I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
15
,
No
.
4
,
A
u
g
u
s
t
20
25
,
p
p
.
3
8
7
6
~
3
8
8
7
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
4
.
pp
3
8
7
6
-
388
7
3876
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
O
ptimi
zing
co
nv
o
lutiona
l neural n
e
twork
hy
per
pa
ra
meters
to
enha
nce liver
seg
menta
tion a
ccura
cy
in
medica
l i
ma
g
ing
I
wa
n P
urna
m
a
1
,
Ag
us
P
er
da
na
Wind
a
rt
o
2
,
So
lik
hu
n
3
1
I
n
f
o
r
mat
i
o
n
T
e
c
h
n
o
l
o
g
y
S
t
u
d
y
P
r
o
g
r
a
m,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
a
n
d
Te
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
si
t
a
s
La
b
u
h
a
n
b
a
t
u
,
R
a
n
t
a
u
p
r
a
p
a
t
,
I
n
d
o
n
e
si
a
2
M
a
s
t
e
r
's Pr
o
g
r
a
m,
I
n
f
o
r
m
a
t
i
c
s S
t
u
d
y
P
r
o
g
r
a
m
,
S
TI
K
O
M
T
u
n
a
s
B
a
n
g
sa,
P
e
mat
a
n
g
si
a
n
t
a
r
,
I
n
d
o
n
e
si
a
3
I
n
f
o
r
mat
i
c
s E
n
g
i
n
e
e
r
i
n
g
P
r
o
g
r
a
m,
S
TI
K
O
M
T
u
n
a
s
B
a
n
g
sa,
P
e
ma
t
a
n
g
si
a
n
t
a
r
,
I
n
d
o
n
e
s
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
n
1
6
,
2
0
2
4
R
ev
is
ed
Ma
r
2
1
,
2
0
2
5
Acc
ep
ted
Ma
r
2
3
,
2
0
2
5
Li
v
e
r
se
g
m
e
n
tati
o
n
in
m
e
d
ica
l
i
m
a
g
in
g
is
a
c
ru
c
ial
ste
p
in
v
a
ri
o
u
s
c
li
n
ica
l
a
p
p
li
c
a
ti
o
n
s,
su
c
h
a
s
d
ise
a
se
d
iag
n
o
sis,
s
u
rg
ica
l
p
lan
n
in
g
,
a
n
d
e
v
a
l
u
a
ti
o
n
o
f
re
sp
o
n
se
t
o
t
h
e
ra
p
y
,
wh
ic
h
re
q
u
i
re
a
h
ig
h
d
e
g
re
e
o
f
p
re
c
isio
n
f
o
r
a
c
c
u
ra
te
re
su
lt
s.
T
h
is
re
se
a
rc
h
f
o
c
u
se
s
o
n
in
c
re
a
sin
g
t
h
e
a
c
c
u
ra
c
y
o
f
li
v
e
r
se
g
m
e
n
tatio
n
b
y
o
p
ti
m
izi
n
g
h
y
p
e
rp
a
ra
m
e
ters
in
th
e
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
two
rk
(CNN
)
m
o
d
e
l
u
si
n
g
th
e
d
e
v
e
lo
p
e
d
Re
s
Ne
t
a
rc
h
it
e
c
tu
re
.
Th
e
u
n
i
q
u
e
n
e
ss
o
f
t
h
is
re
se
a
rc
h
li
e
s
in
th
e
a
p
p
l
ica
ti
o
n
o
f
h
y
p
e
rp
a
ra
m
e
ter
o
p
ti
m
iza
ti
o
n
m
e
th
o
d
s
su
c
h
a
s
ra
n
d
o
m
se
a
rc
h
a
n
d
Ba
y
e
sia
n
o
p
t
imiz
a
ti
o
n
,
wh
ich
a
ll
o
w
b
r
o
a
d
e
r
a
n
d
m
o
r
e
e
fficie
n
t
e
x
p
l
o
ra
ti
o
n
t
h
a
n
c
o
n
v
e
n
t
io
n
a
l
a
p
p
ro
a
c
h
e
s.
Th
e
re
su
lt
s
s
h
o
w
t
h
a
t
t
h
e
De
e
p
Lab
V3
P
lu
s
m
o
d
e
l
(t
h
e
p
r
o
p
o
se
d
m
o
d
e
l)
sig
n
ifi
c
a
n
tl
y
o
u
t
p
e
rfo
r
m
s
th
e
sta
n
d
a
rd
Re
sN
e
t
in
t
h
e
ima
g
e
se
g
m
e
n
tatio
n
tas
k
.
De
e
p
Lab
V
3
P
lu
s
sh
o
ws
e
x
c
e
ll
e
n
t
p
e
rf
o
rm
a
n
c
e
with
a
n
M
Io
U sc
o
re
o
f
0
.
9
6
5
,
a
P
A S
c
o
re
o
f
0
.
9
2
9
,
a
n
d
a
m
e
a
g
e
r
lo
ss
v
a
l
u
e
o
f
0
.
0
1
1
.
Th
e
se
re
su
lt
s
sh
o
w
t
h
a
t
De
e
p
La
b
V3
P
l
u
s
is
a
b
le
to
re
c
o
g
n
ize
a
n
d
p
re
d
ict
se
g
m
e
n
tatio
n
a
re
a
s
v
e
ry
a
c
c
u
ra
tely
a
n
d
c
o
n
siste
n
tl
y
a
n
d
m
in
imiz
e
p
re
d
ictio
n
e
rro
rs
e
ffe
c
ti
v
e
ly
.
In
c
o
n
c
l
u
sio
n
,
th
e
re
su
lt
s
o
f
th
is
stu
d
y
s
h
o
w
a
sig
n
ifi
c
a
n
t
imp
ro
v
e
m
e
n
t
i
n
se
g
m
e
n
tatio
n
a
c
c
u
ra
c
y
,
with
t
h
e
o
p
ti
m
ize
d
m
o
d
e
l
p
ro
v
id
i
n
g
b
e
tt
e
r
p
e
rfo
rm
a
n
c
e
in
th
e
e
v
a
lu
a
ti
o
n
.
K
ey
w
o
r
d
s
:
C
o
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
Dee
p
L
ab
V3
Plu
s
Hy
p
er
p
ar
a
m
eter
o
p
tim
izatio
n
L
iv
er
s
eg
m
en
tatio
n
R
esNet
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
:
I
wan
Pu
r
n
am
a
I
n
f
o
r
m
atio
n
T
ec
h
n
o
lo
g
y
Stu
d
y
Pro
g
r
am
,
Facu
lty
o
f
Scien
ce
an
d
T
ec
h
n
o
lo
g
y
,
Un
i
v
er
s
itas
L
ab
u
h
an
b
atu
Stre
et
Sis
in
g
am
an
g
ar
aja
No
.
1
2
6
A
KM
3
.
5
Aek
T
ap
a
,
B
ak
a
r
an
B
atu
,
Su
b
d
is
tr
ict
R
an
tau
Sel.,
R
eg
en
cy
L
ab
u
h
an
b
atu
,
No
r
th
Su
m
atr
a
2
1
4
1
8
,
I
n
d
o
n
esia
E
m
ai
l
: iwa
n
p
u
r
n
am
a
2
0
1
4
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
L
i
v
e
r
s
e
g
m
e
n
t
a
ti
o
n
i
n
m
e
d
i
c
a
l
i
m
a
g
i
n
g
p
l
a
y
s
a
v
i
t
a
l
r
o
l
e
in
v
a
r
i
o
u
s
c
l
i
n
i
c
a
l
a
p
p
l
i
c
at
i
o
n
s
,
i
n
c
l
u
d
i
n
g
d
i
a
g
n
o
s
i
s
o
f
l
i
v
e
r
d
is
e
as
e
s
,
s
u
r
g
i
c
a
l
p
l
a
n
n
i
n
g
,
a
n
d
m
o
n
i
t
o
r
i
n
g
r
e
s
p
o
n
s
e
t
o
t
h
e
r
a
p
y
[
1
]
–
[
3
]
.
A
c
c
u
r
a
t
e
s
e
g
m
e
n
t
a
ti
o
n
o
f
t
h
i
s
o
r
g
a
n
i
s
c
r
i
t
i
c
al
b
e
c
a
u
s
e
e
r
r
o
r
s
i
n
l
i
v
e
r
id
e
n
t
i
f
i
c
a
ti
o
n
a
n
d
d
e
l
i
n
e
a
t
i
o
n
c
a
n
r
e
s
u
l
t
i
n
i
n
c
o
r
r
e
ct
d
i
a
g
n
o
s
i
s
o
r
i
n
a
p
p
r
o
p
r
i
a
t
e
s
u
r
g
i
c
a
l
p
la
n
n
i
n
g
,
w
h
i
c
h
i
n
t
u
r
n
c
an
h
a
r
m
t
h
e
p
a
t
ie
n
t
.
T
h
e
r
e
f
o
r
e
,
r
e
l
i
a
b
l
e
a
n
d
a
c
c
u
r
a
te
s
e
g
m
e
n
t
a
ti
o
n
m
e
t
h
o
d
s
a
r
e
n
e
ed
e
d
t
o
s
u
p
p
o
r
t
a
p
p
r
o
p
r
i
a
t
e
c
l
i
n
ic
a
l
d
e
c
is
i
o
n
s
[
4
]
,
[
5
]
.
I
n
r
e
c
e
n
t
y
e
a
r
s
,
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
NN
)
h
a
v
e
e
m
e
r
g
e
d
a
s
a
p
o
t
e
n
t
t
o
o
l
i
n
m
e
d
i
c
a
l
i
m
a
g
e
s
e
g
m
e
n
t
a
ti
o
n
t
as
k
s
[
6
]
.
C
N
Ns
a
r
e
c
a
p
a
b
l
e
o
f
c
a
p
t
u
r
i
n
g
co
m
p
l
e
x
f
e
a
t
u
r
e
s
f
r
o
m
m
e
d
i
c
al
i
m
a
g
es
a
n
d
h
a
v
e
d
e
m
o
n
s
t
r
a
t
e
d
e
x
c
el
l
e
n
t
p
e
r
f
o
r
m
a
n
c
e
i
n
v
a
r
i
o
u
s
s
e
g
m
e
n
t
at
i
o
n
t
a
s
k
s
.
O
n
e
o
f
t
h
e
f
a
m
o
u
s
an
d
i
n
f
l
u
e
n
t
i
a
l
C
NN
a
r
c
h
i
t
e
c
t
u
r
es
f
o
r
m
e
d
i
c
a
l
i
m
a
g
e
s
e
g
m
e
n
t
a
t
i
o
n
is
R
es
N
e
t
,
w
h
ic
h
i
s
s
p
e
c
i
f
i
ca
l
l
y
d
e
s
i
g
n
e
d
t
o
a
d
d
r
e
s
s
c
h
a
l
l
e
n
g
es
i
n
i
m
a
g
e
s
e
g
m
e
n
ta
t
i
o
n
w
it
h
h
i
g
h
r
e
s
o
l
u
t
i
o
n
a
n
d
c
o
m
p
l
e
x
d
e
t
ai
ls
[
7
]
–
[
1
0
]
.
P
r
e
v
i
o
u
s
l
it
e
r
a
t
u
r
e
s
t
u
d
i
es
c
o
n
d
u
c
t
e
d
b
y
[
1
1
]
i
n
t
h
i
s
s
t
u
d
y
s
h
o
w
e
d
t
h
a
t
t
h
e
t
h
r
e
e
C
NN
s
ac
h
i
e
v
e
d
a
n
a
v
e
r
a
g
e
D
i
c
e
s
c
o
r
e
a
b
o
v
e
9
0
%
i
n
l
i
v
e
r
s
e
g
m
e
n
t
a
ti
o
n
w
i
t
h
c
o
n
t
r
a
s
t
c
o
m
p
u
t
e
d
t
o
m
o
g
r
a
p
h
y
(
C
T
)
i
m
a
g
e
s
o
f
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
Op
timiz
in
g
co
n
vo
lu
tio
n
a
l
n
e
u
r
a
l n
etw
o
r
k
h
yp
erp
a
r
a
mete
r
s
to
en
h
a
n
ce
liver
…
(
I
w
a
n
P
u
r
n
a
ma
)
3877
o
r
d
i
n
a
r
y
q
u
a
l
i
t
y
.
H
o
w
e
v
e
r
,
t
h
e
r
e
s
u
l
ts
a
ls
o
s
h
o
w
e
d
t
h
at
t
h
e
t
h
r
e
e
C
NN
s
h
a
d
d
e
c
r
e
as
e
d
p
e
r
f
o
r
m
a
n
c
e
i
n
l
i
v
e
r
s
e
g
m
e
n
t
a
ti
o
n
i
n
C
T
i
m
a
g
e
s
wit
h
l
o
w
d
o
s
e
s
a
n
d
n
o
c
o
n
t
r
a
s
t
.
T
h
i
s
s
u
g
g
e
s
t
s
t
h
a
t
li
v
e
r
s
e
g
m
en
t
a
t
i
o
n
o
n
l
o
w
-
d
o
s
e
,
no
-
c
o
n
t
r
a
s
t
C
T
i
m
a
g
e
s
i
s
s
ti
l
l
a
c
h
a
l
l
e
n
g
i
n
g
t
as
k
a
n
d
r
e
q
u
i
r
e
s
f
u
r
t
h
e
r
i
m
p
r
o
v
e
m
e
n
t
b
e
f
o
r
e
i
t
c
a
n
b
e
a
p
p
l
i
e
d
in
c
l
i
n
i
c
al
p
r
a
c
t
i
c
e
.
M
e
a
n
w
h
i
l
e
,
in
t
h
e
f
o
l
l
o
w
i
n
g
s
t
u
d
y
c
o
n
d
u
c
t
e
d
b
y
[
1
2
]
,
t
h
e
r
e
s
u
l
t
s
o
f
r
e
s
e
a
r
c
h
o
n
l
i
v
e
r
t
u
m
o
r
s
e
g
m
e
n
t
a
ti
o
n
w
e
r
e
b
a
s
e
d
o
n
3
D
C
NN
w
i
t
h
m
u
lt
i
p
l
e
s
c
a
l
es
.
