C
omp
u
te
r
S
c
i
e
n
c
e
an
d
I
n
for
mati
on
T
e
c
h
n
ol
ogi
e
s
V
ol
.
7
,
N
o
.
1
,
M
a
r
c
h
20
26
,
p
p.
20
~
29
IS
S
N
:
2722
-
3221
,
D
O
I
:
10.
1
1591
/
c
s
i
t
.
v
7
i
1
.
pp
20
-
29
20
Jou
r
n
al
h
o
m
e
pa
ge
:
ht
t
p:
/
/
i
ae
s
pr
i
m
e
.
c
om
/
i
nd
e
x
.
php
/
c
s
i
t
D
e
v
e
l
o
p
me
n
t
a
n
d
p
e
r
f
o
r
m
a
n
c
e
e
v
a
l
u
a
t
i
o
n
o
f
a
CNN
m
o
d
e
l
f
o
r
sea
g
r
a
ss sp
e
c
i
e
s
c
l
a
ss
i
f
i
c
a
t
i
o
n
i
n
B
i
n
t
a
n
,
In
d
o
n
e
s
i
a
N
u
r
u
l
H
ayaty
1
,
H
ol
l
an
d
a
A
r
i
e
f
K
u
s
u
ma
2
1
D
e
p
a
rt
m
e
n
t
o
f
In
fo
rm
a
t
i
c
s
E
n
g
i
n
e
e
ri
n
g
,
F
a
c
u
l
t
y
o
f
E
n
g
i
n
e
e
ri
n
g
a
n
d
M
a
ri
t
i
m
e
T
e
c
h
n
o
l
o
g
y
,
U
n
i
v
e
rs
i
t
y
o
f
M
a
ri
t
i
m
e
Ra
j
a
A
l
i
H
a
j
i
,
T
a
n
j
u
n
g
p
i
n
a
n
g
,
In
d
o
n
e
s
i
a
2
D
e
p
a
rt
m
e
n
t
o
f
E
l
e
c
t
ri
c
a
l
E
n
g
i
n
e
e
ri
n
g
,
F
a
c
u
l
t
y
o
f
E
n
g
i
n
e
e
ri
n
g
a
n
d
M
a
ri
t
i
m
e
T
e
c
h
n
o
l
o
g
y
,
U
n
i
v
e
rs
i
t
y
o
f
M
a
ri
t
i
m
e
Ra
j
a
A
l
i
H
a
j
i
,
T
a
n
j
u
n
g
p
i
n
a
n
g
,
In
d
o
n
e
s
i
a
A
r
ti
c
l
e
I
n
fo
A
BS
TR
A
C
T
Ar
t
i
c
l
e
h
i
s
t
or
y
:
Re
c
e
i
v
e
d
M
a
y
29
,
2025
Re
vi
s
e
d
A
u
g
1
2,
2025
A
c
c
e
pt
e
d
N
o
v
28
,
2025
T
hi
s
s
t
ud
y
pr
e
s
e
nt
s
t
he
d
e
ve
l
op
m
e
nt
a
nd
e
v
a
l
u
a
t
i
on
o
f
a
c
on
vol
u
t
i
ona
l
ne
ur
a
l
ne
t
w
or
k
(
C
N
N
)
m
od
e
l
f
o
r
a
u
t
o
m
a
t
e
d
s
e
a
gr
a
s
s
s
pe
c
i
e
s
c
l
a
s
s
i
f
i
c
a
t
i
on
i
n
B
i
n
t
a
n,
I
ndon
e
s
i
a
.
T
he
ob
j
e
c
t
i
ve
o
f
t
h
i
s
r
e
s
e
a
r
c
h
i
s
t
o
e
xa
m
i
ne
how
di
f
f
e
r
e
n
t
t
r
a
i
n
-
v
a
l
i
d
a
t
i
on
da
t
a
s
p
l
i
t
r
a
t
i
os
a
f
f
e
c
t
m
ode
l
a
c
c
u
r
a
c
y
a
n
d
ge
ne
r
a
l
i
z
a
t
i
o
n
pe
r
f
o
r
m
a
n
c
e
.
T
he
C
N
N
w
a
s
t
r
a
i
ne
d
unde
r
f
o
ur
c
o
nf
i
gur
a
t
i
ons
(
60
:
40
,
70:
30
,
8
0:
2
0,
a
nd
90:
10)
t
o
a
na
l
y
z
e
t
h
e
i
n
f
l
ue
nc
e
of
t
r
a
i
ni
ng
d
a
t
a
vol
u
m
e
on
l
e
a
r
ni
n
g
c
on
ve
r
ge
n
c
e
a
n
d
pr
e
di
c
t
i
ve
c
a
p
a
bi
l
i
t
y
.
T
h
e
r
e
s
ul
t
s
i
nd
i
c
a
t
e
t
h
a
t
a
l
l
c
on
f
i
gur
a
t
i
on
s
a
c
h
i
e
ve
d
h
i
gh
va
l
i
da
t
i
o
n
a
c
c
u
r
a
c
y,
w
i
t
h
t
he
be
s
t
p
e
r
f
or
m
a
nc
e
r
e
a
c
hi
ng
98
.
53
%
w
h
e
n
us
i
ng
t
h
e
90:
10
s
p
l
i
t
.
E
va
l
ua
t
i
on
on
uns
e
e
n
da
t
a
de
m
o
ns
t
r
a
t
e
d
t
ha
t
t
h
e
60:
40
c
on
f
i
gur
a
t
i
o
n
p
r
ov
i
d
e
d
t
he
m
os
t
c
ons
i
s
t
e
n
t
a
nd
r
e
l
i
a
bl
e
ge
n
e
r
a
l
i
z
a
t
i
on
.
P
e
r
f
or
m
a
nc
e
va
r
i
a
t
i
ons
w
e
r
e
a
l
s
o
a
f
f
e
c
t
e
d
b
y
t
h
e
m
or
p
hol
og
i
c
a
l
s
i
m
i
l
a
r
i
t
y
be
t
w
e
e
n
t
h
e
c
l
a
s
s
i
f
i
e
d
s
p
e
c
i
e
s
,
w
hi
c
h
i
nc
r
e
a
s
e
s
t
he
c
ha
l
l
e
ng
e
i
n
c
or
r
e
c
t
l
y
di
s
t
i
ng
ui
s
hi
ng
c
e
r
t
a
i
n
c
l
a
s
s
e
s
.
O
ve
r
a
l
l
,
t
h
e
f
i
n
di
ngs
c
on
f
i
r
m
t
h
e
e
f
f
e
c
t
i
v
e
ne
s
s
of
C
N
N
-
b
a
s
e
d
c
l
a
s
s
i
f
i
c
a
t
i
on
f
or
s
up
po
r
t
i
ng
m
a
r
i
ne
b
i
od
i
ve
r
s
i
t
y
m
on
i
t
or
i
ng
a
nd
und
e
r
l
i
ne
t
h
e
i
m
po
r
t
a
nc
e
of
d
a
t
a
s
e
t
c
om
pos
i
t
i
o
n
i
n
a
c
hi
e
v
i
ng
o
pt
i
m
a
l
pe
r
f
or
m
a
nc
e
.
F
u
t
u
r
e
i
m
pr
o
ve
m
e
n
t
s
w
i
l
l
f
oc
u
s
on
e
x
pa
nd
i
n
g
da
t
a
va
r
i
a
b
i
l
i
t
y
t
o
e
n
ha
n
c
e
r
obu
s
t
ne
s
s
i
n
r
e
a
l
-
w
o
r
l
d
s
c
e
na
r
i
o
s
.
Ke
y
w
or
d
s
:
D
e
e
p
l
e
a
rn
i
ng
E
a
r
l
y
s
t
opp
i
ng
Im
a
g
e
c
l
a
s
s
i
fi
c
a
t
i
on
M
a
ri
n
e
b
i
od
i
ve
rs
i
t
y
M
orphol
o
gi
c
a
l
s
i
m
i
l
a
r
i
t
y
T
hi
s
i
s
an
op
e
n
ac
c
e
s
s
ar
t
i
c
l
e
u
nde
r
t
he
C
C
B
Y
-
SA
l
i
c
e
n
s
e
.
Cor
r
e
s
pon
di
n
g
Au
t
h
or
:
N
urul
H
a
ya
t
y
D
e
pa
r
t
m
e
nt
of
Info
rm
a
t
i
c
s
E
ng
i
ne
e
ri
ng,
F
a
c
ul
t
y
of
E
ng
i
ne
e
r
i
ng
a
nd
M
a
ri
t
i
m
e
T
e
c
hnol
ogy
U
ni
ve
rs
i
t
y
of
M
a
r
i
t
i
m
e
R
a
j
a
A
l
i
H
a
j
i
T
a
n
j
ungp
i
na
ng,
K
e
pul
a
ua
n
R
i
a
u
,
Indon
e
s
i
a
E
m
a
i
l
:
nuru
l
.
h
a
y
a
t
y
@
u
m
ra
h
.
a
c
.
i
d
1.
I
N
TR
O
D
U
C
TI
O
N
S
e
a
gr
a
s
s
pl
a
ys
a
c
ruc
i
a
l
ro
l
e
i
n
c
o
a
s
t
a
l
e
c
os
ys
t
e
m
s
,
p
rov
i
di
ng
ha
bi
t
a
t
for
m
a
ri
n
e
b
i
ot
a
,
c
a
rbon
s
e
que
s
t
r
a
t
i
on
,
a
nd
c
oa
s
t
a
l
pr
ot
e
c
t
i
on
[1]
,
[2]
.
In
Bi
n
t
a
n
,
s
e
a
gr
a
s
s
e
c
os
ys
t
e
m
s
c
ont
r
i
but
e
s
i
gn
i
fi
c
a
n
t
l
y
t
o
fi
s
he
r
i
e
s
a
nd
m
a
ri
ne
t
ou
ri
s
m
.
S
e
a
gra
s
s
be
ds
i
n
Bi
n
t
a
n
c
a
n
i
nc
re
a
s
e
fi
s
h
re
s
o
urc
e
a
va
i
l
a
bi
l
i
t
y
by
9
,
04
9.
3
kg
pe
r
he
c
t
a
re
a
nn
ua
l
l
y
,
e
qu
i
va
l
e
n
t
t
o
a
ha
b
i
t
a
t
va
l
ue
of
Rp
.
16
6,
963
,
204
.
72
.
S
e
a
g
ra
s
s
a
l
s
o
a
c
t
s
a
s
a
b
l
u
e
c
a
rb
on
s
i
nk,
w
i
t
h
t
he
h
i
ghe
s
t
c
a
rb
on
s
t
ora
ge
foun
d
i
n
P
e
ngud
a
ng
a
t
245.
03
gC
/
m
²
or
348.
26
M
gC/
h
a
.
T
he
i
de
nt
i
f
i
c
a
t
i
on
of
s
e
a
g
ra
s
s
e
s
i
n
c
o
a
s
t
a
l
e
c
os
ys
t
e
m
s
f
a
c
e
s
e
ve
r
a
l
c
ha
l
l
e
ng
e
s
a
nd
o
ppor
t
uni
t
i
e
s
for
i
m
pro
ve
m
e
n
t
.
M
a
nua
l
i
de
nt
i
fi
c
a
t
i
on
of
s
e
a
gra
s
s
s
p
e
c
i
e
s
po
s
e
s
c
h
a
l
l
e
n
ge
s
due
t
o
m
orpho
l
og
i
c
a
l
s
i
m
i
l
a
ri
t
y,
va
ri
a
bl
e
i
m
a
g
i
ng
c
o
ndi
t
i
ons
,
a
nd
t
h
e
ne
e
d
for
e
x
pe
r
t
t
a
x
onom
i
c
know
l
e
dg
e
.
T
hi
s
re
s
e
a
r
c
h
prop
os
e
s
a
n
a
ut
o
m
a
t
e
d
a
ppro
a
c
h
us
i
ng
c
onvo
l
ut
i
on
a
l
ne
ura
l
ne
t
w
orks
(CN
N
s
)
t
o
i
m
prov
e
e
ff
i
c
i
e
n
c
y
a
n
d
a
c
c
ur
a
c
y
i
n
ec
o
l
ogi
c
a
l
s
ur
ve
ys
.
T
ra
di
t
i
on
a
l
m
a
nu
a
l
m
e
t
h
ods
,
w
hi
l
e
c
o
m
m
o
nl
y
us
e
d
,
a
r
e
t
i
m
e
-
c
ons
u
m
i
ng
,
c
os
t
l
y
,
a
nd
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
D
e
v
e
l
opm
e
nt
and
p
e
r
f
or
m
an
c
e
e
v
al
u
at
i
on
of
a
c
nn
m
ode
l
f
or
s
e
agr
as
s
s
p
e
c
i
e
s
…
(
Nur
ul
H
ay
a
t
y
)
21
re
qui
re
a
hi
g
h
l
e
ve
l
of
s
p
e
c
i
a
l
i
z
e
d
e
xp
e
rt
i
s
e
.
