ca
n
ce
r
d
ea
th
s
d
u
e
t
o
i
ts
as
y
m
p
to
m
a
tic
p
r
o
g
r
e
s
s
io
n
a
n
d
m
o
s
tl
y
l
at
e
s
t
a
g
es
o
f
d
et
ec
ti
o
n
[
6
]
.
I
t
is
w
ell
r
ec
o
g
n
iz
ed
t
h
at
th
e
ea
r
l
y
d
ia
g
n
o
s
is
o
f
l
u
n
g
ca
n
ce
r
s
i
g
n
if
ic
an
tl
y
im
p
r
o
v
es
a
p
ati
e
n
t'
s
ch
a
n
ce
s
o
f
s
u
r
v
i
v
a
l.
T
h
o
r
a
cic
c
o
m
p
u
te
d
t
o
m
o
g
r
a
p
h
y
(
CT
)
s
ca
n
s
a
r
e
th
e
u
s
u
a
l
m
et
h
o
d
f
o
r
th
e
d
e
te
cti
o
n
o
f
l
u
n
g
n
o
d
u
l
es
[
7
]
.
H
o
we
v
er
,
d
u
e
to
a
la
r
g
e
n
u
m
b
er
o
f
C
T
s
li
ce
s
a
n
d
v
ar
ia
b
il
ities
o
f
t
h
e
n
o
d
u
les
,
in
t
er
p
r
et
in
g
t
h
e
C
T
s
c
a
n
is
a
t
im
e
-
co
n
s
u
m
in
g
t
as
k
wit
h
a
h
i
g
h
p
o
s
s
i
b
il
it
y
o
f
m
is
t
ak
es.
Hen
ce
,
r
ese
ar
c
h
o
n
au
t
o
m
at
e
d
d
ee
p
le
a
r
n
in
g
-
b
ase
d
d
ia
g
n
o
s
is
is
m
er
ite
d
.
So
f
a
r
,
C
NNs
h
av
e
d
em
o
n
s
t
r
at
e
d
g
r
ea
t
ca
p
a
b
i
lit
y
i
n
th
o
r
ac
ic
i
m
a
g
i
n
i
n
g
,
r
e
v
e
ali
n
g
s
u
b
tle
f
e
at
u
r
es
i
n
v
is
i
b
le
t
o
c
la
s
s
ic
m
e
t
h
o
d
s
[
8
]
.
Ho
w
e
v
e
r
,
t
h
e
ch
all
e
n
g
e
o
f
t
h
e
lac
k
o
f
an
n
o
ta
te
d
C
T
d
at
ase
ts
e
x
ac
e
r
b
ates
o
v
er
f
i
tti
n
g
a
n
d
p
o
o
r
m
o
d
el
g
e
n
e
r
al
iz
ati
o
n
[
9
]
.
Un
d
e
r
u
til
iza
ti
o
n
o
f
m
e
d
ic
al
im
a
g
i
n
g
d
at
asets
is
f
o
c
u
s
e
d
o
n
,
i
m
p
r
o
v
i
n
g
d
at
aset
d
i
v
er
s
i
ty
t
h
r
o
u
g
h
g
e
n
e
r
a
ti
v
e
m
o
d
els
a
n
d
,
m
o
r
e
p
r
ec
is
el
y
,
g
e
n
e
r
a
ti
v
e
a
d
v
er
s
ar
ial
n
e
tw
o
r
k
s
(
G
ANs)
[
1
0
]
–
[
1
2
]
.
On
t
h
e
o
n
e
h
a
n
d
,
o
n
e
s
h
o
u
l
d
n
o
t
f
o
r
g
et
t
h
a
t
GAN
-
g
en
er
ate
d
im
ag
es
m
ig
h
t
i
n
t
r
o
d
u
ce
ar
t
if
ici
al
a
r
t
if
ac
ts
o
r
b
i
ases
o
f
s
u
b
t
le
f
e
at
u
r
es
i
n
c
ase
o
f
a
l
ac
k
o
f
r
i
g
o
r
o
u
s
v
ali
d
ati
o
n
,
an
d
th
e
u
s
e
o
f
l
im
i
te
d
p
u
b
lic
d
ata
li
k
e
l
u
n
g
im
a
g
e
d
a
ta
b
as
e
c
o
n
s
o
r
t
iu
m
an
d
i
m
a
g
e
d
at
ab
ase
r
es
o
u
r
ce
i
n
iti
ati
v
e
(
L
I
DC
-
I
DR
I
)
m
a
y
r
es
tr
ict
e
x
t
er
n
a
l
g
en
er
al
izat
io
n
a
cr
o
s
s
d
iv
er
s
e
cli
n
i
ca
l
p
o
p
u
l
ati
o
n
s
.
Nev
er
th
eless
,
GAN
-
b
ased
au
g
m
en
tatio
n
h
as
p
r
o
v
en
an
ef
f
ec
tiv
e
to
o
l
f
o
r
ex
te
n
d
in
g
tr
a
in
in
g
d
ata,
im
p
r
o
v
in
g
r
o
b
u
s
tn
ess
,
an
d
e
n
h
an
cin
g
p
e
r
f
o
r
m
an
ce
in
b
o
t
h
lu
n
g
n
o
d
u
le
class
if
icatio
n
an
d
s
eg
m
e
n
tatio
n
.
B
ey
o
n
d
th
e
is
s
u
es
o
f
d
ata
s
ca
r
city
,
r
elian
ce
o
n
s
in
g
le
m
o
d
els
r
estricts
d
iag
n
o
s
tic
r
eliab
ilit
y
.
E
n
s
em
b
le
lear
n
in
g
d
ea
ls
with
th
e
ab
o
v
e
is
s
u
e
b
y
co
m
b
in
in
g
th
e
o
u
tp
u
t
s
o
f
m
u
ltip
le
m
o
d
els
in
o
r
d
er
to
r
ed
u
ce
v
a
r
ian
ce
an
d
b
ias
wh
ile
en
h
a
n
cin
g
a
cc
u
r
ac
y
[
1
3
]
.
T
h
e
r
esear
c
h
e
r
s
ad
o
p
t
n
eu
r
al
n
etwo
r
k
e
n
s
em
b
les
to
ad
d
r
ess
m
o
d
el
-
s
p
ec
if
ic
b
iases
an
d
h
e
n
ce
en
h
a
n
ce
d
ia
g
n
o
s
tic
co
n
f
i
d
en
ce
[
1
4
]
.
Var
io
u
s
s
tate
-
of
-
th
e
-
ar
t
b
ac
k
b
o
n
es,
in
clu
d
in
g
R
esNet,
Den
s
eNe
t,
an
d
E
f
f
icien
tNet,
p
r
o
v
i
d
e
c
o
m
p
lem
en
tar
y
s
tr
en
g
th
s
.
R
esNet
-
1
5
2
with
r
esid
u
al
lin
k
s
d
ee
p
en
s
f
ea
tu
r
e
lear
n
in
g
[
1
5
]
,
wh
ile
Den
s
eNe
t
-
1
6
9
en
co
u
r
a
g
es
f
ea
tu
r
e
r
eu
s
e
b
y
m
ain
tain
in
g
ef
f
icien
t
g
r
ad
ien
t
f
lo
w
[
1
6
]
.
E
f
f
icien
tNet
-
B
7
b
alan
ce
s
d
e
p
th
,
wid
th
,
a
n
d
r
eso
l
u
tio
n
a
n
d
th
u
s
ac
h
iev
e
s
to
p
ac
c
u
r
ac
y
at
a
r
ea
s
o
n
ab
le
co
m
p
u
tatio
n
al
co
s
t
[
1
7
]
.
W
eig
h
ted
av
e
r
ag
in
g
,
s
tack
in
g
,
an
d
v
o
tin
g
a
r
e
en
s
em
b
l
e
tech
n
iq
u
es wh
ich
f
u
r
th
er
p
u
s
h
th
e
p
e
r
f
o
r
m
an
ce
b
o
u
n
d
ar
ies
o
n
th
e
class
if
icatio
n
o
f
b
r
ea
s
t
ca
n
ce
r
,
p
n
eu
m
o
n
ia,
C
OVI
D
-
1
9
,
an
d
b
r
ain
tu
m
o
r
s
[
1
8
]
,
[
1
9
]
.
R
ec
en
t
wo
r
k
s
ten
d
to
g
o
to
wa
r
d
s
h
y
b
r
id
e
n
s
em
b
les
th
at
m
er
g
e
C
NNs
with
r
ec
u
r
r
en
t
n
eu
r
a
l
n
etwo
r
k
s
(
R
NN
s
)
o
r
t
r
an
s
f
o
r
m
e
r
s
f
o
r
ca
p
tu
r
in
g
tem
p
o
r
al
an
d
s
p
atial
in
f
o
r
m
atio
n
.
Mo
s
t
r
ec
e
n
tly
,
t
h
e
r
e
is
a
tr
en
d
to
war
d
h
ig
h
er
im
ag
e
f
id
elity
an
d
r
ich
er
im
ag
e
v
ar
iatio
n
b
y
au
g
m
en
tatio
n
,
in
clu
d
in
g
GANs,
v
ar
iatio
n
al
au
to
en
c
o
d
er
s
(
VAE
s
)
,
an
d
d
if
f
u
s
io
n
m
o
d
els
[
2
0
]
–
[
2
3
]
.
T
h
e
s
u
cc
ess
o
f
GAN
-
em
b
e
d
d
ed
d
ee
p
m
o
d
els
in
d
if
f
er
en
t
ap
p
licatio
n
s
s
tr
en
g
th
e
n
s
th
e
p
o
wer
o
f
th
is
a
u
g
m
e
n
ted
g
e
n
er
ativ
e
f
r
am
ewo
r
k
,
esp
ec
iall
y
in
th
e
d
o
m
ain
o
f
m
ed
ical
im
ag
in
g
[
2
4
]
.
T
h
e
wo
r
k
p
r
o
p
o
s
es,
f
o
r
th
e
f
ir
s
t
tim
e,
th
e
co
n
s
en
s
u
s
-
g
u
id
e
d
ad
ap
tiv
e
b
len
d
i
n
g
(
C
GAB)
f
r
am
ewo
r
k
f
o
r
ea
r
ly
-
s
tag
e
lu
n
g
ca
n
ce
r
p
r
ed
ictio
n
:
a
n
o
v
el
ad
a
p
tiv
e
e
n
s
em
b
le
ar
ch
itectu
r
e
t
h
at
is
d
esig
n
ed
to
en
h
an
ce
r
o
b
u
s
tn
ess
an
d
ac
cu
r
ac
y
in
th
o
r
ac
ic
C
T
an
aly
s
is
.