T
h
i
s
s
t
u
d
y
a
i
m
s
t
o
o
v
e
r
c
o
m
e
t
h
e
d
i
f
f
i
c
u
l
t
i
es
o
f
l
i
v
e
r
t
u
m
o
r
s
e
g
m
e
n
t
a
ti
o
n
f
r
o
m
C
T
i
m
a
g
e
s
,
w
h
i
c
h
i
n
cl
u
d
e
l
o
w
c
o
n
t
r
a
s
t
b
et
w
e
e
n
l
i
v
e
r
t
u
m
o
r
s
a
n
d
h
e
a
l
t
h
y
t
is
s
u
e
,
a
s
w
e
l
l
a
s
t
h
e
c
o
m
p
l
e
x
s
h
a
p
e
,
s
iz
e
,
a
n
d
l
o
c
a
ti
o
n
o
f
l
i
v
e
r
t
u
m
o
r
i
m
a
g
e
s
.
T
h
e
m
e
t
h
o
d
d
e
v
e
l
o
p
e
d
i
s
a
C
N
N
-
b
as
e
d
l
i
v
e
r
a
n
d
l
i
v
e
r
t
u
m
o
r
s
e
g
m
en
t
a
t
i
o
n
a
l
g
o
r
i
t
h
m
,
w
h
i
c
h
is
eq
u
i
p
p
e
d
w
i
t
h
a
t
h
r
e
e
-
d
i
m
e
n
s
io
n
a
l
d
u
a
l
p
a
t
h
w
a
y
(
T
D
P
-
C
N
N
)
t
o
m
a
i
n
t
ai
n
a
b
a
l
a
n
c
e
o
f
s
e
g
m
e
n
t
a
ti
o
n
p
e
r
f
o
r
m
a
n
c
e
a
n
d
c
o
m
p
u
t
i
n
g
r
e
s
o
u
r
c
e
r
e
q
u
i
r
e
m
e
n
t
s
.
E
x
p
e
r
i
m
e
n
t
a
l
r
es
u
lt
s
s
h
o
w
t
h
at
t
h
e
p
r
o
p
o
s
e
d
a
l
g
o
r
it
h
m
p
e
r
f
o
r
m
s
w
el
l
i
n
li
v
e
r
a
n
d
l
i
v
e
r
tu
m
o
r
s
e
g
m
e
n
t
a
ti
o
n
.
H
o
w
e
v
e
r
,
s
o
m
e
i
m
p
r
o
v
e
m
e
n
t
s
c
a
n
b
e
m
a
d
e
,
e
s
p
e
c
i
a
l
l
y
i
n
t
h
e
s
e
g
m
e
n
t
a
ti
o
n
o
f
s
m
a
l
l
l
i
v
e
r
tu
m
o
r
s
.
I
n
a
d
d
i
t
i
o
n
,
t
h
e
u
s
e
o
f
f
u
l
l
y
c
o
n
n
e
c
t
e
d
c
o
n
d
i
t
i
o
n
a
l
r
a
n
d
o
m
f
i
e
l
d
s
(
FC
-
C
R
F
)
h
e
l
p
s
i
m
p
r
o
v
e
l
i
v
e
r
t
u
m
o
r
s
e
g
m
e
n
t
a
t
i
o
n
r
e
s
u
lt
s
.
T
h
i
s
m
e
t
h
o
d
s
h
o
w
e
d
b
e
tt
e
r
r
e
s
u
l
t
s
t
h
a
n
s
e
v
e
r
al
o
t
h
e
r
li
v
e
r
a
n
d
l
i
v
e
r
t
u
m
o
r
s
e
g
m
e
n
t
a
t
i
o
n
al
g
o
r
i
t
h
m
s
i
n
t
h
e
2
0
1
7
M
I
C
C
A
I
c
o
m
p
e
ti
t
i
o
n
.
F
r
o
m
t
h
e
p
r
e
v
i
o
u
s
r
e
s
e
a
r
c
h
t
h
a
t
h
a
s
b
e
e
n
d
e
s
c
r
i
b
e
d
,
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
t
h
e
C
N
N
m
o
d
e
l
i
s
v
e
r
y
d
e
p
e
n
d
e
n
t
o
n
c
h
o
o
s
i
n
g
t
h
e
c
o
r
r
e
c
t
h
y
p
e
r
p
a
r
a
m
e
t
e
r
s
,
s
u
c
h
as
l
e
a
r
n
i
n
g
r
at
e
,
b
at
c
h
s
i
z
e
,
n
u
m
b
e
r
o
f
f
i
l
te
r
s
,
a
n
d
k
e
r
n
e
l
s
i
z
e
[
1
3
]
,
[
1
4
]
.
H
y
p
e
r
p
a
r
a
m
e
t
e
r
s
t
h
a
t
a
r
e
n
o
t
o
p
t
i
m
al
c
a
n
c
a
u
s
e
t
h
e
m
o
d
e
l
t
o
e
x
p
e
r
i
e
n
c
e
o
v
e
r
f
i
t
ti
n
g
o
r
u
n
d
e
r
f
i
t
t
i
n
g
,
t
h
e
r
e
b
y
r
e
d
u
c
i
n
g
s
e
g
m
e
n
t
a
ti
o
n
a
c
c
u
r
a
c
y
[
1
5
]
,
[
1
6
]
.
T
h
e
r
e
f
o
r
e
,
h
y
p
e
r
p
a
r
a
m
e
t
e
r
o
p
t
i
m
i
z
a
ti
o
n
i
s
a
c
r
u
c
i
a
l
s
t
e
p
i
n
d
e
v
e
l
o
p
i
n
g
m
o
r
e
a
c
c
u
r
a
t
e
a
n
d
r
e
l
i
a
b
l
e
C
N
N
m
o
d
e
l
s
.
H
y
p
e
r
p
a
r
a
m
et
e
r
o
p
t
i
m
i
za
t
i
o
n
m
e
t
h
o
d
s
s
u
c
h
a
s
g
r
i
d
s
e
a
r
c
h
[
1
7
]
–
[
1
9
]
,
a
n
d
B
a
y
es
i
a
n
o
p
t
i
m
i
z
a
ti
o
n
h
a
v
e
b
e
e
n
p
r
o
p
o
s
e
d
t
o
f
i
n
d
o
p
t
im
a
l
h
y
p
e
r
p
a
r
a
m
e
t
e
r
c
o
m
b
i
n
a
t
i
o
n
s
[
2
0
]
–
[
2
3
]
.
G
r
i
d
s
e
a
r
c
h
e
x
p
l
o
r
e
s
t
h
e
e
n
t
i
r
e
h
y
p
e
r
p
a
r
a
m
e
t
e
r
s
p
a
c
e
s
y
s
t
e
m
a
t
i
c
al
l
y
b
u
t
o
f
t
e
n
r
e
q
u
i
r
e
s
a
v
e
r
y
l
o
n
g
t
i
m
e
a
n
d
s
i
g
n
i
f
ic
a
n
t
c
o
m
p
u
t
a
t
i
o
n
a
l
r
es
o
u
r
c
e
s
.
I
n
co
n
t
r
a
s
t
,
r
a
n
d
o
m
s
e
a
r
c
h
a
n
d
B
ay
e
s
i
a
n
o
p
t
i
m
i
z
a
t
i
o
n
o
f
f
e
r
a
m
o
r
e
e
f
f
i
c
i
e
n
t
a
p
p
r
o
a
c
h
b
y
r
a
n
d
o
m
l
y
e
x
p
l
o
r
i
n
g
t
h
e
h
y
p
e
r
p
a
r
a
m
e
t
e
r
s
p
a
c
e
o
r
u
s
i
n
g
p
r
o
b
a
b
i
l
i
s
ti
c
m
o
d
e
l
s
t
o
p
r
e
d
i
c
t
t
h
e
m
o
s
t
p
r
o
m
i
s
i
n
g
h
y
p
e
r
p
a
r
a
m
e
t
e
r
c
o
m
b
i
n
a
t
i
o
n
s
[
2
4
]
.
B
as
e
d
o
n
t
h
e
l
i
t
e
r
a
t
u
r
e
s
t
u
d
i
e
s
t
h
a
t
h
a
v
e
b
e
e
n
d
e
s
c
r
i
b
e
d
,
t
h
i
s
r
e
s
e
a
r
c
h
f
o
c
u
s
es
o
n
t
h
e
a
p
p
l
i
c
a
t
i
o
n
o
f
h
y
p
e
r
p
a
r
a
m
e
t
e
r
o
p
t
i
m
i
z
a
t
i
o
n
m
e
t
h
o
d
s
t
o
i
n
c
r
e
a
s
e
t
h
e
a
c
c
u
r
ac
y
o
f
l
i
v
e
r
s
e
g
m
e
n
t
a
t
i
o
n
u
s
i
n
g
a
C
N
N
m
o
d
e
l
w
i
t
h
R
es
n
e
t
a
r
c
h
i
t
e
c
t
u
r
e
[
2
5
]
–
[
2
8
]
.
B
y
u
s
i
n
g
r
a
n
d
o
m
s
e
a
r
c
h
a
n
d
B
a
y
e
s
i
a
n
o
p
t
i
m
i
z
a
ti
o
n
m
e
t
h
o
d
s
,
t
h
is
r
es
e
a
r
c
h
a
i
m
s
t
o
f
i
n
d
t
h
e
o
p
t
i
m
a
l
c
o
m
b
i
n
a
t
io
n
o
f
h
y
p
e
r
p
a
r
a
m
e
t
e
r
s
t
h
a
t
c
an
s
i
g
n
i
f
i
c
a
n
tl
y
i
m
p
r
o
v
e
m
o
d
e
l
p
e
r
f
o
r
m
a
n
c
e
.
T
h
e
r
e
s
u
l
ts
o
f
t
h
is
r
e
s
e
a
r
c
h
a
r
e
e
x
p
e
c
t
e
d
t
o
m
a
k
e
a
s
i
g
n
i
f
ica
n
t
c
o
n
t
r
i
b
u
t
i
o
n
t
o
t
h
e
f
i
e
l
d
o
f
m
e
d
i
c
a
l
i
m
a
g
e
s
e
g
m
e
n
t
a
ti
o
n
a
n
d
i
m
p
r
o
v
e
c
l
i
n
i
c
a
l
a
p
p
l
i
c
at
i
o
n
s
i
n
d
i
a
g
n
o
s
i
s
a
n
d
p
a
t
i
e
n
t
t
r
e
a
t
m
e
n
t
p
l
a
n
n
in
g
[
1
2
]
,
[
2
9
]
,
[
3
0
]
.
E
f
f
e
c
t
i
v
e
h
y
p
e
r
p
a
r
a
m
e
t
e
r
o
p
t
im
i
z
a
t
i
o
n
i
m
p
r
o
v
e
s
s
e
g
m
e
n
t
a
tio
n
a
c
c
u
r
a
c
y
a
n
d
t
h
e
m
o
d
el'
s
ab
i
l
i
t
y
t
o
g
e
n
e
r
a
l
iz
e
t
o
n
e
w
d
a
t
a
,
t
h
e
r
e
b
y
e
n
s
u
r
i
n
g
i
t
s
u
s
e
i
n
a
v
a
r
i
e
t
y
o
f
c
l
i
n
i
c
a
l
c
o
n
d
i
t
i
o
n
s
[
3
1
]
,
[
3
2
]
.
T
h
u
s
,
t
h
i
s
r
e
s
e
a
r
c
h
f
o
c
u
s
e
s
n
o
t
o
n
l
y
o
n
i
m
p
r
o
v
i
n
g
t
h
e
m
o
d
e
l
'
s
p
e
r
f
o
r
m
a
n
c
e
b
u
t
a
l
s
o
o
n
i
t
s
p
r
a
c
t
i
c
al
a
p
p
l
i
c
a
b
il
i
t
y
i
n
r
e
a
l
m
ed
i
c
a
l
e
n
v
i
r
o
n
m
e
n
t
s
[
3
3
]
.
2.
M
E
T
H
O
D
T
h
is
r
esear
ch
was
co
n
d
u
cte
d
to
im
p
r
o
v
e
t
h
e
ac
cu
r
ac
y
o
f
li
v
er
s
eg
m
en
tatio
n
in
m
ed
ical
i
m
ag
es
b
y
an
aly
zin
g
an
d
o
p
tim
izin
g
C
NN
h
y
p
er
p
ar
am
eter
s
.
T
h
e
s
eg
m
en
tatio
n
r
esu
lts
o
b
tain
ed
p
r
o
v
id
e
v
alu
ab
le
in
s
ig
h
ts
in
to
th
e
ef
f
ec
tiv
en
ess
o
f
v
ar
io
u
s
C
NN
m
o
d
el
o
p
tim
izatio
n
tech
n
iq
u
es.
T
h
ese
f
in
d
in
g
s
ca
n
im
p
r
o
v
e
o
u
r
u
n
d
er
s
tan
d
i
n
g
o
f
h
o
w
to
ef
f
e
ctiv
ely
tu
n
e
C
NN
h
y
p
er
p
ar
a
m
eter
s
to
im
p
r
o
v
e
s
eg
m
en
tat
io
n
ac
cu
r
ac
y
in
th
e
co
n
tex
t o
f
m
ed
ical
im
ag
es.
2
.
1
.
Da
t
a
c
o
llect
io
n m
et
ho
d
I
n
an
ef
f
o
r
t
to
o
b
tain
r
elev
an
t
d
ata
an
d
in
f
o
r
m
atio
n
,
th
e
m
eth
o
d
u
s
ed
in
d
ata
co
llectio
n
is
o
p
en
-
s
o
u
r
ce
d
at
a,
wh
ich
is
av
ailab
le
o
n
th
e
K
a
g
g
le.
co
m
web
s
ite.
