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.
41
,
No
.
1
,
J
an
u
ar
y
20
26
,
p
p
.
2
0
0
~
2
0
9
I
SS
N:
2
502
-
4
7
52
,
DOI
: 1
0
.
1
1
5
9
1
/ijee
cs
.v
41.
i
1
.
p
p
200
-
2
0
9
200
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs
.
ia
esco
r
e.
co
m
A nov
el appro
a
ch
f
o
r de
tecting
dia
betic
r
etinopa
thy
using
two
-
stream
CNNs
mo
del
P
ha
m
T
hi Viet
H
uo
ng
1
,
L
e
Duc
T
hi
nh
2
,
T
ra
n T
hi O
a
nh
1
,
T
ra
n Xua
n B
a
ch
2
,
H
o
a
ng
Q
ua
ng
H
uy
2
,
T
ra
n Anh Vu
2
1
I
n
t
e
r
n
a
t
i
o
n
a
l
S
c
h
o
o
l
,
V
i
e
t
n
a
m N
a
t
i
o
n
a
l
U
n
i
v
e
r
si
t
y
,
H
a
n
o
i
,
V
i
e
t
n
a
m
2
S
c
h
o
o
l
o
f
E
l
e
c
t
r
i
c
a
l
a
n
d
E
l
e
c
t
r
o
n
i
c
E
n
g
i
n
e
e
r
i
n
g
,
H
a
n
o
i
U
n
i
v
e
r
si
t
y
o
f
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
H
a
n
o
i
,
V
i
e
t
n
a
m
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Oct
30
,
2
0
2
5
R
ev
is
ed
Dec
4
,
2
0
2
5
Acc
ep
ted
Dec
13
,
2
0
2
5
M
a
jo
r
c
a
u
se
s
o
f
v
is
u
a
l
imp
a
irm
e
n
t,
p
a
rti
c
u
larly
d
ia
b
e
ti
c
re
ti
n
o
p
a
th
y
(DR
)
a
n
d
a
g
e
d
-
re
late
d
m
a
c
u
lar
d
e
g
e
n
e
ra
ti
o
n
(AMD),
h
a
s
p
o
se
d
s
ig
n
ifi
c
a
n
t
c
h
a
ll
e
n
g
e
s
fo
r
c
li
n
ica
l
d
ia
g
n
o
sis
a
n
d
trea
tme
n
t.
Early
d
e
tec
ti
o
n
a
n
d
p
ro
m
p
t
in
terv
e
n
t
io
n
c
a
n
h
e
l
p
p
re
v
e
n
t
se
v
e
re
c
o
n
se
q
u
e
n
c
e
s
fo
r
p
a
ti
e
n
ts.
Th
e
stu
d
y
p
re
se
n
ts
a
n
o
v
e
l
a
p
p
ro
a
c
h
f
o
r
d
e
tec
ti
n
g
e
y
e
d
ise
a
se
s
u
sin
g
a
t
wo
-
stre
a
m
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
two
r
k
(C
NN
)
m
o
d
e
l.
Th
e
first
stre
a
m
p
ro
c
e
ss
e
s
p
re
-
p
ro
c
e
ss
e
d
fu
n
d
u
s
ima
g
e
s,
wh
i
l
e
th
e
se
c
o
n
d
stre
a
m
a
n
a
ly
z
e
s
h
i
g
h
-
p
a
ss
fil
tere
d
fu
n
d
u
s
ima
g
e
s
i
n
t
h
e
sp
a
ti
a
l
fre
q
u
e
n
c
y
d
o
m
a
in
.
T
o
a
ss
e
ss
th
e
m
o
d
e
l
’
s p
e
rfo
rm
a
n
c
e
,
we
u
se
th
e
APT
OS
2
0
1
9
d
a
tas
e
t,
wh
ich
wa
s o
rig
i
n
a
ll
y
c
o
m
p
il
e
d
f
o
r
t
h
e
As
ia
P
a
c
ifi
c
Te
le
-
Op
h
th
a
lmo
lo
g
y
S
o
c
iet
y
2
0
1
9
Bli
n
d
n
e
ss
De
tec
ti
o
n
c
o
m
p
e
ti
ti
o
n
a
n
d
is
p
u
b
li
c
l
y
a
v
a
il
a
b
le
o
n
Ka
g
g
le.
Ou
r
m
e
th
o
d
sh
o
ws
p
r
o
m
ise
a
s
a
n
e
a
rly
sc
re
e
n
in
g
t
o
o
l
fo
r
DR
d
e
tec
ti
o
n
wi
th
a
n
a
c
c
u
ra
c
y
o
f
0
.
9
8
6
.
K
ey
w
o
r
d
s
:
Diab
etic
r
etin
o
p
ath
y
Fo
u
r
ier
tr
an
s
f
o
r
m
Fu
n
d
u
s
p
h
o
to
g
r
a
p
h
y
L
o
o
s
e
p
air
in
g
tr
ain
in
g
T
wo
-
s
tr
ea
m
C
NN
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
:
T
r
an
An
h
V
u
Sch
o
o
l o
f
E
lectr
ical
a
n
d
E
lectr
o
n
ic
E
n
g
i
n
ee
r
in
g
,
Han
o
i U
n
i
v
er
s
ity
o
f
Scien
ce
an
d
T
ec
h
n
o
l
o
g
y
Han
o
i,
Vietn
am
E
m
ail: v
u
.
tr
an
a
n
h
@
h
u
s
t.e
d
u
.
v
n
1.
I
NT
RO
D
UCT
I
O
N
R
etin
al
p
ath
o
lo
g
ies
in
cr
ea
s
in
g
ly
r
ep
r
esen
t
a
g
lo
b
al
h
ea
lth
c
o
n
ce
r
n
,
d
em
a
n
d
in
g
s
ig
n
if
ican
t
atten
tio
n
an
d
r
eso
u
r
ce
s
.
Am
o
n
g
th
ese
co
n
d
itio
n
s
,
r
etin
o
p
at
h
y
is
a
lead
in
g
ca
u
s
e
o
f
s
ev
er
e
v
is
u
al
im
p
air
m
en
t
an
d
b
lin
d
n
ess
wo
r
ld
wid
e
[
1
]
.
B
y
2
0
4
0
,
ag
e
-
r
elate
d
m
ac
u
la
r
d
e
g
e
n
er
atio
n
(
AM
D)
is
p
r
e
d
icted
t
o
af
f
ec
t
n
ea
r
ly
3
0
0
m
illi
o
n
p
eo
p
le.
Me
an
wh
ile,
o
n
e
o
f
th
e
m
ai
n
ca
u
s
es
o
f
v
is
io
n
lo
s
s
in
in
d
iv
id
u
als
is
d
iab
eti
c
r
etin
o
p
at
h
y
(
DR
)
,
wh
ich
h
as
b
ee
n
id
e
n
tifie
d
a
s
a
wo
r
ld
wid
e
e
p
id
em
ic.
T
h
is
u
n
d
er
s
co
r
es
th
e
cr
itical
n
ee
d
f
o
r
p
r
ev
en
tiv
e
m
ea
s
u
r
es,
ea
r
ly
d
iag
n
o
s
is
,
an
d
in
n
o
v
ativ
e
tr
ea
tm
en
ts
to
ad
d
r
ess
th
ese
d
eb
ilit
atin
g
ey
e
co
n
d
itio
n
s
[
2
]
,
[
3
]
.
Me
d
icin
e
an
d
b
io
lo
g
y
in
cr
ea
s
in
g
ly
d
ep
en
d
o
n
au
to
m
ated
s
y
s
tem
s
f
o
r
an
aly
s
is
an
d
d
ia
g
n
o
s
is
[
4
]
.
Fo
r
in
s
tan
ce
,
co
m
p
u
ter
-
aid
e
d
d
iag
n
o
s
is
(
C
AD)
o
f
r
etin
al
i
m
ag
es
ass
i
s
ts
h
ea
lth
ca
r
e
p
r
o
v
id
er
s
in
id
en
tify
in
g
d
is
ea
s
es,
in
it
iatin
g
ea
r
ly
tr
ea
tm
en
ts
,
an
d
r
ed
u
cin
g
in
co
n
s
is
ten
cies
in
im
ag
e
in
ter
p
r
etatio
n
[
5
]
.
Au
to
m
ated
an
al
y
s
is
also
o
f
f
er
s
n
u
m
er
o
u
s
b
en
ef
its
o
v
e
r
m
a
n
u
al
in
s
p
ec
tio
n
,
in
cl
u
d
in
g
co
s
t
ef
f
ic
ien
cy
,
o
b
jectiv
ity
,
r
eliab
ilit
y
,
an
d
r
ed
u
c
ed
d
ep
e
n
d
en
ce
o
n
s
k
illed
s
p
ec
ialis
ts
f
o
r
im
ag
e
ev
alu
atio
n
[
6
]
,
[
7
]
.
B
ef
o
r
e
th
e
g
r
o
win
g
o
f
d
ee
p
lear
n
in
g
(
DL
)
tech
n
iq
u
es
,
C
AD
s
y
s
tem
s
we
r
e
u
s
ed
f
o
r
im
ag
e
r
esto
r
atio
n
an
d
en
h
an
ce
m
en
t,
am
o
n
g
o
th
er
p
h
ases
o
f
th
e
r
etin
al
d
iag
n
o
s
tics
p
r
o
ce
s
s
.
So
m
e
C
AD
te
ch
n
iq
u
es
aim
to
r
ep
licate
th
e
p
r
o
ce
d
u
r
es
th
at
p
h
y
s
ician
s
d
o
to
id
en
tify
r
etin
al
d
is
o
r
d
er
s
,
wh
ich
in
clu
d
e
s
eg
m
en
tin
g
im
ag
es,
ex
tr
ac
tin
g
f
ea
tu
r
es
,
an
d
u
s
in
g
m
ac
h
in
e
lear
n
in
g
to
class
if
y
th
e
r
esu
lts
[
8
]
.
DL
tech
n
iq
u
es,
esp
ec
ially
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
C
NNs)
an
d
m
o
r
e
r
ec
en
tly
tr
an
s
f
o
r
m
er
m
o
d
els,
h
av
e
s
h
o
wn
r
em
ar
k
a
b
le
s
u
cc
ess
in
th
e
au
to
m
ated
d
iag
n
o
s
is
o
f
r
etin
al
d
is
o
r
d
er
s
.
Nu
m
er
o
u
s
ef
f
o
r
ts
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
n
o
ve
l a
p
p
r
o
a
c
h
fo
r
d
etec
tin
g
d
ia
b
etic
r
etin
o
p
a
th
y
u
s
in
g
t
w
o
-
s
tr
ea
m
C
N
N
s
mo
d
el
(
P
h
a
m
Th
i V
iet
Hu
o
n
g
)
201
h
av
e
f
o
c
u
s
ed
o
n
id
e
n
tify
in
g
p
r
ev
alen
t
r
etin
al
co
n
d
itio
n
s
,
s
u
ch
as
AM
D,
DR
,
an
d
g
lau
co
m
a.
Fo
r
in
s
tan
ce
,
a
m
eth
o
d
u
tili
zin
g
two
p
r
etr
ai
n
ed
C
NNs
(
VGG1
6
an
d
a
cu
s
to
m
C
NN)
was
d
ev
elo
p
ed
t
o
d
iag
n
o
s
e
DR
b
y
an
aly
zin
g
th
e
lik
elih
o
o
d
o
f
lesi
o
n
p
atch
es
[
9
]
.
Similar
ly
,
a
s
y
s
tem
wh
ich
co
m
b
in
ed
th
r
ee
C
NN
m
o
d
els
—
I
n
ce
p
tio
n
-
v
3
,
R
esNet1
5
2
,
a
n
d
I
n
ce
p
tio
n
-
R
esNet
-
v
2
was
b
u
ilt
to
d
etec
t
DR
in
f
u
n
d
u
s
im
ag
es
[
1
0
]
.
Ad
d
itio
n
ally
,
a
DL
-
b
ased
d
iag
n
o
s
tic
to
o
l
h
as
b
ee
n
c
r
ea
ted
to
s
cr
ee
n
p
atien
ts
f
o
r
a
r
an
g
e
o
f
co
m
m
o
n
r
etin
al
d
is
ea
s
es
[
1
1
]
.
T
h
e
ap
p
r
o
ac
h
es
s
till
h
av
e
s
o
m
e
wea
k
n
es
s
es,
s
u
ch
as
lo
s
s
o
f
im
a
g
e
in
f
o
r
m
atio
n
d
u
r
in
g
p
r
o
ce
s
s
in
g
o
r
h
ig
h
c
o
m
p
u
tatio
n
al
d
em
an
d
s
an
d
co
m
p
lex
ity
.
T
h
e
APTO
S
d
ataset
h
as
b
e
en
wid
ely
u
s
ed
in
v
ar
io
u
s
m
eth
o
d
s
f
o
r
class
if
y
in
g
e
y
e
d
is
ea
s
es.