Alg
o
r
ith
m
ically
,
C
GAB
s
y
n
th
esizes
th
e
o
u
tp
u
ts
o
f
m
u
ltip
le
d
ee
p
C
NNs,
n
am
ely
R
esNet
-
1
5
2
,
Den
s
eNe
t
-
1
6
9
,
an
d
E
f
f
icien
tNet
-
B
7
,
th
r
o
u
g
h
a
h
ie
r
ar
ch
ical
en
s
em
b
le
m
ec
h
an
is
m
th
at
ass
ig
n
s
weig
h
ts
d
y
n
am
ically
b
ased
o
n
th
e
co
n
f
id
en
ce
o
f
ea
ch
m
o
d
el
an
d
th
e
in
ter
-
m
o
d
e
l
co
n
s
en
s
u
s
.
Un
lik
e
th
eir
class
ical
s
tatic
co
u
n
ter
p
ar
ts
,
C
GA
B
ch
an
g
es
its
weig
h
tin
g
o
n
a
p
er
-
s
am
p
le
b
asis
,
d
am
p
en
in
g
th
e
in
f
l
u
en
ce
o
f
t
h
e
u
n
ce
r
tain
p
r
ed
ictio
n
s
wh
il
e
en
h
an
cin
g
th
e
r
eliab
ilit
y
o
f
th
e
f
in
al
d
ec
is
io
n
s
.
T
h
is
p
r
o
ce
s
s
o
f
ad
ap
tatio
n
b
y
lo
o
k
in
g
at
co
n
s
en
s
u
s
av
o
id
s
o
v
er
f
itti
n
g
an
d
in
co
n
s
is
ten
cy
o
f
in
d
iv
id
u
al
C
NNs
in
class
if
y
in
g
lu
n
g
n
o
d
u
les
ef
f
ec
tiv
ely
.
I
m
p
r
o
v
ed
p
r
ec
is
io
n
an
d
s
tab
ilit
y
co
m
p
ar
ed
to
s
tan
d
-
alo
n
e
C
NNs
an
d
tr
ad
itio
n
al
en
s
em
b
les
wer
e
o
b
s
er
v
ed
u
p
o
n
th
e
v
alid
atio
n
o
f
L
I
DC
-
I
DR
I
[
2
5
]
.
C
GAB,
th
er
ef
o
r
e,
o
v
er
co
m
es
lim
itatio
n
s
in
b
o
th
d
ata
a
n
d
v
ar
iab
ilit
y
in
f
o
r
ec
asts
,
h
e
n
ce
p
r
o
v
id
in
g
a
s
ca
lab
le
an
d
in
ter
p
r
eta
b
le
AI
f
r
am
ewo
r
k
f
o
r
tr
u
ly
d
ep
en
d
ab
le
d
etec
tio
n
o
f
l
u
n
g
c
an
ce
r
u
s
in
g
C
T
im
ag
in
g
.
2.
M
E
T
H
O
D
T
h
is
p
ap
er
p
r
o
p
o
s
es a
n
en
d
-
to
-
en
d
lu
n
g
ca
n
ce
r
d
etec
tio
n
f
r
a
m
ewo
r
k
b
y
in
teg
r
atin
g
GAN
-
d
r
iv
en
d
ata
au
g
m
en
tatio
n
with
th
e
C
GAB
en
s
em
b
le.
T
h
e
wo
r
k
f
lo
w,
as
illu
s
tr
ated
in
Fig
u
r
e
1
,
co
n
s
is
ts
o
f
d
ee
p
co
n
v
o
l
u
tio
n
al
g
en
e
r
ativ
e
ad
v
e
r
s
ar
ial
n
etwo
r
k
(
DC
GAN
)
-
d
r
iv
en
s
y
n
th
etic
d
ata
g
en
e
r
atio
n
,
f
ea
tu
r
es
ex
tr
ac
ted
u
s
in
g
R
esNet
-
1
5
2
,
Den
s
eNe
t
-
1
6
9
,
a
n
d
E
f
f
icien
tNet
-
B
7
,
an
d
ad
ap
tiv
e
e
n
s
em
b
le
f
u
s
io
n
th
at
in
co
r
p
o
r
ates
au
x
iliar
y
co
n
f
lict
r
eso
lu
tio
n
an
d
g
r
ad
ie
n
t
-
weig
h
ted
class
ac
tiv
atio
n
m
ap
p
in
g
(
Gr
ad
-
C
AM
)
–
b
ased
in
ter
p
r
etab
ilit
y
.
Af
ter
th
at,
t
h
e
s
u
g
g
ested
m
et
h
o
d
was
test
ed
o
n
th
e
a
n
n
o
tated
L
I
DC
-
I
DR
I
th
o
r
ac
ic
C
T
d
ataset,
an
d
it su
cc
ess
f
u
lly
id
e
n
tifie
d
lu
n
g
n
o
d
u
les.
2
.
1
.
G
AN
-
ba
s
ed
da
t
a
a
ug
m
e
nta
t
io
n
T
o
ad
d
r
ess
d
ataset
s
ca
r
city
an
d
class
im
b
alan
ce
in
th
e
L
I
DC
-
I
DR
I
co
llectio
n
,
im
ag
e
au
g
m
en
tatio
n
was
p
er
f
o
r
m
ed
with
a
DC
GA
N
[
2
6
]
.
C
o
m
p
ar
e
d
to
o
th
er
g
en
er
ativ
e
m
o
d
els,
s
u
ch
as
c
o
n
d
itio
n
al
g
en
er
ati
v
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dee
p
lea
r
n
in
g
en
s
emb
les fo
r
lu
n
g
ca
n
ce
r
d
etec
tio
n
in
th
o
r
a
c
ic
C
T sca
n
s
…
(
B
in
ee
s
h
Mo
o
z
h
ip
p
u
r
a
th
)
1607
ad
v
er
s
ar
ial
n
etwo
r
k
s
(
cGA
Ns)
[
2
7
]
f
o
r
lab
el
co
n
d
itio
n
ed
s
y
n
th
esis
,
C
y
cleG
AN
s
[
2
8
]
f
o
r
d
o
m
ain
tr
an
s
latio
n
,
an
d
Sty
leGAN
s
[
2
9
]
f
o
r
h
ig
h
-
f
id
elity
g
en
er
ati
o
n
,
DC
GAN
p
r
esen
ted
a
g
o
o
d
t
r
ad
e
-
o
f
f
b
etwe
en
im
ag
e
q
u
ality
an
d
co
m
p
u
tatio
n
al
ef
f
icien
c
y
f
o
r
g
en
er
atin
g
s
y
n
th
etic
lu
n
g
n
o
d
u
les.
T
h
is
n
etwo
r
k
was
co
m
p
o
s
ed
o
f
f
o
u
r
co
n
v
o
l
u
tio
n
al
lay
e
r
s
with
l
ea
k
y
r
ec
tifie
d
lin
ea
r
u
n
it
(
R
eL
U)
,
f
o
u
r
tr
a
n
s
p
o
s
ed
c
o
n
v
o
lu
tio
n
a
l
lay
er
s
with
b
atch
n
o
r
m
aliza
tio
n
an
d
R
eL
U
ac
tiv
atio
n
in
th
e
g
en
e
r
ato
r
,
a
n
d
a
f
in
al
s
ig
m
o
id
ac
tiv
atio
n
in
th
e
d
is
cr
im
in
ato
r
.
E
v
er
y
lay
er
was
in
itialized
wi
th
Xav
ier
.
Usi
n
g
th
e
Ad
am
o
p
tim
izer
,
th
e
n
etwo
r
k
was
tr
ain
ed
f
o
r
2
0
0
ep
o
ch
s
with
a
b
atch
s
ize
o
f
6
4
,
β₁=0
.
5
,
β₂=0
.
9
9
9
,
a
n
d
a
lear
n
in
g
r
ate
o
f
0
.
0
0
0
2
.
Sp
ec
tr
al
n
o
r
m
a
lizatio
n
,
o
n
e
-
s
id
e
d
lab
el
s
m
o
o
th
in
g
,
alo
n
g
with
g
r
ad
ie
n
t
p
e
n
alties,
was
ad
o
p
ted
to
im
p
r
o
v
e
s
tab
ilit
y
an
d
p
r
ev
en
t
m
o
d
e
co
llap
s
in
g
.
Sy
n
t
h
etic
im
ag
es
r
ea
ch
ed
an
Fré
ch
et
in
ce
p
tio
n
d
is
tan
ce
(
FID
)
o
f
1
7
.
6
,
wh
ile
t
-
d
is
tr
ib
u
ted
s
to
ch
asti
c
n
eig
h
b
o
r
em
b
ed
d
i
n
g
(
t
-
SNE)
v
is
u
aliza
tio
n
s
co
n
f
ir
m
ed
s
ig
n
i
f
ican
t
o
v
er
lap
with
r
ea
l
im
ag
es
in
laten
t
s
p
ac
e,
in
d
icativ
e
o
f
r
ea
l
is
tic
v
ar
iab
ilit
y
.
T
h
ese
s
am
p
les
b
alan
ce
d
u
n
d
er
r
e
p
r
esen
ted
n
o
d
u
le
class
es
in
th
e
au
g
m
en
ted
s
et
D′,
in
cr
ea
s
in
g
t
h
e
r
o
b
u
s
tn
ess
an
d
g
en
e
r
aliza
tio
n
ab
ilit
y
o
f
th
e
d
o
wn
s
tr
ea
m
C
GA
B
en
s
em
b
le.
′
=
∪
{
(
,
)
}
=
1
(
1
)
W
h
er
e
D
is
th
e
o
r
ig
in
al
d
ataset,
d
en
o
tes
th
e
C
T
im
ag
es,
d
en
o
tes
th
e
co
r
r
esp
o
n
d
in
g
lab
els,
an
d
M
d
en
o
tes s
y
n
th
etic
s
am
p
les.
Fig
u
r
e
1
.
Ov
e
r
v
iew
o
f
th
e
C
GAB
-
b
ased
lu
n
g
ca
n
ce
r
d
etec
ti
o
n
p
ip
elin
e
2
.
2
.
CG
AB
ens
em
ble
T
h
is
s
tu
d
y
s
h
o
ws
C
GA
B
as
a
n
ew
m
eth
o
d
f
o
r
en
s
em
b
le
m
o
d
elin
g
.
I
n
C
GAB,
m
o
d
el
co
n
f
id
en
ce
an
d
in
ter
-
m
o
d
el
ag
r
e
em
en
t
ar
e
u
s
ed
f
o
r
p
r
ed
ictio
n
s
,
wh
ich
g
u
a
r
an
tees
r
o
b
u
s
tn
ess
an
d
p
r
ec
is
io
n
o
n
a
p
er
-
s
am
p
le
b
asis
.
