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.
39
,
No
.
2
,
A
u
g
u
s
t
2
0
2
5
,
p
p
.
1
236
~
1
2
4
8
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/
ijeec
s
.
v
39
.i
2
.
p
p
1
2
3
6
-
1
2
4
8
1236
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Dev
elo
pment of
R
esNet
-
1
8
archi
tec
ture
to lesio
n ide
n
tif
ica
tion
in breas
t
ultras
o
u
nd ima
g
es
Sil
f
ia
Andi
n
i
1
,
Su
m
ij
a
n
2
,
I
s
k
a
nd
a
r
F
it
ri
2
1
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
c
s
M
a
n
a
g
e
m
e
n
t
,
U
n
i
v
e
r
s
i
t
a
s Pu
t
r
a
I
n
d
o
n
e
s
i
a
Y
P
T
K
P
a
d
a
n
g
,
W
e
s
t
S
u
ma
t
e
r
a
,
I
n
d
o
n
e
si
a
2
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
o
n
Te
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
si
t
a
s
P
u
t
r
a
I
n
d
o
n
e
s
i
a
Y
P
T
K
,
P
a
d
a
n
g
,
W
e
s
t
S
u
ma
t
e
r
a
,
I
n
d
o
n
e
s
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
4
,
2
0
2
4
R
ev
is
ed
Ma
r
2
4
,
2
0
2
5
Acc
ep
ted
J
u
l 2
,
2
0
2
5
Bre
a
st
u
lt
ra
so
u
n
d
(USG
)
is
wid
e
l
y
u
se
d
fo
r
e
a
rly
b
re
a
st
c
a
n
c
e
r
d
e
t
e
c
ti
o
n
,
b
u
t
c
h
a
ll
e
n
g
e
s
su
c
h
a
s
n
o
ise
,
lo
w
c
o
n
tras
t,
a
n
d
re
so
lu
ti
o
n
li
m
it
a
ti
o
n
s
h
i
n
d
e
r
a
c
c
u
ra
te
les
io
n
i
d
e
n
ti
fica
ti
o
n
.
Th
is
st
u
d
y
p
r
o
p
o
se
s
a
m
o
d
ifi
e
d
re
sid
u
a
l
n
e
two
rk
-
18
(
Re
sN
e
t
-
18
)
a
rc
h
it
e
c
t
u
re
f
o
r
b
re
a
st
les
io
n
se
g
m
e
n
tati
o
n
,
a
ime
d
a
t
imp
ro
v
in
g
d
e
tec
ti
o
n
a
c
c
u
ra
c
y
.
Th
e
m
e
th
o
d
o
l
o
g
y
i
n
v
o
lv
e
s
p
re
p
ro
c
e
ss
in
g
ste
p
s
in
c
lu
d
i
n
g
re
d
g
re
e
n
b
l
u
e
(
RG
B
)
to
G
ra
y
sc
a
le
c
o
n
v
e
rsio
n
,
c
o
n
tras
t
stre
tch
in
g
,
a
n
d
m
e
d
ian
fil
terin
g
to
e
n
h
a
n
c
e
ima
g
e
q
u
a
li
t
y
.
Th
e
m
o
d
ifi
e
d
Re
sN
e
t
-
1
8
m
o
d
e
l
i
n
tro
d
u
c
e
s
a
d
d
i
ti
o
n
a
l
c
o
n
v
o
lu
t
io
n
a
l
lay
e
rs
to
re
fi
n
e
fe
a
tu
re
e
x
trac
ti
o
n
.
T
h
e
p
ro
p
o
se
d
m
o
d
e
l
wa
s
train
e
d
a
n
d
v
a
li
d
a
ted
o
n
3
0
b
re
a
st
u
lt
ra
so
u
n
d
ima
g
e
s,
wit
h
e
v
a
lu
a
ti
o
n
m
e
tri
c
s
in
c
lu
d
in
g
a
c
c
u
ra
c
y
,
s
e
n
siti
v
it
y
,
a
n
d
s
p
e
c
ifi
c
it
y
.
E
x
p
e
rime
n
tal
re
s
u
lt
s
i
n
d
ica
te
th
a
t
th
e
m
o
d
ifi
e
d
a
rc
h
it
e
c
tu
re
o
u
t
p
e
rfo
rm
s
t
h
e
b
a
se
li
n
e
m
o
d
e
l,
a
c
h
iev
in
g
a
n
a
v
e
ra
g
e
a
c
c
u
ra
c
y
o
f
0
.
9
7
0
9
3
,
se
n
siti
v
it
y
o
f
0
.
9
0
0
5
6
,
a
n
d
s
p
e
c
ifi
c
it
y
o
f
0
.
9
7
7
0
5
.
Va
li
d
a
ti
o
n
b
y
a
ra
d
io
l
o
g
y
sp
e
c
ialist
c
o
n
firms
t
h
e
m
o
d
e
l
’
s
c
li
n
ica
l
re
lev
a
n
c
e
.
T
h
e
se
fin
d
in
g
s
su
g
g
e
st
th
a
t
t
h
e
e
n
h
a
n
c
e
d
Re
sN
e
t
-
1
8
m
o
d
e
l
h
a
s
t
h
e
p
o
ten
t
ial
t
o
a
ss
ist
ra
d
i
o
lo
g
ists
i
n
m
o
re
a
c
c
u
ra
tely
id
e
n
ti
fy
i
n
g
b
re
a
st
les
io
n
s.
F
u
t
u
re
re
se
a
rc
h
sh
o
u
l
d
fo
c
u
s
o
n
e
x
p
a
n
d
in
g
t
h
e
d
a
tas
e
t,
i
n
teg
ra
ti
n
g
m
u
lt
i
-
m
o
d
a
l
ima
g
i
n
g
,
a
n
d
o
p
ti
m
izin
g
m
o
d
e
l
g
e
n
e
ra
li
z
a
b
il
it
y
f
o
r
re
a
l
-
ti
m
e
c
li
n
ica
l
a
p
p
li
c
a
ti
o
n
s.
T
h
e
stu
d
y
c
o
n
tri
b
u
tes
t
o
a
d
v
a
n
c
in
g
a
rti
fic
ial
in
telli
g
e
n
c
e
(
AI
)
-
d
ri
v
e
n
b
re
a
st
c
a
n
c
e
r
d
iag
n
o
stics
,
su
p
p
o
rti
n
g
e
a
rly
d
e
te
c
ti
o
n
,
a
n
d
imp
ro
v
in
g
p
a
ti
e
n
t
o
u
tco
m
e
s.
K
ey
w
o
r
d
s
:
B
r
ea
s
t
I
d
en
ti
f
icatio
n
L
esio
n
R
esNet
-
18
Ultr
aso
u
n
d
im
ag
es
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
:
Sil
f
ia
An
d
in
i
Dep
ar
tm
en
t o
f
I
n
f
o
r
m
atics
Ma
n
ag
em
en
t,
U
n
iv
er
s
itas
Pu
tr
a
I
n
d
o
n
esia Y
PTK
Pad
an
g
2
5
2
2
1
,
Pad
an
g
,
W
est Su
m
ater
a,
I
n
d
o
n
esia
E
m
ail: sil
f
iaan
d
in
i6
8
@
u
p
iy
p
t
k
.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
B
r
ea
s
t
ca
n
ce
r
r
em
ain
s
o
n
e
o
f
th
e
lead
in
g
ca
u
s
es
o
f
m
o
r
tality
am
o
n
g
wo
m
e
n
wo
r
l
d
wid
e,
with
ea
r
ly
d
etec
tio
n
p
lay
in
g
a
cr
u
cial
r
o
l
e
in
im
p
r
o
v
in
g
s
u
r
v
iv
al
r
ates
[
1
]
.
I
n
d
e
v
elo
p
in
g
co
u
n
tr
ies
lik
e
I
n
d
o
n
esia,
late
-
s
tag
e
d
iag
n
o
s
is
is
a
p
r
ev
alen
t
is
s
u
e
d
u
e
to
lim
ited
ac
ce
s
s
to
ad
v
an
ce
d
d
iag
n
o
s
tic
to
o
ls
an
d
v
ar
iab
ilit
y
in
u
ltra
s
o
u
n
d
im
ag
i
n
g
in
ter
p
r
etat
io
n
[
2
]
.
Ultr
aso
u
n
d
(
USG)
is
wid
ely
u
s
ed
f
o
r
ea
r
l
y
b
r
ea
s
t
c
an
ce
r
s
cr
ee
n
in
g
d
u
e
to
its
n
o
n
-
in
v
asiv
e
n
atu
r
e
an
d
ab
ilit
y
to
d
etec
t
ab
n
o
r
m
alities
in
d
en
s
e
b
r
ea
s
t
tis
s
u
e
[
3
]
.
Ho
wev
er
,
th
e
ac
cu
r
ac
y
o
f
u
ltra
s
o
u
n
d
-
b
ased
d
iag
n
o
s
is
is
o
f
ten
h
i
n
d
er
ed
b
y
s
ev
er
al
ch
allen
g
es,
i
n
clu
d
i
n
g
im
a
g
e
n
o
is
e,
lo
w
co
n
tr
ast,
an
d
o
p
er
ato
r
d
ep
en
d
en
cy
,
m
ak
in
g
it d
if
f
icu
lt to
d
is
tin
g
u
is
h
m
alig
n
an
t f
r
o
m
b
e
n
ig
n
lesi
o
n
s
[
4
]
.
T
h
e
r
ap
i
d
d
e
v
elo
p
m
e
n
t
o
f
d
ig
ital
im
ag
e
p
r
o
ce
s
s
in
g
tec
h
n
o
lo
g
y
h
as
g
r
ea
tly
b
en
ef
ite
d
m
ed
ical
im
ag
in
g
,
f
ac
ilit
atin
g
m
o
r
e
ac
c
u
r
at
e
an
d
r
eliab
le
d
iag
n
o
s
es
[
5
]
.
I
m
ag
e
p
r
o
ce
s
s
in
g
tech
n
iq
u
es
aim
to
en
h
a
n
ce
im
ag
e
q
u
ality
an
d
ex
tr
ac
t
m
ea
n
in
g
f
u
l
in
f
o
r
m
atio
n
f
o
r
b
et
ter
an
aly
s
is
[
6
]
.
Am
o
n
g
th
e
ad
v
an
ce
d
m
eth
o
d
s
,
co
n
v
o
l
u
tio
n
al
n
e
u
r
al
n
etwo
r
k
s
(
C
NNs)
h
av
e
s
h
o
wn
r
em
ar
k
ab
le
p
er
f
o
r
m
an
ce
i
n
im
ag
e
a
n
aly
s
is
an
d
m
ed
ical
ap
p
licatio
n
s
d
u
e
to
th
eir
ab
ilit
y
to
lear
n
h
ier
ar
c
h
ical
f
ea
tu
r
e
s
d
ir
ec
tly
f
r
o
m
im
ag
e
d
ata
[
7
]
.
C
NNs
co
n
s
i
s
t
o
f
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:
2502
-
4
7
5
2
Dev
elo
p
men
t o
f R
esN
et
-
1
8
a
r
ch
itectu
r
e
to
lesi
o
n
id
en
tifi
ca
tio
n
in
b
r
ea
s
t
…
(
S
ilfi
a
A
n
d
in
i
)
1237
m
u
ltip
le
lay
er
s
,
in
clu
d
in
g
co
n
v
o
lu
tio
n
al
an
d
p
o
o
lin
g
la
y
er
s
,
wh
ich
h
elp
ca
p
t
u
r
e
s
p
atial
f
e
atu
r
es
at
d
if
f
er
en
t
lev
els
o
f
g
r
an
u
lar
ity
.
On
e
o
f
th
e
p
o
p
u
lar
C
NN
ar
c
h
itectu
r
es
u
s
ed
f
o
r
im
ag
e
an
aly
s
is
i
s
r
esid
u
al
n
e
two
r
k
(
R
esNet
)
[
8
]
.
R
esNet,
in
tr
o
d
u
ce
d
to
o
v
e
r
co
m
e
t
h
e
v
an
is
h
in
g
g
r
ad
ie
n
t
p
r
o
b
lem
,
allo
ws
d
ee
p
er
n
etwo
r
k
s
b
y
in
co
r
p
o
r
atin
g
s
h
o
r
tcu
t
co
n
n
ec
tio
n
s
,
wh
ich
im
p
r
o
v
e
tr
ain
in
g
an
d
p
er
f
o
r
m
a
n
ce
.
R
esNet
-
1
8
,
a
s
p
ec
if
ic
v
ar
ian
t
o
f
R
esNet,
h
as
b
ee
n
wid
ely
a
d
o
p
ted
i
n
m
ed
ical
im
a
g
e
an
al
y
s
is
d
u
e
to
its
b
alan
ce
b
etwe
en
co
m
p
le
x
ity
an
d
p
er
f
o
r
m
an
ce
.
R
esear
ch
er
s
h
av
e
b
ee
n
lev
er
a
g
in
g
a
n
d
m
o
d
if
y
in
g
R
esNet
-
1
8
f
o
r
v
ar
io
u
s
task
s
,
s
u
ch
as
lesi
o
n
d
etec
ti
o
n
an
d
class
if
icatio
n
,
to
en
h
a
n
ce
ac
cu
r
ac
y
an
d
r
elia
b
ilit
y
.
T
h
e
f
lex
i
b
ilit
y
o
f
th
e
R
esNet
ar
ch
itectu
r
e,
alo
n
g
with
d
ata
au
g
m
en
tatio
n
an
d
tr
an
s
f
er
lear
n
i
n
g
tech
n
i
q
u
es,
en
ab
les th
e
d
ev
elo
p
m
e
n
t o
f
m
o
d
els ca
p
ab
le
o
f
h
an
d
lin
g
d
if
f
er
en
t
ch
allen
g
es
in
m
e
d
ical
im
a
g
i
n
g
,
in
clu
d
i
n
g
n
o
is
e
r
ed
u
ctio
n
,
f
ea
tu
r
e
e
x
tr
ac
tio
n
,
an
d
h
ig
h
-
ac
cu
r
ac
y
class
if
icatio
n
.
Pre
v
io
u
s
s
tu
d
ies
h
av
e
ex
p
l
o
r
ed
th
e
p
o
ten
tial
o
f
C
NN
-
b
ased
ar
ch
itectu
r
es,
p
ar
ticu
lar
l
y
R
esNet
-
1
8
,
f
o
r
b
r
ea
s
t
lesi
o
n
d
etec
tio
n
an
d
class
if
icatio
n
.
W
an
g
an
d
Hu
a
n
g
[
9
]
in
tr
o
d
u
ce
d
a
f
r
am
ewo
r
k
u
s
in
g
C
Dee
p
3
M,
a
C
NN
-
b
ased
m
o
d
el,
wh
ic
h
d
em
o
n
s
tr
ated
p
r
o
m
is
in
g
r
esu
lts
in
b
r
ea
s
t
t
u
m
o
r
s
eg
m
en
tati
o
n
.
