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.
40
,
No
.
2
,
N
o
v
em
b
e
r
2
0
2
5
,
p
p
.
85
0
~
85
9
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
40
.
i
2
.
pp
850
-
85
9
850
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
A sy
stema
tic
ev
a
lua
tion o
f
pre
-
trai
ned encode
r archi
tect
u
res for
multimo
da
l brain
t
umo
r seg
men
tat
io
n using
U
-
N
et
-
b
a
sed
a
rchitec
t
ures
M
a
rwa
Abba
s
1
,
As
hra
f
A.
M
.
K
ha
la
f
2
,
H
us
s
ein Mo
g
a
hed
1
,
Aziz
a
I
.
H
us
s
ein
3
,
L
a
m
y
a
G
a
ber
1
,
M.
M
o
ura
d
M
a
bro
o
k
4
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
a
n
d
S
y
st
e
m
s E
n
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
,
M
i
n
i
a
U
n
i
v
e
r
si
t
y
,
Al
-
M
i
n
i
a
,
Eg
y
p
t
2
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
C
o
m
mu
n
i
c
a
t
i
o
n
s
En
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
,
M
i
n
i
a
U
n
i
v
e
r
si
t
y
,
Al
-
M
i
n
i
a
,
E
g
y
p
t
3
El
e
c
t
r
i
c
a
l
a
n
d
C
o
mp
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
D
e
p
a
r
t
me
n
t
,
Ef
f
a
t
U
n
i
v
e
r
s
i
t
y
,
Je
d
d
a
h
,
A
r
a
b
S
a
u
d
i
4
S
p
a
c
e
C
o
mm
u
n
i
c
a
t
i
o
n
D
e
p
a
r
t
me
n
t
,
F
a
c
u
l
t
y
o
f
N
a
v
i
g
a
t
i
o
n
S
c
i
e
n
c
e
a
n
d
S
p
a
c
e
Te
c
h
n
o
l
o
g
y
,
B
e
n
i
-
S
u
e
f
U
n
i
v
e
r
si
t
y
,
B
e
n
i
-
S
u
e
f
,
E
g
y
p
t
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
an
19
,
2
0
2
5
R
ev
is
ed
J
u
l
28
,
2
0
2
5
Acc
ep
ted
Oct
14
,
2
0
2
5
A
c
c
u
ra
t
e
b
ra
i
n
t
u
m
o
r
se
g
m
e
n
t
a
t
i
o
n
f
r
o
m
m
e
d
i
c
a
l
i
m
a
g
i
n
g
i
s
c
r
i
t
ic
a
l
f
o
r
e
a
r
l
y
d
i
a
g
n
o
s
i
s
a
n
d
e
f
f
e
c
t
i
v
e
t
re
a
tm
e
n
t
p
l
a
n
n
i
n
g
.
D
e
e
p
l
e
a
r
n
i
n
g
m
e
t
h
o
d
s
,
p
a
r
t
i
c
u
l
a
r
l
y
U
-
Ne
t
-
b
a
s
e
d
a
rc
h
i
t
e
c
tu
r
e
s
,
h
a
v
e
d
e
m
o
n
s
t
ra
t
e
d
s
t
r
o
n
g
p
e
r
f
o
r
m
a
n
c
e
i
n
t
h
i
s
d
o
m
a
i
n
.
H
o
w
e
v
e
r
,
p
r
i
o
r
s
t
u
d
i
e
s
h
a
v
e
p
r
i
m
a
r
i
l
y
f
o
c
u
se
d
o
n
l
i
m
i
t
e
d
e
n
c
o
d
e
r
b
a
c
k
b
o
n
e
s
,
o
v
e
r
l
o
o
k
i
n
g
t
h
e
p
o
t
e
n
t
i
a
l
a
d
v
a
n
t
a
g
e
s
o
f
a
l
t
e
r
n
a
t
i
v
e
p
r
e
t
r
a
i
n
e
d
m
o
d
e
l
s
.
T
h
i
s
s
t
u
d
y
p
r
e
s
e
n
t
s
a
s
y
s
t
e
m
a
t
ic
e
v
a
l
u
a
t
i
o
n
o
f
t
w
e
l
v
e
p
r
e
t
r
a
i
n
e
d
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
R
e
s
Ne
t
3
4
,
R
e
s
Ne
t
5
0
,
R
e
s
Ne
t
1
0
1
,
VG
G
1
6
,
V
G
G
1
9
,
D
e
n
se
N
e
t1
2
1
,
I
n
c
e
p
t
i
o
n
R
e
s
N
e
t
V
2
,
I
n
c
e
p
t
i
o
n
V
3
,
M
o
b
i
l
e
N
e
t
V
2
,
E
f
f
i
c
ie
n
t
N
e
t
B
1
,
S
E
-
R
e
s
Ne
t
3
4
,
a
n
d
S
E
-
R
e
s
Ne
t
1
8
—
u
s
e
d
a
s
e
n
c
o
d
e
r
b
a
c
k
b
o
n
e
s
i
n
t
h
e
U
-
N
e
t
f
ra
m
e
w
o
r
k
f
o
r
i
d
e
n
t
i
f
i
c
a
t
i
o
n
a
n
d
e
x
tr
a
c
t
i
o
n
o
f
t
u
m
o
r
-
a
f
f
e
c
t
e
d
b
ra
i
n
a
re
a
s
u
s
i
n
g
t
h
e
B
ra
T
S
2
0
1
9
m
u
l
t
i
m
o
d
a
l
M
R
I
d
a
t
a
s
e
t
.
M
o
d
e
l
p
e
r
f
o
r
m
a
n
c
e
wa
s
a
s
se
ss
e
d
t
h
r
o
u
g
h
c
r
o
s
s
-
v
a
l
i
d
a
t
i
o
n
,
i
n
c
o
r
p
o
r
a
t
i
n
g
f
a
u
l
t
d
e
t
e
c
t
i
o
n
t
o
e
n
h
a
n
c
e
re
l
i
a
b
i
l
i
t
y
.
Th
e
M
o
b
i
l
e
N
e
t
V
2
-
b
a
s
e
d
U
-
N
e
t
c
o
n
f
i
g
u
r
a
t
i
o
n
o
u
t
p
e
r
f
o
r
m
e
d
a
l
l
o
t
h
e
r
a
rc
h
i
t
e
c
t
u
re
s
,
a
c
h
i
e
v
i
n
g
9
9
%
c
r
o
s
s
-
v
a
l
i
d
a
t
i
o
n
a
c
c
u
r
a
c
y
a
n
d
9
9
.
3
%
t
e
s
t
a
c
c
u
r
a
c
y
.
A
d
d
i
t
i
o
n
a
l
l
y
,
i
t
a
c
h
i
e
v
e
d
a
J
a
c
c
a
r
d
c
o
e
f
f
i
c
i
e
n
t
o
f
8
3
.
4
5
%
,
a
n
d
D
i
c
e
c
o
e
f
f
i
c
ie
n
t
s
o
f
9
0
.
3
%
(
W
h
o
l
e
T
u
m
o
r
)
,
8
6
.
0
7
%
(
T
u
m
o
r
C
o
r
e
)
,
a
n
d
8
1
.
9
3
%
(
E
n
h
a
n
c
i
n
g
T
u
m
o
r
)
,
w
i
t
h
a
l
o
w
-
te
s
t
l
o
s
s
o
f
0
.
0
2
8
2
.
T
h
e
s
e
r
e
s
u
l
t
s
d
e
m
o
n
s
t
r
a
te
t
h
a
t
M
o
b
i
l
e
Ne
t
V
2
i
s
a
h
i
g
h
l
y
e
f
f
e
c
t
i
v
e
e
n
c
o
d
e
r
b
a
c
k
b
o
n
e
f
o
r
U
-
N
e
t
i
n
e
x
t
r
a
c
t
i
n
g
t
a
s
k
s
f
o
r
t
u
m
o
r
-
a
f
f
e
c
t
e
d
b
r
a
i
n
r
e
g
io
n
s
u
s
i
n
g
m
u
l
t
i
m
o
d
a
l
m
e
d
ic
a
l
i
m
a
g
i
n
g
d
a
t
a
.
K
ey
w
o
r
d
s
:
B
r
ain
tu
m
o
r
ca
n
ce
r
E
n
co
d
er
s
Me
d
ical
im
ag
in
g
Pre
-
tr
ain
ed
U
-
Net
s
eg
m
en
tatio
n
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
:
Ma
r
wa
Ab
b
as
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
a
n
d
Sy
s
tem
s
E
n
g
in
ee
r
in
g
,
Facu
lty
o
f
E
n
g
i
n
ee
r
in
g
,
Min
ia
Un
iv
er
s
ity
Al
-
Min
ia,
6
1
1
1
1
,
E
g
y
p
t
E
m
ail: M
ar
wa.
ab
as@
m
u
.
ed
u
.
eg
1.
I
NT
RO
D
UCT
I
O
N
B
r
ain
tu
m
o
r
s
,
d
ef
in
e
d
b
y
ab
er
r
an
t
n
eu
r
al
ce
ll
p
r
o
life
r
atio
n
,
co
n
s
titu
te
o
n
e
o
f
th
e
m
o
s
t
d
if
f
icu
lt
p
r
o
b
lem
s
i
n
n
e
u
r
o
-
o
n
c
o
lo
g
y
,
with
s
ig
n
if
ican
t
g
l
o
b
al
in
ci
d
e
n
ce
(
e.
g
.
,
2
3
,
8
2
0
n
ew
ca
s
es
r
e
p
o
r
ted
in
th
e
US
in
2
0
1
9
[
1
]
,
[
2
]
)
.
Acc
u
r
ate
d
ia
g
n
o
s
is
an
d
m
an
ag
e
m
en
t
o
f
t
h
ese
tu
m
o
r
s
r
ely
h
ea
v
ily
o
n
ad
v
an
ce
d
m
e
d
ical
im
ag
in
g
m
o
d
alities
,
p
ar
ticu
lar
ly
m
u
ltip
ar
am
etr
ic
MRI
(
T
1
,
T
1
c,
T
2
,
FLAI
R
)
,
wh
ich
p
r
o
v
id
es
s
u
p
er
io
r
s
o
f
t
-
tis
s
u
e
co
n
tr
ast
f
o
r
d
elin
ea
tin
g
tu
m
o
r
b
o
u
n
d
ar
ies
co
m
p
ar
ed
to
C
T
o
r
u
ltra
s
o
u
n
d
[
3
]
,
[
4
]
.
W
h
ile
th
ese
im
ag
in
g
tech
n
iq
u
es
ar
e
in
d
is
p
en
s
ab
le,
m
an
u
al
s
eg
m
en
tatio
n
o
f
tu
m
o
r
s
u
b
r
eg
io
n
s
(
en
h
an
cin
g
tu
m
o
r
,
n
ec
r
o
tic
co
r
e,
an
d
p
er
itu
m
o
r
al
ed
em
a)
r
em
ain
s
lab
o
r
-
i
n
ten
s
iv
e,
tim
e
-
c
o
n
s
u
m
in
g
,
a
n
d
s
u
b
ject
to
in
te
r
-
r
ater
v
a
r
iab
ilit
y
,
u
n
d
er
s
co
r
i
n
g
th
e
c
r
itical
n
ee
d
f
o
r
au
t
o
m
ated
,
h
ig
h
-
p
r
ec
is
io
n
co
m
p
u
tatio
n
al
m
eth
o
d
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
A
s
ystema
tic
ev
a
lu
a
tio
n
o
f p
r
e
-
tr
a
in
ed
en
co
d
er a
r
ch
itectu
r
es fo
r
mu
ltimo
d
a
l b
r
a
in
tu
mo
r
…
(
Ma
r
w
a
A
b
b
a
s
)
851
T
r
ad
itio
n
al
m
ac
h
in
e
lear
n
in
g
ap
p
r
o
ac
h
es
f
o
r
b
r
ai
n
tu
m
o
r
s
eg
m
en
tatio
n
o
f
ten
s
tr
u
g
g
le
with
li
m
ited
f
ea
tu
r
e
g
en
er
aliza
b
ilit
y
an
d
d
ep
en
d
en
ce
o
n
h
an
d
c
r
af
ted
f
ea
tu
r
e
en
g
in
ee
r
in
g
.
I
n
co
n
tr
ast,
d
ee
p
lear
n
in
g
(
DL
)
,
p
ar
ticu
lar
ly
c
o
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
C
NNs),
h
as
s
h
o
wn
p
o
wer
f
u
l
s
u
cc
ess
in
m
ed
i
ca
l
im
ag
e
a
n
aly
s
is
th
r
o
u
g
h
au
to
n
o
m
o
u
s
ac
q
u
is
itio
n
o
f
h
ie
r
ar
c
h
ical
f
ea
tu
r
es.