T
h
e
r
esear
ch
i
n
[
1
2
]
e
m
p
h
asi
ze
d
tr
ain
in
g
o
n
a
f
o
cu
s
ed
s
u
b
s
et
o
f
ch
allen
g
in
g
ca
s
es
wh
ile
m
in
im
izin
g
th
e
in
f
lu
en
ce
o
f
a
lar
g
e
n
u
m
b
er
o
f
ea
s
y
n
eg
ativ
es
th
at
co
u
ld
o
v
er
wh
elm
th
e
d
etec
to
r
d
u
r
in
g
tr
ain
in
g
.
A
n
eu
r
o
n
in
tr
in
s
ic
lear
n
in
g
f
r
am
ew
o
r
k
was
in
tr
o
d
u
ce
d
i
n
[
1
3
]
to
id
en
tify
d
is
tr
a
cto
r
s
in
th
e
C
NN
f
ea
tu
r
e
s
p
ac
e.
I
t e
m
p
lo
y
e
d
a
n
o
v
el
d
is
tr
ac
to
r
-
awa
r
e
lo
s
s
f
u
n
ctio
n
to
cr
ea
te
a
clea
r
d
is
tin
ctio
n
b
etwe
en
th
e
o
r
ig
in
al
im
ag
e
a
n
d
its
d
is
tr
ac
to
r
with
in
th
e
f
ea
tu
r
e
s
p
ac
e.
T
h
e
a
p
p
r
o
ac
h
in
[
1
4
]
lev
e
r
ag
ed
c
o
n
tr
asti
v
e
le
ar
n
in
g
f
o
r
f
ea
t
u
r
e
em
b
ed
d
in
g
to
ad
d
r
ess
class
im
b
alan
ce
,
r
ep
lacin
g
th
e
tr
ad
it
io
n
al
cr
o
s
s
-
en
tr
o
p
y
lo
s
s
.
T
h
e
r
esear
ch
in
[
1
5
]
is
co
n
d
u
cte
d
th
r
o
u
g
h
an
iter
ativ
e
tr
ain
in
g
p
r
o
ce
s
s
to
g
et
an
id
en
tity
m
atr
ix
f
o
r
th
e
cr
o
s
s
-
co
r
r
el
atio
n
.
Desp
ite
s
ig
n
if
ican
t
ad
v
an
ce
m
en
ts
,
a
cr
itical
r
esear
ch
g
a
p
p
er
s
is
ts
in
d
ev
elo
p
in
g
a
m
o
d
el
th
at
ca
n
ef
f
ec
tiv
ely
in
teg
r
ate
m
u
lti
-
s
p
ec
tr
al
f
ea
tu
r
e
in
f
o
r
m
atio
n
wh
ile
m
ain
tain
in
g
co
m
p
u
tatio
n
al
ef
f
icien
cy
.
Sp
ec
if
ically
,
th
e
p
r
im
ar
y
s
cie
n
tific
p
r
o
b
lem
we
ad
d
r
ess
is
th
e
lack
o
f
a
r
o
b
u
s
t
m
o
d
el
ar
ch
itectu
r
e
th
at
ca
n
:
(
i
)
p
r
e
s
er
v
e
th
e
f
u
ll
co
n
tex
t
o
f
th
e
o
r
ig
in
al
f
u
n
d
u
s
im
ag
e
;
(
ii
)
i
s
o
late
an
d
em
p
h
asize
cr
itical
h
ig
h
-
f
r
e
q
u
en
c
y
p
ath
o
lo
g
ical
f
ea
tu
r
es
(
s
u
ch
as
m
icr
o
an
eu
r
y
s
m
s
an
d
h
em
o
r
r
h
ag
es)
u
s
in
g
a
d
ed
icate
d
ch
a
n
n
el,
an
d
(
iii
)
ac
h
iev
e
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
co
m
p
ar
ed
to
co
m
p
u
tatio
n
ally
h
e
av
y
e
n
s
em
b
le
m
eth
o
d
s
.
An
ef
f
icien
t
s
tr
ateg
y
f
o
r
f
u
s
in
g
th
ese
d
is
tin
ct
f
ea
tu
r
e
s
ets is
es
s
en
tial to
r
eso
lv
e
th
is
is
s
u
e.
T
o
o
v
e
r
co
m
e
th
is
r
esear
ch
g
ap
,
a
n
o
v
el
m
eth
o
d
o
lo
g
y
is
p
r
o
p
o
s
ed
f
o
r
d
etec
tin
g
DR
u
s
in
g
a
two
-
s
tr
ea
m
C
NN
m
o
d
el.
T
h
e
k
ey
i
n
n
o
v
atio
n
a
n
d
m
ain
c
o
n
tr
i
b
u
tio
n
o
f
th
is
r
esear
ch
is
th
e
in
teg
r
atio
n
o
f
t
h
e
p
r
o
p
o
s
ed
ar
ch
itectu
r
al
d
esig
n
an
d
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
/f
u
s
io
n
m
eth
o
d
o
l
o
g
y
.
T
h
e
f
i
r
s
t
s
tr
ea
m
p
r
o
ce
s
s
es
th
e
s
tan
d
ar
d
p
r
e
-
p
r
o
ce
s
s
ed
f
u
n
d
u
s
im
ag
e
(
with
d
ata
a
u
g
m
en
ta
tio
n
)
to
ca
p
tu
r
e
g
l
o
b
a
l,
l
o
w
-
f
r
eq
u
en
c
y
f
ea
t
u
r
es
(
e.
g
.
,
o
p
tic
d
is
c,
g
e
n
er
al
v
ascu
latu
r
e)
.
T
h
e
s
ec
o
n
d
s
tr
ea
m
p
r
o
ce
s
s
es
th
e
h
ig
h
-
p
ass
f
ilter
e
d
f
u
n
d
u
s
im
a
g
e
i
n
th
e
s
p
atial
f
r
eq
u
en
c
y
d
o
m
ain
(
d
er
i
v
ed
v
ia
f
o
u
r
ier
tr
a
n
s
f
o
r
m
)
.
T
h
is
s
tep
is
cr
u
cial
as
it
is
o
lates
an
d
em
p
h
asizes
th
e
cr
iti
ca
l,
h
ig
h
-
f
r
eq
u
en
cy
d
etails
r
elate
d
to
D
R
p
ath
o
lo
g
y
,
f
o
r
cin
g
th
e
n
etwo
r
k
to
f
o
cu
s
o
n
s
m
all,
clin
ically
r
elev
an
t
lesi
o
n
s
.
Featu
r
e
f
u
s
io
n
s
tr
ateg
y
af
ter
g
lo
b
al
av
e
r
ag
e
p
o
o
lin
g
(
GAP)
:
th
e
o
u
t
p
u
ts
f
r
o
m
th
e
two
s
tr
ea
m
s
ar
e
co
n
ca
ten
ate
d
af
ter
th
e
GAP
lay
er
.
T
h
is
s
tr
ateg
ic
f
u
s
io
n
p
o
in
t
en
s
u
r
es
th
at
ea
ch
s
tr
ea
m
h
as
alr
ea
d
y
ex
tr
ac
ted
r
o
b
u
s
t,
r
e
f
in
ed
,
h
ig
h
-
lev
el
s
em
an
tic
f
ea
tu
r
es
b
ef
o
r
e
m
er
g
in
g
,
lead
in
g
to
a
r
ich
er
f
in
al
r
ep
r
esen
tatio
n
f
o
r
class
if
icatio
n
.
T
h
is
co
m
b
in
ed
d
u
al
-
s
tr
ea
m
m
eth
o
d
o
lo
g
y
s
u
cc
ess
f
u
lly
p
r
eser
v
es
i
m
ag
e
in
teg
r
ity
wh
ile
ac
tiv
ely
g
u
id
i
n
g
th
e
m
ac
h
in
e
’
s
f
o
cu
s
to
war
d
s
p
ec
if
ic
p
ath
o
l
o
g
ical
f
ea
tu
r
es,
a
ch
iev
in
g
an
im
p
r
ess
iv
e
ac
cu
r
ac
y
o
f
0
.
9
8
6
o
n
th
e
APTO
S 2
0
1
9
d
ataset.
2.
M
E
T
H
O
D
2
.
1
.
Da
t
a
s
et
I
n
th
is
s
tu
d
y
,
we
em
p
lo
y
th
e
APTO
S
2
0
1
9
B
D
d
ataset
[
1
6
]
to
tr
ai
n
a
n
d
e
v
alu
ate
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
’
s
p
e
r
f
o
r
m
an
ce
.
T
h
is
d
ataset
co
m
p
r
is
es
3
,
6
6
2
s
am
p
les,
in
clu
d
in
g
1
,
8
0
5
n
o
r
m
al
im
ag
es
an
d
1
,
8
5
7
im
ag
es
d
ep
ictin
g
DR
,
co
llecte
d
f
r
o
m
a
lar
g
e
p
o
p
u
latio
n
in
I
n
d
ia.
T
h
e
co
llectio
n
p
r
o
ce
s
s
was
o
r
g
an
ized
b
y
t
h
e
Ar
av
in
d
E
y
e
Ho
s
p
ital
in
I
n
d
ia,
f
ea
tu
r
in
g
f
u
n
d
u
s
p
h
o
to
s
ta
k
en
u
n
d
er
d
if
f
e
r
en
t
c
o
n
d
itio
n
s
an
d
o
v
er
d
iv
er
s
e
p
er
io
d
s
.
A
team
o
f
tr
ain
e
d
m
ed
ical
ex
p
er
ts
ca
r
ef
u
lly
r
e
v
iewe
d
an
d
lab
eled
th
e
s
am
p
les
b
ased
o
n
th
e
I
n
ter
n
atio
n
al
C
lin
ical
Diab
etic
R
etin
o
p
ath
y
Dis
ea
s
e
Sev
er
ity
Scale
(
I
C
D
R
SS
)
.
T
h
e
APTO
S
2
0
1
9
B
D
d
ataset
is
ca
teg
o
r
ized
in
to
f
iv
e
g
r
o
u
p
s
b
ased
o
n
th
is
cr
iter
io
n
:
p
r
o
life
r
ativ
e
DR
,
m
ild
DR
,
m
o
d
er
ate
DR
,
s
ev
er
e
DR
,
an
d
n
o
DR
.
2
.
2
.
Wo
r
k
f
lo
w
T
h
e
wo
r
k
f
lo
w
o
f
th
e
m
o
d
el
is
p
r
esen
ted
in
Fig
u
r
e
1
.
T
h
e
d
etec
tio
n
s
y
s
tem
u
tili
ze
s
two
in
p
u
t
ch
an
n
els.
As
s
h
o
wn
in
th
e
wo
r
k
f
lo
w
d
iag
r
am
,
d
ata
au
g
m
e
n
tatio
n
is
ap
p
lied
f
i
r
s
t
to
ad
d
r
ess
d
ata
lim
itatio
n
s
an
d
m
itig
ate
o
v
er
f
itti
n
g
is
s
u
es.
Af
ter
th
at,
th
e
im
a
g
es a
r
e
d
iv
id
ed
in
to
two
c
h
an
n
els.
W
h
ile
th
e
s
ec
o
n
d
ch
an
n
el
o
f
f
er
s
ad
d
itio
n
al
in
f
o
r
m
atio
n
to
i
m
p
r
o
v
e
class
if
icatio
n
ac
cu
r
ac
y
,
th
e
f
ir
s
t
ch
an
n
el
p
r
o
ce
s
s
es
th
e
o
r
ig
in
a
l
im
ag
es.
B
y
tr
an
s
f
o
r
m
i
n
g
th
e
au
g
m
en
te
d
im
ag
es
in
to
th
e
s
p
atial
f
r
eq
u
en
c
y
d
o
m
ain
in
th
e
s
ec
o
n
d
ch
a
n
n
el,
we
ca
n
is
o
late
an
d
r
e
m
o
v
e
th
e
lo
w
-
f
r
eq
u
en
cy
d
ata.
Sin
ce
th
ese
lo
w
f
r
eq
u
e
n
cies
en
co
d
e
g
en
e
r
al
i
n
f
o
r
m
atio
n
(
lik
e
th
e
s
h
ap
e
o
f
th
e
o
p
tic
n
er
v
e
h
ea
d
,
m
ac
u
la,
an
d
m
ain
b
lo
o
d
v
ess
els),
th
eir
r
em
o
v
al
s
er
v
es
to
s
h
ar
p
en
th
e
f
o
c
u
s
o
n
s
p
ec
if
ic
in
d
icato
r
s
o
f
DR
.
Fin
ally
,
t
h
e
two
c
h
an
n
els
ar
e
f
ed
in
to
a
class
if
icatio
n
b
lo
c
k
,
wh
ich
is
a
two
-
s
tr
ea
m
C
NN
m
o
d
el,
to
d
etec
t th
e
d
is
ea
s
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
41
,
No
.
1
,
J
an
u
ar
y
20
26
:
200
-
2
0
9
202
Fig
u
r
e
1
.
T
h
e
wo
r
k
f
lo
w
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
2
.
3
.