T
h
is
ap
p
r
o
ac
h
h
ig
h
lig
h
ted
th
e
ca
p
ab
ilit
y
o
f
d
ee
p
lear
n
in
g
in
ass
is
tin
g
co
m
p
r
eh
e
n
s
iv
e
d
iag
n
o
s
es,
s
h
o
win
g
s
ig
n
if
ican
t
f
lex
ib
ilit
y
in
r
ea
l
-
tim
e
ca
n
ce
r
p
r
ed
ictio
n
[
9
]
.
Similar
ly
,
Das
an
d
R
an
a
[
1
0
]
in
v
esti
g
ated
v
ar
io
u
s
R
esNet
ar
ch
itectu
r
es
(
R
esNet
-
1
8
,
-
3
4
,
-
5
0
,
-
1
0
1
,
an
d
-
1
5
2
)
to
d
et
ec
t
b
r
ea
s
t
ca
n
ce
r
f
r
o
m
u
ltra
s
o
u
n
d
im
a
g
es.
T
h
ei
r
r
esu
lts
s
h
o
wed
th
at
R
esNet
m
o
d
els,
esp
ec
ially
with
tr
an
s
f
er
lear
n
in
g
,
co
u
l
d
ac
h
iev
e
h
i
g
h
p
r
ec
is
io
n
,
r
ec
all,
an
d
ac
cu
r
ac
y
,
in
d
icatin
g
th
ei
r
ef
f
ec
tiv
en
ess
f
o
r
m
ed
ical
im
ag
e
an
aly
s
is
.
I
n
an
o
th
er
s
t
u
d
y
,
Yu
et
a
l.
[
1
1
]
p
r
o
p
o
s
ed
an
au
g
m
en
ted
d
ata
f
r
am
ewo
r
k
ca
lled
s
ca
lin
g
an
d
co
n
tr
ast
lim
ited
ad
ap
tiv
e
h
is
t
o
g
r
am
eq
u
aliza
ti
on
d
ata
au
g
m
en
tatio
n
(
SC
DA
)
co
m
b
in
ed
with
R
esNet
-
5
0
,
y
ield
in
g
a
n
a
v
er
ag
e
s
p
ec
if
ic
ity
o
f
9
8
.
5
5
%
a
n
d
s
en
s
itiv
ity
o
f
9
2
.
8
3
%.
T
h
is
en
h
an
ce
d
th
e
o
v
e
r
all
m
o
d
el
ac
cu
r
ac
y
to
9
5
.
7
4
%
[
1
1
]
.
Fu
r
t
h
er
m
o
r
e,
Dai
et
a
l.
[
1
2
]
p
r
esen
ted
a
r
ea
l
-
tim
e
d
etec
ti
o
n
n
etwo
r
k
,
STA
-
Net
,
b
u
ilt
o
n
R
esNet
-
1
8
f
o
r
v
id
eo
u
ltra
s
o
u
n
d
b
r
ea
s
t
lesi
o
n
d
etec
tio
n
.
T
h
eir
m
o
d
el
r
ea
ch
e
d
an
im
p
r
ess
iv
e
5
4
f
r
a
m
es
p
er
s
ec
o
n
d
p
r
o
ce
s
s
in
g
r
ate
with
a
n
av
er
ag
e
p
r
ec
is
io
n
o
f
3
8
.
7
m
AP,
o
u
tp
er
f
o
r
m
i
n
g
ex
is
tin
g
m
eth
o
d
s
an
d
d
em
o
n
s
tr
atin
g
th
e
p
o
te
n
tial
o
f
C
NN
-
b
ased
r
ea
l
-
tim
e
an
aly
s
is
[
1
2
]
.
Sev
er
al
co
m
p
u
tatio
n
al
a
p
p
r
o
ac
h
es
h
av
e
b
ee
n
p
r
o
p
o
s
ed
to
en
h
a
n
ce
b
r
ea
s
t
lesi
o
n
d
et
ec
tio
n
i
n
u
ltra
s
o
u
n
d
im
ag
es.
T
r
a
d
itio
n
a
l
m
ac
h
in
e
lear
n
in
g
tech
n
iq
u
es,
s
u
ch
as
tex
tu
r
e
an
al
y
s
is
an
d
h
an
d
cr
a
f
ted
f
ea
t
u
r
e
ex
tr
ac
tio
n
,
h
a
v
e
s
h
o
wn
p
r
o
m
i
s
e
b
u
t
o
f
ten
lac
k
r
o
b
u
s
tn
ess
d
u
e
to
v
ar
iatio
n
s
in
im
ag
e
q
u
al
ity
.
R
ec
en
tly
,
d
ee
p
lear
n
in
g
m
o
d
els,
p
ar
ticu
lar
l
y
C
NNs,
h
av
e
b
ee
n
wid
ely
ad
o
p
ted
f
o
r
m
ed
ical
i
m
ag
e
an
aly
s
is
.
Stu
d
ies
b
y
W
an
g
an
d
Hu
a
n
g
[
9
]
an
d
Yu
et
a
l.
[
1
1
]
d
em
o
n
s
tr
ated
th
e
ef
f
ec
t
iv
en
ess
o
f
C
NN
-
b
ased
s
eg
m
en
tatio
n
m
o
d
els
in
im
p
r
o
v
in
g
lesi
o
n
d
etec
tio
n
ac
cu
r
ac
y
.
Ho
wev
e
r
,
th
ese
m
eth
o
d
s
s
till
f
ac
e
ch
allen
g
es
r
elate
d
to
lim
ited
tr
ain
in
g
d
ata,
o
v
er
f
itti
n
g
,
an
d
th
e
n
ee
d
f
o
r
im
p
r
o
v
e
d
f
ea
tu
r
e
ex
tr
ac
tio
n
tailo
r
ed
to
b
r
ea
s
t u
ltra
s
o
u
n
d
im
ag
in
g
.
B
u
ild
in
g
o
n
th
e
f
o
u
n
d
atio
n
o
f
ex
is
tin
g
r
esear
ch
,
th
is
s
tu
d
y
aim
s
to
f
u
r
th
er
d
e
v
elo
p
a
s
eg
m
en
tatio
n
m
eth
o
d
f
o
r
id
e
n
tify
in
g
b
r
e
ast
lesi
o
n
s
in
u
ltra
s
o
u
n
d
im
a
g
es
u
s
in
g
a
m
o
d
if
ied
R
esNet
-
1
8
ar
ch
itectu
r
e.
T
h
e
f
o
cu
s
will
b
e
o
n
en
h
a
n
cin
g
th
e
R
esNet
-
1
8
m
o
d
el
b
y
ad
d
i
n
g
an
d
m
o
d
if
y
in
g
co
n
v
o
lu
ti
o
n
al
lay
er
s
to
cr
ea
te
a
n
o
v
el
m
o
d
el
ca
p
ab
le
o
f
p
r
o
d
u
cin
g
m
o
r
e
ac
cu
r
ate
r
esu
lts
.
T
h
e
d
ev
elo
p
ed
m
o
d
el
will
u
n
d
er
g
o
ev
alu
atio
n
,
tr
ain
in
g
,
an
d
v
alid
atio
n
to
m
e
asu
r
e
its
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
.
T
h
e
g
o
al
is
t
o
id
en
tif
y
t
h
e
b
est
-
p
er
f
o
r
m
in
g
m
o
d
el
with
o
p
tim
al
v
alu
es
f
o
r
th
ese
p
ar
a
m
eter
s
,
wh
ich
wo
u
l
d
r
e
p
r
esen
t
a
s
ig
n
if
ican
t
co
n
tr
ib
u
tio
n
to
th
e
f
ield
.
T
h
e
n
o
v
elty
o
f
th
is
r
esear
ch
lies
in
th
e
m
o
d
if
icatio
n
s
m
ad
e
to
th
e
o
r
ig
in
al
R
esNet
-
1
8
ar
ch
itectu
r
e
an
d
its
ap
p
licatio
n
in
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
e
an
aly
s
is
.
T
h
e
an
ticip
ated
o
u
tco
m
e
is
th
at
th
is
en
h
an
ce
d
m
o
d
el
ca
n
b
e
im
p
lem
en
ted
in
m
e
d
ic
al
u
ltra
s
o
u
n
d
eq
u
i
p
m
en
t
to
s
u
p
p
o
r
t
r
a
d
io
lo
g
is
ts
an
d
o
n
co
l
o
g
is
ts
in
d
iag
n
o
s
in
g
b
r
ea
s
t
lesi
o
n
s
m
o
r
e
ac
cu
r
ately
an
d
ef
f
icien
tly
.
T
h
e
r
esu
lts
co
u
ld
lead
to
b
etter
clin
ical
r
ec
o
m
m
en
d
atio
n
s
an
d
im
p
r
o
v
e
d
ea
r
ly
d
etec
tio
n
,
u
ltima
tely
co
n
tr
ib
u
tin
g
to
lo
wer
m
o
r
tality
r
ates f
r
o
m
b
r
ea
s
t c
an
ce
r
.
T
h
i
s
s
t
u
d
y
a
i
m
s
t
o
a
d
d
r
e
s
s
t
h
ese
c
h
a
l
l
e
n
g
e
s
b
y
d
e
v
e
l
o
p
i
n
g
a
n
i
m
p
r
o
v
e
d
s
e
g
m
e
n
t
a
t
i
o
n
m
e
t
h
o
d
b
a
s
e
d
o
n
a
m
o
d
i
f
i
e
d
R
es
N
e
t
-
1
8
a
r
c
h
i
tec
t
u
r
e
.
B
y
i
n
t
r
o
d
u
c
i
n
g
a
d
d
i
t
i
o
n
a
l
c
o
n
v
o
l
u
t
i
o
n
a
l
l
a
y
e
r
s
,
w
e
s
e
e
k
t
o
e
n
h
a
n
c
e
t
h
e
m
o
d
e
l
’
s
c
a
p
a
b
i
l
it
y
i
n
e
x
t
r
a
c
t
i
n
g
c
r
i
t
i
c
a
l
f
e
a
t
u
r
es
,
t
h
e
r
e
b
y
i
m
p
r
o
v
i
n
g
t
h
e
a
c
c
u
r
a
c
y
,
s
e
n
s
i
t
i
v
ity
,
a
n
d
s
p
e
c
i
f
i
c
i
t
y
o
f
b
r
e
a
s
t
l
e
s
i
o
n
i
d
e
n
t
i
f
i
c
a
ti
o
n
.
T
h
e
p
r
i
m
a
r
y
o
b
j
e
c
ti
v
e
s
o
f
t
h
is
r
es
e
a
r
c
h
a
r
e:
(
1
)
t
o
d
e
v
el
o
p
a
d
e
e
p
l
e
a
r
n
i
n
g
m
o
d
e
l
t
h
a
t
e
n
h
a
n
c
es
b
r
e
a
s
t
l
es
i
o
n
s
e
g
m
e
n
t
a
t
i
o
n
a
c
c
u
r
a
c
y
,
(
2
)
t
o
e
v
a
l
u
a
t
e
t
h
e
m
o
d
e
l
’
s
p
e
r
f
o
r
m
a
n
c
e
c
o
m
p
a
r
e
d
t
o
s
t
a
n
d
a
r
d
R
es
N
e
t
-
1
8
a
n
d
e
x
is
t
in
g
a
p
p
r
o
a
c
h
e
s
,
a
n
d
(
3
)
t
o
e
x
p
l
o
r
e
i
ts
p
o
te
n
t
i
al
a
p
p
l
ic
a
t
i
o
n
in
c
l
i
n
i
ca
l
d
i
a
g
n
o
s
t
ic
w
o
r
k
f
l
o
w
s
.
B
y
o
p
t
i
m
i
z
i
n
g
l
es
io
n
d
e
t
e
c
t
i
o
n
i
n
u
l
t
r
a
s
o
u
n
d
i
m
ag
i
n
g
,
t
h
i
s
s
t
u
d
y
as
p
i
r
e
s
t
o
c
o
n
t
r
i
b
u
t
e
t
o
e
a
r
l
y
b
r
e
a
s
t
c
a
n
c
e
r
d
i
a
g
n
o
s
is
,
u
l
ti
m
a
t
e
l
y
i
m
p
r
o
v
i
n
g
p
a
t
i
e
n
t
o
u
t
c
o
m
e
s
.
2.
M
E
T
H
O
D
2
.
1
.
Resea
rc
h f
ra
m
ewo
r
k
I
n
Fig
u
r
e
1
is
th
e
r
esear
ch
f
r
a
m
ewo
r
k
c
o
n
d
u
cted
in
th
is
s
tu
d
y
.
T
h
e
f
r
a
m
ewo
r
k
ab
o
v
e
is
d
iv
id
ed
in
to
f
o
u
r
s
tag
es
co
n
s
is
tin
g
o
f
th
e
f
ir
s
t
s
tag
e,
n
am
ely
im
ag
e
in
p
u
t.
T
h
e
im
ag
e
in
p
u
t
s
tag
e
is
th
e
f
ir
s
t
s
tag
e
in
th
is
s
tu
d
y
th
at
en
ter
s
im
ag
e
d
ata
in
to
th
e
s
y
s
tem
to
b
e
an
aly
ze
d
i
n
th
e
s
tu
d
y
.
T
h
e
s
ec
o
n
d
s
tag
e
is
p
r
ep
r
o
ce
s
s
in
g
o
r
p
r
ep
r
o
ce
s
s
in
g
wh
ich
co
n
s
is
ts
o
f
th
r
ee
p
r
e
p
r
o
ce
s
s
in
g
s
u
b
-
s
tep
s
,
n
am
ely
th
e
co
lo
r
c
o
n
v
e
r
s
io
n
o
f
th
e
im
ag
e
f
r
o
m
r
ed
g
r
ee
n
b
lu
e
(
R
GB
)
to
Gr
ay
s
ca
le,
th
e
n
im
p
r
o
v
in
g
im
ag
e
q
u
ality
(
im
a
g
e
en
h
an
ce
m
en
t)
u
s
in
g
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
2
,
Au
g
u
s
t
20
25
:
1
2
3
6
-
1
248
1238
co
n
tr
ast
s
tr
etch
in
g
m
eth
o
d
,
th
en
r
ed
u
cin
g
n
o
is
e
(
n
o
is
e
r
ed
u
ctio
n
)
u
s
in
g
th
e
m
ed
ian
f
ilter
m
eth
o
d
.