A
m
o
n
g
DL
a
r
ch
itectu
r
es,
U
-
Ne
t
h
as
em
er
g
ed
as
th
e
d
e
f
ac
to
s
tan
d
ar
d
f
o
r
b
io
m
e
d
ical
s
eg
m
en
tatio
n
,
u
tili
zin
g
an
en
co
d
er
-
d
ec
o
d
er
co
n
f
ig
u
r
atio
n
co
m
b
in
e
d
with
s
k
ip
co
n
n
ec
tio
n
s
,
wh
ich
h
el
p
m
ain
tain
s
p
atial
d
etails
d
u
r
in
g
th
e
le
ar
n
i
n
g
p
r
o
ce
s
s
[
5
]
,
[
6
]
.
Ho
wev
er
,
U
-
Net
im
p
lem
en
tatio
n
s
f
ac
e
ch
allen
g
es,
in
clu
d
in
g
h
ig
h
co
m
p
u
t
atio
n
al
co
m
p
lex
ity
an
d
d
ep
e
n
d
en
ce
o
n
s
izab
le
an
n
o
tated
d
atasets
,
a
m
ajo
r
o
b
s
tacle
in
th
e
d
o
m
ain
o
f
m
e
d
ical
im
ag
in
g
d
u
e
to
th
e
s
ca
r
city
o
f
lab
eled
d
ata.
T
o
ad
d
r
ess
th
ese
lim
itatio
n
s
,
tr
an
s
f
er
lear
n
in
g
(
T
L
)
h
as
b
ee
n
in
cr
ea
s
in
g
ly
a
d
o
p
ted
,
em
p
l
o
y
in
g
p
r
e
-
tr
ain
ed
C
NNs
(
e.
g
.
,
R
e
s
Net,
VGG,
Den
s
eNe
t)
as
en
co
d
er
s
to
en
h
an
ce
f
ea
tu
r
e
ex
tr
ac
t
io
n
wh
ile
r
ed
u
cin
g
tr
ain
in
g
tim
e
[
7
]
,
[
8
]
.
R
ec
en
t s
tu
d
ies h
ig
h
lig
h
t th
e
ef
f
icac
y
o
f
T
L
-
in
teg
r
ated
U
-
Net
v
ar
ian
ts
.
Nawaz
et
a
l.
[
9
]
ac
h
iev
e
d
0
.
8
1
–
0
.
8
8
Dice
s
im
ilar
ity
co
ef
f
ic
ien
t
f
o
r
t
u
m
o
r
s
u
b
r
eg
i
o
n
s
u
s
in
g
VGG1
9
-
U
-
Net.
W
h
ile
Gh
o
s
h
et
a
l
.
[
1
0
]
h
av
e
s
u
g
g
ested
a
k
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
m
et
h
o
d
an
d
U
-
Net
ar
ch
itectu
r
e
em
p
lo
y
in
g
VGG
-
1
6
as
a
b
ac
k
b
o
n
e
n
etwo
r
k
f
o
r
s
e
g
m
e
n
tin
g
b
r
ain
tu
m
o
r
s
.
T
h
ei
r
m
o
d
el
ac
h
iev
ed
a
9
2
%
Dice
s
co
r
e
o
n
th
e
T
C
GA
-
L
GG
d
ata
s
et.
L
in
et
a
l.
[
1
1
]
r
ep
o
r
ted
im
p
r
o
v
e
d
ef
f
icien
cy
with
E
f
f
icie
n
tNetV2
-
U
-
Net
f
o
r
m
u
lti
-
s
eq
u
en
ce
MRI.
R
ab
b
y
et
a
l.
[
1
2
]
attain
ed
an
8
6
%
Dice
s
co
r
e
u
s
in
g
I
n
ce
p
tio
n
V
3
-
U
-
Net,
wh
ile
3
D
ex
ten
s
io
n
s
f
u
r
th
er
im
p
r
o
v
e
d
m
u
lti
-
m
o
d
al
f
u
s
io
n
.
Saif
u
llah
et
a
l.
[
1
3
]
s
u
g
g
ested
a
co
m
b
in
ed
m
o
d
el
th
at
in
teg
r
ates
R
esNe
t5
0
with
De
ep
L
ab
V3
,
ac
h
iev
i
n
g
a
9
6
.
9
%
Dice
s
co
r
e
o
n
th
e
Fig
s
h
ar
e
d
ataset.
Au
th
o
r
s
in
[
1
4
]
.
Pro
p
o
s
ed
a
f
r
am
ewo
r
k
th
at
u
tili
ze
s
th
e
E
f
f
i
cie
n
tNetB
4
as
its
f
ea
tu
r
e
ex
t
r
ac
tio
n
b
ac
k
b
o
n
e.
E
f
f
icien
tNetB
4
em
p
lo
y
s
a
m
eth
o
d
o
f
m
i
x
tu
r
e
s
ca
lin
g
th
at
en
h
an
ce
s
th
e
n
etwo
r
k
'
s
wid
th
,
d
ep
th
,
an
d
r
eso
lu
tio
n
to
g
et
a
g
o
o
d
b
alan
ce
b
etwe
en
p
er
f
o
r
m
an
ce
an
d
co
m
p
u
tatio
n
al
ef
f
icien
cy
.
T
h
eir
m
o
d
el
s
co
r
ed
9
3
.
3
9
%
Dic
e
s
co
r
e
o
n
th
e
Fig
s
h
ar
e
d
ataset.
An
o
th
er
s
tu
d
y
[
1
5
]
p
r
esen
ts
a
cu
s
to
m
ap
p
r
o
ac
h
u
s
in
g
C
o
n
v
o
lu
tio
n
al
Neu
r
al
Netwo
r
k
s
(
C
NNs)
with
Den
s
eNe
t2
0
1
to
d
etec
t
an
d
ca
teg
o
r
ize
Acu
te
ly
m
p
h
o
b
last
ic
leu
k
ae
m
ia
ca
s
es
u
s
in
g
3
5
6
2
b
lo
o
d
s
m
ea
r
im
ag
es f
r
o
m
8
9
p
atien
ts
,
ac
h
iev
in
g
9
8
%
s
eg
m
en
tatio
n
ac
cu
r
ac
y
an
d
9
7
.
0
9
% test ac
cu
r
ac
y
.
Desp
ite
th
ese
ad
v
an
ce
s
,
a
s
y
s
tem
atic
co
m
p
ar
is
o
n
o
f
p
r
e
-
tr
ain
ed
en
c
o
d
er
s
f
o
r
b
r
ain
tu
m
o
r
s
eg
m
en
tatio
n
r
em
ain
s
u
n
d
er
e
x
p
lo
r
ed
,
p
ar
ticu
lar
ly
r
eg
ar
d
in
g
th
eir
co
m
p
u
tatio
n
al
ef
f
icie
n
cy
,
s
eg
m
en
tatio
n
ac
cu
r
ac
y
,
a
n
d
ab
ilit
y
t
o
g
en
e
r
a
lize
ac
r
o
s
s
tu
m
o
r
ca
teg
o
r
ies.
T
h
is
s
tu
d
y
p
r
esen
ts
a
c
o
m
p
r
eh
en
s
iv
e
ev
alu
ati
o
n
o
f
1
2
p
r
e
-
tr
ain
ed
C
NN
ar
c
h
itectu
r
es
as
U
-
Net
en
co
d
er
s
f
o
r
m
u
ltim
o
d
al
b
r
ai
n
tu
m
o
r
s
eg
m
e
n
tatio
n
,
u
s
in
g
th
e
B
r
aT
S 2
0
1
9
d
ataset.
Key
in
n
o
v
atio
n
s
in
clu
d
e:
a)
Ar
ch
itectu
r
al
B
en
ch
m
ar
k
i
n
g
:
Pre
cise
co
m
p
ar
is
o
n
o
f
R
esNet3
4
/5
0
/1
0
1
,
VGG1
6
/1
9
,
Den
s
eNe
t1
2
1
,
I
n
ce
p
tio
n
R
esNetV2
/V3
,
Mo
b
i
leNe
tV2
,
an
d
E
f
f
icien
tNetB
1
,
q
u
an
tify
in
g
p
er
f
o
r
m
an
ce
v
ia
Dice
,
J
ac
ca
r
d
,
an
d
co
m
p
u
tatio
n
al
m
et
r
ics.
b)
No
v
el
SE
-
R
esNet
I
n
teg
r
atio
n
,
wh
ich
u
tili
z
es
s
q
u
ee
ze
-
an
d
-
ex
citatio
n
b
lo
ck
s
to
im
p
r
o
v
e
f
ea
tu
r
e
ex
tr
ac
tio
n
,
SE
-
R
esNet1
8
/3
4
w
as f
ir
s
t u
s
ed
as U
-
Net
en
co
d
er
s
.
c)
C
o
m
p
u
tatio
n
al
-
ac
cu
r
ac
y
tr
ad
e
-
o
f
f
an
al
y
s
is
:
I
d
en
tific
atio
n
o
f
Mo
b
ileNetV2
as
an
o
p
t
im
al
en
co
d
er
,
b
alan
cin
g
ac
c
u
r
ac
y
a
n
d
ef
f
icien
cy
.
T
h
is
m
an
u
s
cr
ip
t'
s
s
ec
tio
n
s
ar
e
p
r
ep
ar
ed
as
f
o
llo
ws:
Sectio
n
2
d
em
o
n
s
tr
ates
th
e
s
u
g
g
ested
f
r
am
ewo
r
k
an
d
v
ar
i
o
u
s
f
r
am
ew
o
r
k
s
em
p
lo
y
ed
i
n
th
e
r
esear
ch
,
alo
n
g
with
im
p
lem
en
tatio
n
s
p
ec
if
ics
an
d
ass
ess
m
en
t
cr
iter
ia.
Sectio
n
3
o
u
tlin
es
th
e
o
u
tco
m
es
an
d
s
u
b
s
eq
u
e
n
t
an
aly
s
is
f
o
r
twelv
e
p
r
e
-
tr
ain
e
d
e
n
co
d
er
s
.
Sectio
n
4
illu
s
tr
ates th
e
co
n
clu
s
io
n
o
f
th
is
wo
r
k
.
2.
M
E
T
H
O
D
T
h
e
ex
p
er
im
e
n
tal
m
eth
o
d
o
l
o
g
y
u
s
ed
in
th
is
s
tu
d
y
is
ex
p
lai
n
ed
in
th
is
s
ec
tio
n
.
I
n
itially
,
a
p
u
b
licly
ac
ce
s
s
ib
le
m
u
ltimo
d
al
MRI
B
r
aT
S
2
0
1
9
d
ataset
was
s
elec
ted
.
A
s
eq
u
en
ce
o
f
p
r
e
p
r
o
ce
s
s
in
g
s
tep
s
was
u
s
ed
to
en
s
u
r
e
co
m
p
atib
ilit
y
with
t
h
e
d
ee
p
lear
n
in
g
m
o
d
els.
T
h
ese
tr
an
s
f
o
r
m
atio
n
s
s
tan
d
ar
d
ized
t
h
e
im
a
g
e
d
im
en
s
io
n
s
to
2
2
4
x
2
2
4
p
ix
els,
in
clu
d
i
n
g
r
o
tatio
n
an
d
co
n
t
r
a
s
t
ad
ju
s
tm
en
t.
Fo
llo
win
g
d
ata
p
r
ep
r
o
ce
s
s
in
g
,
th
e
ch
o
s
en
d
ee
p
lea
r
n
in
g
m
o
d
els
wer
e
im
p
lem
en
ted
,
a
n
d
th
eir
p
er
f
o
r
m
an
ce
was
s
u
b
s
eq
u
en
t
ly
ev
alu
ated
u
s
in
g
r
elev
an
t
m
etr
ics.
Ad
d
itio
n
ally
,
th
is
s
ec
tio
n
o
f
f
er
s
a
th
o
r
o
u
g
h
ex
p
lan
atio
n
o
f
th
e
d
ataset,
a
b
r
ief
ex
am
in
atio
n
o
f
th
e
d
ee
p
lear
n
in
g
ar
c
h
itectu
r
es
th
at
h
a
v
e
b
ee
n
ap
p
lied
i
n
ex
ec
u
tio
n
,
a
n
d
a
n
in
clu
s
iv
e
ev
alu
atio
n
o
f
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
.
2
.
1
.
St
a
nd
a
rd
U
-
Net
m
o
del
T
h
e
U
-
Net
ar
ch
itectu
r
e
[
1
6
]
,
cr
ea
ted
f
o
r
th
e
s
eg
m
en
tatio
n
o
f
b
io
m
e
d
ical
im
ag
es,
is
a
f
u
lly
co
n
v
o
l
u
tio
n
al
n
etwo
r
k
(
FC
N)
d
is
tin
g
u
is
h
ed
b
y
its
u
n
iq
u
e
U
-
sh
ap
ed
to
p
o
lo
g
y
.
