I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
36
,
No
.
2
,
N
o
v
em
b
e
r
20
24
,
p
p
.
1
1
4
9
~1
162
I
SS
N:
2
502
-
4
7
52
,
DOI
: 1
0
.
1
1
5
9
1
/ijee
cs
.v
36.
i
2
.
p
p
1
1
4
9
-
1
1
6
2
1149
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs
.
ia
esco
r
e.
co
m
Co
nv
o
lutiona
l neural networks
bre
a
st
ca
ncer c
la
ss
ifi
ca
tion
using
P
a
lestinia
n
ma
mm
o
g
ra
m da
tas
et
H
a
nin
Sa
a
da
h
1
,
Am
a
ni Yo
us
ef
O
wda
1
,
M
a
j
di O
wda
2
1
D
e
p
a
r
t
me
n
t
o
f
N
a
t
u
r
a
l
E
n
g
i
n
e
e
r
i
n
g
a
n
d
Te
c
h
n
o
l
o
g
y
S
c
i
e
n
c
e
s
,
F
a
c
u
l
t
y
o
f
G
r
a
d
u
a
t
e
S
t
u
d
i
e
s,
A
r
a
b
A
meri
c
a
n
U
n
i
v
e
r
si
t
y
,
R
a
ma
l
l
a
h
,
P
a
l
e
s
t
i
n
e
2
F
a
c
u
l
t
y
o
f
D
a
t
a
S
c
i
e
n
c
e
,
U
N
ES
C
O
C
h
a
i
r
o
n
D
a
t
a
S
c
i
e
n
c
e
f
o
r
S
u
s
t
a
i
n
a
b
l
e
D
e
v
e
l
o
p
m
e
n
t
,
A
r
a
b
A
meri
c
a
n
U
n
i
v
e
r
si
t
y
,
R
a
ma
l
l
a
h
,
P
a
l
e
s
t
i
n
e
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Feb
29
,
2
0
2
4
R
ev
is
ed
Au
g
1
,
2
0
2
4
Acc
ep
ted
Au
g
5
,
2
0
2
4
Bre
a
st
c
a
n
c
e
r
is
wid
e
sp
re
a
d
a
c
ro
ss
th
e
g
lo
b
e
.
It’s
th
e
p
rima
ry
c
a
u
se
o
f
d
e
a
th
in
c
a
n
c
e
r
fa
talit
ies
.
Ac
c
o
rd
in
g
t
o
th
e
P
a
les
ti
n
ia
n
M
i
n
istry
o
f
He
a
lt
h
a
n
n
u
a
l
re
p
o
rt,
it
ra
n
k
e
d
a
s
t
h
e
th
ird
re
p
o
rted
d
e
a
th
o
f
a
ll
re
p
o
rted
c
a
n
c
e
r
d
e
a
th
s
in
th
e
Wes
t
Ba
n
k
.
M
a
m
m
o
g
ra
m
sc
re
e
n
in
g
is
th
e
m
o
st
c
o
m
m
o
n
tec
h
n
i
q
u
e
t
o
d
iag
n
o
se
b
re
a
st
a
b
n
o
rm
a
li
ti
e
s,
b
u
t
th
e
re
is
a
c
h
a
ll
e
n
g
e
in
th
e
lac
k
o
f
sk
il
led
e
x
p
e
rts
a
b
le t
o
a
c
c
u
ra
tely
i
n
terp
re
t
m
a
m
m
o
g
ra
m
s.
M
a
c
h
in
e
lea
rn
in
g
p
la
y
s a
n
imp
o
rtan
t
r
o
le
in
m
e
d
ica
l
ima
g
e
p
ro
c
e
ss
in
g
p
a
rti
c
u
larly
in
e
a
rly
d
e
tec
ti
o
n
wh
e
n
th
e
trea
tme
n
t
is
les
s
e
x
p
e
n
siv
e
a
n
d
a
v
a
il
a
b
le.
In
t
h
is
p
a
p
e
r
we
p
ro
p
o
se
d
d
iffere
n
t
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
(CNN
)
m
o
d
e
ls
to
d
e
tec
t
b
re
a
st
a
b
n
o
rm
a
li
ti
e
s
with
p
ro
m
is
in
g
re
su
l
ts.
S
i
x
CNN
m
o
d
e
ls
we
re
u
se
d
in
th
is
re
se
a
rc
h
o
n
a
u
n
i
q
u
e
(fir
st
-
h
a
n
d
)
d
a
tas
e
t
c
o
ll
e
c
ted
fr
o
m
th
e
P
a
les
ti
n
ian
M
in
istr
y
o
f
He
a
lt
h
.
T
h
e
m
o
d
e
ls
a
re
VG
G
1
6
,
VG
G
1
9
,
De
n
se
Ne
t1
2
1
,
Re
sN
e
t5
0
,
Xc
e
p
ti
o
n
,
a
n
d
Eff
i
c
ien
tNe
tB7
.
C
o
n
se
q
u
e
n
tl
y
,
De
n
se
Ne
t1
2
1
o
u
t
p
e
rfo
rm
e
d
o
t
h
e
r
m
o
d
e
ls
wit
h
0
.
8
3
a
n
d
0
.
8
5
f
o
r
tes
ti
n
g
a
c
c
u
ra
c
y
a
n
d
a
re
a
u
n
d
e
r
c
u
r
v
e
(AU
C)
re
sp
e
c
ti
v
e
ly
.
As
a
fu
t
u
re
w
o
rk
,
t
h
e
o
u
t
p
e
rfo
rm
e
d
m
o
d
e
l
c
a
n
b
e
c
o
m
b
in
e
d
wi
th
o
th
e
r
p
a
ti
e
n
t
d
a
ta
li
k
e
g
e
n
e
ti
c
in
f
o
rm
a
ti
o
n
,
m
e
d
ica
l
h
isto
r
y
,
a
n
d
li
fe
sty
le
fa
c
to
rs
to
e
v
a
lu
a
te
th
e
r
isk
o
f
d
e
v
e
l
o
p
i
n
g
sp
e
c
ifi
c
d
ise
a
se
s.
Th
is
wo
u
ld
in
c
re
a
se
th
e
su
r
v
iv
a
l
ra
te
a
n
d
e
n
a
b
le
p
ro
a
c
ti
v
e
m
e
a
su
re
s.
K
ey
w
o
r
d
s
:
B
r
ea
s
t c
an
ce
r
C
o
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
I
m
ag
e
p
r
o
ce
s
s
in
g
Ma
ch
in
e
lear
n
in
g
Ma
m
m
o
g
r
am
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
:
Am
an
i Y
o
u
s
ef
Owd
a
Dep
ar
tm
en
t o
f
Natu
r
al
E
n
g
in
e
er
in
g
an
d
T
ec
h
n
o
lo
g
y
Scien
ce
s
,
Facu
lty
o
f
Gr
ad
u
ate
Stu
d
ies
Ar
ab
Am
er
ican
Un
iv
er
s
ity
P6
0
0
,
R
am
allah
,
Palest
in
e
E
m
ail:
a
m
an
i.o
wd
a@
aa
u
p
.
ed
u
1.
I
NT
RO
D
UCT
I
O
N
B
r
ea
s
t
ca
n
ce
r
is
p
r
ev
alen
t
g
lo
b
ally
,
an
d
it’s
th
e
f
ir
s
t
ca
u
s
e
o
f
ca
n
ce
r
-
r
elate
d
f
atalities
in
wo
m
en
,
p
ar
ticu
lar
ly
im
p
ac
tin
g
th
o
s
e
in
lo
w
-
an
d
m
id
d
le
-
in
co
m
e
co
m
m
u
n
ities
[
1
]
.
I
n
th
e
W
est B
an
k
,
b
r
ea
s
t c
an
ce
r
was
th
e
th
ir
d
r
ep
o
r
te
d
d
ea
th
o
f
all
r
ep
o
r
ted
ca
n
ce
r
d
ea
th
s
in
2
0
2
1
as
s
h
o
wn
in
Fig
u
r
e
1
,
an
d
it
ac
co
u
n
ted
f
o
r
3
0
%
o
f
d
ea
th
s
am
o
n
g
wo
m
e
n
o
f
r
e
p
r
o
d
u
ctiv
e
ag
e
[
2
]
.
W
h
ile
b
r
e
ast
ca
n
ce
r
is
ty
p
ically
d
iag
n
o
s
ed
m
o
r
e
f
r
eq
u
en
tly
in
wo
m
en
ag
ed
f
if
ty
y
ea
r
s
o
r
o
ld
er
,
th
e
r
e
is
a
co
n
ce
r
n
i
n
g
in
cr
ea
s
e
in
its
o
cc
u
r
r
e
n
ce
am
o
n
g
y
o
u
n
g
e
r
wo
m
e
n
.
C
u
r
r
en
tly
,
b
r
ea
s
t
ca
n
ce
r
r
a
n
k
s
as
th
e
s
ec
o
n
d
lead
in
g
ca
u
s
e
o
f
ca
n
ce
r
-
r
elate
d
d
ea
th
s
in
wo
m
en
ag
ed
less
th
an
4
0
g
lo
b
ally
[
3
]
.
Ap
p
r
o
x
im
ate
ly
3
0
%
o
f
b
r
ea
s
t
ca
n
ce
r
ca
s
es
ar
e
p
r
ev
en
tab
le
th
r
o
u
g
h
m
o
d
if
iab
le
r
is
k
f
ac
to
r
s
lik
e
ex
ce
s
s
b
o
d
y
weig
h
t,
p
h
y
s
ic
al
in
ac
tiv
ity
,
an
d
alco
h
o
l
in
tak
e,
a
n
d
c
an
b
e
f
u
r
th
er
r
ed
u
ce
d
th
r
o
u
g
h
m
am
m
o
g
r
a
p
h
y
s
cr
ee
n
in
g
an
d
ad
v
an
ce
s
in
tr
ea
tm
en
t
[
4
]
.
T
h
r
o
u
g
h
o
u
t
a
wo
m
an
'
s
life
,
th
e
b
r
ea
s
t
u
n
d
er
g
o
es
n
u
m
er
o
u
s
tr
an
s
f
o
r
m
atio
n
s
,
s
p
an
n
in
g
f
r
o
m
in
f
a
n
cy
,
a
d
o
lescen
ce
,
an
d
m
o
th
e
r
h
o
o
d
i
n
clu
d
in
g
p
r
e
g
n
an
c
y
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
36
,
No
.
2
,
No
v
em
b
er
20
24
:
1
1
4
9
-
1
1
6
2
1150
b
r
ea
s
tf
ee
d
in
g
,
all
th
e
way
t
o
t
h
e
m
en
o
p
au
s
al
s
tag
e.
T
h
e
b
r
e
ast
is
a
g
lan
d
u
lar
o
r
g
an
lo
ca
te
d
o
n
th
e
ch
est
wall
o
f
f
em
ales.
T
h
o
u
g
h
n
o
t
an
ato
m
ically
ca
teg
o
r
ized
as
p
ar
t
o
f
th
e
b
o
d
y
'
s
u
p
p
er
lim
b
,
it
r
esid
es
in
th
e
p
ec
t
o
r
al
r
eg
io
n
,
with
its
b
lo
o
d
s
u
p
p
ly
an
d
ly
m
p
h
atic
d
r
ain
a
g
e
p
r
im
a
r
ily
d
ir
ec
te
d
to
war
d
t
h
e
ar
m
p
i
t.
I
t’
s
a
s
p
ec
ialized
ac
ce
s
s
o
r
y
g
lan
d
o
f
t
h
e
s
k
in
th
at
s
ec
r
etes
m
ilk
an
d
ex
is
ts
in
b
o
th
m
ales
an
d
f
em
ales
[
5
]
.
I
n
m
ales
an
d
im
m
atu
r
e
f
em
ales,
th
eir
s
tr
u
ct
u
r
e
is
s
im
ilar
.
T
h
er
e
ar
e
m
an
y
s
y
m
p
to
m
s
ass
o
ciate
d
with
b
r
ea
s
t
ca
n
ce
r
li
k
e
s
h
ap
e,
s
ize,
co
lo
r
,
p
ai
n
,
o
r
s
wellin
g
in
th
e
b
r
ea
s
t o
r
n
ip
p
le.