T
h
e
th
ir
d
s
tag
e
in
th
is
s
tu
d
y
is
p
r
o
ce
s
s
i
n
g
wh
ich
co
n
s
is
ts
o
f
s
ev
en
s
u
b
-
s
tep
s
o
f
th
e
p
r
o
ce
s
s
,
n
am
el
y
th
e
f
ir
s
t
r
eg
io
n
o
f
in
ter
est
(
R
OI
)
,
th
e
s
ec
o
n
d
tr
ain
in
g
an
d
v
alid
atio
n
o
f
R
OI
d
ata,
th
e
th
ir
d
im
ag
e
s
eg
m
en
tatio
n
,
th
e
f
o
u
r
t
h
ap
p
licatio
n
o
f
th
e
R
esNet
-
1
8
ar
ch
itectu
r
e
an
d
th
e
d
ev
elo
p
m
en
t
o
f
th
e
R
esNet
-
1
8
ar
c
h
itectu
r
e,
th
e
f
if
t
h
ev
alu
atio
n
,
tr
ain
in
g
an
d
v
ali
d
atio
n
o
f
th
e
R
esNet
-
1
8
m
o
d
el
b
ef
o
r
e
d
ev
elo
p
m
en
t
a
n
d
a
f
ter
d
ev
elo
p
m
en
t
(
ev
alu
atio
n
,
tr
ai
n
in
g
a
n
d
v
al
id
atio
n
ar
e
ca
r
r
ie
d
o
u
t
o
n
t
h
r
ee
v
alu
es,
n
am
el
y
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
)
,
th
e
s
ix
th
m
o
d
el
test
u
s
in
g
test
d
ata
b
ef
o
r
e
d
ev
el
o
p
m
en
t
a
n
d
a
f
ter
d
ev
elo
p
m
en
t
,
th
e
s
ev
en
t
h
is
th
e
ev
alu
atio
n
o
f
test
r
esu
lt
d
ata
b
ef
o
r
e
d
e
v
elo
p
m
e
n
t
an
d
af
t
er
d
ev
elo
p
m
en
t
(
ev
alu
atio
n
o
f
test
r
esu
lt
d
ata
is
ca
r
r
ied
o
u
t
o
n
t
h
r
ee
v
alu
es,
n
a
m
ely
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
)
.
T
h
e
f
o
u
r
th
s
tag
e
in
th
is
s
tu
d
y
is
th
e
r
esu
lts
o
f
lesi
o
n
d
etec
tio
n
in
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
es
f
r
o
m
ar
ch
itectu
r
es
th
at
h
av
e
n
o
t
b
ee
n
d
ev
elo
p
ed
with
ar
ch
itectu
r
es
th
at
h
av
e
b
ee
n
d
ev
elo
p
e
d
.
T
h
is
s
tu
d
y
f
o
llo
w
s
a
s
tr
u
ctu
r
ed
r
esear
ch
f
r
am
e
wo
r
k
d
esig
n
e
d
to
en
h
an
ce
b
r
ea
s
t
lesi
o
n
id
en
tific
atio
n
in
u
ltra
s
o
u
n
d
im
ag
es
th
r
o
u
g
h
a
m
o
d
if
ied
R
esNet
-
1
8
ar
ch
itectu
r
e.
T
h
e
f
r
am
ewo
r
k
co
n
s
is
ts
o
f
f
o
u
r
m
ain
s
tag
es:
d
ata
ac
q
u
is
itio
n
,
p
r
ep
r
o
ce
s
s
in
g
,
m
o
d
el
d
ev
el
o
p
m
en
t,
an
d
e
v
alu
atio
n
.
T
h
e
r
esear
ch
wo
r
k
f
lo
w
is
illu
s
tr
ated
in
Fig
u
r
e
1
.
Fig
u
r
e
1
.
R
esear
ch
f
r
am
ewo
r
k
2
.
2
.
I
np
ut
im
a
g
e
(
brea
s
t
ultr
a
s
o
un
d im
a
g
e)
T
h
e
u
ltra
s
o
u
n
d
im
ag
es
u
s
ed
in
th
is
s
tu
d
y
wer
e
co
llected
f
r
o
m
Pr
o
f
.
Dr
.
MA
Ha
n
if
iah
SM
B
atu
s
an
g
k
ar
Ho
s
p
ital,
o
b
tain
e
d
u
s
in
g
a
Ph
ilip
s
Af
f
in
iti
5
0
G
u
ltra
s
o
u
n
d
m
ac
h
in
e.
T
h
e
d
ataset
co
m
p
r
is
es
3
0
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
es
f
r
o
m
p
atien
ts
d
iag
n
o
s
ed
with
s
u
s
p
ec
ted
lesi
o
n
s
.
E
th
ical
a
p
p
r
o
v
a
l
was
o
b
tain
ed
,
an
d
th
e
im
ag
es
wer
e
a
n
o
n
y
m
ized
b
ef
o
r
e
p
r
o
ce
s
s
in
g
.
T
h
e
d
ataset
in
clu
d
es
v
ar
i
o
u
s
lesi
o
n
ty
p
es
to
en
s
u
r
e
m
o
d
el
r
o
b
u
s
tn
ess
.
T
h
e
u
ltra
s
o
u
n
d
i
m
ag
e
p
r
o
d
u
ce
d
b
y
th
e
d
e
v
ice
is
in
th
e
f
o
r
m
o
f
a
d
ig
ital
f
ile
f
o
r
m
at
with
th
e
ex
ten
s
io
n
*
.
jp
g
,
in
th
e
f
o
r
m
o
f
an
R
GB
im
ag
e
b
ec
au
s
e
th
e
u
ltra
s
o
u
n
d
im
a
g
e
f
r
o
m
a
p
atie
n
t
id
en
tifie
d
with
a
b
r
ea
s
t tu
m
o
r
is
R
GB
,
as in
Fig
u
r
e
2
.
Fig
u
r
e
2
.
USG
im
ag
e
o
f
b
r
ea
s
t le
s
io
n
s
with
af
f
in
iti 5
0
G
Ph
il
ip
B
r
an
d
Ma
ch
in
e
at
Pr
o
f
.
Dr
.
MA
.
Han
if
iah
SM
B
atu
s
an
g
k
ar
Ho
s
p
ital
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:
2502
-
4
7
5
2
Dev
elo
p
men
t o
f R
esN
et
-
1
8
a
r
ch
itectu
r
e
to
lesi
o
n
id
en
tifi
ca
tio
n
in
b
r
ea
s
t
…
(
S
ilfi
a
A
n
d
in
i
)
1239
2
.
3
.
P
re
pro
ce
s
s
ing
(
co
lo
r
co
nv
er
s
io
n
(
RG
B
t
o
G
ra
y
s
ca
le
)
,
im
a
g
e
enha
ncem
ent
,
no
is
e
re
du
ct
io
n
)
Pre
p
r
o
ce
s
s
in
g
is
th
e
in
itial s
te
p
in
d
ata
p
r
o
ce
s
s
in
g
th
at
aim
s
to
clea
n
an
d
f
o
r
m
at
d
ata
b
ef
o
r
e
it is
u
s
ed
in
an
aly
s
is
.
W
ith
p
r
ep
r
o
ce
s
s
in
g
,
d
ata
b
ec
o
m
es
m
o
r
e
co
n
s
i
s
ten
t
an
d
h
ig
h
q
u
ality
,
ca
n
in
cr
ea
s
e
th
e
ac
cu
r
ac
y
an
d
ef
f
ec
tiv
e
n
ess
o
f
th
e
m
o
d
el
b
u
ilt.
a)
C
o
lo
r
c
o
n
v
e
r
s
io
n
(
R
GB
to
Gr
ay
s
ca
le)
.
T
h
e
im
a
g
e
p
r
e
-
p
r
o
ce
s
s
in
g
s
tag
e
is
a
r
esear
ch
s
t
ep
ca
r
r
ied
o
u
t
b
ef
o
r
e
th
e
m
ain
p
r
o
ce
s
s
in
g
p
r
o
ce
s
s
o
f
th
e
im
ag
e
b
ein
g
s
tu
d
i
ed
[
1
3
]
.
T
h
e
m
ai
n
p
u
r
p
o
s
e
o
f
th
is
s
tag
e
is
to
p
r
o
v
id
e
t
h
e
b
est
an
d
m
o
s
t
ac
cu
r
ate
im
ag
e
d
ata
b
e
f
o
r
e
th
e
m
a
in
p
r
o
ce
s
s
in
g
p
r
o
ce
s
s
is
ca
r
r
ied
o
u
t,
s
o
th
at
af
ter
th
e
m
ain
p
r
o
ce
s
s
in
g
p
r
o
ce
s
s
th
e
r
esu
lts
o
b
tain
ed
ar
e
b
etter
,
m
o
r
e
p
r
ec
is
e,
an
d
m
o
r
e
ac
cu
r
ate
[
1
4
]
,
[
1
5
]
.
I
n
th
is
s
tu
d
y
,
th
e
f
ir
s
t
i
m
ag
e
p
r
e
-
p
r
o
ce
s
s
in
g
s
tag
e
was
ca
r
r
ied
o
u
t,
n
am
ely
co
n
v
e
r
tin
g
th
e
in
p
u
t
im
ag
e
co
lo
r
f
r
o
m
R
GB
to
Gr
ay
s
ca
le.
T
h
is
s
tep
is
tak
en
s
o
th
at
th
e
r
esu
ltin
g
im
ag
e
o
n
ly
co
n
s
is
ts
o
f
wh
ite,
b
lack
an
d
g
r
a
y
.
b)
I
m
ag
e
e
n
h
an
ce
m
e
n
t.
T
h
e
s
ec
o
n
d
s
tep
in
p
r
ep
r
o
ce
s
s
in
g
is
i
m
ag
e
en
h
an
ce
m
e
n
t,
wh
ich
is
a
s
tep
tak
en
to
clar
if
y
an
d
s
h
ar
p
en
ce
r
tain
c
h
ar
ac
ter
is
tics
/f
ea
tu
r
es
o
f
th
e
im
ag
e
s
o
th
at
th
e
im
ag
e
is
ea
s
ier
to
p
er
ce
iv
e
o
r
an
aly
ze
m
o
r
e
ca
r
e
f
u
lly
[
1
6
]
.
I
n
ad
d
itio
n
,
th
e
aim
is
to
h
ig
h
lig
h
t a
ce
r
tain
ch
ar
ac
ter
is
tic
in
th
e
im
ag
e
o
r
t
o
im
p
r
o
v
e
t
h
e
ap
p
ea
r
an
ce
asp
e
ct
[
1
7
]
.
T
h
is
s
tep
is
ca
r
r
ied
o
u
t
af
ter
th
e
R
GB
to
Gr
ay
s
ca
le
co
lo
r
p
r
e
-
p
r
o
ce
s
s
in
g
s
tag
e
f
o
r
t
h
e
USG
B
r
ea
s
t
L
esio
n
im
ag
e
q
u
ality
im
p
r
o
v
e
m
en
t
s
tag
e
[
1
8
]
.
T
h
e
m
eth
o
d
u
s
ed
t
o
im
p
r
o
v
e
t
h
is
im
ag
e
q
u
ality
is
th
e
co
n
tr
ast
s
tr
etch
in
g
m
eth
o
d
.
T
h
is
co
n
tr
ast
s
tr
etch
in
g
m
eth
o
d
ca
n
b
e
u
s
ed
to
im
p
r
o
v
e
th
e
q
u
ality
o
f
d
i
g
ital
im
ag
es
r
elate
d
to
lig
h
tin
g
,
n
am
ely
b
y
ad
j
u
s
tin
g
t
h
e
b
r
i
g
h
tn
ess
lev
el
an
d
co
n
tr
ast o
f
a
d
ig
ital im
ag
e
s
o
th
at
it c
an
b
e
u
s
ed
f
o
r
th
e
n
ex
t
p
r
e
-
p
r
o
ce
s
s
in
g
s
tag
e.
c)
No
is
e
r
ed
u
ctio
n
.
T
h
e
th
ir
d
s
tep
in
p
r
ep
r
o
ce
s
s
in
g
is
im
ag
e
n
o
is
e
r
ed
u
ctio
n
,
wh
ich
is
a
s
te
p
tak
en
t
o
clea
n
o
r
elim
in
ate
o
r
r
ed
u
ce
n
o
is
e
f
r
o
m
th
e
im
a
g
e
s
o
th
at
th
e
i
m
ag
e
is
ea
s
ier
to
p
er
ce
iv
e
o
r
an
aly
ze
m
o
r
e
ca
r
ef
u
lly
[
1
9
]
,
[
2
0
]
.
I
n
ad
d
itio
n
,
th
e
aim
is
to
h
ig
h
lig
h
t
a
ce
r
tain
f
ea
tu
r
e
in
th
e
im
ag
e
o
r
t
o
im
p
r
o
v
e
th
e
ap
p
ea
r
an
ce
asp
ec
t
[
2
1
]
.
T
h
is
s
t
ep
is
ca
r
r
ied
o
u
t
a
f
ter
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
s
tag
e
o
f
im
p
r
o
v
i
n
g
th
e
im
ag
e
q
u
ality
o
f
th
e
USG
B
r
ea
s
t
L
es
io
n
im
ag
e.
T
h
e
m
eth
o
d
u
s
ed
t
o
im
p
r
o
v
e
im
a
g
e
q
u
ality
is
th
e
m
ed
ian
f
ilter
m
eth
o
d
.
T
h
is
m
ed
ian
f
ilter
m
eth
o
d
ca
n
b
e
u
s
ed
to
r
ed
u
ce
i
m
ag
e
n
o
is
e
ass
o
ciate
d
wit
h
b
lack
o
r
wh
ite
s
p
o
ts
th
at
ar
e
ir
r
eg
u
la
r
ly
p
o
s
itio
n
ed
a
n
d
ir
r
e
g
u
lar
ly
s
h
a
p
ed
.
2
.
4
.
P
r
o
ce
s
s
ing
T
h
e
f
ir
s
t
s
tep
in
th
e
p
r
o
ce
s
s
in
g
s
tag
e
is
th
e
R
OI
.
R
OI
in
d
ig
ital
im
ag
er
y
r
ef
e
r
s
to
a
s
p
ec
if
ic
ar
ea
with
in
an
im
ag
e
th
at
is
o
f
p
r
i
m
ar
y
f
o
c
u
s
o
r
in
ter
est
in
f
u
r
th
er
an
aly
s
is
o
r
p
r
o
ce
s
s
in
g
[
2
2
]
.
R
OI
is
u
s
ed
to
lim
it
o
u
r
atten
tio
n
to
th
e
m
o
s
t
r
elev
an
t
o
r
in
ter
esti
n
g
ar
ea
s
with
in
an
im
ag
e,
th
er
e
b
y
r
ed
u
cin
g
th
e
tim
e
an
d
r
eso
u
r
ce
s
r
eq
u
ir
ed
to
an
aly
ze
o
r
p
r
o
ce
s
s
th
e
en
tire
im
a
g
e
[
2
3
]
.