T
h
is
m
eth
o
d
o
v
er
c
o
m
es
s
o
m
e
o
f
th
e
m
ajo
r
p
r
o
b
lem
s
in
e
n
s
em
b
le
lear
n
in
g
,
s
u
ch
as
co
n
f
lictin
g
p
r
ed
ictio
n
s
an
d
b
alan
cin
g
th
e
co
n
tr
ib
u
tio
n
o
f
ea
ch
m
o
d
el.
T
h
e
C
GAB f
r
am
ewo
r
k
is
h
av
i
n
g
th
e
f
o
llo
win
g
k
ey
co
m
p
o
n
e
n
ts
:
i)
Mo
d
el
-
s
p
ec
if
ic
c
o
n
f
id
e
n
ce
s
c
o
r
es:
ea
ch
b
ase
m
o
d
el,
d
en
o
te
d
as
(
)
wh
er
e
=
1
,
2
,
3
(
r
ep
r
esen
tin
g
R
esNet
-
1
5
2
,
Den
s
eNe
t
-
1
6
9
,
an
d
E
f
f
icien
tNet
-
B
7
)
,
o
u
tp
u
ts
a
clas
s
p
r
ed
ictio
n
̂
alo
n
g
with
a
co
n
f
id
en
ce
s
co
r
e.
T
h
e
co
n
f
i
d
en
ce
s
co
r
e,
(
)
,
is
d
er
iv
ed
f
r
o
m
th
e
s
o
f
tm
a
x
p
r
o
b
a
b
ilit
y
o
f
th
e
p
r
ed
icted
class
(
|
)
an
d
th
e
en
tr
o
p
y
o
f
th
e
o
u
tp
u
t d
is
tr
ib
u
tio
n
.
(
)
=
(
∣
)
−
∑
(
∣
)
(
∣
)
(
2
)
W
h
er
e
=
(
|
)
an
d
iter
ates o
v
er
all
p
o
s
s
ib
le
class
e
s
.
ii)
C
o
n
s
en
s
u
s
m
ea
s
u
r
em
en
t:
th
e
lev
el
o
f
ag
r
ee
m
en
t
b
etwe
e
n
th
e
b
ase
m
o
d
els
is
q
u
an
tifie
d
u
s
in
g
a
co
n
s
en
s
u
s
s
co
r
e,
(
)
,
wh
ich
m
ea
s
u
r
es h
o
w
m
an
y
m
o
d
els p
r
ed
ict
th
e
s
am
e
class
as in
(
3
)
.
(
)
=
ℎ
(
3
)
A
h
ig
h
er
c
o
n
s
en
s
u
s
in
d
icate
s
s
tr
o
n
g
er
ag
r
ee
m
e
n
t a
m
o
n
g
th
e
m
o
d
els.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
2
,
Ap
r
il 2
0
2
6
:
1
6
0
5
-
1
6
1
2
1608
iii)
Ad
ap
tiv
e
b
len
d
in
g
:
th
e
p
r
ed
i
ctio
n
s
f
r
o
m
th
e
b
ase
m
o
d
els
ar
e
co
m
b
in
ed
ad
a
p
tiv
ely
,
with
ea
ch
m
o
d
el
weig
h
ted
b
y
its
co
n
f
i
d
en
ce
s
c
o
r
e
an
d
th
e
c
o
n
s
en
s
u
s
s
co
r
e.
F
o
r
a
g
iv
en
s
am
p
le
,
th
e
weig
h
t
o
f
m
o
d
el
m
,
(
)
,
is
co
m
p
u
ted
.
(
)
=
(
)
⋅
(
)
∑
(
)
⋅
(
)
3
=
1
(
4
)
T
h
is
en
s
u
r
es
th
at
m
o
d
els
with
h
ig
h
er
co
n
f
id
en
ce
an
d
s
tr
o
n
g
er
ag
r
ee
m
e
n
t
c
o
n
tr
ib
u
te
m
o
r
e
to
th
e
f
in
a
l
p
r
ed
ictio
n
.
iv
)
Div
er
s
ity
r
eg
u
lar
izatio
n
:
to
av
o
id
o
v
e
r
-
r
elian
ce
o
n
a
n
y
s
in
g
le
m
o
d
el,
a
d
iv
er
s
ity
r
eg
u
lar
izatio
n
ter
m
is
in
tr
o
d
u
ce
d
d
u
r
in
g
tr
ain
in
g
to
m
ax
im
ize
v
ar
iab
ilit
y
am
o
n
g
th
e
m
o
d
els’
p
r
ed
ictio
n
s
.
T
h
is
e
n
s
u
r
es
th
at
th
e
en
s
em
b
le
lev
er
ag
es th
e
c
o
m
p
l
em
en
tar
y
s
tr
en
g
t
h
s
o
f
all
b
ase
m
o
d
els.
v)
Fin
al
p
r
ed
ictio
n
:
th
e
en
s
em
b
l
e
p
r
ed
ictio
n
̂
is
ca
lcu
lated
as
a
weig
h
ted
co
m
b
in
atio
n
o
f
t
h
e
p
r
ed
ictio
n
s
f
r
o
m
th
e
b
ase
m
o
d
els
.
̂
=
∑
(
)
⋅
̂
3
=
1
(
5
)
W
h
er
e
(
)
is
th
e
weig
h
t a
s
s
ig
n
ed
to
th
e
m
th
m
o
d
el.
T
h
e
C
GAB
f
r
am
ewo
r
k
'
s
tr
ain
in
g
p
r
o
ce
d
u
r
e
was
d
esig
n
e
d
t
o
g
u
a
r
an
tee
d
ep
en
d
ab
le
a
n
d
s
tr
o
n
g
p
er
f
o
r
m
an
ce
.
T
h
e
m
er
g
e
d
d
ataset
was
s
ep
ar
ated
in
to
tr
ain
in
g
,
v
alid
atio
n
,
an
d
test
in
g
s
u
b
s
ets
wh
ile
p
r
eser
v
in
g
class
b
alan
ce
af
ter
th
e
GAN
-
b
ased
au
g
m
en
t
atio
n
p
r
ev
i
o
u
s
ly
d
is
cu
s
s
ed
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
3
.
1
.
Co
m
pu
t
a
t
io
na
l c
o
s
t
a
nd
infe
re
nce
t
im
e
E
f
f
icien
cy
was
m
ea
s
u
r
ed
wit
h
a
T
esla
T
4
GPU
(
1
5
GB
)
.
T
h
e
C
GAB
ap
p
r
o
ac
h
p
r
o
ce
s
s
ed
ea
ch
C
T
s
ca
n
in
0
.
8
3
s
ec
o
n
d
s
,
u
s
in
g
2
.
3
GB
o
f
m
em
o
r
y
,
s
ig
n
if
ic
an
tly
o
u
tp
e
r
f
o
r
m
in
g
tr
a
n
s
f
o
r
m
er
-
b
ased
m
eth
o
d
s
,
wh
ich
tak
e
o
v
er
1
.
5
s
ec
o
n
d
s
with
m
o
r
e
th
an
4
GB
o
f
m
em
o
r
y
.
T
h
is
ef
f
icien
cy
co
u
p
led
with
ac
cu
r
ac
y
s
u
g
g
ests
g
r
ea
t su
itab
ilit
y
f
o
r
r
ea
l
-
tim
e
clin
ical
ap
p
licatio
n
s
.
3
.
2
.
Co
m
pa
riso
n wit
h sta
t
e
-
of
-
t
he
-
a
r
t
mo
dels
Fo
r
b
en
c
h
m
ar
k
i
n
g
t
h
e
C
GAB
f
r
am
ewo
r
k
,
v
a
r
io
u
s
C
NN
ar
ch
itectu
r
es,
i
n
clu
d
in
g
Den
s
eNe
t
-
1
2
1
,
Go
o
g
L
eNe
t,
E
f
f
icien
tNet
-
B
7
,
Alex
Net,
an
d
R
esNet
-
1
5
2
,
with
t
r
an
s
f
o
r
m
er
-
b
ased
m
o
d
els
lik
e
v
is
io
n
tr
an
s
f
o
r
m
er
(
ViT
)
-
B
/1
6
an
d
s
win
tr
an
s
f
o
r
m
er
,
wer
e
ev
alu
ated
in
co
m
p
ar
is
o
n
to
ad
v
an
ce
d
s
elf
-
s
u
p
er
v
is
ed
(
Mo
C
o
-
v
3
R
esNet
-
1
5
2
)
an
d
h
y
b
r
id
C
NN
-
R
NN
ap
p
r
o
ac
h
es.
I
n
s
u
m
m
ar
y
,
as p
r
o
v
i
d
ed
b
y
T
ab
les 1
an
d
2
,
ViT
,
an
d
s
win
r
ea
c
h
ed
c
o
m
p
ar
a
b
l
e
r
ec
all
o
f
9
8
.
1
2
%
a
n
d
9
8
.
0
5
%,
r
esp
ec
tiv
ely
,
w
h
ile
C
GAB
o
u
tp
er
f
o
r
m
ed
all
b
en
ch
m
ar
k
s
,
with
9
7
.
3
5
%
ac
cu
r
ac
y
,
9
8
.
4
6
%
F1
-
s
co
r
e
,
a
n
d
0
.
9
8
5
r
ec
eiv
e
r
o
p
er
atin
g
ch
ar
ac
ter
is
tic
(
R
OC
)
-
ar
ea
u
n
d
e
r
th
e
cu
r
v
e
(
AUC)
.
Me
an
wh
ile,
C
GAB
ex
h
ib
ited
h
ig
h
er
co
m
p
u
tatio
n
al
ef
f
icie
n
cy
,
r
eq
u
ir
in
g
o
n
ly
0
.
8
3
s
p
er
s
ca
n
a
n
d
2
.
3
GB
p
ea
k
GPU
m
em
o
r
y
,
s
u
b
s
tan
tially
lo
wer
th
an
th
at
o
f
m
o
r
e
th
a
n
1
.
5
s
an
d
g
r
ea
te
r
th
an
4
GB
o
f
t
r
an
s
f
o
r
m
er
-
b
ased
m
o
d
els.
Ov
er
all,
C
GA
B
r
ea
ch
ed
th
e
b
est
b
alan
ce
b
etwe
en
p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
an
d
ef
f
icien
cy
,
f
u
r
th
er
u
n
d
er
lin
in
g
its
p
o
ten
tial
as
a
clin
ically
d
ep
lo
y
ab
le
s
y
s
tem
.