T
h
e
av
ailab
ilit
y
o
f
th
es
e
r
eso
u
r
ce
s
allo
ws
r
esear
ch
er
s
to
ex
p
l
o
r
e
r
elev
a
n
t
d
ata
an
d
b
u
ild
r
esear
ch
s
e
r
ies
th
at
s
u
it
th
eir
n
ee
d
s
.
T
h
e
d
ata
s
o
u
r
ce
f
r
o
m
K
a
g
g
le.
co
m
is
o
n
e
o
f
th
e
k
ey
asp
ec
ts
in
o
b
tain
in
g
a
q
u
ality
d
ataset
f
o
r
th
is
r
esear
ch
.
2
.
2
.
F
l
o
wcha
rt
o
f
CNN
m
o
d
el
o
ptim
iza
t
io
n
I
n
th
is
r
esear
ch
,
C
NN
wer
e
tr
ain
ed
u
s
in
g
d
if
f
e
r
en
t
o
p
t
im
izer
s
an
d
d
if
f
er
en
t
b
atch
s
izes
f
o
r
co
m
p
ar
is
o
n
.
T
h
e
au
th
o
r
s
ad
o
p
ted
th
e
ap
p
r
o
ac
h
o
f
u
s
in
g
a
ty
p
ical
d
ee
p
lear
n
in
g
f
r
am
ewo
r
k
s
u
ch
as
T
en
s
o
r
Flo
w.
B
y
ex
p
lo
itin
g
t
h
e
ad
v
a
n
tag
es
o
f
th
is
f
r
am
e
wo
r
k
,
th
ey
ca
n
m
an
ag
e
a
n
d
tr
ain
C
NN
m
o
d
els
ef
f
icien
tly
.
T
o
p
r
o
v
id
e
a
clea
r
v
is
u
al
r
ep
r
esen
tatio
n
o
f
th
e
o
p
tim
izatio
n
f
lo
w
o
f
th
e
C
NN
m
o
d
el
u
s
ed
in
th
is
r
esear
ch
,
we
p
r
esen
t a
d
etailed
f
lo
w
d
iag
r
a
m
in
Fig
u
r
e
1
.
F
i
g
u
r
e
1
d
is
p
l
a
y
s
t
h
e
s
t
e
p
s
i
n
v
o
l
v
e
d
i
n
t
h
e
C
N
N
m
o
d
e
l
o
p
t
im
i
z
a
t
i
o
n
p
r
o
c
es
s
,
f
r
o
m
t
h
e
d
a
ta
c
o
l
l
ec
t
i
o
n
s
t
a
g
e
t
o
t
h
e
e
v
a
l
u
a
t
i
o
n
o
f
t
h
e
r
e
s
u
l
ts
.
T
h
i
s
d
i
a
g
r
a
m
h
e
l
p
s
r
e
a
d
e
r
s
e
a
s
il
y
u
n
d
e
r
s
t
a
n
d
t
h
e
s
t
e
p
s
i
n
v
o
l
v
e
d
i
n
t
r
a
i
n
i
n
g
a
n
d
o
p
t
i
m
i
z
i
n
g
a
C
N
N
m
o
d
el
a
n
d
h
o
w
c
h
a
n
g
i
n
g
t
h
e
o
p
ti
m
i
ze
r
a
n
d
b
a
t
c
h
s
i
z
e
a
f
f
e
c
ts
t
h
e
m
o
d
e
l
p
e
r
f
o
r
m
a
n
c
e
.
T
h
u
s
,
i
t
is
a
u
s
e
f
u
l
v
i
s
u
al
g
u
i
d
e
t
o
c
l
a
r
i
f
y
t
h
e
e
x
p
e
r
i
m
e
n
t
a
l
p
r
o
c
e
s
s
a
n
d
f
i
n
d
i
n
g
s
i
n
t
h
is
r
es
e
a
r
c
h
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
4
,
Au
g
u
s
t
20
25
:
3
8
7
6
-
3887
3878
P
r
e
p
r
o
c
e
s
s
i
n
g
I
m
a
ge
E
nha
nc
e
m
e
nt
G
r
a
ys
c
a
l
e
C
ha
ng
e
I
m
a
ge
S
i
z
e
D
a
t
a
A
ugm
e
nt
a
t
i
on
F
l
i
ppi
ng
R
ot
a
t
e
C
h
e
s
t
X
r
a
y
d
a
t
a
s
e
t
S
p
l
i
t
D
a
t
a
T
r
a
i
ni
ng
D
a
t
a
0
.
70
V
a
l
i
da
t
i
on
D
a
t
a
0
.
16
P
r
e
pa
r
i
ng
D
a
t
a
S
e
g
m
e
n
t
a
t
i
o
n
T
r
a
i
ni
ng
D
a
t
a
V
a
l
i
da
t
i
on
D
a
t
a
I
m
a
ge
S
e
gm
e
nt
a
t
i
on
D
a
t
a
s
e
t
R
e
s
N
e
t
D
a
t
a
V
i
s
ua
l
i
z
a
t
i
on
T
r
a
i
ni
ng
D
a
t
a
s
e
t
P
r
op
os
e
d
M
e
t
ho
d
D
e
e
pL
a
bV
3
P
l
us
P
r
e
di
c
t
i
on
r
e
s
ul
t
T
e
s
t
i
ng
d
a
t
a
R
e
s
u
l
t
s
T
e
s
t
i
ng
D
a
t
a
0
.
14
H
yp
e
r
pa
r
a
m
e
t
e
r
D
e
t
e
r
m
i
na
t
i
on
a
nd
H
ype
r
pa
r
a
m
e
t
e
r
O
pt
i
m
i
z
a
t
i
on
E
va
l
ua
t
i
on a
nd
A
na
l
ys
i
s
o
f
R
e
s
ul
t
s
Fig
u
r
e
1
.
Mo
d
el
o
p
tim
izatio
n
f
lo
wch
ar
t
Fig
u
r
e
1
is
th
e
f
lo
wch
a
r
t
a
b
o
v
e
d
e
p
ictin
g
th
e
w
o
r
k
f
l
o
w
in
o
p
tim
izin
g
th
e
C
NN
m
o
d
el.
T
h
e
f
o
llo
win
g
is
a
d
etailed
e
x
p
lan
atio
n
o
f
ea
c
h
s
tep
in
th
e
f
lo
wc
h
ar
t,
s
tar
tin
g
with
th
e
d
ata
co
l
lectio
n
p
r
o
ce
s
s
th
at
will
b
e
u
s
ed
f
o
r
m
o
d
el
tr
ai
n
in
g
.
T
h
e
n
ex
t
s
tag
e
is
d
ata
p
r
e
p
r
o
ce
s
s
in
g
,
in
clu
d
in
g
n
o
r
m
aliz
atio
n
,
s
ca
lin
g
,
an
d
r
em
o
v
al
o
f
ir
r
elev
an
t
o
r
m
is
s
in
g
d
ata.
Af
ter
th
at,
d
ata
a
u
g
m
en
tatio
n
is
ca
r
r
ied
o
u
t
to
in
cr
e
ase
an
d
en
r
ich
th
e
d
ataset
with
v
ar
iatio
n
s
s
u
ch
a
s
r
o
tatio
n
,
s
h
if
t,
an
d
m
ir
r
o
r
in
g
.
T
h
e
d
ata
was
th
en
tr
an
s
f
o
r
m
ed
an
d
d
iv
id
e
d
in
t
o
th
r
ee
p
a
r
ts
:
7
0
%
f
o
r
m
o
d
el
tr
ain
in
g
,
1
6
%
f
o
r
m
o
d
el
v
alid
at
io
n
d
u
r
in
g
tr
ain
in
g
to
av
o
id
o
v
er
f
itti
n
g
,
an
d
1
4
%
f
o
r
f
o
llo
w
-
u
p
tr
ai
n
in
g
.
Nex
t,
th
e
C
NN
n
etwo
r
k
ar
ch
itect
u
r
e
th
at
will
b
e
u
s
ed
is
d
esig
n
ed
,
f
o
llo
we
d
b
y
d
eter
m
in
in
g
an
d
o
p
tim
izin
g
t
h
e
m
o
d
el
h
y
p
er
p
a
r
am
eter
s
,
s
u
ch
as
n
u
m
b
er
o
f
lay
e
r
s
,
b
atch
s
ize,
an
d
lear
n
in
g
r
ate.
T
h
e
tr
ain
in
g
d
ataset
is
u
s
ed
to
tr
ain
th
e
C
NN
m
o
d
el;
th
en
,
th
e
m
o
d
el
r
esu
lts
ar
e
ev
alu
ated
an
d
a
n
aly
ze
d
u
s
in
g
v
alid
atio
n
d
ata
to
m
ea
s
u
r
e
m
o
d
el
p
er
f
o
r
m
a
n
ce
.
Op
tio
n
ally
,
ex
ter
n
al
v
alid
atio
n
ca
n
b
e
p
er
f
o
r
m
ed
u
s
in
g
a
d
i
f
f
er
en
t
o
r
a
d
d
itio
n
al
d
ataset
to
en
s
u
r
e
th
e
g
en
e
r
aliza
tio
n
o
f
t
h
e
m
o
d
el.
T
h
e
f
in
al
s
tag
e
is
th
e
d
is
cu
s
s
io
n
an
d
co
n
clu
s
io
n
o
f
th
e
r
esu
lts
,
wh
er
e
th
is
p
r
o
ce
s
s
en
s
u
r
es
t
h
at
th
e
r
esu
ltin
g
m
o
d
el
h
as
o
p
tim
al
an
d
r
eliab
le
p
er
f
o
r
m
an
ce
.
T
h
is
f
lo
wch
a
r
t
s
h
o
ws
a
s
y
s
te
m
atic
ap
p
r
o
ac
h
to
b
u
ild
in
g
an
d
o
p
tim
izin
g
a
C
NN
m
o
d
el,
f
r
o
m
d
ata
co
llectio
n
to
ev
alu
atin
g
an
d
v
alid
atin
g
th
e
f
in
al
r
esu
lt
s
.
T
h
is
co
m
p
r
e
h
en
s
iv
e
ap
p
r
o
a
ch
en
s
u
r
es
th
at
t
h
e
r
esu
ltin
g
m
o
d
el
h
as o
p
tim
al
a
n
d
r
eliab
le
p
e
r
f
o
r
m
an
ce
.
2
.
3
.
O
pti
m
ized
CNN
a
rc
hite
ct
ure
T
h
e
C
NN
ar
ch
itectu
r
e
in
th
is
r
esear
ch
u
s
es
th
e
Dee
p
L
ab
V3
Plu
s
m
o
d
el.
Dee
p
L
ab
V3
Plu
s
is
a
co
m
p
ellin
g
an
d
c
o
m
p
lex
m
o
d
el
u
s
ed
f
o
r
im
ag
e
s
eg
m
en
tat
io
n
.
I
t
co
m
b
in
es
ad
v
a
n
ce
d
f
e
atu
r
es
f
r
o
m
s
ev
er
al
ex
is
tin
g
ar
ch
itectu
r
es,
s
u
ch
as
R
esNet,
an
d
is
eq
u
ip
p
ed
wit
h
s
ev
er
a
l
u
n
iq
u
e
tech
n
iq
u
es
t
o
im
p
r
o
v
e
ac
cu
r
ac
y
an
d
p
er
f
o
r
m
a
n
ce
.
Fig
u
r
e
2
s
h
o
ws th
e
o
p
tim
izatio
n
m
o
d
el.
Fig
u
r
e
2
I
n
g
en
er
al,
th
e
Dee
p
L
ab
V3
Plu
s
m
o
d
el
h
as
th
e
f
o
llo
win
g
ar
ch
itectu
r
e
f
o
r
its
co
n
v
o
lu
tio
n
al
p
ar
t (
C
NN)
:
a.
R
esNet
B
ac
k
b
o
n
e:
T
h
is
p
ar
t
is
th
e
ess
en
tia
l
p
ar
t
t
h
at
p
r
o
v
id
es
in
p
u
t
im
ag
e
f
ea
tu
r
es.
Dee
p
L
ab
V3
Plu
s
u
s
e
s
R
esNet
a
s
a
b
ac
k
b
o
n
e,
w
h
ich
h
as
a
d
ee
p
s
tr
u
ctu
r
e
with
R
esid
u
al
b
lo
ck
s
th
at
e
n
ab
les
b
etter
lear
n
in
g
e
v
en
f
o
r
in
ten
s
e
n
etwo
r
k
s
.
b.
Atr
o
u
s
s
p
atial
p
y
r
am
id
co
llectio
n
(
ASPP
)
:
T
h
is
i
s
an
es
s
en
t
ial
p
ar
t
o
f
th
e
Dee
p
L
ab
m
o
d
e
l.
ASPP
allo
ws
th
e
n
etwo
r
k
t
o
ex
p
a
n
d
its
v
is
u
al
r
an
g
e
with
o
u
t in
cr
ea
s
in
g
th
e
n
u
m
b
e
r
o
f
p
ar
am
eter
s
to
o
m
u
ch
.
T
h
is
allo
ws
th
e
n
etwo
r
k
t
o
o
b
tain
b
r
o
ad
er
co
n
tex
tu
al
in
f
o
r
m
at
io
n
,
wh
ich
is
h
elp
f
u
l in
im
a
g
e
s
eg
m
en
tat
io
n
task
s
.
c.
Dec
o
d
er
(
Dee
p
L
ab
V
3
Plu
s
)
:
On
ce
im
p
o
r
tan
t
f
ea
tu
r
es
ar
e
o
b
tain
ed
f
r
o
m
ASPP
,
th
e
Dec
o
d
er
f
u
n
ctio
n
s
to
r
ef
in
e
an
d
r
ec
o
v
er
d
ata
to
p
r
o
d
u
ce
m
o
r
e
p
r
ec
is
e
s
eg
m
e
n
tatio
n
.