Nip
p
le
in
v
er
s
io
n
tu
r
n
in
g
in
war
d
o
r
d
i
s
c
h
ar
g
e
o
th
er
th
a
n
m
ilk
,
an
d
r
ed
n
ess
o
r
th
ick
n
ess
in
th
e
u
n
d
er
a
r
m
ar
ea
.
Sy
m
p
to
m
s
ca
n
b
e
p
h
y
s
ically
o
b
v
io
u
s
o
r
tan
g
ib
le,
wh
ile
o
th
er
s
m
ay
b
e
s
u
b
tle
o
r
in
tan
g
ib
le
[
6
]
.
Ma
n
y
o
f
th
ese
s
y
m
p
to
m
s
m
ay
o
cc
u
r
at
th
e
s
am
e
tim
e.
T
h
e
ca
n
ce
r
clin
ician
s
ca
lled
th
is
ca
s
e
s
y
m
p
to
m
clu
s
ter
s
[
7
]
.
A
s
tu
d
y
in
J
o
r
d
a
n
s
h
o
wed
th
at
th
e
r
e
ar
e
f
iv
e
m
a
in
s
y
m
p
to
m
clu
s
ter
s
am
o
n
g
b
r
ea
s
t
ca
n
ce
r
wo
m
en
in
J
o
r
d
an
[
8
]
.
T
h
e
clu
s
ter
s
ar
e
f
atig
u
e,
p
ai
n
,
tr
ea
tm
en
t
s
i
d
e
ef
f
ec
ts
,
p
s
y
ch
o
l
o
g
ical,
n
a
u
s
ea
,
an
d
v
o
m
itin
g
.
On
th
e
o
th
e
r
h
a
n
d
,
b
r
ea
s
t
ca
n
ce
r
h
as
m
an
y
ass
o
ciate
d
r
is
k
f
ac
to
r
s
th
at
af
f
ec
t
i
n
cr
ea
s
ed
r
is
k
,
an
d
th
ese
r
is
k
f
ac
to
r
s
ca
n
b
e
ca
teg
o
r
ize
d
in
to
two
ca
teg
o
r
ies;
life
s
ty
le
a
n
d
g
e
n
etic
[
3
]
,
[
9
]
.
Ar
li
et
a
l
.
[
1
0
]
,
co
n
d
u
cte
d
a
cr
o
s
s
‐
s
ec
tio
n
al
s
tu
d
y
co
n
d
u
cted
in
T
u
r
k
e
y
id
e
n
tifie
d
m
u
ltip
le
b
ir
th
s
,
a
s
h
o
r
t
b
r
ea
s
tf
ee
d
in
g
p
e
r
io
d
,
o
v
er
weig
h
t,
l
o
w
s
o
cio
ec
o
n
o
m
ic
lev
el,
an
d
lo
w
lev
el
o
f
ed
u
ca
ti
o
n
as
th
e
m
o
s
t
s
ig
n
if
ican
t
f
ac
to
r
s
.
E
ar
ly
d
etec
tio
n
is
th
e
co
r
n
e
r
s
to
n
e
o
f
c
o
n
tr
o
llin
g
b
r
ea
s
t
ca
n
ce
r
an
d
im
p
r
o
v
in
g
o
u
tc
o
m
es
an
d
s
u
r
v
i
v
al.
T
h
e
W
o
r
ld
Hea
lth
Or
g
an
iza
tio
n
(
W
HO)
d
ef
in
ed
th
e
ea
r
ly
d
iag
n
o
s
is
o
f
ca
n
ce
r
s
as
th
e
d
etec
tio
n
o
f
in
itial
p
h
ases
in
wo
m
e
n
with
s
y
m
p
t
o
m
s
,
as
th
is
f
ac
ilit
ates
s
tr
aig
h
tf
o
r
war
d
an
d
co
s
t
-
ef
f
ec
tiv
e
tr
ea
tm
en
t,
lead
in
g
to
elev
ated
r
ates
o
f
r
ec
o
v
er
y
[
1
1
]
.
T
h
e
Pan
Am
er
ican
He
alth
Or
g
an
izatio
n
(
PAHO)
co
n
s
id
er
ed
r
aisi
n
g
awa
r
en
ess
th
r
o
u
g
h
p
r
o
g
r
am
s
an
d
ed
u
ca
tio
n
ab
o
u
t
b
r
ea
s
t
ca
n
ce
r
as
th
e
m
o
s
t
im
p
o
r
tan
t
k
ey
elem
en
t
th
at
im
p
r
o
v
es o
u
tco
m
es,
an
d
wo
m
en
s
h
o
u
ld
b
e
em
p
o
wer
ed
t
o
ac
ce
s
s
ca
n
ce
r
s
er
v
ices tim
ely
[
1
2
]
.
C
u
r
r
en
tly
,
th
er
e
ar
e
m
u
ltip
l
e
b
r
ea
s
t
s
cr
ee
n
in
g
tech
n
iq
u
es
av
ailab
le,
s
u
ch
as
m
am
m
o
g
r
ap
h
y
,
u
ltra
s
o
u
n
d
im
ag
in
g
,
p
o
s
itro
n
em
is
s
io
n
to
m
o
g
r
a
p
h
y
(
PET
)
,
co
m
p
u
ted
to
m
o
g
r
ap
h
y
(
C
T
)
,
m
ag
n
etic
r
eso
n
a
n
ce
im
ag
in
g
(
MRI)
,
an
d
m
icr
o
wa
v
e
im
ag
in
g
[
1
3
]
.
B
y
f
ar
,
s
tu
d
ies
h
av
e
p
r
o
v
en
th
at
m
am
m
o
g
r
ap
h
y
is
th
e
m
o
s
t
s
u
f
f
ici
en
t
ev
id
e
n
ce
an
d
ef
f
ec
t
iv
e
in
r
e
d
u
cin
g
m
o
r
tality
[
1
4
]
,
[
1
5
]
.
Ma
m
m
o
g
r
a
p
h
y
is
a
r
e
co
m
m
en
d
e
d
in
itial
s
cr
ee
n
in
g
im
ag
i
n
g
m
et
h
o
d
th
a
t
u
s
es
a
b
r
ea
s
t
-
s
p
ec
if
ic
X
-
r
ay
i
m
ag
in
g
m
o
d
ality
t
o
p
r
o
d
u
ce
i
m
ag
es
o
f
t
h
e
b
r
ea
s
t
in
v
ar
io
u
s
p
o
s
itio
n
s
,
h
elp
in
g
t
o
d
etec
t
ab
n
o
r
m
alit
ies
an
d
lesi
o
n
s
[
1
6
]
.
A
cr
o
s
s
-
s
ec
tio
n
al
s
tu
d
y
was
ca
r
r
ied
o
u
t
in
Palest
in
e
in
2
0
1
6
to
ass
ess
m
am
m
o
g
r
am
s
cr
ee
n
in
g
am
o
n
g
wo
m
en
at
r
is
k
[
1
5
]
.
T
h
e
s
tu
d
y
r
ev
ea
led
t
h
at
5
0
%
o
f
wo
m
e
n
h
ad
u
n
d
er
g
o
n
e
at
least
o
n
e
m
am
m
o
g
r
am
,
b
u
t
o
n
ly
2
1
%
h
ad
r
ec
eiv
ed
tim
ely
m
am
m
o
g
r
am
s
.
Acc
o
r
d
in
g
to
th
e
2
0
2
1
h
ea
lth
an
n
u
al
r
e
p
o
r
t
r
elea
s
ed
b
y
th
e
Palest
in
ian
Min
is
tr
y
o
f
Hea
lth
,
3
9
.
6
%
o
f
all
ca
s
es
ex
am
in
ed
ac
r
o
s
s
th
e
g
o
v
er
n
o
r
ates,
to
talin
g
5
,
8
6
4
,
wer
e
f
o
u
n
d
to
b
e
a
b
n
o
r
m
al,
with
a
t
o
tal
o
f
2
,
3
2
2
ca
s
es
id
en
tifie
d
[
2
]
.
T
h
e
b
r
ea
s
t
im
ag
in
g
r
ep
o
r
ti
n
g
an
d
d
ata
s
y
s
tem
(
B
I
-
R
AD
S)
was
in
tr
o
d
u
ce
d
b
y
th
e
Am
er
ic
an
C
o
lleg
e
o
f
R
ad
io
lo
g
y
(
AC
R
)
in
1
9
9
3
[
1
7
]
.
I
t
’
s
a
s
tan
d
ar
d
s
ca
le
f
o
r
m
am
m
o
g
r
am
r
e
p
o
r
tin
g
,
ai
m
ed
at
en
h
an
cin
g
co
m
m
u
n
icatio
n
am
o
n
g
h
ea
lt
h
ca
r
e
p
r
o
v
id
er
s
,
m
i
n
im
izin
g
am
b
ig
u
ity
s
u
r
r
o
u
n
d
in
g
m
a
m
m
o
g
r
am
r
esu
lts
,
f
ac
ilit
atin
g
ca
s
e
m
an
ag
em
en
t,
an
d
aid
in
g
in
th
e
m
o
n
ito
r
i
n
g
o
f
o
u
tco
m
es
[
1
8
]
.
B
I
-
R
ADS
s
y
s
tem
is
co
m
m
o
n
l
y
u
s
ed
in
m
am
m
o
g
r
a
m
s
,
MRIs,
an
d
u
ltra
s
o
u
n
d
s
.
I
t
em
p
lo
y
s
a
n
u
m
er
ic
s
ca
le
th
at
r
an
g
es
f
r
o
m
0
to
6
,
with
ea
ch
ca
teg
o
r
y
d
e
n
o
tin
g
d
is
tin
ct
lev
els
o
f
s
u
s
p
icio
n
co
n
ce
r
n
in
g
b
r
ea
s
t
ca
n
ce
r
.
T
ab
le
1
s
h
o
ws
d
if
f
er
en
t
B
I
-
R
ADS
ca
teg
o
r
ies
an
d
th
e
ass
ig
n
ed
d
e
s
cr
ip
tio
n
s
[
1
9
]
−
[
2
3
]
.
T
h
er
e
ar
e
s
ev
en
ca
te
g
o
r
ies
in
th
e
B
I
-
R
ADS
s
ca
le
s
tar
tin
g
f
r
o
m
ze
r
o
to
s
ix
.
E
ac
h
ca
teg
o
r
y
h
as
a
s
p
ec
ial
f
in
d
in
g
,
m
an
ag
em
en
t,
an
d
th
e
p
er
ce
n
tag
e
o
f
t
h
ese
ab
n
o
r
m
alities
b
ein
g
ca
n
ce
r
o
u
s
.
T
h
e
f
ir
s
t
ca
teg
o
r
y
(
ze
r
o
)
r
ep
r
esen
ts
an
i
n
co
m
p
lete
ass
ess
m
en
t
an
d
n
o
d
iag
n
o
s
is
ca
n
b
e
ex
tr
ac
ted
f
r
o
m
th
e
im
ag
e.
T
h
e
in
co
m
p
lete
ass
ess
m
en
t
m
ay
b
e
attr
ib
u
ted
to
th
e
m
am
m
o
g
r
am
d
e
v
ice
m
alf
u
n
ctio
n
in
g
o
r
th
e
b
r
e
ast
p
o
s
itio
n
an
d
i
n
th
is
ca
s
e,
th
e
wo
m
an
is
ask
ed
to
r
e
p
ea
t
th
e
s
c
r
ee
n
in
g
.
C
ateg
o
r
ies
o
f
o
n
e
to
t
h
r
ee
m
o
s
t
p
r
o
b
ab
l
y
h
av
e
b
en
i
g
n
o
r
n
o
r
m
al
f
in
d
in
g
s
an
d
h
av
e
less
th
an
a
2
%
ch
an
ce
o
f
b
ein
g
m
alig
n
an
t.
Fu
r
th
er
m
o
r
e,
th
e
c
h
an
ce
th
at
th
e
ab
n
o
r
m
alities
a
r
e
ca
n
ce
r
o
u
s
in
cr
ea
s
es
as
th
e
s
ca
le
in
cr
ea
s
es
f
r
o
m
4
to
6
,
an
d
th
e
p
r
o
b
a
b
ilit
y
ap
p
r
o
ac
h
es
1
0
0
%
i
n
th
e
s
ix
th
g
r
o
u
p
,
w
h
er
e
t
h
e
a
b
n
o
r
m
alities
h
av
e
b
ee
n
d
iag
n
o
s
ed
as m
alig
n
an
t.