Da
t
a
a
ug
m
ent
a
t
io
n
T
o
g
u
ar
an
tee
th
e
n
u
m
b
er
o
f
tr
ain
in
g
ex
am
p
les
is
lar
g
e
en
o
u
g
h
f
o
r
th
e
two
-
s
tr
ea
m
m
o
d
el,
we
em
p
lo
y
ed
d
ata
a
u
g
m
en
tatio
n
d
u
r
in
g
th
e
tr
ai
n
in
g
p
r
o
ce
s
s
[
1
7
]
,
a
p
p
ly
in
g
at
least
o
n
e
au
g
m
en
tatio
n
t
o
ea
ch
tr
ain
in
g
im
a
g
e
b
e
f
o
r
e
in
p
u
ttin
g
it
in
to
th
e
two
-
s
tr
ea
m
C
NN
m
o
d
el.
T
h
ese
au
g
m
en
tatio
n
s
,
im
p
lem
en
ted
u
s
in
g
th
e
Alb
u
m
en
tatio
n
s
lib
r
ar
y
[
1
8
]
,
in
clu
d
ed
o
p
tical
d
is
to
r
tio
n
,
g
r
id
d
is
to
r
tio
n
,
p
iece
wis
e
af
f
i
n
e
tr
an
s
f
o
r
m
atio
n
s
,
h
o
r
izo
n
tal
a
n
d
v
e
r
tical
f
lip
s
,
r
an
d
o
m
r
o
tatio
n
s
,
s
h
if
ts
,
s
ca
lin
g
,
r
ed
,
g
r
ee
n
,
an
d
b
lu
e
(
R
GB
)
v
alu
e
s
h
if
ts
,
an
d
r
an
d
o
m
ad
ju
s
tm
en
ts
to
b
r
ig
h
tn
ess
an
d
co
n
tr
ast.
As
a
r
esu
lt,
th
e
d
ataset
ex
p
a
n
d
ed
to
2
9
,
2
9
6
s
am
p
les,
co
m
p
r
is
in
g
1
4
,
4
4
0
n
o
r
m
al
im
ag
es
an
d
1
4
,
8
5
6
d
is
ea
s
e
im
ag
es.
T
h
is
au
g
m
e
n
tatio
n
p
r
o
ce
s
s
s
ig
n
if
ican
tly
en
h
an
ce
d
d
ata
d
iv
e
r
s
ity
,
r
e
d
u
cin
g
o
v
er
f
itti
n
g
a
n
d
im
p
r
o
v
in
g
t
h
e
m
o
d
el
’
s
g
en
er
aliza
t
io
n
ca
p
a
b
ilit
ies
f
o
r
d
etec
tin
g
an
d
class
if
y
in
g
DR
.
2
.
4
.
H
i
g
h
-
pa
s
s
f
ilte
r
I
n
th
e
tr
ain
in
g
an
d
v
alid
atio
n
p
h
ases
,
m
o
d
if
ie
d
v
e
r
s
io
n
s
o
f
th
e
o
r
i
g
in
al
im
ag
es
wer
e
u
tili
z
ed
.
T
h
ese
p
h
o
to
s
wer
e
s
ca
led
an
d
cr
o
p
p
ed
.
T
h
e
APTO
S2
0
1
9
d
atase
t
s
h
o
ws
er
r
o
n
eo
u
s
ass
o
ciatio
n
s
b
etwe
en
d
is
ea
s
e
s
tag
es
an
d
a
n
u
m
b
er
o
f
i
m
ag
e
m
eta
-
f
ea
tu
r
es,
in
clu
d
in
g
b
r
i
g
h
tn
ess
,
zo
o
m
lev
el,
cr
o
p
p
in
g
t
y
p
e,
an
d
r
eso
lu
tio
n
.
T
o
p
r
e
v
en
t
C
NNs
f
r
o
m
o
v
er
f
itti
n
g
to
th
ese
m
eta
-
f
ea
tu
r
es
an
d
to
m
i
n
im
ize
co
r
r
elatio
n
s
b
etwe
en
th
e
im
ag
e
co
n
ten
t
a
n
d
s
u
ch
attr
i
b
u
tes,
e
x
ten
s
iv
e
au
g
m
en
tatio
n
s
we
r
e
ap
p
lied
.
Su
b
s
eq
u
en
tly
,
h
i
g
h
-
p
ass
f
ilter
in
g
u
s
in
g
f
o
u
r
ier
t
r
an
s
f
o
r
m
was im
p
lem
en
ted
o
n
th
e
f
ir
s
t stre
am
’
s
in
p
u
ts
to
g
en
er
ate
in
p
u
ts
f
o
r
th
e
s
ec
o
n
d
s
tr
ea
m
:
2
.
4
.
1
.
F
o
urier
t
ra
ns
f
o
r
m
T
h
e
f
o
u
r
ier
tr
an
s
f
o
r
m
is
ca
p
ab
le
o
f
d
ec
o
m
p
o
s
in
g
im
a
g
es
b
y
a
n
aly
zin
g
v
ar
iatio
n
s
i
n
in
ten
s
ity
(
g
r
ay
lev
els)
ac
r
o
s
s
s
p
at
ial
d
is
tan
ce
s
,
r
ath
er
th
an
ex
a
m
in
in
g
s
ig
n
als
th
at
ch
an
g
e
o
v
er
tim
e
[
1
9
]
.
T
h
is
p
r
o
ce
s
s
tr
an
s
f
o
r
m
s
th
e
tim
e
d
o
m
ain
o
f
th
e
im
ag
es
in
to
th
e
s
p
ati
al
d
o
m
ain
.
I
n
th
e
co
n
tex
t
o
f
s
o
u
n
d
,
f
r
eq
u
e
n
cy
r
ep
r
esen
ts
th
e
(
in
v
e
r
s
e)
r
eg
u
la
r
ity
o
f
s
in
e
wav
e
r
e
p
etitio
n
s
.
Similar
ly
,
in
im
ag
es,
s
p
atial
f
r
eq
u
en
c
y
d
escr
ib
es
th
e
(
in
v
e
r
s
e)
r
eg
u
lar
ity
o
f
f
l
u
ctu
atio
n
s
in
i
n
ten
s
ity
v
alu
es
.
T
h
e
s
p
atial
f
r
eq
u
e
n
cy
is
d
e
ter
m
in
ed
u
s
in
g
th
e
f
o
llo
win
g
f
o
r
m
u
la:
(
,
)
=
∬
(
,
)
∙
−
2
(
+
)
∞
−
∞
(
1
)
wh
er
e
(
,
)
:
g
r
ay
lev
el
v
alu
e
at
a
p
ix
el
in
t
h
e
s
p
atial
d
o
m
ain
,
(
,
)
:
v
alu
e
i
n
th
e
s
p
atial
f
r
eq
u
en
cy
d
o
m
ain
,
a
n
d
(
,
)
: sp
atial
f
r
eq
u
en
c
y
co
o
r
d
in
ates
.
2
.
4
.
2
.
Sp
ec
t
rum
Hig
h
s
p
atial
f
r
eq
u
en
cies
co
r
r
esp
o
n
d
to
im
a
g
e
f
ea
tu
r
es
th
a
t
s
h
o
w
r
ap
id
in
ten
s
ity
v
ar
iati
o
n
s
o
v
er
s
h
o
r
t
d
is
tan
ce
s
[
1
9
]
.
I
n
c
o
n
tr
a
s
t,
f
ea
tu
r
es
with
g
r
ad
u
al
i
n
ten
s
ity
ch
an
g
es
o
v
er
lo
n
g
e
r
d
is
tan
ce
s
ar
e
ass
o
ciate
d
with
lo
w
s
p
atial
f
r
eq
u
en
cies.
I
m
ag
e
p
o
p
u
lar
ity
ca
n
b
e
d
e
ter
m
in
ed
b
y
a
p
p
ly
in
g
th
e
f
o
u
r
ier
tr
a
n
s
f
o
r
m
a
n
d
ca
lcu
latin
g
th
e
n
atu
r
al
lo
g
ar
ith
m
o
f
th
e
r
esu
lt.
(
,
)
=
l
og
(
1
+
|
(
,
)
|
2
)
(
2
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
n
o
ve
l a
p
p
r
o
a
c
h
fo
r
d
etec
tin
g
d
ia
b
etic
r
etin
o
p
a
th
y
u
s
in
g
t
w
o
-
s
tr
ea
m
C
N
N
s
mo
d
el
(
P
h
a
m
Th
i V
iet
Hu
o
n
g
)
203
2
.
4
.
3
.
Ana
ly
zing
s
pa
t
ia
l f
re
q
uency
do
m
a
in
T
h
e
s
p
atial
f
r
eq
u
en
c
y
is
d
eter
m
in
ed
b
y
t
h
e
d
is
tan
ce
f
r
o
m
th
e
ce
n
ter
o
f
th
e
d
is
cr
ete
f
o
u
r
ie
r
tr
an
s
f
o
r
m
(
DFT)
[
2
0
]
.
A
s
p
atial
f
r
eq
u
en
cy
o
f
ze
r
o
co
r
r
esp
o
n
d
s
to
th
e
im
ag
e
’
s
av
er
ag
e
b
r
ig
h
tn
ess
.
A
p
o
in
t
lo
ca
ted
two
p
ix
els
f
r
o
m
th
e
ce
n
ter
r
e
p
r
esen
ts
a
s
in
u
s
o
id
al
co
m
p
o
n
e
n
t
o
f
th
e
im
ag
e
th
at
co
m
p
letes
two
cy
cles
o
f
t
h
e
s
in
e
wav
e.
C
o
n
s
eq
u
en
tly
,
th
e
d
is
tan
ce
in
th
e
im
ag
e,
d
en
o
ted
a
s
,
an
d
its
co
r
r
esp
o
n
d
i
n
g
s
p
at
ial
f
r
eq
u
e
n
c
y
,
,
ca
n
b
e
ex
p
r
ess
ed
as:
=
∆
(
3
)
wh
er
e
∆
: th
e
im
ag
e
d
im
en
s
io
n
(
in
p
ix
el)
.
2
.
4
.
4
.
H
ig
hp
a
s
s
f
ilte
ring
As
an
aly
ze
d
,
in
f
u
n
d
u
s
p
h
o
to
g
r
ap
h
y
,
th
e
lo
w
f
r
eq
u
e
n
cies
co
r
r
esp
o
n
d
to
r
eg
io
n
s
n
ea
r
th
e
ce
n
ter
o
f
th
e
s
p
ec
tr
u
m
,
wh
ile
h
ig
h
f
r
eq
u
en
cies
ar
e
ass
o
ciate
d
with
th
e
o
u
ter
ar
ea
s
o
f
th
e
s
p
ec
tr
u
m
.
T
o
ap
p
ly
h
i
g
h
-
p
ass
f
ilter
in
g
,
a
th
r
esh
o
ld
0
is
d
ef
in
ed
,
r
ep
r
esen
tin
g
th
e
d
is
tan
ce
f
r
o
m
th
e
ce
n
ter
o
f
t
h
e
s
p
ec
tr
u
m
to
a
p
o
in
t
in
th
e
s
p
ec
tr
al
s
p
ac
e.
Fo
r
ea
ch
p
o
in
t
in
th
e
s
p
ec
tr
u
m
,
if
its
d
i
s
ta
n
ce
f
r
o
m
th
e
ce
n
ter
is
less
t
h
an
0
,
it
will
b
e
r
em
o
v
ed
.
T
o
ca
lcu
late
d
is
tan
ce
s
in
a
2
D
s
p
ac
e,
th
e
E
u
clid
ea
n
d
is
tan
ce
is
u
s
ed
.
T
h
e
f
o
r
m
u
la
f
o
r
E
u
clid
ea
n
d
is
tan
ce
b
etwe
en
two
p
o
i
n
ts
=
(
1
,
2
,
…
,
)
an
d
=
(
1
,
2
,
…
,
)
is
:
(
,
)
=
√
∑
(
−
)
2
=
1
(
4
)
wh
er
e
-
,
:co
m
p
o
n
e
n
ts
o
f
p
o
i
n
ts
P a
n
d
Q
r
esp
ec
tiv
ely
,
a
n
d
n
: n
u
m
b
er
o
f
d
im
e
n
s
io
n
s
o
f
s
p
ac
e
.
An
illu
s
tr
atio
n
o
f
u
s
in
g
h
ig
h
-
p
ass
f
ilter
in
g
f
o
r
f
u
n
d
u
s
im
ag
es
is
p
r
esen
ted
in
Fig
u
r
es
2
an
d
3
.
Fig
u
r
e
2
(
a)
s
h
o
ws
th
e
f
r
eq
u
en
cy
s
p
ec
tr
u
m
o
f
th
e
im
ag
e
b
e
f
o
r
e
f
ilter
in
g
,
an
Fig
u
r
e
2
(
b
)
s
h
o
ws
th
e
f
r
eq
u
en
cy
s
p
ec
tr
u
m
af
ter
h
ig
h
-
p
ass
f
ilter
in
g
.
Fig
u
r
e
3
(
a)
is
th
e
f
u
n
d
u
s
i
m
ag
e
b
ef
o
r
e
p
r
o
ce
s
s
in
g
.