I
n
Dee
p
L
ab
V3
Plu
s
,
t
h
er
e
ar
e
s
ev
er
al
m
ec
h
an
is
m
s
,
s
u
ch
as
s
k
ip
co
n
n
ec
tio
n
s
,
th
at
h
elp
s
tr
en
g
th
en
th
e
r
ep
r
esen
tatio
n
.
Ou
tp
u
t
lay
er
:
Fin
ally
,
th
e
m
o
d
el
p
r
o
d
u
ce
s
o
u
tp
u
t
in
t
h
e
f
o
r
m
o
f
th
e
ex
p
ec
ted
p
i
x
el
s
e
g
m
en
tatio
n
,
wh
ich
co
r
r
esp
o
n
d
s
to
th
e
n
u
m
b
er
o
f
class
es
s
p
ec
if
ied
in
th
e
clas
s
.
Ov
er
all,
Dee
p
L
ab
V3
Plu
s
is
an
in
ten
s
e
an
d
co
m
p
lex
m
o
d
el
th
at
lev
er
ag
es
ad
v
an
ce
d
tech
n
o
l
o
g
ies in
C
NNs f
o
r
im
ag
e
s
eg
m
en
tatio
n
task
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
Op
timiz
in
g
co
n
vo
lu
tio
n
a
l
n
e
u
r
a
l n
etw
o
r
k
h
yp
erp
a
r
a
mete
r
s
to
en
h
a
n
ce
liver
…
(
I
w
a
n
P
u
r
n
a
ma
)
3879
Fig
u
r
e
2
.
Pro
p
o
s
ed
m
o
d
el
o
p
tim
izatio
n
(
Dee
p
L
ab
V
3
Plu
s
)
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
B
ased
o
n
th
e
tr
ain
in
g
r
esu
lts
,
v
alid
atio
n
r
esu
lts
,
an
d
test
in
g
r
esu
lts
,
an
aly
s
is
was
ca
r
r
ie
d
o
u
t
u
s
in
g
v
ar
io
u
s
h
y
p
er
p
ar
a
m
eter
s
an
d
d
if
f
er
en
t
ar
c
h
itectu
r
es
to
s
ee
th
e
im
ag
e
s
eg
m
en
tatio
n
r
e
s
u
lts
in
liv
er
im
ag
e
s
eg
m
en
tatio
n
.
T
h
is
an
aly
s
is
aim
s
to
id
en
tify
th
e
b
est
co
n
f
ig
u
r
atio
n
th
at
p
r
o
d
u
ce
s
th
e
m
o
s
t
ac
cu
r
ate
an
d
co
n
s
is
ten
t
s
eg
m
en
tatio
n
.
B
y
c
o
m
p
ar
in
g
th
e
p
er
f
o
r
m
an
ce
o
f
ea
ch
m
o
d
el
th
r
o
u
g
h
r
elev
a
n
t
ev
alu
atio
n
m
etr
ics,
s
u
ch
as
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
a
n
d
r
ec
all,
it
ca
n
b
e
d
eter
m
in
e
d
wh
ich
m
et
h
o
d
p
r
o
v
id
es
o
p
t
im
al
r
esu
lts
in
th
e
task
o
f
liv
er
im
ag
e
s
eg
m
en
tatio
n
.
T
h
e
r
esu
l
ts
o
f
th
ese
ex
p
er
im
en
ts
will
p
r
o
v
id
e
v
alu
ab
l
e
in
s
ig
h
ts
in
to
th
e
ef
f
ec
tiv
en
ess
o
f
v
ar
i
o
u
s
ap
p
r
o
ac
h
es
in
p
r
o
ce
s
s
in
g
an
d
an
aly
zin
g
m
ed
ical
im
ag
es
an
d
lead
to
im
p
r
o
v
em
en
ts
in
th
e
p
er
f
o
r
m
an
ce
o
f
s
eg
m
e
n
tatio
n
m
o
d
els in
th
e
f
u
tu
r
e.
3
.
1
.
Resul
t
T
h
e
m
o
d
el
tr
ain
in
g
r
esu
lts
in
th
e
ar
ticle
in
clu
d
e
s
o
m
e
im
p
o
r
tan
t
in
f
o
r
m
atio
n
e
x
p
l
ain
in
g
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
o
p
tim
iz
ed
C
NN
m
o
d
el
f
o
r
liv
e
r
s
eg
m
en
tatio
n
i
n
3
D
m
ed
ical
im
ag
es
u
s
in
g
t
h
e
Dee
p
L
ab
V3
Plu
s
m
o
d
el
as
th
e
p
r
o
p
o
s
ed
m
o
d
el
a
n
d
R
esNet.
I
n
Fig
u
r
e
3
(
s
ee
in
ap
p
e
n
d
ix
)
,
th
e
f
o
llo
win
g
is
a
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
4
,
Au
g
u
s
t
20
25
:
3
8
7
6
-
3887
3880
ex
p
lan
atio
n
o
f
th
e
r
esu
lts
f
o
r
ea
ch
ep
o
c
h
:
Yo
u
ca
n
s
ee
th
e
tr
ain
in
g
p
r
o
ce
s
s
p
r
esen
ted
i
n
Fig
u
r
e
3
.
I
n
t
h
e
tr
ain
in
g
p
r
o
ce
s
s
s
h
o
wn
in
Fig
u
r
e
3
,
it
ca
n
b
e
s
ee
n
th
at
af
ter
th
e
8
th
ep
o
c
h
,
tr
ai
n
in
g
is
s
to
p
p
ed
.
T
h
is
is
ca
u
s
ed
b
y
th
e
lo
s
s
v
alu
e
n
o
t
d
ec
r
ea
s
in
g
f
o
r
f
iv
e
co
n
s
ec
u
tiv
e
ep
o
c
h
s
,
wh
ich
in
d
icate
s
th
at
th
e
m
o
d
el
h
as
r
ea
ch
e
d
co
n
v
er
g
en
ce
o
r
t
h
er
e
is
a
p
o
s
s
ib
ilit
y
th
at
th
e
m
o
d
el
is
o
v
er
f
i
ttin
g
.
T
h
e
to
tal
tr
ain
in
g
tim
e
f
o
r
Dee
p
L
ab
V3
Plu
s
is
2
8
0
,
5
9
6
m
in
u
tes,
wh
ile
f
o
r
th
e
o
th
er
m
o
d
els,
it
o
n
ly
s
to
p
s
at
ep
o
ch
5
with
a
tim
e
o
f
9
0
m
in
u
tes.
T
h
is
d
if
f
er
en
ce
in
tr
ai
n
in
g
tim
e
is
q
u
ite
s
ig
n
if
ican
t
an
d
s
h
o
ws
th
e
ex
is
ten
ce
o
f
tim
e
in
ef
f
icien
cies
i
n
Dee
p
L
ab
V3
Plu
s
tr
ain
in
g
,
wh
i
ch
tak
es
m
u
ch
lo
n
g
e
r
.
H
o
wev
er
,
it
h
as
ex
ce
llen
t
p
r
e
d
ictio
n
ac
cu
r
ac
y
r
esu
lts
co
m
p
ar
ed
to
t
h
e
s
tan
d
ar
d
R
esNet
m
o
d
el.
T
im
e
ef
f
icien
c
y
is
an
ess
en
tial
asp
ec
t
o
f
th
e
m
o
d
el
tr
ain
in
g
p
r
o
ce
s
s
,
esp
ec
ially
in
co
n
tex
ts
th
at
r
eq
u
ir
e
s
ig
n
if
ican
t
co
m
p
u
tin
g
r
eso
u
r
ce
s
an
d
a
lo
n
g
tim
e.
E
x
ce
s
s
iv
e,
tim
e
-
co
n
s
u
m
in
g
tr
ain
in
g
n
o
t
o
n
ly
in
cr
ea
s
es
o
p
er
atio
n
al
co
s
ts
b
u
t
ca
n
also
s
lo
w
d
o
wn
th
e
m
o
d
el
d
ev
elo
p
m
en
t
iter
atio
n
cy
cle.
T
h
e
r
ef
o
r
e,
f
in
d
in
g
a
b
alan
ce
b
etwe
en
m
o
d
el
q
u
ality
an
d
tr
ain
in
g
tim
e
ef
f
ici
en
cy
is
cr
u
cial.
Ho
wev
er
,
in
co
n
tr
ast
to
De
ep
L
ab
V3
Plu
s
,
th
e
R
es
N
et
m
o
d
el
o
n
ly
r
e
q
u
ir
es
f
iv
e
ep
o
c
h
s
b
ef
o
r
e
tr
ain
in
g
is
s
to
p
p
ed
.
Ho
wev
er
,
th
e
tr
ain
in
g
r
esu
lts
f
r
o
m
R
es
N
et
co
u
ld
b
e
m
o
r
e
s
atis
f
ac
to
r
y
,
in
d
icatin
g
th
at
th
is
m
o
d
el
h
as
n
o
t
b
ee
n
ab
le
to
p
r
o
d
u
ce
th
e
ex
p
e
cted
p
e
r
f
o
r
m
a
n
ce
in
a
s
h
o
r
ter
tr
ai
n
in
g
tim
e.
T
h
is
s
u
g
g
ests
th
at
alth
o
u
g
h
R
es
N
et
is
m
o
r
e
ef
f
icien
t
in
tr
ai
n
in
g
tim
e,
t
h
e
q
u
ality
o
f
th
e
r
esu
lts
o
b
tain
ed
is
d
if
f
e
r
en
t
f
r
o
m
Dee
p
L
ab
V3
Plu
s
.
I
n
o
r
d
er
to
i
n
cr
ea
s
e
th
e
ef
f
icien
c
y
an
d
e
f
f
e
ctiv
en
ess
o
f
t
h
e
tr
ain
i
n
g
p
r
o
ce
s
s
,
co
n
s
id
er
th
e
u
s
e
o
f
tech
n
iq
u
es
s
u
c
h
as
m
o
r
e
a
d
ap
tiv
e
ea
r
ly
s
to
p
p
in
g
,
th
e
u
s
e
o
f
r
eg
u
lar
izatio
n
to
r
ed
u
ce
o
v
er
f
itti
n
g
,
o
r
ev
e
n
th
e
ex
p
lo
r
atio
n
o
f
m
o
r
e
ef
f
icien
t
b
u
t
s
till
h
ig
h
-
p
er
f
o
r
m
an
ce
m
o
d
el
ar
ch
itectu
r
es.
T
h
r
o
u
g
h
th
is
ap
p
r
o
ac
h
,
an
o
p
tim
al
b
alan
ce
ca
n
b
e
ac
h
ie
v
ed
b
etwe
en
tr
ain
in
g
tim
e
an
d
t
h
e
q
u
ality
o
f
th
e
r
esu
lts
o
b
tain
ed
.
I
n
Fig
u
r
e
4
,
we
ca
n
s
ee
th
e
m
I
OU,
PA
s
co
r
e,
a
n
d
l
o
s
s
cu
r
v
es o
n
Fig
u
r
e
4
(
a)
Dee
p
L
ab
V3
Plu
s
an
d
Fig
u
r
e
4
(
b
)
R
esNet.
W
ell,
h
er
e
is
a
co
m
p
r
eh
en
s
iv
e
d
escr
ip
tio
n
o
f
t
h
e
tr
ain
in
g
r
esu
lts
o
f
two
m
o
d
els,
Dee
p
L
ab
V3
Plu
s
an
d
R
esNet,
b
ased
o
n
th
e
cu
r
v
es sh
o
wn
in
th
e
f
ig
u
r
e.
a.
Dee
p
L
ab
V3
Plu
s
(
a)
−
M
e
a
n
in
ters
e
c
ti
o
n
o
v
e
r
u
n
io
n
(I
o
U) l
e
a
rn
in
g
c
u
rv
e
:
Th
e
trai
n
in
g
c
u
rv
e
sh
o
ws
a
si
g
n
ifi
c
a
n
t
i
n
c
re
a
se
in
th
e
in
i
ti
a
l
e
p
o
c
h
a
n
d
sta
b
il
ize
s
a
fter
th
e
3
r
d
e
p
o
c
h
.
T
h
e
v
a
li
d
a
ti
o
n
c
u
r
v
e
fo
ll
o
ws
th
e
train
i
n
g
c
u
r
v
e
with
li
tt
le d
iffere
n
c
e
,
in
d
ica
ti
n
g
th
a
t
t
h
e
m
o
d
e
l
d
o
e
s n
o
t
s
u
ffe
r
fro
m
sig
n
ifi
c
a
n
t
o
v
e
rfit
ti
n
g
.
Afte
r
th
e
8
th
e
p
o
c
h
,
t
ra
in
in
g
wa
s
sto
p
p
e
d
b
e
c
a
u
se
th
e
l
o
ss
v
a
lu
e
d
id
n
o
t
d
e
c
re
a
se
fo
r
fiv
e
c
o
n
se
c
u
ti
v
e
e
p
o
c
h
s,
in
d
ica
ti
n
g
t
h
a
t
t
h
e
m
o
d
e
l
h
a
d
re
a
c
h
e
d
c
o
n
v
e
rg
e
n
c
e
.
−
P
ix
e
l
a
c
c
u
ra
c
y
(P
A) l
e
a
rn
in
g
c
u
r
v
e
:
T
h
e
tr
ain
in
g
cu
r
v
e
s
h
o
ws
a
s
ig
n
if
ican
t
in
cr
ea
s
e
in
th
e
f
ir
s
t
two
ep
o
ch
s
,
th
en
s
tab
ilizes
u
n
til
th
e
8
th
ep
o
c
h
.
T
h
e
v
alid
atio
n
cu
r
v
e
s
h
o
ws
a
s
u
b
s
tan
tial
an
d
s
tab
le
in
itial
in
cr
ea
s
e
a
f
ter
t
h
e
2
nd
e
p
o
ch
.
C
lo
s
e
tr
ain
in
g
an
d
v
alid
atio
n
cu
r
v
es
in
d
icate
th
e
m
o
d
el
h
as
co
n
s
is
ten
t
p
er
f
o
r
m
an
ce
b
etwe
en
tr
ain
in
g
an
d
v
alid
atio
n
d
ata.