E
nv
i
ron
m
e
nt
a
l
fa
c
t
ors
,
s
u
c
h
a
s
va
r
i
a
t
i
ons
i
n
w
a
t
e
r
c
l
a
ri
t
y
,
l
i
ght
c
ondi
t
i
o
ns
,
a
nd
s
e
a
gr
a
s
s
m
orpho
l
ogy
,
f
urt
h
e
r
c
o
m
pl
i
c
a
t
e
t
he
i
de
n
t
i
f
i
c
a
t
i
on
proc
e
s
s
[3
]
.
M
a
c
h
i
ne
l
e
a
rni
ng
a
n
d
i
m
a
g
e
pro
c
e
s
s
i
n
g
t
e
c
hn
i
qu
e
s
a
re
e
m
e
r
gi
ng
a
s
e
ffe
c
t
i
ve
s
o
l
ut
i
ons
t
o
ove
r
c
o
m
e
t
he
s
e
l
i
m
i
t
a
t
i
ons
.
E
ns
e
m
b
l
e
-
ba
s
e
d
m
a
c
hi
n
e
l
e
a
rn
i
ng
m
e
t
hods
,
p
a
rt
i
c
u
l
a
r
l
y
rot
a
t
i
on
f
ore
s
t
s
,
h
a
v
e
ou
t
pe
rfor
m
e
d
t
ra
d
i
t
i
ona
l
m
a
x
i
m
u
m
l
i
k
e
l
i
hood
c
l
a
s
s
i
fi
e
rs
i
n
m
a
ppi
n
g
s
e
a
gr
a
s
s
us
i
ng
S
e
n
t
i
ne
l
-
2
i
m
a
ge
ry
[4]
.
CN
N
s
t
yp
i
c
a
l
l
y
s
urpa
s
s
t
h
e
pe
rf
orm
a
n
c
e
of
t
ra
di
t
i
on
a
l
c
l
a
s
s
i
f
i
c
a
t
i
on
m
e
t
hod
s
w
he
n
a
pp
l
i
e
d
t
o
l
a
rge
-
s
c
a
l
e
da
t
a
s
e
t
s
,
ow
i
ng
t
o
t
he
i
r
a
b
i
l
i
t
y
t
o
a
ut
o
m
a
t
i
c
a
l
l
y
l
e
a
rn
a
nd
e
xt
r
a
c
t
hi
e
ra
rc
h
i
c
a
l
fe
a
t
u
re
s
di
r
e
c
t
l
y
fr
om
r
a
w
i
m
a
ge
d
a
t
a
[5]
,
[6]
.
V
a
ri
ous
a
p
proa
c
h
e
s
for
s
e
a
gr
a
s
s
de
t
e
c
t
i
on
a
nd
m
a
p
pi
ng
h
a
ve
b
e
e
n
i
de
n
t
i
f
i
e
d,
i
n
c
l
u
di
ng
s
t
i
l
l
i
m
a
ge
,
vi
d
e
o
da
t
a
,
a
c
o
us
t
i
c
i
m
a
g
e
,
a
nd
s
p
e
c
t
r
a
l
i
m
a
ge
d
a
t
a
-
b
a
s
e
d
t
e
c
h
ni
que
s
[
6]
–
[8]
.
T
he
t
ra
ns
i
t
i
on
fro
m
t
r
a
d
i
t
i
ona
l
m
a
nu
a
l
a
ppro
a
c
h
e
s
t
o
di
g
i
t
a
l
i
m
a
gi
ng
a
nd
m
a
c
h
i
ne
l
e
a
rni
ng
t
e
c
hni
q
ue
s
re
pr
e
s
e
n
t
s
a
t
r
a
ns
for
m
a
t
i
v
e
s
t
e
p
forw
a
rd
i
n
s
e
a
gr
a
s
s
a
n
d
m
a
r
i
ne
v
e
ge
t
a
t
i
o
n
m
oni
t
or
i
n
g
e
ffo
rt
s
.
Re
c
e
nt
s
t
ud
i
e
s
ha
ve
i
n
c
re
a
s
i
ng
l
y
a
p
pl
i
e
d
CN
N
s
for
s
e
a
gr
a
s
s
c
l
a
s
s
i
f
i
c
a
t
i
on
a
nd
c
o
ns
e
rv
a
t
i
on.
CN
N
-
ba
s
e
d
m
o
de
l
s
ha
ve
be
e
n
d
e
v
e
l
op
e
d
t
o
de
t
e
c
t
a
nd
c
l
a
s
s
i
f
y
s
e
a
gra
s
s
s
p
e
c
i
e
s
from
und
e
rw
a
t
e
r
i
m
a
ge
ry
,
a
c
h
i
e
v
i
ng
hi
g
h
l
e
v
e
l
s
of
a
c
c
ur
a
c
y
—
oft
e
n
e
x
c
e
e
di
ng
90%
[
9],
[10]
.
In
a
ddi
t
i
o
n
t
o
u
nde
rw
a
t
e
r
a
pp
l
i
c
a
t
i
ons
,
CN
N
s
ha
v
e
a
l
s
o
b
e
e
n
ut
i
l
i
z
e
d
for
a
n
a
l
y
z
i
n
g
h
i
gh
-
r
e
s
ol
u
t
i
o
n
s
a
t
e
l
l
i
t
e
i
m
a
ge
ry
t
o
m
a
p
b
e
nt
h
i
c
h
a
bi
t
a
t
s
a
nd
m
oni
t
or
s
e
a
gr
a
s
s
d
i
s
t
ri
but
i
on
i
n
s
h
a
l
l
ow
m
a
ri
n
e
e
n
vi
ron
m
e
n
t
s
[
11]
.
F
o
r
i
ns
t
a
nc
e
,
N
o
m
a
n
e
t
al
.
[12]
i
m
p
l
e
m
e
n
t
e
d
a
CN
N
-
b
a
s
e
d
a
pp
roa
c
h
t
ha
t
a
c
hi
e
ve
d
a
n
a
c
c
ur
a
c
y
of
99
.
33%
i
n
s
e
a
gra
s
s
c
l
a
s
s
i
fi
c
a
t
i
o
n.
O
z
a
e
t
a
e
t
al
.
[13]
i
n
t
rodu
c
e
d
a
de
e
p
l
e
a
rn
i
ng
m
e
t
ho
d
us
i
ng
di
ff
e
re
nt
i
a
b
l
e
a
r
c
hi
t
e
c
t
u
re
s
e
a
rc
h
,
re
a
c
h
i
ng
93.
72
%
a
c
c
u
ra
c
y
i
n
c
l
a
s
s
i
fy
i
ng
f
i
v
e
s
e
a
gra
s
s
s
p
e
c
i
e
s
fro
m
t
h
e
P
hi
l
i
pp
i
ne
s
.
M
e
a
nw
h
i
l
e
,
R
e
us
[1
4]
fo
c
us
e
d
on
s
e
a
gr
a
s
s
s
e
g
m
e
n
t
a
t
i
o
n,
de
m
ons
t
ra
t
i
ng
t
h
e
e
ff
e
c
t
i
v
e
ne
s
s
of
CN
N
fe
a
t
ur
e
s
i
n
e
s
t
i
m
a
t
i
ng
s
e
a
gr
a
s
s
c
ove
r
a
g
e
,
w
i
t
h
a
n
a
c
c
u
ra
c
y
of
9
4.
5
%.
Re
c
e
nt
re
s
e
a
r
c
h
ha
s
foc
us
e
d
o
n
e
nh
a
nc
i
ng
t
h
e
p
e
rfor
m
a
nc
e
of
CN
N
s
i
n
c
l
a
s
s
i
fi
c
a
t
i
on
t
a
s
ks
t
hrough
da
t
a
opt
i
m
i
z
a
t
i
on.
I
n
[
15]
,
t
he
i
m
p
a
c
t
of
di
f
fe
r
e
nt
t
ra
i
n
-
t
e
s
t
s
pl
i
t
r
a
t
i
os
o
n
CN
N
a
c
c
ur
a
c
y
fo
r
E
E
G
e
m
o
t
i
o
n
re
c
og
ni
t
i
on
w
a
s
i
nv
e
s
t
i
g
a
t
e
d
,
r
e
ve
a
l
i
ng
t
h
a
t
a
n
80:
2
0
s
pl
i
t
pro
duc
e
d
o
pt
i
m
a
l
re
s
ul
t
s
.
L
i
ke
w
i
s
e
,
[16
]
de
m
o
ns
t
ra
t
e
d
t
ha
t
m
odi
f
yi
ng
da
t
a
s
e
t
c
on
fi
gur
a
t
i
ons
—
s
u
c
h
a
s
c
l
a
s
s
b
a
l
a
n
c
e
a
nd
da
t
a
prop
ort
i
on
—
s
ubs
t
a
nt
i
a
l
l
y
i
m
prov
e
d
CN
N
pe
r
form
a
n
c
e
i
n
c
h
e
s
t
X
-
r
a
y
i
m
a
g
e
c
l
a
s
s
i
fi
c
a
t
i
on
.
A
c
o
m
pre
he
ns
i
ve
s
t
udy
by
A
ba
di
e
t
a
l
.
[17]
e
va
l
u
a
t
e
d
c
l
a
s
s
i
m
ba
l
a
nc
e
e
ff
e
c
t
s
i
n
CN
N
m
ode
l
s
,
c
on
c
l
ud
i
ng
t
ha
t
ov
e
rs
a
m
pl
i
ng
s
t
ra
t
e
g
i
e
s
offe
r
e
d
t
h
e
m
os
t
s
t
a
bl
e
s
o
l
ut
i
o
n
t
o
i
m
b
a
l
a
nc
e
i
s
s
ue
s
w
i
t
hou
t
i
nd
uc
i
ng
ove
rfi
t
t
i
ng.
F
ur
t
he
r
m
or
e
,
P
r
e
c
h
e
l
t
[
18]
propos
e
d
a
ra
t
i
on
a
l
e
-
b
a
s
e
d
CN
N
m
ode
l
fo
r
t
e
x
t
c
l
a
s
s
i
f
i
c
a
t
i
o
n,
w
h
i
c
h
i
n
t
e
g
ra
t
e
d
s
e
nt
e
n
c
e
-
l
e
v
e
l
j
us
t
i
fi
c
a
t
i
ons
a
nd
a
c
h
i
e
v
e
d
h
i
gh
a
c
c
ur
a
c
y
a
c
ros
s
m
ul
t
i
pl
e
b
e
nc
hm
a
rk
d
a
t
a
s
e
t
s
.
T
o
g
e
t
h
e
r,
t
h
e
s
e
f
i
nd
i
ngs
und
e
rs
c
ore
t
h
e
pi
vot
a
l
ro
l
e
o
f
bot
h
d
a
t
a
s
t
ru
c
t
ur
i
ng
a
nd
a
r
c
hi
t
e
c
t
u
ra
l
de
s
i
g
n
i
n
m
a
xi
m
i
z
i
ng
t
he
e
ff
e
c
t
i
v
e
ne
s
s
of
CN
N
s
for
c
l
a
s
s
i
fi
c
a
t
i
on
t
a
s
ks
.
T
he
s
e
fi
nd
i
ngs
h
i
ghl
i
gh
t
t
h
e
s
i
gn
i
fi
c
a
nt
po
t
e
nt
i
a
l
of
d
e
e
p
l
e
a
rni
n
g
t
e
c
hn
i
qu
e
s
i
n
e
nh
a
nc
i
ng
t
he
a
c
c
ura
c
y
a
nd
a
u
t
om
a
t
i
on
of
s
e
a
gr
a
s
s
c
l
a
s
s
i
f
i
c
a
t
i
on.
A
s
s
e
a
gra
s
s
e
c
os
ys
t
e
m
s
fa
c
e
i
n
c
r
e
a
s
i
n
g
t
hre
a
t
s
f
rom
hum
a
n
a
c
t
i
v
i
t
i
e
s
a
nd
c
l
i
m
a
t
e
c
ha
nge
,
s
uc
h
a
dva
n
c
e
m
e
nt
s
a
r
e
c
r
i
t
i
c
a
l
for
i
m
prov
i
ng
t
h
e
e
ff
i
c
i
e
n
c
y
a
n
d
s
c
a
l
a
bi
l
i
t
y
of
m
oni
t
ori
ng
a
nd
c
ons
e
rva
t
i
on
e
f
fort
s
.
By
ut
i
l
i
z
i
ng
a
r
e
pre
s
e
nt
a
t
i
ve
da
t
a
s
e
t
of
s
e
a
gr
a
s
s
i
m
a
g
e
s
,
m
a
c
hi
n
e
l
e
a
rni
n
g
m
o
de
l
s
c
a
n
b
e
opt
i
m
i
z
e
d
t
o
a
ut
o
m
a
t
i
c
a
l
l
y
de
t
e
c
t
a
nd
c
l
a
s
s
i
fy
s
e
a
gr
a
s
s
s
pe
c
i
e
s
,
t
hus
i
m
pro
vi
ng
t
he
e
ff
i
c
i
e
n
c
y
of
t
h
e
m
on
i
t
or
i
ng
pro
c
e
s
s
.
T
hi
s
r
e
s
e
a
rc
h
fo
c
us
e
s
on
t
he
de
ve
l
opm
e
n
t
of
a
m
a
c
h
i
ne
l
e
a
rni
n
g
m
ode
l
t
o
c
l
a
s
s
i
fy
t
hr
e
e
s
e
a
g
ra
s
s
s
pe
c
i
e
s
found
i
n
t
he
c
oa
s
t
a
l
r
e
gi
on
of
B
i
nt
a
n.