R
OI
ca
n
b
e
a
r
ec
tan
g
le,
s
q
u
ar
e,
cir
cle,
o
r
o
th
er
s
h
ap
e,
d
ep
e
n
d
in
g
o
n
th
e
n
ee
d
s
an
d
ty
p
e
o
f
an
aly
s
is
b
ein
g
p
er
f
o
r
m
ed
.
R
OI
ca
n
b
e
m
an
u
ally
d
ef
in
ed
b
y
th
e
u
s
er
b
y
d
r
awin
g
a
b
o
x
o
r
o
t
h
er
s
h
ap
e
ar
o
u
n
d
th
e
ar
ea
o
f
i
n
ter
est,
o
r
in
s
o
m
e
ca
s
es,
R
O
I
ca
n
b
e
d
eter
m
i
n
ed
u
s
in
g
a
co
m
p
u
ter
alg
o
r
ith
m
th
at
d
etec
ts
s
p
ec
if
ic
f
ea
tu
r
es
o
r
o
b
jects
with
in
th
e
im
ag
e
.
T
h
e
s
ec
o
n
d
s
tep
in
th
e
p
r
o
ce
s
s
in
g
s
tag
e
is
tr
ain
in
g
a
n
d
v
alid
atin
g
th
e
R
OI
d
ata.
T
r
ain
in
g
an
d
v
alid
atio
n
d
ata
ar
e
two
i
m
p
o
r
ta
n
t
s
tag
es
in
th
e
p
r
o
ce
s
s
o
f
d
ev
el
o
p
in
g
a
m
ac
h
in
e
lear
n
i
n
g
m
o
d
el.
R
OI
r
ef
er
s
to
a
s
p
ec
if
ic
p
ar
t
o
f
th
e
d
ata
th
at
is
th
e
f
o
cu
s
o
f
t
h
e
an
al
y
s
is
.
T
r
ain
in
g
d
ata
is
th
e
d
ata
s
et
u
s
ed
t
o
tr
ain
t
h
e
m
ac
h
in
e
lear
n
in
g
m
o
d
el.
At
th
is
s
tag
e
,
th
e
m
o
d
el
will
lear
n
p
atter
n
s
an
d
r
elatio
n
s
h
ip
s
in
th
e
d
ata
to
m
ak
e
p
r
e
d
ictio
n
s
o
r
m
ak
e
d
ec
is
io
n
s
.
I
n
th
e
co
n
tex
t
o
f
R
OI
,
tr
ai
n
in
g
d
ata
will
in
clu
d
e
r
elev
a
n
t
d
ata
o
n
ly
f
r
o
m
ce
r
tain
p
a
r
ts
th
at
ar
e
c
o
n
s
id
er
ed
i
m
p
o
r
ta
n
t
o
r
in
ter
esti
n
g
to
th
e
R
OI
.
Valid
atio
n
d
ata
is
a
s
ep
ar
ate
d
ata
s
et
u
s
ed
to
ev
alu
ate
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
tr
ain
e
d
m
o
d
el.
Valid
a
tio
n
d
ata
is
n
o
t
u
s
ed
d
u
r
in
g
tr
ain
in
g
,
b
u
t
is
u
s
ed
to
p
r
o
v
id
e
an
u
n
b
iased
ass
es
s
m
en
t
o
f
th
e
m
o
d
el
’
s
ab
ilit
y
to
g
en
er
alize
t
o
n
ew
d
ata.
I
n
th
e
co
n
te
x
t
o
f
R
OI
,
v
alid
atio
n
d
ata
also
f
o
cu
s
es
o
n
s
p
ec
if
ic
p
ar
t
s
th
at
ar
e
co
n
s
id
er
e
d
im
p
o
r
tan
t.
2
.
4
.
1
.
Seg
m
ent
a
t
io
n
T
h
e
th
ir
d
s
tep
i
n
th
e
p
r
o
ce
s
s
in
g
s
tag
e
is
im
a
g
e
s
eg
m
e
n
tatio
n
.
I
m
a
g
e
s
eg
m
e
n
tatio
n
is
t
h
e
p
r
o
ce
s
s
o
f
b
r
ea
k
in
g
d
o
wn
o
r
g
r
o
u
p
in
g
b
ased
o
n
th
e
c
h
ar
ac
ter
is
tics
o
f
p
ix
els
in
th
e
im
ag
e
[
2
4
]
.
I
m
a
g
e
s
eg
m
en
tatio
n
ca
n
b
e
in
th
e
f
o
r
m
o
f
s
ep
ar
atin
g
th
e
f
o
r
eg
r
o
u
n
d
f
r
o
m
th
e
b
ac
k
g
r
o
u
n
d
o
r
g
r
o
u
p
i
n
g
p
ix
el
r
eg
io
n
s
b
ased
o
n
s
im
ilar
ity
in
co
lo
r
o
r
s
h
ap
e
[
2
5
]
.
T
h
e
p
u
r
p
o
s
e
o
f
s
eg
m
en
tati
o
n
is
to
f
ac
ilit
ate
im
ag
e
an
aly
s
is
b
y
f
o
cu
s
in
g
o
n
ce
r
tain
ar
ea
s
o
r
o
b
jects.
T
h
e
i
m
ag
e
s
eg
m
en
tatio
n
p
r
o
ce
s
s
is
b
ased
o
n
t
h
e
s
im
ilar
ity
o
f
co
lo
r
b
etwe
en
th
e
co
lo
r
o
f
ea
ch
p
i
x
el
an
d
th
e
b
ac
k
g
r
o
u
n
d
co
l
o
r
i
n
th
e
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
e.
Mu
lti
-
th
r
esh
o
ld
in
g
is
an
im
a
g
e
s
eg
m
en
tatio
n
m
et
h
o
d
wh
er
e
p
ix
els
in
th
e
im
ag
e
ar
e
d
iv
i
d
e
d
in
to
s
ev
er
al
class
es
o
r
g
r
o
u
p
s
u
s
in
g
m
o
r
e
th
an
o
n
e
th
r
esh
o
ld
v
alu
e.
T
h
e
m
ain
p
u
r
p
o
s
e
o
f
m
u
lti
-
th
r
esh
o
ld
i
n
g
is
to
s
ep
ar
ate
p
ix
els
in
to
m
o
r
e
th
an
two
g
r
o
u
p
s
,
ac
co
r
d
in
g
to
th
e
lev
el
o
f
p
ix
el
in
ten
s
ity
.
T
h
is
i
s
u
s
ef
u
l
wh
en
th
e
im
ag
e
h
as
m
o
r
e
th
an
two
ty
p
es
o
f
o
b
jects
o
r
s
tr
u
ctu
r
es
with
d
if
f
e
r
en
t
lev
el
s
o
f
in
ten
s
ity
.
At
th
is
s
tag
e
t
h
e
im
ag
e
is
co
n
v
e
r
ted
in
t
o
a
b
in
ar
y
im
ag
e
u
s
in
g
s
ev
er
al
th
r
esh
o
ld
s
,
s
o
th
at
p
ix
els ca
n
b
e
class
if
ied
in
to
s
ev
er
al
s
eg
m
en
ts
ac
co
r
d
in
g
to
th
e
t
h
r
esh
o
ld
u
s
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
2
,
Au
g
u
s
t
20
25
:
1
2
3
6
-
1
248
1240
A.
ResNet
-
1
8
a
rc
hite
ct
ure
bef
o
re
a
nd
a
f
t
er
dev
elo
pm
ent
R
esNet
-
1
8
is
p
ar
t
o
f
th
e
R
esNet
f
am
il
y
d
ev
elo
p
ed
to
ad
d
r
ess
co
m
m
o
n
p
r
o
b
lem
s
in
d
e
ep
n
eu
r
al
n
etwo
r
k
s
,
s
u
c
h
as
v
an
is
h
in
g
g
r
a
d
ien
ts
an
d
th
e
d
if
f
icu
lty
o
f
tr
ain
in
g
v
er
y
d
ee
p
n
etwo
r
k
s
[
2
6
]
.
R
esNet
in
tr
o
d
u
ce
s
t
h
e
c
o
n
ce
p
t
o
f
r
esid
u
al
lear
n
in
g
,
wh
ich
allo
ws
l
ay
er
s
in
a
n
e
u
r
al
n
etwo
r
k
to
l
ea
r
n
to
id
en
tif
y
th
e
d
if
f
er
en
ce
(
r
esid
u
al)
b
etwe
en
th
e
in
p
u
t
an
d
t
h
e
d
esire
d
o
u
tp
u
t,
r
ath
er
th
a
n
tr
y
in
g
t
o
lear
n
a
d
ir
ec
t
m
ap
[
2
7
]
.
Fo
r
m
o
r
e
d
etails,
th
e
co
m
p
let
e
s
tr
u
ctu
r
e
o
f
th
e
R
esNet
-
1
8
ar
ch
itectu
r
e
ca
n
b
e
s
ee
n
in
T
ab
le
1
.
T
h
e
m
o
d
if
ie
d
R
esNet
ar
ch
itectu
r
e
with
th
e
ad
d
itio
n
o
f
co
n
v
o
lu
tio
n
lay
er
s
c
an
b
e
s
ee
n
in
T
ab
le
2
.
T
ab
le
1
.
R
esNet
-
18
ar
ch
itectu
r
e
b
ef
o
r
e
an
d
af
ter
d
e
v
elo
p
m
e
n
t
R
e
sN
e
t
-
1
8
a
r
c
h
i
t
e
c
t
u
r
e
b
e
f
o
r
e
d
e
v
e
l
p
o
me
n
t
R
e
sN
e
t
-
1
8
a
r
c
h
i
t
e
c
t
u
r
e
a
f
t
e
r
d
e
v
e
l
p
o
men
t
La
y
e
r
Ty
p
e
La
y
e
r
Ty
p
e
1
I
n
p
u
t
l
a
y
e
r
1
I
n
p
u
t
l
a
y
e
r
2
C
o
n
v
1
(
L
a
p
i
s
a
n
K
o
n
v
o
l
u
si
Aw
a
l
)
2
C
o
n
v
1
(
L
a
p
i
s
a
n
K
o
n
v
o
l
u
si
Aw
a
l
)
3
C
o
n
v
2
_
x
(
R
e
si
d
u
a
l
B
l
o
c
k
1
)
3
C
o
n
v
2
_
x
(
R
e
si
d
u
a
l
B
l
o
c
k
1
)
4
C
o
n
v
3
_
x
(
R
e
si
d
u
a
l
B
l
o
c
k
2
)
4
C
o
n
v
2
D
+
R
e
LU
(
A
d
d
e
d
)
5
C
o
n
v
4
_
x
(
R
e
si
d
u
a
l
B
l
o
c
k
3
)
5
C
o
n
v
3
_
x
(
R
e
si
d
u
a
l
B
l
o
c
k
2
)
6
C
o
n
v
5
_
x
(
R
e
si
d
u
a
l
B
l
o
c
k
4
)
6
C
o
n
v
2
D
+
R
e
LU
(
A
d
d
e
d
)
7
F
u
l
l
y
c
o
n
n
e
c
t
e
d
l
a
y
e
r
7
C
o
n
v
4
_
x
(
R
e
si
d
u
a
l
B
l
o
c
k
3
)
8
O
u
t
p
u
t
l
a
y
e
r
8
C
o
n
v
2
D
+
R
e
LU
(
A
d
d
e
d
)
9
C
o
n
v
5
_
x
(
R
e
si
d
u
a
l
B
l
o
c
k
4
)
10
F
u
l
l
y
c
o
n
n
e
c
t
e
d
l
a
y
e
r
11
O
u
t
p
u
t
l
a
y
e
r
T
ab
le
2
.
I
n
p
u
t im
a
g
e
P
a
t
i
e
n
t
B
r
e
a
s
t
u
l
t
r
a
s
o
u
n
d
i
m
a
g
e
P
a
t
i
e
n
t
B
r
e
a
s
t
u
l
t
r
a
s
o
u
n
d
i
m
a
g
e
P
a
t
i
e
n
t
B
r
e
a
s
t
u
l
t
r
a
s
o
u
n
d
i
m
a
g
e
1
4
7
2
5
8
3
6
I
n
th
e
b
ef
o
r
e
d
ev
elo
p
m
en
t
R
esNet
-
1
8
d
iag
r
am
,
ea
ch
s
tag
e
co
n
s
is
ts
o
f
two
co
n
v
o
lu
ti
o
n
al
lay
er
s
f
o
llo
wed
b
y
m
ax
p
o
o
lin
g
(
i
n
th
e
en
co
d
er
s
ec
tio
n
)
a
n
d
u
p
s
am
p
lin
g
(
i
n
th
e
d
ec
o
d
er
s
ec
tio
n
)
with
s
k
ip
co
n
n
ec
tio
n
s
co
n
n
ec
tin
g
th
e
c
o
r
r
esp
o
n
d
i
n
g
lay
er
s
f
r
o
m
th
e
e
n
co
d
er
to
t
h
e
d
ec
o
d
er
.
I
n
th
e
Af
ter
d
ev
elo
p
m
en
t
R
esNet
-
1
8
d
iag
r
am
,
ad
d
itio
n
a
l
co
n
v
o
lu
tio
n
al
la
y
er
s
(
lab
ele
d
as
“Co
n
v
2
D
+
R
eL
U
(
Ad
d
e
d
)
”)
a
r
e
in
s
er
ted
at
s
o
m
e
s
tag
es
to
im
p
r
o
v
e
th
e
f
e
atu
r
e
ex
tr
ac
tio
n
ca
p
a
b
ilit
y
.
T
h
e
ad
d
itio
n
o
f
c
o
n
v
o
lu
tio
n
al
lay
er
s
in
th
e
R
esNet
-
1
8
ar
ch
itectu
r
e
is
ex
p
ec
ted
to
im
p
r
o
v
e
t
h
e
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
in
b
r
ea
s
t
lesi
o
n
s
eg
m
e
n
tatio
n
.
Ad
d
itio
n
al
co
n
v
o
l
u
tio
n
al
la
y
er
s
allo
w
th
e
m
o
d
el
to
lear
n
m
o
r
e
c
o
m
p
l
ex
an
d
d
iv
er
s
e
f
ea
tu
r
es.
E
ac
h
co
n
v
o
lu
tio
n
al
lay
er
ca
p
tu
r
es
v
ar
io
u
s
p
atter
n
s
an
d
tex
tu
r
es
f
r
o
m
th
e
in
p
u
t
d
ata.