Alth
o
u
g
h
cu
r
r
en
t
v
alid
atio
n
is
lim
ited
to
L
I
DC
-
I
DR
I
an
d
p
a
r
tially
n
ati
o
n
al
lu
n
g
s
cr
ee
n
in
g
tr
ial
(
NL
S
T
)
d
atasets
,
s
tr
o
n
g
g
en
er
aliza
b
ilit
y
o
f
C
GAB
to
o
th
er
C
T
d
atasets
-
f
o
r
ex
am
p
l
e,
lu
n
g
n
o
d
u
le
a
n
aly
s
is
2
0
1
6
(
L
UNA1
6
)
an
d
t
h
e
ca
n
ce
r
im
ag
i
n
g
a
r
ch
iv
e
(
T
C
I
A
)
,
ca
n
b
e
i
n
f
er
r
e
d
f
r
o
m
th
e
ad
ap
tiv
e
weig
h
tin
g
m
ec
h
an
is
m
o
f
C
GAB,
wh
ich
wo
u
ld
war
r
an
t
f
u
r
th
er
in
v
esti
g
atio
n
in
m
u
lti
-
ce
n
ter
s
tu
d
ies.
Fig
u
r
es
2
an
d
3
d
ep
ict
co
m
p
ar
ativ
e
p
er
f
o
r
m
an
ce
tr
en
d
s
,
with
Fig
u
r
e
2
s
h
o
win
g
C
GAB’s
s
u
p
er
io
r
ac
cu
r
ac
y
a
n
d
p
r
ec
is
io
n
,
a
n
d
Fig
u
r
e
3
co
n
f
ir
m
in
g
its
h
ig
h
e
r
r
ec
all
an
d
r
e
d
u
ce
d
f
alse n
eg
ati
v
es.
T
ab
le
1
.
B
en
ch
m
a
r
k
in
g
o
f
C
G
AB
ag
ain
s
t state
-
of
-
th
e
-
ar
t m
o
d
els
M
o
d
e
l
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
sc
o
r
e
(
%)
R
O
C
-
AUC
S
p
e
c
i
f
i
c
i
t
y
(
%)
S
e
n
s
i
t
i
v
i
t
y
(
%)
I
n
f
e
r
e
n
c
e
t
i
me
(
s/
sca
n
)
G
P
U
memo
r
y
(
G
B
)
D
e
n
seN
e
t
-
1
2
1
9
6
.
2
3
9
6
.
6
3
9
6
.
2
4
9
6
.
4
3
0
.
9
7
1
9
5
.
9
8
9
6
.
2
4
0
.
6
2
1
.
8
R
e
sN
e
t
-
152
9
7
.
1
9
9
8
.
3
1
9
7
.
2
9
9
7
.
7
9
0
.
9
7
8
9
6
.
9
5
9
7
.
2
9
0
.
7
2
2
.
0
Ef
f
i
c
i
e
n
t
N
e
t
-
B7
9
6
.
9
5
9
8
.
0
5
9
6
.
8
1
9
7
.
4
2
0
.
9
7
6
9
6
.
7
9
6
.
8
1
0
.
8
8
2
.
4
G
o
o
g
L
e
N
e
t
9
6
.
5
1
9
7
.
1
2
9
6
.
3
9
6
.
7
0
.
9
7
9
5
.
9
9
6
.
3
0
.
5
5
1
.
6
A
l
e
x
N
e
t
9
4
.
8
7
9
5
.
2
5
9
4
.
5
6
9
4
.
9
0
.
9
6
1
9
4
.
1
2
9
4
.
5
6
0
.
4
1
.
3
V
i
T
-
B
/
1
6
9
7
.
0
5
9
8
.
1
9
8
.
1
2
9
8
.
1
1
0
.
9
8
1
9
6
.
8
3
9
8
.
1
2
1
.
5
2
4
.
5
S
w
i
n
t
r
a
n
sf
o
r
mer
9
7
.
1
2
9
8
.
2
9
8
.
0
5
9
8
.
1
2
0
.
9
8
2
9
6
.
9
5
9
8
.
0
5
1
.
4
7
4
.
2
C
N
N
–
R
N
N
h
y
b
r
i
d
9
6
.
8
9
9
7
.
8
8
9
6
.
7
5
9
7
.
3
1
0
.
9
7
4
9
6
.
2
9
6
.
7
5
1
.
1
3
.
1
M
o
C
o
-
v
3
+
R
e
sN
e
t
-
152
9
7
.
0
2
9
8
.
1
5
9
7
.
8
5
98
0
.
9
8
9
6
.
7
7
9
7
.
8
5
1
.
0
8
3
.
0
C
G
A
B
(
p
r
o
p
o
s
e
d
)
9
7
.
3
5
9
8
.
7
3
9
8
.
2
9
8
.
4
6
0
.
9
8
5
9
7
.
1
9
8
.
2
0
.
8
3
2
.
3
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dee
p
lea
r
n
in
g
en
s
emb
les fo
r
lu
n
g
ca
n
ce
r
d
etec
tio
n
in
th
o
r
a
c
ic
C
T sca
n
s
…
(
B
in
ee
s
h
Mo
o
z
h
ip
p
u
r
a
th
)
1609
T
ab
le
2
.
C
o
m
p
a
r
is
o
n
o
f
en
s
e
m
b
le
f
r
am
ewo
r
k
s
f
o
r
lu
n
g
ca
n
ce
r
d
etec
tio
n
S
t
u
d
y
D
a
t
a
s
e
t
A
r
c
h
i
t
e
c
t
u
r
e
/
G
A
N
v
a
r
i
a
n
t
K
e
y
m
e
t
r
i
c
s/
c
o
n
t
r
i
b
u
t
i
o
n
En
se
mb
l
e
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
r
i
s
k
st
r
a
t
i
f
i
c
a
t
i
o
n
o
f
i
n
v
a
si
v
e
l
u
n
g
a
d
e
n
o
c
a
r
c
i
n
o
ma
[
3
0
]
.
Th
i
n
-
sl
i
c
e
C
T
(
mu
l
t
i
c
e
n
t
e
r
)
M
u
l
t
i
-
v
i
e
w
3
D
C
N
N
e
n
s
e
mb
l
e
A
c
c
u
r
a
c
y
=
9
5
.
2
%
,
A
U
C
=
0
.
9
7
2
D
e
e
p
l
e
a
r
n
i
n
g
e
n
sem
b
l
e
2
D
C
N
N
f
o
r
l
u
n
g
c
a
n
c
e
r
d
e
t
e
c
t
i
o
n
[
3
1
]
.
LU
N
A
1
6
En
se
mb
l
e
o
f
2
D
C
N
N
s
A
c
c
u
r
a
c
y
=
9
4
.
8
%,
sen
s
i
t
i
v
i
t
y
=
9
3
.
6
%
,
sp
e
c
i
f
i
c
i
t
y
=
9
5
.
1
%
En
se
mb
l
e
o
f
C
N
N
mo
d
e
l
s
f
o
r
l
u
n
g
n
o
d
u
l
e
d
e
t
e
c
t
i
o
n
[
3
2
]
.
LI
D
C
-
I
D
R
I
H
y
b
r
i
d
C
N
N
e
n
s
e
mb
l
e
A
c
c
u
r
a
c
y
=
9
6
.
1
%,
A
U
C
=
0
.
9
7
5
C
T
i
ma
g
e
s
G
A
N
-
b
a
s
e
d
a
u
g
m
e
n
t
a
t
i
o
n
w
i
t
h
A
d
a
I
N
f
o
r
l
u
n
g
n
o
d
u
l
e
s
d
e
t
e
c
t
i
o
n
[
3
3
]
.
LU
N
A
1
6
(
p
u
l
mo
n
a
r
y
n
o
d
u
l
e
s)
c
G
A
N
+
A
d
a
I
N
P
u
l
m
o
n
a
r
y
n
o
d
u
l
e
s
d
e
t
e
c
t
i
o
n
i
s
i
mp
r
o
v
e
d
;
t
h
e
a
u
g
me
n
t
a
t
i
o
n
e
f
f
e
c
t
i
s
v
a
l
i
d
a
t
e
d
q
u
a
n
t
i
t
a
t
i
v
e
l
y
Lu
n
g
c
a
n
c
e
r
C
T
i
m
a
g
e
g
e
n
e
r
a
t
i
o
n
f
r
o
m
a
f
r
e
e
-
f
o
r
m
s
k
e
t
c
h
u
si
n
g
st
y
l
e
-
b
a
se
d
p
i
x
2
p
i
x
f
o
r
d
a
t
a
a
u
g
me
n
t
a
t
i
o
n
[
3
4
]
.
P
r
i
v
a
t
e
C
T
+
l
e
si
o
n
sk
e
t
c
h
e
s
S
t
y
l
e
P
i
x
2
p
i
x
(
p
i
x
2
p
i
x
+
S
t
y
l
e
G
A
N
st
y
l
e
b
l
o
c
k
s)
I
mp
r
o
v
e
d
d
i
v
e
r
si
t
y
o
f
l
e
si
o
n
sh
a
p
e
s
;
e
f
f
e
c
t
i
v
e
d
a
t
a
a
u
g
me
n
t
a
t
i
o
n
d
e
mo
n
s
t
r
a
t
e
d
C
G
A
B
(
t
h
i
s wo
r
k
)
.
LI
D
C
-
I
D
R
I
+
G
A
N
a
u
g
me
n
t
a
t
i
o
n
(
p
r
o
p
o
s
e
d
)
R
e
sN
e
t
-
1
5
2
+
D
e
n
s
e
N
e
t
-
1
6
9
+
Ef
f
i
c
i
e
n
t
N
e
t
-
B
7
+
C
G
A
B
b
l
e
n
d
i
n
g
B
e
t
t
e
r
g
e
n
e
r
a
l
i
z
a
t
i
o
n
a
n
d
r
o
b
u
st
n
e
s
s
a
r
e
s
h
o
w
n
t
h
r
o
u
g
h
h
i
g
h
e
r
me
t
r
i
c
s.
A
c
c
u
r
a
c
y
9
7
.
3
5
%
,
p
r
e
c
i
si
o
n
9
8
.
7
3
%
,
r
e
c
a
l
l
9
8
.
2
0
%
Fig
u
r
e
2
.
C
o
m
p
a
r
is
o
n
o
f
ac
cu
r
ac
y
an
d
p
r
ec
is
io
n
ac
r
o
s
s
m
o
d
e
ls
Fig
u
r
e
3
.
C
o
m
p
a
r
is
o
n
o
f
r
ec
al
l a
cr
o
s
s
m
o
d
els
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
2
,
Ap
r
il 2
0
2
6
:
1
6
0
5
-
1
6
1
2
1610
3
.
3
.