S
e
a
g
ra
s
s
i
m
a
g
e
s
t
a
k
e
n
fro
m
t
he
s
t
u
dy
l
o
c
a
t
i
ons
w
i
l
l
u
nde
r
go
a
s
e
ri
e
s
of
pre
p
roc
e
s
s
i
ng
s
t
e
ps
a
n
d
w
i
l
l
be
us
e
d
t
o
t
r
a
i
n
t
he
m
a
c
h
i
ne
l
e
a
rni
ng
m
ode
l
.
T
hrou
gh
t
h
i
s
a
pp
roa
c
h
,
i
t
i
s
e
xp
e
c
t
e
d
t
ha
t
t
h
e
r
e
s
e
a
r
c
h
w
i
l
l
m
a
ke
a
m
e
a
n
i
ngf
ul
c
ont
r
i
b
ut
i
on
i
n
s
i
m
pl
i
fy
i
ng
t
he
i
de
n
t
i
f
i
c
a
t
i
on
of
s
e
a
gr
a
s
s
s
pe
c
i
e
s
a
c
c
ur
a
t
e
l
y
a
nd
e
ffi
c
i
e
nt
l
y.
T
he
obj
e
c
t
i
v
e
of
t
hi
s
s
t
ud
y
i
s
t
o
de
s
i
gn
a
nd
e
v
a
l
u
a
t
e
t
he
p
e
rfor
m
a
n
c
e
of
a
m
a
c
hi
n
e
l
e
a
rni
n
g
m
o
de
l
i
n
c
l
a
s
s
i
fy
i
ng
s
e
a
gr
a
s
s
s
pe
c
i
e
s
ba
s
e
d
o
n
t
h
e
a
va
i
l
a
bl
e
i
m
a
ge
d
a
t
a
s
e
t
.
D
e
s
pi
t
e
grow
i
ng
r
e
s
e
a
rc
h
o
n
s
e
a
g
ra
s
s
c
l
a
s
s
i
f
i
c
a
t
i
on,
t
he
re
r
e
m
a
i
ns
a
l
a
c
k
of
l
oc
a
l
i
z
e
d
,
s
p
e
c
i
e
s
-
s
p
e
c
i
fi
c
m
o
de
l
s
t
ra
i
ne
d
on
c
ur
a
t
e
d
da
t
a
s
e
t
s
fro
m
In
done
s
i
a
.
T
hi
s
s
t
udy
a
ddr
e
s
s
e
s
t
h
a
t
g
a
p
by
c
ons
t
ru
c
t
i
ng
a
r
obus
t
CN
N
t
r
a
i
n
e
d
on
ha
nd
-
l
a
b
e
l
e
d
Bi
n
t
a
n
s
e
a
gr
a
s
s
i
m
a
ge
ry,
w
i
t
h
pe
rfor
m
a
n
c
e
e
va
l
ua
t
i
on
a
c
r
os
s
v
a
ri
e
d
da
t
a
c
om
p
os
i
t
i
ons
a
nd
de
pl
o
ym
e
nt
on
m
ob
i
l
e
-
fri
e
nd
l
y
pl
a
t
for
m
s
s
uc
h
a
s
T
e
ns
orF
l
o
w
L
i
t
e
(
T
F
L
i
t
e
)
for
l
i
ght
w
e
i
g
ht
i
n
fe
r
e
nc
e
i
n
f
i
e
l
d
e
nvi
r
onm
e
nt
s
.
2.
M
ET
H
O
D
2.
1
.
R
e
s
e
ar
c
h
f
r
am
e
w
or
k
T
hi
s
s
t
ud
y
e
m
pl
o
ys
a
s
upe
r
vi
s
e
d
de
e
p
l
e
a
rn
i
ng
a
ppro
a
c
h
u
t
i
l
i
z
i
ng
a
CN
N
a
rc
hi
t
e
c
t
ur
e
t
o
c
l
a
s
s
i
fy
t
hre
e
s
pe
c
i
e
s
of
s
e
a
g
ra
s
s
fou
nd
i
n
t
h
e
c
oa
s
t
a
l
w
a
t
e
rs
of
Bi
nt
a
n:
H
al
o
dul
e
un
i
ne
r
v
i
s
,
Sy
r
i
ngod
i
um
i
s
oe
t
i
f
o
l
i
um
,
a
nd
T
hal
as
s
i
a
he
m
pr
i
c
h
i
i
.
CN
N
s
a
re
p
a
rt
i
c
ul
a
rl
y
s
ui
t
a
bl
e
f
or
i
m
a
ge
-
ba
s
e
d
c
l
a
s
s
i
f
i
c
a
t
i
on
t
a
s
ks
du
e
t
o
t
he
i
r
a
bi
l
i
t
y
t
o
a
u
t
om
a
t
i
c
a
l
l
y
e
xt
r
a
c
t
s
pa
t
i
a
l
f
e
a
t
ur
e
s
from
ra
w
pi
x
e
l
d
a
t
a
[16]
.
T
he
re
s
e
a
r
c
h
m
e
t
hodo
l
ogy
c
ons
i
s
t
s
of
ni
n
e
pr
i
m
a
ry
s
t
a
g
e
s
,
a
s
i
l
l
us
t
ra
t
e
d
i
n
F
i
gu
r
e
1
.
I
m
a
ge
da
t
a
w
e
re
c
o
l
l
e
c
t
e
d
from
a
c
ur
a
t
e
d
r
e
pos
i
t
ory
a
nd
groupe
d
i
n
t
o
s
pe
c
i
e
s
-
s
pe
c
i
f
i
c
c
l
a
s
s
e
s
.
P
r
e
pro
c
e
s
s
i
n
g
i
n
c
l
u
d
e
d
re
s
i
z
i
ng
,
nor
m
a
l
i
z
a
t
i
on
,
a
nd
a
ugm
e
nt
a
t
i
on
t
o
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
7
,
N
o
.
1
,
M
a
rc
h
20
26
:
20
-
29
22
s
t
a
nd
a
rdi
z
e
i
npu
t
s
a
nd
e
nha
n
c
e
da
t
a
s
e
t
di
v
e
rs
i
t
y,
f
ol
l
ow
e
d
by
l
a
b
e
l
e
nc
o
di
ng
t
o
pre
pa
r
e
da
t
a
fo
r
m
od
e
l
t
ra
i
ni
ng
.
T
he
da
t
a
s
e
t
w
a
s
s
p
l
i
t
us
i
ng
f
our
t
ra
i
n
-
t
e
s
t
c
onfi
gura
t
i
o
ns
(60
:
40,
70:
3
0,
8
0:
20
,
90
:
10)
w
i
t
h
s
t
ra
t
i
f
i
e
d
s
a
m
pl
i
ng
t
o
m
a
i
n
t
a
i
n
c
l
a
s
s
ba
l
a
n
c
e
.
T
h
e
CN
N
a
r
c
hi
t
e
c
t
ur
e
c
o
m
pri
s
e
d
c
on
vol
u
t
i
o
na
l
a
nd
po
ol
i
ng
l
a
ye
rs
fol
l
ow
e
d
by
ful
l
y
c
o
nne
c
t
e
d
l
a
ye
rs
w
i
t
h
dropou
t
re
g
ul
a
ri
z
a
t
i
on.
T
he
m
ode
l
w
a
s
t
r
a
i
n
e
d
us
i
ng
t
he
A
da
m
opt
i
m
i
z
e
r
a
nd
c
a
t
e
g
ori
c
a
l
c
ros
s
-
e
n
t
ropy
l
os
s
,
w
i
t
h
e
a
rl
y
s
t
o
p
pi
ng
a
nd
c
he
c
kpo
i
nt
i
ng
t
o
pre
ve
n
t
o
ve
rf
i
t
t
i
ng
a
nd
pre
s
e
rv
e
t
h
e
b
e
s
t
-
p
e
rfor
m
i
ng
w
e
i
g
ht
s
.
M
o
de
l
p
e
rf
or
m
a
n
c
e
w
a
s
e
v
a
l
u
a
t
e
d
us
i
ng
a
c
c
u
r
a
c
y
a
n
d
c
on
fu
s
i
on
m
a
t
r
i
x
a
n
a
l
ys
i
s
on
i
n
d
e
p
e
n
de
n
t
t
e
s
t
d
a
t
a
.
T
h
e
op
t
i
m
a
l
m
od
e
l
w
a
s
t
h
e
n
c
o
n
v
e
r
t
e
d
i
n
t
o
T
e
ns
o
rF
l
o
w
L
i
t
e
fo
r
m
a
t
t
o
s
u
p
po
r
t
l
i
g
ht
w
e
i
g
h
t
d
e
p
l
o
y
m
e
n
t
.
F
i
n
a
l
l
y
,
i
n
f
e
r
e
n
c
e
w
a
s
p
e
r
fo
r
m
e
d
o
n
u
ns
e
e
n
i
m
a
g
e
s
,
a
n
d
pr
e
d
i
c
t
i
o
ns
w
e
r
e
c
o
m
p
a
r
e
d
a
g
a
i
ns
t
g
ro
u
nd
-
t
r
u
t
h
l
a
be
l
s
.
F
i
gure
1
.
R
e
s
e
a
rc
h
w
ork
fl
ow
for
s
e
a
gra
s
s
s
p
e
c
i
e
s
c
l
a
s
s
i
f
i
c
a
t
i
on
us
i
n
g
CN
N
A
l
l
e
xp
e
ri
m
e
n
t
s
w
e
re
e
x
e
c
u
t
e
d
on
G
oogl
e
Col
a
b
us
i
ng
G
P
U
a
c
c
e
l
e
r
a
t
i
on
.
T
e
ns
orF
l
ow
a
nd
K
e
r
a
s
w
e
re
a
pp
l
i
e
d
fo
r
m
ode
l
d
e
v
e
l
op
m
e
nt
,
w
h
i
l
e
N
u
m
P
y,
P
a
n
da
s
,
a
nd
P
IL
s
uppor
t
e
d
pre
proc
e
s
s
i
ng
a
nd
d
a
t
a
m
a
n
i
pul
a
t
i
on
.
T
h
e
c
om
pl
e
t
e
w
or
kfl
ow
,
i
m
pl
e
m
e
nt
a
t
i
on
s
c
r
i
pt
s
,
a
nd
m
od
e
l
f
i
l
e
s
a
r
e
a
v
a
i
l
a
b
l
e
i
n
a
n
ope
n
-
a
c
c
e
s
s
G
i
t
H
ub
re
pos
i
t
ory,
e
ns
uri
n
g
r
e
prod
uc
i
bi
l
i
t
y
a
nd
s
c
a
l
a
bi
l
i
t
y
.
2.
2
.
D
ata
c
o
l
l
e
c
t
i
on
an
d
d
atas
e
t
c
omp
os
i
t
i
on
T
he
da
t
a
s
e
t
us
e
d
i
n
t
hi
s
s
t
udy
c
o
ns
i
s
t
s
o
f
c
l
os
e
-
up
p
hot
ogra
phs
of
i
ndi
vi
du
a
l
s
e
a
gra
s
s
l
e
a
ve
s
re
pre
s
e
nt
i
ng
t
hr
e
e
t
rop
i
c
a
l
s
p
e
c
i
e
s
:
H
.
uni
n
e
r
v
i
s
,
S
.
i
s
oe
t
i
f
ol
i
um
,
and
T
.
h
e
m
pr
i
c
hi
i
.
A
l
l
s
a
m
pl
e
s
w
e
r
e
c
ol
l
e
c
t
e
d
from
t
he
c
oa
s
t
a
l
w
a
t
e
rs
of
B
i
nt
a
n
Is
l
a
nd
,
Indo
ne
s
i
a
.
U
nl
i
ke
i
n
-
s
i
t
u
unde
rw
a
t
e
r
i
m
a
ge
ry
,
e
a
c
h
s
a
m
pl
e
w
a
s
m
a
nu
a
l
l
y
e
xt
r
a
c
t
e
d
fro
m
i
t
s
h
a
bi
t
a
t
,
c
l
e
a
ne
d
of
de
br
i
s
,
a
nd
pho
t
ogra
phe
d
a
g
a
i
ns
t
a
p
l
a
i
n
w
hi
t
e
ba
c
k
ground
und
e
r
c
on
t
ro
l
l
e
d
l
i
gh
t
i
ng
c
ond
i
t
i
ons
t
o
m
i
ni
m
i
z
e
s
h
a
dow
s
a
nd
noi
s
e
a
s
s
how
n
i
n
F
i
g
ure
2
.
E
a
c
h
s
pe
c
i
e
s
w
a
s
i
ni
t
i
a
l
l
y
r
e
pr
e
s
e
nt
e
d
by
40
uni
q
ue
l
e
a
f
i
m
a
ge
s
,
r
e
s
ul
t
i
ng
i
n
a
t
ot
a
l
of
120
ori
g
i
na
l
i
m
a
ge
s
a
c
ros
s
a
l
l
c
l
a
s
s
e
s
.