T
h
er
e
a
r
e
two
m
ain
p
ar
ts
to
th
is
ar
ch
itectu
r
e:
a)
C
o
n
tex
tu
al
in
f
o
r
m
atio
n
an
d
a
b
s
tr
ac
t
f
ea
tu
r
es
ar
e
ca
p
tu
r
ed
b
y
th
e
C
o
n
tr
ac
tin
g
Path
(
E
n
c
o
d
er
)
,
wh
ic
h
g
r
ad
u
al
ly
d
o
wn
s
am
p
les
th
e
i
n
p
u
t
im
ag
e.
C
o
n
v
o
lu
tio
n
al
lay
er
s
ar
e
ap
p
lied
r
ep
ea
ted
ly
,
an
d
th
en
p
o
o
lin
g
o
p
er
atio
n
s
ar
e
p
er
f
o
r
m
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.
40
,
No
.
2
,
No
v
em
b
er
20
25
:
8
5
0
-
85
9
852
b)
T
h
e
ex
p
an
d
in
g
p
ath
(
Dec
o
d
er
)
im
p
r
o
v
es
s
p
atial
r
eso
lu
tio
n
b
y
u
p
s
am
p
lin
g
f
ea
tu
r
e
m
ap
s
f
r
o
m
th
e
co
n
tr
ac
tin
g
p
ath
an
d
u
s
in
g
s
k
i
p
co
n
n
ec
tio
n
s
to
m
er
g
e
lo
w
s
p
atial
r
eso
lu
tio
n
co
n
tex
tu
al
i
n
f
o
r
m
atio
n
with
h
ig
h
s
p
atial
r
eso
lu
tio
n
l
o
ca
lized
f
ea
tu
r
es,
as p
r
esen
te
d
in
Fig
u
r
e
1
.
Fig
u
r
e
1
.
U
-
n
et
m
o
d
el
s
tr
u
ctu
r
e
v
is
u
aliza
tio
n
2
.
2
.
P
r
o
po
s
ed
m
o
dels
f
o
r
en
co
ders
T
h
i
s
f
r
a
m
e
w
o
r
k
i
n
v
o
l
v
e
s
a
n
u
p
s
a
m
p
l
i
n
g
p
a
t
h
(
d
e
c
o
d
e
r
)
i
n
t
e
n
d
e
d
t
o
r
e
s
t
o
r
e
s
p
a
t
i
al
r
e
s
o
lu
t
i
o
n
a
n
d
i
m
p
r
o
v
e
t
h
e
s
e
g
m
e
n
ta
t
i
o
n
o
u
t
p
u
t
,
a
n
d
a
d
o
w
n
s
a
m
p
l
i
n
g
p
a
t
h
(
e
n
c
o
d
e
r
)
r
e
s
p
o
n
s
i
b
l
e
f
o
r
f
e
a
t
u
r
e
e
x
t
r
a
ct
i
o
n
.
I
n
i
t
i
a
ll
y
,
t
h
e
d
e
c
o
d
e
r
c
o
m
p
o
n
en
t
w
i
t
h
i
n
t
h
e
U
-
N
e
t
a
r
c
h
i
t
e
c
t
u
r
e
w
as
h
e
l
d
c
o
n
s
t
a
n
t
,
w
h
i
l
e
t
h
e
e
n
c
o
d
e
r
c
o
m
p
o
n
e
n
t
w
a
s
r
e
p
l
ac
e
d
w
it
h
p
r
e
-
t
r
a
i
n
e
d
m
o
d
e
l
s
.
T
h
e
p
r
o
p
o
s
e
d
s
t
u
d
y
u
ti
l
i
z
e
d
1
2
p
r
e
-
t
r
a
i
n
e
d
e
n
c
o
d
e
r
a
lg
o
r
i
t
h
m
s
,
i
n
c
l
u
d
i
n
g
R
es
N
e
t
3
4
,
R
es
Ne
t
5
0
,
R
es
N
et
1
0
1
,
V
G
G
1
6
,
V
G
G
1
9
,
D
e
n
s
e
N
e
t
1
2
1
,
I
n
c
e
p
t
i
o
n
R
e
s
N
e
tV
2
,
I
n
c
e
p
t
i
o
n
V
3
,
M
o
b
i
l
e
n
e
t
V
2
,
E
f
f
i
ci
e
n
t
N
e
t
B
1
,
S
E
R
e
s
N
e
t
3
4
,
a
n
d
S
E
R
es
N
et
1
8
a
s
e
n
c
o
d
e
r
s
w
i
t
h
t
h
e
f
i
x
e
d
d
e
c
o
d
e
r
p
a
r
t
i
n
t
h
e
U
-
N
e
t
a
r
c
h
i
t
e
ct
u
r
e
o
n
l
y
.
F
i
g
u
r
e
2
p
r
e
s
e
n
ts
a
c
o
m
p
r
e
h
e
n
s
i
v
e
d
e
s
c
r
i
p
t
i
o
n
o
f
t
h
e
r
es
e
a
r
c
h
w
o
r
k
f
l
o
w
,
a
n
d
T
a
b
l
e
1
c
o
m
p
a
r
e
s
p
r
e
-
t
r
a
i
n
e
d
e
n
c
o
d
e
r
s
i
n
t
e
r
m
s
o
f
t
h
e
n
u
m
b
e
r
o
f
l
a
y
e
r
s
.
Fig
u
r
e
2
.
Ov
e
r
v
iew
o
f
th
e
in
v
esti
g
atio
n
'
s
wo
r
k
f
lo
w
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
A
s
ystema
tic
ev
a
lu
a
tio
n
o
f p
r
e
-
tr
a
in
ed
en
co
d
er a
r
ch
itectu
r
es fo
r
mu
ltimo
d
a
l b
r
a
in
tu
mo
r
…
(
Ma
r
w
a
A
b
b
a
s
)
853
T
ab
le
1
.
C
o
m
p
a
r
is
o
n
o
f
p
r
e
-
t
r
ain
ed
en
co
d
er
s
in
ter
m
s
o
f
n
u
m
b
er
o
f
lay
er
s
M
o
d
e
l
N
u
mb
e
r
o
f
La
y
e
r
s
D
e
n
seN
e
t
-
1
2
1
1
2
1
R
e
sN
e
t
3
4
34
R
e
sN
e
t
5
0
50
R
e
sN
e
t
1
0
1
1
0
1
V
G
G
1
6
16
V
G
G
1
9
19
I
n
c
e
p
t
i
o
n
-
v3
48
I
n
c
e
p
t
i
o
n
-
R
e
s
n
e
t
V
2
1
6
4
S
E
R
e
sN
e
t
1
8
1
8
3
S
E
R
e
sN
e
t
3
4
3
1
1
Ef
f
i
c
i
e
n
t
N
e
t
B
2
3
7
3
M
o
b
i
l
e
n
e
t
V
2
1
9
6
I
m
ag
eNe
t
weig
h
ts
wer
e
u
s
ed
to
in
itialize
th
ese
b
ac
k
b
o
n
e
e
n
co
d
er
s
,
an
d
th
e
en
co
d
er
lay
e
r
s
wer
e
all
k
ep
t
f
r
o
ze
n
.
T
h
e
n
,
th
e
1
2
p
r
e
-
tr
ain
ed
en
co
d
er
s
wer
e
tr
ain
ed
o
n
th
e
s
p
ec
if
ic
d
ataset
to
g
en
er
ate
th
eir
m
o
d
els.
T
h
e
p
r
o
ce
d
u
r
es th
at
c
o
m
p
o
s
e
th
e
tr
ain
in
g
p
h
ase
ar
e
e
x
p
lain
e
d
in
Alg
o
r
ith
m
1
.
Alg
o
r
ith
m
1
.
Step
s
(
T
r
ain
in
g
Ph
ase)
Step 1: Brain MRI images are read from the
selected dataset.
St
ep
2:
Th
e
im
ag
es
ar
e
re
si
ze
d
to
ma
tc
h
th
e
in
pu
t
di
me
ns
io
ns
re
qu
ir
ed
by
th
e
tr
an
sf
er
learning models.
St
ep
3:
Th
e
da
ta
se
t
is
d
is
tr
ib
ut
ed
in
to
70
%
fo
r
tr
ai
ni
ng
,
10
%f
or
va
li
da
ti
on
,
an
d
20
%
fo
r
testing.
St
ep
4:
Fe
at
ur
es
ar
e
ex
tr
ac
te
d
fr
om
e
ac
h
of
th
e
tw
el
ve
mo
de
ls
pr
e
-
trained
o
n
the
benchmark
dataset.
Step 5: Adjusting and optimizing the associated function to train pre
-
trained models.
Step 6: Fully trained models are generated.
2
.
3
.
T
he
inp
ut
da
t
a
s
et
T
h
is
s
tu
d
y
u
s
es
twelv
e
p
r
e
-
tr
a
in
ed
m
o
d
els
as
b
ac
k
b
o
n
e
e
n
c
o
d
er
s
in
an
e
x
p
er
im
e
n
tal
s
tu
d
y
u
s
in
g
th
e
MI
C
C
AI
B
r
aT
S
2
0
1
9
[
1
7
]
ch
allen
g
e
d
ataset
with
3
3
5
ca
s
e
s
.
T
h
e
d
ataset
in
clu
d
es
n
ativ
e
,
p
o
s
t
-
co
n
tr
ast,
T
2
-
weig
h
ted
,
a
n
d
T
2
-
FLAI
R
s
c
an
v
o
lu
m
es.
T
h
e
s
tu
d
y
id
en
tifie
s
n
ec
r
o
tic
a
n
d
n
o
n
-
en
h
a
n
cin
g
t
u
m
o
r
co
r
e
NC
R
/NE
T
(
lab
el
1
)
,
p
er
itu
m
o
r
al
ed
em
a
E
D,
a
n
d
e
n
h
an
ci
n
g
tu
m
o
r
E
T
lab
els,
tr
a
n
s
f
o
r
m
in
g
th
em
in
to
3
-
ch
an
n
el
v
o
lu
m
es.
E
ac
h
im
ag
in
g
d
ataset
s
eg
m
e
n
ts
th
r
ee
la
b
els.
T
h
e
illu
s
tr
atio
n
in
Fig
u
r
e
3
d
is
p
lay
s
f
o
u
r
d
if
f
er
en
t m
o
d
alities
an
d
th
e
m
ask
.
Fig
u
r
e
3
.
B
r
aT
S 2
0
1
9
m
o
d
alit
ies an
d
g
en
er
ate
d
m
ask
2
.
4
.
T
ra
ini
ng
a
nd
ex
perim
e
nta
l c
o
nfig
ura
t
io
n
Mo
d
el
tr
ain
in
g
was
co
n
d
u
cte
d
u
s
in
g
a
n
Nv
id
ia
T
esla
P1
0
0
GPU
with
1
6
GB
GDDR6
V
R
AM
.
T
h
e
f
r
am
ewo
r
k
u
s
ed
in
th
e
ex
p
er
i
m
en
ts
is
T
en
s
o
r
Flo
w.
T
h
e
d
a
taset
'
s
d
is
tr
ib
u
tio
n
is
8
0
%
ch
o
s
en
at
r
an
d
o
m
f
o
r
tr
ain
in
g
(
7
0
%
f
o
r
tr
ain
i
n
g
a
n
d
1
0
%
f
o
r
v
alid
atio
n
)
,
with
t
h
e
r
em
ain
in
g
2
0
%
g
o
i
n
g
to
w
ar
d
test
in
g
.
W
h
ile
tr
ain
in
g
,
th
e
o
p
tim
izer
Ad
a
m
is
u
s
ed
with
a
lear
n
in
g
r
ate
v
alu
e
s
et
at
0
.
0
0
0
1
[
1
8
]
an
d
R
elu
f
o
r
ac
tiv
a
tio
n
f
u
n
ctio
n
with
5
0
ep
o
c
h
s
.
T
h
e
lo
s
s
f
u
n
ctio
n
th
at
is
u
tili
ze
d
is
b
in
ar
y
_
c
r
o
s
s
-
en
tr
o
p
y
.
T
h
e
d
ef
in
itio
n
o
f
b
i
n
ar
y
-
cr
o
s
s
-
en
tr
o
p
y
[
1
9
]
is
s
h
o
wn
in
(
1
)
.
=
−
1
∑
{
=
1
+
(
1
-
)
(
1
−
)
}
(
1
)
2
.
5
.
E
v
a
lua
t
i
o
n m
et
rics
Fo
r
m
o
d
el
ev
alu
atio
n
,
b
y
d
i
v
id
in
g
th
e
to
tal
n
u
m
b
er
o
f
p
r
e
d
ictio
n
s
b
y
th
e
s
u
m
o
f
ac
cu
r
at
e
p
o
s
itiv
e
an
d
n
e
g
ativ
e
p
r
ed
ictio
n
s
,
th
e
ac
cu
r
a
cy
s
co
r
e
is
d
eter
m
in
ed
as
in
(
2
)
.