Abla
t
io
n study
:
im
pa
c
t
o
f
G
AN
a
ug
m
ent
a
t
io
n a
nd
s
t
a
t
is
t
ica
l sig
nifica
nce
T
o
s
ee
h
o
w
GAN
-
b
ased
au
g
m
en
tatio
n
s
h
if
ted
p
er
f
o
r
m
a
n
c
e,
we
co
m
p
ar
ed
t
h
e
C
GAB
en
s
em
b
le
th
at
was
tr
ain
ed
o
n
ly
o
n
th
e
o
r
ig
i
n
al
L
I
DC
-
I
DR
I
d
ata
to
th
e
s
a
m
e
s
etu
p
,
b
u
t
th
is
tim
e
tr
ain
ed
o
n
an
au
g
m
en
ted
d
ataset
D
'
,
in
clu
d
in
g
th
e
C
T
s
ca
n
s
g
en
er
ated
b
y
DC
GA
N.
Fro
m
T
ab
le
3
,
we
o
b
s
er
v
e
th
at
au
g
m
en
tatio
n
b
o
o
s
ted
all
th
e
p
er
f
o
r
m
an
ce
m
etr
ics:
ac
cu
r
ac
y
in
cr
ea
s
ed
f
r
o
m
9
5
.
9
1
%
to
9
7
.
3
5
%,
R
OC
-
AUC
in
cr
ea
s
ed
f
r
o
m
0
.
9
7
1
to
0
.
9
8
3
,
an
d
F1
-
s
co
r
e
f
r
o
m
9
6
.
5
2
%
to
9
8
.
4
6
%.
Fiv
e
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
co
m
b
in
ed
with
t
-
test
s
co
n
f
ir
m
e
d
th
at
g
ain
s
ar
e
s
tatis
tically
s
ig
n
if
ican
t
(
p
<0
.
0
1
)
an
d
ar
e
n
o
t
lik
el
y
d
u
e
to
r
an
d
o
m
f
l
u
ctu
atio
n
.
Mo
r
eo
v
er
,
au
g
m
en
tatio
n
im
p
r
o
v
ed
s
en
s
itiv
ity
to
m
in
o
r
ity
le
s
io
n
ty
p
es
an
d
h
elp
ed
b
alan
ce
class
es,
im
p
r
o
v
in
g
g
en
er
aliza
tio
n
.
Fu
r
th
e
r
,
t
-
SNE
v
is
u
aliza
tio
n
s
in
d
icate
d
th
at
s
y
n
th
etic
s
am
p
les
ex
p
an
d
ed
th
e
laten
t
f
ea
tu
r
e
s
p
ac
e
with
o
u
t
lo
s
in
g
alig
n
m
e
n
t
with
r
ea
l
d
ata.
I
n
a
n
u
ts
h
el
l,
GAN
au
g
m
en
tatio
n
s
ig
n
if
ic
an
tly
s
tr
en
g
th
e
n
ed
th
e
r
o
b
u
s
tn
ess
an
d
r
eliab
ilit
y
o
f
th
e
C
GAB en
s
em
b
le
f
r
o
m
b
o
th
q
u
a
n
titativ
e
an
d
s
tatis
tical
s
tan
d
p
o
in
ts
.
T
ab
le
3
.
Ab
latio
n
s
tu
d
y
r
esu
lt
s
M
o
d
e
l
s
e
t
u
p
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
sc
o
r
e
(
%)
R
O
C
-
AUC
p
-
v
a
l
u
e
(
v
s.
w
i
t
h
o
u
t
G
A
N
)
C
G
A
B
(
w
i
t
h
o
u
t
GAN
a
u
g
me
n
t
a
t
i
o
n
)
9
5
.
9
1
9
7
.
0
4
9
6
.
2
2
9
6
.
5
2
0
.
9
7
1
–
C
G
A
B
(
w
i
t
h
GAN
a
u
g
me
n
t
a
t
i
o
n
)
9
7
.
3
5
9
8
.
7
3
9
8
.
2
9
8
.
4
6
0
.
9
8
3
<
0
.
0
1
3
.
4
.
T
heo
re
t
ica
l
c
o
ntr
ibu
t
io
n
T
h
e
C
GAB
s
y
s
tem
g
o
es
b
ey
o
n
d
s
tatic
en
s
em
b
le
lea
r
n
in
g
liter
atu
r
e
th
at
u
s
es
weig
h
ted
v
o
tin
g
o
r
m
ajo
r
ity
r
u
les.
W
ith
th
e
ca
lib
r
atio
n
o
f
c
o
n
f
id
e
n
ce
,
th
e
f
r
a
m
ewo
r
k
in
teg
r
ates
th
e
en
tr
o
p
y
-
ad
ju
s
ted
S
o
f
t
M
ax
s
co
r
es,
co
n
s
en
s
u
s
m
o
d
elin
g
t
h
at
q
u
an
tifie
s
in
ter
-
m
o
d
el
ag
r
ee
m
en
t
u
s
in
g
a
f
o
r
m
al
c
o
n
s
en
s
u
s
s
co
r
e,
an
d
th
en
th
e
r
eso
lu
tio
n
o
f
co
n
f
lictin
g
ac
tio
n
s
th
r
o
u
g
h
an
a
u
x
iliar
y
d
ec
is
io
n
m
o
d
el
wh
en
a
co
n
s
e
n
s
u
s
is
wea
k
.
T
h
is
d
esig
n
f
alls
with
in
B
ay
esian
e
n
s
em
b
le
p
r
in
ci
p
les,
b
u
t
with
an
en
h
a
n
ce
m
en
t
o
f
ad
ap
tab
ilit
y
o
n
a
p
er
s
am
p
le.
Un
lik
e
class
ical
en
s
em
b
les
th
at
o
p
tim
ize
a
g
lo
b
al
weig
h
t,
C
GA
B
em
p
lo
y
s
a
co
n
tex
tu
all
y
d
y
n
a
m
ic
ad
h
esio
n
b
len
d
in
g
a
p
p
r
o
ac
h
.
T
h
is
p
o
s
it
io
n
s
C
GA
B
as
a
s
y
s
tem
f
o
r
lu
n
g
ca
n
ce
r
d
etec
tio
n
,
b
u
t
m
o
r
e
im
p
o
r
tan
tly
,
as
an
en
s
em
b
le
ar
ch
itectu
r
e
th
at
c
an
b
e
ap
p
lied
t
o
o
th
er
d
o
m
ain
s
s
u
ch
as
p
ath
o
lo
g
y
,
r
em
o
te
s
en
s
in
g
,
an
d
au
to
n
o
m
o
u
s
d
r
iv
i
n
g
.
3
.
5
.
E
x
pla
ina
bil
it
y
a
nd
i
nte
rpre
t
a
bil
it
y
T
h
e
ac
ce
p
tan
ce
o
f
a
n
AI
-
ass
is
ted
d
iag
n
o
s
is
is
m
ea
s
u
r
ed
th
r
o
u
g
h
in
ter
p
r
etab
ilit
y
,
wh
ich
i
s
ess
en
tial
f
o
r
clin
ical
ju
d
g
m
en
ts
.
T
h
e
s
tu
d
y
,
in
v
o
l
v
in
g
Gr
ad
-
C
AM
,
h
el
p
ed
to
f
ea
tu
r
e
k
ey
ar
ea
s
in
th
o
r
ac
ic
C
T
s
ca
n
s
th
at
in
f
lu
en
ce
d
m
o
d
el
p
r
e
d
ictio
n
s
.
Gen
er
atio
n
o
f
s
tab
le
an
d
clin
i
ca
lly
p
r
o
v
e
n
atten
tio
n
p
atter
n
s
is
th
e
r
esu
lt
o
f
th
e
C
GA
B
f
r
am
ewo
r
k
,
w
h
ich
will im
p
r
o
v
e
th
e
u
n
d
er
s
tan
d
in
g
o
f
r
ad
io
lo
g
is
ts
.
4.
CO
NCLU
SI
O
N
T
h
e
C
GAB
en
s
em
b
le
f
r
am
ewo
r
k
,
e
n
h
an
ce
d
b
y
GAN
-
b
ased
s
y
n
th
etic
au
g
m
e
n
tatio
n
,
is
th
e
r
esu
lt
o
f
th
is
s
tu
d
y
f
o
r
th
e
i
d
en
tific
a
tio
n
o
f
lu
n
g
ca
n
ce
r
in
th
o
r
ac
ic
C
T
s
ca
n
s
.
T
h
is
m
o
d
el
attain
ed
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
m
ea
s
u
r
es
b
y
u
s
in
g
DC
GAN
-
g
en
er
ated
C
T
im
ag
es,
wh
ich
ad
d
r
ess
ed
th
e
p
r
o
b
lem
s
o
f
d
ata
s
ca
r
city
an
d
m
o
d
el
o
v
er
f
itti
n
g
an
d
p
r
ed
ictio
n
s
f
r
o
m
R
esNet
-
1
5
2
,
Den
s
eNe
t
-
1
6
9
,
an
d
E
f
f
icien
tNet
-
B
7
wer
e
b
len
d
ed
o
n
th
e
L
I
DC
-
I
DR
I
d
ataset.
Ab
latio
n
s
tu
d
ies
s
h
o
w
co
n
s
is
ten
t
g
ain
s
o
v
er
s
in
g
le
C
NN
ar
ch
itectu
r
es
an
d
tr
a
d
itio
n
al
weig
h
te
d
en
s
e
m
b
les.
B
ias
m
ay
b
e
in
tr
o
d
u
ce
d
b
ec
a
u
s
e
o
f
GAN
g
en
er
ated
im
ag
es.
Stab
ilit
y
in
th
e
tr
ain
in
g
p
r
o
ce
s
s
is
a
ch
all
en
g
e.
E
x
p
a
n
s
io
n
o
f
test
in
g
ac
r
o
s
s
a
b
r
o
ad
er
p
o
p
u
latio
n
is
r
eq
u
ir
ed
f
o
r
ex
ter
n
a
l
v
alid
atio
n
,
o
th
er
th
an
NL
ST
d
ataset.
Fu
tu
r
e
e
n
h
an
ce
m
en
ts
ca
n
b
e
f
o
cu
s
ed
o
n
d
ata
d
iv
e
r
s
ity
u
s
in
g
m
o
d
els
s
u
ch
as
Sty
leGAN
,
d
if
f
u
s
io
n
m
o
d
els,
c
GANs
an
d
in
te
g
r
atin
g
t
r
a
n
s
f
o
r
m
er
-
b
ased
h
y
b
r
id
s
.
Dev
elo
p
in
g
ex
p
lain
ab
le
en
s
em
b
le
s
tr
ateg
i
es
will
h
elp
f
o
r
im
p
r
o
v
in
g
clin
ical
tr
u
s
t
an
d
th
e
d
e
p
lo
y
m
e
n
t o
f
r
ea
l
-
tim
e
clin
ical
p
ip
elin
es.