Fig
u
r
e
1
.
Pro
p
o
r
tio
n
al
d
is
tr
ib
u
tio
n
o
f
th
e
m
o
s
t
r
ep
o
r
ted
c
a
n
c
er
d
ea
th
s
o
f
W
est B
an
k
in
Palest
in
e
,
2
0
2
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
C
o
n
vo
lu
tio
n
a
l
n
eu
r
a
l n
etw
o
r
ks b
r
ea
s
t c
a
n
ce
r
cla
s
s
ifica
tio
n
u
s
in
g
…
(
Ha
n
in
S
a
a
d
a
h
)
1151
T
ab
le
1
.
B
I
-
R
ADS
c
ateg
o
r
ies
C
a
t
e
g
o
r
y
F
i
n
d
i
n
g
s/
d
i
a
g
n
o
si
s
M
a
n
a
g
e
m
e
n
t
Li
k
e
l
i
h
o
o
d
o
f
d
e
v
e
l
o
p
i
n
g
b
r
e
a
s
t
c
a
n
c
e
r
0
I
n
c
o
mp
l
e
t
e
a
ssess
me
n
t
A
d
d
i
t
i
o
n
a
l
i
ma
g
i
n
g
i
s re
q
u
i
r
e
d
-
1
N
e
g
a
t
i
v
e
o
r
n
o
r
ma
l
f
i
n
d
i
n
g
s
R
o
u
t
i
n
e
f
o
l
l
o
w
-
u
p
e
v
e
r
y
y
e
a
r
0%
2
B
e
n
i
g
n
R
o
u
t
i
n
e
f
o
l
l
o
w
-
u
p
e
v
e
r
y
y
e
a
r
0%
3
P
r
o
b
a
b
l
y
b
e
n
i
g
n
R
o
u
t
i
n
e
f
o
l
l
o
w
-
u
p
a
t
s
h
o
r
t
i
n
t
e
r
v
a
l
s
(
e
v
e
r
y
6
m
o
n
t
h
s)
<
=
2
%
4
S
u
sp
i
c
i
o
u
s f
i
n
d
i
n
g
s f
o
r
mal
i
g
n
a
n
c
y
B
i
o
p
s
y
s
h
o
u
l
d
b
e
c
o
n
s
i
d
e
r
e
d
>
2
%
t
o
<
=
9
5
%
5
H
i
g
h
l
y
s
u
g
g
e
st
i
v
e
o
f
mal
i
g
n
a
n
c
y
Th
e
d
o
c
t
o
r
’
s d
e
c
i
si
o
n
a
n
d
a
p
p
r
o
p
r
i
a
t
e
a
c
t
i
o
n
s
h
o
u
l
d
b
e
t
a
k
e
n
.
B
i
o
p
s
y
i
s re
q
u
i
r
e
d
>
=
9
5
%
6
P
r
o
v
e
n
mal
i
g
n
a
n
c
y
A
p
p
r
o
p
r
i
a
t
e
t
h
e
r
a
p
y
/
su
r
g
e
r
y
1
0
0
%
2.
P
RO
B
L
E
M
ST
A
T
E
M
E
NT
A
s
ig
n
if
ican
t
v
o
lu
m
e
o
f
m
am
m
o
g
r
am
s
is
g
ath
er
ed
th
r
o
u
g
h
an
e
x
ten
s
iv
e
m
am
m
o
g
r
ap
h
y
s
cr
ee
n
in
g
p
r
o
g
r
a
m
an
d
n
ec
ess
itates
ev
alu
atio
n
f
r
o
m
p
r
o
f
icien
t
r
ad
i
o
lo
g
is
ts
,
wh
o
ar
e
q
u
alif
ied
b
u
t
o
v
er
lo
a
d
ed
with
ex
ce
s
s
iv
e
wo
r
k
lo
ad
s
[
2
4
]
.
An
im
p
o
r
tan
t
ch
allen
g
e
in
th
e
f
i
eld
o
f
m
am
m
o
g
r
ap
h
y
is
th
e
s
h
o
r
tag
e
o
f
s
k
illed
d
o
m
ain
ex
p
er
ts
ca
p
ab
le
o
f
ac
cu
r
ately
in
ter
p
r
etin
g
m
am
m
o
g
r
am
s
.
T
h
is
s
ca
r
city
o
f
r
ad
io
l
o
g
is
ts
an
d
clin
ician
s
s
p
ec
ialized
in
b
r
ea
s
t
im
ag
in
g
ca
n
lead
to
d
elay
s
in
d
iag
n
o
s
is
,
in
cr
ea
s
ed
wo
r
k
lo
a
d
s
f
o
r
e
x
is
tin
g
ex
p
er
ts
,
an
d
p
o
ten
tial
er
r
o
r
s
in
th
e
i
n
ter
p
r
e
tatio
n
p
r
o
ce
s
s
.
T
h
e
s
h
o
r
tag
e
c
an
b
e
attr
ib
u
ted
to
v
ar
io
u
s
f
ac
to
r
s
lik
e
in
ad
eq
u
ate
tr
ain
in
g
.
B
y
a
d
d
r
ess
in
g
th
e
s
h
o
r
tag
e
o
f
d
o
m
ain
e
x
p
er
ts
an
d
im
p
lem
en
tin
g
s
tr
ateg
ies
to
im
p
r
o
v
e
in
ter
p
r
etatio
n
p
r
o
ce
s
s
es,
h
ea
lth
ca
r
e
s
y
s
tem
s
ca
n
p
r
o
v
i
d
e
m
o
r
e
tim
ely
an
d
ac
cu
r
ate
d
iag
n
o
s
es,
u
ltima
tely
im
p
r
o
v
in
g
p
atien
t
o
u
tco
m
es
in
th
e
r
ea
lm
o
f
m
a
m
m
o
g
r
a
p
h
y
.
T
h
er
e
f
o
r
e,
th
er
e
is
an
u
n
m
et
n
ee
d
to
d
ev
elo
p
m
ac
h
in
e
lea
r
n
in
g
(
ML
)
m
o
d
els
to
ass
is
t
r
ad
io
lo
g
is
ts
with
m
am
m
o
g
r
ap
h
ic
in
t
er
p
r
etatio
n
,
an
d
ML
m
o
d
el
d
e
v
elo
p
m
en
t
r
e
q
u
ir
es
in
ter
d
is
cip
lin
ar
y
r
esear
ch
th
at
in
teg
r
ates m
ed
ical
s
cien
ce
an
d
en
g
in
ee
r
in
g
[
2
5
]
.
3.
L
I
T
E
R
AT
U
RE
R
E
VI
E
W
ML
p
lay
s
a
s
ig
n
if
ican
t
r
o
le
in
ea
r
ly
d
etec
tio
n
ac
r
o
s
s
v
ar
io
u
s
f
ield
s
,
in
clu
d
in
g
h
ea
lth
ca
r
e,
an
d
h
as
em
er
g
ed
as
a
p
o
wer
f
u
l
to
o
l
in
m
am
m
o
g
r
ap
h
y
in
ter
p
r
eta
tio
n
,
r
ev
o
lu
tio
n
izin
g
th
e
f
iel
d
o
f
b
r
ea
s
t
ca
n
ce
r
s
cr
ee
n
in
g
a
n
d
d
iag
n
o
s
is
[
2
6
]
.
B
y
u
tili
zin
g
ML
tech
n
iq
u
es,
m
ed
ical
p
r
o
f
ess
io
n
als
ca
n
d
ev
elo
p
m
o
d
els
th
a
t
h
elp
d
etec
t
d
is
ea
s
es
at
th
eir
ea
r
ly
s
tag
es,
wh
e
n
in
ter
v
en
tio
n
an
d
tr
ea
tm
e
n
t
ar
e
m
o
s
t
ef
f
ec
tiv
e
[
2
7
]
.
Fo
r
ex
am
p
le,
in
ca
n
ce
r
d
etec
ti
o
n
,
ML
alg
o
r
ith
m
s
ca
n
an
aly
z
e
m
ed
ical
im
ag
in
g
d
ata,
s
u
ch
as m
am
m
o
g
r
am
s
o
r
MRIs,
to
id
en
tify
s
u
s
p
icio
u
s
lesi
o
n
s
o
r
ab
n
o
r
m
alities
th
a
t
m
ay
r
eq
u
ir
e
f
u
r
th
er
in
v
esti
g
atio
n
[
2
8
]
.
W
h
ile
class
ical
ML
h
as
d
em
o
n
s
tr
at
ed
its
ef
f
icien
cy
,
th
e
p
r
e
v
ailin
g
ap
p
r
o
ac
h
am
o
n
g
r
esear
ch
er
s
is
d
ee
p
lear
n
in
g
(
DL
)
,
wh
ich
o
f
f
e
r
s
m
o
r
e
p
o
ten
t
tech
n
iq
u
es,
p
ar
ticu
lar
l
y
in
th
e
m
ed
ical
im
ag
i
n
g
d
o
m
ain
,
n
o
tab
ly
in
m
am
m
o
g
r
a
p
h
y
.
Dif
f
er
en
t
s
tu
d
ies
an
d
m
o
d
els
wer
e
d
ev
elo
p
ed
to
d
etec
t
b
r
ea
s
t
ca
n
ce
r
u
s
in
g
DL
tech
n
iq
u
es
[
2
9
]
−
[
3
1
]
.
R
esear
ch
er
s
u
s
ed
m
an
y
p
r
e
-
tr
ain
e
d
C
NN
alg
o
r
i
th
m
s
in
th
is
r
eg
a
r
d
lik
e
Alex
Net,
VGG,
R
esNet,
Go
o
g
leNe
t,
an
d
I
n
ce
p
tio
n
[
3
2
]
.
T
ab
le
2
s
u
m
m
ar
izes so
m
e
o
f
th
e
s
tu
d
ies co
n
d
u
cte
d
r
eg
a
r
d
in
g
b
r
ea
s
t c
an
ce
r
.
T
ab
le
2
.
Su
m
m
a
r
y
o
f
s
tu
d
ies i
n
b
r
ea
s
t c
an
ce
r
Ref
e
r
e
n
c
e
D
a
t
a
s
e
t
/
i
n
p
u
t
Ta
r
g
e
t
e
d
c
l
a
sses
A
l
g
o
r
i
t
h
ms
O
u
t
c
o
m
e
[
3
3
]
C
o
l
l
e
c
t
e
d
d
a
t
a
se
t
o
f
d
i
g
i
t
a
l
scr
e
e
n
i
n
g
mammo
g
r
a
ms
D
e
n
si
t
y
a
n
d
r
i
s
k
(
d
e
n
se,
n
o
d
e
n
s
e
,
l
o
w
r
i
sk
,
a
n
d
h
i
g
h
r
i
sk
)
Lo
g
i
s
t
i
c
r
e
g
r
e
ssi
o
n
,
R
e
sN
e
t
1
8
A
U
C
=
0
.
7
0
[
3
4
]
C
B
I
S
-
D
D
S
M
,
I
N
b
r
e
a
st
R
O
I
a
n
n
o
t
a
t
i
o
n
s
V
G
G
1
6
,
R
e
sN
e
t
5
0
A
U
C
(
D
D
S
M
)
=
0
.
8
8
A
U
C
(
I
N
b
r
e
a
st
)
=
0
.
9
5
[
3
5
]
1
5
mi
c
r
o
c
a
l
c
i
f
i
c
a
t
i
o
n
s
f
e
a
t
u
r
e
s
a
n
d
2
6
b
r
e
a
s
t
m
a
sses
f
e
a
t
u
r
e
s
B
r
e
a
s
t
l
e
si
o
n
s
a
n
d
mi
c
r
o
c
a
l
c
i
f
i
c
a
t
i
o
n
s
A
st
a
c
k
e
d
a
u
t
o
e
n
c
o
d
e
r
(
S
A
E)
w
i
t
h
n
l
a
y
e
r
s,
S
V
M
A
c
c
u
r
a
c
y
=
8
5
.
