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.
42
,
No
.
3
,
J
u
n
e
2
0
2
6
,
p
p
.
798
~
80
8
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
42
.i
3
.
pp
798
-
80
8
798
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
O
ptimi
za
tion o
f
g
lio
ma
seg
menta
ti
o
n
u
sing
3D U
-
Ne
t++ in MRI
surg
ica
l plan
ning
a
nd pa
tient
sa
fet
y
outco
mes
Ahm
ed
B
o
un
eg
t
a
,
M
us
t
a
ph
a
K
helifi,
M
o
ha
m
m
ed
B
ela
dg
ha
m
La
b
o
r
a
t
o
r
y
o
f
I
n
f
o
r
ma
t
i
o
n
P
r
o
c
e
ss
i
n
g
a
n
d
Te
l
e
c
o
mm
u
n
i
c
a
t
i
o
n
(
LTI
T)
,
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
Te
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
si
t
y
o
f
TA
H
R
I
M
o
h
a
mm
e
d
,
B
e
c
h
a
r
,
A
l
g
e
r
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
y
1
9
,
2
0
2
5
R
ev
is
ed
Ma
r
3
1
,
2
0
2
6
Acc
ep
ted
Ma
y
2
6
,
2
0
2
6
Th
e
m
a
in
g
o
a
l
o
f
th
is
st
u
d
y
is
t
o
d
e
v
e
lo
p
a
n
d
e
v
a
lu
a
te
a
n
o
v
e
l
3
D
U
-
Ne
t++
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
two
r
k
f
o
r
a
c
c
u
ra
te
se
g
m
e
n
tatio
n
o
f
g
li
o
m
a
su
b
-
re
g
io
n
s
i
n
M
RI
sc
a
n
s,
a
imin
g
to
e
n
h
a
n
c
e
su
r
g
ica
l
p
lan
n
in
g
,
targ
e
ted
ra
d
io
t
h
e
ra
p
y
,
a
n
d
p
a
ti
e
n
t
sa
fe
ty
.
P
re
c
ise
se
g
m
e
n
tatio
n
o
f
g
l
io
m
a
s
u
b
-
re
g
i
o
n
s
is
a
p
e
rsiste
n
t
c
h
a
ll
e
n
g
e
in
n
e
u
r
o
-
o
n
c
o
l
o
g
y
d
u
e
to
su
b
sta
n
ti
a
l
m
o
rp
h
o
l
o
g
ica
l
v
a
riab
il
it
y
a
c
ro
ss
p
a
ti
e
n
ts.
T
o
a
d
d
re
ss
th
is,
we
in
tr
o
d
u
c
e
a
n
a
u
to
m
a
ti
c
se
g
m
e
n
tatio
n
m
o
d
e
l
b
a
se
d
o
n
a
3
D
U
-
Ne
t++
a
rc
h
it
e
c
tu
re
wi
th
d
e
n
se
sk
ip
c
o
n
n
e
c
ti
o
n
s,
w
h
ich
im
p
ro
v
e
s
sp
a
t
ial
fe
a
tu
re
e
x
trac
ti
o
n
a
n
d
th
e
d
e
li
n
e
a
ti
o
n
o
f
tu
m
o
r
b
o
u
n
d
a
ries
.
Util
izin
g
v
o
lu
m
e
tri
c
d
a
ta
fro
m
th
e
Bra
TS
2
0
2
0
b
e
n
c
h
m
a
rk
,
t
h
e
m
o
d
e
l
a
u
to
m
a
t
ica
ll
y
se
g
m
e
n
ts
th
re
e
c
li
n
ica
ll
y
re
le
v
a
n
t
su
b
stru
c
t
u
re
s:
tu
m
o
r
c
o
re
,
th
e
e
n
h
a
n
c
in
g
tu
m
o
r,
a
n
d
wh
o
le
tu
m
o
r.
Th
e
in
teg
ra
ti
o
n
o
f
d
e
n
se
c
o
n
n
e
c
ti
o
n
s
with
3
D
c
o
n
v
o
l
u
ti
o
n
a
l
lay
e
rs
fa
c
il
it
a
tes
th
e
d
e
tec
ti
o
n
o
f
su
b
tl
e
ti
ss
u
e
v
a
r
iatio
n
s,
in
c
lu
d
i
n
g
n
e
c
ro
sis
a
n
d
e
d
e
m
a
.
Qu
a
n
ti
tati
v
e
e
v
a
lu
a
ti
o
n
d
e
m
o
n
stra
te
s
th
a
t
t
h
e
p
ro
p
o
se
d
3
D
U
-
Ne
t++
su
rp
a
ss
e
s
c
o
n
v
e
n
ti
o
n
a
l
a
rc
h
it
e
c
tu
re
s
su
c
h
a
s
sta
n
d
a
rd
U
-
Ne
t
a
n
d
De
e
p
M
e
d
ic
in
Dic
e
c
o
e
fficie
n
t,
se
n
siti
v
it
y
,
a
n
d
sp
e
c
ifi
c
it
y
,
y
ield
in
g
m
o
re
h
o
m
o
g
e
n
e
o
u
s
a
n
d
c
o
n
ti
n
u
o
u
s
se
g
m
e
n
tati
o
n
s
w
h
il
e
re
d
u
c
i
n
g
m
a
n
u
a
l
a
n
d
se
m
i
-
a
u
to
m
a
ti
c
a
n
n
o
tati
o
n
e
ffo
rts.
Th
is
a
p
p
ro
a
c
h
su
p
p
o
rts
a
d
v
a
n
c
e
d
c
li
n
ica
l
d
e
c
isio
n
-
m
a
k
in
g
a
n
d
wo
r
k
flo
w
a
u
to
m
a
ti
o
n
,
a
n
d
o
ffe
rs
p
o
ten
ti
a
l
fo
r
a
p
p
li
c
a
ti
o
n
t
o
o
th
e
r
t
u
m
o
r
t
y
p
e
s
o
r
i
n
teg
ra
ti
o
n
i
n
to
re
a
l
-
ti
m
e
c
li
n
ica
l
p
ra
c
ti
c
e
.
K
ey
w
o
r
d
s
:
3
D
co
n
v
o
lu
tio
n
E
n
h
an
cin
g
tu
m
o
r
Glio
m
a
s
e
g
m
en
tatio
n
MRI
T
u
m
o
r
c
o
r
e
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
:
Ah
m
ed
B
o
u
n
e
g
ta
L
ab
o
r
ato
r
y
o
f
I
n
f
o
r
m
atio
n
Pro
ce
s
s
in
g
an
d
T
elec
o
m
m
u
n
icati
o
n
(
L
T
I
T
)
Dep
ar
tm
en
t o
f
E
lectr
ical
E
n
g
i
n
ee
r
in
g
,
Facu
lty
o
f
T
ec
h
n
o
l
o
g
y
Un
iv
er
s
ity
o
f
T
AHRI
Mo
h
am
m
ed
B
ec
h
ar
,
Alg
er
ia
E
m
ail:
b
o
u
n
e
g
ta.
ah
m
e
d
@
u
n
iv
-
b
ec
h
ar
.
d
z
1.
I
NT
RO
D
UCT
I
O
N
A
lar
g
e
p
r
o
p
o
r
tio
n
o
f
ce
n
tr
a
l
n
er
v
o
u
s
s
y
s
tem
ca
n
ce
r
s
ar
e
g
lio
m
as,
th
e
m
o
s
t
p
r
e
v
ale
n
t
ty
p
e
o
f
p
r
im
ar
y
b
r
ai
n
tu
m
o
r
[
1
]
,
[
2
]
.
T
h
e
p
r
ec
is
e
s
eg
m
en
tatio
n
o
f
th
ese
tu
m
o
r
s
o
n
MRI
is
ess
en
tial
f
o
r
s
u
r
g
ical
p
lan
n
in
g
,
tr
ea
tm
en
t,
an
d
p
atie
n
t
f
o
llo
w
-
u
p
[
1
]
,
[
3
]
.
Do
cto
r
s
m
u
s
t
ca
r
ef
u
lly
d
is
tin
g
u
is
h
b
e
twee
n
th
e
n
ec
r
o
tic
co
r
e,
t
h
e
en
h
an
ce
d
tu
m
o
r
m
a
s
s
,
an
d
th
e
ed
em
a,
as
t
h
is
d
i
r
ec
tly
in
f
lu
en
ce
s
th
e
p
r
o
g
n
o
s
is
an
d
th
e
r
ap
eu
tic
o
p
tio
n
s
[
2
]
-
[
6
]
.
C
u
r
r
en
tly
,
ad
v
an
ce
d
3
D
tech
n
iq
u
es
ar
e
ess
en
tial
f
o
r
ef
f
ec
tiv
e
p
lan
n
in
g
[
6
]
-
[
8
]
,
b
u
t
m
a
n
u
al
an
d
s
em
i
-
au
to
m
atic
m
eth
o
d
s
r
em
ain
co
m
p
le
x
,
tim
e
-
co
n
s
u
m
in
g
,
a
n
d
h
ig
h
ly
o
p
er
ato
r
-
d
ep
en
d
e
n
t
[
8
]
-
[1
0
]
.
Au
to
m
ated
3
D
tu
m
o
r
s
eg
m
e
n
tatio
n
f
r
o
m
MRI
h
as
m
ad
e
s
ig
n
if
ican
t
p
r
o
g
r
ess
th
an
k
s
to
th
e
em
er
g
e
n
ce
o
f
ad
v
an
ce
d
d
ee
p
lear
n
in
g
ar
c
h
itectu
r
es,
im
p
r
o
v
in
g
th
e
ac
cu
r
ac
y
o
f
g
lio
m
a
tr
ea
tm
e
n
t.
Me
n
ze
et
a
l.
[
1
]
estab
lis
h
ed
th
e
B
r
aT
S
ch
all
en
g
e
as
a
b
e
n
ch
m
ar
k
b
y
p
r
o
v
id
in
g
m
u
ltimo
d
al
d
atab
as
es
th
at
s
tim
u
lated
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
Op
timiz
a
tio
n
o
f g
lio
ma
s
eg
me
n
ta
tio
n
u
s
in
g
3
D
U
-
N
et++
in
MRI
s
u
r
g
ica
l p
la
n
n
in
g
…
(
A
h
med
B
o
u
n
e
g
ta
)
799
alg
o
r
ith
m
ic
in
n
o
v
atio
n
.
B
u
ild
in
g
o
n
th
ese
r
eso
u
r
ce
s
,
R
ese
ar
ch
er
[
2
]
,
[
8
]
,
th
en
en
r
ich
ed
th
em
with
ex
p
er
t
r
ad
io
m
ic
an
n
o
tatio
n
s
,
o
f
f
er
in
g
m
o
r
e
co
m
p
r
e
h
en
s
iv
e
d
atasets
f
o
r
m
o
d
el
tr
ain
in
g
an
d
v
alid
ati
o
n
.
I
n
itial
m
o
d
els
o
f
3
D
U
-
Net+
+
in
ar
ticles
b
y
o
th
er
au
th
o
r
s
[
1
1
]
,
[
1
2
]
h
av
e
d
em
o
n
s
tr
ated
e
x
ce
p
tio
n
al
p
er
f
o
r
m
an
ce
.
T
h
e
y
lev
er
ag
e
im
p
r
o
v
e
d
s
k
ip
co
n
n
ec
tio
n
s
an
d
d
en
s
e
f
ea
tu
r
e
p
r
o
p
a
g
atio
n
to
ac
h
iev
e
Dice
co
ef
f
icien
ts
th
at
e
n
h
an
ce
wh
o
le
tu
m
o
r
s
eg
m
e
n
tatio
n
,
with
p
r
ec
is
io
n
v
alu
es
s
lig
h
tly
ex
c
ee
d
in
g
0
.
9
9
o
n
th
e
B
r
aT
S2
0
2
0
d
ataset.
T
h
ese
ar
ch
itectu
r
es
ad
d
r
ess
th
e
f
u
n
d
a
m
en
tal
ch
allen
g
e
o
f
v
o
l
u
m
etr
ic
f
ea
tu
r
e
ex
tr
ac
tio
n
wh
ile
m
ain
tain
in
g
c
o
m
p
u
tatio
n
al
ef
f
icien
c
y
th
r
o
u
g
h
o
p
tim
iz
ed
co
n
n
ec
tio
n
s
ch
em
es
.
T
h
e
3
D
U
-
Net,
d
ev
elo
p
e
d
by
th
e
a
u
th
o
r
s
o
f
[
1
3
]
,
was
s
p
ec
if
ically
d
esig
n
ed
to
ad
d
r
ess
d
en
s
e
v
o
lu
m
etr
ic
s
eg
m
en
tatio
n
f
r
o
m
s
p
ar
s
e
an
n
o
tatio
n
s
.
T
h
is
laid
th
e
g
r
o
u
n
d
wo
r
k
f
o
r
s
u
b
s
eq
u
en
t
ar
ch
itectu
r
al
im
p
r
o
v
em
en
ts
.
Atten
tio
n
-
b
ased
ar
ch
itectu
r
es
also
a
d
d
r
ess
tis
s
u
e
h
eter
o
g
en
eity
.
Gu
a
n
et
a
l.
[
1
4
]
ac
h
iev
e
d
Dice
s
co
r
es
o
f
0
.
6
8
(
wh
o
le
tu
m
o
r
)
,
0
.
8
5
(
tu
m
o
r
co
r
e)
,
an
d
0
.
7
0
(
en
h
an
cin
g
tu
m
o
r
)
o
n
B
r
aT
S2
0
2
0
b
y
in
co
r
p
o
r
atin
g
Sq
u
ee
z
e
-
an
d
-
E
x
cite
f
ilter
s
an
d
atten
tio
n
g
u
id
an
ce
m
ec
h
an
is
m
s
.
B
SAU
-
Net’
s
[
1
5
]
s
p
atia
l
an
d
c
h
an
n
el
atten
tio
n
m
o
d
u
les
r
e
f
in
ed
th
e
b
o
u
n
d
ar
ies,
ac
h
iev
in
g
a
Dice
s
co
r
e
o
f
>0
.
7
5
with
a
s
en
s
itiv
ity
o
f
0
.
9
0
an
d
a
p
r
ec
is
io
n
o
f
0
.
7
8
.
T
h
e
co
n
clu
s
io
n
s
o
f
a
r
ticles
[
1
6
]
-
[
1
8
]
co
n
s
is
ten
tly
r
an
k
th
e
n
n
U
-
Net
v
ar
ian
ts
am
o
n
g
th
e
b
e
s
t
.
I
m
p
lem
en
tatio
n
s
in
co
r
p
o
r
atin
g
r
esid
u
al
b
lo
ck
s
,
atten
tio
n
g
ates,
an
d
a
Ha
u
s
d
o
r
f
f
d
is
tan
ce
lo
s
s
y
ield
ed
an
av
er
ag
e
Dice
o
f
0
.
8
3
an
d
an
HD9
5
o
f
3
.
8
m
m
f
o
r
a
d
u
lt
g
lio
m
as
(
Dice
=
0
.
7
1
,
HD
9
5
=
8
.