−
Lo
ss
lea
rn
in
g
c
u
r
v
e
:
T
h
e
tr
ain
in
g
lo
s
s
cu
r
v
e
s
h
o
ws
a
d
r
asti
c
d
ec
r
ea
s
e
in
th
e
f
ir
s
t
two
ep
o
ch
s
an
d
s
tar
ts
to
s
tab
ilize
u
n
til
th
e
8
th
e
p
o
ch
.
T
h
e
v
alid
atio
n
lo
s
s
cu
r
v
e
also
s
h
o
ws
a
s
ig
n
if
ican
t
r
e
d
u
cti
o
n
in
th
e
f
ir
s
t
two
ep
o
c
h
s
an
d
s
tab
ilizes
af
ter
th
at.
T
h
e
s
tab
ilit
y
o
f
th
e
v
ali
d
atio
n
lo
s
s
c
u
r
v
e
s
h
o
ws
th
at
th
e
m
o
d
el
d
o
es
n
o
t
ex
p
er
ien
ce
s
u
b
s
tan
tial o
v
er
f
itti
n
g
.
b.
Re
sN
e
t
(b
)
−
M
e
a
n
in
ters
e
c
ti
o
n
o
v
e
r
u
n
io
n
(I
o
U) l
e
a
rn
in
g
c
u
rv
e
:
T
h
e
tr
ain
i
n
g
c
u
r
v
e
s
h
o
ws
s
ta
b
l
e
an
d
h
i
g
h
I
o
U
v
alu
es
th
r
o
u
g
h
o
u
t
f
i
v
e
e
p
o
ch
s
.
T
h
e
v
alid
atio
n
cu
r
v
e
s
h
o
ws
a
d
r
asti
c
d
ec
lin
e
af
ter
th
e
f
ir
s
t
ep
o
ch
a
n
d
r
e
m
ain
s
lo
w,
in
d
icatin
g
th
e
m
o
d
el
ca
n
n
o
t
g
e
n
er
alize
th
e
v
alid
atio
n
d
ata
well.
T
h
e
lo
w
v
alid
atio
n
cu
r
v
e
s
u
g
g
ests
th
is
m
o
d
el
m
ay
b
e
o
v
er
f
itti
n
g
t
h
e
t
r
ain
in
g
d
ata.
−
P
ix
e
l
a
c
c
u
ra
c
y
(P
A) l
e
a
rn
in
g
c
u
r
v
e
:
T
h
e
tr
ain
in
g
c
u
r
v
e
s
h
o
ws
h
i
g
h
an
d
s
tab
le
PA
v
alu
es
th
r
o
u
g
h
o
u
t
th
e
f
iv
e
ep
o
c
h
s
.
T
h
e
v
alid
atio
n
cu
r
v
e
s
ig
n
if
ica
n
tly
d
ec
r
ea
s
es
af
ter
th
e
f
ir
s
t
ep
o
ch
an
d
r
e
m
ain
s
lo
w.
L
ik
e
th
e
I
o
U
c
u
r
v
e,
th
e
m
o
d
el
is
ex
p
er
ien
cin
g
o
v
e
r
f
itti
n
g
a
n
d
c
an
n
o
t g
e
n
er
alize
well.
−
Lo
ss
lea
rn
in
g
c
u
r
v
e
:
T
h
e
tr
ain
in
g
lo
s
s
cu
r
v
e
s
h
o
ws
s
tab
le
an
d
r
elativ
ely
lo
w
v
alu
es
th
r
o
u
g
h
o
u
t
th
e
f
iv
e
ep
o
ch
s
.
T
h
e
v
alid
atio
n
lo
s
s
cu
r
v
e
s
h
o
ws
lar
g
e
f
lu
ctu
atio
n
s
,
with
h
ig
h
er
v
alu
es
th
an
th
e
tr
ain
in
g
cu
r
v
e.
Flu
ctu
atio
n
s
an
d
h
ig
h
v
alid
atio
n
lo
s
s
v
alu
es in
d
icate
th
e
m
o
d
el
is
u
n
s
tab
le
an
d
u
n
ab
le
to
g
e
n
er
alize
o
n
v
ali
d
atio
n
d
ata.
Dee
p
L
ab
V3
Plu
s
s
h
o
ws
b
etter
r
esu
lts
co
m
p
ar
e
d
to
R
esNet.
T
h
is
m
o
d
el
s
h
o
ws
g
o
o
d
s
ta
b
ilit
y
an
d
g
en
er
aliza
tio
n
ab
ilit
y
,
as
s
ee
n
f
r
o
m
th
e
tr
ain
i
n
g
a
n
d
v
alid
at
io
n
cu
r
v
es,
w
h
ich
a
r
e
alm
o
s
t
p
ar
allel
an
d
s
tab
le.
E
v
en
t
h
o
u
g
h
it
r
eq
u
ir
es
lo
n
g
e
r
tr
ain
in
g
tim
e
(
2
8
0
,
5
9
6
m
in
u
tes),
th
e
r
esu
lts
ar
e
m
o
r
e
c
o
n
s
is
ten
t
an
d
r
eliab
le.
R
esNet,
h
o
wev
er
,
s
h
o
ws
p
o
o
r
p
er
f
o
r
m
an
ce
o
n
v
alid
atio
n
d
a
ta,
wit
h
s
tr
o
n
g
in
d
icatio
n
s
o
f
o
v
er
f
itti
n
g
.
Desp
ite
th
e
m
u
ch
s
h
o
r
ter
tr
ain
in
g
tim
e
(
9
0
m
in
u
tes),
th
e
r
esu
lts
o
b
tain
ed
co
u
ld
h
a
v
e
b
ee
n
b
etter
an
d
m
o
r
e
r
eliab
le
f
o
r
p
r
ed
ictio
n
s
o
n
n
ew
d
ata.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
Op
timiz
in
g
co
n
vo
lu
tio
n
a
l
n
e
u
r
a
l n
etw
o
r
k
h
yp
erp
a
r
a
mete
r
s
to
en
h
a
n
ce
liver
…
(
I
w
a
n
P
u
r
n
a
ma
)
3881
(a
)
(b
)
Fig
u
r
e
4
.
Gr
a
p
h
o
f
m
ea
n
in
ter
s
ec
tio
n
o
v
er
u
n
io
n
lear
n
in
g
cu
r
v
e,
p
ix
el
ac
c
u
r
ac
y
lear
n
in
g
c
u
r
v
e,
a
n
d
lo
s
s
lear
n
in
g
cu
r
v
e
C
XR
im
ag
e
s
eg
m
en
tatio
n
tr
ain
i
n
g
p
r
o
ce
s
s
u
s
in
g
ar
ch
itectu
r
e:
(
a)
Dee
p
L
ab
V3
Plu
s
an
d
(
b
)
R
esNet
3
.
2
.
Dis
cus
s
io
n
I
n
th
is
s
ec
tio
n
,
we
will
an
aly
z
e
an
d
d
is
cu
s
s
th
e
r
esear
ch
r
es
u
lts
o
b
tain
ed
f
r
o
m
t
r
ain
in
g
two
m
o
d
els,
Dee
p
L
ab
V3
Plu
s
an
d
R
esNet.
T
h
is
an
aly
s
is
in
clu
d
es
ev
alu
at
in
g
th
e
p
e
r
f
o
r
m
an
ce
o
f
b
o
t
h
m
o
d
els
b
ased
o
n
th
e
r
esu
ltin
g
lear
n
in
g
cu
r
v
es,
in
clu
d
in
g
m
ea
n
in
ter
s
ec
tio
n
o
v
er
u
n
io
n
(
I
o
U)
,
p
ix
el
ac
cu
r
ac
y
(
PA)
,
an
d
l
o
s
s
.
T
h
is
d
is
cu
s
s
io
n
will
al
s
o
in
clu
d
e
a
co
m
p
ar
is
o
n
b
etwe
en
th
e
e
f
f
i
cien
cy
o
f
tr
ain
in
g
tim
e
an
d
th
e
q
u
ality
o
f
r
esu
lts
o
b
tain
ed
f
r
o
m
b
o
th
m
o
d
els.
B
y
u
n
d
e
r
s
tan
d
in
g
th
e
d
if
f
e
r
en
ce
s
an
d
ad
v
a
n
tag
es
o
f
ea
ch
m
o
d
el,
we
ca
n
d
r
aw
m
o
r
e
p
r
ec
is
e
co
n
clu
s
io
n
s
a
b
o
u
t
wh
ich
m
o
d
el
is
m
o
r
e
ef
f
ec
t
iv
e
to
u
s
e
in
a
s
p
ec
if
ic
a
p
p
lic
atio
n
co
n
tex
t.
Nex
t,
we
will
also
d
is
cu
s
s
th
e
im
p
licatio
n
s
o
f
th
ese
r
esu
lts
f
o
r
f
u
tu
r
e
u
s
e
o
f
co
m
p
u
tatio
n
al
r
e
s
o
u
r
ce
s
an
d
m
o
d
el
d
ev
elo
p
m
e
n
t
s
tr
ateg
ies.
I
n
Fi
g
u
r
e
5
,
th
e
p
r
e
d
ictio
n
r
esu
lts
o
f
th
e
f
o
llo
win
g
two
m
o
d
el
s
ar
e
co
m
p
ar
ed
o
n
Fig
u
r
e
5
(
a)
Dee
p
L
ab
V3
Plu
s
a
n
d
Fig
u
r
e
5
(
b
)
R
esNet.
I
n
Fig
u
r
e
5
,
Dee
p
L
ab
V3
Plu
s
:
T
h
is
m
o
d
el
s
h
o
ws
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
in
im
ag
e
s
eg
m
e
n
tatio
n
.
T
h
e
p
r
ed
ictio
n
r
esu
lts
ar
e
v
er
y
c
lo
s
e
to
th
e
g
r
o
u
n
d
tr
u
th
m
a
s
k
,
d
em
o
n
s
tr
atin
g
th
e
ab
ilit
y
o
f
th
is
m
o
d
el
to
r
ec
o
g
n
ize
an
d
d
elin
ea
te
o
b
jec
ts
in
im
ag
es
with
h
ig
h
p
r
ec
is
io
n
.
Dee
p
L
ab
V3
Plu
s
s
u
cc
ee
d
e
d
in
ca
p
tu
r
i
n
g
t
h
e
d
etails
an
d
s
tr
u
ctu
r
e
o
f
o
b
je
cts
v
er
y
well,
p
r
o
d
u
cin
g
ac
c
u
r
ate
s
eg
m
en
tatio
n
.
R
esNet:
T
h
is
m
o
d
el
s
h
o
ws
ab
y
s
m
al
p
er
f
o
r
m
an
ce
in
im
a
g
e
s
eg
m
en
tatio
n
task
s
.
T
h
e
p
r
e
d
ictio
n
r
esu
lts
ar
e
f
a
r
f
r
o
m
th
e
g
r
o
u
n
d
tr
u
th
m
ask
an
d
s
h
o
w
m
an
y
er
r
o
r
s
.
R
esNet
ap
p
ea
r
s
to
n
ee
d
to
im
p
r
o
v
e
i
n
r
ec
o
g
n
izin
g
an
d
d
elin
ea
tin
g
o
b
jects
in
im
a
g
es,
r
esu
ltin
g
in
in
ac
cu
r
ate
an
d
n
o
is
y
s
eg
m
en
tatio
n
.
C
o
m
p
ar
is
o
n
:
T
h
e
p
er
f
o
r
m
a
n
ce
d
if
f
er
en
ce
b
etwe
e
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
4
,
Au
g
u
s
t
20
25
:
3
8
7
6
-
3887
3882
Dee
p
L
ab
V3
Plu
s
an
d
R
esNet
is
v
er
y
s
ig
n
if
ica
n
t.
Dee
p
L
a
b
V3
Plu
s
is
m
o
r
e
ef
f
ec
tiv
e
in
s
eg
m
en
tin
g
im
a
g
es
co
m
p
ar
ed
to
R
esNet.
Pre
v
io
u
s
lear
n
in
g
cu
r
v
e
an
aly
s
is
s
h
o
ws
th
at
Dee
p
L
ab
V3
Plu
s
h
as
b
etter
g
en
er
aliza
tio
n
ca
p
ab
ilit
ies
an
d
d
o
es
n
o
t
ex
p
er
ien
ce
o
v
er
f
itti
n
g
li
k
e
R
es
Net.
T
h
ese
r
esu
lts
em
p
h
asize
th
e
im
p
o
r
tan
ce
o
f
ch
o
o
s
in
g
a
s
u
itab
le
m
o
d
el
f
o
r
im
ag
e
s
eg
m
en
tatio
n
task
s
,
w
h
er
e
Dee
p
L
ab
V3
Plu
s
is
p
r
o
v
e
n
m
o
r
e
r
eliab
le
a
n
d
ef
f
ec
tiv
e
th
an
R
esNet.
T
o
p
r
o
v
id
e
a
clea
r
er
p
ictu
r
e
o
f
th
e
p
er
f
o
r
m
an
c
e
o
f
th
e
tw
o
m
o
d
els
th
at
h
av
e
b
ee
n
test
ed
in
th
is
r
esear
ch
,
b
el
o
w
we
p
r
esen
t
a
t
ab
le
o
f
ev
alu
atio
n
r
esu
lts
wh
ich
in
clu
d
es
th
r
ee
m
ain
m
etr
ic
s
:
m
ea
n
in
ter
s
ec
tio
n
o
v
er
u
n
io
n
(
MI
o
U
)
s
co
r
e,
p
ix
el
ac
cu
r
ac
y
(
PA)
s
co
r
e,
a
n
d
l
o
s
s
v
alu
e.
T
h
ese
m
etr
ics
p
r
o
v
id
e
a
co
m
p
r
eh
en
s
iv
e
v
iew
o
f
h
o
w
well
ea
c
h
m
o
d
el
p
er
f
o
r
m
s
im
ag
e
s
eg
m
en
tatio
n
.