8
%
[
3
6
]
I
N
b
r
e
a
st
M
a
l
i
g
n
a
n
t
o
r
b
e
n
i
g
n
l
e
si
o
n
s
F
a
st
e
r
R
-
C
N
N
A
U
C
=
0
.
9
5
[
3
7
]
M
I
A
S
,
I
N
b
r
e
a
st
,
I
mag
e
N
e
t
,
a
n
d
p
r
i
v
a
t
e
d
a
t
a
b
a
s
e
s
N
o
r
mal
a
n
d
s
u
sp
e
c
t
e
d
R
O
I
(
1
)
C
o
n
v
N
e
t
+
S
V
M
(
2
)
V
G
G
N
e
t
1
6
,
(
3
)
V
G
G
N
e
t
1
9
,
(
4
)
G
o
o
g
L
e
N
e
t
,
(
5
)
M
o
b
i
l
e
N
e
t
V
2
,
(
6
)
R
e
sN
e
t
5
0
,
(
7
)
D
e
n
seN
e
t
1
2
1
A
U
C
:
(
1
)
9
1
.
4
,
(
2
)
8
6
.
4
,
(
3
)
8
7
.
6
,
(
4
)
7
9
.
4
,
(
5
)
6
6
.
8
,
(
6
)
7
5
.
8
,
(
7
)
8
1
.
7
[
3
8
]
M
I
A
S
a
n
d
p
r
i
v
a
t
e
d
a
t
a
se
t
B
e
n
i
g
n
a
n
d
m
a
l
i
g
n
a
n
t
R
O
I
C
o
n
v
N
e
t
A
U
C
=
0
.
9
9
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
36
,
No
.
2
,
No
v
em
b
er
20
24
:
1
1
4
9
-
1
1
6
2
1152
I
n
p
r
e
v
io
u
s
ly
m
en
tio
n
ed
s
tu
d
i
es,
th
e
d
atasets
u
s
ed
ar
e
co
m
m
o
n
p
u
b
lic
d
atasets
lik
e
MI
AS,
DDSM,
an
d
I
Nb
r
ea
s
t.
I
n
th
is
s
tu
d
y
,
d
i
f
f
er
en
t
p
r
e
-
tr
ain
e
d
C
NN
m
o
d
e
ls
wer
e
im
p
lem
en
ted
o
n
a
lo
c
al
f
ir
s
t
-
h
an
d
d
ataset
f
r
o
m
t
h
e
W
est
B
an
k
in
Palest
i
n
e.
T
h
e
m
o
d
els
ar
e:
VGG
1
6
,
VGG1
9
,
Den
s
eNe
t1
2
1
,
Xce
p
t
io
n
,
R
esNet5
0
,
an
d
E
f
f
icen
tNetB
7
.
T
h
e
m
ain
task
was
to
d
etec
t
ab
n
o
r
m
alities
in
th
e
b
r
ea
s
t
an
d
class
if
y
th
em
in
to
n
o
r
m
al
an
d
ab
n
o
r
m
al.
T
h
is
co
n
tr
i
b
u
tio
n
will
ad
d
v
alu
e
t
o
h
ea
lth
ca
r
e
in
c
lin
ical
u
s
e
s
in
ce
th
er
e
is
n
o
s
u
ch
co
m
p
u
te
r
aid
e
d
d
etec
tio
n
(
C
AD)
s
y
s
tem
o
r
d
e
cisi
o
n
s
u
p
p
o
r
t
s
y
s
tem
(
DSS)
im
p
lem
en
ted
in
t
h
e
m
a
m
m
o
g
r
am
d
ep
a
r
tm
en
ts
at
th
e
Palest
in
ian
Mo
H
to
g
iv
e
t
h
e
clin
ician
s
a
s
ec
o
n
d
o
p
in
io
n
wh
ich
will
im
p
r
o
v
e
th
e
ea
r
ly
d
etec
tio
n
o
f
b
r
ea
s
t
ab
n
o
r
m
ality
.
T
h
e
f
o
llo
win
g
s
ec
tio
n
s
p
r
e
s
en
t
th
e
m
eth
o
d
o
lo
g
y
f
o
r
im
p
lem
en
tin
g
th
e
d
ee
p
lear
n
in
g
m
o
d
els
to
d
iag
n
o
s
e
b
r
ea
s
t
ab
n
o
r
m
alitie
s
an
d
class
if
y
th
em
in
to
n
o
r
m
al
an
d
ab
n
o
r
m
al.
T
h
e
m
e
th
o
d
o
lo
g
y
in
clu
d
es
d
escr
ib
in
g
th
e
co
llected
d
ataset
an
d
h
o
w
th
e
im
a
g
es
wer
e
p
r
e
-
p
r
o
ce
s
s
ed
.
T
h
e
co
m
in
g
s
ec
tio
n
s
also
illu
s
tr
ate
s
o
m
e
ex
tr
ac
ted
in
s
ig
h
ts
f
r
o
m
th
e
d
ata
in
th
e
ex
p
lo
r
at
o
r
y
d
ata
an
aly
s
is
(
E
DA)
s
tag
e.
I
n
ad
d
itio
n
,
th
e
im
p
lem
en
ted
m
o
d
els in
th
is
s
t
u
d
y
b
esid
es th
e
p
er
f
o
r
m
a
n
ce
m
ea
s
u
r
es a
r
e
p
r
esen
ted
.
Fin
ally
,
th
e
r
esu
lts
will b
e
p
r
esen
ted
an
d
d
is
cu
s
s
ed
in
a
d
d
itio
n
to
th
e
c
o
n
clu
s
io
n
a
n
d
f
u
tu
r
e
wo
r
k
.
4.
M
E
T
H
O
D
T
h
is
s
ec
tio
n
p
r
esen
ts
th
e
m
et
h
o
d
o
lo
g
y
u
s
ed
in
th
is
s
tu
d
y
to
d
etec
t
b
r
ea
s
t
ab
n
o
r
m
alities
an
d
class
if
y
th
em
in
to
n
o
r
m
al
an
d
ab
n
o
r
m
al.
Fig
u
r
e
2
s
h
o
ws
th
e
o
u
tlin
e
wo
r
k
f
lo
w
f
o
r
th
e
m
eth
o
d
o
lo
g
y
an
d
Fig
u
r
e
3
ill
u
s
tr
ates
th
e
m
o
d
ellin
g
ap
p
r
o
ac
h
.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
co
n
s
is
ts
o
f
s
ix
s
tag
es
to
ad
d
r
ess
th
e
r
esear
ch
o
b
jectiv
e.
T
h
e
f
ir
s
t
s
tag
e
talk
s
ab
o
u
t
th
e
d
ataset
an
d
h
o
w
it’
s
lab
eled
wh
ile
th
e
s
ec
o
n
d
s
tag
e
is
ab
o
u
t
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
in
clu
d
in
g
th
e
tech
n
iq
u
es
u
s
ed
t
o
p
r
ep
ar
e
th
e
m
a
m
m
o
g
r
am
im
a
g
es
to
b
e
r
ea
d
y
f
o
r
th
e
n
ex
t
s
tag
e.
Ad
d
itio
n
ally
,
s
tag
e
th
r
ee
ex
tr
ac
ts
th
e
in
s
ig
h
ts
f
r
o
m
th
e
d
ataset
an
d
s
tag
e
f
o
u
r
in
clu
d
es
s
p
litt
in
g
th
e
d
ataset
in
to
s
u
b
s
ets.
I
n
th
e
m
o
d
ellin
g
s
tag
e,
th
e
d
ataset
is
f
ed
in
to
s
ix
C
NN
m
o
d
els
w
h
ile
in
th
e
la
s
t
s
tag
e,
th
e
m
o
d
els
ar
e
test
ed
u
s
in
g
d
if
f
e
r
en
t p
e
r
f
o
r
m
an
ce
m
ea
s
u
r
es.
All stag
es a
r
e
p
r
esen
ted
in
th
e
f
o
llo
win
g
s
u
b
s
ec
tio
n
s
.
Fig
u
r
e
2
.
R
esear
ch
m
eth
o
d
o
l
o
g
y
co
n
d
u
cted
in
th
is
r
esear
ch
Fig
u
r
e
3
.
T
h
e
m
o
d
ellin
g
ap
p
r
o
ac
h
b
lo
ck
d
iag
r
am
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
C
o
n
vo
lu
tio
n
a
l
n
eu
r
a
l n
etw
o
r
ks b
r
ea
s
t c
a
n
ce
r
cla
s
s
ifica
tio
n
u
s
in
g
…
(
Ha
n
in
S
a
a
d
a
h
)
1153
4
.
1
.
Da
t
a
s
et
T
h
e
f
ir
s
t
s
tag
e
is
d
ata
co
llectio
n
,
th
e
d
ata
was
c
o
llected
f
r
o
m
th
e
B
eth
leh
em
Ma
m
m
o
g
r
a
p
h
y
C
en
ter
.
I
t
is
o
n
e
o
f
th
e
ce
n
te
r
s
o
f
th
e
Palest
in
ian
Min
is
tr
y
o
f
Hea
lth
th
at
p
o
s
s
ess
es
th
e
n
ec
ess
ar
y
d
ata
th
at
m
ee
ts
th
e
co
n
d
itio
n
s
f
o
r
c
o
n
d
u
ctin
g
s
u
c
h
s
tu
d
ies.
A
to
tal
o
f
8
1
3
m
a
m
m
o
g
r
am
im
ag
es
wer
e
co
llec
ted
f
r
o
m
1
7
4
lad
ies
wh
o
wer
e
s
cr
ee
n
ed
at
th
e
ce
n
t
er
b
ef
o
r
e
J
u
ly
2
0
2
3
.
T
h
e
co
llected
d
ataset
in
clu
d
ed
im
ag
es
tak
en
f
r
o
m
b
o
th
b
r
ea
s
t
s
id
es.
E
ac
h
s
id
e
is
im
ag
ed
in
d
if
f
er
en
t
o
r
ien
tatio
n
s
with
two
d
if
f
er
en
t
v
iews n
am
ely
m
ed
io
later
al
o
b
liq
u
e
(
ML
O)
an
d
cr
an
ial
ca
u
d
al
(
C
C
)
.
T
h
e
ML
O
v
iew
is
tak
en
f
r
o
m
th
e
ce
n
ter
o
f
th
e
ch
est
o
u
twar
d
,
wh
er
e
as
th
e
C
C
v
iew
i
s
tak
en
f
r
o
m
ab
o
v
e
th
e
b
r
ea
s
t.
T
h
e
ML
O
v
iew
o
f
f
e
r
s
a
m
o
r
e
ex
ten
s
iv
e
p
er
s
p
ec
tiv
e
o
f
t
h
e
u
p
p
e
r
-
o
u
ter
q
u
a
d
r
an
t,
p
r
o
v
id
in
g
an
o
p
tim
al
v
is
u
aliza
tio
n
o
f
t
h
e
b
r
ea
s
t'
s
la
ter
al
s
id
e
.
T
h
e
im
a
g
es
wer
e
c
o
n
v
er
ted
b
y
th
e
ce
n
ter
s
taf
f
f
r
o
m
DI
C
OM
to
J
PG
f
o
r
m
at
with
a
s
tan
d
ar
d
s
ize
o
f
4
,
7
1
0
×5
,
8
4
4
p
ix
els
an
d
wer
e
lab
eled
u
s
in
g
a
B
I
-
R
ADS
s
co
r
e
f
r
o
m
1
to
6
ac
co
r
d
in
g
to
th
e
d
o
cto
r
’
s
f
in
d
i
n
g
s
.
4
.
2
.
P
re
-
p
ro
ce
s
s
ing
I
n
th
e
s
ec
o
n
d
s
tag
e,
t
h
e
im
ag
es
wer
e
lab
eled
as
n
o
r
m
al
an
d
ab
n
o
r
m
al
in
s
tead
o
f
B
I
-
R
AD
s
co
r
es.
I
m
ag
es
o
f
s
co
r
es
1
to
3
wer
e
lab
eled
as
n
o
r
m
al,
an
d
th
e
i
m
ag
es
o
f
s
co
r
es
4
to
6
as
ab
n
o
r
m
al.
Af
ter
th
at,
a
r
an
d
o
m
s
am
p
le
was
tak
en
f
r
o
m
th
e
m
ain
d
ataset.