7
m
m
in
p
ed
iatr
ics)
.
R
esear
ch
er
[
1
9
]
,
[
2
0
]
f
o
r
k
n
o
wled
g
e
d
is
till
atio
n
in
p
o
s
to
p
er
ativ
e
s
eg
m
en
tatio
n
ac
h
iev
ed
a
Dice
o
f
0
.
8
3
6
,
h
ig
h
lig
h
tin
g
its
u
tili
ty
f
o
r
s
u
r
g
ical
p
lan
n
in
g
.
E
x
ter
n
al
v
a
lid
atio
n
s
r
ep
o
r
t
a
Dice
o
f
0
.
8
1
to
0
.
8
6
ac
r
o
s
s
in
s
titu
tio
n
s
.
E
n
s
em
b
le
an
d
h
y
b
r
id
s
tr
ateg
ies
en
h
an
ce
r
o
b
u
s
tn
ess
.
E
n
s
em
b
leUNe
ts
,
co
m
b
i
n
in
g
s
ev
er
al
v
ar
ian
ts
o
f
U
-
Ne
t,
ac
h
iev
ed
a
Dice
s
co
r
e
o
f
0
.
9
3
an
d
HD9
5
o
f
1
8
m
m
o
n
B
r
aT
S2
0
2
1
,
s
u
r
p
ass
in
g
C
aPT
k
,
2
DVNe
t,
an
d
R
esNet5
0
[
2
1
]
,
[
2
2
]
.
A
3
D
U
-
Net
-
L
STM
f
u
s
io
n
in
teg
r
atin
g
tem
p
o
r
al
MRI
p
atter
n
s
ac
h
iev
ed
an
ac
cu
r
ac
y
o
f
9
8
.
9
%.
T
h
e
d
u
al
-
m
o
d
ality
s
ets ac
h
iev
ed
D
ice
=
0
.
9
7
7
3
an
d
a
v
e
r
ag
e
I
o
U
=
0
.
6
0
0
8
.
Desp
ite
s
u
b
s
tan
tial
p
r
o
g
r
ess
i
n
au
to
m
ate
d
b
r
ain
tu
m
o
r
s
eg
m
en
tatio
n
,
s
ig
n
if
ica
n
t
ch
allen
g
es
r
em
ain
in
tr
an
s
latin
g
an
d
g
e
n
er
alizin
g
th
ese
m
eth
o
d
s
to
clin
ical
p
r
ac
tice.
Alth
o
u
g
h
s
tate
-
of
-
th
e
-
ar
t
m
o
d
els
ac
h
iev
e
h
ig
h
Dice
s
co
r
es
o
n
cu
r
ated
d
atasets
lik
e
B
r
aT
S,
th
eir
p
er
f
o
r
m
an
ce
o
f
ten
d
r
o
p
s
to
0
.
5
9
–
0
.
6
7
in
r
ea
l
clin
ical
s
ettin
g
s
.
T
h
is
d
ec
lin
e
in
p
er
f
o
r
m
an
ce
in
d
icate
s
a
s
ig
n
if
ic
an
t
s
h
if
t
b
etwe
en
th
e
d
o
m
ai
n
s
o
f
tr
ai
n
in
g
an
d
d
ep
lo
y
m
e
n
t
en
v
ir
o
n
m
en
ts
.
Var
iatio
n
s
in
ac
q
u
is
itio
n
p
ar
am
eter
s
,
s
ca
n
n
e
r
h
ar
d
war
e,
an
d
im
ag
in
g
ar
tifa
cts
co
m
p
r
o
m
is
e
th
e
r
o
b
u
s
tn
ess
o
f
th
ese
s
y
s
tem
s
.
Del
in
ea
tin
g
en
h
an
cin
g
tu
m
o
r
s
is
esp
ec
ially
p
r
o
b
lem
atic:
m
o
s
t
m
eth
o
d
s
ac
h
ie
v
e
Dice
s
co
r
es o
f
o
n
ly
0
.
4
8
to
0
.
7
8
an
d
s
h
o
w
h
ig
h
Ha
u
s
d
o
r
f
f
d
is
tan
ce
s
with
o
u
t
s
p
ec
ialized
l
o
s
s
f
u
n
ctio
n
s
,
r
ev
ea
lin
g
p
e
r
s
is
t
en
t
lim
itatio
n
s
in
co
n
to
u
r
r
ef
in
em
e
n
t.
Po
s
to
p
e
r
ativ
e
an
d
in
tr
ao
p
er
ativ
e
s
eg
m
en
tatio
n
ar
e
ev
en
m
o
r
e
ch
allen
g
in
g
d
u
e
to
b
r
ain
m
o
v
em
en
t,
s
u
r
g
ical
ar
tifa
cts,
a
n
d
s
tr
ict
r
ea
l
-
tim
e
r
eq
u
ir
em
e
n
ts
.
Fin
ally
,
th
e
lac
k
o
f
s
tan
d
ar
d
ized
ev
alu
atio
n
m
et
r
ics
alig
n
ed
with
clin
ical
n
ee
d
s
p
er
p
etu
ates
th
e
g
ap
b
etwe
en
r
esear
ch
b
en
ch
m
ar
k
s
an
d
p
r
ac
tical
u
tili
ty
,
u
n
d
er
s
co
r
in
g
th
e
n
ee
d
f
o
r
m
o
d
els
an
d
p
r
o
to
co
ls
t
h
at
ca
n
b
e
r
eliab
ly
in
teg
r
ated
in
to
d
iv
er
s
e
clin
ical
wo
r
k
f
l
o
ws
.
T
h
is
s
tu
d
y
ex
am
in
ed
a
m
u
lti
-
o
b
jectiv
e
lo
s
s
f
u
n
ctio
n
th
at
co
m
b
in
es
weig
h
ted
Dice
,
co
n
to
u
r
-
s
en
s
itiv
e
Hau
s
d
o
r
f
f
,
a
n
d
f
o
ca
l
m
o
d
u
l
atio
n
to
b
alan
ce
o
v
e
r
lap
ac
cu
r
ac
y
,
c
o
n
to
u
r
f
id
elity
,
an
d
class
im
b
alan
ce
m
an
ag
em
en
t.
T
h
e
f
u
n
ctio
n
i
n
co
r
p
o
r
ates
s
k
ip
co
n
n
ec
tio
n
s
an
d
i
n
teg
r
ates
co
m
p
lex
s
p
atial
d
ata.
On
th
e
B
r
aT
S2
0
2
0
d
ataset,
o
u
r
o
p
tim
ized
3
D
U
-
Net+
+
m
o
d
el
ac
h
i
ev
ed
a
lo
s
s
o
f
0
.
0
1
4
6
,
an
ac
c
u
r
ac
y
o
f
9
9
.
5
1
%,
a
p
r
ec
is
io
n
o
f
9
9
.
5
3
%,
an
d
an
av
er
ag
e
in
ter
s
ec
tio
n
o
v
er
u
n
i
o
n
(
I
o
U)
o
f
8
2
.
8
6
%.
T
h
is
p
er
f
o
r
m
an
ce
s
u
r
p
ass
es
b
aselin
e
U
-
Net
m
o
d
els
[
1
3
]
,
[
2
3
]
an
d
Dee
p
Me
d
ic
m
o
d
els
[
3
]
,
[
2
4
]
.
Ad
d
itio
n
ally
,
o
u
r
m
o
d
el
d
em
o
n
s
tr
ates
s
u
p
er
io
r
b
o
u
n
d
ar
y
ac
cu
r
ac
y
with
Hau
s
d
o
r
f
f
d
is
tan
ce
s
lo
w
er
th
an
th
o
s
e
o
f
n
n
U
-
Net
v
ar
i
an
ts
(
ap
p
r
o
x
im
ately
3
.
8
m
m
)
.
Mo
d
el
p
er
f
o
r
m
an
ce
d
ep
en
d
s
o
n
ar
c
h
itectu
r
al
s
o
p
h
is
ticatio
n
,
lo
s
s
f
u
n
ctio
n
u
n
d
er
s
tan
d
in
g
,
tr
ain
in
g
d
ata
v
ar
iety
,
a
n
d
clin
ical
e
v
a
lu
atio
n
m
etr
ics.
T
h
e
p
r
o
p
o
s
e
d
lo
s
s
f
u
n
ctio
n
en
s
u
r
es
h
ig
h
p
r
ec
is
io
n
an
d
lo
w
b
o
u
n
d
ar
y
e
r
r
o
r
,
b
o
t
h
cr
u
cial
f
o
r
s
u
r
g
ical
p
la
n
n
in
g
an
d
r
ad
io
t
h
er
ap
y
g
u
id
a
n
ce
.
2.
WO
RK
UNDER
T
AK
E
N
a
n
d
E
NV
I
RO
NM
E
N
T
2
.
1
.
Su
mm
a
ry
o
f
t
ec
hn
ic
a
l
a
s
pect
s
I
n
teg
r
atin
g
3
D
U
-
Net+
+
ar
c
h
itectu
r
e
with
ad
v
an
ce
d
atten
tio
n
m
ec
h
a
n
is
m
s
an
d
o
p
ti
m
ized
s
k
ip
co
n
n
ec
tio
n
s
s
ig
n
if
ican
tly
im
p
r
o
v
es
g
lio
m
a
s
eg
m
en
tatio
n
ac
cu
r
ac
y
an
d
p
r
e
o
p
er
ativ
e
p
la
n
n
in
g
in
MRI
-
g
u
id
e
d
n
eu
r
o
s
u
r
g
er
y
[
2
]
,
[
1
3
]
,
[
1
6
]
,
[
2
5
]
.
T
h
is
f
r
am
ewo
r
k
en
h
a
n
ce
s
tu
m
o
r
m
a
r
g
in
d
elin
ea
tio
n
a
n
d
en
ab
les
co
n
s
is
ten
t
d
etec
tio
n
o
f
in
tr
a
-
tu
m
o
r
al
s
u
b
r
eg
io
n
s
,
in
clu
d
i
n
g
n
ec
r
o
ti
c
co
r
es
an
d
p
r
o
life
r
ativ
e
z
o
n
es,
ess
en
tial
f
o
r
tar
g
eted
th
er
a
p
eu
tic
s
tr
ateg
i
es
[
2
5
]
-
[
2
7
]
.
T
ask
-
s
p
ec
if
ic
U
-
Net+
+
ad
ap
tatio
n
s
[
1
1
]
,
[2
8]
d
em
o
n
s
tr
ate
co
m
p
u
tatio
n
al
e
f
f
icien
cy
an
d
r
o
b
u
s
tn
ess
in
s
eg
m
en
tin
g
h
eter
o
g
en
e
o
u
s
b
r
ain
t
u
m
o
r
s
u
b
co
m
p
o
n
en
ts
.
T
h
is
alig
n
s
with
o
u
r
s
tu
d
y
'
s
o
b
jectiv
e
to
en
h
an
ce
m
e
d
ical
im
ag
in
g
wo
r
k
f
lo
ws
th
r
o
u
g
h
p
r
ec
is
e,
au
to
m
ated
s
eg
m
en
tatio
n
p
r
o
to
co
ls
.
2
.
2
.
Wo
r
k
u
nd
er
t
a
k
en
T
h
is
s
tu
d
y
illu
s
tr
ates
th
at
i
n
te
g
r
atin
g
in
tr
ao
p
er
ativ
e
r
ad
io
l
o
g
ical
m
o
n
ito
r
in
g
(
I
R
M)
in
to
p
r
eo
p
er
ativ
e
g
lio
m
a
s
eg
m
en
tatio
n
p
r
esen
ts
s
ig
n
if
ican
t
ch
allen
g
es
d
u
e
to
t
u
m
o
r
h
eter
o
g
en
eity
a
n
d
ir
r
eg
u
lar
ity
in
MRI
d
ata.
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.
42
,
No
.
3
,
J
u
n
e
20
2
6
:
7
9
8
-
80
8
800
T
o
ad
d
r
ess
th
is
,
we
ap
p
l
y
a
3
D
U
-
Net+
+
m
o
d
e
l
with
d
en
s
e
s
k
ip
co
n
n
e
ctio
n
s
an
d
d
ee
p
s
u
p
er
v
is
io
n
to
im
p
r
o
v
e
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
s
eg
m
e
n
t
th
e
en
h
an
ci
n
g
tu
m
o
r
(
E
T
)
,
tu
m
o
r
co
r
e
(
T
C
)
,
an
d
p
er
itu
m
o
r
al
ed
em
a.
Usi
n
g
th
e
B
r
aT
S2
0
2
0
d
ataset,
we
o
p
tim
ize
p
er
f
o
r
m
an
ce
with
Dice
Similar
ity
C
o
ef
f
icien
t
an
d
9
5
th
-
pe
r
ce
n
tile
Hau
s
d
o
r
f
f
d
is
tan
ce
(
HD9
5
)
m
etr
ics.
T
h
is
au
to
m
ated
ap
p
r
o
ac
h
d
eliv
er
s
ac
cu
r
ate
s
eg
m
e
n
tatio
n
ess
en
tial
f
o
r
n
eu
r
o
s
u
r
g
ical
p
lan
n
i
n
g
,
f
ac
ilit
ates d
ec
is
io
n
-
m
ak
in
g
,
a
n
d
im
p
r
o
v
es su
r
g
ical
o
u
tco
m
es
.
T
h
is
s
tu
d
y
ad
d
r
ess
es
th
e
ch
all
en
g
es
o
f
g
lio
m
a
s
eg
m
en
tatio
n
f
o
r
p
r
eo
p
er
ativ
e
p
lan
n
in
g
,
wi
th
a
f
o
cu
s
o
n
in
tr
a
o
p
er
ativ
e
r
ad
io
l
o
g
ical
m
o
n
ito
r
in
g
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
u
tili
ze
s
a
3
D
U
-
Net+
+
a
r
c
h
itectu
r
e
f
ea
tu
r
in
g
d
en
s
e
s
k
ip
co
n
n
ec
tio
n
s
an
d
d
ee
p
s
u
p
er
v
is
io
n
t
o
im
p
r
o
v
e
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
th
e
s
eg
m
en
tatio
n
o
f
tu
m
o
r
s
u
b
r
eg
io
n
s
(
s
ee
T
ab
le
1
f
o
r
ar
ch
itectu
r
al
d
etails).
A
co
m
p
o
s
ite
lo
s
s
f
u
n
ctio
n
is
em
p
l
o
y
ed
t
o
ca
p
tu
r
e
s
tr
u
ctu
r
al
f
ea
tu
r
es
an
d
r
e
d
u
ce
v
o
x
el
-
lev
el
m
is
class
if
icatio
n
s
.
T
h
e
m
o
d
el’
s
p
er
f
o
r
m
a
n
ce
is
e
v
alu
ate
d
o
n
th
e
B
r
aT
S2
0
2
0
d
ataset,
as
illu
s
tr
ated
in
Fig
u
r
e
1
,
d
em
o
n
s
tr
atin
g
its
p
o
t
en
tial
to
en
h
a
n
ce
clin
ical
d
e
cisi
o
n
-
m
ak
in
g
a
n
d
o
p
tim
ize
s
u
r
g
ical
o
u
tco
m
es
.