MI
o
U
Sco
r
e
is
a
m
etr
ic
u
s
ed
to
m
ea
s
u
r
e
h
o
w
well
a
m
o
d
el
p
r
ed
icts
s
eg
m
en
tatio
n
b
y
co
m
p
ar
in
g
th
e
o
v
e
r
lap
ar
ea
b
etwe
en
p
r
e
d
ictio
n
s
an
d
g
r
o
u
n
d
tr
u
th
.
PA
Sco
r
e
m
ea
s
u
r
es
p
ix
el
ac
cu
r
ac
y
in
s
eg
m
en
tatio
n
,
n
a
m
ely
th
e
p
er
ce
n
tag
e
o
f
co
r
r
ec
tly
clas
s
if
ied
p
ix
els.
L
o
s
s
Valu
e
s
h
o
ws h
o
w
well
th
e
m
o
d
el
m
in
im
izes p
r
ed
ictio
n
er
r
o
r
s
d
u
r
in
g
tr
ain
in
g
.
T
ab
le
1
s
h
o
ws th
e
p
er
f
o
r
m
an
ce
co
m
p
ar
is
o
n
r
esu
lts
b
etwe
en
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
(
Dee
p
L
ab
V3
Plu
s
)
an
d
th
e
R
esNet
m
o
d
el.
T
h
e
d
ata
in
th
is
tab
le
s
h
o
ws
s
ig
n
if
ican
t
d
if
f
e
r
en
ce
s
in
s
eg
m
en
tatio
n
ca
p
ab
ilit
ies
b
etwe
en
th
e
two
m
o
d
els,
wh
ich
will
b
e
d
is
cu
s
s
ed
f
u
r
th
er
in
t
h
e
an
aly
s
is
an
d
d
is
cu
s
s
io
n
s
ec
tio
n
.
(
a)
(
b
)
Fig
u
r
e
5
.
I
n
f
er
e
n
ce
an
d
p
er
f
o
r
m
an
ce
an
aly
s
is
o
f
AI
m
o
d
els
with
Gr
ad
C
AM
(
a)
Dee
p
L
ab
V3
Plu
s
an
d
(
b
)
R
esNet
T
ab
l
e
1
.
C
o
m
p
a
r
is
o
n
o
f
Dee
p
L
ab
V3
Plu
s
an
d
R
esNet
an
aly
s
is
M
e
t
h
o
d
M
I
o
U
s
c
o
r
e
P
A
s
c
o
r
e
Lo
ss
v
a
l
u
e
P
r
o
p
o
se
d
m
o
d
e
l
(
D
e
e
p
L
a
b
V
3
P
l
u
s)
0
.
9
6
5
0
.
9
2
9
0
.
0
1
1
R
e
sN
e
t
0
.
0
6
0
0
.
1
1
7
0
.
9
2
1
T
ab
le
1
r
esear
ch
r
esu
lts
s
h
o
w
th
at
th
e
Dee
p
L
ab
V3
Plu
s
m
o
d
el
is
s
ig
n
if
ican
tly
s
u
p
er
io
r
to
R
esNet
in
th
e
im
ag
e
s
eg
m
en
tatio
n
task
.
Dee
p
L
ab
V3
Plu
s
ac
h
iev
ed
a
n
MI
o
U
s
co
r
e
o
f
0
.
9
6
5
a
n
d
a
PA
s
co
r
e
o
f
0
.
9
2
9
,
m
u
ch
h
ig
h
er
th
a
n
R
esNet,
wh
ich
o
n
ly
ac
h
iev
ed
a
n
MI
o
U
s
co
r
e
o
f
0
.
0
6
0
an
d
a
PA Sco
r
e
o
f
0
.
1
1
7
.
I
n
ad
d
itio
n
,
th
e
lo
s
s
v
alu
e
o
f
Dee
p
L
ab
V3
Plu
s
is
v
er
y
lo
w
(
0
.
0
1
1
)
co
m
p
ar
e
d
to
R
esNet
(
0
.
9
2
1
)
,
in
d
icatin
g
th
at
Dee
p
L
ab
V3
Plu
s
is
m
o
r
e
ef
f
ec
tiv
e
in
m
in
im
izin
g
p
r
ed
ictio
n
er
r
o
r
s
.
I
n
c
o
n
clu
s
io
n
,
Dee
p
L
ab
V3
Plu
s
is
m
o
r
e
ac
cu
r
ate
a
n
d
e
f
f
icien
t f
o
r
im
ag
e
s
eg
m
en
tatio
n
,
wh
ile
R
esNet
s
h
o
ws ab
y
s
m
al
p
er
f
o
r
m
an
ce
i
n
th
is
co
n
tex
t.
Alth
o
u
g
h
th
is
s
tu
d
y
s
u
cc
ess
f
u
lly
im
p
r
o
v
e
d
th
e
liv
er
s
eg
m
en
tatio
n
ac
cu
r
ac
y
u
s
in
g
th
e
Dee
p
L
ab
V3
Plu
s
m
o
d
el
with
h
y
p
er
p
ar
am
eter
o
p
tim
izatio
n
,
th
er
e
ar
e
s
till
s
o
m
e
u
n
an
s
wer
ed
q
u
esti
o
n
s
an
d
o
p
p
o
r
tu
n
ities
f
o
r
f
u
tu
r
e
r
esear
ch
.
On
e
asp
ec
t
th
at
n
ee
d
s
to
b
e
f
u
r
th
e
r
e
x
p
lo
r
e
d
is
th
e
g
e
n
er
aliza
tio
n
o
f
th
e
m
o
d
el
to
v
ar
io
u
s
v
ar
iatio
n
s
o
f
m
ed
ical
im
ag
es,
s
u
ch
as
lo
w
-
q
u
ality
o
r
h
ig
h
-
n
o
is
e
im
ag
es.
I
n
ad
d
itio
n
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
Op
timiz
in
g
co
n
vo
lu
tio
n
a
l
n
e
u
r
a
l n
etw
o
r
k
h
yp
erp
a
r
a
mete
r
s
to
en
h
a
n
ce
liver
…
(
I
w
a
n
P
u
r
n
a
ma
)
3883
ad
ap
tin
g
th
is
m
o
d
el
to
s
eg
m
en
t
o
th
er
o
r
g
an
s
o
r
in
m
u
lti
-
o
r
g
an
s
ce
n
ar
i
o
s
co
u
ld
b
e
a
p
r
o
m
is
in
g
r
esear
ch
d
ir
ec
tio
n
.
T
h
e
u
s
e
o
f
m
o
r
e
s
o
p
h
is
ticated
d
ata
au
g
m
en
tat
io
n
tech
n
iq
u
es
an
d
f
ew
-
s
h
o
t
o
r
s
elf
-
s
u
p
er
v
is
ed
lear
n
in
g
-
b
ased
ap
p
r
o
ac
h
es
c
an
also
b
e
in
v
esti
g
ated
to
f
u
r
th
er
i
m
p
r
o
v
e
th
e
m
o
d
el
p
er
f
o
r
m
a
n
ce
with
o
u
t
r
eq
u
ir
in
g
lar
g
e
am
o
u
n
ts
o
f
tr
ai
n
in
g
d
ata.
4.
CO
NCLU
SI
O
N
B
ased
o
n
th
e
r
esear
ch
r
esu
lt
s
,
th
e
Dee
p
L
ab
V3
Plu
s
m
o
d
e
l
is
s
ig
n
if
ican
tly
s
u
p
er
io
r
to
R
esNet
in
im
ag
e
s
eg
m
en
tatio
n
task
s
.
D
ee
p
L
ab
V3
Plu
s
ac
h
iev
e
d
an
MI
o
U
s
co
r
e
o
f
0
.
9
6
5
,
a
PA
s
co
r
e
o
f
0
.
9
2
9
,
an
d
a
m
ea
g
er
lo
s
s
v
alu
e
o
f
0
.
0
1
1
.
T
h
is
s
h
o
ws its
ab
ilit
y
to
r
ec
o
g
n
i
z
e
an
d
p
r
ed
ict
s
eg
m
en
tatio
n
a
r
ea
s
v
er
y
ac
cu
r
ately
an
d
co
n
s
is
ten
tly
,
as
well
as
m
in
im
ize
p
r
e
d
ictio
n
er
r
o
r
s
ef
f
ec
tiv
ely
.
I
n
co
n
t
r
ast,
R
es
Net
s
h
o
ws
ab
y
s
m
al
p
er
f
o
r
m
an
ce
with
a
n
MI
o
U
s
co
r
e
o
f
0
.
0
6
0
,
a
PA
s
co
r
e
o
f
0
.
1
1
7
,
a
n
d
a
h
ig
h
lo
s
s
v
alu
e
o
f
0
.
9
2
1
,
in
d
icatin
g
its
in
co
m
p
eten
ce
in
im
a
g
e
s
eg
m
en
tatio
n
task
s
.
T
h
is
s
ig
n
if
ican
t
d
if
f
er
en
ce
c
o
n
f
ir
m
s
th
at
Dee
p
L
ab
V3
Plu
s
is
a
m
o
r
e
ef
f
ec
tiv
e
an
d
e
f
f
icien
t c
h
o
ice
f
o
r
a
p
p
licatio
n
s
th
at
r
e
q
u
i
r
e
p
r
ec
is
e
im
ag
e
s
eg
m
e
n
tatio
n
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
e
au
th
o
r
s
d
ec
lar
e
th
at
n
o
f
u
n
d
in
g
was r
ec
eiv
e
d
f
o
r
th
is
s
tu
d
y
.
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
I
wan
Pu
r
n
am
a
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
A
g
u
s
P
e
r
d
a
n
a
W
i
n
d
a
r
t
o
✓
✓
✓
✓
✓
✓
✓
✓
So
lik
h
u
n
✓
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
T
h
e
au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
ter
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
av
ailab
le
f
r
o
m
th
e
au
th
o
r
,
Ka
g
g
le,
u
p
o
n
r
ea
s
o
n
ab
le
r
eq
u
est.
RE
F
E
R
E
NC
E
S
[
1
]
S
.
A
r
o
r
a
,
S
.
K
.
Tr
i
p
a
t
h
y
,
R
.
G
u
p
t
a
,
a
n
d
R
.
S
r
i
v
a
s
t
a
v
a
,
“
E
x
p
l
o
i
t
i
n
g
mu
l
t
i
m
o
d
a
l
C
N
N
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
a
u
t
o
m
a
t
e
d
t
e
e
t
h
s
e
g
m
e
n
t
a
t
i
o
n
o
n
d
e
n
t
a
l
p
a
n
o
r
a
m
i
c
X
-
r
a
y
i
ma
g
e
s
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
I
n
s
t
i
t
u
t
i
o
n
o
f
M
e
c
h
a
n
i
c
a
l
En
g
i
n
e
e
rs,
Pa
r
t
H
:
J
o
u
r
n
a
l
o
f
E
n
g
i
n
e
e
ri
n
g
i
n
Me
d
i
c
i
n
e
,
M
a
r
.
2
0
2
3
,
v
o
l
.
2
3
7
,
n
o
.
3
,
p
p
.
3
9
5
–
4
0
5
.
d
o
i
:
1
0
.
1
1
7
7
/
0
9
5
4
4
1
1
9
2
3
1
1
5
7
1
3
7
.
[
2
]
X
.
M
e
n
g
,
Z
.
X
i
n
g
,
a
n
d
S
.
L
u
,
“
A
d
e
e
p
l
e
a
r
n
i
n
g
-
b
a
s
e
d
l
u
n
g
n
o
d
u
l
e
d
e
n
si
t
y
c
l
a
ss
i
f
i
c
a
t
i
o
n
a
n
d
se
g
m
e
n
t
a
t
i
o
n
m
e
t
h
o
d
a
n
d
i
t
s
e
f
f
e
c
t
i
v
e
n
e
ss
u
n
d
e
r
d
i
f
f
e
r
e
n
t
C
T
r
e
c
o
n
s
t
r
u
c
t
i
o
n
a
l
g
o
r
i
t
h
ms,
”
N
a
t
i
o
n
a
l
M
e
d
i
c
a
l
J
o
u
rn
a
l
o
f
C
h
i
n
a
,
v
o
l
.
1
0
1
,
n
o
.
7
,
p
p
.
4
7
6
–
4
8
0
,
2
0
2
1
,
d
o
i
:
1
0
.
3
7
6
0
/
c
m
a
.
j
.
c
n
1
1
2
1
3
7
-
2
0
2
0
1
1
2
3
-
0
3
1
7
1
.
[
3
]
S
.
N
a
g
o
o
r
a
n
d
S
.
V
.
J
i
n
n
y
,
“
A
d
u
a
l
f
u
z
z
y
w
i
t
h
h
y
b
r
i
d
d
e
e
p
l
e
a
r
n
i
n
g
a
r
c
h
i
t
e
c
t
u
r
e
b
a
se
d
o
n
C
N
N
w
i
t
h
h
y
b
r
i
d
me
t
a
h
e
u
r
i
s
t
i
c
a
l
g
o
r
i
t
h
m
f
o
r
e
f
f
e
c
t
i
v
e
s
e
g
m
e
n
t
a
t
i
o
n
a
n
d
c
l
a
ssi
f
i
c
a
t
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
n
f
o
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
5
,
n
o
.
1
,
p
p
.
5
3
1
–
5
4
3
,
J
a
n
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s4
1
8
7
0
-
0
2
2
-
0
1
1
0
6
-
5.
[
4
]
J.
A
mo
s
e
,
P
.
M
a
n
i
m
e
g
a
l
a
i
,
a
n
d
R
.
P
o
n
S
e
l
c
h
i
y
a
,
“
C
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
a
d
v
e
n
t
i
t
i
o
u
s
l
u
n
g
s
o
u
n
d
s:
W
h
e
e
z
e
,
c
r
a
c
k
l
e
u
s
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s,”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
S
y
s
t
e
m
s a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
i
n
E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
1
,
n
o
.
3
,
p
p
.