T
h
e
s
e
i
m
a
ge
s
w
e
r
e
s
t
ore
d
i
n
J
P
E
G
f
orm
a
t
a
nd
e
x
hi
bi
t
e
d
v
a
ri
a
t
i
ons
i
n
ori
e
n
t
a
t
i
on
,
s
i
z
e
,
s
ha
pe
,
a
nd
t
e
x
t
ure
—
fa
c
t
o
rs
t
ha
t
c
ont
r
i
bu
t
e
t
o
s
pe
c
i
e
s
d
i
ff
e
re
n
t
i
a
t
i
on.
H
.
u
n
i
n
e
r
v
i
s
S
.
i
s
o
e
t
i
f
o
l
i
u
m
T
.
h
e
m
p
r
i
c
h
i
i
F
i
gure
2
.
E
xa
m
p
l
e
of
e
a
c
h
s
e
a
gra
s
s
c
l
a
s
s
d
a
t
a
2.
3
.
I
mag
e
au
g
me
n
tat
i
on
an
d
p
r
e
p
r
o
c
e
s
s
i
n
g
T
he
l
i
m
i
t
e
d
ra
w
da
t
a
s
e
t
,
c
o
ns
i
s
t
i
ng
of
onl
y
40
i
m
a
g
e
s
pe
r
s
pe
c
i
e
s
,
w
a
s
e
xp
a
nd
e
d
us
i
ng
off
l
i
ne
a
ugm
e
n
t
a
t
i
on
pr
i
or
t
o
m
ode
l
de
v
e
l
o
pm
e
nt
.
E
a
c
h
i
m
a
ge
und
e
rw
e
n
t
1
6
t
ra
ns
form
a
t
i
on
va
ri
a
t
i
o
ns
,
re
s
u
l
t
i
ng
i
n
17
t
ot
a
l
v
e
rs
i
o
ns
p
e
r
s
a
m
pl
e
,
i
n
c
l
ud
i
ng
t
he
or
i
gi
na
l
.
T
hi
s
proc
e
s
s
i
nc
r
e
a
s
e
d
t
h
e
da
t
a
s
e
t
t
o
680
i
m
a
ge
s
pe
r
c
l
a
s
s
,
y
i
e
l
di
ng
2,
0
40
i
m
a
g
e
s
a
c
r
os
s
t
h
e
t
hr
e
e
s
p
e
c
i
e
s
.
A
u
gm
e
n
t
a
t
i
o
n
i
nc
l
ud
e
d
ge
o
m
e
t
ri
c
t
r
a
ns
for
m
a
t
i
ons
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
D
e
v
e
l
opm
e
nt
and
p
e
r
f
or
m
an
c
e
e
v
al
u
at
i
on
of
a
c
nn
m
ode
l
f
or
s
e
agr
as
s
s
p
e
c
i
e
s
…
(
Nur
ul
H
ay
a
t
y
)
23
(fl
i
ps
,
rot
a
t
i
on
,
re
s
i
z
e
,
z
oo
m
,
s
hi
ft
s
)
,
ph
ot
o
m
e
t
ri
c
a
dj
us
t
m
e
nt
s
(br
i
gh
t
ne
s
s
,
c
on
t
ra
s
t
),
a
nd
di
s
t
o
rt
i
on
s
(G
a
us
s
i
a
n
no
i
s
e
a
nd
s
he
a
r)
.
F
o
l
l
ow
i
ng
a
ug
m
e
nt
a
t
i
o
n,
a
l
l
i
m
a
ge
s
w
e
r
e
i
m
po
rt
e
d
i
nt
o
t
h
e
G
oogl
e
Col
a
b
e
nvi
r
onm
e
nt
for
pr
e
pro
c
e
s
s
i
ng
.
I
m
a
g
e
s
w
e
r
e
re
s
i
z
e
d
t
o
1
00×
100
pi
x
e
l
s
,
nor
m
a
l
i
z
e
d
t
o
a
p
i
xe
l
va
l
u
e
ra
ng
e
of
[0,
1]
,
a
nd
a
s
s
i
gn
e
d
nu
m
e
ri
c
c
l
a
s
s
l
a
b
e
l
s
for
c
o
m
pa
t
i
bi
l
i
t
y
w
i
t
h
t
he
m
od
e
l
'
s
c
a
t
e
gor
i
c
a
l
l
os
s
fun
c
t
i
on.
T
h
e
da
t
a
s
e
t
w
a
s
t
h
e
n
s
pl
i
t
i
n
t
o
t
ra
i
ni
ng
a
nd
va
l
i
d
a
t
i
on
s
ubs
e
t
s
u
s
i
ng
s
t
ra
t
i
fi
e
d
s
a
m
pl
i
ng
w
hi
l
e
m
a
i
n
t
a
i
ni
n
g
c
l
a
s
s
ba
l
a
nc
e
a
c
ros
s
four
t
ra
i
n
–
t
e
s
t
ra
t
i
os
:
60
:
40
,
70:
30
,
80:
20,
a
nd
90:
10
.
2.
4
.
C
N
N
mod
e
l
a
r
c
h
i
t
e
c
tu
r
e
T
he
CN
N
us
e
d
i
n
t
hi
s
s
t
udy
w
a
s
de
v
e
l
o
pe
d
t
o
c
l
a
s
s
i
fy
s
e
a
gra
s
s
s
pe
c
i
e
s
b
a
s
e
d
on
s
i
n
gl
e
-
l
e
a
f
i
n
put
i
m
a
ge
s
.
T
h
e
a
rc
h
i
t
e
c
t
ur
e
w
a
s
i
m
p
l
e
m
e
n
t
e
d
us
i
ng
t
h
e
S
e
que
nt
i
a
l
A
P
I
pro
vi
d
e
d
by
T
e
ns
orF
l
ow
'
s
K
e
r
a
s
l
i
br
a
ry
[17
]
a
nd
c
ons
i
s
t
s
of
c
o
nvol
ut
i
ona
l
l
a
ye
rs
fo
r
fe
a
t
ur
e
e
x
t
ra
c
t
i
on
fo
l
l
ow
e
d
by
d
e
ns
e
l
a
y
e
rs
for
f
i
n
a
l
c
l
a
s
s
i
fi
c
a
t
i
on
.
T
h
e
m
od
e
l
a
c
c
e
p
t
s
RG
B
i
m
a
g
e
s
r
e
s
i
z
e
d
t
o
1
0
0×
100
×
3.
F
e
a
t
ure
l
e
a
rni
n
g
i
s
p
e
rfor
m
e
d
t
hro
ugh
t
hre
e
c
onvol
ut
i
on
–
m
a
x
pool
i
ng
b
l
o
c
ks
w
i
t
h
i
n
c
re
a
s
i
ng
fi
l
t
e
r
de
pt
h
(
32,
64,
a
nd
128
f
i
l
t
e
rs
)
a
nd
r
e
c
t
i
f
i
e
d
l
i
n
e
a
r
un
i
t
(
R
e
L
U
)
a
c
t
i
v
a
t
i
ons
,
e
a
c
h
fol
l
ow
e
d
by
a
2
×
2
m
a
x
-
pool
i
ng
l
a
y
e
r.
T
he
f
i
na
l
c
on
vol
u
t
i
o
n
s
t
a
g
e
produc
e
s
a
10
×
10×
128
fe
a
t
ure
m
a
p
,
w
hi
c
h
i
s
fl
a
t
t
e
n
e
d
be
fore
e
n
t
e
r
i
ng
t
he
c
l
a
s
s
i
fi
e
r
s
t
a
g
e
.
T
h
e
c
l
a
s
s
i
fi
e
r
c
ons
i
s
t
s
of
a
D
e
ns
e
l
a
y
e
r
w
i
t
h
128
n
e
urons
,
fol
l
ow
e
d
by
a
0
.
4
dr
opout
r
a
t
e
t
o
m
i
t
i
g
a
t
e
ove
rfi
t
t
i
ng.
T
he
o
ut
pu
t
l
a
y
e
r
c
on
t
a
i
ns
t
hr
e
e
ne
urons
,
r
e
pr
e
s
e
nt
i
ng
t
he
s
e
a
gra
s
s
c
l
a
s
s
e
s
,
w
i
t
h
a
s
of
t
m
a
x
a
c
t
i
va
t
i
on
for
pr
oba
b
i
l
i
t
y
-
ba
s
e
d
pr
e
di
c
t
i
on.
In
t
o
t
a
l
,
t
he
m
o
de
l
c
ont
a
i
ns
a
pprox
i
m
a
t
e
l
y
1.
7
m
i
l
l
i
on
t
r
a
i
n
a
bl
e
pa
r
a
m
e
t
e
rs
.
2.
5
.
M
od
e
l
t
r
ai
n
i
n
g
T
he
CN
N
m
ode
l
w
a
s
t
r
a
i
ne
d
us
i
ng
t
h
e
T
e
ns
orF
l
ow
-
K
e
ra
s
f
ra
m
e
w
or
k
i
n
t
h
e
G
oog
l
e
Co
l
a
b
e
nvi
r
onm
e
nt
.
T
r
a
i
ni
ng
w
a
s
p
e
rfor
m
e
d
o
n
t
he
a
ug
m
e
n
t
e
d
d
a
t
a
s
e
t
c
ont
a
i
n
i
ng
680
i
m
a
ge
s
p
e
r
c
l
a
s
s
,
re
s
u
l
t
i
ng
i
n
a
t
ot
a
l
of
2
,
040
i
m
a
ge
s
.
S
e
ve
r
a
l
t
ra
i
n
-
va
l
i
d
a
t
i
on
s
pl
i
t
ra
t
i
os
w
e
re
e
x
pl
or
e
d
,
n
a
m
e
l
y
60:
4
0,
70
:
30
,
8
0:
20
,
a
nd
90:
10
,
w
i
t
h
c
l
a
s
s
b
a
l
a
nc
e
pr
e
s
e
rv
e
d
t
hro
ugh
s
t
ra
t
i
f
i
e
d
s
a
m
pl
i
ng.
T
h
e
da
t
a
w
a
s
s
pl
i
t
i
n
t
o
t
ra
i
ni
ng
a
nd
t
e
s
t
i
ng
s
ubs
e
t
s
unde
r
t
he
s
e
four
c
o
nfi
g
ura
t
i
ons
,
a
s
s
how
n
i
n
T
a
bl
e
1.
E
a
c
h
c
om
p
os
i
t
i
on
m
a
i
n
t
a
i
n
e
d
a
t
o
t
a
l
of
2,
0
40
i
m
a
ge
s
but
v
a
ri
e
d
t
h
e
a
l
l
oc
a
t
i
on
be
t
w
e
e
n
t
ra
i
ni
n
g
a
nd
t
e
s
t
i
n
g
s
e
t
s
.
T
he
m
o
de
l
w
a
s
c
o
m
pi
l
e
d
us
i
n
g
t
h
e
A
d
a
m
op
t
i
m
i
z
e
r
,
s
e
l
e
c
t
e
d
for
i
t
s
a
da
pt
i
ve
l
e
a
rn
i
ng
r
a
t
e
a
nd
s
t
a
b
l
e
c
onv
e
rge
nc
e
be
h
a
vi
o
r.
T
he
l
e
a
rni
n
g
ra
t
e
w
a
s
s
e
t
t
o
0
.
001
,
w
i
t
h
s
t
a
n
da
rd
h
yp
e
rp
a
r
a
m
e
t
e
rs
:
β
₁
=
0.
9
,
β
₂
=
0.
9
99,
a
nd
ε
=
1
e
-
7.
T
h
e
c
h
os
e
n
l
os
s
fu
nc
t
i
on
w
a
s
s
pa
rs
e
c
a
t
e
gori
c
a
l
c
r
os
s
e
nt
r
opy
,
s
ui
t
a
bl
e
for
m
u
l
t
i
-
c
l
a
s
s
c
l
a
s
s
i
fi
c
a
t
i
on
w
i
t
h
i
nt
e
ge
r
-
e
n
c
od
e
d
l
a
b
e
l
s
.
M
od
e
l
pe
rfor
m
a
nc
e
du
ri
ng
t
r
a
i
ni
ng
w
a
s
m
on
i
t
or
e
d
us
i
ng
t
h
e
a
c
c
ura
c
y
m
e
t
r
i
c
.
T
ra
i
ni
n
g
w
a
s
c
ond
uc
t
e
d
i
n
m
i
ni
-
b
a
t
c
he
s
of
32
s
a
m
pl
e
s
,
a
s
de
fi
n
e
d
by
t
he
ba
t
c
h_s
i
z
e
p
a
ra
m
e
t
e
r
i
n
t
he
d
a
t
a
g
e
n
e
ra
t
or
.
A
l
t
ho
ugh
t
h
e
m
a
xi
m
u
m
num
be
r
of
t
ra
i
ni
ng
e
p
oc
hs
w
a
s
s
e
t
t
o
250
,
a
n
e
a
r
l
y
s
t
opp
i
ng
s
t
ra
t
e
gy
w
a
s
i
m
pl
e
m
e
nt
e
d
t
o
dyn
a
m
i
c
a
l
l
y
t
e
rm
i
na
t
e
t
r
a
i
n
i
ng
.