T
h
e
s
am
p
le
in
clu
d
ed
2
0
0
im
a
g
es
ca
teg
o
r
ized
in
to
tw
o
g
r
o
u
p
s
;
1
0
0
f
o
r
t
h
e
n
o
r
m
al
g
r
o
u
p
,
an
d
1
0
0
f
o
r
th
e
a
b
n
o
r
m
a
l
g
r
o
u
p
.
I
n
ad
d
itio
n
,
a
s
m
all
s
u
b
s
et
o
f
1
8
im
ag
es
(
1
0
n
o
r
m
al,
8
ab
n
o
r
m
al)
wer
e
also
tak
en
f
r
o
m
th
e
o
r
ig
in
al
d
ataset
to
b
e
f
ed
in
to
th
e
tr
ain
ed
m
o
d
els
f
o
r
e
x
ter
n
al
test
p
u
r
p
o
s
es.
T
h
e
last
d
ataset
ac
ts
as th
e
ex
ter
n
al
d
ataset.
Mo
r
eo
v
er
,
th
e
im
ag
es
in
th
e
r
an
d
o
m
s
am
p
le
wen
t
th
r
o
u
g
h
d
if
f
er
en
t
tech
n
iq
u
es
s
h
o
wn
in
Fig
u
r
e
4
t
o
r
ed
u
ce
th
e
n
o
is
e
an
d
clea
n
th
e
im
ag
es
to
p
r
ep
ar
e
th
em
f
o
r
t
h
e
n
ex
t
s
tag
e.
T
h
ese
tech
n
iq
u
es
ar
e:
1
)
Ma
n
u
ally
cr
o
p
p
in
g
tex
t
an
d
p
ec
to
r
al
m
u
s
cle.
2
)
C
o
n
v
er
t
im
ag
es
f
r
o
m
R
GB
to
Gr
ay
s
ca
le
to
r
ed
u
ce
th
e
d
im
en
s
io
n
s
an
d
en
h
an
ce
co
m
p
u
tatio
n
al
tim
e.
3
)
R
esizin
g
f
r
o
m
th
e
o
r
ig
in
al
s
ize
to
2
2
4
×
2
2
4
to
b
e
co
m
p
at
ib
le
with
th
e
in
p
u
t
s
ize
o
f
th
e
p
r
e
-
tr
ain
ed
m
o
d
el
u
s
ed
later
.
4
)
No
r
m
alizin
g
i
m
ag
es
to
g
et
a
s
tan
d
ar
d
r
an
g
e
f
r
o
m
0
to
1
t
o
s
im
p
lify
th
e
p
r
o
ce
s
s
.
5
)
Fin
ally
,
au
g
m
e
n
tatio
n
was
im
p
lem
en
ted
o
n
th
e
r
an
d
o
m
s
am
p
le
to
r
e
d
u
ce
o
v
e
r
f
itti
n
g
th
at
m
ig
h
t
h
ap
p
e
n
in
th
e
s
m
all
d
ataset.
T
h
e
a
u
g
m
en
tatio
n
s
tep
r
esu
lted
in
cr
ea
tin
g
1
0
im
ag
es
f
r
o
m
ea
ch
im
ag
e
s
o
th
e
f
in
al
n
u
m
b
e
r
o
f
im
ag
es
was
1
,
000
f
o
r
n
o
r
m
a
l
an
d
1
,
000
f
o
r
a
b
n
o
r
m
al.
T
h
e
to
o
ls
u
s
ed
in
th
is
s
tag
e
an
d
th
e
n
ex
t
s
tag
es
ar
e
Py
th
o
n
an
d
J
u
p
y
ter
No
teb
o
o
k
an
d
d
if
f
er
e
n
t
lib
r
ar
ies
s
u
ch
as
C
V2
,
Ker
as,
T
en
s
o
r
f
lo
w,
an
d
Sk
lear
n
.
T
h
e
s
e
lib
r
ar
ies ar
e
co
m
m
o
n
ly
u
s
e
d
in
class
if
icatio
n
an
d
p
r
ed
icti
o
n
task
s
.
As
a
r
esu
lt
o
f
th
is
s
tag
e,
2
,
0
0
0
p
r
ep
r
o
ce
s
s
ed
im
ag
es
wer
e
cr
ea
ted
f
r
o
m
th
e
im
ag
es
in
th
e
r
an
d
o
m
d
ataset;
1
0
im
a
g
es
f
r
o
m
ea
ch
im
ag
e.
Fig
u
r
e
5
is
a
p
ictu
r
e
th
a
t
s
h
o
ws
a
s
am
p
le
im
ag
e
f
r
o
m
th
e
r
a
n
d
o
m
d
atase
t
b
ef
o
r
e
an
d
af
ter
p
r
e
p
r
o
ce
s
s
in
g
.
T
h
is
p
ictu
r
e
in
clu
d
es
3
r
o
ws
o
f
im
ag
es.
T
h
e
im
ag
e
lo
ca
ted
in
th
e
f
ir
s
t
r
o
w
o
n
th
e
lef
t
is
th
e
o
r
i
g
in
al
im
a
g
e
b
ef
o
r
e
p
r
e
-
p
r
o
ce
s
s
in
g
.
T
h
e
i
m
ag
e
in
clu
d
ed
an
n
o
tatio
n
s
s
u
ch
as
th
e
p
atien
t'
s
d
etails,
th
e
v
iew
m
o
d
e
(
ML
O
)
,
o
th
er
u
n
wa
n
ted
tex
t,
an
d
th
e
p
ec
to
r
al
m
u
s
cle.
I
n
ad
d
itio
n
,
th
e
im
ag
e
o
n
th
e
to
p
r
i
g
h
t
is
af
ter
cr
o
p
p
in
g
t
h
e
an
n
o
tatio
n
s
,
th
e
p
ec
t
o
r
al
m
u
s
c
le,
an
d
co
n
v
er
tin
g
th
e
c
o
lo
r
to
g
r
ay
s
ca
le.
Fin
ally
,
th
e
ten
im
a
g
es
in
t
h
e
s
ec
o
n
d
an
d
th
ir
d
r
o
ws
ar
e
r
esized
i
m
ag
es
in
to
2
2
4
×
2
2
4
an
d
th
e
n
n
o
r
m
alize
d
,
an
d
au
g
m
en
ted
.
Mo
r
eo
v
e
r
,
th
e
im
ag
es
in
th
e
s
ec
o
n
d
a
n
d
th
ir
d
r
o
ws
ar
e
th
e
1
0
im
ag
es
th
at
w
er
e
cr
ea
ted
f
r
o
m
th
e
o
r
ig
in
al
in
t
h
e
au
g
m
en
tatio
n
p
r
o
ce
s
s
b
y
u
s
in
g
tech
n
iq
u
es
lik
e
r
o
tatio
n
,
wid
th
an
d
h
eig
h
t
s
h
if
t,
s
h
ea
r
,
zo
o
m
,
h
o
r
izo
n
tal
a
n
d
v
e
r
tical
f
lip
.
Fig
u
r
e
4
.
Data
p
r
e
-
p
r
o
ce
s
s
in
g
m
eth
o
d
o
l
o
g
y
i
n
th
is
r
esear
ch
4
.
3
.
E
x
plo
ra
t
o
ry
da
t
a
a
na
l
y
s
is
T
h
is
s
tag
e
h
elp
s
in
u
n
d
e
r
s
tan
d
in
g
th
e
r
aw
d
ata
an
d
co
n
v
er
t
in
g
it
to
u
s
ef
u
l
in
f
o
r
m
atio
n
a
n
d
in
s
ig
h
ts
th
at
ca
n
b
e
u
s
ed
in
m
ak
in
g
im
p
o
r
tan
t
d
ec
is
io
n
s
.
Dif
f
er
en
t
v
is
u
aliza
tio
n
s
h
av
e
b
ee
n
ex
tr
ac
t
ed
f
r
o
m
th
e
d
ataset
u
s
ed
in
th
is
s
tu
d
y
.
Sev
en
co
l
u
m
n
s
wer
e
u
s
ed
to
d
escr
ib
e
th
e
d
ata;
B
I
R
AD,
s
id
e,
y
ea
r
o
f
b
ir
th
,
ag
e
in
y
ea
r
s
,
v
iew,
class
,
an
d
ag
e
g
r
o
u
p
.
B
I
R
AD
i
s
th
e
g
lo
b
al
s
co
r
e
f
o
r
m
am
m
o
g
r
am
s
f
r
o
m
1
to
6
,
th
e
s
id
e
in
d
icate
s
wh
eth
er
th
e
im
ag
e
is
tak
en
f
r
o
m
a
lad
y
th
at
h
as
ab
n
o
r
m
alit
ies
o
n
th
e
lef
t,
r
ig
h
t,
o
r
b
o
th
s
id
es
o
f
th
e
b
r
ea
s
t.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
36
,
No
.
2
,
No
v
em
b
er
20
24
:
1
1
4
9
-
1
1
6
2
1154
T
h
e
v
iew
in
d
icate
s
a
C
C
o
r
ML
O
v
iew,
clas
s
is
th
e
tar
g
et
co
lu
m
n
th
at
s
ay
s
if
th
e
im
ag
e
is
n
o
r
m
al
o
r
ab
n
o
r
m
al,
an
d
f
in
ally
ag
e
g
r
o
u
p
ca
teg
o
r
izes
th
e
im
a
g
es
in
to
6
g
r
o
u
p
s
wh
ich
a
r
e:
30
-
3
9
,
4
0
-
4
9
,
5
0
-
5
9
,
6
0
-
6
9
,
70
-
7
9
,
8
0
-
89
.
Fig
u
r
es
6
t
o
8
ill
u
s
tr
ate
s
o
m
e
o
f
th
e
m
o
s
t im
p
o
r
tan
t in
s
ig
h
ts
.
Fig
u
r
e
6
is
a
ch
ar
t
to
illu
s
tr
a
te
th
e
d
is
tr
ib
u
tio
n
o
f
s
cr
ee
n
e
d
im
ag
es
am
o
n
g
th
e
ag
e
g
r
o
u
p
an
d
t
h
e
n
o
r
m
ality
class
.
T
h
e
m
ajo
r
ity
o
f
s
cr
ee
n
ed
im
ag
es
ar
e
r
elat
ed
to
ag
e
g
r
o
u
p
5
0
to
5
9
.
T
h
is
m
ea
n
s
th
at
t
h
e
lar
g
est
g
r
o
u
p
o
f
wo
m
en
wh
o
h
av
e
u
n
d
e
r
g
o
n
e
a
m
am
m
o
g
r
am
ar
e
b
etwe
en
th
e
ag
es
o
f
f
if
t
y
an
d
f
if
ty
-
n
in
e.
T
h
is
ag
e
g
r
o
u
p
s
h
o
ws
a
clo
s
e
n
u
m
b
e
r
o
f
im
ag
es
b
etwe
en
t
h
e
n
o
r
m
al
an
d
a
b
n
o
r
m
al
class
es
wh
er
e
th
er
e
ar
e
4
3
im
a
g
es
in
th
e
ab
n
o
r
m
al
an
d
4
1
in
th
e
n
o
r
m
al
.
On
th
e
o
t
h
er
h
a
n
d
,
th
er
e
ar
e
a
s
m
all
n
u
m
b
er
o
f
w
o
m
en
in
th
e
s
am
p
le
w
h
o
ar
e
u
n
d
e
r
t
h
e
ag
e
o
f
th
i
r
ty
-
n
in
e
an
d
t
h
o
s
e
w
h
o
a
r
e
o
v
er
s
ev
en
ty
.
T
h
ese
ag
e
g
r
o
u
p
s
a
r
e
s
h
o
wn
m
o
r
e
in
th
e
a
b
n
o
r
m
al
class
th
an
th
e
n
o
r
m
al.
T
h
is
m
ea
n
s
th
at
th
e
ab
n
o
r
m
alities
ar
e
u
s
u
ally
d
ev
el
o
p
ed
in
wo
m
en
in
o
l
d
er
ag
es.
Fig
u
r
e
5
.
Ma
m
m
o
g
r
a
m
im
ag
e
b
ef
o
r
e
a
n
d
af
te
r
p
r
e
-
p
r
o
c
ess
in
g
Fig
u
r
e
6
.