T
h
ese
im
p
r
o
v
em
e
n
ts
o
n
th
e
f
r
am
ewo
r
k
will
m
ak
e
C
GAB
a
cr
o
s
s
-
d
o
m
ain
to
o
l
ac
r
o
s
s
m
ed
ical
im
ag
in
g
an
d
b
e
y
o
n
d
.
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
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dee
p
lea
r
n
in
g
en
s
emb
les fo
r
lu
n
g
ca
n
ce
r
d
etec
tio
n
in
th
o
r
a
c
ic
C
T sca
n
s
…
(
B
in
ee
s
h
Mo
o
z
h
ip
p
u
r
a
th
)
1611
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
B
in
ee
s
h
Mo
o
zh
ip
p
u
r
ath
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
J
ay
ap
an
d
ian
Nata
r
ajan
✓
✓
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
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.
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
in
g
s
o
f
th
is
s
tu
d
y
ar
e
av
ailab
l
e
f
r
o
m
co
r
r
esp
o
n
d
in
g
au
th
o
r
,
[
J
N]
,
u
p
o
n
r
ea
s
o
n
ab
le
r
eq
u
est
.
RE
F
E
R
E
NC
E
S
[
1
]
M
.
D
a
v
e
a
n
d
N
.
P
a
t
e
l
,
“
A
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
i
n
h
e
a
l
t
h
c
a
r
e
a
n
d
e
d
u
c
a
t
i
o
n
,
”
Bri
t
i
s
h
D
e
n
t
a
l
J
o
u
r
n
a
l
,
v
o
l
.
2
3
4
,
n
o
.
1
0
,
p
p
.
7
6
1
–
7
6
4
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
3
8
/
s4
1
4
1
5
-
0
2
3
-
5
8
4
5
-
2.
[
2
]
Y
.
A
.
F
a
h
i
m,
I
.
W
.
H
a
sa
n
i
,
S
.
K
a
b
b
a
,
a
n
d
W
.
M
.
R
a
g
a
b
,
“
A
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
i
n
h
e
a
l
t
h
c
a
r
e
a
n
d
m
e
d
i
c
i
n
e
:
c
l
i
n
i
c
a
l
a
p
p
l
i
c
a
t
i
o
n
s
,
t
h
e
r
a
p
e
u
t
i
c
a
d
v
a
n
c
e
s,
a
n
d
f
u
t
u
r
e
p
e
r
s
p
e
c
t
i
v
e
s
,
”
E
u
ro
p
e
a
n
J
o
u
r
n
a
l
o
f
Me
d
i
c
a
l
Re
s
e
a
r
c
h
,
v
o
l
.
30
,
p
p
.
1
–
21
,
2
0
2
5
,
d
o
i
:
1
0
.
1
1
8
6
/
s
4
0
0
0
1
-
025
-
0
3
1
9
6
-
w
.
[
3
]
H
.
Y
u
,
L
.
T.
Y
a
n
g
,
Q
.
Zh
a
n
g
,
D
.
A
r
mstro
n
g
,
a
n
d
M
.
J.
D
e
e
n
,
“
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
f
o
r
me
d
i
c
a
l
i
m
a
g
e
a
n
a
l
y
si
s
:
st
a
t
e
-
of
-
t
h
e
-
a
r
t
,
c
o
mp
a
r
i
s
o
n
s
,
i
m
p
r
o
v
e
me
n
t
a
n
d
p
e
r
sp
e
c
t
i
v
e
s,
”
N
e
u
ro
c
o
m
p
u
t
i
n
g
,
v
o
l
.
4
4
4
,
p
p
.
9
2
–
1
1
0
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
n
e
u
c
o
m.
2
0
2
0
.
0
4
.
1
5
7
.
[
4
]
A
.
S
.
P
a
n
a
y
i
d
e
s
e
t
a
l
.
,
“
A
I
i
n
me
d
i
c
a
l
i
ma
g
i
n
g
i
n
f
o
r
ma
t
i
c
s:
c
u
r
r
e
n
t
c
h
a
l
l
e
n
g
e
s
a
n
d
f
u
t
u
r
e
d
i
r
e
c
t
i
o
n
s,
”
I
EE
E
J
o
u
r
n
a
l
o
f
B
i
o
m
e
d
i
c
a
l
a
n
d
H
e
a
l
t
h
I
n
f
o
rm
a
t
i
c
s
,
v
o
l
.
2
4
,
n
o
.
7
,
p
p
.
1
8
3
7
–
1
8
5
7
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/
J
B
H
I
.
2
0
2
0
.
2
9
9
1
0
4
3
.
[
5
]
S
.
H
u
a
n
g
,
J.
Y
a
n
g
,
N
.
S
h
e
n
,
Q
.
X
u
,
a
n
d
Q
.
Z
h
a
o
,
“
A
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
i
n
l
u
n
g
c
a
n
c
e
r
d
i
a
g
n
o
s
i
s
a
n
d
p
r
o
g
n
o
si
s
:
c
u
r
r
e
n
t
a
p
p
l
i
c
a
t
i
o
n
a
n
d
f
u
t
u
r
e
p
e
r
s
p
e
c
t
i
v
e
,
”
S
e
m
i
n
a
rs
i
n
C
a
n
c
e
r B
i
o
l
o
g
y
,
v
o
l
.
8
9
,
p
p
.
3
0
–
3
7
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
sem
c
a
n
c
e
r
.
2
0
2
3
.
0
1
.
0
0
6
.
[
6
]
G
.
Zh
a
n
g
e
t
a
l
.
,
“
A
u
t
o
m
a
t
i
c
n
o
d
u
l
e
d
e
t
e
c
t
i
o
n
f
o
r
l
u
n
g
c
a
n
c
e
r
i
n
C
T
i
ma
g
e
s
:
a
r
e
v
i
e
w
,
”
C
o
m
p
u
t
e
rs
i
n
Bi
o
l
o
g
y
a
n
d
M
e
d
i
c
i
n
e
,
v
o
l
.
1
0
3
,
p
p
.
2
8
7
–
3
0
0
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
b
i
o
m
e
d
.
2
0
1
8
.
1
0
.
0
3
3
.
[
7
]
S
.
F
.
A
h
m
e
d
e
t
a
l
.
,
“
D
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
l
i
n
g
t
e
c
h
n
i
q
u
e
s
:
c
u
r
r
e
n
t
p
r
o
g
r
e
ss,
a
p
p
l
i
c
a
t
i
o
n
s,
a
d
v
a
n
t
a
g
e
s
,
a
n
d
c
h
a
l
l
e
n
g
e
s,”
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
Re
v
i
e
w
,
v
o
l
.
5
6
,
n
o
.
1
1
,
p
p
.
1
3
5
2
1
–
1
3
6
1
7
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
4
6
2
-
0
2
3
-
1
0
4
6
6
-
8.
[
8
]
I
.
G
o
o
d
f
e
l
l
o
w
e
t
a
l
.
,
“
G
e
n
e
r
a
t
i
v
e
a
d
v
e
r
sari
a
l
n
e
t
w
o
r
k
s,”
C
o
m
m
u
n
i
c
a
t
i
o
n
s
o
f
t
h
e
AC
M
,
v
o
l
.
6
3
,
n
o
.
1
1
,
p
p
.
1
3
9
–
1
4
4
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
4
5
/
3
4
2
2
6
2
2
.
[
9
]
V
.
S
o
r
i
n
,
Y
.
B
a
r
a
s
h
,
E
.
K
o
n
e
n
,
a
n
d
E.
K
l
a
n
g
,
“
C
r
e
a
t
i
n
g
a
r
t
i
f
i
c
i
a
l
i
m
a
g
e
s
f
o
r
r
a
d
i
o
l
o
g
y
a
p
p
l
i
c
a
t
i
o
n
s
u
s
i
n
g
g
e
n
e
r
a
t
i
v
e
a
d
v
e
r
sar
i
a
l
n
e
t
w
o
r
k
s
(
G
A
N
s)
–
a
s
y
s
t
e
m
a
t
i
c
r
e
v
i
e
w
,
”
A
c
a
d
e
m
i
c
Ra
d
i
o
l
o
g
y
,
v
o
l
.
2
7
,
n
o
.
8
,
p
p
.
1
1
7
5
–
1
1
8
5
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
c
r
a
.
2
0
1
9
.
1
2
.
0
2
4
.
[
1
0
]
M
.
A
.
G
a
n
a
i
e
,
M
.
H
u
,
A
.
K
.
M
a
l
i
k
,
M
.
Ta
n
v
e
e
r
,
a
n
d
P
.
N
.
S
u
g
a
n
t
h
a
n
,
“
E
n
s
e
mb
l
e
d
e
e
p
l
e
a
r
n
i
n
g
:
a
r
e
v
i
e
w
,
”
E
n
g
i
n
e
e
r
i
n
g
Ap
p
l
i
c
a
t
i
o
n
s
o
f
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
1
1
5
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
n
g
a
p
p
a
i
.
2
0
2
2
.
1
0
5
1
5
1
.
[
1
1
]
O
.
S
a
g
i
a
n
d
L.
R
o
k
a
c
h
,
“
E
n
sem
b
l
e
l
e
a
r
n
i
n
g
:
a
su
r
v
e
y
,
”
WIR
Es
D
a
t
a
Mi
n
i
n
g
a
n
d
K
n
o
w
l
e
d
g
e
D
i
s
c
o
v
e
r
y
,
v
o
l
.
8
,
n
o
.
4
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
0
2
/
w
i
d
m.
1
2
4
9
.
[
1
2
]
M
.
K
.
P
a
n
d
a
,
B
.
N
.
S
u
b
u
d
h
i
,
T.
V
e
e
r
a
k
u
m
a
r
,
a
n
d
V
.
Ja
k
h
e
t
i
y
a
,
“
M
o
d
i
f
i
e
d
R
e
sN
e
t
-
1
5
2
n
e
t
w
o
r
k
w
i
t
h
h
y
b
r
i
d
p
y
r
a
mi
d
a
l
p
o
o
l
i
n
g
f
o
r
l
o
c
a
l
c
h
a
n
g
e
d
e
t
e
c
t
i
o
n
,
”
I
E
EE
T
ra
n
s
a
c
t
i
o
n
s
o
n
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
5
,
n
o
.
4
,
p
p
.
1
5
9
9
–
1
6
1
2
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
TA
I
.
2
0
2
3
.
3
2
9
9
9
0
3
.
[
1
3
]
P
.
P
.
D
a
l
v
i
,
D
.
R
.
E
d
l
a
,
a
n
d
B
.
R
.