2
.
3
.
E
nv
iro
nm
ent
:
da
t
a
s
et
&
t
er
m
s
o
f
us
e
a
nd
des
cr
ipti
o
n
2
.
3
.1
.
Co
m
pu
t
a
t
i
o
na
l
en
v
iro
nm
ent
All
co
m
p
u
tatio
n
al
ex
p
er
im
e
n
ts
wer
e
co
n
d
u
cted
o
n
a
h
ig
h
-
p
er
f
o
r
m
an
ce
PC
eq
u
ip
p
ed
wi
th
an
I
n
tel
C
o
r
e
i7
p
r
o
ce
s
s
o
r
,
1
2
GB
R
AM
,
an
d
I
n
tel
HD
Gr
ap
h
ics
3
0
0
0
GPU,
r
u
n
n
i
n
g
a
6
4
-
b
i
t
o
p
er
atin
g
s
y
s
tem
.
Fo
r
co
m
p
u
tatio
n
ally
in
ten
s
iv
e
task
s
,
an
NVI
DI
A
T
esla
P1
0
0
GPU
av
aila
b
le
o
n
Ka
g
g
le
was
u
tili
ze
d
.
T
h
e
s
o
f
twar
e
e
n
v
ir
o
n
m
en
t
c
o
m
p
r
is
ed
Py
th
o
n
3
.
8
a
n
d
T
e
n
s
o
r
Flo
w
2
.
4
to
en
s
u
r
e
r
o
b
u
s
t
m
ac
h
in
e
lea
r
n
in
g
p
er
f
o
r
m
an
ce
.
2
.
3
.
2
.
Usa
g
e
a
g
re
em
ent
T
h
e
B
r
aT
S
-
2
0
2
0
d
atasets
ar
e
f
r
ee
ly
av
ailab
le
f
o
r
r
esear
ch
p
u
r
p
o
s
es,
p
r
o
v
id
ed
th
at
ap
p
r
o
p
r
ia
te
ac
k
n
o
wled
g
m
en
t
is
g
iv
en
b
y
citin
g
t
h
e
m
ain
r
e
f
er
en
ce
w
o
r
k
s
.
Sp
ec
if
ically
,
u
s
er
s
m
u
s
t
cite
th
e
r
eq
u
ir
e
d
m
an
u
s
cr
ip
ts
to
co
m
p
l
y
with
th
e
d
ataset
u
s
ag
e
p
o
licy
[
1
]
,
[
2
]
,
[
2
9
]
.
T
h
is
en
s
u
r
es
p
r
o
p
e
r
r
ec
o
g
n
itio
n
o
f
th
e
o
r
ig
in
al
d
e
v
elo
p
er
s
an
d
u
p
h
o
l
d
s
th
e
in
teg
r
i
ty
o
f
ac
ad
em
ic
r
e
s
ea
r
ch
u
tili
zin
g
B
r
aT
S d
ata
.
2
.
3
.
3
.
O
v
er
v
iew
o
f
t
he
B
ra
T
S
-
20
20
t
ra
ini
ng
da
t
a
s
et
T
h
e
d
ataset
co
n
tain
s
MRI
s
ca
n
s
o
f
h
ig
h
-
an
d
lo
w
-
g
r
ad
e
g
lio
m
as
th
at
wer
e
n
o
r
m
alize
d
,
c
o
-
r
eg
is
ter
ed
,
an
d
r
esam
p
led
f
o
r
u
n
if
o
r
m
ity
.
B
o
th
h
ig
h
-
g
r
a
d
e
an
d
l
o
w
-
g
r
ad
e
g
lio
m
as
in
clu
d
ed
an
n
o
tated
r
eg
io
n
s
an
d
clin
ical
d
ata.
T
u
m
o
r
s
eg
m
e
n
ta
tio
n
u
tili
ze
d
f
o
u
r
MRI
m
o
d
alities
:
T
1
,
T
1
-
Gd
,
T
2
,
an
d
FLA
I
R
.
Patien
t
ag
e
an
d
s
u
r
v
iv
al
d
u
r
atio
n
d
ata
s
u
p
p
o
r
t
th
e
d
ev
elo
p
m
en
t
o
f
p
r
o
g
n
o
s
tic
m
o
d
els.
T
h
e
B
r
aT
S
d
ataset
co
m
p
r
is
es
m
u
ltimo
d
al
s
c
an
s
f
r
o
m
1
9
in
s
titu
tio
n
s
,
r
ep
r
esen
tin
g
n
ativ
e
T
1
,
p
o
s
t
-
co
n
tr
ast
T
1
-
weig
h
ted
(
T
1
-
Gd
)
,
T2
-
weig
h
ted
,
an
d
T
2
-
FLAI
R
v
o
lu
m
es.
B
r
aT
S2
0
2
0
is
p
a
r
ticu
lar
ly
ef
f
ec
tiv
e
f
o
r
b
r
ain
t
u
m
o
r
s
eg
m
en
tatio
n
an
d
s
u
r
v
iv
al
p
r
e
d
ictio
n
[
2
9
]
.
2
.
4
.
M
et
ho
do
lo
g
ic
a
l
o
v
er
v
iew
T
h
e
s
tu
d
y
em
p
lo
y
s
a
3
D
U
-
Net+
+
m
o
d
el
to
s
eg
m
en
t
g
lio
m
as
in
MRI
v
o
lu
m
es
u
s
in
g
3
D
co
n
v
o
l
u
tio
n
s
an
d
d
en
s
e
s
k
ip
c
o
n
n
ec
tio
n
s
f
o
r
ef
f
icien
t
m
u
lti
-
s
ca
le
f
ea
tu
r
e
ex
tr
ac
tio
n
[
1
6
]
,
[
3
0
]
.
A
p
r
o
g
r
ess
iv
e
u
p
s
am
p
lin
g
p
ath
r
ef
in
es
s
p
at
ial
r
eso
lu
tio
n
to
d
elin
ea
te
n
ec
r
o
tic/
co
r
e,
e
d
em
a,
a
n
d
en
h
an
cin
g
s
u
b
r
eg
io
n
s
[
8
]
,
[
3
1
]
.
Usi
n
g
th
e
B
r
aT
S2
0
2
0
d
ataset,
1
0
0
s
lices
p
er
v
o
l
u
m
e
ar
e
s
elec
ted
,
a
n
d
p
er
f
o
r
m
a
n
ce
is
m
ea
s
u
r
e
d
b
y
Dice
co
ef
f
icien
t,
ac
cu
r
ac
y
,
s
e
n
s
itiv
ity
,
an
d
s
p
ec
if
icity
[
1
0
]
,
[
3
1
]
.
T
h
e
ar
ch
itectu
r
e
’
s
ex
te
n
s
iv
e
s
k
ip
p
ath
way
s
o
p
tim
ize
f
ea
tu
r
e
r
e
u
s
e,
lo
wer
m
em
o
r
y
r
e
q
u
ir
em
e
n
ts
,
an
d
d
e
liv
er
a
co
s
t
-
ef
f
ec
tiv
e,
h
ig
h
-
ac
cu
r
ac
y
m
eth
o
d
f
o
r
g
lio
m
a
s
eg
m
en
tatio
n
[
2
3
]
,
[
3
1
]
,
s
u
p
p
o
r
tin
g
p
r
eo
p
e
r
ativ
e
n
eu
r
o
s
u
r
g
ical
p
lan
n
i
n
g
[
2
]
,
[
2
9
]
.
W
o
r
k
f
l
o
w
s
tag
es
in
clu
d
e
d
ata
p
r
e
p
r
o
ce
s
s
in
g
,
m
o
d
el
d
esig
n
[
2
]
,
[
1
3
]
,
lo
s
s
f
u
n
ctio
n
an
d
m
etr
ic
s
elec
tio
n
,
a
n
d
u
ltima
tely
,
tr
ain
in
g
with
ev
alu
ati
o
n
[
1
8
]
,
[
2
6
]
.
2
.
5
.
T
im
eline
o
f
m
od
el
i
m
pl
em
ent
a
t
io
ns
T
h
e
3
D
U
-
Net+
+
m
o
d
el
is
ca
t
eg
o
r
ized
u
n
d
er
th
e
y
ea
r
2
0
1
8
i
n
th
e
im
ag
e
tim
elin
e
s
h
o
wn
in
Fig
u
r
e
1
.
T
h
is
y
ea
r
is
ch
ar
a
cter
ize
d
b
y
im
p
r
o
v
em
e
n
ts
in
s
o
p
h
is
t
icate
d
n
eu
r
al
n
etwo
r
k
s
,
p
ar
t
icu
lar
ly
U
-
Net+
+,
g
en
er
ativ
e
n
eu
r
al
n
etwo
r
k
s
(
GNNs)
,
an
d
Dee
p
L
ab
,
wh
i
ch
s
ig
n
if
y
p
r
o
g
r
ess
in
in
tr
i
ca
te
s
eg
m
en
tatio
n
m
o
d
els
f
o
r
m
ed
ical
an
d
s
cien
tific
p
u
r
p
o
s
es.
T
h
e
3
D
U
-
Net+
+
m
o
d
el,
a
v
ar
iatio
n
o
f
U
-
Net+
+,
is
an
o
p
tim
ized
ar
c
h
itectu
r
e
d
esig
n
ed
to
im
p
r
o
v
e
f
ea
tu
r
e
ex
t
r
ac
tio
n
an
d
s
eg
m
en
tatio
n
p
r
ec
is
io
n
in
th
r
ee
-
d
im
en
s
io
n
al
m
ed
ical
im
a
g
in
g
.
T
h
e
o
r
g
an
izatio
n
o
f
th
is
c
h
ar
t
in
Fig
u
r
e
1
c
o
m
es
af
ter
th
e
f
i
r
s
t
attem
p
t
at
m
ath
em
atica
l
m
o
d
elin
g
o
f
n
eu
r
o
n
s
in
1
9
4
3
b
r
o
u
g
h
t
b
y
Mc
C
u
llo
ch
an
d
Pit
ts
,
an
d
th
e
Per
ce
p
tr
o
n
in
1
9
5
8
,
c
r
ea
ted
b
y
Fra
n
k
R
o
s
en
b
latt.
T
h
is
was o
n
e
o
f
th
e
f
ir
s
t b
in
ar
y
class
if
icatio
n
alg
o
r
ith
m
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:
2502
-
4
7
5
2
Op
timiz
a
tio
n
o
f g
lio
ma
s
eg
me
n
ta
tio
n
u
s
in
g
3
D
U
-
N
et++
in
MRI
s
u
r
g
ica
l p
la
n
n
in
g
…
(
A
h
med
B
o
u
n
e
g
ta
)
801
Fig
u
r
e
1
.
NE
T
e
v
o
lu
tio
n
ch
ar
t
(
d
ee
p
lear
n
in
g
ca
s
e)
3.
NE
T
WO
RK
ARCH
I
T
E
CT
URE
T
h
e
co
m
p
o
s
itio
n
an
d
ar
ch
itect
u
r
e
d
i
m
en
s
io
n
s
o
f
th
e
im
p
lem
en
ted
U
-
Net+
+
m
o
d
el
u
s
ed
i
n
th
is
s
tu
d
y
ar
e
p
r
esen
ted
with
d
etail
in
T
a
b
le
1
.
T
h
is
ex
p
lan
atio
n
h
ig
h
lig
h
ts
th
e
n
etwo
r
k
’
s
s
tr
u
ctu
r
al
d
e
s
ig
n
,
in
clu
d
in
g
t
h
e
r
ep
r
esen
tatio
n
o
f
f
ea
t
u
r
e
m
ap
s
an
d
th
e
n
u
m
b
e
r
o
f
c
h
an
n
els ass
ig
n
ed
to
ea
ch
lay
er
,
p
r
o
v
id
in
g
a
co
m
p
r
eh
e
n
s
iv
e
u
n
d
er
s
tan
d
i
n
g
o
f
m
o
d
el’
s
c
o
m
p
lex
ity
.
Sp
ec
ial
em
p
h
asis
is
p
lace
d
o
n
th
e
3
D
U
-
N
et+
+
ar
ch
itectu
r
e,
wh
er
e
ea
ch
lay
e
r
'
s
r
o
le
an
d
co
n
n
ec
tio
n
ar
e
illu
s
tr
ated
to
r
ef
le
ct
th
e
ad
v
an
ce
d
d
esig
n
o
f
th
e
ex
ec
u
ted
m
o
d
el.
T
ab
le
1
.