Nu
m
b
er
o
f
s
cr
ee
n
e
d
im
ag
es v
er
s
u
s
class
an
d
ag
e
g
r
o
u
p
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
C
o
n
vo
lu
tio
n
a
l
n
eu
r
a
l n
etw
o
r
ks b
r
ea
s
t c
a
n
ce
r
cla
s
s
ifica
tio
n
u
s
in
g
…
(
Ha
n
in
S
a
a
d
a
h
)
1155
Mo
r
eo
v
er
,
Fig
u
r
e
7
r
ep
r
esen
ts
a
ch
ar
t
o
f
t
h
e
n
u
m
b
er
o
f
s
cr
ee
n
ed
im
ag
es
v
er
s
u
s
class
an
d
s
id
e.
I
t
ca
n
b
e
n
o
ted
th
at
th
e
lar
g
est
n
u
m
b
er
o
f
im
ag
es
wer
e
tak
en
f
r
o
m
wo
m
en
with
th
e
p
o
s
s
ib
ilit
y
o
f
ca
n
ce
r
ce
lls
o
n
th
e
lef
t
s
id
e
o
f
t
h
e
b
r
ea
s
t,
an
d
th
i
s
g
r
o
u
p
co
n
s
titu
tes
h
alf
o
f
th
e
s
am
p
le,
wh
ile
th
e
n
u
m
b
er
o
f
im
ag
es
tak
en
f
r
o
m
wo
m
en
with
t
h
e
p
o
s
s
ib
ilit
y
o
f
ca
n
ce
r
ce
lls
o
n
th
e
r
ig
h
t
s
id
e
is
less
.
Als
o
,
th
er
e
ar
e
a
g
o
o
d
n
u
m
b
er
o
f
im
a
g
es
t
ak
en
f
r
o
m
wo
m
e
n
s
u
f
f
er
in
g
f
r
o
m
p
o
s
s
ib
le
ca
n
ce
r
ce
lls
o
n
b
o
th
s
id
es
o
f
th
e
b
r
ea
s
t.
B
o
th
cl
ass
es
s
h
o
w
a
clo
s
e
n
u
m
b
er
o
f
im
a
g
es tak
en
o
n
ea
ch
s
id
e.
Similar
ly
,
th
e
ch
ar
t
in
Fig
u
r
e
8
s
h
o
ws
th
e
n
u
m
b
er
o
f
im
ag
es
tak
en
f
r
o
m
th
e
C
C
an
d
ML
O
v
iews
an
d
h
o
w
th
ey
a
r
e
d
is
tr
ib
u
ted
b
etw
ee
n
th
e
n
o
r
m
al
an
d
ab
n
o
r
m
al
class
es.
I
n
th
e
a
b
n
o
r
m
al
s
am
p
le,
th
e
n
u
m
b
er
s
o
f
th
e
C
C
an
d
ML
O
im
ag
es
a
r
e
alm
o
s
t
eq
u
al
w
h
er
e
th
e
r
e
ar
e
5
1
an
d
4
9
im
ag
es
f
o
r
th
e
C
C
an
d
ML
O
r
esp
ec
tiv
e
ly
.
L
ik
ewise,
in
th
e
n
o
r
m
al
s
am
p
le,
t
h
er
e
is
a
n
o
ticea
b
le
d
if
f
er
e
n
ce
in
th
e
n
u
m
b
er
s
,
as
th
e
n
u
m
b
e
r
o
f
im
ag
es in
th
e
C
C
v
iew
was
h
ig
h
er
a
n
d
h
a
d
5
7
im
ag
es wh
i
le
th
e
ML
O
v
iew
h
ad
4
3
im
a
g
es.
Fig
u
r
e
7
.
Nu
m
b
er
o
f
s
cr
ee
n
e
d
im
ag
es v
er
s
u
s
class
an
d
s
id
e
Fig
u
r
e
8
.
Nu
m
b
er
o
f
s
cr
ee
n
e
d
im
ag
es v
er
s
u
s
class
an
d
v
iew
4
.
4
.
Da
t
a
s
pli
t
t
ing
T
h
e
p
r
o
ce
s
s
o
f
s
p
litt
in
g
th
e
d
ata
was
in
th
e
f
o
u
r
th
s
tag
e
w
h
er
e
th
e
2
,
0
0
0
im
a
g
es
af
ter
au
g
m
en
tatio
n
wer
e
s
p
lit
in
to
two
s
u
b
s
ets.
T
h
e
f
ir
s
t
s
u
b
s
et
is
u
s
ed
to
tr
ain
th
e
m
o
d
el
an
d
co
n
s
titu
tes
8
0
%
o
f
th
e
d
ataset,
an
d
2
0
% f
o
r
t
h
e
test
in
g
s
u
b
s
e
t w
h
ich
is
u
s
ed
to
test
th
e
r
ea
l a
cc
u
r
ac
y
an
d
p
er
f
o
r
m
an
ce
o
f
th
e
m
o
d
el.
4
.
5
.
M
o
delin
g
I
n
th
is
p
a
p
er
,
a
tr
an
s
f
er
lea
r
n
i
n
g
ap
p
r
o
ac
h
was
f
o
llo
wed
an
d
s
ix
DL
p
r
e
-
tr
ain
ed
m
o
d
els
wer
e
u
s
ed
.
T
h
e
tr
ain
in
g
s
u
b
s
et
was th
e
in
p
u
t f
o
r
all
m
o
d
els.
T
h
e
f
o
llo
w
in
g
s
u
b
s
ec
tio
n
s
b
r
ief
ly
p
r
esen
t
th
e
m
o
d
els u
s
ed
in
th
is
r
esear
ch
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
36
,
No
.
2
,
No
v
em
b
er
20
24
:
1
1
4
9
-
1
1
6
2
1156
4
.
6
.
1
.
DenseNet
1
2
1
Den
s
ely
co
n
n
ec
ted
co
n
v
o
lu
tio
n
al
n
etwo
r
k
(
Den
s
eNe
t)
is
o
n
e
o
f
th
e
C
NN
m
o
d
els
th
at
ac
h
iev
ed
h
ig
h
p
er
f
o
r
m
an
ce
in
t
h
e
liter
atu
r
e
co
m
p
ar
ed
to
o
th
e
r
p
r
e
-
tr
ain
e
d
m
o
d
els.
I
t’
s
c
o
n
s
tr
u
cted
f
r
o
m
1
2
0
c
o
n
v
o
lu
tio
n
s
an
d
4
Av
g
Po
o
l
an
d
in
clu
d
e
s
f
o
u
r
co
m
p
o
n
en
ts
wh
ich
ar
e
co
n
n
ec
tiv
ity
,
Den
s
eBl
o
ck
s
,
g
r
o
wth
r
ate,
an
d
b
o
ttlen
ec
k
l
ay
er
s
.
Den
s
eNe
t
tak
es
th
e
in
p
u
t
f
r
o
m
all
p
r
ev
i
o
u
s
lay
er
s
’
co
n
n
ec
ted
o
u
tp
u
ts
to
th
e
d
en
s
e
b
lo
ck
lay
er
[
3
9
]
.
Fig
u
r
e
9
p
r
esen
ts
t
h
e
s
ch
em
atic
lay
o
u
t
o
f
t
h
e
De
n
s
eNe
t
[
4
0
]
.
T
h
e
co
n
n
ec
tiv
ity
p
atter
n
is
th
e
m
ai
n
id
ea
b
eh
i
n
d
t
h
e
Den
s
eNe
t
s
tr
u
ctu
r
e
wh
e
r
e
th
e
f
ea
tu
r
e
m
a
p
s
f
r
o
m
p
r
e
v
io
u
s
lay
e
r
s
ar
e
c
o
n
ca
ten
ated
o
n
to
th
e
in
p
u
ts
o
f
f
u
tu
r
e
lay
er
s
.
C
o
n
tin
u
ally
co
n
ca
ten
atin
g
r
esu
lts
in
v
er
y
d
ee
p
in
p
u
ts
.
T
h
e
r
e
ar
e
m
an
y
ac
tiv
atio
n
f
u
n
ctio
n
s
s
u
ch
as
r
ec
tifie
d
lin
ea
r
u
n
it
(
R
eL
U
)
ar
e
u
tili
ze
d
t
o
in
cr
ea
se
th
e
n
o
n
-
lin
ea
r
ity
in
th
e
p
o
o
lin
g
lay
er
s
wh
en
th
e
f
ea
t
u
r
e
m
a
p
s
ar
e
f
e
d
in
to
th
e
co
m
in
g
lay
er
s
.
In
(
1
)
illu
s
tr
ates
th
e
m
ath
em
atic
al
f
o
r
m
u
la
f
o
r
th
is
s
ch
em
a.
=
(
[
0
,
1
,
…
,
−
1
]
)
(
1
)
W
h
er
e
l
th
is
th
e
lay
er
th
at
r
ec
eiv
es
th
e
co
n
ca
ten
ated
f
ea
tu
r
e
m
ap
s
x
0
,
x
1
,
...,
x
l
-
1
f
r
o
m
all
p
r
e
v
io
u
s
lay
er
s
.
H
l
is
a
co
m
p
o
s
ite
f
u
n
ctio
n
o
f
o
p
er
ati
o
n
s
lik
e
b
atch
n
o
r
m
aliza
tio
n
(
B
N)
,
a
R
eL
U
,
an
d
a
3
×3
co
n
v
o
lu
tio
n
(
c
o
n
v
)
th
at
g
en
er
ates
th
e
K
lev
els
f
o
r
th
e
m
ap
p
in
g
f
ea
t
u
r
es
in
th
e
co
m
in
g
lay
er
wh
ich
d
eter
m
in
es
th
e
g
r
o
wth
r
ate
in
th
e
n
etwo
r
k
.
T
h
e
g
r
o
wth
r
ate
K
is
ca
lcu
lated
in
(
2
)
.
=
0
+
∗
(
−
1
)
(
2
)
W
h
er
e
k
0
is
th
e
n
u
m
b
e
r
o
f
c
h
a
n
n
els in
th
e
in
p
u
t la
y
er
.
Fig
u
r
e
9
.
Den
s
eNe
t
s
ch
em
atic
lay
o
u
t
4
.
6
.
2
.
VG
G
1
6
a
nd
VG
G
1
9
VGG
s
tan
d
s
f
o
r
v
is
u
al
g
eo
m
etr
y
g
r
o
u
p
.
VGG1
6
an
d
VGG1
9
ar
e
also
C
N
N
m
o
d
els
th
at
ar
e
co
m
m
o
n
l
y
u
s
ed
with
im
ag
e
class
if
icatio
n
an
d
d
etec
tio
n
.
T
h
eir
ar
c
h
itectu
r
e
is
th
e
s
am
e
an
d
th
e
o
n
l
y
d
if
f
er
en
ce
is
th
e
n
u
m
b
er
o
f
lay
er
s
.
VGG1
6
is
c
o
n
s
tr
u
cted
f
r
o
m
1
3
c
o
n
v
o
lu
tio
n
al
la
y
er
s
an
d
3
f
u
lly
co
n
n
ec
te
d
lay
er
s
wh
ile
in
VGG1
9
th
er
e
ar
e
1
6
co
n
v
o
lu
tio
n
al
lay
er
s
an
d
3
f
u
lly
co
n
n
ec
ted
[
4
1
]
.
T
h
e
im
ag
e
s
h
ap
e
s
h
o
u
ld
b
e
2
2
4
×
2
4
4
in
b
o
th
m
o
d
els.
4
.
6
.
3
.
ResNet
5
0
R
esid
u
al
n
etwo
r
k
;
R
esNet
is
a
C
NN
m
o
d
el
th
at
is
u
s
ed
to
d
etec
t
o
b
jects.
I
t
was
th
e
r
esu
lt
o
f
ad
d
in
g
r
esid
u
al
s
tr
u
ctu
r
e
to
th
e
C
NN
to
f
ix
th
e
g
r
ad
ien
t
d
eg
r
a
d
atio
n
an
d
d
is
ap
p
ea
r
an
ce
is
s
u
e
th
at
f
r
eq
u
en
tly
o
cc
u
r
s
in
th
e
tr
ain
in
g
p
r
o
ce
s
s
[
4
2
]
.