P
u
r
u
sh
o
t
h
a
ma,
“
D
i
a
g
n
o
s
i
s o
f
c
o
r
o
n
a
v
i
r
u
s
d
i
se
a
se
f
r
o
m
c
h
e
s
t
X
-
r
a
y
i
ma
g
e
s
u
si
n
g
D
e
n
seN
e
t
-
169
a
r
c
h
i
t
e
c
t
u
r
e
,
”
S
N
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
4
,
n
o
.
3
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s
4
2
9
7
9
-
0
2
2
-
0
1
6
2
7
-
7.
[
1
4
]
M
.
La
t
h
a
,
P
.
S
.
K
u
mar
,
R
.
R
.
C
h
a
n
d
r
i
k
a
,
T.
R
.
M
a
h
e
s
h
,
V
.
V
.
K
u
mar
,
a
n
d
S
.
G
u
l
u
w
a
d
i
,
“
R
e
v
o
l
u
t
i
o
n
i
z
i
n
g
b
r
e
a
s
t
u
l
t
r
a
s
o
u
n
d
d
i
a
g
n
o
st
i
c
s
w
i
t
h
Ef
f
i
c
i
e
n
t
N
e
t
-
B
7
a
n
d
e
x
p
l
a
i
n
a
b
l
e
A
I
,
”
BM
C
M
e
d
i
c
a
l
I
m
a
g
i
n
g
,
v
o
l
.
2
4
,
n
o
.
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
8
6
/
s1
2
8
8
0
-
024
-
0
1
4
0
4
-
3.
[
1
5
]
S
.
M
o
t
a
m
e
d
,
P
.
R
o
g
a
l
l
a
,
a
n
d
F
.
K
h
a
l
v
a
t
i
,
“
D
a
t
a
a
u
g
m
e
n
t
a
t
i
o
n
u
si
n
g
g
e
n
e
r
a
t
i
v
e
a
d
v
e
r
s
a
r
i
a
l
n
e
t
w
o
r
k
s
(
G
A
N
s)
f
o
r
G
A
N
-
b
a
se
d
d
e
t
e
c
t
i
o
n
o
f
p
n
e
u
mo
n
i
a
a
n
d
C
O
V
I
D
-
1
9
i
n
c
h
e
s
t
X
-
r
a
y
i
ma
g
e
s
,
”
I
n
f
o
rm
a
t
i
c
s
i
n
M
e
d
i
c
i
n
e
U
n
l
o
c
k
e
d
,
v
o
l
.
2
7
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
mu
.
2
0
2
1
.
1
0
0
7
7
9
.
[
1
6
]
W
.
L.
B
i
e
t
a
l
.
,
“
A
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
i
n
c
a
n
c
e
r
i
ma
g
i
n
g
:
c
l
i
n
i
c
a
l
c
h
a
l
l
e
n
g
e
s
a
n
d
a
p
p
l
i
c
a
t
i
o
n
s
,
”
C
A:
A
C
a
n
c
e
r
J
o
u
rn
a
l
f
o
r
C
l
i
n
i
c
i
a
n
s
,
v
o
l
.
6
9
,
n
o
.
2
,
p
p
.
1
2
7
–
1
5
7
,
2
0
1
9
,
d
o
i
:
1
0
.
3
3
2
2
/
c
a
a
c
.
2
1
5
5
2
.
[
1
7
]
V
i
j
a
y
a
l
a
x
mi
,
M
.
F
a
t
a
h
i
,
a
n
d
O
.
S
p
e
c
k
,
“
M
a
g
n
e
t
i
c
r
e
s
o
n
a
n
c
e
i
ma
g
i
n
g
(
M
R
I
)
:
a
r
e
v
i
e
w
o
f
g
e
n
e
t
i
c
d
a
ma
g
e
i
n
v
e
st
i
g
a
t
i
o
n
s
,
”
Mu
t
a
t
i
o
n
R
e
se
a
rc
h
/
Re
v
i
e
w
s
i
n
M
u
t
a
t
i
o
n
Re
s
e
a
rc
h
,
v
o
l
.
7
6
4
,
p
p
.
5
1
–
6
3
,
2
0
1
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
mr
r
e
v
.
2
0
1
5
.
0
2
.
0
0
2
.
[
1
8
]
M
.
K
a
y
a
a
n
d
Y
.
Ç
.
-
K
a
y
a
,
“
A
n
o
v
e
l
e
n
sem
b
l
e
l
e
a
r
n
i
n
g
f
r
a
m
e
w
o
r
k
b
a
sed
o
n
a
g
e
n
e
t
i
c
a
l
g
o
r
i
t
h
m
f
o
r
t
h
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
p
n
e
u
mo
n
i
a
,
”
E
n
g
i
n
e
e
r
i
n
g
A
p
p
l
i
c
a
t
i
o
n
s o
f
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
1
3
3
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
n
g
a
p
p
a
i
.
2
0
2
4
.
1
0
8
4
9
4
.
[
1
9
]
J.
N
a
i
d
o
o
,
S
.
C
.
S
h
e
l
m
e
r
d
i
n
e
,
C
.
F
.
U
.
-
C
h
a
r
c
a
p
e
,
a
n
d
A
.
S
.
S
o
d
h
i
,
“
A
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
i
n
p
a
e
d
i
a
t
r
i
c
t
u
b
e
r
c
u
l
o
si
s
,
”
Pe
d
i
a
t
ri
c
Ra
d
i
o
l
o
g
y
,
v
o
l
.
5
3
,
n
o
.
9
,
p
p
.
1
7
3
3
–
1
7
4
5
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
2
4
7
-
023
-
0
5
6
0
6
-
9.
[
2
0
]
A
.
S
.
Lu
n
d
e
r
v
o
l
d
a
n
d
A
.
L
u
n
d
e
r
v
o
l
d
,
“
A
n
o
v
e
r
v
i
e
w
o
f
d
e
e
p
l
e
a
r
n
i
n
g
i
n
me
d
i
c
a
l
i
m
a
g
i
n
g
f
o
c
u
si
n
g
o
n
M
R
I
,
”
Z
e
i
t
sc
h
ri
f
t
f
ü
r
Me
d
i
z
i
n
i
sc
h
e
Ph
y
si
k
,
v
o
l
.
2
9
,
n
o
.
2
,
p
p
.
1
0
2
–
1
2
7
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
z
e
m
e
d
i
.
2
0
1
8
.
1
1
.
0
0
2
.
[
2
1
]
S
.
K
u
m
a
r
,
H
.
K
u
mar,
G
.
K
u
m
a
r
,
S
.
P
.
S
i
n
g
h
,
A
.
B
i
j
a
l
w
a
n
,
a
n
d
M
.
D
i
w
a
k
a
r
,
“
A
me
t
h
o
d
i
c
a
l
e
x
p
l
o
r
a
t
i
o
n
o
f
i
m
a
g
i
n
g
mo
d
a
l
i
t
i
e
s
f
r
o
m
d
a
t
a
s
e
t
t
o
d
e
t
e
c
t
i
o
n
t
h
r
o
u
g
h
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
p
a
r
a
d
i
g
ms
i
n
p
r
o
m
i
n
e
n
t
l
u
n
g
d
i
sea
se
d
i
a
g
n
o
s
i
s:
a
r
e
v
i
e
w
,
”
B
MC
Me
d
i
c
a
l
I
m
a
g
i
n
g
,
v
o
l
.
2
4
,
n
o
.
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
8
6
/
s
1
2
8
8
0
-
0
2
4
-
0
1
1
9
2
-
w.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
2
,
Ap
r
il 2
0
2
6
:
1
6
0
5
-
1
6
1
2
1612
[
2
2
]
A
.
K
e
b
a
i
l
i
,
J.
L
.
-
La
h
o
r
g
u
e
,
a
n
d
S
.
R
u
a
n
,
“
D
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
e
s
f
o
r
d
a
t
a
a
u
g
m
e
n
t
a
t
i
o
n
i
n
m
e
d
i
c
a
l
i
ma
g
i
n
g
:
a
r
e
v
i
e
w
,
”
J
o
u
rn
a
l
o
f
I
m
a
g
i
n
g
,
v
o
l
.
9
,
n
o
.
4
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
j
i
ma
g
i
n
g
9
0
4
0
0
8
1
.
[
2
3
]
C.
-
H
.
Li
n
,
C
.
-
J.
L
i
n
,
Y
.
-
C
.
L
i
,
a
n
d
S
.
-
H
.
W
a
n
g
,
“
U
si
n
g
g
e
n
e
r
a
t
i
v
e
a
d
v
e
r
sar
i
a
l
n
e
t
w
o
r
k
s
a
n
d
p
a
r
a
met
e
r
o
p
t
i
m
i
z
a
t
i
o
n
o
f
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 f
o
r
l
u
n
g
t
u
m
o
r
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
Ap
p
l
i
e
d
S
c
i
e
n
c
e
s
,
v
o
l
.
1
1
,
n
o
.
2
,
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
1
1
0
2
0
4
8
0
.
[
2
4
]
S
.
I
sl
a
m
e
t
a
l
.
,
“
G
e
n
e
r
a
t
i
v
e
a
d
v
e
r
sari
a
l
n
e
t
w
o
r
k
s
(
G
A
N
s)
i
n
m
e
d
i
c
a
l
i
m
a
g
i
n
g
:
a
d
v
a
n
c
e
m
e
n
t
s,
a
p
p
l
i
c
a
t
i
o
n
s
a
n
d
c
h
a
l
l
e
n
g
e
s
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
2
,
p
p
.
3
5
7
2
8
–
3
5
7
5
3
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
4
.
3
3
7
0
8
4
8
.
[
2
5
]
S
.
G
.
A
r
ma
t
o
I
I
I
e
t
a
l
.
,
“
D
a
t
a
f
r
o
m
t
h
e
l
u
n
g
i
m
a
g
e
d
a
t
a
b
a
s
e
c
o
n
s
o
r
t
i
u
m
(
L
I
D
C
)
a
n
d
i
ma
g
e
d
a
t
a
b
a
s
e
r
e
so
u
r
c
e
i
n
i
t
i
a
t
i
v
e
(
I
D
R
I
)
:
a
c
o
m
p
l
e
t
e
d
r
e
f
e
r
e
n
c
e
d
a
t
a
b
a
s
e
o
f
l
u
n
g
n
o
d
u
l
e
s
o
n
C
T
sc
a
n
s,
”
T
h
e
C
a
n
c
e
r
I
m
a
g
i
n
g
Arc
h
i
v
e
,
2
0
1
5
,
d
o
i
:
1
0
.
7
9
3
7
/
K
9
/
TC
I
A
.
2
0
1
5
.
LO
9
Q
L
9
S
X
.
[
2
6
]
M
.
A
h
a
r
o
n
u
a
n
d
L.
K
.