An
o
th
er
m
etr
ics
ar
e
th
e
Dice
co
ef
f
icien
t
an
d
th
e
J
ac
ca
r
d
co
ef
f
icien
t,
wh
ich
ca
n
b
e
ca
lcu
lated
ac
co
r
d
i
n
g
to
(
3
)
an
d
(
4
)
.
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.
40
,
No
.
2
,
No
v
em
b
er
20
25
:
8
5
0
-
85
9
854
=
+
+
+
+
(
2)
=
2
2
+
+
(
3
)
=
+
+
(
4
)
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
s
u
g
g
ested
s
tu
d
y
u
s
ed
p
r
e
-
tr
ain
ed
d
ee
p
lear
n
in
g
ar
ch
itectu
r
es a
s
an
en
co
d
er
b
ac
k
b
o
n
e
f
o
r
Un
et
to
s
eg
m
en
t
b
r
ai
n
tu
m
o
r
s
f
r
o
m
m
u
ltimo
d
al
MRI
im
ag
es
with
I
m
ag
eNe
t
weig
h
ts
.
He
r
e,
1
2
p
r
e
-
tr
ain
e
d
ar
ch
itectu
r
es
ar
e
ap
p
lie
d
as
en
co
d
er
s
f
o
r
Un
et:
R
esN
et3
4
,
R
esNet5
0
,
R
esNet1
0
1
,
VGG1
6
,
VGG1
9
,
Den
s
eNe
t1
2
1
,
I
n
ce
p
tio
n
R
esN
etV2
,
I
n
ce
p
tio
n
V3
,
Mo
b
ilen
etV2
,
an
d
E
f
f
icien
t
Net
B
1
,
u
tili
zin
g
th
e
B
r
aT
S
2
0
1
9
b
r
ain
t
u
m
o
r
s
eg
m
en
tatio
n
m
u
ltim
o
d
al
d
ataset.
I
n
ad
d
itio
n
,
a
n
e
v
alu
atio
n
o
f
th
e
two
n
o
v
el
a
r
ch
itectu
r
es,
SERes
Ne
t
1
8
an
d
SERes
Net
3
4
,
was
also
co
n
d
u
cted
.
T
h
e
r
esu
lts
o
f
tr
ain
in
g
an
d
v
alid
ati
o
n
ac
cu
r
ac
y
f
o
r
p
r
e
-
tr
ain
ed
en
c
o
d
er
s
ar
e
s
h
o
wn
i
n
Fig
u
r
e
4
.
Fig
u
r
e
5
e
v
alu
ates
a
n
d
co
n
tr
asts
th
e
p
r
e
-
t
r
ain
ed
m
o
d
els'
p
er
f
o
r
m
an
ce
in
ter
m
s
o
f
th
e
Dice
s
im
ilar
ity
co
ef
f
icien
t.
T
a
b
le
2
c
o
m
p
ar
es
p
r
e
-
tr
ain
e
d
en
co
d
er
s
f
o
r
U
-
n
e
t
m
o
d
els
r
eg
ar
d
in
g
tr
ain
in
g
tim
e,
tr
ain
a
b
le
p
ar
am
eter
s
,
an
d
n
o
n
-
tr
ain
a
b
le
p
ar
a
m
eter
s
.
T
h
e
test
lo
s
s
an
d
ac
c
u
r
ac
y
g
r
a
p
h
s
f
o
r
1
2
d
if
f
er
en
t m
o
d
els u
s
in
g
p
r
e
-
t
r
ain
ed
en
co
d
er
s
ar
e
co
m
p
a
r
ed
in
Fig
u
r
e
6
.
T
ab
le
3
c
o
m
p
ar
e
s
th
e
J
ac
ca
r
d
an
d
Dice
co
ef
f
icien
ts
f
o
r
s
eg
m
en
tin
g
b
r
ain
tu
m
o
r
s
with
U
-
n
et
with
p
r
ev
io
u
s
wo
r
k
.
3
.
1
.
Resul
t
s
a
na
ly
s
is
Acc
o
r
d
in
g
to
th
e
r
esu
lts
in
Mo
b
ilen
etV2
(
Fig
u
r
e
4
(
l)
)
w
as
th
e
m
o
s
t
ac
cu
r
ate
an
d
lea
s
t
lo
s
s
p
r
e
-
tr
ain
ed
m
o
d
el
u
s
ed
to
s
eg
m
e
n
t
b
r
ain
tu
m
o
r
s
f
r
o
m
th
e
m
u
l
tim
o
d
al
d
ataset
(
Fig
u
r
es
4
(
a)
-
4
(
c)
)
.
Fo
llo
wed
b
y
VGG1
6
an
d
VGG1
9
as
s
h
o
wn
in
Fig
u
r
e
s
4
(
d
)
an
d
4
(
e)
.
E
f
f
icien
t
Net
B
1
ca
m
e
in
th
i
r
d
p
lace
,
as
s
h
o
wn
i
n
Fig
u
r
e
4
(
k
)
,
f
o
llo
wed
b
y
Den
s
e1
2
1
n
etwo
r
k
e
n
co
d
er
s
as in
F
ig
u
r
e
4
(
h
)
.
I
n
ce
p
tio
n
R
esNetV
2
an
d
I
n
ce
p
tio
n
V3
(
Fig
u
r
e
4
(
f
)
)
wer
e
also
co
n
s
i
d
er
ed
,
b
u
t
I
n
ce
p
ti
o
n
V3
(
Fig
u
r
e
4
(
f
)
)
was
s
u
p
er
io
r
in
lo
s
s
.
R
esNet3
4
(
Fig
u
r
e
4
(
g
)
)
was
co
n
s
id
er
ed
th
e
b
est
o
f
th
e
th
r
ee
R
esNet
m
o
d
els
(
R
esNet3
4
,
R
esNe
t5
0
,
an
d
R
esNet1
0
1
)
,
wh
ile
SERes
Ne
t 1
8
(F
ig
u
r
e
4
(
i)
)
an
d
SERes
Ne
t 3
4
(
Fig
u
r
e
4
(
j)
)
h
ad
th
e
wo
r
s
t r
esu
lts
.
Fig
u
r
e
4
.
T
r
ain
in
g
a
n
d
v
alid
atio
n
ac
cu
r
ac
y
:
(
a)
R
esNet3
4
,
(
b
)
R
esNet
-
5
0
,
(
c)
R
esNet
-
1
0
1
,
(
d
)
VGG1
6
,
(
e)
VGG1
9
,
(
f
)
I
n
ce
p
tio
n
V3
,
(
g
)
I
n
ce
p
tio
n
R
esNet,
(
h
)
Den
s
e1
2
1
,
(
i)
SERes
Net
1
8
,
(
j)
SE
R
esNet
3
4
,
(
k
)
E
f
f
ic
ien
t N
et
B
1
an
d
(
l
)
M
o
b
ilen
etV2
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
A
s
ystema
tic
ev
a
lu
a
tio
n
o
f p
r
e
-
tr
a
in
ed
en
co
d
er a
r
ch
itectu
r
es fo
r
mu
ltimo
d
a
l b
r
a
in
tu
mo
r
…
(
Ma
r
w
a
A
b
b
a
s
)
855
C
o
n
ce
r
n
in
g
Fig
u
r
e
s
5
(
a
)
-
5
(
l
)
,
th
e
U
-
Net
s
eg
m
en
tatio
n
m
o
d
el
u
tili
zin
g
Mo
b
ileNetV2
as
i
ts
en
co
d
er
b
ac
k
b
o
n
e
h
a
d
th
e
h
i
g
h
est
d
ic
e
co
ef
f
icien
t
co
m
p
ar
e
d
to
o
t
h
er
U
-
Net
m
o
d
els,
as
ex
h
i
b
ited
in
Fig
u
r
e
5
-
l.
T
h
e
wo
r
s
t
ca
s
e
is
th
e
I
n
ce
p
tio
n
-
R
esn
et
-
Un
et
m
o
d
el.
Fu
r
th
er
m
o
r
e,
T
ab
le
2
s
h
o
ws
th
at
th
e
Mo
b
ilen
etV2
-
U
n
et
ac
h
iev
es
th
e
m
i
n
im
u
m
tr
ain
i
n
g
tim
e
a
m
o
n
g
all
m
o
d
els
with
p
r
e
-
tr
ain
e
d
e
n
co
d
er
s
,
a
n
d
h
as
th
e
f
ewe
s
t
p
ar
am
eter
s
r
ep
r
esen
ted
in
t
r
ain
ab
le
an
d
n
o
n
-
tr
ain
ab
le
p
ar
am
eter
s
.
So
,
Mo
b
ileNet
V2
o
f
f
er
s
h
i
g
h
er
p
er
f
o
r
m
an
ce
with
f
ewe
r
p
ar
a
m
eter
s
an
d
less
co
m
p
u
tatio
n
al
ex
p
en
s
e
th
an
its
p
r
ed
ec
ess
o
r
.
Fig
u
r
e
5
.
T
r
ain
in
g
a
n
d
Valid
at
io
n
Di
ce
co
ef
f
icien
t
:
(
a)
R
esNet
-
3
4
,
(
b
)
R
esNet
-
5
0
,
(
c)
R
esNet
-
1
0
1
,
(
d
)
VGG
-
1
6
,
(
e)
VGG
-
1
9
,
(
f
)
I
n
ce
p
tio
n
V3
,
(
g
)
I
n
ce
p
tio
n
R
e
s
Net,
(
h
)
Den
s
e1
2
1
,
(
i)
SERes
Net
1
8
,
(
j)
SERes
Net
3
4
,
(
k
)
E
f
f
ic
ien
t
Net
B
1
an
d
(
l)
Mo
b
ilen
etV2
T
ab
le
2
.
C
o
m
p
a
r
is
o
n
o
f
p
r
e
-
t
r
ain
ed
en
co
d
er
s
f
o
r
U
-
n
et
f
r
am
ewo
r
k
s
f
o
r
non
-
tr
ain
ab
le
p
ar
a
m
eter
s
,
tr
ain
ab
le
p
ar
am
eter
s
,
an
d
tr
ain
in
g
tim
e
M
o
d
e
l
s
t
r
u
c
t
u
r
e
Tr
a
i
n
i
n
g
t
i
me
(
)
To
t
a
l
p
a
r
a
me
t
e
r
s
(
mi
l
l
i
o
n
s)
Tr
a
i
n
a
b
l
e
p
a
r
a
m
e
t
e
r
s
N
o
n
-
t
r
a
i
n
a
b
l
e
p
a
r
a
m
e
t
e
r
s
R
e
sN
e
t
3
4
-
U
n
e
t
3
1
9
9
.
6
7
6
2
4
,
4
5
6
,
5
8
9
3
,
1
6
7
,
4
9
5
2
1
,
2
8
9
,
0
9
4
R
e
sN
e
t
5
0
-
U
n
e
t
3
7
5
6
.
3
6
1
3
2
,
5
6
1
,
5
4
9
9
,
0
5
9
,
0
7
9
2
3
,
5
0
2
,
4
7
0
R
e
sN
e
t
1
0
1
-
U
n
e
t
4
5
6
2
.
6
5
9
5
1
,
6
0
5
,
9
0
1
9
,
1
1
1
,
3
0
3
4
2
,
4
9
4
,
5
9
8
VGG
-
16
-
U
n
e
t
3
2
6
1
.
1
4
4
2
3
,
7
5
2
,
7
0
8
9
,
0
3
3
,
9
8
8
1
4
,
7
1
8
,
7
2
0
VGG
-
19
-
U
n
e
t
3
3
2
5
.
9
8
1
2
9
,
0
6
2
,
4
0
4
9
,
0
3
3
,
9
8
8
2
0
,
0
2
8
,
4
1
6
I
n
c
e
p
t
i
o
n
V
3
-
U
n
e
t
3
2
8
7
.
7
8
6
2
9
,
9
3
3
,
5
4
0
8
,
1
4
5
,
9
8
8
2
1
,
7
8
7
,
5
5
2
I
n
c
e
p
t
i
o
n
-
R
e
s
n
e
t
V
2
-
U
n
e
t
4
6
5
2
.
0
4
5
6
2
,
0
6
1
,
9
8
8
7
,
7
5
3
,
5
4
0
5
4
,
3
0
8
,
4
4
8
D
e
n
seN
e
t
1
2
1
-
U
n
e
t
3
4
1
6
.
7
1
9
1
2
,
1
4
5
,
4
1
2
5
,
1
8
9
,
5
7
2
6
,
9
5
5
,
8
4
0
S
E
R
e
sN
e
t
1
8
-
U
n
e
t
3
4
0
0
.
8
5
8
1
4
,
4
3
0
,
0
8
5
3
,
1
6
0
,
0
7
1
1
1
,
2
7
0
,
0
1
4
S
E
R
e
sN
e
t
3
4
-
U
n
e
t
3
3
5
6
.
3
1
5
2
4
,
6
1
7
,
7
8
5
3
,
1
6
7
,
4
9
5
2
1
,
4
5
0
,
2
9
0
Ef
f
i
c
i
e
n
t
N
e
t
B
2
-
U
n
e
t
3
7
6
5
.