1
1
4
3
–
1
1
5
2
,
2
0
2
3
.
[
5
]
D
.
W
.
K
i
m,
M
.
S
.
K
i
m,
J
.
Le
e
,
a
n
d
P
.
P
a
r
k
,
“
A
d
a
p
t
i
v
e
l
e
a
r
n
i
n
g
-
r
a
t
e
b
a
c
k
p
r
o
p
a
g
a
t
i
o
n
n
e
u
r
a
l
n
e
t
w
o
r
k
a
l
g
o
r
i
t
h
m
b
a
s
e
d
o
n
t
h
e
mi
n
i
mi
z
a
t
i
o
n
o
f
me
a
n
-
s
q
u
a
r
e
d
e
v
i
a
t
i
o
n
f
o
r
i
m
p
u
l
s
i
v
e
n
o
i
s
e
s,
”
I
EE
E
Ac
c
e
ss
,
v
o
l
.
8
,
p
p
.
9
8
0
1
8
–
9
8
0
2
6
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
0
.
2
9
9
7
0
1
0
.
[
6
]
H
.
Li
s
t
i
y
o
n
o
,
Z
.
B
u
d
i
a
r
so
,
a
n
d
A
.
P
.
W
i
n
d
a
r
t
o
,
“
A
n
o
p
t
i
m
i
z
e
d
C
N
N
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
a
c
c
u
r
a
t
e
3
D
l
i
v
e
r
se
g
me
n
t
a
t
i
o
n
i
n
m
e
d
i
c
a
l
i
ma
g
e
s,
”
J
o
u
r
n
a
l
o
f
I
m
a
g
e
a
n
d
G
ra
p
h
i
c
s
,
v
o
l
.
1
3
,
n
o
.
1
,
p
p
.
1
5
–
2
3
,
2
0
2
5
,
d
o
i
:
1
0
.
1
8
1
7
8
/
j
o
i
g
.
1
3
.
1
.
1
5
-
2
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
4
,
Au
g
u
s
t
20
25
:
3
8
7
6
-
3887
3884
[
7
]
P
.
A
l
k
h
a
i
r
i
,
W
.
W
a
n
a
y
u
mi
n
i
,
a
n
d
B
.
H
.
H
a
y
a
d
i
,
“
A
n
a
l
y
si
s
o
f
t
h
e
a
d
a
p
t
i
v
e
l
e
a
r
n
i
n
g
r
a
t
e
a
n
d
m
o
me
n
t
u
m
e
f
f
e
c
t
s
o
n
p
r
e
d
i
c
t
i
o
n
p
r
o
b
l
e
ms
i
n
i
n
c
r
e
a
si
n
g
t
h
e
t
r
a
i
n
i
n
g
t
i
me
o
f
t
h
e
b
a
c
k
p
r
o
p
a
g
a
t
i
o
n
a
l
g
o
r
i
t
h
m,
”
i
n
AI
P
C
o
n
f
e
re
n
c
e
Pr
o
c
e
e
d
i
n
g
s
,
2
0
2
4
,
p
.
2
0
0
4
9
.
d
o
i
:
1
0
.
1
0
6
3
/
5
.
0
2
0
3
3
7
4
.
[
8
]
K
.
H
.
N
i
r
a
l
g
i
k
a
r
a
n
d
M
.
A
.
B
u
l
sar
a
,
“
F
r
e
q
u
e
n
t
p
a
t
t
e
r
n
d
a
t
a
m
i
n
i
n
g
b
a
se
d
c
l
u
s
t
e
r
i
n
g
a
n
d
c
l
a
s
si
f
i
c
a
t
i
o
n
u
si
n
g
f
u
z
z
y
b
a
c
k
p
r
o
p
a
g
a
t
i
o
n
r
e
s
n
e
t
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
t
w
o
r
k
s,”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
S
y
s
t
e
m
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
i
n
En
g
i
n
e
e
r
i
n
g
,
v
o
l
.
1
0
,
n
o
.
2
s
,
p
p
.
2
0
1
–
2
0
4
,
2
0
2
2
.
[
9
]
L.
I
.
K
e
s
u
ma
,
“
E
l
r
e
i
:
E
n
sem
b
l
e
l
e
a
r
n
i
n
g
o
f
R
e
sN
e
t
,
e
f
f
i
c
i
e
n
t
n
e
t
,
a
n
d
I
n
c
e
p
t
i
o
n
-
V
3
f
o
r
l
u
n
g
d
i
s
e
a
s
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
b
a
se
d
o
n
c
h
e
st
X
-
R
a
y
i
m
a
g
e
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
E
n
g
i
n
e
e
ri
n
g
a
n
d
S
y
s
t
e
m
s
,
v
o
l
.
1
6
,
n
o
.
5
,
p
p
.
1
4
9
–
1
6
1
,
O
c
t
.
2
0
2
3
,
d
o
i
:
10
.
2
2
2
6
6
/
i
j
i
e
s
2
0
2
3
.
1
0
3
1
.
1
4
.
[
1
0
]
S
.
S
h
i
v
a
d
e
k
a
r
,
B
.
K
a
t
a
r
i
a
,
S
.
H
u
n
d
e
k
a
r
i
,
K
.
W
a
n
j
a
l
e
,
V
.
P
.
B
a
l
p
a
n
d
e
,
a
n
d
R
.
S
u
r
y
a
w
a
n
s
h
i
,
“
D
e
e
p
l
e
a
r
n
i
n
g
b
a
se
d
i
m
a
g
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
l
u
n
g
s
r
a
d
i
o
g
r
a
p
h
y
f
o
r
d
e
t
e
c
t
i
n
g
C
O
V
I
D
-
1
9
u
s
i
n
g
a
d
e
e
p
C
N
N
a
n
d
R
e
sN
e
t
5
0
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
S
y
s
t
e
m
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
i
n
E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
1
,
n
o
.
1
s
,
p
p
.
2
4
1
–
2
5
0
,
2
0
2
3
.
[
1
1
]
H
.
S
.
H
o
a
n
g
,
C
.
P
h
u
o
n
g
P
h
a
m,
D
.
F
r
a
n
k
l
i
n
,
T
.
V
a
n
W
a
l
s
u
m,
a
n
d
M
.
H
a
Lu
u
,
“
A
n
e
v
a
l
u
a
t
i
o
n
o
f
C
N
N
-
b
a
s
e
d
l
i
v
e
r
s
e
g
me
n
t
a
t
i
o
n
met
h
o
d
s
u
s
i
n
g
m
u
l
t
i
-
t
y
p
e
s
o
f
C
T
a
b
d
o
m
i
n
a
l
i
ma
g
e
s
f
r
o
m
m
u
l
t
i
p
l
e
m
e
d
i
c
a
l
c
e
n
t
e
r
s,”
i
n
Pr
o
c
e
e
d
i
n
g
s
-
2
0
1
9
1
9
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
S
y
m
p
o
s
i
u
m
o
n
C
o
m
m
u
n
i
c
a
t
i
o
n
s
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
,
I
S
C
I
T
2
0
1
9
,
2
0
1
9
,
p
p
.
2
0
–
2
5
.
d
o
i
:
1
0
.
1
1
0
9
/
I
S
C
I
T.
2
0
1
9
.
8
9
0
5
1
6
6
.
[
1
2
]
L.
M
e
n
g
,
Y
.
Ti
a
n
,
a
n
d
S
.
B
u
,
“
Li
v
e
r
t
u
m
o
r
se
g
m
e
n
t
a
t
i
o
n
b
a
se
d
o
n
3
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
w
i
t
h
d
u
a
l
s
c
a
l
e
,
”
J
o
u
rn
a
l
o
f
Ap
p
l
i
e
d
C
l
i
n
i
c
a
l
Me
d
i
c
a
l
Ph
y
si
c
s
,
v
o
l
.
2
1
,
n
o
.
1
,
p
p
.
1
4
4
–
1
5
7
,
J
a
n
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
2
/
a
c
m
2
.
1
2
7
8
4
.
[
1
3
]
D
.
R
u
a
n
,
F
.
Z
h
a
n
g
,
L.
Zh
a
n
g
,
a
n
d
J.
Y
a
n
,
“
O
p
t
i
ma
l
m
o
d
i
f
i
c
a
t
i
o
n
s
i
n
C
N
N
f
o
r
b
e
a
r
i
n
g
f
a
u
l
t
c
l
a
ss
i
f
i
c
a
t
i
o
n
a
n
d
a
d
a
p
t
a
t
i
o
n
a
c
r
o
ss
d
i
f
f
e
r
e
n
t
w
o
r
k
i
n
g
c
o
n
d
i
t
i
o
n
s
,
”
J
o
u
r
n
a
l
o
f
Vi
b
ra
t
i
o
n
E
n
g
i
n
e
e
r
i
n
g
&
T
e
c
h
n
o
l
o
g
i
e
s
,
v
o
l
.
1
2
,
n
o
.
3
,
p
p
.
4
0
7
5
–
4
0
9
5
,
M
a
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
4
2
4
1
7
-
0
2
3
-
0
1
1
0
6
-
0.
[
1
4
]
K
.
M
.
S
u
n
n
e
t
c
i
,
“
F
a
c
e
M
a
sk
D
e
t
e
c
t
i
o
n
U
s
i
n
g
G
o
o
g
Le
N
e
t
C
N
N
-
B
a
s
e
d
S
V
M
C
l
a
ss
i
f
i
e
r
s,
”
G
a
z
i
U
n
i
v
e
rsi
t
y
J
o
u
rn
a
l
o
f
S
c
i
e
n
c
e
,
v
o
l
.
3
6
,
n
o
.
2
,
p
p
.
6
4
5
–
6
5
8
,
2
0
2
3
,
d
o
i
:
1
0
.
3
5
3
7
8
/
g
u
j
s.
1
0
0
9
3
5
9
.
[
1
5
]
A
.
K
h
a
l
i
l
i
F
a
k
h
r
a
b
a
d
i
,
M
.
J
.
S
h
a
h
b
a
z
z
a
d
e
h
,
N
.
Ja
l
a
l
i
,
a
n
d
M
.
Es
l
a
m
i
,
“
A
h
y
b
r
i
d
i
n
c
e
p
t
i
o
n
-
d
i
l
a
t
e
d
-
R
e
sN
e
t
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
d
e
e
p
l
e
a
r
n
i
n
g
-
b
a
s
e
d
p
r
e
d
i
c
t
i
o
n
o
f
C
O
V
I
D
-
1
9
s
e
v
e
r
i
t
y
,
”
S
c
i
e
n
t
i
f
i
c
Re
p
o
r
t
s
,
v
o
l
.
1
5
,
n
o
.
1
,
p
.
6
4
9
0
,
F
e
b
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
3
8
/
s4
1
5
9
8
-
025
-
9
1
3
2
2
-
3.
[
1
6
]
A
.
K
a
u
r
a
n
d
D
.
G
u
p
t
a
,
“
A
H
y
b
r
i
d
C
l
a
ssi
f
i
c
a
t
i
o
n
M
o
d
e
l
f
o
r
P
r
e
d
i
c
t
i
o
n
o
f
A
c
a
d
e
mi
c
P
e
r
f
o
r
ma
n
c
e
o
f
S
t
u
d
e
n
t
s
:
A
n
ED
M
A
p
p
l
i
c
a
t
i
o
n
,
”
L
e
c
t
u
r
e
N
o
t
e
s
i
n
E
l
e
c
t
ri
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
1
0
4
0
LN
EE,
p
p
.
5
9
–
7
1
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
9
8
1
-
99
-
2
2
7
1
-
0
_
6
.
[
1
7
]
K
.
C
h
o
n
g
a
n
d
N
.
S
h
a
h
,
“
C
o
m
p
a
r
i
s
o
n
o
f
n
a
i
v
e
B
a
y
e
s
a
n
d
S
V
M
c
l
a
ssi
f
i
c
a
t
i
o
n
i
n
g
r
i
d
-
se
a
r
c
h
h
y
p
e
r
p
a
r
a
me
t
e
r
t
u
n
e
d
a
n
d
n
on
-
h
y
p
e
r
p
a
r
a
me
t
e
r
t
u
n
e
d
h
e
a
l
t
h
c
a
r
e
s
t
o
c
k
mar
k
e
t
se
n
t
i
me
n
t
a
n
a
l
y
si
s
,
”
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
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
3
,
n
o
.
1
2
,
2
0
2
2
,
d
o
i
:
1
0
.
1
4
5
6
9
/
I
JA
C
S
A
.
2
0
2
2
.
0
1
3
1
2
1
3
.
[
1
8
]
Y
.
Li
,
J.
Zh
a
n
g
,
T.
Li
,
H
.
Li
u
,
J
.
L
i
,
a
n
d
Y
.
W
a
n
g
,
“
G
e
o
g
r
a
p
h
i
c
a
l
t
r
a
c
e
a
b
i
l
i
t
y
o
f
w
i
l
d
B
o
l
e
t
u
s
e
d
u
l
i
s
b
a
s
e
d
o
n
d
a
t
a
f
u
si
o
n
o
f
F
T
-
M
I
R
a
n
d
I
C
P
-
A
ES
c
o
u
p
l
e
d
w
i
t
h
d
a
t
a
m
i
n
i
n
g
me
t
h
o
d
s
(
S
V
M
)
,
”
S
p
e
c
t
r
o
c
h
i
m
i
c
a
A
c
t
a
P
a
rt
A:
Mo
l
e
c
u
l
a
r
a
n
d
Bi
o
m
o
l
e
c
u
l
a
r
S
p
e
c
t
ro
s
c
o
p
y
,
v
o
l
.
1
7
7
,
p
p
.
2
0
–
2
7
,
A
p
r
.
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
sa
a
.
2
0
1
7
.
0
1
.
0
2
9
.
[
1
9
]
A
.
H
.
S
h
a
h
i
d
a
n
d
M
.
P
.