Fig
u
r
e
3
(
b
)
is
th
e
im
ag
e
af
ter
h
ig
h
-
p
ass
f
ilter
in
g
,
wh
i
ch
clar
if
ies
im
p
o
r
tan
t
d
etails.
Af
ter
h
ig
h
-
p
ass
f
ilter
in
g
,
al
l
o
f
th
e
im
p
o
r
ta
n
t
in
f
o
r
m
atio
n
ab
o
u
t
DR
is
h
ig
h
lig
h
ted
.
(
a)
(
b
)
Fig
u
r
e
2
.
I
m
ag
es
o
f
(
a)
o
r
ig
in
a
l a
n
d
(
b
)
af
ter
h
ig
h
-
p
ass
f
ilter
in
g
in
s
p
atial
f
r
e
q
u
en
c
y
d
o
m
ain
(
a)
(
b
)
Fig
u
r
e
3
.
I
ll
u
s
tr
atio
n
o
f
(
a)
b
e
f
o
r
e
an
d
(
b
)
af
ter
h
i
g
h
-
p
ass
f
ilter
in
g
2
.
5
.
T
wo
-
s
t
re
a
m
net
wo
r
k
a
r
chit
ec
t
ure
I
n
th
is
s
tu
d
y
,
a
two
-
s
tr
ea
m
C
NN
is
d
esig
n
ed
to
p
r
o
c
ess
two
d
is
tin
ct
f
ea
tu
r
e
cla
s
s
es
in
p
u
t.
T
h
e
ar
ch
itectu
r
e
o
f
th
is
m
o
d
e
l
is
i
llu
s
tr
ated
in
Fig
u
r
e
4
,
wi
th
R
esNet
s
er
v
in
g
as
it
s
b
ac
k
b
o
n
e.
T
h
is
d
esig
n
co
n
ce
p
t
s
h
ar
es
s
im
ilar
ities
with
th
e
two
-
s
tr
ea
m
n
etwo
r
k
ap
p
r
o
ac
h
p
r
e
v
io
u
s
ly
e
x
p
lo
r
e
d
in
th
e
d
o
m
ain
o
f
h
u
m
an
ac
tiv
ity
r
ec
o
g
n
itio
n
in
v
id
eo
s
[
2
1
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
41
,
No
.
1
,
J
an
u
ar
y
20
26
:
200
-
2
0
9
204
Fig
u
r
e
4
.
T
wo
-
s
tr
ea
m
C
NNs m
o
d
el
T
h
e
f
ir
s
t
R
esNet
s
tr
ea
m
f
o
cu
s
es
o
n
e
x
tr
ac
tin
g
f
ea
tu
r
es
f
r
o
m
f
u
n
d
u
s
im
a
g
es,
wh
ile
t
h
e
s
ec
o
n
d
s
tr
ea
m
p
r
o
v
id
es
s
u
p
p
lem
en
tar
y
i
n
f
o
r
m
atio
n
to
s
u
p
p
o
r
t
th
e
f
ir
s
t
ch
an
n
el.
B
o
th
b
r
a
n
ch
es
o
f
th
e
t
wo
-
s
tr
ea
m
C
NN
ar
e
b
ased
o
n
R
esNet
-
34
.
W
h
ile
o
t
h
er
C
NN
ar
ch
itectu
r
es
co
u
ld
h
av
e
b
ee
n
em
p
lo
y
ed
,
R
esNet
-
3
4
was
s
elec
ted
f
o
r
s
ev
er
al
r
ea
s
o
n
s
:
it
h
as
f
ewe
r
p
ar
am
eter
s
,
r
eq
u
ir
es
less
d
ata
f
o
r
tr
ain
i
n
g
,
a
n
d
h
as
p
r
o
v
e
n
ef
f
ec
tiv
e
in
v
a
r
io
u
s
f
u
n
d
u
s
im
ag
e
id
en
tific
atio
n
ta
s
k
s
[
2
2
]
,
[
2
3
]
.
A
s
in
g
le
-
m
o
d
al
C
NN
cr
ea
te
s
s
u
cc
ess
iv
e
2
D
f
ea
tu
r
e
m
a
p
s
lay
er
b
y
lay
er
in
o
r
d
e
r
to
ex
tr
ac
t
in
f
o
r
m
atio
n
f
r
o
m
a
n
im
ag
e.
T
h
ese
f
ea
tu
r
e
m
a
p
s
k
ee
p
s
o
m
e
s
p
atial
in
f
o
r
m
atio
n
f
r
o
m
th
e
o
r
ig
in
al
in
p
u
t
im
ag
e
ev
en
if
th
ey
g
et
s
m
aller
an
d
s
m
aller
.
T
h
u
s
,
af
ter
th
e
GAP
l
ay
er
,
we
ch
o
o
s
e
to
u
n
d
e
r
tak
e
s
p
atially
in
v
ar
ian
t
f
u
s
io
n
.
B
y
r
ed
u
cin
g
ea
ch
f
ea
t
u
r
e
m
ap
to
a
s
in
g
le
s
ca
lar
v
alu
e,
th
e
GAP
lay
er
ef
f
icien
tly
elim
in
ates
s
p
atial
in
f
o
r
m
atio
n
.
I
n
th
e
f
ir
s
t
s
tr
ea
m
,
th
e
f
i
n
al
co
n
v
o
lu
tio
n
al
b
lo
c
k
o
f
R
esNet
-
3
4
p
r
o
d
u
ce
s
5
1
2
f
e
atu
r
e
m
ap
s
,
r
ep
r
esen
ted
as
(
=
{
,
1
,
.
.
.
,
,
512
}
)
.
E
ac
h
f
ea
tu
r
e
m
ap
h
as
d
im
en
s
io
n
s
o
f
m
×
m
,
wh
er
e
th
e
s
ize
o
f
m
d
ep
en
d
s
o
n
th
e
i
n
p
u
t
im
ag
e
s
i
ze
.
Fo
r
an
in
p
u
t
s
ize
o
f
4
4
8
×
4
4
8
,
m
eq
u
als
1
4
.
T
o
ex
tr
ac
t
a
f
ea
tu
r
e
v
alu
e
at
a
ce
r
tain
p
o
s
itio
n
(
x
,
y
)
in
a
f
ea
tu
r
e
m
ap
,
,
we
u
s
e
t
h
e
n
o
tatio
n
,
(
,
)
.
Similar
ly
,
th
e
s
ec
o
n
d
s
tr
ea
m
b
r
an
c
h
o
u
tp
u
ts
f
ea
t
u
r
e
m
ap
s
r
e
p
r
esen
ted
as
(
=
{
,
1
,
.
.
.
,
,
512
}
)
.
B
o
th
an
d
ar
e
p
ass
ed
th
r
o
u
g
h
th
e
GAP
lay
er
s
im
u
ltan
eo
u
s
ly
,
r
esu
ltin
g
in
two
5
1
2
-
d
im
e
n
s
io
n
al
f
ea
tu
r
e
v
ec
to
r
s
d
en
o
ted
as
̅
=
(
,
1
̅
̅
̅
̅
̅
,
.
.
.
,
,
512
̅
̅
̅
̅
̅
̅
̅
)
an
d
̅
=
(
,
1
̅
̅
̅
̅
̅
,
.
.
.
,
,
512
̅
̅
̅
̅
̅
̅
̅
)
.
T
h
e
f
u
s
io
n
m
o
d
u
le
lay
er
co
m
b
in
es
th
e
h
i
g
h
-
lev
el
o
u
tp
u
t
f
ea
tu
r
es
f
r
o
m
two
s
ep
a
r
ate
s
tr
ea
m
s
,
ef
f
ec
tiv
ely
u
tili
zin
g
th
e
in
f
o
r
m
atio
n
f
r
o
m
ea
ch
to
im
p
r
o
v
e
o
v
er
all
p
er
f
o
r
m
a
n
ce
.
T
h
is
f
u
s
i
o
n
lay
er
ar
c
h
itectu
r
e
p
r
o
v
id
es
two
k
ey
ad
v
an
tag
es
.
First,
it
ca
p
tu
r
es
a
m
u
ch
r
ich
er
r
ep
r
esen
tatio
n
o
f
th
e
o
r
i
g
in
al
d
ata
with
o
u
t
r
eq
u
ir
in
g
th
e
tr
ain
in
g
o
f
m
u
ltip
le
class
if
ier
s
,
a
s
s
ee
n
in
lat
e
-
f
u
s
io
n
m
eth
o
d
s
.
Seco
n
d
,
in
co
n
tr
ast
to
ea
r
ly
-
f
u
s
io
n
ap
p
r
o
ac
h
es
wh
e
r
e
f
u
s
i
o
n
o
cc
u
r
s
at
th
e
in
itial
lay
er
,
th
e
C
NN
lay
er
s
in
ea
ch
in
d
iv
i
d
u
al
s
tr
ea
m
co
llect
m
o
r
e
u
s
ef
u
l
an
d
r
ef
in
e
d
in
f
o
r
m
atio
n
f
r
o
m
th
e
r
aw
f
ea
tu
r
es.
As
illu
s
tr
ated
in
Fig
u
r
e
4
,
th
e
f
ir
s
t
s
tr
ea
m
p
r
o
ce
s
s
es
f
u
n
d
u
s
im
ag
es,
with
ea
ch
elem
en
t
o
f
th
e
o
u
t
p
u
t
ten
s
o
r
co
r
r
esp
o
n
d
in
g
to
th
ese
im
ag
es,
wh
ile
th
e
s
ec
o
n
d
s
tr
ea
m
h
an
d
les h
ig
h
-
p
ass
f
ilter
ed
im
ag
es,
with
ea
ch
ten
s
o
r
elem
en
t r
ep
r
esen
tin
g
th
is
f
ilter
ed
d
ata.
T
h
e
f
u
s
ed
f
ea
tu
r
e
ten
s
o
r
,
r
e
p
r
esen
ted
as
=
(
,
)
,
co
m
b
in
es
th
e
s
tr
en
g
th
s
o
f
b
o
th
s
tr
ea
m
s
.
Af
ter
th
at,
th
is
ten
s
o
r
is
s
en
t
in
to
later
C
NN
lay
er
s
f
o
r
ad
d
itio
n
al
p
r
o
ce
s
s
in
g
.
L
astl
y
,
th
e
class
if
icatio
n
is
ca
r
r
ied
o
u
t
b
y
a
So
f
tMa
x
lay
er
an
d
a
f
u
ll
y
co
n
n
ec
ted
(
FC
)
lay
er
.
T
h
is
m
eth
o
d
im
p
r
o
v
es
p
r
ed
ictio
n
ac
c
u
r
ac
y
b
y
allo
wi
n
g
ad
a
p
tiv
e
tu
n
in
g
an
d
b
alan
ci
n
g
th
e
s
ig
n
if
ica
n
ce
o
f
d
if
f
er
e
n
t f
ea
tu
r
e
s
ets.
I
n
tr
ain
in
g
an
d
test
in
g
p
h
ase,
t
h
e
cr
o
s
s
-
en
tr
o
p
y
lo
s
s
f
u
n
ctio
n
s
h
o
wn
b
elo
w
is
u
s
ed
:
=
−
∑
∙
l
og
(
)
−
1
=
0
(
5
)
W
h
er
e
,
−
=
[
0
,
…
,
−
1
]
r
ep
r
esen
ts
a
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
,
d
en
o
tes
th
e
p
r
o
b
a
b
ilit
y
th
at
a
s
am
p
le
b
elo
n
g
s
to
class
.
−
=
[
0
,
…
,
−
1
]
d
en
o
tes th
e
o
n
e
-
h
o
t r
ep
r
esen
tatio
n
o
f
class
lab
els,
an
d
is
th
e
n
u
m
b
e
r
o
f
class
es.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
I
n
th
is
p
ar
t,
we
co
n
d
u
ct
ex
p
er
im
en
ts
to
v
alid
ate
th
e
p
r
o
p
o
s
ed
m
o
d
el
an
d
d
em
o
n
s
tr
ate
th
e
ef
f
ec
tiv
en
ess
.
Sectio
n
3
.
1
p
r
e
s
en
ts
th
e
ex
p
er
im
en
tal
r
esu
lt
s
u
s
in
g
th
e
APTO
S
d
ataset,
wh
ile
3
.
2
ex
p
l
o
r
es
ad
d
itio
n
al
ex
p
er
im
en
ts
to
ev
al
u
ate
th
e
m
o
d
el
’
s
co
m
p
atib
ilit
y
wi
th
o
th
er
d
atasets
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
n
o
ve
l a
p
p
r
o
a
c
h
fo
r
d
etec
tin
g
d
ia
b
etic
r
etin
o
p
a
th
y
u
s
in
g
t
w
o
-
s
tr
ea
m
C
N
N
s
mo
d
el
(
P
h
a
m
Th
i V
iet
Hu
o
n
g
)
205
3
.
1
.