T
w
o
c
ond
i
t
i
ons
w
e
re
a
ppl
i
e
d
:
i)
A
c
us
t
om
c
a
l
l
ba
c
k
t
h
a
t
s
t
opp
e
d
t
r
a
i
n
i
ng
on
c
e
va
l
i
d
a
t
i
on
a
c
c
u
ra
c
y
r
e
a
c
he
d
98
%,
ii)
A
bui
l
t
-
i
n
E
a
r
l
yS
t
op
pi
ng
c
a
l
l
ba
c
k
t
ha
t
m
on
i
t
or
e
d
v
a
l
i
d
a
t
i
on
l
os
s
w
i
t
h
a
pa
t
i
e
nc
e
o
f
15
e
p
oc
hs
,
r
e
s
t
or
i
ng
t
he
be
s
t
w
e
i
gh
t
s
up
on
t
e
r
m
i
na
t
i
on
.
T
hi
s
d
ua
l
s
t
r
a
t
e
gy
a
l
l
ow
e
d
t
he
m
o
de
l
t
o
a
vo
i
d
o
ve
rf
i
t
t
i
ng
a
nd
re
d
uc
e
d
t
ra
i
ni
ng
t
i
m
e
by
ha
l
t
i
n
g
t
h
e
p
roc
e
s
s
w
he
n
op
t
i
m
a
l
pe
rf
orm
a
n
c
e
w
a
s
r
e
a
c
h
e
d.
A
m
od
e
l
c
he
c
kp
oi
nt
m
e
c
h
a
ni
s
m
w
a
s
a
l
s
o
i
n
c
l
u
de
d
t
o
re
t
a
i
n
t
he
w
e
i
gh
t
s
c
orre
s
p
ondi
n
g
t
o
t
h
e
h
i
gh
e
s
t
va
l
i
d
a
t
i
on
a
c
c
ur
a
c
y
obs
e
rve
d
du
ri
ng
t
r
a
i
ni
ng
.
T
a
b
l
e
1
.
D
i
s
t
ri
b
ut
i
on
o
f
t
ra
i
ni
n
g
a
nd
t
e
s
t
i
n
g
i
m
a
ge
s
for
di
f
fe
r
e
nt
da
t
a
s
pl
i
t
c
o
nfi
gu
ra
t
i
ons
D
a
t
a
s
p
l
i
t
T
ra
i
n
i
n
g
i
m
a
g
e
s
T
e
s
t
i
n
g
i
m
a
g
e
s
T
o
t
a
l
6
0
:
4
0
1
,
2
2
4
816
2
,
0
4
0
7
0
:
3
0
1
,
4
2
8
612
2
,
0
4
0
8
0
:
2
0
1
,
6
3
2
408
2
,
0
4
0
9
0
:
1
0
1
,
8
3
6
204
2
,
0
4
0
2.
6
.
M
od
e
l
e
val
u
at
i
on
M
ode
l
e
va
l
u
a
t
i
on
w
a
s
c
ondu
c
t
e
d
us
i
n
g
t
h
e
va
l
i
d
a
t
i
on
d
a
t
a
s
e
t
s
c
orr
e
s
pond
i
ng
t
o
e
a
c
h
t
ra
i
n
-
t
e
s
t
s
p
l
i
t
c
onfi
gura
t
i
o
n
(60:
40,
7
0:
30
,
80
:
20,
a
nd
90
:
10)
.
A
t
t
he
e
nd
o
f
e
a
c
h
t
r
a
i
n
i
ng
s
e
s
s
i
o
n,
t
he
m
od
e
l
w
a
s
e
va
l
ua
t
e
d
us
i
ng
t
he
m
od
e
l
.
e
v
a
l
u
a
t
e
()
fun
c
t
i
on,
w
h
i
c
h
re
t
u
rn
e
d
t
h
e
f
i
na
l
v
a
l
i
da
t
i
o
n
l
os
s
a
nd
a
c
c
ur
a
c
y
fo
r
t
he
b
e
s
t
-
pe
rfor
m
i
n
g
m
ode
l
c
he
c
kpo
i
nt
.
T
h
e
t
ra
i
ni
n
g
h
i
s
t
ory
,
w
h
i
c
h
i
n
c
l
ud
e
d
l
os
s
a
nd
a
c
c
ura
c
y
va
l
ue
s
for
bot
h
t
r
a
i
n
i
ng
a
nd
va
l
i
d
a
t
i
on
s
e
t
s
a
c
ros
s
a
l
l
e
poc
hs
,
w
a
s
s
t
ore
d
a
nd
vi
s
ua
l
i
z
e
d.
L
i
n
e
p
l
o
t
s
w
e
r
e
ge
ne
r
a
t
e
d,
s
how
i
ng
t
h
e
p
rogre
s
s
i
on
of
l
os
s
a
nd
a
c
c
ura
c
y
t
hr
oughou
t
t
h
e
t
ra
i
ni
n
g
proc
e
s
s
.
T
he
s
e
v
i
s
ua
l
i
z
a
t
i
ons
o
ffe
r
e
d
i
ns
i
g
ht
s
i
nt
o
t
h
e
m
o
de
l
’s
c
onv
e
rg
e
nc
e
be
h
a
vi
o
r,
s
t
a
b
i
l
i
t
y
,
a
nd
a
ny
i
ndi
c
a
t
i
ons
of
o
ve
rf
i
t
t
i
ng
o
r
unde
rfi
t
t
i
ng.
A
b
a
r
c
h
a
rt
w
a
s
a
l
s
o
c
re
a
t
e
d
t
o
s
u
m
m
a
r
i
z
e
t
h
e
fi
n
a
l
a
c
c
ura
c
y
a
n
d
l
os
s
v
a
l
u
e
s
a
c
ros
s
a
l
l
s
p
l
i
t
c
onf
i
gur
a
t
i
ons
.
T
hi
s
c
om
p
a
r
a
t
i
ve
vi
s
ua
l
i
z
a
t
i
on
c
l
a
ri
f
i
e
d
how
d
i
ff
e
re
n
t
t
r
a
i
n
i
ng
d
a
t
a
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
7
,
N
o
.
1
,
M
a
rc
h
20
26
:
20
-
29
24
proport
i
ons
a
ffe
c
t
e
d
t
h
e
m
ode
l
’s
a
bi
l
i
t
y
t
o
ge
ne
r
a
l
i
z
e
.
A
l
l
pl
ot
s
w
e
re
s
a
v
e
d
a
nd
l
a
b
e
l
e
d
a
c
c
ord
i
ng
t
o
t
he
i
r
re
s
p
e
c
t
i
ve
da
t
a
s
pl
i
t
c
o
nfi
gu
ra
t
i
ons
,
e
ns
uri
ng
re
pro
duc
i
bi
l
i
t
y
a
nd
t
ra
ns
pa
r
e
nc
y
i
n
pos
t
-
t
ra
i
ni
ng
a
na
l
ys
i
s
.
2.
7
.
M
od
e
l
s
avi
n
g
an
d
i
n
fe
r
e
n
c
e
U
pon
c
o
m
p
l
e
t
i
o
n
of
t
h
e
t
r
a
i
ni
ng
p
ro
c
e
s
s
for
e
a
c
h
d
a
t
a
s
p
l
i
t
c
on
fi
gu
ra
t
i
on
(
60
:
4
0,
70
:
30
,
8
0
:
20
,
a
nd
90:
10)
,
t
h
e
b
e
s
t
-
pe
rf
or
m
i
ng
m
od
e
l
fr
om
e
a
c
h
c
o
nf
i
g
ur
a
t
i
o
n
w
a
s
s
a
ve
d
i
n.
t
f
l
i
t
e
for
m
a
t
a
n
d
s
t
or
e
d
i
n
a
de
d
i
c
a
t
e
d
G
oog
l
e
D
ri
v
e
di
re
c
t
or
y
.
T
h
i
s
e
ns
u
re
d
c
ons
i
s
t
e
nc
y
dur
i
n
g
s
ubs
e
q
ue
nt
e
v
a
l
u
a
t
i
o
n
a
nd
e
na
bl
e
d
pot
e
n
t
i
a
l
de
pl
oy
m
e
n
t
w
i
t
hou
t
r
e
q
ui
ri
ng
re
t
r
a
i
n
i
ng
.
Inf
e
re
nc
e
w
a
s
p
e
rfo
rm
e
d
us
i
ng
11
pr
e
v
i
o
us
l
y
uns
e
e
n
t
e
s
t
i
m
a
ge
s
,
w
hi
c
h
w
e
re
no
t
i
n
c
l
ud
e
d
i
n
e
i
t
h
e
r
t
h
e
t
r
a
i
ni
ng
or
va
l
i
d
a
t
i
o
n
s
e
t
s
.
T
h
e
s
e
i
m
a
g
e
s
w
e
r
e
ob
t
a
i
n
e
d
fr
o
m
a
pub
l
i
c
l
y
a
va
i
l
a
bl
e
G
i
t
H
u
b
re
pos
i
t
ory
a
nd
u
nd
e
rw
e
n
t
t
h
e
s
a
m
e
p
re
pr
oc
e
s
s
i
ng
s
t
e
ps
a
s
t
h
e
t
r
a
i
n
i
ng
d
a
t
a
.
E
a
c
h
s
a
v
e
d
m
od
e
l
w
a
s
re
l
o
a
de
d
a
nd
a
p
pl
i
e
d
t
o
c
l
a
s
s
i
fy
t
h
e
t
e
s
t
i
m
a
g
e
s
,
a
nd
t
h
e
p
re
di
c
t
e
d
l
a
b
e
l
s
w
e
r
e
c
o
m
pa
re
d
a
g
a
i
ns
t
t
h
e
gr
oun
d
t
r
ut
h
t
o
d
e
t
e
r
m
i
ne
c
o
rr
e
c
t
n
e
s
s
.
T
o
fa
c
i
l
i
t
a
t
e
pe
rf
or
m
a
nc
e
e
v
a
l
ua
t
i
on
,
v
i
s
u
a
l
re
s
ul
t
s
w
e
r
e
ge
n
e
ra
t
e
d
b
y
di
s
pl
a
y
i
n
g
e
a
c
h
t
e
s
t
i
m
a
g
e
a
l
o
ngs
i
d
e
i
t
s
pr
e
d
i
c
t
i
on
.
Cor
r
e
c
t
c
l
a
s
s
i
fi
c
a
t
i
o
ns
w
e
re
m
a
rk
e
d
i
n
g
re
e
n,
w
he
re
a
s
i
n
c
or
re
c
t
pr
e
d
i
c
t
i
o
ns
w
e
re
m
a
rk
e
d
i
n
r
e
d
,
a
n
d
a
l
l
o
ut
put
s
w
e
r
e
a
r
c
h
i
v
e
d
f
or
d
oc
um
e
nt
a
t
i
on
pu
rp
os
e
s
.
3.
R
ES
U
LTS
A
N
D
D
I
S
C
U
S
S
I
O
N
3.
1
.
Tr
ai
n
i
n
g
an
d
v
al
i
d
a
ti
on
p
e
r
for
man
c
e
T
he
CN
N
m
od
e
l
w
a
s
e
v
a
l
u
a
t
e
d
un
de
r
f
our
t
r
a
i
n
–
va
l
i
d
a
t
i
on
s
pl
i
t
c
onf
i
gur
a
t
i
ons
:
60:
40
,
70
:
30
,
80
:
20
,
a
nd
90
:
10
.
T
he
a
c
c
ur
a
c
y
a
n
d
l
os
s
c
urv
e
s
prov
i
d
e
i
ns
i
ght
i
n
t
o
t
he
m
o
de
l
’s
l
e
a
rn
i
ng
s
t
a
b
i
l
i
t
y
a
nd
ge
n
e
r
a
l
i
z
a
t
i
o
n
be
ha
vi
or
a
c
ros
s
t
he
s
e
c
onf
i
gur
a
t
i
ons
.
A
s
s
how
n
i
n
F
i
gu
re
3,
a
l
l
a
c
c
ura
c
y
c
urv
e
s
i
nc
re
a
s
e
d
s
t
e
a
di
l
y
duri
ng
t
he
e
a
r
l
y
e
po
c
hs
be
fo
re
re
a
c
hi
ng
s
t
a
bi
l
i
z
a
t
i
on
.
T
h
e
90:
10
s
p
l
i
t
de
m
o
ns
t
ra
t
e
d
t
he
f
a
s
t
e
s
t
c
onv
e
rg
e
nc
e
,
a
c
h
i
e
v
i
ng
s
t
a
bl
e
a
c
c
ur
a
c
y
w
i
t
hi
n
four
e
po
c
hs
,
w
he
re
a
s
t
h
e
6
0:
40
s
pl
i
t
re
qui
r
e
d
27
e
po
c
hs
a
nd
e
xhi
bi
t
e
d
gre
a
t
e
r
fl
uc
t
ua
t
i
o
n,
i
nd
i
c
a
t
i
ng
s
l
ow
e
r
l
e
a
rni
ng
s
t
a
b
i
l
i
t
y
.
L
os
s
p
a
t
t
e
rns
f
ur
t
h
e
r
s
u
ppo
rt
t
h
e
s
e
obs
e
rv
a
t
i
o
ns
.
T
he
t
r
a
i
n
i
ng
a
nd
va
l
i
d
a
t
i
o
n
l
os
s
c
ur
ve
s
for
t
h
e
9
0:
10
a
nd
80
:
2
0
c
on
fi
gur
a
t
i
ons
re
m
a
i
n
e
d
c
l
os
e
l
y
a
l
i
gn
e
d
a
s
s
how
n
i
n
F
i
gur
e
3,
s
ug
ge
s
t
i
ng
e
ff
e
c
t
i
v
e
l
e
a
r
ni
ng
w
i
t
h
re
d
u
c
e
d
ov
e
r
fi
t
t
i
n
g
.