B
y
ad
d
in
g
m
o
r
e
lay
e
r
s
,
th
e
m
o
d
el
ca
n
ca
p
tu
r
e
h
ig
h
er
-
le
v
el
f
ea
tu
r
es,
s
u
c
h
as
f
in
e
ed
g
es,
te
x
tu
r
e
p
atter
n
s
,
a
n
d
m
o
r
e
co
m
p
lex
s
h
a
p
es
th
at
m
ay
b
e
p
r
esen
t
i
n
b
r
ea
s
t
lesi
o
n
im
ag
es.
W
ith
m
o
r
e
co
n
v
o
l
u
tio
n
al
lay
er
s
,
th
e
m
o
d
el
ca
n
id
e
n
tify
f
in
er
d
etai
ls
in
th
e
im
a
g
e.
T
h
is
is
esp
ec
ially
im
p
o
r
tan
t
in
m
ed
ical
s
eg
m
en
tatio
n
,
wh
er
e
s
m
all
d
etails
ca
n
in
d
icat
e
th
e
p
r
esen
ce
o
r
ch
ar
ac
ter
is
tics
o
f
s
ig
n
if
ican
t
lesi
o
n
s
.
Mo
r
e
d
etailed
f
ea
t
u
r
e
d
etec
tio
n
im
p
r
o
v
es
th
e
m
o
d
el
’
s
ab
ilit
y
to
d
is
tin
g
u
is
h
b
etwe
en
n
o
r
m
al
tis
s
u
e
an
d
tis
s
u
e
co
n
tain
in
g
lesi
o
n
s
.
B.
Co
nv
o
lutio
n
l
a
y
er
s
ResNet
-
18
T
h
e
co
n
v
o
l
u
tio
n
lay
er
in
R
esNet
-
1
8
is
th
e
co
r
e
o
f
t
h
e
a
r
ch
it
ec
tu
r
e
th
at
p
r
o
ce
s
s
es
th
e
in
p
u
t
im
ag
e,
th
e
d
etails o
f
wh
ich
ar
e
as f
o
llo
ws:
-
I
n
itial
co
n
v
o
lu
tio
n
lay
er
(
C
o
n
v
1
)
:
t
h
is
is
th
e
f
ir
s
t
co
n
v
o
l
u
tio
n
lay
er
th
at
u
s
es
6
4
f
ilter
s
with
a
lar
g
e
k
er
n
el
(
7
x
7
)
[
2
8
]
.
T
h
is
k
er
n
el
s
ize
is
lar
g
er
th
an
u
s
u
al
to
c
ap
tu
r
e
m
o
r
e
c
o
n
tex
t
o
f
th
e
o
v
er
all
im
ag
e
at
th
e
b
eg
in
n
in
g
o
f
t
h
e
n
etwo
r
k
.
Pad
d
in
g
o
f
3
is
u
s
ed
t
o
en
s
u
r
e
th
e
o
u
tp
u
t
h
as
th
e
d
e
s
ir
ed
d
im
en
s
io
n
.
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:
2502
-
4
7
5
2
Dev
elo
p
men
t o
f R
esN
et
-
1
8
a
r
ch
itectu
r
e
to
lesi
o
n
id
en
tifi
ca
tio
n
in
b
r
ea
s
t
…
(
S
ilfi
a
A
n
d
in
i
)
1241
Strid
e
2
is
u
s
ed
to
r
ed
u
ce
th
e
im
ag
e
s
ize.
Fo
llo
wed
b
y
a
m
ax
p
o
o
lin
g
o
p
er
atio
n
th
at
f
u
r
th
er
r
ed
u
ce
s
th
e
s
p
atial
d
im
en
s
io
n
.
-
R
esid
u
al
b
lo
ck
s
(
C
o
n
v
2
_
x
,
C
o
n
v
3
_
x
,
C
o
n
v
4
_
x
,
C
o
n
v
5
_
x
)
:
e
ac
h
r
esid
u
al
b
lo
ck
c
o
n
s
is
t
s
o
f
two
co
n
v
o
l
u
tio
n
lay
e
r
s
,
with
f
ilter
s
o
f
s
ize
3
x
3
[
1
]
.
Pad
d
i
n
g
1
is
u
s
ed
to
m
ain
tain
th
e
s
am
e
o
u
t
p
u
t
d
im
en
s
io
n
as
th
e
in
p
u
t
(
ex
ce
p
t
wh
en
s
tr
id
e
is
ap
p
lied
)
.
B
atch
n
o
r
m
al
izatio
n
is
ap
p
lied
a
f
ter
ea
c
h
c
o
n
v
o
lu
ti
o
n
to
s
tab
ilize
th
e
o
u
tp
u
t
d
is
tr
ib
u
tio
n
.
R
eL
U
ac
tiv
atio
n
f
u
n
cti
o
n
is
ap
p
lied
af
ter
n
o
r
m
aliza
tio
n
.
T
h
ese
co
n
v
o
l
u
tio
n
al
lay
er
s
ar
e
d
esig
n
ed
to
wo
r
k
with
s
h
o
r
tcu
t
co
n
n
ec
tio
n
s
th
at
allo
w
th
e
o
u
tp
u
t
o
f
o
n
e
b
lo
c
k
to
b
e
p
ass
ed
to
th
e
n
ex
t
b
lo
ck
with
o
u
t m
o
d
if
icatio
n
(
o
r
with
m
in
o
r
s
ize
ad
ju
s
tm
en
ts
if
n
ec
e
s
s
ar
y
)
.
C.
M
o
del
ResNet
-
18
T
h
is
s
u
b
-
ch
ap
ter
will
d
is
cu
s
s
in
d
etail
th
e
R
esNet
-
1
8
m
o
d
el,
wh
ich
is
o
n
e
o
f
th
e
m
o
s
t
in
f
lu
en
tial
d
ee
p
lear
n
in
g
ar
ch
itectu
r
es.
T
h
e
R
esNet
-
1
8
m
o
d
el
as
a
wh
o
le
is
a
d
ee
p
lear
n
in
g
m
o
d
el
co
n
s
is
tin
g
o
f
1
8
lay
er
s
,
in
clu
d
in
g
co
n
v
o
l
u
tio
n
al
lay
er
s
an
d
r
esid
u
al
b
lo
ck
s
.
S
o
m
e
im
p
o
r
tan
t
p
o
i
n
ts
ab
o
u
t
th
is
m
o
d
el
ca
n
b
e
ex
p
lain
ed
as f
o
llo
ws:
-
E
f
f
ec
tiv
en
ess
o
f
r
esid
u
al
lea
r
n
in
g
:
t
h
is
m
o
d
el
u
tili
ze
s
s
h
o
r
tcu
t
c
o
n
n
ec
tio
n
s
to
s
k
ip
o
n
e
o
r
m
o
r
e
co
n
v
o
l
u
tio
n
al
lay
e
r
s
[
29
]
.
T
h
is
allo
ws
th
e
n
etwo
r
k
to
m
o
r
e
ea
s
ily
lear
n
th
e
id
en
tity
f
u
n
ctio
n
,
wh
ic
h
m
ea
n
s
th
at
th
e
n
etwo
r
k
o
n
l
y
n
ee
d
s
to
lear
n
th
e
d
if
f
er
en
ce
(
r
esid
u
al)
b
etwe
en
th
e
i
n
p
u
t
a
n
d
th
e
d
esire
d
o
u
tp
u
t.
-
T
r
ain
in
g
ad
v
a
n
tag
e:
d
u
e
to
th
e
r
esid
u
al
ar
ch
itectu
r
e,
R
esNe
t
-
1
8
is
ea
s
ier
to
tr
ain
ev
en
wh
en
th
e
n
etwo
r
k
is
v
er
y
d
ee
p
[
3
0
]
.
T
h
is
o
v
er
c
o
m
es
a
co
m
m
o
n
p
r
o
b
lem
in
t
r
ad
itio
n
al
n
e
u
r
al
n
etwo
r
k
s
wh
er
e
th
e
d
ee
p
er
th
e
n
etwo
r
k
,
th
e
h
a
r
d
er
it is
to
tr
ain
it e
f
f
ec
tiv
ely
d
u
e
to
v
an
is
h
in
g
g
r
ad
ien
ts
.
-
Ap
p
licatio
n
s
:
R
esNe
t
-
1
8
is
c
o
m
m
o
n
l
y
u
s
ed
f
o
r
im
ag
e
cla
s
s
if
icatio
n
task
s
o
n
lar
g
e
d
atasets
s
u
ch
as
I
m
ag
eNe
t
[3
1
]
.
Ho
wev
er
,
d
u
e
to
its
f
lex
ib
le
ar
ch
itectu
r
e,
R
esNet
-
1
8
ca
n
also
b
e
ad
ap
ted
f
o
r
v
ar
i
o
u
s
o
th
er
task
s
s
u
c
h
as
o
b
ject
d
etec
tio
n
,
im
ag
e
s
eg
m
e
n
tatio
n
,
a
n
d
m
o
r
e.
Usi
n
g
tr
a
n
s
f
er
lea
r
n
i
n
g
,
t
h
e
m
o
d
el
ca
n
b
e
ap
p
lied
to
s
m
aller
d
ata
s
ets an
d
tailo
r
ed
f
o
r
d
if
f
e
r
en
t
ap
p
licatio
n
s
.
-
E
f
f
icien
cy
:
R
esNet
-
1
8
is
lig
h
ter
an
d
less
co
m
p
lex
th
a
n
lar
g
er
R
esNet
v
ar
ian
ts
(
s
u
ch
as
R
esNet
-
5
0
o
r
R
esNet
-
1
0
1
)
,
m
ak
in
g
it
f
aster
to
tr
ain
a
n
d
m
o
r
e
e
f
f
icien
t
in
m
em
o
r
y
u
s
ag
e
[
3
2
]
.
T
h
is
m
a
k
es
it
s
u
itab
le
f
o
r
ap
p
licatio
n
s
th
at
r
eq
u
ir
e
d
e
ep
n
et
wo
r
k
s
b
u
t a
r
e
lim
ited
in
co
m
p
u
tin
g
r
eso
u
r
ce
s
.
-
R
esNet
-
1
8
o
f
f
er
s
a
b
alan
ce
b
etwe
en
d
ep
th
an
d
ef
f
icien
c
y
,
m
ak
in
g
it
a
p
o
p
u
lar
ch
o
i
ce
f
o
r
m
an
y
co
m
p
u
ter
v
is
io
n
ap
p
licatio
n
s
,
esp
ec
ially
in
s
itu
atio
n
s
wh
er
e
r
eso
u
r
ce
s
ar
e
lim
ited
o
r
s
p
ee
d
is
a
p
r
io
r
ity
[3
3
].
D.
E
v
a
l
ua
t
io
n a
nd
det
ec
t
io
n
Af
ter
cr
ea
tin
g
th
e
R
esNet
-
1
8
ar
ch
itectu
r
e,
co
n
v
o
l
u
tio
n
la
y
er
s
an
d
m
o
d
els,
th
e
n
ex
t
s
t
ep
in
th
is
r
esear
ch
is
to
ev
alu
ate,
tr
ain
an
d
v
alid
ate
th
e
m
o
d
el.
Mo
d
el
tr
ain
in
g
is
d
o
n
e
to
tr
ain
th
e
r
esu
ltin
g
m
o
d
el.
Mo
d
el
ev
alu
atio
n
is
d
o
n
e
t
o
f
in
d
o
u
t
wh
eth
er
th
e
r
es
u
lt
in
g
m
o
d
el
is
in
ac
co
r
d
a
n
ce
with
th
e
r
esear
ch
o
b
jectiv
es.
Mo
d
el
tr
ain
in
g
is
d
o
n
e
to
f
in
d
o
u
t
wh
eth
er
th
e
r
esu
ltin
g
m
o
d
el
ca
n
b
e
ap
p
lied
to
th
e
ca
s
e
in
th
is
s
tu
d
y
.
Mo
d
el
v
alid
atio
n
is
d
o
n
e
to
f
in
d
o
u
t
wh
eth
er
th
e
r
esu
ltin
g
m
o
d
el
is
c
o
r
r
ec
t
(
v
alid
)
o
r
n
o
t.
T
esti
n
g
t
h
e
n
ew
R
esNet
-
1
8
m
o
d
el
u
s
in
g
t
est d
ata
to
d
etec
t b
r
ea
s
t le
s
io
n
s
is
th
e
p
r
o
ce
s
s
o
f
ev
alu
atin
g
t
h
e
p
er
f
o
r
m
an
ce
o
f
a
s
eg
m
en
tatio
n
m
o
d
el
th
at
h
as
b
ee
n
tr
ain
ed
u
s
in
g
a
n
ew
d
ataset
th
at
h
as
n
ev
er
b
ee
n
s
ee
n
b
y
th
e
m
o
d
el.
T
h
is
test
d
ata
is
u
s
ed
to
ass
es
s
th
e
m
o
d
el
’
s
g
en
er
aliza
tio
n
ab
ilit
y
,
n
a
m
ely
its
ab
ilit
y
t
o
p
r
o
v
id
e
ac
c
u
r
ate
p
r
ed
ictio
n
s
o
n
d
ata
th
at
h
as
n
ev
er
b
ee
n
u
s
ed
d
u
r
in
g
th
e
tr
ain
in
g
an
d
v
alid
atio
n
p
r
o
ce
s
s
.
Her
e
ar
e
th
e
alg
o
r
ith
m
1
in
th
is
test
in
g
p
r
o
ce
s
s
:
Alg
o
r
ith
m
1
.
Mo
d
el
test
in
g
u
s
in
g
test
d
ata
1.
T
est
d
ata
p
r
ep
a
r
atio
n
:
th
e
test
d
ata
is
a
s
et
o
f
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
es
c
o
n
tain
in
g
lesi
o
n
s
,
an
d
th
is
d
ata
is
s
ep
ar
ate
f
r
o
m
t
h
e
tr
ain
in
g
an
d
v
alid
atio
n
d
ata.
Ma
k
e
s
u
r
e
th
is
d
ata
is
r
ep
r
esen
tativ
e
an
d
h
as
th
e
co
r
r
ec
t
an
n
o
tatio
n
s
.
2.
T
est
d
ata
p
r
ep
r
o
ce
s
s
in
g
:
s
im
il
ar
to
th
e
tr
ain
i
n
g
an
d
v
alid
atio
n
d
ata,
th
e
test
d
ata
n
ee
d
s
to
b
e
p
r
ep
r
o
ce
s
s
ed
in
th
e
s
am
e
way
,
s
u
ch
as c
o
n
v
er
tin
g
to
g
r
ay
s
ca
le,
n
o
r
m
alizin
g
,
an
d
d
ata
au
g
m
en
tatio
n
i
f
n
e
ce
s
s
ar
y
.
3.
Pre
d
ictio
n
u
s
in
g
th
e
m
o
d
el:
u
s
e
th
e
tr
ain
ed
R
esNet
-
1
8
m
o
d
el
to
p
er
f
o
r
m
s
eg
m
en
tatio
n
p
r
ed
ictio
n
s
o
n
th
e
test
d
ata.