E
x
perim
ent
r
esu
lt
s
Fo
r
th
is
p
ar
t,
th
e
APTO
S
d
ataset
was
s
p
lit
in
to
two
p
ar
ts
:
tr
ain
in
g
an
d
test
in
g
,
with
a
7
:3
r
atio
,
r
esp
ec
tiv
ely
.
I
n
itially
,
th
e
p
er
f
o
r
m
a
n
ce
o
f
co
n
v
en
tio
n
al
C
NN
m
o
d
els
was
n
o
t
p
a
r
ticu
lar
ly
im
p
r
ess
iv
e,
ac
h
iev
in
g
a
m
ax
im
u
m
ac
cu
r
a
cy
o
f
0
.
8
8
an
d
an
F1
s
co
r
e
o
f
0
.
8
7
.
T
o
im
p
r
o
v
e
th
ese
r
esu
lts
,
we
o
p
ted
f
o
r
a
two
-
s
tr
ea
m
m
o
d
el,
an
ticip
atin
g
b
etter
p
er
f
o
r
m
an
ce
c
o
m
p
a
r
ed
to
tr
a
d
itio
n
al
s
in
g
le
-
s
tr
ea
m
C
NN
m
o
d
els.
W
e
test
ed
s
ev
er
al
co
m
m
o
n
ar
ch
it
ec
tu
r
es,
in
clu
d
in
g
I
n
ce
p
tio
n
,
Mo
b
i
leNe
t,
VGG,
an
d
R
esNet.
R
esu
lts
ar
e
s
h
o
wn
in
T
ab
le
1
.
Hig
h
-
p
ass
f
ilter
in
g
in
o
n
e
c
h
an
n
el
p
lay
s
a
cr
itical
r
o
le
in
tr
ain
in
g
t
h
e
two
-
s
tr
ea
m
m
o
d
el,
as
i
t
d
ir
ec
tly
in
f
lu
e
n
ce
s
th
e
m
o
d
el
’
s
p
er
f
o
r
m
a
n
ce
,
g
e
n
er
aliza
tio
n
,
an
d
tr
ain
in
g
ef
f
icien
cy
.
T
ab
le
2
s
u
m
m
ar
izes
th
e
m
o
d
el
’
s
r
o
b
u
s
t
p
er
f
o
r
m
an
ce
.
T
h
e
class
if
ier
ex
h
ib
its
ex
ce
llen
t
p
r
ec
is
io
n
,
ac
h
iev
in
g
0
.
9
9
f
o
r
th
e
“
No
r
m
al
”
ca
teg
o
r
y
a
n
d
0
.
9
8
2
f
o
r
“
DR
”
.
Fu
r
th
er
m
o
r
e,
its
r
ec
a
ll
r
ates
(
0
.
9
8
5
f
o
r
“
No
r
m
al
”
an
d
0
.
9
8
8
f
o
r
“
DR
”
)
u
n
d
er
s
co
r
e
th
e
m
o
d
el
’
s
h
ig
h
s
en
s
itiv
ity
in
ca
p
t
u
r
in
g
p
o
s
itiv
e
ca
s
es.
T
h
e
F1
s
co
r
es,
a
co
m
p
r
eh
en
s
iv
e
m
ea
s
u
r
e
th
at
h
ar
m
o
n
izes
p
r
ec
is
io
n
an
d
r
ec
all,
a
r
e
u
n
if
o
r
m
l
y
s
tr
o
n
g
at
0
.
9
8
6
f
o
r
b
o
th
class
es.
T
h
is
co
n
s
is
ten
c
y
p
o
in
ts
to
an
o
p
tim
al
b
alan
ce
b
etwe
en
m
in
im
izin
g
f
alse
p
o
s
itiv
es
an
d
f
alse
n
eg
ativ
es,
co
n
f
ir
m
in
g
th
e
m
o
d
el
’
s
h
ig
h
ef
f
ec
tiv
en
ess
an
d
r
eliab
ilit
y
in
ac
cu
r
ately
d
is
tin
g
u
is
h
in
g
b
etwe
e
n
n
o
r
m
al
an
d
DR
ca
s
e
.
T
ab
le
1
.
Per
f
o
r
m
an
ce
o
f
d
if
f
er
en
t m
o
d
els
M
e
t
h
o
d
s
A
c
c
u
r
a
c
y
F1
-
sc
o
r
e
1
-
st
r
e
a
m c
o
n
v
e
n
t
i
o
n
a
l
C
N
N
mo
d
e
l
0
.
8
8
0
.
8
7
2
-
st
r
e
a
m
C
N
N
mo
d
e
l
I
n
c
e
p
t
i
o
n
_
v
3
0
.
9
8
4
0
.
9
8
3
M
o
b
i
l
e
N
e
t
0
.
9
6
0
.
9
5
6
VGG
0
.
9
5
9
0
.
9
6
R
e
s
N
et
0
.
9
8
6
0
.
9
8
6
T
ab
le
2
.
C
lass
if
icatio
n
p
er
f
o
r
m
an
ce
D
i
sea
s
e
s
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
sc
o
r
e
N
o
r
mal
0
.
9
9
0
.
9
8
5
0
.
9
8
6
DR
0
.
9
8
2
0
.
9
8
8
0
.
9
8
6
As
s
h
o
wn
in
Fig
u
r
e
5
,
th
e
ar
e
a
u
n
d
e
r
th
e
cu
r
v
e
(
AUC)
an
d
r
ec
eiv
er
o
p
er
atin
g
ch
ar
ac
te
r
is
tic
(
R
OC
)
cu
r
v
es
h
i
g
h
lig
h
ts
th
e
m
o
d
el
’
s
s
tab
ilit
y
.
A
h
ig
h
AUC
s
co
r
e,
ty
p
ically
p
air
ed
with
a
well
-
f
o
r
m
ed
R
OC
cu
r
v
e,
d
em
o
n
s
tr
ates
th
at
th
e
m
o
d
el
m
ain
tain
s
s
tr
o
n
g
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
v
ar
io
u
s
th
r
esh
o
ld
s
.
T
h
is
s
tab
ili
ty
is
cr
itical
f
o
r
e
n
s
u
r
in
g
co
n
s
is
ten
t
class
i
f
icatio
n
p
e
r
f
o
r
m
an
ce
,
as
it
r
e
f
lects
th
e
m
o
d
el
’
s
a
b
ilit
y
to
r
eliab
ly
d
is
tin
g
u
is
h
b
etwe
en
p
o
s
itiv
e
an
d
n
eg
ativ
e
in
s
tan
ce
s
with
m
in
im
al
v
ar
ian
ce
in
p
r
ed
ictio
n
s
.
T
h
er
ef
o
r
e
,
t
h
e
s
tr
o
n
g
AUC an
d
R
OC
cu
r
v
es p
r
o
v
id
e
co
n
f
id
en
ce
in
th
e
m
o
d
el
’
s
r
o
b
u
s
tn
ess
an
d
r
eliab
ilit
y
.
Fig
u
r
e
5
.
AUC an
d
R
OC
cu
r
v
e
T
ab
le
3
co
m
p
a
r
es
th
e
s
u
g
g
e
s
ted
ap
p
r
o
a
ch
t
o
v
a
r
io
u
s
s
tate
-
of
-
th
e
-
ar
t
m
eth
o
d
s
a
n
d
s
h
o
ws
th
at
it
p
er
f
o
r
m
s
b
etter
in
ter
m
s
o
f
a
cc
u
r
ac
y
a
n
d
F1
-
s
co
r
e.
T
h
e
s
u
g
g
ested
m
eth
o
d
s
h
o
ws
r
em
a
r
k
ab
le
ac
cu
r
ac
y
i
n
ca
teg
o
r
izin
g
ca
s
es
in
th
e
d
ata
s
et,
with
an
ac
cu
r
ac
y
o
f
0
.
9
8
6
.
Fu
r
th
e
r
m
o
r
e
,
th
e
m
o
d
el
’
s
c
ap
ac
ity
to
s
tr
ik
e
an
ap
p
r
o
p
r
iate
b
alan
ce
b
etwe
en
a
cc
u
r
ac
y
an
d
r
ec
all
is
d
em
o
n
s
tr
ated
b
y
its
F1
-
s
co
r
e
o
f
0
.
9
8
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
41
,
No
.
1
,
J
an
u
ar
y
20
26
:
200
-
2
0
9
206
T
ab
le
3
.
C
o
m
p
a
r
is
o
n
with
th
e
s
tate
-
of
-
th
e
-
ar
t a
p
p
r
o
ac
h
es with
APTO
S d
ataset
M
e
t
h
o
d
s
A
c
c
u
r
a
c
y
F1
-
sc
o
r
e
FG
-
S
S
L
[
1
5
]
0
.
8
5
8
0
.
7
2
0
P
r
o
C
o
[
2
4
]
0
.
8
3
7
0
.
6
7
4
C
L
+
r
e
s
a
mp
l
e
[
1
4
]
0
.
8
1
6
0
.
6
0
8
C
L
[
1
4
]
0
.
8
2
5
0
.
6
5
2
D
A
N
I
L
[
1
3
]
0
.
8
2
5
0
.
6
6
0
F
o
c
a
l
l
o
ss
[
1
2
]
0
.
8
1
5
0
.
6
2
9
Pr
o
p
o
sed
m
e
t
h
o
d
0
.
9
8
6
0
.
9
8
6
I
n
co
n
tr
ast,
alter
n
ativ
e
m
eth
o
d
s
s
u
ch
as
FG
-
SS
L
,
Pro
C
o
,
C
L
+
r
esam
p
le,
C
L
,
an
d
DA
NI
L
y
ield
lo
wer
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
th
ese
m
etr
ics.
Fo
r
in
s
tan
ce
,
w
h
ile
FG
-
SS
L
ac
h
iev
es
a
r
ea
s
o
n
ab
le
ac
c
u
r
ac
y
o
f
0
.
8
5
8
,
its
F1
-
s
co
r
e
is
s
ig
n
if
ican
tly
lo
wer
at
0
.
7
2
0
.
Similar
ly
,
th
e
p
r
o
p
o
s
ed
m
eth
o
d
s
u
r
p
ass
es
Pro
C
o
,
C
L
+
r
esam
p
le,
C
L
,
an
d
DANI
L
in
b
o
th
ac
c
u
r
ac
y
a
n
d
F1
-
s
co
r
e,
f
u
r
th
er
d
em
o
n
s
tr
atin
g
its
s
u
p
er
io
r
ity
in
d
eliv
er
in
g
ac
cu
r
ate
p
r
e
d
ictio
n
s
.
3
.
2
.
Co
m
pa
t
ibi
lity
ev
a
lua
t
io
n
I
n
th
is
s
ec
tio
n
,
we
ev
al
u
ate
o
u
r
m
o
d
el
u
s
in
g
a
d
if
f
er
en
t
d
ataset
—
th
e
DR
d
ataset
[
2
5
]
.
T
h
e
DR
d
ataset,
o
b
tain
ed
f
r
o
m
Kag
g
le
an
d
p
r
o
v
id
e
d
b
y
E
y
ePAC
S,
a
f
r
ee
p
latf
o
r
m
f
o
r
r
etin
o
p
ath
y
s
cr
ee
n
in
g
,
c
o
n
tain
s
3
5
,
1
2
6
f
u
n
d
u
s
im
ag
es.
Of
th
e
s
e,
2
5
,
8
0
5
ar
e
n
o
r
m
al
(
d
is
ea
se
-
f
r
ee
)
,
an
d
9
,
3
2
1
s
h
o
w
s
ig
n
s
o
f
DR
.
T
o
a
d
d
r
ess
th
e
is
s
u
e
o
f
class
im
b
alan
ce
,
9
,
3
2
1
n
o
r
m
al
f
u
n
d
u
s
im
ag
e
s
an
d
9
,
3
2
1
d
is
ea
s
e
-
af
f
ec
ted
f
u
n
d
u
s
im
ag
es
ar
e
s
elec
ted
f
r
o
m
th
e
DR
d
ataset.
T
h
e
d
ataset
is
th
en
d
iv
id
ed
i
n
to
tr
ain
in
g
a
n
d
test
in
g
s
ets,
with
a
7
0
%
tr
ai
n
in
g
an
d
3
0
% testi
n
g
s
p
lit.
T
ab
le
4
d
em
o
n
s
tr
ates
th
at
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
u
tp
er
f
o
r
m
ed
all
o
th
er
s
,
ac
h
iev
in
g
th
e
h
ig
h
est
ac
cu
r
ac
y
o
f
7
6
%,
with
a
0
.
7
6
F1
-
s
co
r
e,
in
d
icatin
g
a
well
-
b
a
lan
ce
d
p
er
f
o
r
m
an
ce
i
n
b
o
th
p
r
ec
is
io
n
an
d
r
ec
all.
Alex
Net,
an
ea
r
ly
C
NN,
ac
h
i
e
v
ed
an
ac
cu
r
ac
y
o
f
7
3
%,
wh
il
e
R
esNet
-
5
0
,
a
d
ee
p
er
C
NN
ar
ch
itectu
r
e,
s
lig
h
tly
ex
ce
ed
ed
it
with
7
5
%.