T
h
e
3
D
U
-
Net
++
ar
c
h
itectu
r
e,
with
r
ep
r
esen
tatio
n
f
ea
tu
r
e
m
ap
s
,
an
d
th
e
n
u
m
b
e
r
o
f
ch
an
n
els is
d
en
o
ted
f
o
r
ea
c
h
lay
er
(
T
h
e
e
x
ec
u
ted
m
o
d
el
)
La
y
e
r
(
T
y
p
e
)
O
u
t
p
u
t
c
o
n
f
i
g
u
r
a
t
i
o
n
P
a
r
a
m N
u
mb
e
r
Li
n
k
e
d
t
o
I
n
p
u
t
_
l
a
y
e
r
_
1
(
I
n
p
u
t
l
a
y
e
r
)
(
N
o
n
e
,
1
2
8
,
1
2
8
,
1
2
8
,
4
)
0
-
c
o
n
v
3
d
(
C
o
n
v
3
D
)
(
N
o
n
e
,
1
2
8
,
1
2
8
,
1
2
8
,
6
4
)
6
,
9
7
6
i
n
p
u
t
_
l
a
y
e
r
_
1
[
0
]
[
0
]
c
o
n
v
3
d
_
1
(
C
o
n
v
3
D
)
(
N
o
n
e
,
1
2
8
,
1
2
8
,
1
2
8
,
6
4
)
1
1
0
,
6
5
6
c
o
n
v
3
d
[
0
]
[
0
]
d
r
o
p
o
u
t
(
D
r
o
p
o
u
t
)
(
N
o
n
e
,
1
2
8
,
1
2
8
,
1
2
8
,
6
4
)
0
c
o
n
v
3
d
_
1
[
0
]
[
0
]
max
_
p
o
o
l
i
n
g
3
d
(
M
a
x
_
P
o
o
l
i
n
g
3
D
)
(
N
o
n
e
,
6
4
,
6
4
,
6
4
,
1
2
8
)
0
d
r
o
p
o
u
t
[
0
]
[
0
]
c
o
n
v
3
d
_
2
(
C
o
n
v
3
D
)
(
N
o
n
e
,
6
4
,
6
4
,
6
4
,
1
2
8
)
2
2
1
,
3
1
2
max
_
p
o
o
l
i
n
g
3
d
[
0
]
[
0
]
c
o
n
v
3
d
_
3
(
C
o
n
v
3
D
)
(
N
o
n
e
,
6
4
,
6
4
,
6
4
,
1
2
8
)
4
4
2
,
4
9
6
c
o
n
v
3
d
_
2
[
0
]
[
0
]
d
r
o
p
o
u
t
_
1
(
D
r
o
p
o
u
t
)
(
N
o
n
e
,
6
4
,
6
4
,
6
4
,
1
2
8
)
0
c
o
n
v
3
d
_
3
[
0
]
[
0
]
M
a
x
_
p
o
o
l
i
n
g
3
d
_
1
(
M
a
x
_
P
o
o
l
i
n
g
3
D
)
(
N
o
n
e
,
3
2
,
3
2
,
3
2
,
1
2
8
)
0
d
r
o
p
o
u
t
_
1
[
0
]
[
0
]
c
o
n
v
3
d
_
4
(
C
o
n
v
3
D
)
(
N
o
n
e
,
3
2
,
3
2
,
3
2
,
2
)
8
8
4
,
9
9
2
max
_
p
o
o
l
i
n
g
3
d
_
1
[
0
]
[
0
]
c
o
n
v
3
d
_
5
(
C
o
n
v
3
D
)
(
N
o
n
e
,
3
2
,
3
2
,
3
2
,
2
)
1
,
7
6
9
,
7
2
8
c
o
n
v
3
d
_
4
[
0
]
[
0
]
d
r
o
p
o
u
t
_
2
(
D
r
o
p
o
u
t
)
(
N
o
n
e
,
3
2
,
3
2
,
3
2
,
2
)
0
c
o
n
v
3
d
_
5
[
0
]
[
0
]
max
_
p
o
o
l
i
n
g
3
d
_
2
(
M
a
x
P
o
o
l
i
n
g
3
D
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
2
)
0
d
r
o
p
o
u
t
_
2
[
0
]
[
0
]
c
o
n
v
3
d
_
6
(
C
o
n
v
3
D
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
5
)
3
,
5
3
9
,
4
5
6
max
_
p
o
o
l
i
n
g
3
d
_
2
[
0
]
[
0
]
c
o
n
v
3
d
_
7
(
C
o
n
v
3
D
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
5
)
7
,
0
7
8
,
4
0
0
c
o
n
v
3
d
_
6
[
0
]
[
0
]
d
r
o
p
o
u
t
_
3
(
D
r
o
p
o
u
t
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
5
)
0
c
o
n
v
3
d
_
7
[
0
]
[
0
]
max
_
p
o
o
l
i
n
g
3
d
_
3
(
M
a
x
P
o
o
l
i
n
g
3
D
)
(
N
o
n
e
,
8
,
8
,
8
,
5
1
2
)
0
d
r
o
p
o
u
t
_
3
[
0
]
[
0
]
c
o
n
v
3
d
_
8
(
C
o
n
v
3
D
)
(
N
o
n
e
,
8
,
8
,
8
,
1
0
2
4
)
1
4
,
1
5
6
,
8
0
0
max
_
p
o
o
l
i
n
g
3
d
_
3
[
0
]
[
0
]
c
o
n
v
3
d
_
9
(
C
o
n
v
3
D
)
(
N
o
n
e
,
8
,
8
,
8
,
1
0
2
4
)
2
8
,
3
1
2
,
5
7
6
c
o
n
v
3
d
_
8
[
0
]
[
0
]
d
r
o
p
o
u
t
_
4
(
D
r
o
p
o
u
t
)
(
N
o
n
e
,
8
,
8
,
8
,
1
0
2
4
)
0
c
o
n
v
3
d
_
9
[
0
]
[
0
]
u
p
_
sa
mp
l
i
n
g
3
d
(
U
p
S
a
m
p
l
i
n
g
3
D
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
1
)
0
d
r
o
p
o
u
t
_
4
[
0
]
[
0
]
c
o
n
v
3
d
_
1
0
(
C
o
n
v
3
D
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
5
)
1
4
,
1
5
6
,
2
8
8
u
p
_
sa
mp
l
i
n
g
3
d
[
0
]
[
0
]
C
o
n
c
a
t
e
n
a
t
e
(
C
o
n
c
a
t
e
n
a
t
e
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
1
)
0
d
r
o
p
o
u
t
_
3
[
0
]
[
0
]
/
c
o
n
v
3
d
_
1
0
[
0
]
[
0
]
c
o
n
v
3
d
_
1
1
(
C
o
n
v
3
D
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
5
)
1
4
,
1
5
6
,
2
8
8
c
o
n
c
a
t
e
n
a
t
e
[
0
]
[
0
]
c
o
n
c
a
t
e
n
a
t
e
(
C
o
n
c
a
t
e
n
a
t
e
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
1
0
2
4
)
0
d
r
o
p
o
u
t
_
3
[
0
]
[
0
]
/
c
o
n
v
3
d
_
1
0
[
0
]
[
0
]
C
o
n
v
3
d
_
1
1
(
C
o
n
v
3
D
)
(
N
o
n
e
,
1
6
,
1
6
,
1
6
,
2
5
6
)
1
4
,
1
5
6
,
2
8
8
c
o
n
c
a
t
e
n
a
t
e
[
0
]
[
0
]
u
p
_
sa
mp
l
i
n
g
3
d
_
3
(
U
p
S
a
m
p
l
i
n
g
3
D
)
(
N
o
n
e
,
1
2
8
,
1
2
8
,
1
2
8
,
1
2
8
)
0
d
r
o
p
o
u
t
_
7
[
0
]
[
0
]
C
o
n
v
3
d
_
1
6
(
C
o
n
v
3
D
)
(
N
o
n
e
,
1
2
8
,
1
2
8
,
1
2
8
,
6
4
)
2
2
1
,
2
4
8
u
p
_
sa
mp
l
i
n
g
3
d
_
3
[
0
]
[
0
]
d
r
o
p
o
u
t
_
8
(
D
r
o
p
o
u
t
)
(
N
o
n
e
,
1
2
8
,
1
2
8
,
1
2
8
,
6
4
)
0
c
o
n
v
3
d
_
1
7
[
0
]
[
0
]
c
o
n
v
3
d
_
1
6
(
C
o
n
v
3
D
)
(
N
o
n
e
,
1
2
8
,
1
2
8
,
1
2
8
,
4
)
2
6
0
d
r
o
p
o
u
t
_
[
0
]
[
0
]
To
t
a
l
p
a
r
a
me
t
e
r
s
:
9
4
,
1
2
6
,
8
5
2
(
3
5
9
.
0
7
M
B
)
,
Tr
a
i
n
a
b
l
e
p
a
r
a
ms
:
9
4
,
1
2
6
,
8
5
2
(
3
5
9
.
0
7
M
B
)
4.
E
XE
CUT
I
O
N
AND
SP
E
CIF
I
CAT
I
O
NS
4
.
1
.
I
nfo
r
m
a
t
io
n
T
h
e
3
D
U
-
Net+
+
m
o
d
el
f
o
r
b
r
ain
tu
m
o
r
s
eg
m
e
n
tatio
n
h
as
two
p
r
im
ar
y
co
m
p
o
n
en
ts
:
d
ata
p
r
ep
ar
atio
n
an
d
tr
ai
n
in
g
co
n
f
ig
u
r
atio
n
.
Data
p
r
ep
ar
atio
n
e
n
tails
im
p
o
r
tin
g
3
D
MRI
v
o
l
u
m
es,
n
o
r
m
alizin
g
v
o
x
el
in
ten
s
ities
,
em
p
lo
y
in
g
au
g
m
en
t
atio
n
tech
n
iq
u
es,
a
n
d
d
elin
ea
tin
g
s
eg
m
e
n
tatio
n
m
ask
s
f
o
r
tu
m
o
r
s
u
b
r
eg
io
n
s
.
B
atch
p
r
o
d
u
ctio
n
in
v
o
lv
es
p
ar
titi
o
n
in
g
th
e
d
atas
et
in
to
tr
ain
in
g
,
v
alid
ati
o
n
,
an
d
test
s
u
b
s
ets:
6
8
%
f
o
r
tr
ain
in
g
,
2
0
%
f
o
r
v
alid
atio
n
,
an
d
1
2
%
f
o
r
test
in
g
.
T
h
e
tr
ain
in
g
d
ataset
was
i
n
ter
n
ally
s
p
lit
f
o
r
v
alid
atio
n
,
u
s
in
g
8
5
% f
o
r
tr
ain
in
g
an
d
1
5
% f
o
r
test
in
g
d
u
r
in
g
th
e
tr
ain
i
n
g
p
h
ase
.
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.
42
,
No
.
3
,
J
u
n
e
20
2
6
:
7
9
8
-
80
8
802
4
.
2
.
E
qu
a
t
io
ns
a
nd
m
a
t
hema
t
ica
l f
o
r
m
ula
t
io
ns
4
.
2
.
1
.
I
m
po
rt
a
nce
in
ima
g
e
s
eg
m
ent
a
t
i
o
n
I
n
s
eg
m
e
n
tatio
n
,
th
e
D
ice
co
ef
f
icien
t
in
d
icate
s
h
o
w
m
u
ch
th
e
s
eg
m
en
t
ed
ar
ea
in
t
h
e
p
r
ed
ictio
n
o
v
er
lap
s
with
th
e
g
r
o
u
n
d
tr
u
th
.
A
h
ig
h
er
D
ice
co
ef
f
icien
t in
d
icate
s
a
b
etter
p
er
f
o
r
m
an
ce
.
4
.
2
.
2
.
M
a
t
hem
a
t
ic
a
l
f
o
r
m
ula
t
io
n
-
T
h
e
D
ice
co
ef
f
icien
t
(
F1
-
s
co
r
e)
D
b
etwe
en
two
s
et
s
A
an
d
B
(
r
ep
r
esen
tin
g
th
e
p
r
ed
ict
ed
s
eg
m
en
tatio
n
an
d
th
e
g
r
o
u
n
d
tr
u
t
h
,
re
-
s
p
ec
tiv
ely
)
is
f
o
r
m
u
lated
as
:
=
2
|
∩
|
|
|
+
|
|
(
1
)
W
h
er
e:
-
∣
A∩B
∣
is
th
e
in
ter
s
ec
tio
n
(
o
v
e
r
lap
)
b
etwe
en
t
h
e
p
r
e
d
icted
an
d
tr
u
e
s
ets,
-
∣
A
∣
an
d
∣
B
∣
ar
e
th
e
s
izes (
p
ix
el
co
u
n
ts
)
o
f
th
e
r
esp
ec
tiv
e
s
ets.
-
I
n
s
eg
m
en
tatio
n
(
class
if
icatio
n
)
task
s
,
s
u
ch
as
tu
m
o
r
b
o
r
d
e
r
id
en
tific
atio
n
,
HD5
is
e
s
s
en
ti
al
as
i
t
p
r
ev
en
ts
s
lig
h
t
m
is
tak
e
s
at
d
i
s
tan
t
p
lace
s
f
r
o
m
s
ig
n
if
ican
tly
d
is
to
r
ti
n
g
th
e
d
is
tan
ce
m
etr
ic,
r
esu
lt
-
in
g
in
a
m
o
r
e
d
ep
en
d
a
b
le
ass
ess
m
en
t o
f
s
p
atial
clo
s
en
ess
.
Ma
t
h
e
m
at
ic
all
y
,
HD
9
5
ca
lc
u
l
a
tes
th
e
Ha
u
s
d
o
r
f
f
d
is
ta
n
ce
at
t
h
e
9
5
t
h
p
e
r
c
e
n
til
e,
f
o
r
m
u
lat
ed
a
s
f
o
l
lo
ws
:
95
(
,
)
=
(
95
(
∈
(
,
)
)
,
95
(
∈
(
,
)
)
(
2
)
w
h
e
r
e
:
d
(
a,
b
)
r
e
p
r
esen
ts
th
e
E
u
clid
ea
n
d
is
tan
ce
b
e
-
twee
n
p
o
in
ts
a
∈
A
an
d
b
∈
B.
-
Sen
s
itiv
ity
,
o
r
tr
u
e
p
o
s
itiv
e
r
ate
,
is
v
ital
in
h
ea
lth
ca
r
e
f
o
r
th
e
p
r
ec
is
e
id
en
tific
atio
n
o
f
p
o
s
itiv
e
ca
s
es,
r
ed
u
cin
g
f
alse n
eg
ativ
es,
an
d
i
s
f
u
n
d
am
en
tal
f
o
r
class
if
icatio
n
jo
b
s
.
-
Sen
s
itiv
ity
is
d
ef
in
ed
as:
=
(
)
(
)
+
(
)
(
3
)
w
h
e
r
e
:
-
T
r
u
e
p
o
s
itiv
es (
T
P)
ar
e
co
r
r
ec
t
ly
id
en
tifie
d
p
o
s
itiv
e
in
s
tan
ce
s
,
-
Fals
e
n
eg
ativ
es (
FN)
ar
e
p
o
s
itiv
e
in
s
tan
ce
s
in
co
r
r
ec
tly
i
d
en
ti
f
ied
as n
eg
ativ
e.
-
Sp
ec
if
icity
(
T
)
,
o
r
tr
u
e
n
e
g
ativ
e
r
ate
,
is
an
ess
en
tial
s
tatis
t
ic
in
f
r
au
d
d
etec
tio
n
,
f
ac
ilit
at
in
g
th
e
p
r
ec
is
e
id
en
tific
atio
n
o
f
n
eg
ativ
e
ca
s
e
s
an
d
m
in
im
izin
g
f
alse p
o
s
itiv
es,
esp
ec
ially
in
s
cr
ee
n
in
g
test
s
.
′
=
(
)
(
)
+
(
)
(
4
)
W
h
e
r
e
:
-
T
r
u
e
n
e
g
ativ
es (
T
N)
ar
e
c
o
r
r
e
ctly
id
en
tifie
d
n
e
g
ativ
e
in
s
tan
ce
s
,
-
Fals
e
p
o
s
itiv
es (
FP
)
ar
e
n
eg
ativ
e
in
s
tan
ce
s
in
co
r
r
ec
tly
i
d
en
ti
f
ied
as p
o
s
itiv
e
.
-
Dice
l
o
s
s
is
a
m
ath
em
atica
l
f
o
r
m
u
la
th
at
o
p
tim
izes
th
e
D
ice
co
ef
f
icien
t
in
s
eg
m
e
n
tatio
n
task
s
b
y
p
en
alizin
g
m
o
d
els f
o
r
l
o
w
o
v
e
r
lap
b
etwe
en
p
r
ed
icted
a
n
d
g
r
o
u
n
d
t
r
u
th
s
eg
m
e
n
ts
.
Dice
lo
s
s
:
=
1
−
2
|
∩
|
|
|
+
|
|
(
5
)
w
h
er
e:
T
h
e
n
o
t
ati
o
n
is
s
im
ilar
to
th
e
Dice
co
ef
f
icien
t b
u
t o
p
ti
m
ized
f
o
r
g
r
ad
ie
n
t
-
b
ased
lear
n
in
g
.
4
.
3
.