R
a
m
a
sam
y
,
“
A
n
i
n
t
e
l
l
i
g
e
n
t
g
e
n
e
r
a
t
i
v
e
a
d
v
e
r
sari
a
l
n
e
t
w
o
r
k
m
u
l
t
i
s
t
a
g
e
l
u
n
g
c
a
n
c
e
r
d
e
t
e
c
t
i
o
n
a
n
d
s
u
b
t
y
p
e
s
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
M
a
c
h
i
n
e
L
e
a
rn
i
n
g
a
n
d
C
y
b
e
rn
e
t
i
c
s
,
v
o
l
.
1
6
,
n
o
.
5
–
6
,
p
p
.
3
8
1
9
–
3
8
4
2
,
2
0
2
5
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
0
4
2
-
024
-
0
2
4
8
4
-
x.
[
2
7
]
T.
-
K
.
Li
n
e
t
a
l
.
,
“
D
a
ma
g
e
sce
n
a
r
i
o
p
r
e
d
i
c
t
i
o
n
f
o
r
c
o
n
c
r
e
t
e
b
r
i
d
g
e
c
o
l
u
m
n
s
u
si
n
g
d
e
e
p
g
e
n
e
r
a
t
i
v
e
n
e
t
w
o
r
k
s
,
”
S
t
ru
c
t
u
ra
l
C
o
n
t
r
o
l
a
n
d
H
e
a
l
t
h
Mo
n
i
t
o
ri
n
g
,
v
o
l
.
2
0
2
4
,
n
o
.
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
5
5
/
2
0
2
4
/
5
5
2
6
5
3
7
.
[
2
8
]
J.
L
e
e
a
n
d
R
.
M
.
N
i
sh
i
k
a
w
a
,
“
I
mp
r
o
v
i
n
g
l
e
s
i
o
n
d
e
t
e
c
t
i
o
n
i
n
mamm
o
g
r
a
m
s
b
y
l
e
v
e
r
a
g
i
n
g
a
C
y
c
l
e
-
GAN
-
b
a
se
d
l
e
si
o
n
r
e
mo
v
e
r
,
”
Bre
a
s
t
C
a
n
c
e
r
Re
se
a
r
c
h
,
v
o
l
.
2
6
,
n
o
.
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
8
6
/
s
1
3
0
5
8
-
0
2
4
-
0
1
7
7
7
-
x.
[
2
9
]
A
.
M
e
l
n
i
k
e
t
a
l
.
,
“
F
a
c
e
g
e
n
e
r
a
t
i
o
n
a
n
d
e
d
i
t
i
n
g
w
i
t
h
S
t
y
l
e
G
A
N
:
a
s
u
r
v
e
y
,
”
I
EEE
T
r
a
n
sa
c
t
i
o
n
s
o
n
P
a
t
t
e
r
n
A
n
a
l
y
si
s
a
n
d
M
a
c
h
i
n
e
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
4
6
,
n
o
.
5
,
p
p
.
3
5
5
7
–
3
5
7
6
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
TPA
M
I
.
2
0
2
4
.
3
3
5
0
0
0
4
.
[
3
0
]
J.
Z
h
o
u
e
t
a
l
.
,
“
A
n
e
n
s
e
mb
l
e
d
e
e
p
l
e
a
r
n
i
n
g
m
o
d
e
l
f
o
r
r
i
s
k
st
r
a
t
i
f
i
c
a
t
i
o
n
o
f
i
n
v
a
si
v
e
l
u
n
g
a
d
e
n
o
c
a
r
c
i
n
o
m
a
u
s
i
n
g
t
h
i
n
-
s
l
i
c
e
C
T
,
”
N
PJ
D
i
g
i
t
a
l
M
e
d
i
c
i
n
e
,
v
o
l
.
6
,
n
o
.
1
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
3
8
/
s
4
1
7
4
6
-
0
2
3
-
0
0
8
6
6
-
z.
[
3
1
]
A
.
A
.
S
h
a
h
,
H
.
A
.
M
.
M
a
l
i
k
,
A
.
M
u
h
a
mm
a
d
,
A
.
A
l
o
u
r
a
n
i
,
a
n
d
Z.
A
.
B
u
t
t
,
“
D
e
e
p
l
e
a
r
n
i
n
g
e
n
sem
b
l
e
2
D
C
N
N
a
p
p
r
o
a
c
h
t
o
w
a
r
d
s
t
h
e
d
e
t
e
c
t
i
o
n
o
f
l
u
n
g
c
a
n
c
e
r
,
”
S
c
i
e
n
t
i
f
i
c
Re
p
o
rt
s
,
v
o
l
.
1
3
,
n
o
.
1
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
3
8
/
s
4
1
5
9
8
-
0
2
3
-
2
9
6
5
6
-
z.
[
3
2
]
N
.
G
a
u
t
a
m,
A
.
B
a
s
u
,
a
n
d
R
.
S
a
r
k
a
r
,
“
Lu
n
g
c
a
n
c
e
r
d
e
t
e
c
t
i
o
n
f
r
o
m
t
h
o
r
a
c
i
c
C
T
s
c
a
n
s
u
s
i
n
g
a
n
e
n
s
e
m
b
l
e
o
f
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s,
”
N
e
u
ra
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
6
,
n
o
.
5
,
p
p
.
2
4
5
9
–
2
4
7
7
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s0
0
5
2
1
-
0
2
3
-
0
9
1
3
0
-
7.
[
3
3
]
M
.
K
r
y
u
c
h
k
o
v
,
N
.
K
h
a
n
z
h
i
n
a
,
I
.
O
smak
o
v
,
a
n
d
P
.
U
l
y
a
n
o
v
,
“
C
T
i
ma
g
e
s
G
A
N
-
b
a
se
d
a
u
g
m
e
n
t
a
t
i
o
n
w
i
t
h
A
d
a
I
N
f
o
r
l
u
n
g
n
o
d
u
l
e
s
d
e
t
e
c
t
i
o
n
,
”
i
n
T
h
i
r
t
e
e
n
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
M
a
c
h
i
n
e
V
i
si
o
n
,
2
0
2
1
.
d
o
i
:
1
0
.
1
1
1
7
/
1
2
.
2
5
8
7
9
4
0
.
[
3
4
]
R
.
T
o
d
a
,
A
.
Te
r
a
m
o
t
o
,
M
.
K
o
n
d
o
,
K
.
I
ma
i
z
u
m
i
,
K
.
S
a
i
t
o
,
a
n
d
H
.
F
u
j
i
t
a
,
“
L
u
n
g
c
a
n
c
e
r
C
T
i
ma
g
e
g
e
n
e
r
a
t
i
o
n
f
r
o
m
a
f
r
e
e
-
f
o
r
m s
k
e
t
c
h
u
si
n
g
s
t
y
l
e
-
b
a
s
e
d
p
i
x
2
p
i
x
f
o
r
d
a
t
a
a
u
g
me
n
t
a
t
i
o
n
,
”
S
c
i
e
n
t
i
f
i
c
R
e
p
o
r
t
s
,
v
o
l
.
1
2
,
n
o
.
1
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
3
8
/
s
4
1
5
9
8
-
022
-
16861
-
5.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Bin
e
e
sh
Mo
o
z
h
ip
p
u
r
a
t
h
is
c
u
rre
n
tl
y
p
u
rsu
in
g
a
P
h
.
D
.
i
n
t
h
e
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
E
n
g
i
n
e
e
rin
g
a
t
CHRIST
Un
i
v
e
rsity
,
Be
n
g
a
lu
ru
,
f
o
c
u
sin
g
o
n
c
a
n
c
e
r
p
re
d
ictio
n
u
si
n
g
m
e
tab
o
lo
m
ics
a
n
d
m
a
c
h
in
e
lea
rn
i
n
g
.
He
c
o
m
p
let
e
d
h
is
M
.
E
.
(Co
m
p
u
ter
a
n
d
Co
m
m
u
n
ica
ti
o
n
)
fro
m
An
n
a
U
n
iv
e
rsit
y
,
Tam
il
Na
d
u
,
i
n
2
0
1
1
.
He
h
o
ld
s
a
Ba
c
h
e
lo
r
o
f
Tec
h
n
o
l
o
g
y
(B.
Tec
h
.
)
d
e
g
re
e
in
In
fo
rm
a
ti
o
n
Tec
h
n
o
lo
g
y
fro
m
th
e
Co
c
h
in
Un
i
v
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
lo
g
y
,
Ke
ra
la,
in
2
0
0
6
.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
m
a
c
h
in
e
lea
rn
i
n
g
,
g
ra
p
h
n
e
u
ra
l
n
e
two
rk
s,
a
n
d
m
e
tab
o
lo
m
ics
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
b
in
e
e
sh
.
m
@re
s.c
h
ristu
n
i
v
e
rsity
.
i
n
.
J
a
y
a
p
a
n
d
ia
n
Na
t
a
r
a
ja
n
is
c
u
rre
n
tl
y
wo
r
k
in
g
a
s
a
n
a
ss
o
c
iat
e
p
ro
fe
ss
o
r
in
t
h
e
De
p
a
rtme
n
t
o
f
C
o
m
p
u
ter
S
c
ien
c
e
a
n
d
E
n
g
in
e
e
rin
g
a
t
CHRIST
Un
i
v
e
rsity
,
Be
n
g
a
lu
r
u
.
He
h
a
s
re
c
e
iv
e
d
h
is
P
h
.
D
.
fro
m
An
n
a
Un
iv
e
rsity
,
Ch
e
n
n
a
i.
He
is
a
n
a
c
ti
v
e
li
fe
m
e
m
b
e
r
o
f
IS
T
E.
He
i
s
c
u
rre
n
tl
y
d
o
i
n
g
h
is
re
se
a
rc
h
i
n
th
e
fiel
d
o
f
c
lo
u
d
c
o
m
p
u
ti
n
g
.
I
n
h
is
1
6
y
e
a
rs
o
f
tea
c
h
in
g
e
x
p
e
rien
c
e
a
n
d
o
n
e
y
e
a
r
o
f
in
d
u
s
try
e
x
p
e
rien
c
e
.
His
re
se
a
rc
h
in
ter
e
sts
a
re
g
rid
c
o
m
p
u
ti
n
g
a
n
d
c
lo
u
d
c
o
m
p
u
ti
n
g
.
He
h
a
s
p
u
b
li
sh
e
d
in
4
b
o
o
k
c
h
a
p
ters
,
3
5
i
n
tern
a
ti
o
n
a
l
jo
u
rn
a
l
a
rti
c
les
,
a
n
d
6
5
i
n
tern
a
ti
o
n
a
l
a
n
d
n
a
ti
o
n
a
l
c
o
n
fe
re
n
c
e
s.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
jay
a
p
a
n
d
ia
n
.
n
@c
h
ristu
n
iv
e
rsit
y
.
i
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.