In
c
on
t
r
a
s
t
,
t
h
e
6
0:
40
a
nd
7
0:
30
s
p
l
i
t
s
d
i
s
p
l
a
ye
d
i
nc
re
a
s
i
n
g
di
ve
rg
e
n
c
e
b
e
t
w
e
e
n
t
r
a
i
ni
ng
a
nd
v
a
l
i
d
a
t
i
on
l
os
s
a
s
t
r
a
i
ni
ng
pro
gr
e
s
s
e
d
,
i
nd
i
c
a
t
i
ng
re
du
c
e
d
ge
n
e
ra
l
i
z
a
t
i
on
—
l
i
k
e
l
y
c
a
us
e
d
b
y
l
i
m
i
t
e
d
t
r
a
i
n
i
ng
s
a
m
p
l
e
d
i
v
e
rs
i
t
y
.
E
a
r
l
y
s
t
opp
i
n
g
w
a
s
e
m
p
l
o
y
e
d
t
o
t
e
rm
i
na
t
e
t
r
a
i
ni
ng
w
h
e
n
no
f
ur
t
h
e
r
i
m
p
ro
ve
m
e
n
t
i
n
va
l
i
da
t
i
on
m
e
t
ri
c
s
w
a
s
o
bs
e
rv
e
d
a
s
p
re
s
e
n
t
e
d
i
n
T
a
b
l
e
2
.
A
l
t
h
oug
h
t
h
e
t
r
a
i
n
i
ng
w
a
s
i
n
i
t
i
a
l
l
y
s
e
t
for
2
50
e
po
c
hs
,
t
h
e
m
e
c
ha
ni
s
m
ha
l
t
e
d
t
r
a
i
ni
ng
s
ubs
t
a
n
t
i
a
l
l
y
e
a
rl
i
e
r
a
c
ros
s
a
l
l
c
on
fi
gu
ra
t
i
ons
—
f
or
e
x
a
m
p
l
e
,
a
t
e
p
oc
h
2
7
(
60
:
40
)
a
nd
e
p
oc
h
12
(9
0:
10
).
M
od
e
l
w
e
i
gh
t
s
w
e
re
r
e
s
t
or
e
d
t
o
t
h
e
o
pt
i
m
a
l
c
he
c
k
po
i
n
t
for
e
a
c
h
c
o
nf
i
g
ur
a
t
i
o
n,
s
u
c
h
a
s
e
po
c
h
1
8
f
or
t
h
e
80
:
2
0
s
p
l
i
t
a
nd
e
p
o
c
h
12
for
t
he
90
:
10
s
pl
i
t
,
e
ns
u
ri
ng
e
va
l
u
a
t
i
on
b
a
s
e
d
o
n
t
h
e
b
e
s
t
-
pe
rfo
rm
i
n
g
s
t
a
t
e
ra
t
h
e
r
t
h
a
n
t
h
e
f
i
n
a
l
t
r
a
i
ni
ng
i
t
e
r
a
t
i
o
n.
T
h
e
ra
pi
d
c
o
nv
e
rg
e
nc
e
s
e
e
n
i
n
t
h
e
90
:
1
0
c
o
nf
i
gu
ra
t
i
on
c
a
n
be
a
t
t
r
i
b
ut
e
d
t
o
t
h
e
r
e
l
a
t
i
v
e
l
y
l
a
r
ge
r
t
ra
i
n
i
n
g
d
a
t
a
p
ro
por
t
i
on
,
w
hi
c
h
a
l
l
ow
e
d
t
he
m
od
e
l
t
o
l
e
a
rn
d
i
s
c
r
i
m
i
n
a
t
i
ve
fe
a
t
ur
e
r
e
p
re
s
e
nt
a
t
i
ons
e
f
fi
c
i
e
nt
l
y
.
M
e
a
nw
hi
l
e
,
c
onf
i
g
ur
a
t
i
o
ns
w
i
t
h
l
o
w
e
r
t
r
a
i
ni
ng
d
a
t
a
vo
l
u
m
e
s
re
qu
i
r
e
d
a
l
on
ge
r
op
t
i
m
i
z
a
t
i
o
n
pe
ri
od
a
nd
de
m
o
ns
t
r
a
t
e
d
g
r
e
a
t
e
r
s
e
ns
i
t
i
v
i
t
y
t
o
ov
e
r
fi
t
t
i
ng
.
F
i
gure
3
.
A
c
c
ur
a
c
y
p
l
ot
s
for
t
r
a
i
n
i
ng
a
n
d
v
a
l
i
da
t
i
o
n
a
c
ros
s
t
h
e
fo
ur
d
a
t
a
s
p
l
i
t
c
onf
i
gur
a
t
i
ons
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
D
e
v
e
l
opm
e
nt
and
p
e
r
f
or
m
an
c
e
e
v
al
u
at
i
on
of
a
c
nn
m
ode
l
f
or
s
e
agr
as
s
s
p
e
c
i
e
s
…
(
Nur
ul
H
ay
a
t
y
)
25
T
a
b
l
e
2
.
E
a
r
l
y
s
t
opp
i
ng
e
ff
i
c
i
e
n
c
y
D
a
t
a
s
p
l
i
t
E
p
o
c
h
s
t
o
p
p
e
d
Be
s
t
e
p
o
c
h
r
e
s
t
o
re
d
6
0
:
4
0
41
26
7
0
:
3
0
37
22
8
0
:
2
0
26
18
9
0
:
1
0
29
12
3.
2
.
M
od
e
l
e
val
u
at
i
on
me
t
r
i
c
s
M
ode
l
pe
rfor
m
a
n
c
e
w
a
s
qu
a
nt
i
t
a
t
i
ve
l
y
a
s
s
e
s
s
e
d
on
t
h
e
v
a
l
i
da
t
i
on
d
a
t
a
s
e
t
s
us
i
ng
t
he
m
ode
l
.
e
v
al
uat
e
()
fun
c
t
i
on,
w
hi
c
h
re
t
urn
e
d
t
h
e
fi
n
a
l
l
os
s
a
n
d
a
c
c
ura
c
y
va
l
ue
s
for
e
a
c
h
t
ra
i
n
-
v
a
l
i
da
t
i
on
s
pl
i
t
c
onfi
gura
t
i
o
n.
E
va
l
ua
t
i
o
n
w
a
s
pe
rf
orm
e
d
on
t
he
c
he
c
kpo
i
nt
t
ha
t
y
i
e
l
d
e
d
t
h
e
hi
g
he
s
t
v
a
l
i
da
t
i
on
p
e
rfo
rm
a
nc
e
,
a
s
d
e
t
e
rm
i
ne
d
by
t
he
e
a
rl
y
s
t
op
pi
ng
m
e
c
ha
n
i
s
m
di
s
c
us
s
e
d
pr
e
vi
ous
l
y.
A
s
p
re
s
e
nt
e
d
i
n
T
a
bl
e
3
,
t
he
80
:
20
a
n
d
90:
10
c
onfi
gura
t
i
o
ns
a
c
h
i
e
v
e
d
t
he
b
e
s
t
re
s
u
l
t
s
,
e
a
c
h
w
i
t
h
a
v
a
l
i
da
t
i
on
a
c
c
ura
c
y
of
98
.
04
%
a
nd
t
he
l
ow
e
s
t
l
os
s
of
0.
0
653
,
i
n
di
c
a
t
i
ng
s
t
rong
ge
ne
r
a
l
i
z
a
t
i
on
a
nd
m
i
n
i
m
a
l
ove
r
fi
t
t
i
ng
.
T
he
70:
3
0
c
onfi
g
ura
t
i
on
fol
l
ow
e
d
c
l
os
e
l
y
w
i
t
h
a
v
a
l
i
da
t
i
o
n
a
c
c
ur
a
c
y
of
95.
59%
a
nd
a
l
os
s
o
f
0.
15
74,
w
hi
l
e
t
he
6
0:
40
c
onf
i
gur
a
t
i
on
re
c
ord
e
d
t
he
l
ow
e
s
t
pe
rfor
m
a
nc
e
,
w
i
t
h
a
v
a
l
i
d
a
t
i
on
a
c
c
ur
a
c
y
of
94
.
61%
,
a
nd
a
l
os
s
of
0
.
19
90
.
In
a
dd
i
t
i
on
t
o
a
c
c
ur
a
c
y
a
nd
l
os
s
,
ot
he
c
l
a
s
s
i
fi
c
a
t
i
on
m
e
t
ri
c
s
s
uc
h
a
s
pr
e
c
i
s
i
on
,
re
c
a
l
l
,
a
n
d
F
1
-
s
c
or
e
w
e
re
a
l
s
o
c
om
put
e
d
(w
e
i
gh
t
e
d
a
v
e
ra
g
e
a
c
ros
s
a
l
l
c
l
a
s
s
e
s
)
t
o
provi
d
e
a
m
o
re
c
om
pre
h
e
ns
i
ve
e
va
l
ua
t
i
o
n
of
m
ode
l
p
e
rfo
rm
a
nc
e
.
Cons
i
s
t
e
n
l
y
,
t
he
80:
20
a
nd
90:
10
c
onf
i
gur
a
t
i
ons
a
c
h
i
ve
d
t
he
hi
g
he
s
t
pre
c
i
s
i
on
(98.
10
%
a
nd
98.
0
5%
,
re
s
p
e
c
t
i
v
e
l
y)
,
re
c
a
l
l
(
98.
0
4%
e
a
c
h),
a
nd
F
1
-
s
c
or
e
(98
.
04
%
e
a
c
h)
,
fur
t
he
r
c
onf
i
r
m
i
ng
t
he
i
r
s
up
e
ri
o
r
p
e
rfor
m
a
n
c
e
.
T
h
e
60
:
40
s
p
l
i
t
y
i
e
l
de
d
t
h
e
l
ow
e
s
t
v
a
l
u
e
s
a
c
ros
s
a
l
l
m
e
t
ri
c
s
,
w
i
t
h
a
pre
c
i
s
i
on
o
f
94.
62
%,
re
c
a
l
l
of
94
.
61%
,
a
nd
F
1
-
s
c
o
re
of
94.
58%
.
T
a
b
l
e
3
.
F
i
na
l
v
a
l
i
da
t
i
on
a
c
c
ura
c
y
a
nd
l
os
s
va
l
u
e
s
for
e
a
c
h
t
r
a
i
n
-
v
a
l
i
da
t
i
o
n
s
p
l
i
t
c
onfi
gura
t
i
o
n
D
a
t
a
s
p
l
i
t
A
c
c
u
ra
c
y
L
o
s
s
P
re
c
i
s
i
o
n
(w
e
i
g
h
t
e
d
a
v
g
)
Re
c
a
l
l
(w
e
i
g
h
t
e
d
a
v
g
)
F1
-
s
c
o
re
(w
e
i
g
h
t
e
d
a
v
g
)
6
0
:
4
0
0
.
9
4
6
1
0
.
1
9
9
0
0
.
9
4
6
2
0
.
9
4
6
1
0
.
9
4
5
8
7
0
:
3
0
0
.
9
5
5
9
0
.
1
5
4
7
0
.
9
5
6
9
0
.
9
5
5
9
0
.
9
5
5
5
8
0
:
2
0
0
.
9
8
0
4
0
.
0
6
5
3
0
.
9
8
1
0
0
.
9
8
0
4
0
.
9
8
0
4
9
0
:
1
0
0
.
9
8
0
4
0
.
0
6
5
3
0
.
9
8
0
5
0
.
9
8
0
4
0
.
9
8
0
4
T
he
s
e
r
e
s
u
l
t
s
d
e
m
o
ns
t
r
a
t
e
a
c
o
ns
i
s
t
e
n
t
t
r
e
n
d
:
i
n
c
r
e
a
s
i
ng
t
h
e
pr
o
po
rt
i
on
o
f
t
ra
i
n
i
n
g
d
a
t
a
t
e
n
ds
t
o
e
nh
a
n
c
e
m
o
de
l
p
e
r
fo
r
m
a
n
c
e
.
W
i
t
h
m
o
re
t
r
a
i
n
i
n
g
e
x
a
m
p
l
e
s
,
t
h
e
CN
N
i
s
a
b
l
e
t
o
l
e
a
rn
m
or
e
r
e
p
r
e
s
e
nt
a
t
i
ve
fe
a
t
ur
e
s
a
nd
g
e
n
e
r
a
l
i
z
e
b
e
t
t
e
r
t
o
u
ns
e
e
n
d
a
t
a
.
A
l
t
o
ug
h
a
l
l
c
o
n
fi
gu
r
a
t
i
o
ns
a
c
h
i
e
v
e
d
o
v
e
r
94
%
a
c
c
u
r
a
c
y
,
t
h
e
g
a
p
o
f
3
.
4
3%
b
e
t
w
e
e
n
t
h
e
6
0
:
4
0
a
n
d
t
h
e
t
op
-
p
e
rf
o
r
m
i
n
g
80
:
20
/
90
:
10
c
o
nf
i
gu
r
a
t
i
o
ns
u
nd
e
rs
c
or
e
s
t
h
e
i
m
p
or
t
a
n
c
e
of
t
r
a
i
n
i
n
g
d
a
t
a
vo
l
u
m
e
i
n
d
e
e
p
l
e
a
rn
i
ng
w
o
rk
f
l
o
w
s
.