5
4
2
1
2
,
6
4
1
,
6
0
4
6
,
1
2
6
,
4
3
6
6
,
5
1
5
,
1
6
8
M
o
b
i
l
e
n
e
t
V
2
-
U
n
e
t
3
2
2
7
.
6
7
5
8
,
0
4
7
,
8
7
6
5
,
8
2
2
,
0
2
0
2
,
2
2
5
,
8
5
6
As
s
h
o
wn
in
Fig
u
r
e
6
,
Mo
b
il
en
etV2
-
Un
et
r
ea
ch
ed
t
h
e
h
ig
h
est
ac
cu
r
ac
y
o
n
th
e
test
s
et
o
f
9
9
.
2
%,
with
a
lo
s
s
o
f
0
.
0
2
8
2
.
T
h
e
V
GG1
9
m
o
d
el
ac
h
iev
ed
a
test
ac
cu
r
ac
y
o
f
9
9
.
1
9
%,
f
o
llo
we
d
b
y
VGG1
6
with
9
8
.
6
8
%
an
d
0
.
0
3
5
8
a
n
d
0
.
0
4
3
7
test
lo
s
s
es,
r
esp
ec
tiv
ely
.
W
ith
in
t
h
e
R
esNet
ar
ch
itectu
r
e
f
am
ily
,
R
esNet3
4
d
em
o
n
s
tr
ated
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
with
9
8
.
4
6
%
test
ac
cu
r
ac
y
an
d
0
.
0
6
5
8
test
lo
s
s
co
m
p
ar
ed
t
o
R
esNet5
0
an
d
R
esNet1
0
1
.
I
n
ad
d
itio
n
,
w
e
p
r
esen
ted
an
e
v
alu
atio
n
o
f
t
h
e
n
ew
a
r
ch
itectu
r
es
r
e
p
r
esen
t
ed
b
y
SE
R
esNet3
4
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.
40
,
No
.
2
,
No
v
em
b
er
20
25
:
8
5
0
-
85
9
856
an
d
SE
R
esNet1
8
p
r
e
-
tr
ain
ed
en
co
d
er
s
,
a
n
d
t
h
ey
ac
h
iev
ed
t
h
e
lo
west
test
ac
cu
r
ac
y
with
9
7
.
2
6
%
an
d
9
5
.
5
3
%.
r
esp
ec
tiv
ely
.
An
d
t
h
e
h
ig
h
est er
r
o
r
r
ate
am
o
n
g
th
e
1
2
m
o
d
el
s
.
T
ab
le
3
p
r
o
v
id
es
a
co
m
p
ar
ativ
e
an
aly
s
is
o
f
th
e
p
r
ev
io
u
s
p
r
e
-
tr
ain
ed
en
co
d
er
s
alo
n
g
s
id
e
th
e
Un
et
an
d
Mo
b
ileNetV2
en
co
d
er
s
a
n
d
t
h
e
two
n
o
v
el
ar
ch
itectu
r
es,
S
E
R
esNet3
4
an
d
SE
R
esNet1
8
,
p
r
o
p
o
s
ed
in
th
is
s
tu
d
y
.
T
h
e
co
m
p
ar
is
o
n
is
f
o
c
u
s
ed
o
n
th
e
ass
ess
m
en
t
o
f
ea
c
h
m
o
d
el’
s
r
esu
lts
b
ased
o
n
m
u
ltip
le
ev
alu
atio
n
m
etr
ics,
e.
g
.
,
J
ac
ca
r
d
,
d
ice
co
ef
f
icien
t
f
o
r
en
h
a
n
cin
g
tu
m
o
r
(
E
T
)
,
tu
m
o
r
c
o
r
e
(
T
C
)
,
a
n
d
wh
o
le
tu
m
o
r
(
W
T
)
.
Mo
b
ilen
etV2
-
Un
et
h
as
th
e
b
e
s
t
d
ice
co
ef
f
icie
n
t
f
o
r
(
E
T
)
o
f
8
1
.
9
3
%,
8
6
.
0
7
%
(
T
C
)
,
an
d
9
0
.
0
3
%
(
W
T
)
a
n
d
th
e
h
ig
h
est
J
ac
ca
r
d
v
alu
e
o
f
8
1
.
6
3
%.
Acc
o
r
d
in
g
to
th
e
r
es
u
lts
in
T
ab
le
3
,
Mo
b
ileNet
V2
as
an
en
c
o
d
er
d
em
o
n
s
tr
ates
a
f
av
o
r
ab
le
tr
a
d
e
-
o
f
f
b
etwe
en
ac
cu
r
ac
y
an
d
co
m
p
u
tatio
n
al
ef
f
icien
c
y
.
Fig
u
r
e
7
g
r
a
p
h
ically
r
ep
r
esen
ts
th
e
to
p
f
o
u
r
p
er
f
o
r
m
in
g
m
o
d
els
o
u
t
o
f
th
e
twelv
e
ev
alu
ated
,
b
ased
o
n
t
h
e
h
ig
h
est
o
b
tain
ed
v
al
u
es
f
o
r
ac
cu
r
ac
y
,
Dice
s
im
ilar
ity
c
o
ef
f
icien
t,
an
d
J
ac
ca
r
d
in
d
ex
.
Fig
u
r
e
6
.
C
o
m
p
a
r
is
o
n
g
r
ap
h
o
f
test
lo
s
s
an
d
ac
cu
r
a
cy
f
o
r
1
2
d
if
f
er
en
t m
o
d
els with
p
r
e
-
tr
ain
ed
en
c
o
d
er
s
Fig
u
r
e
7
.
T
h
e
to
p
f
o
u
r
m
o
d
els with
p
r
e
-
tr
ain
ed
en
co
d
e
r
s
ac
h
i
ev
ed
v
is
u
al
s
eg
m
en
tatio
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
A
s
ystema
tic
ev
a
lu
a
tio
n
o
f p
r
e
-
tr
a
in
ed
en
co
d
er a
r
ch
itectu
r
es fo
r
mu
ltimo
d
a
l b
r
a
in
tu
mo
r
…
(
Ma
r
w
a
A
b
b
a
s
)
857
T
ab
le
3
.
Pre
v
i
o
u
s
wo
r
k
i
n
co
m
p
ar
is
o
n
with
th
e
M
o
b
ilen
et
V2
en
co
d
e
r
with
U
-
n
et
m
o
d
el
s
in
ter
m
s
o
f
J
ac
ca
r
d
an
d
Dice
co
ef
f
icien
ts
f
o
r
E
T
,
T
C
,
an
d
W
T
M
o
d
e
l
Jac
c
a
r
d
(
%)
ET
(
%)
TC
(
%)
W
T
(
%)
R
e
sN
e
t
3
4
-
U
n
e
t
[
2
0
]
6
3
.
5
6
7
3
.
9
6
7
8
.
6
6
8
2
.
3
3
R
e
sN
e
t
5
0
-
U
n
e
t
[
2
1
]
6
3
.
4
2
7
3
.
8
8
7
8
.
4
5
8
2
.
0
2
R
e
sN
e
t
1
0
1
-
U
n
e
t
[
2
2
]
6
1
.
0
8
7
2
.
3
2
7
7
.
3
0
8
1
.
1
4
VGG
-
16
-
U
n
e
t
[
2
3
]
7
2
.
2
3
7
7
.
5
7
8
0
.
5
2
8
4
.
6
1
VGG
-
1
9
U
n
e
t
[
2
4
]
7
4
.
1
3
7
7
.
7
5
8
0
.
6
3
8
5
.
4
1
I
n
c
e
p
t
i
o
n
V
3
-
U
n
e
t
[
1
2
]
7
0
.
1
8
7
5
.
7
7
7
9
.
0
9
8
3
.
4
8
I
n
c
e
p
t
i
o
n
-
R
e
s
n
e
t
V
2
-
U
n
e
t
[
2
5
]
6
4
.
5
8
7
4
.
8
5
7
8
.
4
5
8
2
.
4
5
D
e
n
seN
e
t
1
2
1
-
U
n
e
t
[
2
6
]
6
2
.
0
9
7
5
.
8
8
7
9
.
2
6
8
3
.
6
1
S
E
R
e
sN
e
t
1
8
-
U
n
e
t
6
0
.
6
1
7
2
.
0
6
7
6
.
4
9
8
1
.
0
0
S
E
R
e
sN
e
t
3
4
-
U
n
e
t
6
2
.
5
0
7
2
.
5
7
7
7
.
6
4
8
2
.
0
2
Ef
f
i
c
i
e
n
t
N
e
t
B
1
-
U
n
e
t
[
1
1
]
8
1
.
6
3
7
9
.
2
3
8
3
.
5
7
8
8
.
6
5
M
o
b
i
l
e
n
e
t
V
2
-
U
n
e
t
(
t
h
i
s
w
o
r
k
)
8
3
.
4
5
8
1
.
9
3
8
6
.
0
7
9
0
.
3
4.
CO
NCLU
SI
O
N
T
h
is
ar
ticle
in
v
esti
g
ated
th
e
ef
f
icac
y
o
f
em
p
lo
y
in
g
p
r
e
-
tr
ain
ed
tr
an
s
f
er
lear
n
in
g
m
o
d
els
as
en
co
d
er
s
with
in
th
e
U
-
Net
d
esig
n
f
o
r
s
eg
m
en
tin
g
b
r
ai
n
tu
m
o
r
s
.
W
e
ev
alu
ated
twelv
e
p
r
e
-
tr
ain
ed
e
n
co
d
er
s
,
in
itialized
with
I
m
ag
eNe
t
weig
h
ts
,
u
s
in
g
th
e
B
r
aT
S
2
0
1
9
d
ataset
u
n
d
er
id
en
tical
h
y
p
er
p
ar
am
eter
co
n
f
ig
u
r
atio
n
s
.
T
h
is
s
tu
d
y
ex
ten
d
s
th
e
u
n
d
er
s
tan
d
in
g
o
f
ad
v
an
ce
d
b
ac
k
b
o
n
e
n
etwo
r
k
s
f
o
r
s
em
an
tic
s
eg
m
e
n
tatio
n
,
s
p
ec
if
ically
ad
d
r
ess
in
g
ch
allen
g
es
r
elate
d
to
in
ter
p
r
eta
b
ilit
y
,
co
m
p
u
tat
io
n
al
d
em
a
n
d
s
,
an
d
o
v
e
r
f
itti
n
g
.
Ou
r
c
o
m
p
ar
ativ
e
an
aly
s
is
,
f
o
cu
s
in
g
o
n
m
etr
ics s
u
ch
as
Dice
co
ef
f
icien
t,
test
l
o
s
s
,
ac
cu
r
ac
y
,
an
d
tr
ain
in
g
d
u
r
atio
n
,
r
ev
ea
led
th
at
th
e
h
y
b
r
id
U
-
Net
m
o
d
el
em
p
l
o
y
in
g
M
o
b
ileNetV2
in
th
e
e
n
c
o
d
in
g
p
ath
ex
h
ib
its
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
.
T
h
e
Mo
b
ileNetV2
-
en
h
a
n
ce
d
U
-
Net
r
ea
ch
ed
a
test
ac
cu
r
ac
y
o
f
9
9
.
3
%,
a
cr
o
s
s
-
v
alid
atio
n
ac
cu
r
ac
y
o
f
9
9
%,
a
J
ac
ca
r
d
co
ef
f
icien
t
o
f
8
3
.
4
5
%,
an
d
a
test
lo
s
s
o
f
0
.
0
2
8
2
.
Fu
r
th
er
m
o
r
e,
it
d
em
o
n
s
tr
ated
p
o
wer
f
u
l
s
eg
m
en
tatio
n
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
tu
m
o
r
s
u
b
-
r
e
g
io
n
s
,
with
D
ice
co
ef
f
icien
ts
o
f
8
6
.
0
7
%
f
o
r
th
e
T
C
,
9
0
.
3
%
f
o
r
th
e
W
T
,
an
d
8
1
.
9
3
%
f
o
r
t
h
e
E
T
.
I
m
p
o
r
tan
tl
y
,
th
e
Mo
b
ileNetV2
e
n
co
d
e
r
'
s
lig
h
tweig
h
t
n
atu
r
e
tr
an
s
lates
to
r
ed
u
ce
d
c
o
m
p
u
tatio
n
al
r
eso
u
r
ce
co
n
s
u
m
p
tio
n
.
Acc
o
r
d
in
g
to
t
h
ese
r
esu
lts
,
in
co
r
p
o
r
atin
g
Mo
b
ileNetV2
as
an
en
co
d
e
r
in
to
th
e
U
-
Net
ar
ch
itectu
r
e
p
r
o
v
id
es
a
s
tr
o
n
g
ar
g
u
m
en
t
f
o
r
p
r
ec
is
e
an
d
ef
f
ec
tiv
e
b
r
ain
tu
m
o
r
s
eg
m
en
tat
io
n
.