A
m
o
d
if
ied
v
er
s
io
n
o
f
R
esNet
-
5
0
m
ain
tain
ed
a
co
m
p
etitiv
e
ac
cu
r
ac
y
o
f
7
4
%.
T
h
ese
r
esu
lts
s
u
g
g
est
th
at
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
p
r
o
v
i
d
es
a
s
ig
n
if
ican
t
im
p
r
o
v
em
e
n
t
o
v
er
well
-
estab
lis
h
ed
ar
ch
itectu
r
es su
ch
as Ale
x
Net,
R
esNet
-
5
0
,
an
d
th
eir
r
ev
is
ed
v
er
s
io
n
s
.
T
ab
le
4
.
C
o
m
p
a
r
is
o
n
with
th
e
s
tate
-
of
-
th
e
-
ar
t a
p
p
r
o
ac
h
es with
DR
d
ataset
M
e
t
h
o
d
s
A
c
c
u
r
a
c
y
F1
-
sc
o
r
e
A
l
e
x
N
e
t
[
2
5
]
0
.
7
3
_
R
e
sN
e
t
-
5
0
[
2
5
]
0
.
7
5
_
R
e
v
i
s
e
d
R
e
sN
e
t
-
5
0
[
2
5
]
0
.
7
4
_
Pr
o
p
o
sed
m
e
t
h
o
d
0
.
7
6
0
.
7
6
4.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
s
u
cc
ess
f
u
lly
in
t
r
o
d
u
ce
d
a
n
o
v
el
two
-
s
tr
ea
m
C
NN
m
o
d
el
f
o
r
d
ia
b
etic
r
etin
o
p
ath
y
(
DR
)
d
etec
tio
n
.
T
h
e
p
r
o
p
o
s
ed
ar
c
h
itectu
r
e,
wh
ich
u
tili
ze
s
R
e
s
Net
-
3
4
as
its
b
ac
k
b
o
n
e,
u
n
iq
u
ely
in
teg
r
ates
in
f
o
r
m
atio
n
f
r
o
m
two
ch
an
n
e
ls
:
th
e
o
r
ig
in
al
f
u
n
d
u
s
im
ag
e
an
d
a
h
ig
h
-
p
ass
f
ilter
ed
im
ag
e
in
th
e
f
r
eq
u
en
cy
d
o
m
ain
.
T
h
e
h
ig
h
-
p
ass
ch
an
n
el
en
h
a
n
ce
s
f
ea
tu
r
e
d
is
tin
ctiv
en
ess
b
y
em
p
h
asizin
g
s
u
b
tle
d
is
ea
s
e
-
r
elate
d
d
etails.
E
x
p
er
im
en
tal
r
esu
lts
d
em
o
n
s
tr
ate
th
e
ef
f
ec
tiv
e
n
ess
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el,
ac
h
i
ev
in
g
an
ac
cu
r
ac
y
o
f
0
.
9
8
6
an
d
an
F1
-
s
co
r
e
o
f
0
.
9
8
6
o
n
th
e
APTO
S
2
0
1
9
d
ataset,
h
ig
h
lig
h
tin
g
its
r
eliab
ilit
y
as
an
ea
r
ly
s
cr
ee
n
in
g
to
o
l.
Desp
ite
its
s
tr
o
n
g
p
er
f
o
r
m
a
n
ce
,
th
e
m
o
d
el
s
till
f
ac
es
s
ev
er
al
ch
allen
g
es,
in
clu
d
i
n
g
its
r
elian
ce
o
n
o
p
tim
al
f
ilter
p
ar
am
eter
s
elec
tio
n
an
d
its
cu
r
r
en
t
f
o
c
u
s
o
n
b
in
ar
y
class
if
icatio
n
(
n
o
r
m
al
v
er
s
u
s
DR
)
r
ath
er
th
an
f
u
ll
clin
ical
s
ev
er
ity
g
r
ad
in
g
.
Ne
v
er
th
eless
,
th
is
r
esear
ch
s
h
o
ws
s
ig
n
if
ican
t
clin
ical
p
o
ten
tial
f
o
r
im
p
r
o
v
in
g
ea
r
ly
a
n
d
au
to
m
ated
s
cr
ee
n
in
g
,
p
ar
ticu
lar
ly
in
r
eg
io
n
s
with
lim
ited
ac
ce
s
s
to
o
p
h
th
alm
o
lo
g
y
s
p
ec
ialis
t
s
.
Fu
tu
r
e
wo
r
k
will
f
o
cu
s
o
n
ex
ten
d
i
n
g
th
e
m
o
d
el
to
m
u
lti
-
class
D
R
g
r
ad
in
g
in
a
cc
o
r
d
an
ce
with
th
e
I
C
DR
S
S
s
ev
er
ity
lev
els,
ex
p
lo
r
in
g
ad
a
p
tiv
e
f
ea
t
u
r
e
f
u
s
i
o
n
s
tr
ateg
ies
s
u
ch
as
atten
ti
o
n
m
ec
h
an
is
m
s
to
d
y
n
am
ically
b
ala
n
ce
s
p
atial
an
d
f
r
e
q
u
en
c
y
in
f
o
r
m
atio
n
,
an
d
d
ev
el
o
p
in
g
au
to
m
ated
f
ilter
o
p
tim
izatio
n
tech
n
iq
u
es
to
lear
n
o
p
tim
al
h
i
g
h
-
p
ass
f
ilter
ch
ar
ac
ter
is
tics
d
u
r
in
g
tr
ain
in
g
.
T
h
ese
im
p
r
o
v
e
m
en
ts
ar
e
ex
p
ec
ted
to
en
h
a
n
ce
th
e
m
o
d
el’
s
g
en
er
aliza
tio
n
ca
p
ab
ilit
y
ac
r
o
s
s
d
iv
er
s
e
im
ag
in
g
c
o
n
d
itio
n
s
an
d
f
u
r
th
er
i
n
cr
ea
s
e
its
clin
ical
ap
p
licab
ilit
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
n
o
ve
l a
p
p
r
o
a
c
h
fo
r
d
etec
tin
g
d
ia
b
etic
r
etin
o
p
a
th
y
u
s
in
g
t
w
o
-
s
tr
ea
m
C
N
N
s
mo
d
el
(
P
h
a
m
Th
i V
iet
Hu
o
n
g
)
207
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
is
r
esear
ch
is
f
u
n
d
ed
b
y
I
n
ter
n
atio
n
al
Sch
o
o
l,
Vietn
am
Natio
n
al
Un
iv
er
s
ity
,
Han
o
i
(
VNU
-
I
S)
u
n
d
er
p
r
o
ject
n
u
m
b
e
r
C
S.2
0
2
4
-
06.
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
Ph
am
T
h
i V
iet
Hu
o
n
g
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
L
e
Du
c
T
h
in
h
✓
✓
✓
✓
✓
✓
✓
T
r
an
T
h
i O
an
h
✓
✓
✓
✓
✓
✓
T
r
an
Xu
an
B
ac
h
✓
✓
✓
✓
Ho
an
g
Qu
an
g
Hu
y
✓
✓
✓
T
r
an
An
h
V
u
✓
✓
✓
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
in
g
s
o
f
th
is
s
tu
d
y
ar
e
o
p
e
n
ly
av
ailab
le
in
h
ttp
s
://www.
k
ag
g
le.
co
m
/d
atasets
/m
ar
iah
er
r
er
o
t/ap
to
s
2
0
1
9
.
RE
F
E
R
E
NC
E
S
[
1
]
U
.
S
c
h
mi
d
t
-
Er
f
u
r
t
h
,
A
.
S
a
d
e
g
h
i
p
o
u
r
,
B
.
S
.
G
e
r
e
n
d
a
s
,
S
.
M
.
W
a
l
d
st
e
i
n
,
a
n
d
H
.
B
o
g
u
n
o
v
i
ć
,
“
A
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
i
n
r
e
t
i
n
a
,
”
Pro
g
ress
i
n
R
e
t
i
n
a
l
a
n
d
E
y
e
Re
s
e
a
r
c
h
,
v
o
l
.
6
7
,
p
p
.
1
–
2
9
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
r
e
t
e
y
e
r
e
s
.
2
0
1
8
.
0
7
.
0
0
4
.
[
2
]
T.
D
.
L
.
K
e
e
n
a
n
,
C
.
A
.
C
u
k
r
a
s,
a
n
d
E.
Y
.
C
h
e
w
,
“
A
g
e
-
r
e
l
a
t
e
d
mac
u
l
a
r
d
e
g
e
n
e
r
a
t
i
o
n
:
e
p
i
d
e
m
i
o
l
o
g
y
a
n
d
c
l
i
n
i
c
a
l
a
s
p
e
c
t
s
,
”
Ad
v
a
n
c
e
s i
n
Ex
p
e
r
i
m
e
n
t
a
l
Me
d
i
c
i
n
e
a
n
d
Bi
o
l
o
g
y
,
v
o
l
.
1
2
5
6
,
p
p
.
1
–
3
1
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
0
-
6
6
0
1
4
-
7
_
1
.
[
3
]
Y
.
Z
h
e
n
g
,
M
.
H
e
,
a
n
d
N
.
C
o
n
g
d
o
n
,
“
Th
e
w
o
r
l
d
w
i
d
e
e
p
i
d
e
m
i
c
o
f
d
i
a
b
e
t
i
c
r
e
t
i
n
o
p
a
t
h
y
,
”
I
n
d
i
a
n
J
o
u
r
n
a
l
o
f
O
p
h
t
h
a
l
m
o
l
o
g
y
,
v
o
l
.
6
0
,
n
o
.
5
,
p
.
4
2
8
,
2
0
1
2
,
d
o
i
:
1
0
.
4
1
0
3
/
0
3
0
1
-
4
7
3
8
.
1
0
0
5
4
2
.
[
4
]
J.
H
a
n
,
M
.
K
a
m
b
e
r
,
a
n
d
Jj
t
mk
s
.
P
e
i
,
“
Th
e
M
o
r
g
a
n
K
a
u
f
ma
n
n
s
e
r
i
e
s
i
n
d
a
t
a
m
a
n
a
g
e
me
n
t
s
y
st
e
ms,
”
v
o
l
.
5
,
M
o
r
h
g
a
n
K
a
u
f
ma
n
n
P
u
b
l
i
s
h
e
r
s
,
2
0
1
1
,
p
p
.
8
3
–
1
2
4
.
[
5
]
K
.
M
i
t
t
a
l
a
n
d
V
.
M
.
A
.
R
a
j
a
m,
“
C
o
mp
u
t
e
r
i
z
e
d
r
e
t
i
n
a
l
i
m
a
g
e
a
n
a
l
y
s
i
s
-
a
su
r
v
e
y
,
”
M
u
l
t
i
m
e
d
i
a
T
o
o
l
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
7
9
,
n
o
.
3
1
–
3
2
,
p
p
.
2
2
3
8
9
–
2
2
4
2
1
,
M
a
y
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
042
-
020
-
0
9
0
4
1
-
y.
[
6
]
M
.
D
.
A
b
r
a
m
o
f
f
,
M
.
K
.
G
a
r
v
i
n
,
a
n
d
M
.
S
o
n
k
a
,
“
R
e
t
i
n
a
l
i
m
a
g
i
n
g
a
n
d
i
ma
g
e
a
n
a
l
y
si
s
,
”
I
E
EE
R
e
v
i
e
w
s
i
n
Bi
o
m
e
d
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
3
,
p
p
.
1
6
9
–
2
0
8
,
2
0
1
0
,
d
o
i
:
1
0
.
1
1
0
9
/
R
B
M
E
.
2
0
1
0
.
2
0
8
4
5
6
7
.
[
7
]
M
.
M
.
F
r
a
z
e
t
a
l
.
,
“
B
l
o
o
d
v
e
ss
e
l
se
g
men
t
a
t
i
o
n
me
t
h
o
d
o
l
o
g
i
e
s
i
n
r
e
t
i
n
a
l
i
mag
e
s
-
a
su
r
v
e
y
,
”
C
o
m
p
u
t
e
r
Me
t
h
o
d
s
a
n
d
Pr
o
g
r
a
m
s
i
n
Bi
o
m
e
d
i
c
i
n
e
,
v
o
l
.
1
0
8
,
n
o
.
1
,
p
p
.
4
0
7
–
4
3
3
,
O
c
t
.
2
0
1
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
m
p
b
.
2
0
1
2
.
0
3
.
0
0
9
.
[
8
]
M
.
A
.
O
mar,
M
.
A
.
T
a
h
i
r
,
a
n
d
F
.
K
h
e
l
i
f
i
,
“
M
u
l
t
i
-
l
a
b
e
l
l
e
a
r
n
i
n
g
m
o
d
e
l
f
o
r
i
m
p
r
o
v
i
n
g
r
e
t
i
n
a
l
i
ma
g
e
c
l
a
ssi
f
i
c
a
t
i
o
n
i
n
d
i
a
b
e
t
i
c
r
e
t
i
n
o
p
a
t
h
y
,
”
i
n
2
0
1
7
4
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
C
o
n
t
r
o
l
,
D
e
c
i
s
i
o
n
a
n
d
I
n
f
o
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
i
e
s,
C
o
D
I
T
2
0
1
7
,
A
p
r
.