N
o
t
a
b
l
y
,
t
h
e
8
0:
2
0
a
n
d
9
0
:
1
0
c
on
f
i
g
ur
a
t
i
o
ns
no
t
o
nl
y
d
e
l
i
v
e
r
e
d
t
h
e
b
e
s
t
a
c
c
ur
a
c
y
a
nd
l
o
w
e
s
t
l
o
s
s
,
b
ut
a
l
s
o
a
c
h
i
e
v
e
d
t
h
e
m
o
s
t
b
a
l
a
n
c
e
d
p
r
e
c
i
s
i
o
n
-
r
e
c
a
l
l
t
r
a
d
e
o
f
f
.
T
h
e
s
e
fi
n
di
n
gs
e
m
p
h
a
s
i
z
e
t
h
a
t
a
l
l
o
c
a
t
i
ng
a
l
a
rg
e
r
po
r
t
i
o
n
o
f
d
a
t
a
f
or
t
r
a
i
n
i
n
g
no
t
on
l
y
a
c
c
e
l
e
r
a
t
e
s
c
o
n
v
e
r
g
e
n
c
e
bu
t
a
l
s
o
i
m
p
r
ov
e
s
m
od
e
l
ro
b
us
t
n
e
s
s
a
n
d
s
t
a
b
i
l
i
t
y
,
r
e
a
f
fi
r
m
i
n
g
d
a
t
a
c
o
m
pos
i
t
i
o
n
a
s
a
c
ri
t
i
c
a
l
f
a
c
t
or
i
n
C
N
N
pe
r
fo
r
m
a
n
c
e
.
3.
3
.
I
n
fe
r
e
n
c
e
an
d
e
r
r
or
an
al
ys
i
s
Infe
r
e
nc
e
w
a
s
p
e
rfor
m
e
d
us
i
ng
33
pre
vi
ous
l
y
uns
e
e
n
i
m
a
g
e
s
of
s
e
a
gr
a
s
s
l
e
a
ve
s
,
c
om
pri
s
i
n
g
11
s
a
m
p
l
e
s
for
e
a
c
h
s
p
e
c
i
e
s
.
E
a
c
h
i
m
a
ge
w
a
s
pr
e
pro
c
e
s
s
e
d
f
ol
l
ow
i
ng
t
h
e
s
a
m
e
pro
c
e
d
ure
a
s
du
ri
ng
t
ra
i
ni
ng
,
i
nc
l
udi
n
g
re
s
i
z
i
ng
t
o
1
00×
1
00
pi
x
e
l
s
a
nd
p
i
xe
l
nor
m
a
l
i
z
a
t
i
on.
T
he
b
e
s
t
-
p
e
rfor
m
i
ng
m
o
de
l
s
f
rom
e
a
c
h
d
a
t
a
s
pl
i
t
c
o
nfi
g
ura
t
i
on
(60
:
40
,
70:
30
,
80:
20,
a
nd
9
0:
10
)
w
e
re
re
l
oa
de
d
t
o
g
e
n
e
ra
t
e
pr
e
di
c
t
i
ons
.
Cl
a
s
s
l
a
b
e
l
s
w
e
r
e
de
ri
v
e
d
us
i
ng
t
h
e
ar
gm
a
x
func
t
i
o
n
on
t
h
e
m
od
e
l
out
put
s
a
nd
c
om
p
a
r
e
d
t
o
t
he
grou
nd
t
rut
h.
A
vi
s
u
a
l
s
um
m
a
ry
of
t
h
e
s
e
pre
d
i
c
t
i
ons
i
s
s
how
n
i
n
F
i
gur
e
4
,
G
r
e
e
n
t
e
xt
i
nd
i
c
a
t
e
s
c
orr
e
c
t
pr
e
di
c
t
i
ons
,
w
hi
l
e
r
e
d
t
e
xt
m
a
rks
m
i
s
c
l
a
s
s
i
f
i
c
a
t
i
ons
.
E
a
c
h
g
roup
re
p
re
s
e
nt
s
pre
d
i
c
t
i
ons
fro
m
o
ne
t
ra
i
ni
ng
c
onf
i
gur
a
t
i
on
.
F
i
gure
4
.
In
fe
r
e
nc
e
re
s
u
l
t
s
o
n
u
ns
e
e
n
s
e
a
gr
a
s
s
l
e
a
f
i
m
a
ge
s
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
7
,
N
o
.
1
,
M
a
rc
h
20
26
:
20
-
29
26
A
s
pre
s
e
n
t
e
d
i
n
T
a
b
l
e
4
,
t
he
m
o
de
l
t
r
a
i
ne
d
us
i
ng
t
he
60:
4
0
c
onfi
gura
t
i
o
n
d
e
m
o
ns
t
r
a
t
e
d
t
he
hi
g
he
s
t
i
nfe
r
e
n
c
e
a
c
c
ur
a
c
y
,
c
orr
e
c
t
l
y
c
l
a
s
s
i
fy
i
ng
31
out
of
3
3
t
e
s
s
a
m
pl
e
s
,
r
e
s
ul
t
i
n
g
i
n
pe
r
form
a
n
c
e
of
94
%.
M
o
de
l
s
t
ra
i
ne
d
w
i
t
h
t
he
8
0:
20
a
nd
90
:
10
s
pl
i
t
s
fo
l
l
ow
e
d
c
l
os
e
l
y,
e
a
c
h
a
c
h
i
e
v
i
ng
a
n
a
c
c
u
ra
c
y
of
85
%,
w
h
e
re
a
s
t
h
e
70:
30
c
onf
i
gur
a
t
i
on
y
i
e
l
de
d
t
h
e
l
ow
e
s
t
p
e
rfor
m
a
nc
e
a
t
82
%.
T
he
s
e
out
c
o
m
e
s
d
i
ve
rge
from
t
h
e
v
a
l
i
da
t
i
on
pha
s
e
r
e
s
ul
t
s
,
e
he
r
e
t
h
e
80
:
20
a
nd
90
:
10
c
onfi
gura
t
i
o
ns
e
xhi
b
i
t
e
d
s
up
e
ri
or
a
c
c
ur
a
c
y.
T
hi
s
d
i
s
c
e
pa
n
c
y
i
ndi
c
a
t
e
s
t
h
a
t
i
nf
e
r
e
nc
e
p
e
rfo
rm
a
nc
e
m
a
y
b
e
s
ol
e
l
y
de
pe
nd
e
nt
o
n
t
he
s
i
z
e
of
t
h
e
t
ra
i
ni
ng
s
e
t
but
a
l
s
o
on
t
h
e
e
xt
e
nt
t
o
w
hi
c
h
t
h
e
t
ra
i
n
i
n
d
a
t
a
di
s
t
r
i
bu
t
i
on
c
a
p
t
ure
s
t
h
e
c
h
a
r
a
c
t
e
r
i
s
t
i
c
s
of
t
he
uns
e
e
n
t
e
s
t
i
ns
t
a
nc
e
s
.
T
a
b
l
e
4
.
In
fe
r
e
nc
e
a
c
c
ur
a
c
y
b
a
s
e
d
on
t
h
e
num
b
e
r
of
c
or
re
c
t
p
re
di
c
t
i
ons
f
or
e
a
c
h
da
t
a
s
pl
i
t
c
on
fi
gur
a
t
i
on
P
re
c
i
s
i
o
n
Re
c
a
l
l
F1
-
s
c
o
re
6
0
:
4
0
7
0
:
3
0
8
0
:
2
0
9
0
:
1
0
6
0
:
4
0
7
0
:
3
0
8
0
:
2
0
9
0
:
1
0
6
0
:
4
0
7
0
:
3
0
8
0
:
2
0
9
0
:
1
0
H
.
u
n
i
n
e
r
v
i
s
0
.
9
1
0
.
6
9
0
.
7
1
0
.
7
1
0
.
9
1
0
.
8
2
0
.
9
1
0
.
9
1
0
.
9
1
0
.
7
5
0
.
8
0
.
8
S
.
i
s
o
e
t
i
f
o
l
i
u
m
0
.
9
2
0
.
8
2
0
.
9
0
.
9
0
.
9
2
0
.
7
5
0
.
7
5
0
.
7
5
0
.
9
2
0
.
7
8
0
.
8
2
0
.
8
2
T
.
h
e
m
p
r
i
c
h
i
i
1
1
1
1
1
0
.
9
1
0
.
9
1
0
.
9
1
1
0
.
9
5
0
.
9
5
0
.
9
5
a
c
c
u
ra
c
y
0
.
9
4
0
.
8
2
0
.
8
5
0
.
8
5
m
a
c
ro
a
v
g
0
.
9
4
0
.
8
4
0
.
8
7
0
.
8
7
0
.
9
4
0
.
8
3
0
.
8
6
0
.
8
6
0
.
9
4
0
.
8
3
0
.
8
6
0
.
8
6
w
e
i
g
h
t
e
d
a
v
g
0
.
9
4
0
.
8
4
0
.
8
7
0
.
8
7
0
.
9
4
0
.
8
2
0
.
8
5
0
.
8
5
0
.
9
4
0
.
8
3
0
.
8
6
0
.
8
6
A
m
ore
d
e
t
a
i
l
e
d
p
e
r
-
c
l
a
s
s
a
na
l
ys
i
s
re
v
e
a
l
s
di
s
t
i
nc
t
c
l
a
s
s
i
f
i
c
a
t
i
on
b
e
ha
vi
or
.
T
.
h
e
m
pr
i
c
hi
c
ons
i
s
t
e
nl
y
a
c
h
i
ve
d
p
e
rf
e
c
t
pr
e
c
i
s
i
on
(1
.
00)
a
c
ros
s
a
l
l
c
o
nfi
gu
ra
t
i
on
s
a
nd
m
a
i
nt
a
i
n
e
d
a
hi
g
h
r
e
c
a
l
l
s
c
or
e
(
0.
91
),
c
ul
m
i
n
a
t
i
ng
i
n
a
n
F
1
-
s
c
or
e
of
0
.
95
.
T
he
s
pe
c
i
e
s
’
bro
a
d
a
nd
di
s
t
i
n
gui
s
h
a
bl
e
l
e
a
f
m
orph
ol
ogy
l
i
ke
l
y
fa
c
i
l
i
t
a
t
e
d
i
t
s
c
ons
i
s
t
e
n
t
c
l
a
s
s
i
fi
c
a
t
i
on
.
Conv
e
rs
e
l
y,
S
.
i
s
oe
t
i
f
ol
i
um
a
nd
H
.
u
n
i
n
e
r
v
i
s
–
w
h
i
c
h
s
h
a
re
s
i
m
i
l
a
r
e
l
ong
a
t
e
d
l
e
a
f
s
t
ruc
t
ure
s
a
nd
c
ol
o
ra
t
i
on
–
w
e
r
e
m
or
e
s
us
c
e
p
t
i
b
l
e
t
o
m
i
s
c
l
a
s
s
i
fi
c
a
t
i
on
.
N
e
v
e
rt
h
e
l
e
s
s
,
i
m
pove
m
e
nt
s
w
e
r
e
obs
e
rve
d
unde
r
s
p
e
c
i
fi
c
c
onf
i
gur
a
t
i
ons
.
In
p
a
rt
i
c
ul
a
r,
t
he
60
:
40
s
pl
i
t
e
na
b
l
e
d
S
y
r
i
n
godi
um
t
o
a
c
hi
v
e
pre
c
i
s
i
on
a
nd
re
c
a
l
l
va
l
ue
s
of
0.
9
2,
w
hi
l
e
H
al
odul
e
obt
a
i
n
e
d
i
t
s
hi
ghe
s
t
F
1
-
s
c
or
e
of
0
.
91
und
e
r
t
h
e
s
a
m
e
s
p
l
i
t
.
A
c
ros
s
a
l
l
t
r
a
i
n
–
t
e
s
t
c
onfi
g
ura
t
i
ons
,
t
h
e
c
onfus
i
on
m
a
t
r
i
c
e
s
r
e
ve
a
l
c
ons
i
s
t
e
n
t
c
l
a
s
s
i
f
i
c
a
t
i
on
t
re
n
ds
:
T
.
h
e
m
pr
i
c
hi
i
i
s
re
l
i
a
bl
y
a
nd
a
l
m
os
t
p
e
rfe
c
t
l
y
i
d
e
nt
i
fi
e
d
i
n
e
ve
r
y
s
p
l
i
t
,
i
ndi
c
a
t
i
ng
s
t
rong
m
orpho
l
og
i
c
a
l
di
s
t
i
n
c
t
i
ve
ne
s
s
a
nd
h
i
gh
m
o
de
l
s
e
pa
r
a
b
i
l
i
t
y
.
I
n
c
on
t
ra
s
t
,
m
i
s
c
l
a
s
s
i
f
i
c
a
t
i
ons
o
c
c
ur
pri
m
a
r
i
l
y
b
e
t
w
e
e
n
H
.
un
i
ne
r
v
i
s
a
nd
S
.
i
s
oe
t
i
f
ol
i
um
,
w
hos
e
vi
s
u
a
l
s
i
m
i
l
a
ri
t
i
e
s
l
e
a
d
t
o
r
e
c
i
proc
a
l
e
r
rors
i
n
a
l
l
c
on
fi
gur
a
t
i
ons
.