T
h
is
h
as
s
ig
n
if
ican
t
im
p
licatio
n
s
f
o
r
clin
ical
p
r
ac
tice,
p
o
ten
tially
en
ab
lin
g
f
aste
r
an
d
m
o
r
e
p
r
ec
is
e
d
iag
n
o
s
es,
wh
ich
ca
n
lead
to
en
h
an
ce
d
clin
ical
o
u
tco
m
es
an
d
in
f
o
r
m
ed
tr
ea
tm
en
t
d
ec
i
s
io
n
-
m
ak
in
g
.
T
h
e
m
ain
lim
itatio
n
o
f
p
r
e
-
tr
ain
e
d
m
o
d
els
ar
e
o
f
te
n
c
o
m
p
lex
an
d
lar
g
e,
r
eq
u
ir
in
g
s
ig
n
if
ican
t
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
f
o
r
f
in
e
-
tu
n
in
g
,
wh
ich
ca
n
b
e
ch
allen
g
i
n
g
f
o
r
r
esear
ch
er
s
o
r
p
r
ac
titi
o
n
er
s
with
lim
ited
ac
ce
s
s
to
ad
v
an
ce
d
co
m
p
u
tin
g
r
eso
u
r
ce
s
.
B
u
t
th
e
r
ed
u
ce
d
c
o
m
p
u
tatio
n
al
co
s
t
ass
o
ciate
d
with
Mo
b
ileNetV
2
a
ls
o
m
ak
es
th
is
ap
p
r
o
ac
h
m
o
r
e
ac
ce
s
s
ib
le
f
o
r
r
eso
u
r
ce
-
co
n
s
tr
ain
ed
en
v
ir
o
n
m
en
ts
.
Fu
tu
r
e
r
esear
ch
co
u
ld
i
n
v
e
s
tig
ate
f
u
r
th
er
o
p
tim
izin
g
t
h
e
ar
ch
itectu
r
e,
p
e
r
h
ap
s
th
r
o
u
g
h
th
e
in
co
r
p
o
r
atio
n
o
f
atten
tio
n
m
ec
h
an
is
m
s
o
r
n
o
v
el
lo
s
s
f
u
n
ctio
n
s
,
wh
ich
co
u
l
d
lead
to
ev
en
g
r
ea
te
r
im
p
r
o
v
em
e
n
ts
in
s
eg
m
e
n
tatio
n
p
er
f
o
r
m
a
n
ce
.
Fu
r
th
er
m
o
r
e,
f
u
tu
r
e
wo
r
k
co
u
ld
i
n
v
o
lv
e
t
h
e
ad
o
p
tio
n
o
f
th
e
Mo
b
ileNetV2
-
en
h
an
ce
d
U
-
N
et
f
r
am
ewo
r
k
f
o
r
a
d
d
itio
n
al
m
ed
ical
im
ag
e
s
eg
m
en
tatio
n
r
eq
u
ir
em
e
n
ts
,
lik
e
o
r
g
an
o
r
tu
m
o
r
s
eg
m
en
tatio
n
.
Fu
tu
r
e
r
esear
ch
co
u
ld
also
b
en
ef
it
f
r
o
m
e
x
am
in
in
g
th
e
e
f
f
ec
tiv
en
ess
o
f
t
h
e
m
o
d
el
in
ac
tu
al
tim
e
clin
ical
s
ettin
g
s
an
d
with
lar
g
er
b
r
ain
d
atasets
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
N
T
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
e
s
t.
DATA AV
AI
L
AB
I
L
I
T
Y
Data
av
ailab
ilit
y
d
o
es n
o
t a
p
p
l
y
to
th
is
p
ap
e
r
as n
o
n
ew
d
ata
wer
e
cr
ea
ted
o
r
an
aly
ze
d
in
th
is
s
tu
d
y
.
RE
F
E
R
E
NC
E
S
[
1
]
A
.
H
.
S
a
k
a
e
t
a
l
.
,
“
C
a
n
c
e
r
st
a
t
i
st
i
c
s
f
o
r
A
f
r
i
c
a
n
A
mer
i
c
a
n
a
n
d
b
l
a
c
k
p
e
o
p
l
e
,
2
0
2
5
,
”
C
A:
A
C
a
n
c
e
r
J
o
u
r
n
a
l
f
o
r C
l
i
n
i
c
i
a
n
s
,
v
o
l
.
7
5
,
n
o
.
2
,
p
p
.
1
1
1
–
1
4
0
,
F
e
b
.
2
0
2
5
,
d
o
i
:
1
0
.
3
3
2
2
/
c
a
a
c
.
2
1
8
7
4
.
[
2
]
C
.
T
o
a
d
e
r
e
t
a
l
.
,
“
Lo
w
-
g
r
a
d
e
g
l
i
o
ma
s:
h
i
s
t
o
l
o
g
i
c
a
l
s
u
b
t
y
p
e
s,
mo
l
e
c
u
l
a
r
m
e
c
h
a
n
i
sms
,
a
n
d
t
r
e
a
t
me
n
t
st
r
a
t
e
g
i
e
s,
”
Bra
i
n
S
c
i
e
n
c
e
s
,
v
o
l
.
1
3
,
n
o
.
1
2
,
p
.
1
7
0
0
,
D
e
c
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
b
r
a
i
n
sc
i
1
3
1
2
1
7
0
0
.
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.
40
,
No
.
2
,
No
v
em
b
er
20
25
:
8
5
0
-
85
9
858
[
3
]
C
.
S
t
e
e
n
h
o
u
t
,
L.
D
e
p
r
e
z
,
R
.
H
u
st
i
n
x
,
a
n
d
N
.
W
i
t
h
o
f
s,
“
B
r
a
i
n
t
u
mo
r
a
ss
e
ss
men
t
:
i
n
t
e
g
r
a
t
i
n
g
P
E
T/
c
o
m
p
u
t
e
d
t
o
m
o
g
r
a
p
h
y
a
n
d
M
R
i
ma
g
i
n
g
m
o
d
a
l
i
t
i
e
s,
”
P
ET C
l
i
n
i
c
s
,
v
o
l
.
2
0
,
n
o
.
1
,
p
p
.
1
6
5
–
1
7
4
,
J
a
n
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
p
e
t
.
2
0
2
4
.
0
9
.
0
0
3
.
[
4
]
Y
.
L
i
,
S
.
Ji
a
n
g
,
Z.
Y
a
n
g
,
L.
W
a
n
g
,
Z.
L
i
u
,
a
n
d
Z.
Zh
o
u
,
“
S
e
g
m
e
n
t
a
t
i
o
n
o
f
b
r
a
i
n
t
u
m
o
r
r
e
sec
t
i
o
n
s
i
n
i
n
t
r
a
o
p
e
r
a
t
i
v
e
3
D
u
l
t
r
a
s
o
u
n
d
i
ma
g
e
s
u
si
n
g
a
sem
i
su
p
e
r
v
i
s
e
d
c
r
o
ss
n
n
S
U
-
N
e
t
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
I
m
a
g
i
n
g
S
y
st
e
m
s
a
n
d
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
3
5
,
n
o
.
1
,
J
a
n
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
0
2
/
i
m
a
.
7
0
0
1
8.
[
5
]
S
.
S
a
i
f
u
l
l
a
h
,
R
.
D
r
e
ż
e
w
sk
i
,
A
.
Y
u
d
h
a
n
a
,
M
.
W
i
e
l
g
o
s
z
,
a
n
d
W
.
C
a
e
sare
n
d
r
a
,
“
M
o
d
i
f
i
e
d
U
-
N
e
t
w
i
t
h
a
t
t
e
n
t
i
o
n
g
a
t
e
f
o
r
e
n
h
a
n
c
e
d
a
u
t
o
m
a
t
e
d
b
r
a
i
n
t
u
mo
r
s
e
g
m
e
n
t
a
t
i
o
n
,
”
N
e
u
r
a
l
C
o
m
p
u
t
i
n
g
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
3
7
,
n
o
.
7
,
p
p
.
5
5
2
1
–
5
5
5
8
,
Jan
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
5
2
1
-
0
2
4
-
1
0
9
1
9
-
3.
[
6
]
K
.
V
.
S
h
i
n
y
,
“
B
r
a
i
n
t
u
m
o
r
s
e
g
me
n
t
a
t
i
o
n
a
n
d
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
s
i
n
g
o
p
t
i
m
i
z
e
d
U
-
N
e
t
,
”
I
m
a
g
i
n
g
S
c
i
e
n
c
e
J
o
u
r
n
a
l
,
v
o
l
.
7
2
,
n
o
.
2
,
p
p
.
2
0
4
–
2
1
9
,
M
a
y
2
0
2
4
,
d
o
i
:
1
0
.
1
0
8
0
/
1
3
6
8
2
1
9
9
.
2
0
2
3
.
2
2
0
0
6
1
4
.
[
7
]
M
.
M
.
Ta
y
e
,
“
T
h
e
o
r
e
t
i
c
a
l
u
n
d
e
r
st
a
n
d
i
n
g
o
f
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
:
c
o
n
c
e
p
t
s
,
a
r
c
h
i
t
e
c
t
u
r
e
s
,
a
p
p
l
i
c
a
t
i
o
n
s,
f
u
t
u
r
e
d
i
r
e
c
t
i
o
n
s
,
”
C
o
m
p
u
t
a
t
i
o
n
,
v
o
l
.
1
1
,
n
o
.
3
,
p
.
5
2
,
M
a
r
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
c
o
m
p
u
t
a
t
i
o
n
1
1
0
3
0
0
5
2
.
[
8
]
A
.
H
o
s
n
a
,
E
.
M
e
r
r
y
,
J
.
G
y
a
l
mo
,
Z
.
A
l
o
m
,
Z
.
A
u
n
g
,
a
n
d
M
.
A
.
A
z
i
m,
“
Tr
a
n
sf
e
r
l
e
a
r
n
i
n
g
:
a
f
r
i
e
n
d
l
y
i
n
t
r
o
d
u
c
t
i
o
n
,
”
J
o
u
r
n
a
l
o
f
B
i
g
D
a
t
a
,
v
o
l
.
9
,
n
o
.
1
,
O
c
t
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
8
6
/
s
4
0
5
3
7
-
0
2
2
-
0
0
6
5
2
-
w.
[
9
]
A
.
N
a
w
a
z
,
U
.
A
k
r
a
m,
A
.
A
.
S
a
l
a
m,
A
.
R
.
A
l
i
,
A
.
U
r
R
e
h
ma
n
,
a
n
d
J.
Ze
b
,
“
V
G
G
-
U
N
ET
f
o
r
b
r
a
i
n
t
u
m
o
r
seg
m
e
n
t
a
t
i
o
n
a
n
d
e
n
s
e
mb
l
e
m
o
d
e
l
f
o
r
su
r
v
i
v
a
l
p
r
e
d
i
c
t
i
o
n
,
”
i
n
2
0
2
1
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
R
o
b
o
t
i
c
s
a
n
d
A
u
t
o
m
a
t
i
o
n
i
n
I
n
d
u
s
t
ry,
I
C
R
AI
2
0
2
1
,
O
c
t
.
2
0
2
1
,
p
p
.
1
–
6
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
R
A
I
5
4
0
1
8
.
2
0
2
1
.
9
6
5
1
3
6
7
.
[
1
0
]
S
.
G
h
o
s
h
,
A
.
C
h
a
k
i
,
a
n
d
K
.
S
a
n
t
o
sh
,
“
I
mp
r
o
v
e
d
U
-
N
e
t
a
r
c
h
i
t
e
c
t
u
r
e
w
i
t
h
V
G
G
-
1
6
f
o
r
b
r
a
i
n
t
u
mo
r
se
g
me
n
t
a
t
i
o
n
,
”
P
h
y
s
i
c
a
l
a
n
d
En
g
i
n
e
e
ri
n
g
S
c
i
e
n
c
e
s
i
n
Me
d
i
c
i
n
e
,
v
o
l
.
4
4
,
n
o
.
3
,
p
p
.
7
0
3
–
7
1
2
,
M
a
y
2
0
2
1
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
246
-
021
-
0
1
0
1
9
-
w.
[
1
1
]
S
.
Y
.
Li
n
a
n
d
C
.
L.
Li
n
,
“
B
r
a
i
n
t
u
m
o
r
seg
m
e
n
t
a
t
i
o
n
u
s
i
n
g
U
-
N
e
t
i
n
c
o
n
j
u
n
c
t
i
o
n
w
i
t
h
Ef
f
i
c
i
e
n
t
N
e
t
,
”
Pe
e
r
J
C
o
m
p
u
t
e
r S
c
i
e
n
c
e
,
v
o
l
.
1
0
,
p
.
e
1
7
5
4
,
J
a
n
.
2
0
2
4
,
d
o
i
:
1
0
.
7
7
1
7
/
p
e
e
r
j
-
c
s.
1
7
5
4
.
[
1
2
]
S
.