2
0
1
7
,
v
o
l
.
2
0
1
7
-
Ja
n
u
a
r
y
,
p
p
.
2
0
2
–
2
0
7
,
d
o
i
:
1
0
.
1
1
0
9
/
C
o
D
I
T.
2
0
1
7
.
8
1
0
2
5
9
1
.
[
9
]
G
.
T.
Z
a
g
o
,
R
.
V
.
A
n
d
r
e
ã
o
,
B
.
D
o
r
i
z
z
i
,
a
n
d
E
.
O
.
T
e
a
t
i
n
i
S
a
l
l
e
s
,
“
D
i
a
b
e
t
i
c
r
e
t
i
n
o
p
a
t
h
y
d
e
t
e
c
t
i
o
n
u
s
i
n
g
r
e
d
l
e
s
i
o
n
l
o
c
a
l
i
z
a
t
i
o
n
a
n
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,
”
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
1
6
,
p
.
1
0
3
5
3
7
,
J
a
n
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
b
i
o
m
e
d
.
2
0
1
9
.
1
0
3
5
3
7
.
[
1
0
]
H
.
Ji
a
n
g
,
K
.
Y
a
n
g
,
M
.
G
a
o
,
D
.
Zh
a
n
g
,
H
.
M
a
,
a
n
d
W
.
Q
i
a
n
,
“
A
n
i
n
t
e
r
p
r
e
t
a
b
l
e
e
n
sem
b
l
e
d
e
e
p
l
e
a
r
n
i
n
g
m
o
d
e
l
f
o
r
d
i
a
b
e
t
i
c
r
e
t
i
n
o
p
a
t
h
y
d
i
sea
s
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
i
n
2
0
1
9
4
1
st
A
n
n
u
a
l
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
f
t
h
e
I
EEE
En
g
i
n
e
e
r
i
n
g
i
n
Me
d
i
c
i
n
e
a
n
d
Bi
o
l
o
g
y
S
o
c
i
e
t
y
(
EM
BC
)
,
Ju
l
.
2
0
1
9
,
p
p
.
2
0
4
5
–
2
0
4
8
,
d
o
i
:
1
0
.
1
1
0
9
/
EM
B
C
.
2
0
1
9
.
8
8
5
7
1
6
0
.
[
1
1
]
D
.
S
.
K
e
r
ma
n
y
e
t
a
l
.
,
“
I
d
e
n
t
i
f
y
i
n
g
me
d
i
c
a
l
d
i
a
g
n
o
ses
a
n
d
t
r
e
a
t
a
b
l
e
d
i
se
a
ses
b
y
i
m
a
g
e
-
b
a
se
d
d
e
e
p
l
e
a
r
n
i
n
g
,
”
C
e
l
l
,
v
o
l
.
1
7
2
,
n
o
.
5
,
p
p
.
1
1
2
2
-
1
1
3
1
.
e
9
,
F
e
b
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
e
l
l
.
2
0
1
8
.
0
2
.
0
1
0
.
[
1
2
]
T.
-
Y
.
L
i
n
,
P
.
G
o
y
a
l
,
R
.
G
i
r
s
h
i
c
k
,
K
.
H
e
,
a
n
d
P
.
D
o
l
l
a
r
,
“
F
o
c
a
l
l
o
ss
f
o
r
d
e
n
s
e
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
,
”
i
n
2
0
1
7
I
E
EE
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
s
i
o
n
(
I
C
C
V)
,
O
c
t
.
2
0
1
7
,
p
p
.
2
9
9
9
–
3
0
0
7
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
C
V
.
2
0
1
7
.
3
2
4
.
[
1
3
]
L.
G
o
n
g
,
K
.
M
a
,
a
n
d
Y
.
Zh
e
n
g
,
“
D
i
st
r
a
c
t
o
r
-
a
w
a
r
e
n
e
u
r
o
n
i
n
t
r
i
n
s
i
c
l
e
a
r
n
i
n
g
f
o
r
g
e
n
e
r
i
c
2
D
m
e
d
i
c
a
l
i
ma
g
e
c
l
a
ssi
f
i
c
a
t
i
o
n
s,”
i
n
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
Me
d
i
c
a
l
I
m
a
g
e
C
o
m
p
u
t
i
n
g
a
n
d
C
o
m
p
u
t
e
r
-
Assi
s
t
e
d
I
n
t
e
r
v
e
n
t
i
o
n
,
2
0
2
0
,
p
p
.
5
9
1
–
6
0
1
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
030
-
5
9
7
1
3
-
9
_
5
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
41
,
No
.
1
,
J
an
u
ar
y
20
26
:
200
-
2
0
9
208
[
1
4
]
Y
.
M
a
r
r
a
k
c
h
i
,
O
.
M
a
k
a
n
si
,
a
n
d
T.
B
r
o
x
,
“
F
i
g
h
t
i
n
g
c
l
a
ss
i
m
b
a
l
a
n
c
e
w
i
t
h
c
o
n
t
r
a
st
i
v
e
l
e
a
r
n
i
n
g
,
”
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
Me
d
i
c
a
l
I
m
a
g
e
C
o
m
p
u
t
i
n
g
a
n
d
C
o
m
p
u
t
e
r
-
Assi
s
t
e
d
I
n
t
e
r
v
e
n
t
i
o
n
,
2
0
2
1
,
p
p
.
4
6
6
–
4
7
6
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
030
-
8
7
1
9
9
-
4
_
4
4
.
[
1
5
]
W
.
P
a
r
k
a
n
d
J.
R
y
u
,
“
F
i
n
e
-
g
r
a
i
n
e
d
se
l
f
-
su
p
e
r
v
i
se
d
l
e
a
r
n
i
n
g
w
i
t
h
j
i
g
s
a
w
p
u
z
z
l
e
s
f
o
r
me
d
i
c
a
l
i
m
a
g
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
C
o
m
p
u
t
e
rs
i
n
Bi
o
l
o
g
y
a
n
d
Me
d
i
c
i
n
e
,
v
o
l
.
1
7
4
,
p
.
1
0
8
4
6
0
,
M
a
y
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
4
6
0
.
[
1
6
]
K
a
r
t
h
i
k
,
M
a
g
g
i
e
,
a
n
d
S
.
D
a
n
e
,
“
A
P
TO
S
2
0
1
9
B
l
i
n
d
n
e
ss
D
e
t
e
c
t
i
o
n
.
”
K
a
g
g
l
e
,
2
0
1
9
,
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
k
a
g
g
l
e
.
c
o
m/
c
o
m
p
e
t
i
t
i
o
n
s/
a
p
t
o
s
2
0
1
9
-
b
l
i
n
d
n
e
ss
-
d
e
t
e
c
t
i
o
n
.
[
1
7
]
A
.
K
r
i
z
h
e
v
s
k
y
,
I
.
S
u
t
s
k
e
v
e
r
,
a
n
d
G
.
E.
H
i
n
t
o
n
,
“
I
mag
e
N
e
t
c
l
a
s
si
f
i
c
a
t
i
o
n
w
i
t
h
d
e
e
p
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
,
”
A
d
v
a
n
c
e
s
i
n
N
e
u
ra
l
I
n
f
o
rm
a
t
i
o
n
Pr
o
c
e
ssi
n
g
S
y
s
t
e
m
s
,
v
o
l
.
2
5
,
M
a
y
2
0
1
2
,
d
o
i
:
1
0
.
1
1
4
5
/
3
0
6
5
3
8
6
.
[
1
8
]
A
.
B
u
sl
a
e
v
,
V
.
I
.
I
g
l
o
v
i
k
o
v
,
E
.
K
h
v
e
d
c
h
e
n
y
a
,
A
.
P
a
r
i
n
o
v
,
M
.
D
r
u
z
h
i
n
i
n
,
a
n
d
A
.
A
.
K
a
l
i
n
i
n
,
“
A
l
b
u
me
n
t
a
t
i
o
n
s
:
f
a
st
a
n
d
f
l
e
x
i
b
l
e
i
ma
g
e
a
u
g
m
e
n
t
a
t
i
o
n
s,”
I
n
f
o
rm
a
t
i
o
n
(
S
w
i
t
zerl
a
n
d
)
,
v
o
l
.
1
1
,
n
o
.
2
,
p
.
1
2
5
,
F
e
b
.
2
0
2
0
,
d
o
i
:
1
0
.
3
3
9
0
/
i
n
f
o
1
1
0
2
0
1
2
5
.
[
1
9
]
X
.
Z
h
a
n
g
,
Y
.
K
u
a
n
g
,
a
n
d
J.
Y
a
o
,
“
D
e
t
e
c
t
i
o
n
o
f
mi
c
r
o
a
n
e
u
r
y
sms
i
n
c
o
l
o
r
f
u
n
d
u
s
i
ma
g
e
s
b
a
s
e
d
o
n
l
o
c
a
l
f
o
u
r
i
e
r
t
r
a
n
sf
o
r
m,”
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
8
,
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
8
.
[
2
0
]
A
.
M
.
A
b
d
u
l
-
R
a
h
m
a
n
,
T
.
M
o
l
t
e
n
o
,
a
n
d
A
.
C
.
B
.
M
o
l
t
e
n
o
,
“
F
o
u
r
i
e
r
a
n
a
l
y
si
s
o
f
d
i
g
i
t
a
l
r
e
t
i
n
a
l
i
ma
g
e
s
i
n
e
s
t
i
m
a
t
i
o
n
o
f
c
a
t
a
r
a
c
t
sev
e
r
i
t
y
,
”
C
l
i
n
i
c
a
l
a
n
d
Ex
p
e
r
i
m
e
n
t
a
l
O
p
h
t
h
a
l
m
o
l
o
g
y
,
v
o
l
.
3
6
,
n
o
.
7
,
p
p
.
6
3
7
–
6
4
5
,
S
e
p
.
2
0
0
8
,
d
o
i
:
1
0
.
1
1
1
1
/
j
.
1
4
4
2
-
9
0
7
1
.
2
0
0
8
.
0
1
8
1
9
.
x
.
[
2
1
]
C
.
F
e
i
c
h
t
e
n
h
o
f
e
r
,
A
.
P
i
n
z
,
a
n
d
A
.
Zi
sserm
a
n
,
“
C
o
n
v
o
l
u
t
i
o
n
a
l
t
w
o
-
st
r
e
a
m
n
e
t
w
o
r
k
f
u
s
i
o
n
f
o
r
v
i
d
e
o
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
,
”
i
n
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
I
EEE
C
o
m
p
u
t
e
r
S
o
c
i
e
t
y
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
t
e
r
V
i
si
o
n
a
n
d
P
a
t
t
e
r
n
R
e
c
o
g
n
i
t
i
o
n
,
J
u
n
.
2
0
1
6
,
v
o
l
.
2
0
1
6
-
D
e
c
e
m
b
e
r
,
p
p
.
1
9
3
3
–
1
9
4
1
,
d
o
i
:
1
0
.
1
1
0
9
/
C
V
P
R
.
2
0
1
6
.
2
1
3
.
[
2
2
]
Q
.
W
e
i
,
X
.
L
i
,
H
.
W
a
n
g
,
D
.
D
i
n
g
,
W
.
Y
u
,
a
n
d
Y
.
C
h
e
n
,
“
La
s
e
r
sca
r
d
e
t
e
c
t
i
o
n
i
n
f
u
n
d
u
s
i
ma
g
e
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
,
”
i
n
C
o
m
p
u
t
e
r
V
i
si
o
n
–
AC
C
V
2
0
1
8
:
1
4
t
h
Asi
a
n
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
s
i
o
n
,
2
0
1
9
,
p
p
.
1
9
1
–
2
0
6
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
0
-
2
0
8
7
0
-
7
_
1
2
.
[
2
3
]
X
.
La
i
,
X
.
L
i
,
R
.
Q
i
a
n
,
D
.
D
i
n
g
,
J
.
W
u
,
a
n
d
J.
X
u
,
“
F
o
u
r
mo
d
e
l
s
f
o
r
a
u
t
o
m
a
t
i
c
r
e
c
o
g
n
i
t
i
o
n
o
f
l
e
f
t
a
n
d
r
i
g
h
t
e
y
e
i
n
f
u
n
d
u
s i
ma
g
e
s
,
”
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
(
i
n
c
l
u
d
i
n
g
s
u
b
seri
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
1
2
9
5
LN
C
S
,
S
p
r
i
n
g
e
r
I
n
t
e
r
n
a
t
i
o
n
a
l
P
u
b
l
i
sh
i
n
g
,
2
0
1
9
,
p
p
.
5
0
7
–
5
1
7
.
[
2
4
]
Z
.
Y
a
n
g
e
t
a
l
.