I
ns
t
ruct
io
n
T
h
e
3
D
U
-
Net+
+
m
o
d
el
em
p
l
o
y
s
d
en
s
e
s
k
ip
co
n
n
ec
tio
n
s
,
c
o
m
p
o
s
ite
lo
s
s
f
u
n
ctio
n
s
,
ca
llb
ac
k
s
,
an
d
r
eg
u
lar
izatio
n
to
en
h
a
n
ce
s
eg
m
en
tatio
n
ac
cu
r
ac
y
an
d
g
en
e
r
aliza
tio
n
.
E
v
alu
ated
o
n
a
test
s
et,
it e
n
s
u
r
es p
r
ec
is
e
d
elin
ea
tio
n
o
f
tu
m
o
r
s
u
b
r
e
g
io
n
s
wh
ile
m
ain
tain
in
g
s
p
atial
a
n
d
co
n
tex
tu
al
in
teg
r
ity
.
Ver
tic
al
s
k
ip
co
n
n
ec
tio
n
s
an
d
3
D
co
n
v
o
lu
tio
n
s
f
ac
ilit
ate
ef
f
ec
tiv
e
in
f
o
r
m
atio
n
f
lo
w
an
d
v
o
lu
m
etr
ic
f
ea
tu
r
e
r
ete
n
tio
n
,
m
a
k
in
g
th
e
ar
ch
itectu
r
e
well
-
s
u
ited
f
o
r
ac
cu
r
ate
b
r
ain
tu
m
o
r
s
eg
m
en
tati
o
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
Op
timiz
a
tio
n
o
f g
lio
ma
s
eg
me
n
ta
tio
n
u
s
in
g
3
D
U
-
N
et++
in
MRI
s
u
r
g
ica
l p
la
n
n
in
g
…
(
A
h
med
B
o
u
n
e
g
ta
)
803
5.
CO
M
P
UT
I
NG
E
XP
E
R
I
M
E
NT
S
T
h
is
r
esear
ch
f
o
c
u
s
es
o
n
th
e
s
p
ec
if
ic
m
eth
o
d
o
lo
g
y
f
o
r
co
m
p
letely
au
t
o
m
ated
s
eg
m
en
ta
tio
n
s
tu
d
ies
o
f
b
r
ai
n
tu
m
o
r
s
u
s
in
g
th
e
U
-
N
et+
+
3
D
m
o
d
el
.
T
h
e
p
r
o
ce
s
s
f
o
llo
ws
:
Data
Pre
p
ar
atio
n
an
d
T
r
ain
in
g
C
o
n
f
ig
u
r
atio
n
f
o
r
MRI
Seg
m
en
tatio
n
Utilizin
g
3
D
U
-
Ne
t++
:
MRI
v
o
lu
m
es
ar
e
p
r
ep
r
o
ce
s
s
ed
,
an
d
u
s
er
-
p
r
o
v
id
ed
an
n
o
tatio
n
s
m
ar
k
in
itial
t
u
m
o
r
r
e
g
io
n
s
to
g
u
id
e
s
eg
m
e
n
tatio
n
.
Data
au
g
m
en
tatio
n
tech
n
iq
u
es
lik
e
r
o
tatio
n
an
d
s
ca
lin
g
ex
p
an
d
d
ataset
d
iv
er
s
ity
wh
ile
p
r
eser
v
in
g
an
n
o
tatio
n
ac
cu
r
ac
y
.
A
3
D
U
-
Net+
+
m
o
d
el
with
s
k
ip
c
o
n
n
ec
tio
n
s
is
tr
ai
n
ed
u
s
in
g
Dice
a
n
d
c
r
o
s
s
-
en
tr
o
p
y
l
o
s
s
es
to
b
o
o
s
t
s
eg
m
en
tatio
n
p
er
f
o
r
m
a
n
ce
a
n
d
m
in
im
ize
o
v
er
f
itti
n
g
.
User
f
ee
d
b
ac
k
v
ia
ca
llb
ac
k
s
allo
ws
r
ef
in
em
en
t
o
f
co
m
p
lex
s
lices,
f
u
r
th
er
im
p
r
o
v
in
g
m
o
d
el
ac
cu
r
ac
y
,
wh
ich
i
s
as
s
ess
ed
u
s
in
g
Dice
,
s
en
s
it
i
v
ity
,
an
d
p
r
ec
is
io
n
m
etr
ics to
h
ig
h
lig
h
t t
h
e
b
en
e
f
i
ts
o
f
s
em
i
-
au
to
m
atio
n
i
n
m
ar
k
ed
ar
ea
s
.
Vis
u
aliza
tio
n
to
o
ls
d
is
p
lay
M
R
I
d
ata
u
s
ed
f
o
r
d
ee
p
lear
n
in
g
-
b
ased
b
r
ain
tu
m
o
r
s
eg
m
en
tati
o
n
,
aid
in
g
p
r
ec
is
e
tu
m
o
r
b
o
u
n
d
ar
y
id
en
t
if
icatio
n
an
d
p
r
eo
p
er
ativ
e
p
la
n
n
in
g
f
o
r
in
d
iv
id
u
al
p
atien
ts
.
T
h
e
co
d
e
r
etr
iev
es
m
u
ltip
le
MRI
m
o
d
alities
(
FL
AI
R
,
T
1
,
T
1
ce
,
T
2
)
s
ca
n
s
,
al
o
n
g
with
a
g
r
o
u
n
d
tr
u
th
s
eg
m
en
tatio
n
m
ask
f
o
r
a
s
in
g
le
p
atien
t
in
Fig
u
r
e
2
,
ca
s
e
(
B
r
aT
S2
0
_
T
r
ain
in
g
_
0
0
1
)
.
Usi
n
g
th
e
NiB
ab
el
lib
r
ar
y
,
th
e
co
d
e
r
ea
d
s
ea
ch
MRI
s
ca
n
in
.
n
ii f
o
r
m
at
an
d
e
x
tr
ac
ts
its
d
ata
as a
3
D
n
u
m
p
y
ar
r
ay
.
Fig
u
r
e
2
.
I
ll
u
s
tr
atio
n
o
f
d
if
f
er
e
n
t M
R
I
m
o
d
alities
f
o
r
a
s
in
g
le
p
atien
t c
ase
(
B
r
aT
S2
0
_
T
r
ain
i
n
g
_
0
0
1
)
5
.1
.
F
o
r
f
urt
her
a
na
ly
s
is
o
f
t
he
r
esu
lt
ing
i
m
a
g
es,
o
f
t
he
s
im
ula
t
ed
c
o
de
s
equence
T
h
e
An
im
ated
Vis
u
aliza
tio
n
p
r
esen
ted
in
Fig
u
r
es
3
an
d
4
ill
u
s
tr
ates
MRI
d
ata
em
p
lo
y
e
d
f
o
r
tr
ain
in
g
d
ee
p
lear
n
in
g
m
o
d
els
in
b
r
ai
n
tu
m
o
r
s
eg
m
en
tatio
n
.
T
h
is
d
y
n
am
ic,
lay
er
-
by
-
lay
e
r
GI
F
v
iv
id
ly
s
h
o
wca
s
es
th
e
th
r
ee
-
d
im
en
s
io
n
al
s
p
atial
r
elatio
n
s
h
ip
s
am
o
n
g
an
ato
m
ical
s
tr
u
ctu
r
es,
o
f
f
er
i
n
g
an
in
tu
itiv
e
u
n
d
er
s
tan
d
in
g
o
f
co
m
p
lex
b
r
ai
n
m
o
r
p
h
o
lo
g
y
.
B
y
en
h
an
cin
g
th
e
v
is
ib
ilit
y
o
f
tu
m
o
r
b
o
u
n
d
ar
ies
an
d
ad
jace
n
t
tis
s
u
es,
th
e
v
is
u
aliza
tio
n
s
ig
n
if
ican
tly
co
n
tr
ib
u
tes
to
im
p
r
o
v
i
n
g
s
eg
m
en
t
atio
n
ac
cu
r
ac
y
a
n
d
s
u
p
p
o
r
ts
m
o
r
e
ef
f
ec
tiv
e
p
r
e
-
o
p
er
ativ
e
p
la
n
n
in
g
an
d
cli
n
ical
d
ec
is
io
n
-
m
ak
in
g
.
Fig
u
r
e
3
.
I
n
ter
ac
tiv
e
r
e
p
r
esen
t
atio
n
o
f
c
r
o
s
s
-
s
ec
tio
n
s
in
a
3
D
v
o
lu
m
e
(
GI
F)
Fig
u
r
e
4
.
B
r
ain
_
v
o
lu
m
e
.
g
if
'
f
o
r
th
e
f
ir
s
t sli
ce
(
0
0
1
)
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.
42
,
No
.
3
,
J
u
n
e
20
2
6
:
7
9
8
-
80
8
804
T
h
e
f
o
llo
win
g
g
r
a
p
h
ic
s
h
o
ws
Plo
t_
r
o
i
v
is
u
aliza
tio
n
,
wh
ich
h
elp
s
r
esear
ch
er
s
u
n
d
er
s
tan
d
b
r
ain
s
ca
n
r
ep
r
esen
tatio
n
s
th
at
h
ig
h
lig
h
t
th
e
p
r
esen
ce
o
f
tu
m
o
r
s
(
Fig
u
r
e
5
)
,
ch
ec
k
s
eg
m
en
tatio
n
a
cc
u
r
ac
y
,
an
d
r
ef
in
e
m
o
d
els
f
o
r
b
etter
clin
ical
o
u
tc
o
m
es.
T
h
e
m
o
d
el'
s
tr
ain
in
g
h
is
to
r
y
was
u
s
ed
to
e
v
alu
ate
its
s
eg
m
en
tatio
n
p
e
r
f
o
r
m
an
ce
,
id
en
tif
y
tr
en
d
s
,
an
d
ass
es
s
s
tab
ili
ty
.
Key
in
d
icato
r
s
wer
e
ex
am
in
ed
to
u
n
d
er
s
tan
d
th
e
ad
ap
tatio
n
o
f
th
e
m
o
d
el
d
u
r
in
g
t
r
ain
in
g
.
T
h
e
m
o
d
el'
s
b
eh
av
io
r
was
co
m
p
ar
ed
o
n
tr
ain
in
g
an
d
v
alid
atio
n
d
atasets
to
en
s
u
r
e
its
g
en
er
aliza
b
ilit
y
f
o
r
n
ew
an
d
u
n
p
u
b
lis
h
ed
d
ata
,
in
Fig
u
r
e
6
.
Fig
u
r
e
5
.
Var
io
u
s
r
ep
r
esen
tati
o
n
s
o
f
th
e
b
r
ain
s
ca
n
(
s
p
atial
lo
ca
tio
n
an
d
b
o
u
n
d
a
r
ies o
f
th
e
tu
m
o
r
)
Fig
u
r
e
6
.
T
r
ain
in
g
a
n
d
v
alid
atio
n
m
etr
ics f
o
r
tu
m
o
r
s
eg
m
en
ta
tio
n
m
o
d
el
I
n
th
e
ca
s
e
b
elo
w,
t
h
e
co
d
e
im
p
o
r
ts
MRI
s
ca
n
s
,
g
r
o
u
n
d
tr
u
th
s
eg
m
e
n
tatio
n
m
ask
s
,
an
d
m
o
d
el
p
r
ed
ictio
n
s
,
d
is
p
lay
in
g
th
em
in
a
s
ix
im
ag
es
s
u
b
p
lo
t,
a
n
d
it
m
ay
u
tili
ze
s
h
o
wPr
ed
i
ctsB
y
I
d
f
o
r
p
atien
t
co
m
p
ar
is
o
n
s
as
s
h
o
wn
in
Fig
u
r
e
7
.
T
h
is
p
lo
t
c
o
n
tain
s
two
s
u
b
p
lo
ts
:
T
h
e
le
f
t
o
n
e
s
h
o
ws
th
e
r
esized
g
r
o
u
n
d
tr
u
th
f
o
r
t
h
e
s
p
ec
if
ied
class
an
d
s
lice,
t
h
e
r
ig
h
t
s
u
b
p
lo
t
d
is
p
l
ay
s
th
e
p
r
ed
icted
s
eg
m
en
tatio
n
f
o
r
th
e
s
am
e
class
an
d
s
lice
(
in
th
is
ca
s
e
th
e
4
0
th
s
lice
an
d
th
e
ev
alu
atio
n
class
ar
e
s
et
to
2
(
C
OR
E
r
eg
io
n
o
f
th
e
tu
m
o
r
)
,
in
Fig
u
r
e
8
.
Fig
u
r
e
7
.
T
u
m
o
r
s
eg
m
en
tatio
n
in
MRI:
o
r
ig
in
al
s
ca
n
,
g
r
o
u
n
d
tr
u
th
,
an
d
p
r
ed
icted
s
u
b
r
eg
io
n
s
f
o
r
ed
em
a,
co
r
e,
an
d
en
h
a
n
cin
g
t
u
m
o
r
r
eg
io
n
s
(
m
atr
i
x
s
ize
: 1
2
0
×
1
2
0
)
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
Op
timiz
a
tio
n
o
f g
lio
ma
s
eg
me
n
ta
tio
n
u
s
in
g
3
D
U
-
N
et++
in
MRI
s
u
r
g
ica
l p
la
n
n
in
g
…
(
A
h
med
B
o
u
n
e
g
ta
)
805
Fig
u
r
e
8
.
Su
b
p
lo
ts
id
en
tify
th
e
g
r
o
u
n
d
tr
u
th
an
d
th
e
p
r
ed
icte
d
class
(
m
atr
i
x
s
ize:
1
2
0
x
1
2
0
)
6.
SI
M
UL
A
T
I
O
N
R
E
S
UL
T
S
T
h
e
3
D
U
-
N
et+
+
m
o
d
el
a
ch
ie
v
es
s
u
p
er
io
r
s
e
g
m
e
n
ta
ti
o
n
ac
c
u
r
ac
y
,
wit
h
a
p
ar
tic
u
l
a
r
l
y
l
o
w
lo
s
s
v
al
u
e
(
0
.
0
1
4
6
)
c
o
m
p
a
r
e
d
t
o
s
ta
n
d
a
r
d
U
-
N
et
a
n
d
De
ep
-
M
ed
ic
[
3
]
,
[
3
2
]
,
h
i
g
h
p
r
e
cisi
o
n
(
0
.
9
9
5
1
)
a
n
d
Me
an
I
o
U
(
0
.
8
2
8
6
)
,
i
n
d
ic
ati
n
g
r
eli
ab
le
v
o
x
el
c
ate
g
o
r
iz
ati
o
n
a
n
d
r
e
m
a
r
k
a
b
l
e
o
v
e
r
l
ap
w
it
h
b
a
c
k
g
r
o
u
n
d
s
e
g
m
en
tat
io
n
s
.
I
ts
Dice
c
o
e
f
f
ici
e
n
t
o
f
0
.
6
4
5
2
c
o
m
b
in
e
d
wit
h
h
i
g
h
a
cc
u
r
a
cy
(
0
.
9
9
5
1
)
a
n
d
s
p
ec
i
f
i
cit
y
(
0
.