4.
T
h
r
esh
o
ld
in
g
:
i
f
th
e
m
o
d
el
g
e
n
er
ates p
r
o
b
a
b
ilit
ies,
ap
p
ly
th
r
esh
o
ld
in
g
to
o
b
tain
b
in
ar
y
s
eg
m
en
t m
ap
s
.
5.
Per
f
o
r
m
an
ce
ev
al
u
atio
n
:
co
m
p
ar
e
th
e
m
o
d
el
p
r
ed
ictio
n
r
es
u
lts
with
th
e
o
r
ig
in
al
an
n
o
tatio
n
s
to
ca
lcu
late
ev
alu
atio
n
m
etr
ics
s
u
ch
as
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
s
p
ec
if
i
city
,
I
n
ter
s
ec
tio
n
o
v
er
Un
io
n
(
I
o
U)
,
d
ice
co
ef
f
icien
t,
etc.
6.
R
esu
lts
v
is
u
aliza
tio
n
:
v
is
u
aliz
e
s
o
m
e
ex
am
p
les
o
f
p
r
ed
ictio
n
r
esu
lts
to
v
is
u
ally
ch
e
ck
t
h
e
s
eg
m
en
tatio
n
q
u
ality
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
2
,
Au
g
u
s
t
20
25
:
1
2
3
6
-
1
248
1242
E
v
alu
atio
n
o
f
test
d
ata
ag
ain
s
t th
e
R
esNet
-
1
8
m
o
d
el
f
o
r
d
ete
ctin
g
b
r
ea
s
t le
s
io
n
s
is
a
p
r
o
ce
s
s
in
wh
ich
th
e
p
er
f
o
r
m
an
ce
o
f
a
s
eg
m
e
n
tatio
n
m
o
d
el
is
ass
es
s
ed
u
s
in
g
a
d
ataset
th
at
was
n
o
t
p
r
ev
io
u
s
ly
s
ee
n
d
u
r
in
g
tr
ain
in
g
(
test
d
ata)
.
T
h
is
p
r
o
ce
s
s
in
v
o
lv
es
th
e
u
s
e
o
f
s
ep
ar
ate
test
d
ata
,
co
n
s
i
s
t
en
t
p
r
ep
r
o
ce
s
s
in
g
,
co
m
p
r
eh
e
n
s
iv
e
ev
alu
atio
n
m
et
r
ic
m
ea
s
u
r
em
en
ts
,
a
n
d
in
-
d
ep
t
h
an
aly
s
is
o
f
th
e
r
esu
lts
.
W
ith
p
r
o
p
er
ev
alu
atio
n
,
we
ca
n
o
b
jectiv
ely
ass
ess
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
a
n
d
m
ak
e
b
etter
d
ec
is
io
n
s
ab
o
u
t
its
u
s
e
in
clin
ical
co
n
tex
ts
.
T
h
e
r
e
a
r
e
th
r
ee
p
a
r
a
m
eter
v
alu
es
u
s
ed
to
ev
al
u
ate,
tr
ain
an
d
v
alid
ate
th
e
r
esu
lti
n
g
m
o
d
el,
n
am
ely
ac
cu
r
ac
y
,
s
en
s
itiv
ity
an
d
s
p
ec
if
icity
v
alu
es.
Acc
u
r
ac
y
,
s
en
s
itiv
ity
an
d
s
p
ec
i
f
icity
v
al
u
es
ar
e
e
v
alu
atio
n
m
etr
ics
u
s
ed
to
as
s
ess
th
e
p
er
f
o
r
m
an
ce
o
f
a
class
if
icat
io
n
m
o
d
el,
esp
ec
ially
in
th
e
co
n
tex
t
o
f
m
ed
ical
task
s
s
u
ch
as lesio
n
d
etec
tio
n
in
b
r
e
ast u
ltra
s
o
u
n
d
im
ag
es.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
3
.
1
.
I
np
ut
im
a
g
e
re
s
ult
I
n
th
is
r
esear
ch
we
u
s
e
3
0
US
G
b
r
ea
s
t
l
esio
n
im
ag
es.
I
n
th
i
s
p
ap
er
,
we
o
n
ly
d
is
p
lay
ed
th
e
8
im
ag
es
f
o
r
th
e
s
am
p
le
o
f
th
e
r
esu
lt
o
f
th
is
r
esear
ch
.
i
n
th
e
p
r
o
ce
s
s
in
g
s
tag
e.
W
e
g
et
th
e
im
ag
e
f
r
o
m
Pro
f
.
Dr
.
MA
.
Han
if
iah
SM
B
atu
s
an
g
k
ar
Ho
s
p
ital.
T
ab
le
3
illu
s
tr
ates
th
e
s
et
o
f
in
p
u
t
im
ag
es
u
s
ed
in
th
e
cu
r
r
en
t
s
tu
d
y
,
co
n
s
is
tin
g
o
f
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
es
f
r
o
m
eig
h
t
p
atien
ts
.
E
ac
h
im
a
g
e
s
er
v
es
as
a
v
ital
co
m
p
o
n
en
t
in
th
e
in
v
esti
g
atio
n
aim
ed
at
an
aly
zin
g
an
d
d
etec
tin
g
p
o
ten
tial
b
r
ea
s
t
ab
n
o
r
m
alities
.
T
h
e
im
ag
es
ar
e
m
eth
o
d
ically
o
r
g
an
ized
in
to
two
co
lu
m
n
s
f
o
r
p
atien
ts
an
d
th
ei
r
co
r
r
e
s
p
o
n
d
in
g
u
ltra
s
o
u
n
d
s
ca
n
s
,
e
n
s
u
r
in
g
clar
ity
an
d
s
tr
u
ctu
r
ed
r
e
p
r
esen
tatio
n
.
T
h
e
s
e
s
ca
n
s
,
d
er
iv
ed
f
r
o
m
d
iv
er
s
e
ca
s
es,
f
o
r
m
th
e
f
o
u
n
d
atio
n
o
f
th
e
an
aly
s
is
aim
ed
at
ex
p
lo
r
in
g
d
iag
n
o
s
tic
p
atter
n
s
an
d
en
h
an
cin
g
im
ag
e
-
b
ased
ass
ess
m
en
t
tech
n
iq
u
es.
T
h
e
ar
r
a
n
g
em
e
n
t
p
r
o
v
id
es
an
ac
ce
s
s
ib
le
r
e
f
er
e
n
ce
f
o
r
c
o
m
p
ar
in
g
t
h
e
v
is
u
al
ch
ar
ac
ter
is
tics
o
f
ea
ch
im
ag
e,
f
ac
ilit
atin
g
f
u
r
th
e
r
in
ter
p
r
etatio
n
in
s
u
b
s
eq
u
en
t se
ctio
n
s
o
f
th
e
s
tu
d
y
.
T
ab
le
3
.
Pre
-
p
r
o
ce
s
s
in
g
r
esu
lt
P
a
t
i
e
n
t
B
r
e
a
s
t
u
l
t
r
a
s
o
u
n
d
i
m
a
g
e
R
G
B
t
o
g
r
a
y
s
c
a
l
e
I
mag
e
e
n
h
a
n
c
e
me
n
t
N
o
i
se
r
e
d
u
c
t
i
o
n
1
2
3
4
5
6
7
8
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:
2502
-
4
7
5
2
Dev
elo
p
men
t o
f R
esN
et
-
1
8
a
r
ch
itectu
r
e
to
lesi
o
n
id
en
tifi
ca
tio
n
in
b
r
ea
s
t
…
(
S
ilfi
a
A
n
d
in
i
)
1243
T
ab
le
2
r
e
v
ea
ls
th
e
v
ar
iab
ilit
y
in
u
ltra
s
o
u
n
d
im
ag
e
f
ea
tu
r
es
a
cr
o
s
s
d
if
f
er
en
t p
atien
ts
,
wh
ich
is
p
iv
o
tal
f
o
r
th
o
r
o
u
g
h
d
iag
n
o
s
tic
ev
a
lu
atio
n
s
.
E
ac
h
s
ca
n
d
em
o
n
s
tr
ates
u
n
iq
u
e
attr
ib
u
tes
th
at
ar
e
ess
en
tial
f
o
r
d
is
tin
g
u
is
h
in
g
b
etwe
en
d
if
f
er
e
n
t ty
p
es o
f
b
r
ea
s
t tis
s
u
e
s
tr
u
ctu
r
es.
T
h
ese
v
ar
iatio
n
s
u
n
d
er
s
c
o
r
e
th
e
n
ec
ess
ity
o
f
a
co
m
p
r
eh
en
s
iv
e
ass
ess
m
en
t
m
eth
o
d
th
at
c
o
n
s
id
er
s
d
i
v
er
s
e
im
ag
e
ch
ar
ac
ter
is
tics
,
in
clu
d
in
g
tex
t
u
r
e,
d
en
s
ity
,
an
d
b
o
u
n
d
a
r
y
d
elin
ea
tio
n
s
.
T
h
e
in
clu
s
io
n
o
f
m
u
ltip
le
p
atien
ts
’
u
ltra
s
o
u
n
d
im
ag
es
e
n
r
ich
es
th
e
d
ataset,
p
r
o
v
id
i
n
g
a
m
o
r
e
r
o
b
u
s
t
b
a
s
is
f
o
r
d
ev
elo
p
i
n
g
an
d
v
ali
d
atin
g
im
ag
e
a
n
aly
s
is
alg
o
r
i
th
m
s
.
T
h
is
d
iv
er
s
e
r
ep
r
esen
tatio
n
s
u
p
p
o
r
ts
th
e
s
tu
d
y
’
s
aim
o
f
en
s
u
r
in
g
th
at
a
n
y
p
r
o
p
o
s
ed
d
iag
n
o
s
tic
o
r
d
e
tectio
n
m
o
d
el
ca
n
g
en
er
alize
ef
f
ec
tiv
ely
ac
r
o
s
s
d
if
f
er
en
t
c
ases
,
u
ltima
tely
co
n
tr
ib
u
tin
g
to
m
o
r
e
r
eliab
le
an
d
ac
cu
r
ate
d
iag
n
o
s
tic
o
u
tco
m
es in
b
r
ea
s
t h
ea
lth
ass
ess
m
en
ts
.
3
.
2
.
P
re
-
pro
ce
s
s
ing
re
s
ult
T
h
e
f
ir
s
t
im
ag
e
p
r
e
p
r
o
ce
s
s
in
g
in
th
is
s
tu
d
y
is
a
p
r
o
ce
s
s
ca
r
r
ied
o
u
t
b
e
f
o
r
e
th
e
m
ain
p
r
o
c
ess
in
th
e
s
tu
d
y
.
First
p
r
ep
r
o
ce
s
s
in
g
r
esu
lt
is
co
lo
r
co
n
v
er
s
io
n
f
r
o
m
R
GB
to
g
r
ay
s
ca
le.
T
h
en
th
e
s
ec
o
n
d
p
r
e
p
r
o
ce
s
s
in
g
r
esu
lt
is
im
ag
e
e
n
h
an
ce
m
e
n
t
a
n
d
af
te
r
th
at
th
e
t
h
ir
d
p
r
ep
r
o
c
ess
in
g
r
esu
lt
in
s
n
o
is
e
r
ed
u
cti
o
n
.
T
a
b
le
4
p
r
esen
ts
th
e
co
m
p
r
eh
en
s
iv
e
r
esu
lts
o
f
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
s
tep
s
ap
p
li
ed
to
t
h
e
b
r
ea
s
t
u
ltra
s
o
u
n
d
i
m
ag
es
u
s
ed
in
th
is
s
tu
d
y
.
T
h
e
tab
le
s
y
s
tem
atica
lly
illu
s
tr
ates
th
e
p
r
o
g
r
ess
io
n
o
f
ea
ch
p
atien
t
’
s
u
ltra
s
o
u
n
d
im
a
g
e
th
r
o
u
g
h
v
ar
io
u
s
p
r
e
-
p
r
o
ce
s
s
in
g
s
tag
es,
in
clu
d
in
g
co
n
v
er
s
io
n
f
r
o
m
R
GB
to
g
r
ay
s
ca
le,
im
ag
e
en
h
an
ce
m
en
t,
an
d
n
o
is
e
r
ed
u
ctio
n
.
T
h
ese
s
tep
s
ar
e
cr
iti
ca
l in
im
p
r
o
v
in
g
th
e
q
u
ality
a
n
d
clar
ity
o
f
th
e
in
p
u
t im
ag
es,
th
er
eb
y
f
ac
ilit
atin
g
m
o
r
e
ac
cu
r
ate
a
n
aly
s
is
an
d
i
n
ter
p
r
etatio
n
in
s
u
b
s
eq
u
en
t
p
r
o
ce
s
s
in
g
s
tag
es.
T
h
e
o
r
g
a
n
izatio
n
o
f
th
e
ta
b
le
allo
ws
f
o
r
an
ea
s
y
co
m
p
ar
is
o
n
o
f
im
ag
es
b
ef
o
r
e
a
n
d
a
f
ter
ea
ch
p
r
e
-
p
r
o
ce
s
s
in
g
tec
h
n
iq
u
e,
em
p
h
asizin
g
th
e
tr
an
s
f
o
r
m
atio
n
an
d
im
p
r
o
v
em
en
ts
at
ea
ch
s
tep
.
T
ab
le
4
.
Pro
ce
s
s
in
g
r
esu
lt
P
a
t
i
e
n
t
N
o
i
se
r
e
d
u
c
t
i
o
n
R
O
I
S
e
g
m
e
n
t
a
t
i
o
n
1
2
3
4
5
6
7
8
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
2
,
Au
g
u
s
t
20
25
:
1
2
3
6
-
1
248
1244
A
d
etailed
an
aly
s
is
o
f
T
ab
le
3
u
n
d
e
r
s
co
r
es
th
e
tr
a
n
s
f
o
r
m
ativ
e
im
p
ac
t
o
f
ea
ch
p
r
e
-
p
r
o
ce
s
s
i
n
g
s
tep
o
n
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
es
an
d
th
e
cu
m
u
lativ
e
b
en
e
f
its
th
ey
b
r
in
g
to
th
e
im
a
g
e
an
aly
s
is
p
r
o
ce
s
s
.
T
h
e
in
itial
s
tep
,
R
GB
to
g
r
ay
s
ca
le
co
n
v
er
s
io
n
,
is
ess
en
tial
in
s
i
m
p
lify
in
g
th
e
im
ag
e
d
ata
b
y
elim
in
atin
g
co
lo
r
in
f
o
r
m
atio
n
wh
ile
r
etain
in
g
cr
itical
in
ten
s
ity
d
etails
th
at
ar
e
s
ig
n
if
ican
t
f
o
r
m
ed
ical
im
a
g
e
an
aly
s
is
.