,
“
P
r
o
C
o
:
p
r
o
t
o
t
y
p
e
-
a
w
a
r
e
c
o
n
t
r
a
s
t
i
v
e
l
e
a
r
n
i
n
g
f
o
r
l
o
n
g
-
t
a
i
l
e
d
m
e
d
i
c
a
l
i
m
a
g
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
,
”
i
n
I
n
t
e
r
n
a
t
i
o
n
a
l
c
o
n
f
e
r
e
n
c
e
o
n
m
e
d
i
c
a
l
i
m
a
g
e
c
o
m
p
u
t
i
n
g
a
n
d
c
o
m
p
u
t
e
r
-
a
s
s
i
s
t
e
d
i
n
t
e
r
v
e
n
t
i
o
n
,
2
0
2
2
,
p
p
.
1
7
3
–
1
8
2
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
1
-
1
6
4
5
2
-
1
_
1
7
.
[
2
5
]
C
.
L
.
Li
n
a
n
d
K
.
C
.
W
u
,
“
D
e
v
e
l
o
p
m
e
n
t
o
f
r
e
v
i
se
d
R
e
sN
e
t
-
5
0
f
o
r
d
i
a
b
e
t
i
c
r
e
t
i
n
o
p
a
t
h
y
d
e
t
e
c
t
i
o
n
,
”
B
MC
Bi
o
i
n
f
o
rm
a
t
i
c
s
,
v
o
l
.
2
4
,
n
o
.
1
,
A
p
r
.
2
0
2
3
,
d
o
i
:
1
0
.
1
1
8
6
/
s
1
2
8
5
9
-
0
2
3
-
0
5
2
9
3
-
1.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Dr
.
Ph
a
m
Th
i
Vie
t
H
u
o
n
g
o
b
tai
n
e
d
h
e
r
B
.
S
.
i
n
e
lec
tri
c
a
l
e
n
g
in
e
e
rin
g
fr
o
m
Ha
n
o
i
Un
i
v
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
in
2
0
0
7
.
S
h
e
e
a
rn
e
d
h
e
r
M
.
S
c
.
a
n
d
P
h
.
D.,
b
o
th
in
e
lec
tri
c
a
l
e
n
g
in
e
e
rin
g
,
f
ro
m
t
h
e
Un
iv
e
rsity
o
f
M
a
ss
a
c
h
u
se
tt
s
L
o
we
ll
,
Un
it
e
d
S
tate
s,
in
2
0
1
0
a
n
d
2
0
1
2
,
re
sp
e
c
ti
v
e
ly
.
F
r
o
m
2
0
1
2
to
2
0
1
5
,
sh
e
wa
s
a
re
se
a
rc
h
er
a
t
th
e
M
a
n
n
in
g
S
c
h
o
o
l
o
f
Bu
sin
e
ss
in
Lo
we
ll
,
M
a
ss
a
c
h
u
se
tt
s.
F
ro
m
2
0
1
7
to
2
0
2
0
,
sh
e
wa
s
a
fa
c
u
lt
y
m
e
m
b
e
r
a
t
Vie
t
n
a
m
Na
ti
o
n
a
l
Un
i
v
e
rsity
(VN
U),
U
n
iv
e
rsit
y
o
f
E
n
g
i
n
e
e
rin
g
a
n
d
Tec
h
n
o
l
o
g
y
(VN
U
-
UET)
,
Vie
tn
a
m
.
S
in
c
e
2
0
2
0
,
sh
e
h
a
s
b
e
e
n
wo
rk
in
g
a
t
th
e
I
n
tern
a
ti
o
n
a
l
S
c
h
o
o
l
–
VN
U.
He
r
re
se
a
rc
h
in
tere
sts
in
c
l
u
d
e
d
a
ta
m
in
i
n
g
a
n
d
a
n
a
ly
ti
c
s,
m
a
c
h
in
e
lea
rn
in
g
m
e
th
o
d
o
lo
g
ies
,
wit
h
a
p
p
li
c
a
ti
o
n
s
in
b
io
m
e
d
ica
l
e
n
g
in
e
e
rin
g
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
h
u
o
n
g
p
v
@v
n
u
.
e
d
u
.
v
n
.
Le
Duc
Th
in
h
is
a
lec
tu
re
r
a
t
In
tern
a
ti
o
n
a
l
S
c
h
o
o
l,
Vie
tn
a
m
Na
ti
o
n
a
l
U
n
iv
e
rsit
y
,
Ha
No
i,
Vie
t
n
a
m
.
He
tea
c
h
e
s
m
a
th
e
m
a
ti
c
s
in
E
n
g
li
s
h
f
o
r
b
u
si
n
e
ss
m
a
jo
rs.
He
o
b
tain
e
d
h
is
Ba
c
h
e
lo
r
d
e
g
re
e
in
2
0
0
1
a
n
d
M
a
ste
r
d
e
g
re
e
in
2
0
0
4
,
b
o
t
h
in
m
a
th
e
m
a
ti
c
s
a
t
Ha
No
i
Na
ti
o
n
a
l
Un
iv
e
rsity
o
f
Ed
u
c
a
ti
o
n
,
Vie
tn
a
m
.
He
o
b
tain
e
d
h
is
P
h
.
D
.
d
e
g
re
e
in
2
0
1
2
a
t
th
e
P
e
n
n
sy
l
v
a
n
ia
S
tate
Un
iv
e
rsity
,
USA
.
His
re
se
a
r
c
h
in
tere
sts
in
c
lu
d
e
m
a
c
h
in
e
lea
rn
in
g
,
m
a
th
e
m
a
ti
c
a
l
fin
a
n
c
e
,
e
c
o
n
o
m
ic sta
t
isti
c
s.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
th
i
n
h
ld
@
v
n
u
.
e
d
u
.
v
n
.
Tr
a
n
Th
i
O
a
n
h
g
o
t
th
e
b
a
c
h
e
lo
r
a
n
d
m
a
ste
r
d
e
g
re
e
s
i
n
c
o
m
p
u
ter
sc
ien
c
e
a
t
t
h
e
Un
iv
e
rsity
o
f
E
n
g
i
n
e
e
rin
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
Vie
t
n
a
m
Na
ti
o
n
a
l
Un
iv
e
rsity
,
Ha
n
o
i
i
n
2
0
0
6
a
n
d
2
0
0
9
,
re
sp
e
c
ti
v
e
l
y
.
S
h
e
wa
s
a
wa
rd
e
d
a
Ja
p
a
n
e
se
G
o
v
e
rn
m
e
n
t
S
c
h
o
lars
h
ip
t
o
p
u
rsu
e
P
h
.
D.
i
n
Co
m
p
u
ter
S
c
ien
c
e
a
t
Ja
p
a
n
Ad
v
a
n
c
e
d
In
stit
u
te
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
lo
g
y
(JA
IS
T)
fr
o
m
2
0
1
1
to
2
0
1
4
.
C
u
rre
n
tl
y
,
s
h
e
is
a
l
e
c
tu
re
r
a
t
t
h
e
In
tern
a
ti
o
n
a
l
S
c
h
o
o
l
o
f
Vie
t
n
a
m
Na
ti
o
n
a
l
Un
iv
e
rsity
,
Ha
n
o
i
(VN
U
-
IS
).
He
r
m
a
in
re
se
a
rc
h
in
tere
sts
a
r
e
a
rti
ficia
l
in
telli
g
e
n
c
e
a
n
d
m
a
c
h
in
e
lea
rn
in
g
.
He
r
c
o
n
tri
b
u
ti
o
n
s
to
th
e
field
in
c
lu
d
e
5
0
p
u
b
li
c
a
ti
o
n
s
in
e
ste
e
m
e
d
jo
u
r
n
a
ls
a
n
d
c
o
n
fe
re
n
c
e
s.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
o
a
n
h
tt
@g
m
a
il
.
c
o
m
o
r
tran
th
i
o
a
n
h
@
v
n
u
.
e
d
u
.
v
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:
2
5
0
2
-
4
7
52
A
n
o
ve
l a
p
p
r
o
a
c
h
fo
r
d
etec
tin
g
d
ia
b
etic
r
etin
o
p
a
th
y
u
s
in
g
t
w
o
-
s
tr
ea
m
C
N
N
s
mo
d
el
(
P
h
a
m
Th
i V
iet
Hu
o
n
g
)
209
Mr.
Tr
a
n
Xu
a
n
B
a
c
h
o
b
t
a
in
e
d
h
is
B.
S
c
.
in
th
e
S
c
h
o
o
l
o
f
El
e
c
tri
c
a
l
a
n
d
El
e
c
tro
n
ic
En
g
i
n
e
e
rin
g
,
Ha
n
o
i
Un
iv
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
,
Ha
n
o
i,
Vie
tn
a
m
,
in
2
0
2
1
.
His
m
a
in
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
m
e
d
ica
l
d
a
ta
a
n
a
ly
sis
a
n
d
c
las
sifica
ti
o
n
,
a
s
we
ll
a
s
re
se
a
rc
h
a
n
d
d
e
v
e
lo
p
m
e
n
t
o
f
a
p
p
li
c
a
ti
o
n
s
in
b
io
m
e
d
ica
l
e
n
g
in
e
e
ri
n
g
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
b
a
c
h
tran
x
b
@g
m
a
il
.
c
o
m
.
H
o
a
n
g
Q
u
a
n
g
H
u
y
is
a
lec
tu
re
r
o
f
S
c
h
o
o
l
o
f
E
lec
tri
c
a
l
a
n
d
El
e
c
tro
n
ic
En
g
i
n
e
e
rin
g
,
Ha
n
o
i
Un
i
v
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
,
Ha
n
o
i,
Vie
tn
a
m
.
He
e
a
rn
e
d
h
is
M
.
S
.
d
e
g
re
e
in
e
lec
tro
n
ics
a
n
d
t
e
lec
o
m
m
u
n
ica
ti
o
n
s
fr
o
m
Ha
n
o
i
Un
iv
e
rsity
o
f
S
c
ie
n
c
e
a
n
d
Tec
h
n
o
l
o
g
y
(Vie
tn
a
m
)
i
n
2
0
0
6
,
a
n
d
h
is
B.
S
.
d
e
g
re
e
i
n
e
lec
tro
n
i
c
s
a
n
d
tele
c
o
m
m
u
n
ica
ti
o
n
s
fro
m
t
h
e
sa
m
e
u
n
iv
e
rsit
y
i
n
2
0
0
2
.
His
m
a
in
re
se
a
rc
h
in
tere
sts
i
n
c
lu
d
e
m
e
d
ica
l
d
a
ta
a
n
a
ly
sis
a
n
d
c
las
sifica
ti
o
n
,
re
se
a
rc
h
a
n
d
d
e
v
e
lo
p
m
e
n
t
o
f
a
p
p
li
c
a
ti
o
n
s
f
o
r
re
h
a
b
il
it
a
ti
o
n
,
sm
a
rt
h
e
a
lt
h
,
a
n
d
h
o
sp
i
tal
in
f
o
rm
a
ti
o
n
sy
ste
m
s.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
h
u
y
.
h
o
a
n
g
q
u
a
n
g
@
h
u
st.ed
u
.
v
n
.
Dr
.
Tr
a
n
Anh
Vu
is
a
se
n
io
r
lec
tu
re
r
a
t
S
c
h
o
o
l
o
f
El
e
c
tri
c
a
l
a
n
d
El
e
c
tro
n
ic
En
g
i
n
e
e
rin
g
,
Ha
n
o
i
Un
i
v
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
,
Ha
n
o
i,
Vie
tn
a
m
.
He
e
a
rn
e
d
h
is
P
h
.
D.
d
e
g
re
e
in
e
lec
tri
c
a
l
e
n
g
in
e
e
rin
g
fro
m
th
e
Un
i
v
e
rsity
o
f
M
a
ss
a
c
h
u
se
tt
s
a
t
Lo
we
ll
(USA)
in
2
0
1
4
,
M
S
d
e
g
r
e
e
i
n
b
i
o
m
e
d
i
c
a
l
e
n
g
in
e
e
rin
g
fro
m
T
u
fts
Un
i
v
e
rsity
(USA)
in
2
0
1
0
,
M
S
d
e
g
re
e
in
e
lec
tro
n
ics
a
n
d
tele
c
o
m
m
u
n
ica
ti
o
n
s
fro
m
Ha
n
o
i
Un
iv
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
(Vie
t
n
a
m
)
in
2
0
0
2
,
a
n
d
B
.
S
.
d
e
g
re
e
i
n
e
lec
tro
n
ics
a
n
d
tele
c
o
m
m
u
n
ica
ti
o
n
s
fro
m
Ha
n
o
i
Un
i
v
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
(Vie
tn
a
m
)
in
2
0
0
0
.
His
m
a
in
re
se
a
r
c
h
in
tere
sts
in
c
lu
d
e
th
e
m
e
d
ica
l
d
a
ta
a
n
a
l
y
sis a
n
d
c
las
sifica
ti
o
n
,
re
se
a
rc
h
a
n
d
d
e
v
e
lo
p
m
e
n
t
a
p
p
li
c
a
ti
o
n
s
f
o
r
re
h
a
b
il
it
a
ti
o
n
,
sm
a
rt
h
e
a
lt
h
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
v
u
.
tra
n
a
n
h
@h
u
st
.
e
d
u
.
v
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.