R
esNet5
0
co
n
tain
s
5
0
lay
er
s
th
at
ar
e
d
is
tr
ib
u
ted
o
v
e
r
f
o
u
r
ty
p
es
o
f
lay
e
r
s
wh
ich
ar
e
co
n
v
o
lu
tio
n
al,
p
o
o
lin
g
,
f
u
lly
co
n
n
ec
ted
,
an
d
s
h
o
r
tcu
t
[
4
3
]
.
I
t
ca
n
p
er
f
o
r
m
ef
f
icien
tl
y
ev
en
with
a
lar
g
e
n
u
m
b
er
o
f
la
y
er
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
C
o
n
vo
lu
tio
n
a
l
n
eu
r
a
l n
etw
o
r
ks b
r
ea
s
t c
a
n
ce
r
cla
s
s
ifica
tio
n
u
s
in
g
…
(
Ha
n
in
S
a
a
d
a
h
)
1157
4
.
6
.
4
.
E
f
f
icient
Net
B
7
E
f
f
ic
ien
tNets
is
a
f
am
ily
o
f
m
o
d
els
th
at
o
b
tain
ed
b
etter
e
f
f
icien
cy
a
n
d
ac
c
u
r
ac
y
th
an
t
h
e
p
r
e
v
io
u
s
C
NNs.
E
f
f
icien
tNet
-
B7
in
p
ar
ticu
lar
,
ac
h
iev
ed
to
p
-
1
ac
cu
r
ac
y
o
n
I
m
ag
eNe
t o
f
all
E
f
f
icien
t
Net
f
am
ilies
[
4
4
]
.
4
.
6
.
5
.
Xce
ptio
n
T
h
e
ex
tr
em
e
in
ce
p
tio
n
;
Xce
p
t
io
n
was
p
r
o
p
o
s
ed
b
y
Go
o
g
le
t
h
at
u
tili
ze
s
C
NN
s
tr
u
ctu
r
e
an
d
r
elies
o
n
d
ep
th
wis
e
-
s
ep
ar
ab
le
co
n
v
o
l
u
tio
n
[
4
5
]
.
I
t
was
u
s
ed
in
a
lo
t
o
f
class
if
icatio
n
an
d
r
ec
o
g
n
iti
o
n
task
s
with
h
i
g
h
ac
cu
r
ac
y
.
Acc
o
r
d
in
g
to
th
e
o
f
f
icia
l
d
o
cu
m
en
tatio
n
,
Xce
p
tio
n
ar
c
h
itectu
r
e
in
v
o
lv
es
3
6
c
o
n
v
o
lu
ti
o
n
al
lay
er
s
th
at
f
o
r
m
t
h
e
f
ea
tu
r
e
ex
tr
ac
tio
n
b
ase
o
f
th
e
n
etwo
r
k
[
4
6
]
.
4
.
7
.
T
esting
T
h
e
last
s
tag
e
in
th
i
s
r
esear
ch
was
th
e
tes
tin
g
p
ar
t
wh
er
e
th
e
test
in
g
s
u
b
s
et
wa
s
u
s
ed
as
th
e
in
p
u
t
f
o
r
th
e
tr
ain
ed
m
o
d
el
t
o
g
et
t
h
e
b
est
m
o
d
el
r
esu
lts
in
d
etec
t
in
g
b
r
ea
s
t
ab
n
o
r
m
alities
.
Mo
r
eo
v
er
,
th
e
tr
ain
e
d
m
o
d
els
wer
e
im
p
lem
en
te
d
o
n
th
e
ex
ter
n
al
d
ataset
as
well.
T
o
m
ea
s
u
r
e
th
e
m
o
d
els’
p
er
f
o
r
m
an
ce
,
c
o
m
m
o
n
ML
p
er
f
o
r
m
a
n
ce
m
etr
ics
wer
e
u
s
ed
s
u
ch
as
ac
cu
r
ac
y
,
r
ec
all
,
p
r
ec
is
io
n
,
an
d
F1
-
s
co
r
e.
T
h
es
e
m
ea
s
u
r
es
will
b
e
d
is
cu
s
s
ed
in
th
e
n
ex
t su
b
s
ec
tio
n
s
.
4
.
7
.
1
.
Acc
ura
cy
Acc
u
r
ac
y
is
th
e
s
im
p
lest
an
d
th
e
m
o
s
t
u
s
ed
m
ea
s
u
r
e
in
th
e
class
if
icatio
n
s
tu
d
ies.
I
t
m
ea
s
u
r
es
h
o
w
o
f
ten
th
e
m
o
d
el
co
r
r
ec
tly
p
r
e
d
icts
th
e
o
u
tco
m
e.
I
n
g
en
er
al,
it
r
ep
r
esen
ts
th
e
r
atio
o
f
th
e
co
r
r
ec
t
p
r
ed
ictio
n
s
o
f
all
p
r
ed
ictio
n
s
.
T
h
e
a
cc
u
r
ac
y
f
o
r
m
u
la
is
s
h
o
wn
in
(
3
)
.
=
+
+
+
+
∗
100
(
3
)
W
h
er
e:
TP
:
tr
u
e
p
o
s
itiv
e
;
wh
en
th
e
m
o
d
el
p
r
e
d
icts
th
e
p
o
s
itiv
e
s
am
p
le
co
r
r
ec
tly
.
TN
:
tr
u
e
n
eg
ativ
e
; w
h
en
th
e
m
o
d
el
p
r
e
d
icts
th
e
n
eg
ativ
e
s
am
p
le
co
r
r
ec
tly
.
FP
:
f
alse p
o
s
itiv
e
; w
h
en
th
e
m
o
d
el
p
r
e
d
icts
th
e
p
o
s
itiv
e
s
am
p
le
b
u
t it’
s
n
eg
ativ
e
in
r
ea
lity
.
FN
:
f
alse n
eg
ativ
e
; w
h
en
th
e
m
o
d
el
p
r
e
d
icts
th
e
n
eg
ativ
e
s
a
m
p
le
b
u
t it’
s
p
o
s
itiv
e
in
r
ea
lity
.
4
.
7
.
2
.
Rec
a
ll
R
ec
all
is
th
e
m
etr
ic
th
at
m
ea
s
u
r
es
h
o
w
o
f
ten
th
e
m
o
d
el
id
e
n
tifie
s
“tr
u
e
p
o
s
itiv
es”
co
r
r
ec
t
ly
f
r
o
m
all
g
r
o
u
n
d
t
r
u
th
p
o
s
itiv
es.
I
t’
s
u
s
ed
to
ex
t
r
ac
t
th
e
r
atio
o
f
tr
u
e
p
o
s
itiv
es
to
all
p
o
s
itiv
es
in
th
e
g
r
o
u
n
d
tr
u
t
h
.
Mo
r
eo
v
er
,
it
r
e
p
r
esen
ts
ty
p
e
-
I
I
er
r
o
r
s
wh
ich
o
cc
u
r
wh
e
n
t
h
e
m
o
d
el
ac
ce
p
ts
th
e
f
alse
n
u
ll
h
y
p
o
th
esis
(H
0
)
.
T
h
e
f
o
r
m
u
la
is
s
h
o
wn
in
(
4
)
.
=
+
(
4
)
4
.
7
.
3
.
P
re
cisi
o
n
Pre
cisi
o
n
is
th
e
a
b
ilit
y
o
f
th
e
m
o
d
el
to
p
r
ed
ict
t
h
e
p
o
s
itiv
e
class
co
r
r
ec
tly
.
I
t
r
e
p
r
esen
ts
t
h
e
r
atio
o
f
“tr
u
e
p
o
s
itiv
e”
to
all
p
r
ed
ict
ed
p
o
s
itiv
es.
Ad
d
itio
n
ally
,
it
f
o
cu
s
es
o
n
ty
p
e
-
I
er
r
o
r
wh
i
ch
o
cc
u
r
s
wh
en
th
e
m
o
d
el
r
ejec
ts
th
e
tr
u
e
n
u
ll h
y
p
o
t
h
esis
(H
0
)
b
y
m
is
tak
e.
T
h
e
f
o
r
m
u
la
is
s
h
o
wn
in
(
5
)
.
=
+
(
5
)
4
.
7
.
4
.
F1
-
s
co
re
T
h
e
last
m
ea
s
u
r
e
is
th
e
F1
-
s
co
r
e.
I
t’
s
t
h
e
h
a
r
m
o
n
ic
m
ea
n
o
f
p
r
ec
is
io
n
an
d
r
ec
all.
I
t’
s
u
s
ed
in
b
in
ar
y
an
d
m
u
lti
-
class
class
if
icatio
n
s
.
T
h
e
f
o
r
m
u
la
is
s
h
o
wn
in
(
6
)
.
1
−
=
2
∗
∗
+
(
6
)
4
.
8
.
P
y
t
ho
n
c
o
de
I
n
th
is
s
tu
d
y
,
Py
th
o
n
lan
g
u
ag
e
was
u
s
ed
to
ac
h
iev
e
th
e
g
o
a
l
o
f
th
is
s
tu
d
y
wh
ich
is
d
etec
tin
g
b
r
ea
s
t
ab
n
o
r
m
alities
in
m
am
m
o
g
r
am
im
ag
es.
T
ab
le
3
lis
ts
th
e
d
etailed
s
tep
s
to
r
ep
r
o
d
u
ce
th
e
f
i
n
d
in
g
s
o
b
tain
e
d
in
th
is
s
tu
d
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
36
,
No
.
2
,
No
v
em
b
er
20
24
:
1
1
4
9
-
1
1
6
2
1158
T
ab
le
3
.
Py
th
o
n
f
u
n
ctio
n
s
S
t
e
p
S
t
a
g
e
Th
e
f
u
n
c
t
i
o
n
u
se
d
i
n
P
y
t
h
o
n
1
I
mp
o
r
t
l
i
b
r
a
r
i
e
s
C
V
2
,
O
S
,
K
e
r
a
s,
n
u
mp
y
,
t
e
n
s
o
r
f
l
o
w
,
mat
p
l
o
t
l
i
b
,
sk
l
e
a
r
n
2
Lo
a
d
t
h
e
d
a
t
a
s
e
t
a
n
d
set
t
h
e
sa
v
i
n
g
l
o
c
a
t
i
o
n
d
a
t
a
_
d
i
r
e
c
t
o
r
y
=
“
y
o
u
r
d
a
t
a
set
d
i
r
e
c
t
o
r
y
”
sav
e
_
d
i
r
e
c
t
o
r
y
=
“
t
h
e
d
i
r
e
c
t
o
r
y
o
f
s
a
v
e
d
mo
d
e
l
”
3
P
r
e
p
r
o
c
e
ss
i
n
g
g
r
a
y
sca
l
e
:
c
v
2
.
c
v
t
C
o
l
o
r
r
e
si
z
e
:
c
v
2
.
r
e
si
z
e
n
o
r
m
a
l
i
z
e
:
i
m
g
/
2
5
5
.
0
A
u
g
m
e
n
t
a
t
i
o
n
:
I
mag
e
D
a
t
a
G
e
n
e
r
a
t
o
r
()
4
S
p
l
i
t
t
h
e
d
a
t
a
s
e
t
d
e
f
s
p
l
i
t
_
d
a
t
a
se
t
(
i
ma
g
e
s
,
l
a
b
e
l
s
,
t
e
st
_
si
z
e
=
0
.
2
,
r
a
n
d
o
m
_
s
t
a
t
e
=
4
2
)
5
M
o
d
e
l
l
i
n
g
w
i
t
h
L2
r
e
g
u
l
a
r
i
z
a
t
i
o
n
d
e
f
c
r
e
a
t
e
_
e
f
f
i
c
i
e
n
t
n
e
t
b
7
_
m
o
d
e
l
(
i
n
p
u
t
_
s
h
a
p
e
=
(
2
2
4
,
2
2
4
,
3
)
,
n
u
m_
c
l
a
sses
=
2
,
r
e
g
u
l
a
r
i
z
a
t
i
o
n
_
st
r
e
n
g
t
h
=
0
.