T
h
is
s
tep
r
ed
u
ce
s
co
m
p
u
tatio
n
al
co
m
p
le
x
ity
an
d
en
s
u
r
es
th
at
th
e
f
o
cu
s
is
s
o
lely
o
n
th
e
s
tr
u
ctu
r
al
an
d
tex
tu
r
al
d
etails
o
f
th
e
tis
s
u
e
.
T
h
e
s
u
b
s
eq
u
en
t
im
ag
e
en
h
an
ce
m
en
t
s
tep
s
er
v
es
as
a
v
ital
p
r
o
ce
s
s
th
at
am
p
lifie
s
th
e
v
is
u
al
c
o
n
tr
ast
an
d
s
h
ar
p
n
ess
o
f
th
e
im
ag
es.
T
h
is
en
h
an
ce
m
en
t
r
e
v
ea
ls
m
o
r
e
d
e
f
in
ed
b
o
u
n
d
ar
ies
a
n
d
s
tr
u
ctu
r
es,
m
ak
i
n
g
f
ea
tu
r
es
s
u
ch
as
les
io
n
s
o
r
t
is
s
u
e
ab
n
o
r
m
alities
m
o
r
e
d
is
ce
r
n
ib
le.
T
h
e
co
n
s
is
ten
t
im
p
r
o
v
em
en
t
in
im
ag
e
clar
ity
ac
r
o
s
s
all
p
atien
t
im
ag
es
in
th
is
s
tep
d
em
o
n
s
tr
ates
th
e
r
eliab
ilit
y
an
d
ef
f
e
ctiv
en
ess
o
f
th
e
en
h
an
ce
m
e
n
t
alg
o
r
ith
m
u
s
ed
.
T
h
e
f
in
al
s
t
ag
e,
n
o
is
e
r
e
d
u
ctio
n
,
p
lay
s
a
s
ig
n
if
ican
t
r
o
le
in
r
e
f
in
in
g
th
e
q
u
a
lity
o
f
th
e
p
r
o
ce
s
s
ed
im
ag
es
b
y
m
in
im
iz
in
g
u
n
wan
ted
ar
tifa
cts
an
d
b
a
ck
g
r
o
u
n
d
n
o
is
e.
T
h
is
s
tep
e
n
s
u
r
es
th
at
ess
en
tial
im
ag
e
d
etails
ar
e
p
r
eser
v
e
d
wh
ile
ir
r
elev
an
t
v
is
u
al
in
f
o
r
m
atio
n
is
s
u
p
p
r
ess
ed
.
Ov
er
a
ll,
th
e
p
r
o
g
r
ess
io
n
illu
s
tr
ated
in
T
ab
le
3
in
d
icate
s
a
s
y
s
tem
atic
an
d
co
m
p
r
eh
e
n
s
iv
e
p
r
e
-
p
r
o
ce
s
s
in
g
ap
p
r
o
ac
h
th
at
s
ig
n
if
ican
tly
b
o
o
s
ts
th
e
in
ter
p
r
etab
ilit
y
an
d
q
u
ality
o
f
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
es.
T
h
is
m
u
lti
-
s
tag
e
p
r
e
-
p
r
o
ce
s
s
in
g
p
ip
elin
e
is
v
ital
f
o
r
ac
h
iev
in
g
r
eliab
le
im
ag
e
an
aly
s
is
o
u
tco
m
es,
as
it
p
r
ep
ar
es
th
e
d
ata
in
a
way
th
at
m
ax
im
izes
th
e
ac
cu
r
ac
y
an
d
e
f
f
icien
cy
o
f
s
u
b
s
eq
u
en
t
d
iag
n
o
s
tic
alg
o
r
ith
m
s
.
T
h
e
en
h
an
ce
d
clar
ity
,
r
ed
u
ce
d
n
o
is
e,
an
d
o
p
tim
ized
co
n
tr
ast
p
r
o
v
id
ed
b
y
th
ese
p
r
e
-
p
r
o
ce
s
s
in
g
s
tep
s
lay
th
e
f
o
u
n
d
atio
n
f
o
r
m
o
r
e
r
o
b
u
s
t
an
d
ac
cu
r
ate
f
ea
tu
r
e
ex
tr
a
ctio
n
,
wh
ich
is
cr
i
tical
in
th
e
ea
r
ly
d
etec
tio
n
an
d
an
aly
s
is
o
f
b
r
ea
s
t tis
s
u
e
an
o
m
alies.
3
.
2
.
1
.
P
ro
ce
s
s
ing
re
s
ult
T
h
e
p
r
o
ce
s
s
o
f
ca
lcu
latin
g
th
e
R
OI
v
alu
e
o
n
a
n
im
ag
e
is
t
o
ca
lcu
late
th
e
v
alu
e
o
f
t
h
e
p
lace
to
b
e
an
aly
ze
d
o
n
a
n
im
ag
e
.
I
n
th
is
s
tu
d
y
,
it
is
th
e
lesi
o
n
ar
ea
f
o
u
n
d
i
n
th
e
b
r
ea
s
t
USG
im
ag
e.
T
ab
le
5
s
h
o
wca
s
es
th
e
p
r
o
ce
s
s
in
g
r
esu
lts
o
b
tain
e
d
f
r
o
m
th
e
b
r
ea
s
t
u
ltra
s
o
u
n
d
i
m
ag
es
u
s
ed
in
th
is
s
tu
d
y
.
T
h
e
tab
le
is
s
tr
u
ctu
r
ed
to
d
is
p
lay
th
e
o
u
tc
o
m
es
at
d
if
f
e
r
en
t
s
tag
es
o
f
th
e
im
ag
e
p
r
o
ce
s
s
in
g
wo
r
k
f
lo
w,
in
c
l
u
d
in
g
n
o
is
e
r
ed
u
ctio
n
,
R
OI
ex
tr
ac
tio
n
,
an
d
s
eg
m
en
tatio
n
.
E
ac
h
c
o
lu
m
n
in
th
e
ta
b
le
h
ig
h
lig
h
ts
a
cr
u
cial
s
tep
in
t
h
e
im
ag
e
an
aly
s
is
p
ip
elin
e,
d
em
o
n
s
tr
atin
g
h
o
w
th
e
r
aw
u
ltra
s
o
u
n
d
im
ag
es
ar
e
p
r
o
g
r
ess
iv
ely
r
ef
in
ed
an
d
a
n
aly
ze
d
to
is
o
late
s
ig
n
if
ican
t
r
eg
i
o
n
s
f
o
r
f
u
r
t
h
e
r
d
iag
n
o
s
tic
ev
alu
ati
o
n
.
T
h
is
co
m
p
r
eh
en
s
iv
e
v
iew
o
f
th
e
p
r
o
ce
s
s
in
g
s
tag
es
p
r
o
v
id
es in
s
ig
h
t i
n
to
th
e
e
f
f
ec
t
iv
en
ess
o
f
ea
ch
m
eth
o
d
a
p
p
lie
d
to
en
h
an
ce
an
d
s
eg
m
en
t th
e
im
ag
es a
cc
u
r
ately
.
T
ab
le
5
.
E
v
alu
atio
n
a
n
d
r
esu
lt
P
a
t
i
e
n
t
R
e
sN
e
t
-
1
8
a
r
c
h
i
t
e
c
t
u
r
e
B
e
f
o
r
e
d
e
v
e
l
o
p
me
n
t
A
f
t
e
r
d
e
v
e
l
o
p
me
n
t
A
c
c
u
r
a
c
y
S
e
n
s
i
t
i
f
i
t
y
S
p
e
c
i
f
i
c
i
t
y
A
c
c
u
r
a
c
y
S
e
n
s
i
t
i
f
i
t
y
S
p
e
c
i
f
i
c
i
t
y
1
0
.
9
6
5
8
3
0
.
9
9
3
1
6
0
.
9
5
9
6
1
0
.
9
6
6
7
5
0
.
9
9
3
0
3
0
.
9
6
0
7
8
2
0
.
9
8
6
6
2
0
.
9
2
9
2
4
0
.
9
9
5
1
4
0
.
9
7
4
2
8
0
.
8
0
2
3
1
0
.
9
9
9
8
0
3
0
.
9
7
8
4
8
0
.
9
8
3
7
1
0
.
9
7
7
5
7
0
.
9
7
9
9
2
0
.
9
8
0
6
3
0
.
9
7
9
7
9
4
0
.
9
5
8
9
8
0
.
9
9
0
7
2
0
.
9
2
0
8
1
0
.
9
5
8
7
0
0
.
9
8
8
3
2
0
.
9
2
3
0
9
5
0
.
8
6
9
5
6
0
.
5
6
5
9
0
0
.
9
9
7
0
0
0
.
8
6
3
8
9
0
.
5
4
1
0
3
0
.
9
9
9
3
8
6
0
.
9
7
0
2
0
0
.
9
3
9
6
3
0
.
9
9
2
9
8
0
.
9
7
0
0
9
0
.
9
3
8
3
8
0
.
9
9
3
7
2
7
0
.
9
7
6
5
7
0
.
6
9
2
1
8
0
.
9
9
9
8
9
0
.
9
6
6
9
1
0
.
5
6
3
2
9
1
.
0
0
0
0
0
8
0
.
9
6
4
0
9
0
.
9
4
2
0
3
0
.
9
7
2
2
1
0
.
9
3
5
6
6
0
.
7
9
7
5
2
0
.
9
8
6
5
2
A
d
etailed
ex
am
in
atio
n
o
f
T
a
b
le
4
r
ev
ea
ls
th
e
s
tep
-
by
-
s
tep
im
p
r
o
v
em
en
ts
an
d
tr
an
s
f
o
r
m
atio
n
s
th
at
th
e
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
a
g
es
u
n
d
er
g
o
d
u
r
i
n
g
th
e
p
r
o
ce
s
s
in
g
p
h
ase.
T
h
e
in
itial
n
o
is
e
r
ed
u
c
tio
n
co
l
u
m
n
s
h
o
ws
th
e
im
p
ac
t
o
f
n
o
is
e
m
in
im
iz
atio
n
tech
n
iq
u
es,
wh
ic
h
p
lay
a
cr
itical
r
o
le
in
en
h
a
n
cin
g
im
ag
e
clar
ity
b
y
r
em
o
v
in
g
b
ac
k
g
r
o
u
n
d
ar
tifa
ct
s
an
d
ir
r
ele
v
an
t
v
is
u
al
d
is
tu
r
b
an
ce
s
.
T
h
is
s
tep
en
s
u
r
es
th
at
s
u
b
s
eq
u
en
t
an
aly
s
es
f
o
cu
s
o
n
th
e
m
ea
n
i
n
g
f
u
l
s
tr
u
ctu
r
es
with
in
th
e
im
ag
e
with
o
u
t
in
ter
f
er
en
ce
f
r
o
m
n
o
is
e,
lead
in
g
to
m
o
r
e
r
elia
b
le
f
ea
tu
r
e
e
x
tr
ac
tio
n
a
n
d
in
ter
p
r
etatio
n
.
T
h
e
R
OI
e
x
tr
ac
tio
n
co
lu
m
n
f
u
r
t
h
er
d
em
o
n
s
tr
ates
th
e
a
b
ilit
y
o
f
t
h
e
p
r
o
ce
s
s
in
g
tech
n
i
q
u
es
to
id
e
n
tify
an
d
is
o
late
s
p
ec
if
ic
ar
ea
s
with
in
th
e
im
ag
es
th
at
ar
e
r
el
ev
an
t
f
o
r
d
iag
n
o
s
tic
p
u
r
p
o
s
es.
T
h
e
clea
r
d
em
ar
ca
ti
o
n
o
f
th
ese
r
e
g
io
n
s
h
elp
s
in
n
a
r
r
o
win
g
th
e
f
o
cu
s
to
p
o
ten
tially
s
u
s
p
icio
u
s
ar
ea
s
,
wh
ich
ca
n
s
tr
ea
m
lin
e
f
u
r
th
er
an
aly
s
is
an
d
aid
in
tar
g
eted
m
ed
ical
ass
es
s
m
en
ts
.
T
h
e
f
in
al
co
lu
m
n
,
s
eg
m
en
tatio
n
,
s
h
o
ws
th
e
s
eg
m
en
ted
o
u
tp
u
t
wh
er
e
th
e
r
eg
i
o
n
s
o
f
in
ter
est
ar
e
d
elin
ea
ted
with
p
r
ec
is
io
n
.
T
h
is
s
tep
is
ess
en
tia
l
in
s
ep
ar
atin
g
th
e
s
ig
n
if
ican
t
tis
s
u
e
s
tr
u
ctu
r
es
f
r
o
m
th
e
b
ac
k
g
r
o
u
n
d
,
en
a
b
lin
g
a
m
o
r
e
d
etaile
d
s
tu
d
y
o
f
th
e
id
en
tifie
d
r
eg
io
n
s
.
T
h
e
s
eg
m
en
tatio
n
r
esu
lts
d
em
o
n
s
tr
ate
h
o
w
ef
f
ec
tiv
ely
th
e
ap
p
lied
alg
o
r
ith
m
d
is
tin
g
u
is
h
es
b
etwe
en
d
if
f
er
e
n
t
tis
s
u
e
ty
p
es,
h
ig
h
lig
h
tin
g
a
r
ea
s
th
at
m
ay
r
eq
u
ir
e
clo
s
er
m
ed
ical
ev
alu
atio
n
.
T
h
e
u
n
if
o
r
m
ity
a
n
d
clar
ity
i
n
s
eg
m
en
tatio
n
ac
r
o
s
s
all
eig
h
t
p
atien
t
im
ag
es
u
n
d
er
s
co
r
e
th
e
r
eliab
ilit
y
an
d
ad
ap
tab
ilit
y
o
f
t
h
e
s
eg
m
en
ta
tio
n
p
r
o
ce
s
s
to
v
a
r
io
u
s
ca
s
es.
T
h
is
co
m
p
r
eh
e
n
s
iv
e
p
r
o
ce
s
s
in
g
ap
p
r
o
ac
h
is
in
s
tr
u
m
en
tal
in
s
u
p
p
o
r
tin
g
m
o
r
e
ac
cu
r
ate
d
iag
n
o
s
is
an
d
en
h
an
cin
g
th
e
o
v
er
all
ef
f
ec
tiv
en
ess
o
f
im
ag
e
-
b
ased
m
ed
ical
an
aly
s
is
.
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:
2502
-
4
7
5
2
Dev
elo
p
men
t o
f R
esN
et
-
1
8
a
r
ch
itectu
r
e
to
lesi
o
n
id
en
tifi
ca
tio
n
in
b
r
ea
s
t
…
(
S
ilfi
a
A
n
d
in
i
)
1245
3
.