9
9
8
4
)
,
r
ef
l
ec
ts
e
f
f
ici
e
n
t
an
d
cli
n
i
ca
ll
y
r
elia
b
le
s
eg
m
e
n
ta
ti
o
n
,
p
a
r
t
ic
u
la
r
l
y
f
o
r
r
e
g
i
o
n
s
o
f
n
ec
r
o
ti
c
h
ea
r
t
a
n
d
e
d
e
m
a
.
I
n
a
d
d
iti
o
n
,
th
e
m
o
d
el'
s
h
ig
h
s
e
n
s
it
iv
it
y
(
0
.
9
9
3
7
)
u
n
d
e
r
li
n
es
its
e
f
f
ec
ti
v
e
n
ess
i
n
d
e
tec
ti
n
g
t
u
m
o
r
ca
s
es.
Dic
e
s
co
r
es
f
o
r
s
u
b
r
eg
io
n
s
c
o
n
f
i
r
m
t
h
at
3
D
U
-
Net+
+
o
u
t
p
e
r
f
o
r
m
s
o
t
h
e
r
m
o
d
els
i
n
f
ea
t
u
r
e
e
x
t
r
a
cti
o
n
an
d
b
o
u
n
d
a
r
y
d
e
li
n
e
ati
o
n
in
T
ab
le
2
.
T
ab
le
2
.
C
o
m
p
a
r
ativ
e
p
er
f
o
r
m
an
ce
m
etr
ics o
f
th
r
ee
-
d
im
e
n
s
io
n
al
s
eg
m
en
tatio
n
m
o
d
els f
o
r
m
ed
ical
im
ag
in
g
P
r
o
t
o
t
y
p
e
C
o
n
v
e
n
t
i
o
n
a
l
U
-
N
e
t
[
1
8
]
,
[
2
6
]
P
r
o
t
o
t
y
p
e
C
o
n
v
e
n
t
i
o
n
a
l
U
-
N
e
t
[
1
8
]
,
[
2
6
]
P
r
o
t
o
t
y
p
e
C
o
n
v
e
n
t
i
o
n
a
l
U
-
N
e
t
[
1
8
]
,
[
2
6
]
Lo
ss
0
.
0
2
5
4
0
.
0
1
9
6
0
.
0
2
2
3
0
.
0
1
8
1
0
.
0
1
4
6
A
c
c
u
r
a
c
y
0
.
9
8
7
5
0
.
9
9
1
2
0
.
9
8
9
0
0
.
9
9
2
8
0
.
9
9
5
1
M
e
a
n
I
o
U
0
.
7
8
5
2
0
.
8
0
8
3
0
.
7
9
2
0
0
.
8
1
8
9
0
.
8
2
8
6
D
i
c
e
c
o
e
f
f
i
c
i
e
n
t
0
.
6
0
5
2
0
.
6
2
3
8
0
.
6
1
6
5
0
.
6
4
5
7
0
.
6
4
5
2
P
r
e
c
i
s
i
o
n
0
.
9
8
3
1
0
.
9
8
8
5
0
.
9
8
5
9
0
.
9
9
0
7
0
.
9
9
5
3
S
e
n
s
i
t
i
v
i
t
y
0
.
9
7
8
4
0
.
9
8
4
3
0
.
9
8
2
1
0
.
9
8
7
3
0
.
9
9
3
7
S
p
e
c
i
f
i
c
i
t
y
0
.
9
9
5
4
0
.
9
9
6
7
0
.
9
9
5
9
0
.
9
9
7
2
0
.
9
9
8
4
D
i
c
e
(
n
e
c
r
o
t
i
c
c
o
r
e
)
0
.
6
2
0
4
0
.
6
4
5
3
0
.
6
3
0
9
0
.
6
5
8
3
0
.
6
5
7
3
D
i
c
e
(
e
d
e
ma)
0
.
7
0
5
1
0
.
7
3
2
1
0
.
7
2
1
8
0
.
7
4
8
0
0
.
7
5
0
2
D
i
c
e
(
e
n
h
a
n
c
i
n
g
c
o
r
e
)
0
.
6
9
0
2
0
.
7
1
0
5
0
.
7
0
0
4
0
.
7
2
5
6
0
.
7
3
4
1
7.
DIS
CU
SS
I
O
N
T
h
e
3
D
U
-
Net+
+
m
o
d
el
d
em
o
n
s
tr
ates
s
u
p
er
io
r
p
e
r
f
o
r
m
an
ce
in
au
to
m
ated
b
r
ain
tu
m
o
r
s
eg
m
en
tatio
n
,
co
n
s
is
ten
tly
o
u
tp
er
f
o
r
m
in
g
o
t
h
er
m
o
d
els
ac
r
o
s
s
k
ey
m
etr
ic
s
s
u
ch
as
lo
s
s
,
ac
cu
r
ac
y
,
Me
an
I
n
ter
s
ec
tio
n
o
v
er
Un
io
n
(
Me
an
I
o
U)
,
p
r
ec
is
io
n
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
.
I
t
ac
h
iev
ed
th
e
lo
west
lo
s
s
(
0
.
0
1
4
6
)
an
d
th
e
h
ig
h
est
ac
cu
r
ac
y
(
0
.
9
9
5
1
)
,
in
d
icatin
g
r
em
ar
k
ab
le
r
eliab
ilit
y
in
s
eg
m
en
tin
g
b
r
ai
n
tu
m
o
r
s
.
T
h
e
Me
an
I
o
U
s
co
r
e
o
f
0
.
8
2
8
6
h
ig
h
lig
h
ts
its
p
r
o
f
icien
cy
in
ac
cu
r
ately
d
elin
ea
tin
g
t
u
m
o
r
b
o
u
n
d
a
r
ies,
wh
ich
is
ess
en
tial
f
o
r
e
f
f
ec
tiv
e
s
u
r
g
ical
p
lan
n
i
n
g
.
Dice
c
o
ef
f
i
cien
ts
f
o
r
cr
itical
tu
m
o
r
s
u
b
r
eg
io
n
s
,
0
.
6
5
7
3
f
o
r
th
e
n
ec
r
o
t
ic
co
r
e,
0
.
7
5
0
2
f
o
r
ed
em
a,
an
d
0
.
7
3
4
1
f
o
r
th
e
e
n
h
an
cin
g
co
r
e
,
f
u
r
th
er
d
em
o
n
s
tr
ate
th
e
m
o
d
el’
s
ab
ilit
y
t
o
id
e
n
tify
an
d
s
eg
m
en
t
clin
ically
r
elev
an
t
ar
ea
s
.
Hig
h
s
en
s
itiv
ity
(
0
.
9
9
3
7
)
e
n
s
u
r
es
th
at
tu
m
o
r
r
eg
i
o
n
s
ar
e
co
r
r
ec
tly
id
en
tifie
d
,
m
in
im
izin
g
r
esid
u
al
tu
m
o
r
r
is
k
,
wh
ile
h
ig
h
s
p
ec
if
icity
(
0
.
9
9
8
4
)
r
e
d
u
ce
s
f
alse
p
o
s
itiv
es,
p
r
eser
v
in
g
h
ea
lth
y
tis
s
u
e.
T
h
ese
r
esu
lt
s
h
av
e
s
i
g
n
if
ican
t
im
p
licatio
n
s
f
o
r
s
u
r
g
ical
p
lan
n
in
g
an
d
p
atien
t
s
af
ety
.
T
h
e
m
o
d
el’
s
en
h
an
ce
d
ac
c
u
r
ac
y
en
a
b
les
n
eu
r
o
s
u
r
g
e
o
n
s
to
m
ak
e
m
o
r
e
in
f
o
r
m
e
d
d
ec
is
io
n
s
,
m
ax
im
izi
n
g
tu
m
o
r
r
esectio
n
wh
ile
m
in
im
izin
g
d
am
ag
e
to
h
ea
lth
y
b
r
ain
s
tr
u
ct
u
r
es
-
an
es
p
ec
ially
cr
iti
ca
l
f
ac
to
r
in
co
m
p
lex
ca
s
es
n
ea
r
v
ital
b
r
ain
r
eg
io
n
s
.
Au
to
m
atio
n
o
f
th
e
s
eg
m
en
tatio
n
p
r
o
ce
s
s
also
allev
iates
clin
ician
s
’
wo
r
k
lo
a
d
,
allo
win
g
f
aster
an
d
m
o
r
e
ef
f
icien
t
s
u
r
g
ical
p
lan
n
i
n
g
,
wh
ich
ca
n
lea
d
to
im
p
r
o
v
ed
p
atien
t
o
u
tc
o
m
es
th
r
o
u
g
h
tim
ely
in
ter
v
en
tio
n
s
.
Fro
m
a
p
atien
t
s
af
ety
p
er
s
p
ec
tiv
e,
th
e
m
o
d
el’
s
h
ig
h
s
en
s
itiv
ity
an
d
s
p
ec
if
icity
h
elp
r
ed
u
ce
s
u
r
g
ical
co
m
p
licatio
n
s
a
n
d
th
e
r
is
k
o
f
n
eu
r
o
lo
g
ical
d
ef
icit
s
b
y
p
r
ev
en
tin
g
in
a
d
v
er
ten
t
r
em
o
v
al
o
f
h
ea
lth
y
tis
s
u
e.
Ov
er
all,
3
D
U
-
Net+
+
s
u
p
p
o
r
ts
p
er
s
o
n
alize
d
,
h
ig
h
-
q
u
ality
ca
r
e
b
y
en
ab
lin
g
p
r
ec
is
e
an
d
r
eliab
le
tu
m
o
r
d
elin
ea
tio
n
tailo
r
ed
to
ea
ch
p
a
tien
t’
s
u
n
iq
u
e
a
n
ato
m
y
.
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.
42
,
No
.
3
,
J
u
n
e
20
2
6
:
7
9
8
-
80
8
806
8
.
CO
NCLU
SI
O
N
T
h
e
au
to
m
ated
s
eg
m
e
n
tatio
n
o
f
g
lio
m
a
s
u
b
r
e
g
io
n
s
f
r
o
m
m
u
ltimo
d
al
MRI
v
o
lu
m
es
is
p
r
o
p
o
s
ed
i
n
th
is
p
ap
er
u
s
in
g
an
ad
v
an
ce
d
3
D
U
-
Net+
+
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
etwo
r
k
.
Fo
u
r
MRI
s
eq
u
en
ce
s
(
T
1
,
T
1
-
Gd
,
T
2
,
an
d
FLAI
R
)
ar
e
u
s
ed
in
th
e
p
r
o
ce
d
u
r
e;
th
ese
ar
e
n
o
r
m
alize
d
,
co
-
r
eg
is
ter
ed
,
a
n
d
r
e
s
am
p
led
to
im
p
r
o
v
e
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
g
u
a
r
an
t
ee
d
ata
co
n
s
is
ten
cy
.
T
h
e
m
o
d
el
is
tr
ain
ed
u
s
in
g
a
c
o
m
p
o
s
it
e
lo
s
s
f
u
n
ctio
n
th
at
co
m
b
in
es
Dice
lo
s
s
an
d
Fo
c
al
lo
s
s
to
ad
d
r
ess
class
im
b
a
lan
ce
an
d
en
h
a
n
ce
s
eg
m
e
n
tatio
n
ac
cu
r
ac
y
.
T
h
is
allo
ws
f
o
r
r
o
b
u
s
t
lear
n
in
g
f
r
o
m
b
o
t
h
c
h
allen
g
in
g
an
d
well
-
class
if
ied
v
o
x
els.
T
h
e
3
D
U
-
Net+
+
ar
ch
itectu
r
e
in
co
r
p
o
r
ates
d
ee
p
s
u
p
er
v
is
io
n
an
d
d
e
n
s
e
s
k
ip
co
n
n
ec
tio
n
s
,
wh
ich
en
ab
le
p
r
ec
is
e
d
elin
ea
tio
n
o
f
co
m
p
le
x
tu
m
o
r
b
o
u
n
d
a
r
ies,
s
u
ch
as
n
e
cr
o
tic
co
r
e,
e
d
em
a,
an
d
en
h
a
n
cin
g
tu
m
o
r
r
e
g
io
n
s
.
T
h
e
B
r
a
T
S
2
0
2
0
d
ataset
is
u
s
ed
to
tr
ain
an
d
ass
ess
th
e
m
o
d
el.
I
t
o
u
tp
e
r
f
o
r
m
s
tr
ad
itio
n
al
U
-
Net
an
d
Dee
p
Me
d
ic
m
o
d
els
with
a
r
em
ar
k
ab
l
y
lo
w
lo
s
s
v
alu
e
(
0
.
0
1
4
6
)
,
a
Dice
c
o
ef
f
icien
t
o
f
0
.
6
4
5
2
,
an
ac
c
u
r
ac
y
o
f
0
.
9
9
5
1
,
an
d
a
s
p
ec
if
icity
o
f
0
.
9
9
8
4
.
T
h
ese
f
in
d
i
n
g
s
s
h
o
w
th
at,
wh
ile
p
r
eser
v
in
g
co
m
p
u
tatio
n
al
ef
f
icien
c
y
an
d
ad
a
p
tab
ilit
y
to
d
iv
e
r
s
e
tu
m
o
r
p
r
esen
tatio
n
s
,
th
e
in
c
o
r
p
o
r
ati
o
n
o
f
an
o
p
tim
ized
lo
s
s
f
u
n
ctio
n
in
to
th
e
3
D
U
-
Net+
+
f
r
am
ewo
r
k
p
r
o
v
id
es
r
o
b
u
s
t
an
d
d
e
p
en
d
ab
le
s
eg
m
en
tatio
n
o
f
g
lio
m
a
s
u
b
r
eg
io
n
s
,
aid
s
clin
ical
d
ec
is
io
n
-
m
ak
in
g
in
n
eu
r
o
s
u
r
g
ical
p
lan
n
i
n
g
,
a
n
d
le
s
s
en
s
th
e
n
ee
d
f
o
r
m
an
u
al
a
n
n
o
tatio
n
.
9
.
CH
AL
L
E
NG
E
S AN
D
F
U
T
URE WO
RK
F
u
tu
r
e
r
esear
ch
ca
n
b
en
ef
it
f
r
o
m
en
h
a
n
ce
d
tr
ain
i
n
g
s
tr
ateg
i
es
th
at
lev
er
ag
e
co
n
s
is
ten
tly
an
n
o
tated
d
atasets
s
u
ch
as
B
r
aT
S2
0
2
1
f
o
r
b
r
ain
t
u
m
o
r
s
eg
m
en
tatio
n
an
d
AM
OS
f
o
r
ab
d
o
m
in
al
o
r
g
a
n
d
elin
ea
tio
n
,
wh
ich
ar
e
cr
itical
f
o
r
im
p
r
o
v
in
g
s
eg
m
en
tatio
n
ac
cu
r
ac
y
in
c
o
m
p
lex
tr
a
n
s
itio
n
al
r
eg
io
n
s
.
Fu
r
th
e
r
m
o
r
e,
in
teg
r
atin
g
d
o
m
ain
-
ad
a
p
tiv
e
tr
ain
in
g
,
m
u
lti
-
r
eso
lu
tio
n
in
p
u
ts
,
an
d
h
y
b
r
id
a
r
ch
itect
u
r
es
with
atten
tio
n
m
ec
h
an
is
m
s
m
ay
s
ig
n
if
ica
n
tly
r
ef
in
e
b
o
u
n
d
a
r
y
ac
c
u
r
ac
y
.