F
.
R
a
b
b
y
,
M
.
A
.
A
r
a
f
a
t
,
a
n
d
T
.
H
a
sa
n
,
“
B
T
-
N
e
t
:
A
n
e
n
d
-
to
-
e
n
d
mu
l
t
i
-
t
a
s
k
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
b
r
a
i
n
t
u
mo
r
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
seg
m
e
n
t
a
t
i
o
n
,
a
n
d
l
o
c
a
l
i
z
a
t
i
o
n
f
r
o
m
M
R
I
i
ma
g
e
s,”
Arr
a
y
,
v
o
l
.
2
2
,
p
.
1
0
0
3
4
6
,
J
u
l
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
r
r
a
y
.
2
0
2
4
.
1
0
0
3
4
6
.
[
1
3
]
S
.
S
a
i
f
u
l
l
a
h
a
n
d
R
.
D
r
e
ż
e
w
sk
i
,
“
B
r
a
i
n
t
u
mo
r
se
g
me
n
t
a
t
i
o
n
u
s
i
n
g
e
n
s
e
mb
l
e
C
N
N
-
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
mo
d
e
l
s:
D
e
e
p
La
b
V
3
p
l
u
s
a
n
d
R
e
sN
e
t
5
0
a
p
p
r
o
a
c
h
,
”
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
(
i
n
c
l
u
d
i
n
g
s
u
b
s
e
ri
e
s
L
e
c
t
u
re
N
o
t
e
s
i
n
Art
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
re
N
o
t
e
s
i
n
Bi
o
i
n
f
o
rm
a
t
i
c
s)
,
v
o
l
.
1
4
8
3
5
LN
C
S
,
S
p
r
i
n
g
e
r
N
a
t
u
r
e
S
w
i
t
z
e
r
l
a
n
d
,
2
0
2
4
,
p
p
.
3
4
0
–
3
5
4
.
[
1
4
]
R
.
P
r
e
e
t
h
a
,
M
.
Jas
mi
n
e
P
e
me
e
n
a
P
r
i
y
a
d
a
r
s
i
n
i
,
a
n
d
J
.
S
.
N
i
sh
a
,
“
B
r
a
i
n
t
u
m
o
r
seg
m
e
n
t
a
t
i
o
n
u
s
i
n
g
m
u
l
t
i
-
sca
l
e
a
t
t
e
n
t
i
o
n
U
-
N
e
t
w
i
t
h
Ef
f
i
c
i
e
n
t
N
e
t
B
4
e
n
c
o
d
e
r
f
o
r
e
n
h
a
n
c
e
d
M
R
I
a
n
a
l
y
s
i
s,
”
S
c
i
e
n
t
i
f
i
c
Re
p
o
rt
s
,
v
o
l
.
1
5
,
n
o
.
1
,
M
a
r
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
3
8
/
s
4
1
5
9
8
-
0
2
5
-
9
4
2
6
7
-
9.
[
1
5
]
H
.
R
e
z
a
,
N
.
I
.
Ta
r
e
q
,
M
.
M
.
F
.
R
a
b
b
i
,
S
.
A
.
Ta
n
i
m
,
R
.
A
.
M
.
R
u
d
r
o
,
a
n
d
K
.
N
u
r
,
“
A
c
u
t
e
l
y
m
p
h
o
b
l
a
st
i
c
l
e
u
k
e
mi
a
d
i
a
g
n
o
si
s
a
n
d
su
b
t
y
p
e
s
e
g
me
n
t
a
t
i
o
n
i
n
b
l
o
o
d
s
mea
r
s
u
s
i
n
g
C
N
N
a
n
d
U
-
N
e
t
,
”
I
n
d
o
n
e
s
i
a
n
J
o
u
r
n
a
l
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
ri
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
3
8
,
n
o
.
2
,
p
.
9
5
0
,
M
a
y
2
0
2
5
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s.
v
3
8
.
i
2
.
p
p
9
5
0
-
9
5
9
.
[
1
6
]
W
.
W
e
n
g
a
n
d
X
.
Z
h
u
,
“
I
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
m
a
g
e
se
g
me
n
t
a
t
i
o
n
,
”
I
E
EE
Ac
c
e
ss
,
v
o
l
.
9
,
n
o
.
2
3
4
–
2
4
1
,
p
p
.
1
6
5
9
1
–
1
6
6
0
3
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
1
.
3
0
5
3
4
0
8
.
[
1
7
]
“
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
_
2
0
1
9
,
”
Br
a
T
S
2
0
1
9
,
2
0
1
9
.
h
t
t
p
s:
/
/
w
w
w
.
k
a
g
g
l
e
.
c
o
m
/
d
a
t
a
se
t
s/
a
r
y
a
s
h
a
h
2
k
/
b
r
a
i
n
-
t
u
mo
r
-
seg
m
e
n
t
a
t
i
o
n
-
b
r
a
t
s
-
2
0
1
9
.
[
1
8
]
Z.
Zh
a
n
g
,
“
I
mp
r
o
v
e
d
a
d
a
m
o
p
t
i
m
i
z
e
r
f
o
r
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s,”
i
n
2
0
1
8
I
EEE/
AC
M
2
6
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
S
y
m
p
o
s
i
u
m
o
n
Q
u
a
l
i
t
y
o
f
S
e
r
v
i
c
e
,
I
WQo
S
2
0
1
8
,
Ju
n
.
2
0
1
9
,
p
p
.
1
–
2
,
d
o
i
:
1
0
.
1
1
0
9
/
I
W
Q
o
S
.
2
0
1
8
.
8
6
2
4
1
8
3
.
[
1
9
]
.
U
s
h
a
R
u
b
y
D
r
.
A
,
“
B
i
n
a
r
y
c
r
o
s
s
e
n
t
r
o
p
y
w
i
t
h
d
e
e
p
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
f
o
r
I
m
a
g
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
A
d
v
a
n
c
e
d
T
r
e
n
d
s
i
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
9
,
n
o
.
4
,
p
p
.
5
3
9
3
–
5
3
9
7
,
A
u
g
.
2
0
2
0
,
d
o
i
:
1
0
.
3
0
5
3
4
/
i
j
a
t
c
s
e
/
2
0
2
0
/
1
7
5
9
4
2
0
2
0
.
[
2
0
]
J.
L
.
A
r
r
a
s
t
i
a
e
t
a
l
.
,
“
D
e
e
p
l
y
s
u
p
e
r
v
i
sed
u
n
e
t
f
o
r
s
e
ma
n
t
i
c
s
e
g
m
e
n
t
a
t
i
o
n
t
o
a
ssi
s
t
d
e
r
m
a
t
o
p
a
t
h
o
l
o
g
i
c
a
l
a
s
sessm
e
n
t
o
f
b
a
s
a
l
c
e
l
l
c
a
r
c
i
n
o
ma
,
”
J
o
u
r
n
a
l
o
f
I
m
a
g
i
n
g
,
v
o
l
.
7
,
n
o
.
4
,
p
.
7
1
,
A
p
r
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
j
i
m
a
g
i
n
g
7
0
4
0
0
7
1
.
[
2
1
]
S
.
Le
f
k
o
v
i
t
s
a
n
d
L
.
Le
f
k
o
v
i
t
s,
“
U
-
N
e
t
a
r
c
h
i
t
e
c
t
u
r
e
v
a
r
i
a
n
t
s
f
o
r
b
r
a
i
n
t
u
m
o
r
seg
me
n
t
a
t
i
o
n
o
f
h
i
s
t
o
g
r
a
m
c
o
r
r
e
c
t
e
d
i
ma
g
e
s,
”
Ac
t
a
U
n
i
v
e
rsi
t
a
t
i
s
S
a
p
i
e
n
t
i
a
e
,
I
n
f
o
rm
a
t
i
ca
,
v
o
l
.
1
4
,
n
o
.
1
,
p
p
.
4
9
–
7
4
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
2
4
7
8
/
a
u
si
-
2
0
2
2
-
0
0
0
4
.
[
2
2
]
S
.
K
.
H
.
I
sl
a
m,
J
.
P
r
a
d
h
a
n
,
P
.
V
a
s
h
i
s
t
h
a
,
A
.
P
.
S
i
n
g
h
,
a
n
d
A
.
K
i
s
h
o
r
e
,
“
R
e
sN
e
t
-
C
P
D
S
:
c
o
l
o
n
o
s
c
o
p
y
p
o
l
y
p
d
e
t
e
c
t
i
o
n
a
n
d
seg
m
e
n
t
a
t
i
o
n
u
si
n
g
mo
d
i
f
i
e
d
R
e
sN
e
t
1
0
1
V
2
,
”
i
n
L
e
c
t
u
re
N
o
t
e
s
i
n
N
e
t
w
o
r
k
s
a
n
d
S
y
st
e
m
s
,
v
o
l
.
9
6
3
,
S
p
r
i
n
g
e
r
N
a
t
u
r
e
S
i
n
g
a
p
o
r
e
,
2
0
2
4
,
p
p
.
9
3
–
1
0
5
.
[
2
3
]
S
.
K
a
d
r
y
,
E.
V
e
r
d
ú
,
R
.
D
a
mas
e
v
i
c
i
u
s,
L.
A
b
u
a
l
i
g
a
h
,
V
.
S
i
n
g
h
,
a
n
d
V
.
R
a
j
i
n
i
k
a
n
t
h
,
“
C
N
N
se
g
me
n
t
a
t
i
o
n
o
f
s
k
i
n
m
e
l
a
n
o
ma
i
n
p
r
e
-
p
r
o
c
e
sse
d
d
e
r
mo
s
c
o
p
y
i
ma
g
e
s,
”
Pr
o
c
e
d
i
a
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
2
3
5
,
p
p
.
2
7
7
5
–
2
7
8
2
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
r
o
c
s.2
0
2
4
.
0
4
.
2
6
2
.
[
2
4
]
M
.
A
.
K
h
a
n
e
t
a
l
.
,
“
V
G
G
1
9
n
e
t
w
o
r
k
a
ss
i
st
e
d
j
o
i
n
t
s
e
g
m
e
n
t
a
t
i
o
n
a
n
d
c
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
l
u
n
g
n
o
d
u
l
e
s
i
n
C
T
i
ma
g
e
s,”
D
i
a
g
n
o
st
i
c
s
,
v
o
l
.
1
1
,
n
o
.
1
2
,
p
.
2
2
0
8
,
N
o
v
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
d
i
a
g
n
o
s
t
i
c
s1
1
1
2
2
2
0
8
.
[
2
5
]
V
.
R
a
n
g
a
n
n
a
v
a
r
,
S
.
M
u
d
a
v
i
,
P
.
S
h
a
v
i
,
N
.
P
a
t
i
l
,
S
.
I
t
a
g
a
l
l
i
,
a
n
d
N
.
Y
a
l
i
g
a
r
,
“
A
u
t
o
ma
t
e
d
g
a
st
r
o
i
n
t
e
st
i
n
a
l
t
r
a
c
t
i
ma
g
e
s
e
g
m
e
n
t
a
t
i
o
n
,
”
i
n
2
0
2
4
5
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
f
o
r
Em
e
rg
i
n
g
T
e
c
h
n
o
l
o
g
y
,
I
N
C
ET
2
0
2
4
,
M
a
y
2
0
2
4
,
p
p
.
1
–
6
,
d
o
i
:
1
0
.
1
1
0
9
/
I
N
C
ET6
1
5
1
6
.
2
0
2
4
.
1
0
5
9
3
2
8
5
.
[
2
6
]
N
.
C
i
n
a
r
,
A
.
O
z
c
a
n
,
a
n
d
M
.
K
a
y
a
,
“
A
h
y
b
r
i
d
D
e
n
seN
e
t
1
2
1
-
U
N
e
t
mo
d
e
l
f
o
r
b
r
a
i
n
t
u
m
o
r
se
g
me
n
t
a
t
i
o
n
f
r
o
m
M
R
i
ma
g
e
s,
”
Bi
o
m
e
d
i
c
a
l
S
i
g
n
a
l
Pr
o
c
e
s
si
n
g
a
n
d
C
o
n
t
r
o
l
,
v
o
l
.
7
6
,
p
.
1
0
3
6
4
7
,
J
u
l
.
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
6
4
7
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Ma
r
wa
Abb
a
s
re
c
e
iv
e
d
h
e
r
B.
S
c
.
,
d
e
g
re
e
in
Co
m
p
u
ter
En
g
in
e
e
rin
g
fro
m
M
in
ia
Un
i
v
e
rsity
,
E
g
y
p
t
in
2
0
1
2
.
In
2
0
1
3
,
sh
e
j
o
in
e
d
t
h
e
Co
m
p
u
ters
a
n
d
S
y
ste
m
s
En
g
i
n
e
e
rin
g
d
e
p
a
rtme
n
t,
a
t
M
i
n
ia Un
iv
e
rsity
,
E
g
y
p
t
a
s a
tea
c
h
in
g
a
ss
istan
t.