0
1
)
:
b
a
s
e
_
m
o
d
e
l
=
Ef
f
i
c
i
e
n
t
N
e
t
B
7
(
w
e
i
g
h
t
s=
‘
i
m
a
g
e
n
e
t
,
i
n
c
l
u
d
e
_
t
o
p
=
F
a
l
s
e
,
i
n
p
u
t
_
sh
a
p
e
=
i
n
p
u
t
_
s
h
a
p
e
)
mo
d
e
l
=
mo
d
e
l
s.Se
q
u
e
n
t
i
a
l
(
[
]
)
6
Tr
a
i
n
t
h
e
m
o
d
e
l
d
e
f
t
r
a
i
n
_
mo
d
e
l
(
m
o
d
e
l
,
t
r
a
i
n
_
i
m
a
g
e
s
,
t
r
a
i
n
_
l
a
b
e
l
s,
v
a
l
_
i
m
a
g
e
s,
v
a
l
_
l
a
b
e
l
s,
e
p
o
c
h
s=
2
0
,
b
a
t
c
h
_
s
i
z
e
=
3
2
)
:
e
a
r
l
y
_
s
t
o
p
p
i
n
g
=
c
a
l
l
b
a
c
k
s.E
a
r
l
y
S
t
o
p
p
i
n
g
(
p
a
t
i
e
n
c
e
=
5
,
r
e
s
t
o
r
e
_
b
e
st
_
w
e
i
g
h
t
s=T
r
u
e
)
mo
d
e
l
.
c
o
m
p
i
l
e
(
t
f
.
k
e
r
a
s.
o
p
t
i
mi
z
e
r
s.A
d
a
m(
l
e
a
r
n
i
n
g
_
r
a
t
e
=
0
.
0
0
1
)
,
l
o
ss
=
‘
sp
a
r
se
_
c
a
t
e
g
o
r
i
c
a
l
_
c
r
o
sse
n
t
r
o
p
y
’
,
met
r
i
c
s=[
‘
a
c
c
u
r
a
c
y
’
])
h
i
s
t
o
r
y
=
m
o
d
e
l
.
f
i
t
(
t
r
a
i
n
_
i
m
a
g
e
s,
t
r
a
i
n
_
l
a
b
e
l
s
,
e
p
o
c
h
s=
e
p
o
c
h
s,
v
a
l
i
d
a
t
i
o
n
_
d
a
t
a
=
(
v
a
l
_
i
m
a
g
e
s,
v
a
l
_
l
a
b
e
l
s)
,
b
a
t
c
h
_
s
i
z
e
=
b
a
t
c
h
_
s
i
z
e
,
c
a
l
l
b
a
c
k
s=[
e
a
r
l
y
_
st
o
p
p
i
n
g
]
)
mo
d
e
l
.
sa
v
e
(
‘
t
r
a
i
n
e
d
_
Ef
f
i
c
i
e
n
t
N
e
t
B
7
.
k
e
r
a
s
’
)
r
e
t
u
r
n
m
o
d
e
l
,
h
i
s
t
o
r
y
7
Ev
a
l
u
a
t
i
o
n
c
m =
c
o
n
f
u
si
o
n
_
m
a
t
r
i
x
(
t
e
s
t
_
l
a
b
e
l
s,
p
r
e
d
i
c
t
i
o
n
s)
d
i
s
p
l
a
y
=
C
o
n
f
u
si
o
n
M
a
t
r
i
x
D
i
s
p
l
a
y
(
c
o
n
f
u
s
i
o
n
_
ma
t
r
i
x
=
c
m,
d
i
s
p
l
a
y
_
l
a
b
e
l
s=[
‘
A
b
n
o
r
m
a
l
’
,
‘
N
o
r
ma
l
’
])
d
i
s
p
l
a
y
.
p
l
o
t
(
c
ma
p
=
‘
B
l
u
e
s
’
,
v
a
l
u
e
s_
f
o
r
mat
=
‘
d
’
)
p
l
t
.
t
i
t
l
e
(
‘
C
o
n
f
u
s
i
o
n
M
a
t
r
i
x
2
’
)
p
l
t
.
s
h
o
w
(
)
5.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
s
tu
d
y
im
p
lem
en
ted
a
m
eth
o
d
o
l
o
g
y
ap
p
r
o
ac
h
to
in
v
esti
g
a
te
th
e
ef
f
ec
ts
o
f
th
e
p
o
wer
a
n
d
ef
f
ec
tiv
en
ess
o
f
m
ac
h
in
e
lea
r
n
in
g
m
o
d
els
in
d
etec
tin
g
b
r
e
ast
ab
n
o
r
m
alities
in
lo
ca
l
m
am
m
o
g
r
am
i
m
ag
es
wh
ile
o
th
er
s
f
o
cu
s
ed
o
n
u
s
in
g
th
e
p
u
b
lic
c
o
m
m
o
n
ly
u
s
ed
d
atasets
lik
e
MI
AS
an
d
I
n
b
r
ea
s
t.
T
h
e
ap
p
r
o
ac
h
u
s
ed
C
NN
d
ee
p
lear
n
in
g
m
o
d
els,
n
am
el
y
Den
s
eNe
t1
2
1
,
R
esNet5
0
,
VGG1
6
,
VGG1
9
,
Xce
p
tio
n
,
a
n
d
E
f
f
if
icen
tNetB
7
.
T
h
e
in
p
u
t
o
f
th
ese
m
o
d
el
s
was
a
tr
ain
i
n
g
s
u
b
s
et
o
f
s
ize
1
,
6
0
0
was
u
s
ed
t
o
lear
n
th
e
m
o
d
els,
wh
ile
a
test
in
g
s
u
b
s
et
o
f
s
ize
4
0
0
was
u
s
ed
f
o
r
test
in
g
an
d
e
v
alu
atio
n
.
T
h
e
e
x
p
er
im
e
n
ts
in
th
is
ap
p
r
o
ac
h
wer
e
d
o
n
e
u
s
in
g
th
e
s
am
e
p
ar
am
ete
r
v
alu
es
ac
r
o
s
s
all
m
o
d
els.
T
h
e
m
o
d
el’
s
in
p
u
t
s
h
ap
e
was
2
2
4
×
2
2
4
a
n
d
th
e
test
s
u
b
s
et
s
ize
was
0
.
2
o
f
th
e
o
r
ig
in
al
d
ataset.
I
n
ad
d
itio
n
,
th
e
b
atch
s
ize
wa
s
3
2
with
2
0
ep
o
ch
s
an
d
Ad
am
o
p
tim
izer
was
d
ep
lo
y
ed
with
a
0
.
0
0
1
lear
n
in
g
r
ate.
Als
o
,
t
h
e
ac
tiv
atio
n
f
u
n
ctio
n
s
u
s
ed
i
n
th
e
m
o
d
els
wer
e
R
eL
U
an
d
So
f
tm
ax
.
F
in
ally
,
th
e
ap
p
r
o
ac
h
class
if
ied
th
e
im
ag
es
in
to
n
o
r
m
al
an
d
ab
n
o
r
m
al
an
d
was
ev
alu
ated
u
s
in
g
d
if
f
e
r
en
t
p
e
r
f
o
r
m
an
ce
m
ea
s
u
r
es
s
u
ch
as
co
n
f
u
s
io
n
m
atr
ix
,
ac
cu
r
ac
y
,
r
ec
all,
p
r
ec
is
io
n
,
F1
-
s
co
r
e,
a
n
d
R
OC
cu
r
v
e.
T
h
e
r
esu
lts
f
r
o
m
t
h
is
ap
p
r
o
ac
h
will b
e
p
r
esen
ted
an
d
d
is
cu
s
s
ed
in
th
is
s
ec
tio
n
.
B
ased
o
n
th
e
ex
p
er
im
en
t,
we
f
o
u
n
d
th
at
Den
s
eNe
t1
2
1
o
u
tp
er
f
o
r
m
ed
o
th
e
r
m
o
d
els
in
d
etec
tin
g
ab
n
o
r
m
alities
.
T
ab
le
4
lis
ts
all
o
b
tain
ed
r
esu
lts
f
r
o
m
p
er
f
o
r
m
an
ce
m
ea
s
u
r
es
v
er
s
u
s
m
o
d
els
in
th
e
test
in
g
s
u
b
s
et
an
d
ex
ter
n
al
d
ataset.
I
t
ca
n
b
e
s
ee
n
t
h
at
th
e
ac
cu
r
ac
ies
in
th
e
test
in
g
a
n
d
ex
ter
n
al
s
ets
r
an
g
ed
b
etwe
e
n
0
.
8
3
an
d
0
.
5
0
.
I
n
th
e
test
in
g
s
u
b
s
et,
Den
s
eNe
t1
2
1
o
u
t
p
er
f
o
r
m
ed
all
o
th
er
m
o
d
els
with
th
e
h
ig
h
est
ac
c
u
r
ac
y
,
AUC,
r
ec
all,
p
r
ec
is
io
n
,
an
d
F
1
-
s
co
r
e.
T
h
e
h
ig
h
est
ac
cu
r
ac
y
f
o
r
d
etec
tin
g
b
r
ea
s
t
ab
n
o
r
m
a
liti
es
o
b
tain
ed
was
0
.
8
2
f
r
o
m
Den
s
eNe
t1
2
1
,
wh
ile
E
f
f
icien
tNetB
7
s
h
o
wed
p
r
o
m
is
in
g
an
d
clo
s
e
r
esu
lts
o
f
Den
s
eNe
t1
2
1
.
E
f
f
icien
tNetB
7
ac
cu
r
ac
y
was
0
.
8
1
an
d
g
o
t
0
.
8
0
,
0
.
7
8
,
0
.
8
2
,
an
d
0
.
8
0
as
AUC,
r
ec
all,
p
r
ec
i
s
io
n
,
an
d
F1
-
s
co
r
e
r
esp
ec
tiv
ely
.
Mo
r
e
o
v
er
,
VGG1
6
,
VGG1
9
,
an
d
Xce
p
tio
n
g
o
t
ac
cu
r
ac
ies
o
f
0
.
7
4
,
0
.
6
8
,
a
n
d
0
.
7
8
r
esp
ec
tiv
ely
.
T
h
e
p
er
f
o
r
m
a
n
ce
in
th
ese
m
o
d
els
was
b
etter
in
d
etec
t
in
g
th
e
ab
n
o
r
m
al
class
th
an
th
e
n
o
r
m
al
class
.
Fu
r
th
er
m
o
r
e
,
th
e
lo
west
ac
cu
r
ac
y
was
0
.
5
0
f
r
o
m
R
esNet5
0
wh
er
e
th
e
m
o
d
el
p
r
ed
icted
all
im
ag
es
as
n
o
r
m
al
class
.
L
ik
ewise,
a
s
ec
o
n
d
e
x
p
er
im
en
t
was
co
n
d
u
cted
u
s
in
g
th
e
s
av
ed
t
r
ain
ed
m
o
d
el
f
r
o
m
th
e
f
ir
s
t
ex
p
er
im
en
t
o
n
an
e
x
ter
n
al
im
a
g
e
as
ex
p
lai
n
ed
in
th
e
d
ataset
s
ec
tio
n
ab
o
v
e.
T
h
e
r
esu
lts
o
b
tain
ed
f
r
o
m
th
is
ex
p
er
im
en
t
ar
e
s
h
o
wn
i
n
T
ab
le
4
.
T
h
e
r
esu
lts
h
er
e
ar
e
v
er
y
s
im
ilar
to
t
h
e
r
esu
lts
o
b
tain
ed
f
r
o
m
th
e
f
ir
s
t
ex
p
er
im
en
t.
I
n
th
i
s
ca
s
e,
Den
s
eNe
t1
2
1
p
r
o
d
u
ce
d
th
e
b
est
r
esu
lts
in
ter
m
s
o
f
p
er
f
o
r
m
an
ce
an
d
o
u
tp
e
r
f
o
r
m
ed
o
th
er
m
o
d
els.
Den
s
eNe
t1
2
1
o
b
tain
ed
ac
cu
r
ac
y
,
AUC,
r
ec
all,
p
r
ec
is
io
n
,
an
d
F1
-
s
co
r
e
o
f
0
.
8
3
,
0
.
8
5
,
0
.
7
0
,
1
.
0
,
a
n
d
0
.
8
2
Evaluation Warning : The document was created with Spire.PDF for Python.