2
.
2
.
E
v
a
lua
t
io
n
re
s
ult
Mo
d
el
ev
alu
atio
n
is
ca
lcu
latin
g
th
e
q
u
ality
o
f
th
e
r
esu
ltin
g
m
o
d
el.
T
h
er
e
ar
e
th
r
ee
p
ar
am
eter
s
u
s
ed
,
n
am
ely
ac
cu
r
ac
y
,
s
en
s
itiv
ity
an
d
s
p
ec
if
icity
o
f
th
e
m
o
d
el
cr
ea
ted
.
T
ab
le
6
p
r
esen
ts
th
e
ev
alu
atio
n
r
esu
lts
o
f
th
e
R
esNet
-
1
8
ar
ch
itectu
r
e
b
ef
o
r
e
a
n
d
a
f
ter
f
u
r
th
er
d
ev
el
o
p
m
en
t,
s
h
o
wca
s
in
g
its
p
er
f
o
r
m
an
ce
o
n
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
es
f
r
o
m
eig
h
t
p
atien
ts
.
T
h
e
tab
le
o
u
tlin
es
k
ey
p
er
f
o
r
m
an
ce
m
etr
ics,
in
clu
d
in
g
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
,
f
o
r
ea
ch
p
atien
t.
T
h
ese
m
etr
ics
ar
e
ess
en
tial
f
o
r
u
n
d
er
s
tan
d
in
g
t
h
e
ef
f
ec
tiv
en
ess
o
f
th
e
m
o
d
el
in
co
r
r
ec
tly
i
d
en
tif
y
in
g
a
n
d
d
is
tin
g
u
is
h
in
g
b
etwe
en
d
i
f
f
e
r
en
t
tis
s
u
e
ty
p
es
wit
h
in
th
e
im
a
g
es.
T
h
e
co
m
p
ar
is
o
n
b
etwe
en
th
e
p
r
e
-
an
d
p
o
s
t
-
d
e
v
elo
p
m
e
n
t
p
er
f
o
r
m
an
ce
p
r
o
v
id
es
a
clea
r
v
iew
o
f
th
e
im
p
r
o
v
em
e
n
ts
ac
h
iev
ed
th
r
o
u
g
h
th
e
m
o
d
el
e
n
h
an
ce
m
e
n
t
p
r
o
ce
s
s
,
o
f
f
er
i
n
g
in
s
ig
h
ts
in
to
th
e
r
o
b
u
s
tn
ess
a
n
d
r
eliab
ilit
y
o
f
th
e
u
p
d
ated
R
esNet
-
1
8
ar
ch
itectu
r
e
in
m
ed
ical
im
a
g
e
an
al
y
s
is
[3
4
]
,
[3
5
]
.
T
o
ca
lcu
late
th
e
ac
cu
r
ac
y
,
s
en
s
it
iv
ity
,
an
d
s
p
ec
if
icity
v
alu
es
o
f
th
e
R
esNet
-
1
8
ar
ch
itectu
r
e
b
ef
o
r
e
a
n
d
af
ter
d
ev
elo
p
m
en
t,
we
u
s
e
th
e
f
o
r
m
u
la
b
elo
w:
=
+
+
+
+
[3
6
]
(
1
)
=
+
[3
7
]
(
2
)
=
+
[
26
]
(
3
)
W
h
er
e:
tr
u
e
p
o
s
itiv
e
(
TP
)
:
n
u
m
b
er
o
f
p
o
s
itiv
e
ca
s
es
th
at
wer
e
co
r
r
ec
tly
d
etec
ted
.
,
tr
u
e
n
eg
ativ
e
(
)
:
n
u
m
b
e
r
o
f
n
eg
ati
v
e
ca
s
es
th
at
wer
e
co
r
r
ec
tly
d
etec
te
d
,
f
alse
p
o
s
i
tiv
e
(
)
:
n
u
m
b
e
r
o
f
n
eg
ati
v
e
ca
s
es
th
at
wer
e
d
etec
ted
as
p
o
s
itiv
e
(
f
alse
p
o
s
itiv
e)
,
f
alse
n
eg
ativ
e
(
)
:
n
umb
er
o
f
p
o
s
itiv
e
ca
s
es
th
a
t
wer
e
d
etec
ted
as
n
eg
ativ
e
(
f
alse n
e
g
ativ
e)
.
I
n
th
is
s
tu
d
y
,
th
e
ev
alu
atio
n
o
f
th
e
m
o
d
el
th
at
h
as
b
ee
n
cr
ea
ted
is
ca
r
r
ied
o
u
t
b
y
m
ea
s
u
r
in
g
3
p
ar
am
eter
s
,
n
am
ely
ac
cu
r
ac
y
v
alu
e,
s
en
s
itiv
ity
v
alu
e
an
d
s
p
ec
if
icity
v
alu
e.
B
ased
o
n
T
a
b
l
e
5
,
it
ca
n
b
e
s
ee
n
th
at
th
e
av
e
r
ag
e
v
alu
e
o
f
th
e
R
es
-
Net
a
r
ch
itectu
r
e
m
o
d
el
e
v
alu
atio
n
b
ef
o
r
e
d
ev
el
o
p
m
en
t
is
an
ac
cu
r
ac
y
o
f
0
.
9
6
6
7
1
,
s
en
s
itiv
ity
is
0
.
9
8
4
6
7
a
n
d
s
p
ec
if
icity
is
0
.
9
6
3
1
4
.
W
h
ile
th
e
a
v
er
ag
e
v
al
u
e
o
f
t
h
e
R
es
-
Net
a
r
ch
itectu
r
e
m
o
d
el
ev
alu
atio
n
af
ter
d
ev
elo
p
m
en
t
is
an
ac
cu
r
ac
y
o
f
0
.
9
7
0
9
3
,
s
en
s
itiv
ity
is
0
.
9
0
0
5
6
an
d
s
p
ec
if
icity
is
0
.
9
7
7
0
5
.
W
ith
th
ese
r
esu
lts
,
it
ca
n
b
e
s
ee
n
th
at
th
er
e
is
an
in
cr
ea
s
e
in
th
e
ac
cu
r
ac
y
v
alu
e
th
at
o
cc
u
r
s
is
0
.
0
0
4
2
2
,
a
d
ec
r
ea
s
e
in
s
en
s
itiv
ity
v
alu
e
is
0
.
0
8
4
1
1
an
d
an
in
cr
ea
s
e
in
s
p
ec
if
icity
v
alu
e
is
0
.
0
1
3
9
1
.
W
ith
th
ese
r
e
s
u
lts
,
it
ca
n
b
e
c
o
n
clu
d
e
d
th
at
th
e
m
o
d
el
cr
ea
ted
u
s
in
g
th
e
R
esNet
-
1
8
ar
ch
itectu
r
e
d
ev
elo
p
m
en
t
is
b
etter
th
an
th
e
m
o
d
el
b
ef
o
r
e
th
e
R
esNet
-
1
8
ar
ch
itectu
r
e
d
ev
elo
p
m
en
t
in
d
etec
tin
g
lesi
o
n
s
in
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
a
g
es.
Dete
ctio
n
is
th
e
f
in
al
r
esu
lt
ac
h
iev
ed
in
th
is
s
tu
d
y
.
T
h
e
d
etec
tio
n
in
q
u
esti
o
n
is
th
e
d
etec
tio
n
o
f
lesi
o
n
s
in
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
es.
T
ab
le
6
illu
s
tr
ates
th
e
lesi
o
n
d
etec
tio
n
r
esu
lts
o
n
b
r
ea
s
t
u
ltra
s
o
u
n
d
im
ag
es
f
o
r
eig
h
t
p
atien
ts
u
s
in
g
th
e
R
esNet
-
1
8
ar
ch
itectu
r
e,
co
m
p
ar
i
n
g
th
e
o
u
tp
u
ts
b
e
f
o
r
e
an
d
af
ter
th
e
m
o
d
el
’
s
d
ev
elo
p
m
en
t.
T
h
e
tab
le
p
r
o
v
id
es
a
v
is
u
al
r
ep
r
esen
ta
tio
n
o
f
h
o
w
th
e
en
h
an
ce
d
m
o
d
el
id
en
tifie
s
lesi
o
n
ar
ea
s
,
d
ep
icted
with
y
ello
w
o
u
tlin
es,
o
n
th
e
b
r
ea
s
t
u
lt
r
aso
u
n
d
s
ca
n
s
.
T
h
is
s
id
e
-
by
-
s
id
e
co
m
p
ar
is
o
n
s
h
o
wca
s
es
th
e
im
p
r
o
v
em
en
ts
i
n
d
etec
tio
n
p
er
f
o
r
m
a
n
ce
a
n
d
o
u
tlin
es
t
h
e
r
ef
i
n
ed
ca
p
ab
ilit
y
o
f
t
h
e
m
o
d
el
in
ac
cu
r
ately
d
elin
ea
tin
g
lesi
o
n
b
o
u
n
d
a
r
ies
p
o
s
t
-
d
ev
elo
p
m
en
t.
T
h
e
o
b
ject
iv
e
is
to
ass
e
s
s
th
e
ad
v
an
ce
m
en
ts
m
ad
e
th
r
o
u
g
h
m
o
d
el
o
p
tim
izatio
n
a
n
d
its
im
p
ac
t
o
n
th
e
q
u
ality
o
f
lesi
o
n
d
etec
tio
n
,
wh
ich
is
cr
u
cial
f
o
r
s
u
p
p
o
r
tin
g
m
ed
ical
d
iag
n
o
s
es.
A
th
o
r
o
u
g
h
an
aly
s
is
o
f
T
a
b
le
6
r
ev
ea
ls
n
o
tab
le
im
p
r
o
v
em
en
ts
in
lesi
o
n
d
etec
tio
n
af
ter
th
e
d
ev
elo
p
m
e
n
t
o
f
th
e
R
esNet
-
1
8
ar
ch
itec
tu
r
e.
Acr
o
s
s
all
p
ati
en
ts
,
th
e
p
o
s
t
-
d
ev
elo
p
m
e
n
t
r
esu
lts
ex
h
ib
it
m
o
r
e
p
r
ec
is
e
an
d
co
n
s
is
ten
t
lesi
o
n
b
o
u
n
d
ar
ies,
em
p
h
asizin
g
th
e
en
h
a
n
ce
m
en
ts
ac
h
iev
e
d
th
r
o
u
g
h
m
o
d
e
l
o
p
tim
izatio
n
.
Fo
r
ex
am
p
le,
t
h
e
d
etec
tio
n
in
p
atien
t
1
s
h
o
ws
clea
r
er
an
d
m
o
r
e
ac
cu
r
ate
d
elin
ea
tio
n
o
f
th
e
lesi
o
n
ar
ea
s
co
m
p
ar
ed
to
th
e
p
r
e
-
d
ev
elo
p
m
en
t
s
tag
e,
s
u
g
g
esti
n
g
a
s
ig
n
if
ican
t
ad
v
an
ce
m
en
t
in
th
e
m
o
d
el’
s
ab
ilit
y
to
ac
cu
r
ately
id
en
tify
an
d
o
u
tlin
e
ab
n
o
r
m
alities
.
Patien
ts
3
an
d
4
d
is
p
lay
th
e
m
o
s
t
p
r
o
n
o
u
n
ce
d
d
if
f
er
en
ce
s
,
with
p
o
s
t
-
d
e
v
elo
p
m
en
t im
ag
es in
d
icatin
g
m
o
r
e
co
h
esiv
e
an
d
co
n
tin
u
o
u
s
lesi
o
n
o
u
tlin
es,
r
ed
u
ci
n
g
th
e
o
cc
u
r
r
en
ce
o
f
f
r
a
g
m
en
te
d
o
r
in
co
m
p
lete
b
o
u
n
d
ar
ies
o
b
s
er
v
ed
in
t
h
e
p
r
e
-
d
e
v
elo
p
m
en
t
r
esu
lts
.
T
h
is
im
p
r
o
v
em
e
n
t
is
cr
u
cial
f
o
r
co
m
p
r
eh
e
n
s
iv
e
lesi
o
n
d
etec
tio
n
,
as
well
-
d
ef
in
ed
b
o
u
n
d
a
r
ies
p
r
o
v
id
e
b
etter
g
u
id
an
ce
f
o
r
f
u
r
th
er
clin
ical
e
v
alu
atio
n
s
an
d
p
o
ten
tial
in
ter
v
en
tio
n
s
.
Fo
r
s
o
m
e
p
atien
ts
,
s
u
ch
as
p
atien
ts
2
an
d
5
,
th
e
d
if
f
er
en
ce
s
b
etwe
en
p
r
e
-
an
d
p
o
s
t
-
d
e
v
elo
p
m
en
t
d
ete
ctio
n
r
esu
lts
ap
p
ea
r
m
o
r
e
s
u
b
tle,
in
d
icatin
g
th
at
wh
ile
th
e
m
o
d
el’
s
en
h
an
ce
m
e
n
ts
g
en
er
ally
im
p
r
o
v
ed
p
er
f
o
r
m
an
ce
,
th
e
b
aselin
e
d
etec
tio
n
was
alr
ea
d
y
r
o
b
u
s
t.
No
n
eth
eless
,
th
e
p
o
s
t
-
d
ev
el
o
p
m
en
t
r
esu
lts
d
em
o
n
s
tr
ate
g
r
ea
ter
co
n
s
is
ten
cy
in
lesi
o
n
d
etec
tio
n
,
ev
en
f
o
r
th
ese
ca
s
es.
Patien
t
8
’
s
r
esu
lts
s
h
o
w
a
m
o
r
e
s
u
b
s
tan
tial
en
h
a
n
ce
m
e
n
t,
wh
er
e
th
e
p
r
e
-
d
e
v
elo
p
m
en
t
d
etec
tio
n
p
r
esen
ted
p
a
r
tially
d
is
jo
in
ted
o
u
tlin
es,
wh
ile
th
e
p
o
s
t
-
d
e
v
elo
p
m
en
t
m
o
d
el
p
r
o
d
u
ce
d
s
m
o
o
th
er
an
d
m
o
r
e
d
ef
in
ed
b
o
u
n
d
ar
ies.
T
h
is
s
u
g
g
ests
an
in
cr
ea
s
ed
r
eliab
ilit
y
o
f
th
e
im
p
r
o
v
ed
R
esNet
-
1
8
m
o
d
el
in
h
an
d
lin
g
co
m
p
lex
lesi
o
n
s
tr
u
ctu
r
es
an
d
d
if
f
er
en
tiatin
g
b
etwe
en
lesi
o
n
an
d
n
o
n
-
lesi
o
n
ar
ea
s
m
o
r
e
ef
f
ec
tiv
ely
.
Ov
er
all,
Evaluation Warning : The document was created with Spire.PDF for Python.