S
h
e
re
c
e
iv
e
d
h
e
r
M
.
S
c
.
in
C
o
m
p
u
ter
E
n
g
i
n
e
e
r
in
g
fro
m
M
in
ia
Un
i
v
e
rsity
,
Eg
y
p
t
in
2
0
2
1
.
He
r
re
se
a
rc
h
in
tere
sts
in
c
l
u
d
e
p
a
ra
ll
e
l
p
ro
g
r
a
m
m
in
g
,
m
a
c
h
in
e
lea
rn
in
g
,
d
e
e
p
lea
rn
i
n
g
,
a
n
d
h
ig
h
-
p
e
rfo
rm
a
n
c
e
c
o
m
p
u
ti
n
g
.
Ad
d
re
ss
:
F
a
c
u
lt
y
o
f
E
n
g
i
n
e
e
rin
g
,
M
in
ia
Un
iv
e
rsity
,
M
in
ia,
6
1
1
1
1
,
Eg
y
p
t.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
m
a
rwa
.
a
sb
a
s@
m
u
.
e
d
u
.
e
g
.
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
A
s
ystema
tic
ev
a
lu
a
tio
n
o
f p
r
e
-
tr
a
in
ed
en
co
d
er a
r
ch
itectu
r
es fo
r
mu
ltimo
d
a
l b
r
a
in
tu
mo
r
…
(
Ma
r
w
a
A
b
b
a
s
)
859
As
h
r
a
f
A.
M.
K
h
a
la
f
re
c
e
iv
e
d
t
h
e
B.
S
c
.
a
n
d
M
.
S
c
.
d
e
g
re
e
s
in
e
lec
tri
c
a
l
e
n
g
in
e
e
rin
g
fr
o
m
M
i
n
ia
Un
i
v
e
rsity
,
Eg
y
p
t,
in
1
9
8
9
a
n
d
1
9
9
4
,
re
sp
e
c
ti
v
e
ly
,
th
e
Do
c
t
o
r
o
f
En
g
i
n
e
e
rin
g
d
e
g
re
e
in
sy
ste
m
s
sc
i
e
n
c
e
a
n
d
e
n
g
i
n
e
e
rin
g
fr
o
m
t
h
e
G
ra
d
u
a
te
S
c
h
o
o
l
o
f
Na
tu
ra
l
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
,
Ka
n
a
z
a
wa
Un
iv
e
rsity
,
Ja
p
a
n
,
in
M
a
rc
h
2
0
0
0
,
a
n
d
t
h
e
P
h
.
D.
d
e
g
re
e
in
Eg
y
p
t.
He
is
c
u
rre
n
tl
y
a
P
ro
f
e
ss
o
r
o
f
DSP
a
n
d
th
e
He
a
d
o
f
t
h
e
El
e
c
tri
c
a
l
En
g
in
e
e
rin
g
De
p
a
rtme
n
t,
F
a
c
u
lt
y
o
f
E
n
g
in
e
e
r
in
g
,
M
i
n
ia
U
n
iv
e
rsit
y
.
ORCID
i
s:
0
0
0
0
-
0
0
0
3
-
3
3
4
4
-
5
4
2
0
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
a
sh
ra
f.
k
h
a
laf@m
u
.
e
d
u
.
e
g
.
H
u
ss
e
in
Mo
g
a
h
e
d
wo
rk
s
a
s
a
n
a
ss
istan
t
p
r
o
fe
ss
o
r
a
t
E
g
y
p
t'
s
M
in
ia
Un
i
v
e
rsity
.
He
g
ra
d
u
a
ted
fr
o
m
As
siu
t
U
n
iv
e
rsity
in
Eg
y
p
t
with
a
B.
S
c
.
i
n
Co
m
p
u
ter
a
n
d
Co
n
tr
o
l
En
g
i
n
e
e
rin
g
i
n
1
9
9
9
,
a
n
d
M
i
n
ia
Un
iv
e
rsity
in
E
g
y
p
t
a
wa
rd
e
d
h
i
m
a
n
M
.
S
c
.
in
Co
m
p
u
ter
En
g
i
n
e
e
rin
g
i
n
2
0
0
8
.
He
g
ra
d
u
a
t
e
d
fro
m
To
m
sk
P
o
ly
tec
h
n
ic
Un
i
v
e
rsity
in
Ru
ss
ia
i
n
2
0
1
7
with
a
P
h
.
D.
in
C
o
m
p
u
ter
En
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
(Ad
d
re
ss
:
F
a
c
u
lt
y
o
f
E
n
g
i
n
e
e
rin
g
,
M
in
ia
Un
iv
e
rsity
,
M
i
n
ia,
6
1
1
1
1
,
Eg
y
p
t
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
m
o
g
a
h
e
d
h
s@
m
u
.
e
d
u
.
e
g
.
Az
iza
I.
H
u
ss
e
in
re
c
e
iv
e
d
h
e
r
P
h
.
D.
d
e
g
re
e
in
El
e
c
tri
c
a
l
&
Co
m
p
u
ter
En
g
i
n
e
e
rin
g
fro
m
Ka
n
sa
s
S
tate
Un
iv
e
rsity
,
USA,
i
n
2
0
0
1
a
n
d
h
e
r
M
.
S
c
.
a
n
d
B.
S
c
.
d
e
g
re
e
s
fro
m
As
siu
t
U
n
iv
e
rsit
y
,
Eg
y
p
t
,
i
n
1
9
8
9
a
n
d
1
9
8
3
,
re
sp
e
c
ti
v
e
l
y
.
S
h
e
jo
i
n
e
d
Eff
a
t
Un
i
v
e
rsity
in
S
a
u
d
i
Ara
b
ia
in
2
0
0
4
a
n
d
e
sta
b
li
sh
e
d
t
h
e
first
El
e
c
tri
c
a
l
a
n
d
Co
m
p
u
ter
E
n
g
i
n
e
e
rin
g
p
ro
g
ra
m
fo
r
wo
m
e
n
in
t
h
e
c
o
u
n
try
a
n
d
tau
g
h
t
re
late
d
c
o
u
rse
s.
S
h
e
wa
s
t
h
e
h
e
a
d
o
f
th
e
El
e
c
tri
c
a
l
a
n
d
Co
m
p
u
ter
E
n
g
in
e
e
rin
g
De
p
a
rtme
n
t
a
t
Eff
a
t
U
n
iv
e
rs
it
y
fr
o
m
2
0
0
7
-
2
0
1
0
.
S
h
e
wa
s
th
e
h
e
a
d
o
f
t
h
e
Co
m
p
u
t
e
r
a
n
d
S
y
ste
m
s
En
g
in
e
e
ri
n
g
De
p
a
rtme
n
t,
F
a
c
u
lt
y
o
f
En
g
i
n
e
e
rin
g
,
M
i
n
ia Un
i
v
e
rsity
,
E
g
y
p
t,
fr
o
m
2
0
1
1
-
2
0
1
6
.
S
h
e
wa
s a
p
ro
fe
ss
o
r
a
n
d
c
h
a
ir
o
f
t
h
e
El
e
c
tri
c
a
l
&
Co
m
p
u
ter
E
n
g
in
e
e
rin
g
De
p
a
rtme
n
t
a
n
d
d
irec
to
r
o
f
th
e
M
a
ste
r
o
f
En
e
r
g
y
p
ro
g
ra
m
a
t
Eff
a
t
Un
i
v
e
rs
it
y
S
a
u
d
i
Ara
b
ia
fro
m
2
0
1
6
-
2
0
2
1
.
Cu
rre
n
tl
y
,
s
h
e
is
a
p
r
o
fe
ss
o
r
a
n
d
re
se
a
rc
h
e
r
a
t
th
e
sa
m
e
d
e
p
a
rt
m
e
n
t.
He
r
re
se
a
rc
h
i
n
tere
sts
i
n
c
lu
d
e
m
icro
e
lec
tro
n
ics
,
a
n
a
lo
g
/
d
ig
it
a
l
VLS
I
sy
ste
m
d
e
sig
n
,
RF
c
ircu
it
d
e
sig
n
,
h
i
g
h
-
sp
e
e
d
a
n
a
lo
g
-
to
-
d
i
g
it
a
l
c
o
n
v
e
rters
d
e
sig
n
,
a
n
d
wire
les
s
c
o
m
m
u
n
ica
ti
o
n
s
s
y
ste
m
s
d
e
sig
n
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
a
z
ib
ra
h
im@
e
ffa
tu
n
iv
e
rsity
.
e
d
u
.
sa
.
La
m
y
a
G
a
b
e
r
In
2
0
1
4
,
s
h
e
j
o
in
e
d
th
e
De
p
a
rtme
n
t
o
f
Co
m
p
u
t
e
r
a
n
d
S
y
ste
m
s
En
g
i
n
e
e
rin
g
,
M
i
n
ia
Un
i
v
e
rsity
,
E
g
y
p
t
a
s
a
tea
c
h
in
g
a
ss
istan
t,
a
n
d
sin
c
e
2
0
2
3
,
s
h
e
h
a
s
b
e
e
n
a
n
a
ss
istan
t
p
r
o
fe
ss
o
r.
He
r
re
se
a
rc
h
in
tere
sts
i
n
c
lu
d
e
tes
t
p
a
t
tern
g
e
n
e
ra
ti
o
n
,
fo
rm
a
l
v
e
rifi
c
a
ti
o
n
,
p
a
ra
l
lel
p
ro
g
ra
m
m
in
g
,
a
n
d
h
i
g
h
-
p
e
rfo
rm
a
n
c
e
c
o
m
p
u
ti
n
g
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
lam
y
a
.
g
a
b
e
r@m
u
.
e
d
u
.
e
g
.
Mo
h
a
m
e
d
M
o
u
r
a
d
M
a
b
r
o
o
k
re
c
e
iv
e
d
h
is
B
.
S
c
.
a
n
d
M
.
S
c
.
d
e
g
re
e
s
in
El
e
c
tri
c
a
l
a
n
d
Co
m
m
u
n
ica
ti
o
n
En
g
i
n
e
e
rin
g
fr
o
m
As
siu
t
Un
i
v
e
r
sity
,
Eg
y
p
t
,
in
2
0
0
8
a
n
d
2
0
1
3
,
re
sp
e
c
ti
v
e
ly
.
He
re
c
e
iv
e
d
a
P
h
.
D.
d
e
g
re
e
in
t
h
e
Co
m
m
u
n
ica
ti
o
n
De
p
t.
,
M
e
n
ia
Un
iv
e
rsity
,
Eg
y
p
t
in
2
0
1
7
.
H
e
is
a
n
As
sista
n
t
P
ro
fe
ss
o
r
a
t
S
p
a
c
e
Co
m
m
u
n
ica
ti
o
n
En
g
i
n
e
e
rin
g
De
p
t
i
n
t
h
e
fa
c
u
l
ty
o
f
Na
v
i
g
a
ti
o
n
S
c
ien
c
e
a
n
d
S
p
a
c
e
Tec
h
n
o
l
o
g
y
(NS
S
T),
Be
n
i
-
S
u
e
f
Un
i
v
e
rsity
,
Be
n
i
-
S
u
e
f,
Eg
y
p
t.
He
a
d
o
f
Ap
p
li
e
d
S
c
ien
c
e
s
o
f
S
p
a
c
e
a
n
d
n
a
v
ig
a
ti
o
n
P
ro
g
ra
m
S
c
ien
c
e
2
0
2
1
.
He
is
a
p
a
rt
-
ti
m
e
As
sista
n
t
p
ro
f.
a
t
F
a
c
u
l
ty
o
f
E
n
g
i
n
e
e
rin
g
,
Na
h
d
a
Un
iv
e
rsity
(NU
B),
Be
n
i
-
S
u
e
f,
E
g
y
p
t.
He
p
u
b
li
s
h
e
d
m
o
re
t
h
a
n
3
0
j
o
u
r
n
a
l
a
n
d
c
o
n
fe
re
n
c
e
p
a
p
e
rs
in
t
h
e
field
s
o
f
wire
les
s
c
o
m
m
u
n
ica
ti
o
n
s,
5
G
n
e
two
rk
s
,
c
o
g
n
i
ti
v
e
ra
d
i
o
,
Art
ifi
c
ial
in
telli
g
e
n
c
e
,
c
ircu
it
d
e
sig
n
,
a
n
d
se
n
so
rs.
He
is
a
re
v
iew
e
r
in
m
a
n
y
i
n
tern
a
ti
o
n
a
l
jo
u
r
n
a
ls
re
late
d
to
El
se
v
ier,
S
p
rin
g
e
r,
a
n
d
IEE
E
P
u
b
li
sh
e
rs.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
m
o
h
a
m
e
d
m
o
u
ra
d
2
0
0
8
@
g
m
a
il
.
c
o
m
or
m
o
h
a
m
e
d
.
m
o
u
ra
d
@n
ss
t.
b
su
.
e
d
u
.
e
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.