No
tab
ly
,
lev
er
a
g
in
g
U
-
Net+
+’
s
n
ested
an
d
d
en
s
e
s
k
ip
co
n
n
ec
tio
n
s
en
h
an
ce
s
f
ea
tu
r
e
r
eu
s
ab
ilit
y
an
d
r
ed
u
ce
s
s
em
an
tic
g
ap
s
b
etwe
en
en
co
d
er
an
d
d
ec
o
d
er
,
lead
in
g
to
s
u
p
er
io
r
ac
cu
r
a
cy
an
d
o
v
er
la
p
with
g
r
o
u
n
d
tr
u
th
m
ask
s
.
C
o
m
b
in
i
n
g
U
-
Ne
t++
with
p
o
wer
f
u
l
b
ac
k
b
o
n
e
en
co
d
er
s
lik
e
E
f
f
icien
tNet
ca
n
f
u
r
th
er
b
o
o
s
t
p
er
f
o
r
m
a
n
ce
b
y
e
x
tr
ac
tin
g
r
ich
er
m
u
lti
-
s
ca
le
co
n
tex
tu
al
f
ea
tu
r
es
with
f
ewe
r
p
ar
am
eter
s
,
th
er
e
b
y
im
p
r
o
v
i
n
g
s
eg
m
en
tatio
n
p
r
ec
is
io
n
,
b
o
u
n
d
a
r
y
s
h
ar
p
n
ess
,
an
d
g
e
n
er
aliza
tio
n
ac
r
o
s
s
d
iv
er
s
e
clin
ical
d
atasets
.
ACK
NO
WL
E
DG
M
E
N
T
S
T
h
is
wo
r
k
was
f
in
an
cially
s
u
p
p
o
r
ted
b
y
th
e
Gen
er
al
Dir
ec
to
r
ate
o
f
Scien
tific
R
esear
ch
an
d
T
ec
h
n
o
lo
g
ical
Dev
elo
p
m
e
n
t
(
DGRS
DT
)
u
n
d
er
th
e
au
t
h
o
r
ity
o
f
th
e
Min
is
tr
y
o
f
Hig
h
er
E
d
u
ca
tio
n
an
d
Scien
tific
R
esear
ch
(
ME
S
R
S
)
an
d
s
p
o
n
s
o
r
e
d
b
y
th
e
L
a
b
o
r
ato
r
y
o
f
I
n
f
o
r
m
atio
n
an
d
T
elec
o
m
m
u
n
icatio
n
s
T
ec
h
n
o
lo
g
ies o
f
th
e
Facu
lty
o
f
T
ec
h
n
o
lo
g
y
,
M
o
h
am
m
e
d
T
A
HR
I
Un
iv
er
s
ity
-
B
ec
h
ar
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Ah
m
ed
B
o
u
n
e
g
ta
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Mu
s
tap
h
a
Kh
elif
i
✓
✓
✓
✓
✓
✓
Mo
h
am
m
ed
B
elad
g
h
am
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
T
h
e
au
th
o
r
s
d
ec
lar
e
th
at
th
ey
h
av
e
n
o
k
n
o
wn
f
in
a
n
cial,
p
er
s
o
n
al,
o
r
p
r
o
f
ess
io
n
al
co
n
f
licts
o
f
in
ter
est
th
at
co
u
ld
h
av
e
in
f
lu
en
ce
d
th
e
wo
r
k
p
r
esen
ted
in
th
is
p
a
p
er
.
All
co
n
tr
i
b
u
tio
n
s
wer
e
m
a
d
e
o
b
jectiv
ely
an
d
in
d
e
p
en
d
e
n
tly
.
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
Op
timiz
a
tio
n
o
f g
lio
ma
s
eg
me
n
ta
tio
n
u
s
in
g
3
D
U
-
N
et++
in
MRI
s
u
r
g
ica
l p
la
n
n
in
g
…
(
A
h
med
B
o
u
n
e
g
ta
)
807
RE
F
E
R
E
NC
E
S
[
1
]
B
.
H
.
M
e
n
z
e
e
t
a
l
.
,
“
T
h
e
m
u
l
t
i
m
o
d
a
l
b
r
a
i
n
t
u
mo
r
i
m
a
g
e
s
e
g
m
e
n
t
a
t
i
o
n
b
e
n
c
h
m
a
r
k
(
B
R
A
TS)
,
”
I
EEE
T
ra
n
s
a
c
t
i
o
n
s
o
n
Me
d
i
c
a
l
I
m
a
g
i
n
g
,
v
o
l
.
3
4
,
n
o
.
1
0
,
p
p
.
1
9
9
3
–
2
0
2
4
,
O
c
t
.
2
0
1
5
,
d
o
i
:
1
0
.
1
1
0
9
/
T
M
I
.
2
0
1
4
.
2
3
7
7
6
9
4
.
[
2
]
S
.
B
a
k
a
s
e
t
a
l
.
,
“
A
d
v
a
n
c
i
n
g
t
h
e
c
a
n
c
e
r
g
e
n
o
me
a
t
l
a
s
g
l
i
o
ma
M
R
I
c
o
l
l
e
c
t
i
o
n
s
w
i
t
h
e
x
p
e
r
t
s
e
g
me
n
t
a
t
i
o
n
l
a
b
e
l
s
a
n
d
r
a
d
i
o
mi
c
f
e
a
t
u
r
e
s,
”
S
c
i
e
n
t
i
f
i
c
D
a
t
a
,
v
o
l
.
4
,
n
o
.
1
,
p
.
1
7
0
1
1
7
,
S
e
p
.
2
0
1
7
,
d
o
i
:
1
0
.
1
0
3
8
/
sd
a
t
a
.
2
0
1
7
.
1
1
7
.
[
3
]
R
.
A
.
Ze
i
n
e
l
d
i
n
,
M
.
E.
K
a
r
a
r
,
O
.
B
u
r
g
e
r
t
,
a
n
d
F
.
M
a
t
h
i
s
-
U
l
l
r
i
c
h
,
“
M
u
l
t
i
mo
d
a
l
C
N
N
n
e
t
w
o
r
k
s
f
o
r
b
r
a
i
n
t
u
mo
r
seg
m
e
n
t
a
t
i
o
n
i
n
M
R
I
:
a
B
r
a
TS
2
0
2
2
c
h
a
l
l
e
n
g
e
s
o
l
u
t
i
o
n
,
”
i
n
L
e
c
t
u
re
N
o
t
e
s
i
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
1
3
7
6
9
LN
C
S
,
2
0
2
3
,
p
p
.
1
2
7
–
1
3
7
.
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
031
-
3
3
8
4
2
-
7
_
1
1
.
[
4
]
M
.
A
d
e
w
o
l
e
e
t
a
l
.
,
“
T
h
e
b
r
a
i
n
t
u
mo
r
se
g
me
n
t
a
t
i
o
n
(
B
r
a
TS)
c
h
a
l
l
e
n
g
e
2
0
2
3
:
g
l
i
o
ma
se
g
m
e
n
t
a
t
i
o
n
i
n
s
u
b
-
sa
h
a
r
a
n
A
f
r
i
c
a
p
a
t
i
e
n
t
p
o
p
u
l
a
t
i
o
n
(
B
r
a
TS
-
A
f
r
i
c
a
)
,
”
ArXi
v
.
M
a
y
3
0
,
2
0
2
3
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
w
w
w
.
n
c
b
i
.
n
l
m
.
n
i
h
.
g
o
v
/
p
u
b
m
e
d
/
3
7
3
9
6
6
0
8
%
0
A
h
t
t
p
:
/
/
w
w
w
.
p
u
b
m
e
d
c
e
n
t
r
a
l
.
n
i
h
.
g
o
v
/
a
r
t
i
c
l
e
r
e
n
d
e
r
.
f
c
g
i
?
a
r
t
i
d
=
P
M
C
1
0
3
1
2
8
1
4
[
5
]
A
.
W
.
M
o
a
w
a
d
e
t
a
l
.
,
“
T
h
e
b
r
a
i
n
t
u
m
o
r
s
e
g
m
e
n
t
a
t
i
o
n
(
B
r
a
TS
-
M
ETS
)
c
h
a
l
l
e
n
g
e
2
0
2
3
:
b
r
a
i
n
m
e
t
a
st
a
si
s
seg
m
e
n
t
a
t
i
o
n
o
n
p
re
-
t
r
e
a
t
m
e
n
t
M
R
I
.
”
2
0
2
3
.
d
o
i
:
1
0
.
5
9
2
7
5
/
j
.
m
e
l
b
a
.
2
0
2
5
-
9
b
d
3
.
[
6
]
M
.
C
.
d
e
V
e
r
d
i
e
r
e
t
a
l
.
,
“
T
h
e
2
0
2
4
b
r
a
i
n
t
u
mo
r
s
e
g
me
n
t
a
t
i
o
n
(
B
r
a
TS)
ch
a
l
l
e
n
g
e
:
g
l
i
o
m
a
se
g
me
n
t
a
t
i
o
n
o
n
p
o
s
t
-
t
r
e
a
t
me
n
t
M
R
I
.
”
2
0
2
4
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s
/
2
4
0
5
.
1
8
3
6
8
[
7
]
A
.
C
r
i
mi
,
L.
D
o
d
e
r
o
,
F
.
S
a
m
b
a
t
a
r
o
,
V
.
M
u
r
i
n
o
,
a
n
d
D
.
S
o
n
a
,
“
S
t
r
u
c
t
u
r
a
l
l
y
c
o
n
s
t
r
a
i
n
e
d
e
f
f
e
c
t
i
v
e
b
r
a
i
n
c
o
n
n
e
c
t
i
v
i
t
y
,
”
N
e
u
ro
I
m
a
g
e
,
v
o
l
.
2
3
9
,
p
.
1
1
8
2
8
8
,
O
c
t
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
n
e
u
r
o
i
ma
g
e
.
2
0
2
1
.
1
1
8
2
8
8
.
[
8
]
S
.
B
a
k
a
s
e
t
a
l
.
,
“
I
d
e
n
t
i
f
y
i
n
g
t
h
e
b
e
s
t
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
m
s
f
o
r
b
r
a
i
n
t
u
m
o
r
s
e
g
me
n
t
a
t
i
o
n
,
p
r
o
g
r
e
ssi
o
n
a
sses
sme
n
t
,
a
n
d
o
v
e
r
a
l
l
s
u
r
v
i
v
a
l
p
r
e
d
i
c
t
i
o
n
i
n
t
h
e
B
R
A
TS
ch
a
l
l
e
n
g
e
.
”
2
0
1
9
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s/
1
8
1
1
.
0
2
6
2
9
[
9
]
S
.
B
a
u
e
r
,
R
.
W
i
e
st
,
L.
P
.
N
o
l
t
e
,
a
n
d
M
.
R
e
y
e
s,
“
A
su
r
v
e
y
o
f
M
R
I
-
b
a
sed
me
d
i
c
a
l
i
ma
g
e
a
n
a
l
y
si
s
f
o
r
b
r
a
i
n
t
u
m
o
r
st
u
d
i
e
s,
”
Ph
y
s
i
c
s
i
n
Me
d
i
c
i
n
e
a
n
d
B
i
o
l
o
g
y
,
v
o
l
.
5
8
,
n
o
.
1
3
,
p
p
.
R
9
7
–
R
1
2
9
,
J
u
l
.
2
0
1
3
,
d
o
i
:
1
0
.
1
0
8
8
/
0
0
3
1
-
9
1
5
5
/
5
8
/
1
3
/
R
9
7
.
[
1
0
]
F
.
I
s
e
n
s
e
e
,
P
.
F
.
J
ä
g
e
r
,
P
.
M
.
F
u
l
l
,
P
.
V
o
l
l
m
u
t
h
,
a
n
d
K
.
H
.
M
a
i
e
r
-
H
e
i
n
,
“
n
n
U
-
N
e
t
f
o
r
b
r
a
i
n
t
u
m
o
r
s
e
g
m
e
n
t
a
t
i
o
n
,
”
L
e
c
t
u
r
e
N
o
t
e
s
i
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
i
n
c
l
u
d
i
n
g
s
u
b
s
e
r
i
e
s
L
e
c
t
u
r
e
N
o
t
e
s
i
n
A
r
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
L
e
c
t
u
r
e
N
o
t
e
s
i
n
B
i
o
i
n
f
o
r
m
a
t
i
c
s
)
,
v
o
l
.
1
2
6
5
9
L
N
C
S
.
p
p
.
1
1
8
–
1
3
2
,
2
0
2
1
.
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
0
-
7
2
0
8
7
-
2
_
1
1
.
[
1
1
]
P
.
Li
,
W
.
W
u
,
L.
Li
u
,
F
.
M
i
c
h
a
e
l
S
e
r
r
y
,
J
.
W
a
n
g
,
a
n
d
H
.
H
a
n
,
“
A
u
t
o
m
a
t
i
c
b
r
a
i
n
t
u
m
o
r
s
e
g
m
e
n
t
a
t
i
o
n
f
r
o
m
m
u
l
t
i
p
a
r
a
me
t
r
i
c
M
R
I
b
a
s
e
d
o
n
c
a
sc
a
d
e
d
3
D
U
-
N
e
t
a
n
d
3
D
U
-
N
e
t
+
+
,
”
Bi
o
m
e
d
i
c
a
l
S
i
g
n
a
l
Pr
o
c
e
ssi
n
g
a
n
d
C
o
n
t
ro
l
,
v
o
l
.
7
8
,
p
.
1
0
3
9
7
9
,
S
e
p
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
s
p
c
.
2
0
2
2
.
1
0
3
9
7
9
.
[
1
2
]
Y
.
Z
h
a
n
g
,
X
.
Li
u
,
S
.
W
a
,
Y
.
Li
u
,
J.
K
a
n
g
,
a
n
d
C
.
L
v
,
“
G
e
n
U
-
N
e
t
+
+
:
a
n
a
u
t
o
m
a
t
i
c
i
n
t
r
a
c
r
a
n
i
a
l
b
r
a
i
n
t
u
m
o
r
s
se
g
m
e
n
t
a
t
i
o
n
a
l
g
o
r
i
t
h
m
o
n
3
D
i
m
a
g
e
ser
i
e
s
w
i
t
h
h
i
g
h
p
e
r
f
o
r
ma
n
c
e
,
”
S
y
m
m
e
t
ry
,
v
o
l
.
1
3
,
n
o
.
1
2
,
p
.
2
3
9
5
,
D
e
c
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
s
y
m1
3
1
2
2
3
9
5
.
[
1
3
]
Ö
.
Ç
i
ç
e
k
,
A
.
A
b
d
u
l
k
a
d
i
r
,
S
.
S
.
L
i
e
n
k
a
m
p
,
T
.
B
r
o
x
,
a
n
d
O
.