S
i
n
g
h
,
“
S
c
i
e
n
c
e
d
i
r
e
c
t
a
n
o
v
e
l
a
p
p
r
o
a
c
h
f
o
r
c
o
r
o
n
a
r
y
a
r
t
e
r
y
d
i
sea
s
e
d
i
a
g
n
o
s
i
s
u
si
n
g
h
y
b
r
i
d
p
a
r
t
i
c
l
e
sw
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
b
a
se
d
e
m
o
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
B
i
o
c
y
b
e
r
n
e
t
i
c
s
a
n
d
B
i
o
m
e
d
i
c
a
l
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
4
0
,
n
o
.
4
,
p
p
.
1
5
6
8
–
1
5
8
5
,
O
c
t
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
b
e
.
2
0
2
0
.
0
9
.
0
0
5
.
[
2
0
]
X
.
Z
h
a
n
g
,
N
.
H
a
n
,
a
n
d
J.
Zh
a
n
g
,
“
C
o
mp
a
r
a
t
i
v
e
a
n
a
l
y
si
s
o
f
V
G
G
,
R
e
sN
e
t
,
a
n
d
G
o
o
g
L
e
N
e
t
a
r
c
h
i
t
e
c
t
u
r
e
s
e
v
a
l
u
a
t
i
n
g
p
e
r
f
o
r
ma
n
c
e
,
c
o
m
p
u
t
a
t
i
o
n
a
l
e
f
f
i
c
i
e
n
c
y
,
a
n
d
c
o
n
v
e
r
g
e
n
c
e
r
a
t
e
s,”
A
p
p
l
i
e
d
a
n
d
C
o
m
p
u
t
a
t
i
o
n
a
l
En
g
i
n
e
e
r
i
n
g
,
v
o
l
.
4
4
,
n
o
.
1
,
p
p
.
1
7
2
–
1
8
1
,
2
0
2
4
,
d
o
i
:
1
0
.
5
4
2
5
4
/
2
7
5
5
-
2
7
2
1
/
4
4
/
2
0
2
3
0
6
7
6
.
[
2
1
]
I
.
A
.
A
l
h
a
r
b
i
,
A
.
J.
A
l
mal
k
i
,
a
n
d
C
.
C
.
Z
o
u
,
“
H
y
p
e
r
p
a
r
a
m
e
t
e
r
o
p
t
i
mi
z
a
t
i
o
n
a
n
d
c
o
mp
a
r
i
s
o
n
o
f
s
t
u
d
e
n
t
p
e
r
f
o
r
man
c
e
p
r
e
d
i
c
t
i
o
n
a
l
g
o
r
i
t
h
ms,
”
i
n
2
0
2
1
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
m
p
u
t
a
t
i
o
n
a
l
S
c
i
e
n
c
e
a
n
d
C
o
m
p
u
t
a
t
i
o
n
a
l
I
n
t
e
l
l
i
g
e
n
c
e
(
C
S
C
I
)
,
D
e
c
.
2
0
2
1
,
p
p
.
8
8
9
–
8
9
4
.
d
o
i
:
1
0
.
1
1
0
9
/
C
S
C
I
5
4
9
2
6
.
2
0
2
1
.
0
0
2
0
7
.
[
2
2
]
Y
.
A
l
i
,
E
.
A
w
w
a
d
,
M
.
A
l
-
R
a
z
g
a
n
,
a
n
d
A
.
M
a
a
r
o
u
f
,
“
H
y
p
e
r
p
a
r
a
m
e
t
e
r
sea
r
c
h
f
o
r
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms
f
o
r
o
p
t
i
mi
z
i
n
g
t
h
e
c
o
m
p
u
t
a
t
i
o
n
a
l
c
o
m
p
l
e
x
i
t
y
,
”
Pr
o
c
e
sse
s
,
v
o
l
.
1
1
,
n
o
.
2
,
p
.
3
4
9
,
Ja
n
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
p
r
1
1
0
2
0
3
4
9
.
[
2
3
]
B
.
D
r
ă
g
u
l
e
s
c
u
a
n
d
M
.
B
u
c
o
s,
“
H
y
p
e
r
p
a
r
a
me
t
e
r
t
u
n
i
n
g
u
si
n
g
a
u
t
o
ma
t
e
d
met
h
o
d
s
t
o
i
mp
r
o
v
e
m
o
d
e
l
s
f
o
r
p
r
e
d
i
c
t
i
n
g
st
u
d
e
n
t
su
c
c
e
ss,
”
C
o
m
m
u
n
i
c
a
t
i
o
n
s
i
n
C
o
m
p
u
t
e
r
a
n
d
I
n
f
o
rm
a
t
i
o
n
S
c
i
e
n
c
e
,
v
o
l
.
1
2
8
3
C
C
I
S
,
p
p
.
3
0
9
–
3
2
0
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
0
-
5
9
5
0
6
-
7
_
2
5
.
[
2
4
]
A
.
P
.
W
i
n
d
a
r
t
o
,
P
.
A
l
k
h
a
i
r
i
,
a
n
d
J
.
M
u
sl
i
m,
“
E
n
h
a
n
c
i
n
g
p
r
e
mi
e
r
l
e
a
g
u
e
m
a
t
c
h
o
u
t
c
o
me
p
r
e
d
i
c
t
i
o
n
u
si
n
g
s
u
p
p
o
r
t
v
e
c
t
o
r
m
a
c
h
i
n
e
w
i
t
h
e
n
s
e
mb
l
e
t
e
c
h
n
i
q
u
e
s:
a
c
o
m
p
a
r
a
t
i
v
e
st
u
d
y
o
n
b
a
g
g
i
n
g
a
n
d
b
o
o
st
i
n
g
,
”
J
u
rn
a
l
R
ES
T
I
(
Re
k
a
y
a
s
a
S
i
st
e
m
d
a
n
T
e
k
n
o
l
o
g
i
I
n
f
o
rm
a
s
i
)
,
v
o
l
.
9
,
n
o
.
1
,
p
p
.
9
4
–
1
0
3
,
F
e
b
.
2
0
2
5
,
d
o
i
:
1
0
.
2
9
2
0
7
/
r
e
s
t
i
.
v
9
i
1
.
6
1
7
3
.
[
2
5
]
R
.
S
.
G
a
i
k
w
a
d
a
n
d
S
.
C
.
G
a
n
d
a
g
e
,
“
M
C
N
N
:
V
i
s
u
a
l
se
n
t
i
me
n
t
a
n
a
l
y
si
s
u
s
i
n
g
v
a
r
i
o
u
s
d
e
e
p
l
e
a
r
n
i
n
g
f
r
a
mew
o
r
k
w
i
t
h
d
e
e
p
C
N
N
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
S
y
st
e
m
s
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
i
n
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
1
,
n
o
.
2
s
,
p
p
.
2
6
5
–
2
7
8
,
2
0
2
3
.
[
2
6
]
A
.
Jama
l
i
,
S
.
K
.
R
o
y
,
a
n
d
P
.
G
h
a
m
i
si
,
“
W
e
t
M
a
p
F
o
r
m
e
r
:
A
u
n
i
f
i
e
d
d
e
e
p
C
N
N
a
n
d
v
i
s
i
o
n
t
r
a
n
sf
o
r
mer
f
o
r
c
o
m
p
l
e
x
w
e
t
l
a
n
d
map
p
i
n
g
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
A
p
p
l
i
e
d
E
a
r
t
h
O
b
serv
a
t
i
o
n
a
n
d
G
e
o
i
n
f
o
rm
a
t
i
o
n
,
v
o
l
.
1
2
0
,
p
.
1
0
3
3
3
3
,
Ju
n
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
a
g
.
2
0
2
3
.
1
0
3
3
3
3
.
[
2
7
]
O
.
H
a
b
i
b
i
,
M
.
C
h
e
mm
a
k
h
a
,
a
n
d
M
.
L
a
z
a
a
r
,
“
P
e
r
f
o
r
ma
n
c
e
e
v
a
l
u
a
t
i
o
n
o
f
C
N
N
a
n
d
p
r
e
-
t
r
a
i
n
e
d
m
o
d
e
l
s
f
o
r
M
a
l
w
a
r
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
Ara
b
i
a
n
J
o
u
rn
a
l
f
o
r
S
c
i
e
n
c
e
a
n
d
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
4
8
,
n
o
.
8
,
p
p
.
1
0
3
5
5
–
1
0
3
6
9
,
A
u
g
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
3
3
6
9
-
0
2
3
-
0
7
6
0
8
-
z.
[
2
8
]
I
.
A
.
S
h
o
w
mi
k
,
T
.
F
.
S
a
n
a
m,
a
n
d
H
.
I
mt
i
a
z
,
“
H
u
ma
n
a
c
t
i
v
i
t
y
r
e
c
o
g
n
i
t
i
o
n
f
r
o
m
W
i
-
F
i
C
S
I
d
a
t
a
u
si
n
g
p
r
i
n
c
i
p
a
l
c
o
mp
o
n
e
n
t
-
b
a
s
e
d
w
a
v
e
l
e
t
C
N
N
,
”
D
i
g
i
t
a
l
S
i
g
n
a
l
Pr
o
c
e
s
si
n
g
,
v
o
l
.
1
3
8
,
p
.
1
0
4
0
5
6
,
Ju
n
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
d
s
p
.
2
0
2
3
.
1
0
4
0
5
6
.
[
2
9
]
Q
.
D
o
u
,
H
.
C
h
e
n
,
Y
.
Ji
n
,
L.
Y
u
,
J.
Q
i
n
,
a
n
d
P
.
A
.
H
e
n
g
,
“
3
D
d
e
e
p
l
y
su
p
e
r
v
i
se
d
n
e
t
w
o
r
k
f
o
r
a
u
t
o
ma
t
i
c
l
i
v
e
r
se
g
m
e
n
t
a
t
i
o
n
f
r
o
m C
T
v
o
l
u
mes,
”
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
ser
i
e
s
L
e
c
t
u
r
e
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
.
9
9
0
1
LN
C
S
,
p
p
.
1
4
9
–
1
5
7
,
2
0
1
6
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
3
1
9
-
4
6
7
2
3
-
8
_
1
8
.
[
3
0
]
B
.
C
.
A
n
i
l
a
n
d
P
.
D
a
y
a
n
a
n
d
a
,
“
A
u
t
o
mat
i
c
l
i
v
e
r
t
u
m
o
r
se
g
m
e
n
t
a
t
i
o
n
b
a
se
d
o
n
m
u
l
t
i
-
l
e
v
e
l
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
t
w
o
r
k
s
a
n
d
f
r
a
c
t
a
l
r
e
si
d
u
a
l
n
e
t
w
o
r
k
,
”
I
ETE
J
o
u
r
n
a
l
o
f
R
e
se
a
rc
h
,
v
o
l
.
6
9
,
n
o
.
4
,
p
p
.
1
9
2
5
–
1
9
3
3
,
M
a
y
2
0
2
3
,
d
o
i
:
1
0
.
1
0
8
0
/
0
3
7
7
2
0
6
3
.
2
0
2
1
.
1
8
7
8
0
6
6
.
[
3
1
]
N
.
N
.
P
r
a
k
a
sh
,
V
.
R
a
j
e
s
h
,
D
.
L.
N
a
mak
h
w
a
,
S
.
D
w
a
r
k
a
n
a
t
h
P
a
n
d
e
,
a
n
d
S
.
H
.
A
h
a
mm
a
d
,
“
A
D
e
n
seNe
t
C
N
N
-
b
a
se
d
l
i
v
e
r
l
e
si
o
n
p
r
e
d
i
c
t
i
o
n
a
n
d
c
l
a
ssi
f
i
c
a
t
i
o
n
f
o
r
f
u
t
u
r
e
me
d
i
c
a
l
d
i
a
g
n
o
s
i
s,
”
S
c
i
e
n
t
i
f
i
c
A
f
ri
c
a
n
,
v
o
l
.
2
0
,
p
.
e
0
1
6
2
9
,
Ju
l
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
sc
i
a
f
.
2
0
2
3
.
e
0
1
6
2
9
.
[
3
2
]
M
.
A
h
ma
d
e
t
a
l
.
,
“
A
l
i
g
h
t
w
e
i
g
h
t
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
mo
d
e
l
f
o
r
l
i
v
e
r
s
e
g
m
e
n
t
a
t
i
o
n
i
n
me
d
i
c
a
l
d
i
a
g
n
o
si
s
,
”
C
o
m
p
u
t
a
t
i
o
n
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
N
e
u
r
o
sc
i
e
n
c
e
,
v
o
l
.
2
0
2
2
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
5
5
/
2
0
2
2
/
7
9
5
4
3
3
3
.
[
3
3
]
P
.
A
l
k
h
a
i
r
i
a
n
d
Z
.
S
i
t
u
m
o
r
a
n
g
,
“
A
p
p
l
i
c
a
t
i
o
n
o
f
d
a
t
a
mi
n
i
n
g
t
o
a
n
a
l
y
z
e
e
mp
l
o
y
e
e
sat
i
sf
a
c
t
i
o
n
w
i
t
h
H
R
ser
v
i
c
e
s
w
i
t
h
t
h
e
C
4
.
5
a
l
g
o
r
i
t
h
m,
”
[
i
n
B
a
h
a
s
a
]
,
J
u
r
a
si
k
(
J
u
r
n
a
l
Ri
s
e
t
S
i
st
e
m
I
n
f
o
rm
a
si
d
a
n
T
e
k
n
i
k
I
n
f
o
rm
a
t
i
k
a
)
,
v
o
l
.
7
,
n
o
.
1
,
p
.
4
0
,
2
0
2
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
Op
timiz
in
g
co
n
vo
lu
tio
n
a
l
n
e
u
r
a
l n
etw
o
r
k
h
yp
erp
a
r
a
mete
r
s
to
en
h
a
n
ce
liver
…
(
I
w
a
n
P
u
r
n
a
ma
)
3885
AP
P
E
NDI
X
Pro
p
o
s
ed
m
o
d
el
(
Dee
p
L
a
b
V3
Plu
s
)
R
esNet
Evaluation Warning : The document was created with Spire.PDF for Python.