R
o
n
n
e
b
e
r
g
e
r
,
“
3
D
U
-
n
e
t
:
L
e
a
r
n
i
n
g
d
e
n
s
e
v
o
l
u
m
e
t
r
i
c
s
e
g
m
e
n
t
a
t
i
o
n
f
r
o
m
s
p
a
r
s
e
a
n
n
o
t
a
t
i
o
n
,
”
i
n
L
e
c
t
u
r
e
N
o
t
e
s
i
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
9
9
0
1
L
N
C
S
,
2
0
1
6
,
p
p
.
4
2
4
-
4
3
2
.
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
319
-
4
6
7
2
3
-
8
_
4
9
.
[
1
4
]
X
.
G
u
a
n
e
t
a
l
.
,
“
3
D
A
G
S
E
-
V
N
e
t
:
a
n
a
u
t
o
m
a
t
i
c
b
r
a
i
n
t
u
m
o
r
M
R
I
d
a
t
a
se
g
men
t
a
t
i
o
n
f
r
a
mew
o
r
k
,
”
BM
C
Me
d
i
c
a
l
I
m
a
g
i
n
g
,
v
o
l
.
2
2
,
n
o
.
1
,
p
.
6
,
D
e
c
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
8
6
/
s
1
2
8
8
0
-
0
2
1
-
0
0
7
2
8
-
8.
[
1
5
]
J.
C
a
o
,
J.
L
i
u
,
a
n
d
J.
C
h
e
n
,
“
A
b
r
a
i
n
t
u
m
o
r
se
g
m
e
n
t
a
t
i
o
n
met
h
o
d
b
a
se
d
o
n
a
t
t
e
n
t
i
o
n
me
c
h
a
n
i
sm
,
”
S
c
i
e
n
t
i
f
i
c
R
e
p
o
rt
s
,
v
o
l
.
1
5
,
n
o
.
1
,
p
.
1
5
2
2
9
,
A
p
r
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
3
8
/
s
4
1
5
9
8
-
0
2
5
-
9
8
3
5
5
-
8.
[
1
6
]
G
.
Li
t
j
e
n
s
e
t
a
l
.
,
“
A
s
u
r
v
e
y
o
n
d
e
e
p
l
e
a
r
n
i
n
g
i
n
me
d
i
c
a
l
i
m
a
g
e
a
n
a
l
y
si
s
,
”
Me
d
i
c
a
l
I
m
a
g
e
An
a
l
y
si
s
,
v
o
l
.
4
2
,
p
p
.
6
0
–
8
8
,
D
e
c
.
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
m
e
d
i
a
.
2
0
1
7
.
0
7
.
0
0
5
.
[
1
7
]
F
.
I
sen
see
,
C
.
U
l
r
i
c
h
,
T.
W
a
l
d
,
a
n
d
K
.
H
.
M
a
i
e
r
-
H
e
i
n
,
“
Ex
t
e
n
d
i
n
g
n
n
U
-
N
e
t
i
s
a
l
l
y
o
u
n
e
e
d
.
”
A
u
g
.
2
3
,
2
0
2
2
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s/
2
2
0
8
.
1
0
7
9
1
[
1
8
]
F
.
I
sen
se
e
e
t
a
l
.
,
“
n
n
U
-
N
e
t
R
e
v
i
s
i
t
e
d
:
A
C
a
l
l
f
o
r
R
i
g
o
r
o
u
s
V
a
l
i
d
a
t
i
o
n
i
n
3
D
M
e
d
i
c
a
l
I
mag
e
S
e
g
me
n
t
a
t
i
o
n
.
”
J
u
l
.
2
5
,
2
0
2
4
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s/
2
4
0
4
.
0
9
5
5
6
[
1
9
]
D
.
V
e
i
g
a
-
C
a
n
u
t
o
e
t
a
l
.
,
“
C
o
m
p
a
r
a
t
i
v
e
m
u
l
t
i
c
e
n
t
r
i
c
e
v
a
l
u
a
t
i
o
n
o
f
i
n
t
e
r
-
o
b
s
e
r
v
e
r
v
a
r
i
a
b
i
l
i
t
y
i
n
m
a
n
u
a
l
a
n
d
a
u
t
o
m
a
t
i
c
s
e
g
m
e
n
t
a
t
i
o
n
o
f
n
e
u
r
o
b
l
a
st
i
c
t
u
m
o
r
s
i
n
m
a
g
n
e
t
i
c
r
e
s
o
n
a
n
c
e
i
m
a
g
e
s
,
”
C
a
n
c
e
rs
,
v
o
l
.
1
4
,
n
o
.
1
5
,
p
.
3
6
4
8
,
J
u
l
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
c
a
n
c
e
r
s
1
4
1
5
3
6
4
8
.
[
2
0
]
F
.
I
sen
se
e
e
t
a
l
.
,
“
n
n
U
-
N
e
t
:
s
e
l
f
-
a
d
a
p
t
i
n
g
f
r
a
mew
o
r
k
f
o
r
U
-
N
e
t
-
b
a
se
d
m
e
d
i
c
a
l
i
m
a
g
e
se
g
me
n
t
a
t
i
o
n
.
”
S
e
p
.
2
7
,
2
0
1
8
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s
/
1
8
0
9
.
1
0
4
8
6
[
2
1
]
K
.
K
u
n
d
a
l
,
K
.
V
.
R
a
o
,
A
.
M
a
j
u
m
d
a
r
,
N
.
K
u
mar,
a
n
d
R
.
K
u
mar,
“
C
o
m
p
r
e
h
e
n
si
v
e
b
e
n
c
h
m
a
r
k
i
n
g
o
f
C
N
N
-
b
a
se
d
t
u
m
o
r
seg
m
e
n
t
a
t
i
o
n
m
e
t
h
o
d
s
u
si
n
g
m
u
l
t
i
m
o
d
a
l
M
R
I
d
a
t
a
,
”
C
o
m
p
u
t
e
rs
i
n
B
i
o
l
o
g
y
a
n
d
Me
d
i
c
i
n
e
,
v
o
l
.
1
7
8
,
p
.
1
0
8
7
9
9
,
A
u
g
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
b
i
o
me
d
.
2
0
2
4
.
1
0
8
7
9
9
.
[
2
2
]
N
.
A
y
u
b
,
M
.
W
a
q
a
s
I
q
b
a
l
,
M
.
U
.
S
a
l
e
e
m,
M
.
N
.
A
mi
n
,
O
.
I
mr
a
n
,
a
n
d
H
.
K
h
a
n
,
“
S
p
e
c
t
r
u
m
o
f
e
n
g
i
n
e
e
r
i
n
g
sc
i
e
n
c
e
s
e
f
f
i
c
i
e
n
t
M
L
t
e
c
h
n
i
q
u
e
f
o
r
b
r
a
i
n
t
u
m
o
r
se
g
me
n
t
a
t
i
o
n
,
a
n
d
d
e
t
e
c
t
i
o
n
,
b
a
s
e
d
o
n
M
R
I
sca
n
s
u
si
n
g
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
(
C
N
N
s)
S
p
e
c
t
r
u
m
o
f
E
n
g
i
n
e
e
r
i
n
g
S
c
i
e
n
c
e
s,
”
Sp
e
c
t
r
u
m
o
f
En
g
i
n
e
e
ri
n
g
S
c
i
e
n
c
e
s
,
v
o
l
.
3
,
n
o
.
3
,
p
p
.
1
8
6
–
2
1
3
,
2
0
2
5
.
[
2
3
]
O
.
R
o
n
n
e
b
e
r
g
e
r
,
P
.
F
i
sc
h
e
r
,
a
n
d
T
.
B
r
o
x
,
“
U
-
N
e
t
:
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
t
w
o
r
k
s
f
o
r
b
i
o
m
e
d
i
c
a
l
i
ma
g
e
s
e
g
m
e
n
t
a
t
i
o
n
.
”
M
a
y
1
8
,
2
0
1
5
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s/
1
5
0
5
.
0
4
5
9
7
[
2
4
]
K
.
K
a
m
n
i
t
sas
e
t
a
l
.
,
“
Ef
f
i
c
i
e
n
t
m
u
l
t
i
-
s
c
a
l
e
3
D
C
N
N
w
i
t
h
f
u
l
l
y
c
o
n
n
e
c
t
e
d
C
R
F
f
o
r
a
c
c
u
r
a
t
e
b
r
a
i
n
l
e
s
i
o
n
se
g
me
n
t
a
t
i
o
n
,
”
Me
d
i
c
a
l
I
m
a
g
e
A
n
a
l
y
s
i
s
,
v
o
l
.
3
6
,
p
p
.
6
1
–
7
8
,
F
e
b
.
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
m
e
d
i
a
.
2
0
1
6
.
1
0
.
0
0
4
.
[
2
5
]
J.
N
o
d
i
r
o
v
,
A
.
B
.
A
b
d
u
sa
l
o
mo
v
,
a
n
d
T.
K
.
W
h
a
n
g
b
o
,
“
A
t
t
e
n
t
i
o
n
3
D
U
-
N
e
t
w
i
t
h
mu
l
t
i
p
l
e
s
k
i
p
c
o
n
n
e
c
t
i
o
n
s
f
o
r
se
g
me
n
t
a
t
i
o
n
o
f
b
r
a
i
n
t
u
mo
r
i
ma
g
e
s
,
”
S
e
n
s
o
rs
,
v
o
l
.
2
2
,
n
o
.
1
7
,
p
.
6
5
0
1
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
2
1
7
6
5
0
1
.
[
2
6
]
K
.
X
u
,
M
.
Zh
a
n
g
,
S
.
J
e
g
e
l
k
a
,
a
n
d
K
.
K
a
w
a
g
u
c
h
i
,
“
O
p
t
i
m
i
z
a
t
i
o
n
o
f
g
r
a
p
h
n
e
u
r
a
l
n
e
t
w
o
r
k
s:
i
mp
l
i
c
i
t
a
c
c
e
l
e
r
a
t
i
o
n
b
y
sk
i
p
c
o
n
n
e
c
t
i
o
n
s a
n
d
m
o
r
e
d
e
p
t
h
,
”
i
n
I
n
t
e
r
n
a
t
i
o
n
a
l
c
o
n
f
e
re
n
c
e
o
n
m
a
c
h
i
n
e
l
e
a
rn
i
n
g
,
2
0
2
1
,
p
p
.
1
1
5
9
2
–
1
1
6
0
2
.
[
2
7
]
B
.
B
a
h
e
t
i
e
t
a
l
.
,
“
T
h
e
b
r
a
i
n
t
u
m
o
r
s
e
q
u
e
n
c
e
r
e
g
i
s
t
r
a
t
i
o
n
(
B
r
a
T
S
-
R
e
g
)
c
h
a
l
l
e
n
g
e
:
e
s
t
a
b
l
i
s
h
i
n
g
c
o
r
r
e
s
p
o
n
d
e
n
c
e
b
e
t
w
e
e
n
p
r
e
-
o
p
e
r
a
t
i
v
e
a
n
d
f
o
l
l
o
w
-
u
p
M
R
I
s
c
a
n
s
o
f
d
i
f
f
u
s
e
g
l
i
o
ma
p
a
t
i
e
n
t
s
.
”
A
p
r
.
1
7
,
2
0
2
1
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s
/
2
1
1
2
.
0
6
9
7
9
[
2
8
]
J.
E.
H
.
R
i
v
e
r
a
e
t
a
l
.
,
“
A
n
e
n
sem
b
l
e
a
p
p
r
o
a
c
h
f
o
r
b
r
a
i
n
t
u
m
o
r
se
g
men
t
a
t
i
o
n
a
n
d
s
y
n
t
h
e
si
s
.
”
N
o
v
.
2
6
,
2
0
2
4
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s
/
2
4
1
1
.
1
7
6
1
7
[
2
9
]
F
.
K
o
f
l
e
r
e
t
a
l
.
,
“
B
r
a
TS
T
o
o
l
k
i
t
:
Tr
a
n
sl
a
t
i
n
g
B
r
a
TS
b
r
a
i
n
t
u
mo
r
s
e
g
me
n
t
a
t
i
o
n
a
l
g
o
r
i
t
h
ms
i
n
t
o
c
l
i
n
i
c
a
l
a
n
d
sc
i
e
n
t
i
f
i
c
p
r
a
c
t
i
c
e
,
”
Fro
n
t
i
e
rs
i
n
N
e
u
ro
s
c
i
e
n
c
e
,
v
o
l
.
1
4
,
A
p
r
.
2
0
2
0
,
d
o
i
:
1
0
.
3
3
8
9
/
f
n
i
n
s.
2
0
2
0
.
0
0
1
2
5
.
[
3
0
]
S
.
K
a
ssa
m,
A
.
M
a
r
k
h
a
m,
K
.
V
o
,
Y
.
R
e
v
a
n
a
k
a
r
a
,
M
.
La
m
,
a
n
d
K
.
Z
h
u
,
“
I
n
t
r
a
o
p
e
r
a
t
i
v
e
g
l
i
o
ma
se
g
m
e
n
t
a
t
i
o
n
w
i
t
h
Y
O
LO
+
S
A
M
f
o
r
i
m
p
r
o
v
e
d
a
c
c
u
r
a
c
y
i
n
t
u
mo
r
r
e
se
c
t
i
o
n
.
”
A
u
g
.
2
7
,
2
0
2
4
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s
/
2
4
0
8
.
1
4
8
4
7
[
3
1
]
A
.
I
l
h
a
n
,
B
.
S
e
k
e
r
o
g
l
u
,
a
n
d
R
.
A
b
i
y
e
v
,
“
B
r
a
i
n
t
u
m
o
r
s
e
g
m
e
n
t
a
t
i
o
n
i
n
M
R
I
i
ma
g
e
s
u
s
i
n
g
n
o
n
p
a
r
a
m
e
t
r
i
c
l
o
c
a
l
i
z
a
t
i
o
n
a
n
d
e
n
h
a
n
c
e
me
n
t
m
e
t
h
o
d
s
w
i
t
h
U
-
n
e
t
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
C
o
m
p
u
t
e
r
Ass
i
st
e
d
R
a
d
i
o
l
o
g
y
a
n
d
S
u
r
g
e
ry
,
v
o
l
.
1
7
,
n
o
.
3
,
p
p
.
5
8
9
–
6
0
0
,
M
a
r
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
1
5
4
8
-
022
-
0
2
5
6
6
-
7.
[
3
2
]
A
.
E
.
O
r
h
a
n
a
n
d
X
.
P
i
t
k
o
w
,
“
S
k
i
p
c
o
n
n
e
c
t
i
o
n
s
e
l
i
m
i
n
a
t
e
si
n
g
u
l
a
r
i
t
i
e
s,
”
6
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
L
e
a
r
n
i
n
g
R
e
p
r
e
s
e
n
t
a
t
i
o
n
s
,
I
C
L
R
2
0
1
8
-
C
o
n
f
e
r
e
n
c
e
T
r
a
c
k
Pr
o
c
e
e
d
i
n
g
s
.
M
a
r
.
0
4
,
2
0
1
8
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s/
1
7
0
1
.
0
9
1
7
5
Evaluation Warning : The document was created with Spire.PDF for Python.