T
he
60:
40
s
pl
i
t
pr
ovi
d
e
s
t
h
e
m
os
t
ba
l
a
n
c
e
d
a
nd
s
t
a
b
l
e
p
e
rfor
m
a
n
c
e
,
w
hi
l
e
t
h
e
70
:
30
,
80:
2
0,
a
nd
90
:
10
s
pl
i
t
s
s
how
s
l
i
ght
i
n
c
re
a
s
e
s
i
n
c
o
nfus
i
on
be
t
w
e
e
n
t
he
s
e
t
w
o
s
p
e
c
i
e
s
,
t
hough
ove
ra
l
l
a
c
c
ura
c
y
r
e
m
a
i
ns
hi
gh
.
T
a
k
e
n
t
oge
t
he
r,
t
he
r
e
s
ul
t
s
s
how
t
ha
t
t
h
e
m
o
de
l
i
s
robus
t
a
c
r
os
s
d
a
t
a
s
e
t
propo
rt
i
ons
,
w
i
t
h
e
r
rors
dri
ve
n
no
t
by
d
a
t
a
vol
u
m
e
b
ut
by
t
h
e
i
n
t
ri
ns
i
c
s
i
m
i
l
a
ri
t
y
b
e
t
w
e
e
n
s
p
e
c
i
fi
c
s
pe
c
i
e
s
pa
i
rs
.
A
v
i
s
ua
l
c
onfus
i
o
n
m
a
t
r
i
x
o
n
t
e
s
t
d
a
t
a
i
s
s
how
n
i
n
F
i
g
ure
5
.
F
i
gure
5
.
C
onfus
i
o
n
m
a
t
ri
x
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
D
e
v
e
l
opm
e
nt
and
p
e
r
f
or
m
an
c
e
e
v
al
u
at
i
on
of
a
c
nn
m
ode
l
f
or
s
e
agr
as
s
s
p
e
c
i
e
s
…
(
Nur
ul
H
ay
a
t
y
)
27
A
c
ros
s
a
l
l
t
hr
e
e
s
p
e
c
i
e
s
,
t
h
e
60
:
40
m
od
e
l
c
ons
i
s
t
e
nl
y
re
p
ort
e
d
t
h
e
hi
g
he
s
t
m
e
t
ri
c
s
i
n
t
e
r
m
s
of
pre
c
i
s
i
on
,
r
e
c
a
l
l
,
a
nd
F
1
-
s
c
or
e
,
l
e
a
d
i
ng
t
o
t
he
be
s
t
m
a
c
r
o
a
nd
w
e
i
gh
t
e
d
a
v
e
ra
ge
s
a
t
94%
.
In
c
on
t
ra
s
t
,
t
h
e
re
m
a
i
n
i
ng
c
onfi
g
ura
t
i
ons
(7
0:
30
,
80:
20,
a
nd
90
:
10)
pro
duc
e
d
a
v
e
r
a
g
e
m
a
r
c
o
a
nd
w
e
i
gh
t
e
d
s
c
ore
s
a
round
87%.
T
h
e
s
e
fi
n
di
ngs
s
ugge
s
t
t
ha
t
w
h
i
l
e
l
a
rg
e
r
t
ra
i
ni
ng
d
a
t
a
s
e
t
s
g
e
ne
r
a
l
l
y
e
nh
a
n
c
e
va
l
i
d
a
t
i
on
a
c
c
ur
a
c
y
,
t
h
e
re
pre
s
e
nt
a
t
a
t
i
ve
n
e
s
s
of
t
he
t
r
a
i
n
i
ng
da
t
a
p
l
a
ys
a
c
r
i
t
i
c
a
l
rol
e
i
n
re
a
l
-
w
orl
d
i
n
fe
r
e
nc
e
pe
rf
orm
a
nc
e
.
3.
4
.
D
i
s
c
u
s
s
i
on
an
d
i
mp
l
i
c
at
i
on
s
T
he
C
N
N
m
o
d
e
l
d
e
m
o
ns
t
r
a
t
e
d
f
a
s
t
c
o
nv
e
rg
e
n
c
e
,
s
t
a
b
l
e
t
r
a
i
n
i
ng
b
e
h
a
v
i
or
,
a
n
d
h
i
g
h
c
l
a
s
s
i
f
i
c
a
t
i
on
a
c
c
u
r
a
c
y
a
c
r
os
s
a
l
l
c
on
f
i
g
ur
a
t
i
o
ns
.
T
h
e
i
n
t
e
gr
a
t
i
o
n
o
f
e
a
r
l
y
s
t
op
p
i
n
g
a
nd
dr
op
o
ut
r
e
g
ul
a
r
i
z
a
t
i
on
pl
a
y
e
d
c
ru
c
i
a
l
ro
l
e
s
i
n
pr
e
v
e
n
t
i
n
g
o
v
e
r
f
i
t
t
i
ng
,
a
s
e
vi
d
e
nc
e
d
b
y
t
h
e
c
l
os
e
a
l
i
g
n
m
e
nt
of
t
r
a
i
n
i
n
g
a
n
d
v
a
l
i
d
a
t
i
o
n
c
u
rv
e
s
,
e
s
p
e
c
i
a
l
l
y
i
n
8
0
:
2
0
a
nd
9
0
:
1
0
c
o
nf
i
gu
r
a
t
i
on
s
.
T
he
s
e
fi
n
di
n
gs
a
r
e
c
o
n
s
i
s
t
e
nt
w
i
t
h
b
e
s
t
p
r
a
c
t
i
c
e
s
i
n
C
N
N
-
b
a
s
e
d
i
m
a
g
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
[
18
]
–
[
20
]
.
T
h
e
s
m
o
o
t
h
p
r
og
r
e
s
s
i
o
n
o
f
t
r
a
i
n
i
ng
l
os
s
a
n
d
a
c
c
ur
a
c
y
w
i
t
ho
u
t
os
c
i
l
l
a
t
i
o
n
or
i
n
s
t
a
b
i
l
i
t
y
i
s
c
ons
i
s
t
e
n
t
w
i
t
h
w
e
l
l
-
r
e
g
ul
a
r
i
z
e
d
t
r
a
i
n
i
n
g
b
e
h
a
vi
o
r
[
21
]
.
D
e
s
p
i
t
e
l
i
m
i
t
e
d
d
a
t
a
,
t
h
e
m
od
e
l
e
ff
e
c
t
i
v
e
l
y
l
e
a
r
ne
d
di
s
c
r
i
m
i
n
a
t
i
v
e
f
e
t
ur
e
s
,
y
i
e
l
d
i
n
g
r
e
l
i
a
b
l
e
p
e
r
f
or
m
a
n
c
e
a
c
c
r
s
s
c
o
nf
i
gu
r
a
t
i
on
.
T
hi
s
f
i
n
d
i
n
g
a
l
i
g
ns
w
i
t
h
p
r
i
o
r
s
t
u
d
i
e
s
on
CN
N
a
pl
i
c
a
t
i
o
ns
i
n
m
a
ri
n
e
a
n
d
e
n
v
i
r
on
m
e
n
t
a
l
do
m
a
i
ns
[
22
]
–
[2
5
]
,
w
h
e
r
e
h
i
g
h
p
e
rf
o
r
m
a
n
c
e
w
a
s
a
c
h
i
e
v
e
d
us
i
n
g
m
od
e
r
a
t
e
-
s
i
z
e
d
a
t
a
s
e
t
s
.
I
nf
e
re
n
c
e
re
s
u
l
t
s
fu
r
t
h
e
r
c
o
n
fi
r
m
t
he
m
o
d
e
l
’s
r
ob
us
t
n
e
s
s
,
w
i
t
h
a
l
l
c
on
f
i
g
ur
a
t
i
ons
a
c
h
i
e
v
i
n
g
a
c
c
ur
a
c
y
≥8
6
.
7
%
.
T
h
e
s
e
r
e
s
u
l
t
s
a
r
e
c
ons
i
s
t
e
n
t
w
i
t
h
p
r
i
o
r
s
t
ud
i
e
s
d
e
m
on
s
t
r
a
t
i
n
g
h
i
g
h
CN
N
p
e
r
fo
r
m
a
n
c
e
i
n
m
a
r
i
n
e
c
l
a
s
s
i
f
i
c
a
t
i
on
t
a
s
ks
,
e
v
e
n
w
i
t
h
m
o
d
e
r
a
t
e
da
t
a
s
e
t
s
i
z
e
s
[2
6
]
–
[2
8]
.
M
i
s
c
l
a
s
s
i
fi
c
a
t
i
ons
,
how
e
v
e
r,
e
xp
os
e
a
n
i
m
p
ort
a
nt
l
i
m
i
t
a
t
i
on
i
n
CN
N
pe
rfor
m
a
n
c
e
for
m
or
pho
l
og
i
c
a
l
l
y
s
i
m
i
l
a
r
s
pe
c
i
e
s
.
T
h
e
c
ons
i
s
t
e
n
t
c
onfus
i
on
be
t
w
e
e
n
H
.
uni
ne
r
v
i
s
a
nd
S.
i
s
o
e
t
i
f
ol
i
um
—
a
l
s
o
obs
e
rve
d
i
n
pr
i
or
w
orks
[
9]
—
h
i
ghl
i
gh
t
s
t
he
ne
e
d
for
e
i
t
he
r
m
or
e
d
i
ve
rs
e
t
ra
i
ni
ng
s
a
m
p
l
e
s
o
r
a
dd
i
t
i
ona
l
c
ont
e
xt
(e
.
g
.
,
h
a
b
i
t
a
t
,
ba
c
kgro
und
,
m
ul
t
i
-
a
ngl
e
vi
e
w
s
)
t
o
i
m
pr
ove
s
e
pa
ra
b
i
l
i
t
y
.
O
n
t
he
ot
h
e
r
h
a
nd
,
t
h
e
c
ons
i
s
t
e
nt
l
y
c
o
rre
c
t
c
l
a
s
s
i
fi
c
a
t
i
on
of
T
.
h
e
m
pri
c
hi
i
re
i
nf
orc
e
s
t
he
m
ode
l
’s
s
t
re
ngt
h
w
h
e
n
a
p
pl
i
e
d
t
o
m
orpho
l
og
i
c
a
l
l
y
d
i
s
t
i
nc
t
c
l
a
s
s
e
s
[1
3],
[27]
,
[29]
,
[
30]
.
F
rom
a
n
a
p
pl
i
e
d
pe
rs
pe
c
t
i
ve
,
t
he
pro
pos
e
d
m
ode
l
de
m
on
s
t
ra
t
e
s
s
t
rong
pot
e
n
t
i
a
l
for
s
uppor
t
i
ng
a
ut
o
m
a
t
e
d
s
e
a
gra
s
s
m
oni
t
ori
ng
a
n
d
bi
odi
v
e
rs
i
t
y
a
s
s
e
s
s
m
e
n
t
i
n
B
i
nt
a
n
.
G
i
v
e
n
i
t
s
l
i
ght
w
e
i
g
ht
a
rc
h
i
t
e
c
t
ur
e
,
i
m
p
l
e
m
e
n
t
a
t
i
o
n
on
por
t
a
b
l
e
p
l
a
t
for
m
s
u
c
h
a
s
T
e
ns
orF
l
o
w
L
i
t
e
i
s
fe
a
s
i
b
l
e
a
nd
c
o
ul
d
s
upp
ort
i
n
-
s
i
t
u
i
de
n
t
i
f
i
c
a
t
i
on
us
i
ng
m
obi
l
e
or
e
m
b
e
dd
e
d
d
e
vi
c
e
s
.
N
on
e
t
he
l
e
s
s
,
i
m
pro
ve
m
e
n
t
s
a
r
e
n
e
c
e
s
s
a
ry
for
m
ore
c
ha
l
l
e
ngi
n
g
c
l
a
s
s
i
f
i
c
a
t
i
on
s
c
e
na
r
i
os
.
F
ut
ure
d
i
re
c
t
i
ons
m
a
y
i
n
c
l
ude
i
nt
e
gra
t
i
n
g
m
u
l
t
i
-
a
n
gl
e
or
m
u
l
t
i
-
s
c
a
l
e
i
m
a
ge
r
y
[2
6]
,
a
d
di
ng
c
ont
e
xt
u
a
l
e
nvi
ro
nm
e
nt
a
l
fe
a
t
ure
s
s
uc
h
a
s
d
e
pt
h
or
l
o
c
a
t
i
o
n
m
e
t
a
d
a
t
a
[
31]
,
a
nd
l
e
v
e
ra
g
i
ng
e
ns
e
m
b
l
e
or
s
e
m
i
-
s
upe
r
vi
s
e
d
a
ppro
a
c
h
e
s
t
o
e
nha
nc
e
ge
n
e
r
a
l
i
z
a
t
i
o
n
on
re
a
l
-
w
o
rl
d
und
e
rw
a
t
e
r
da
t
a
[9]
.
T
a
k
e
n
t
oge
t
he
r
,
t
he
f
i
ndi
ngs
pre
s
e
nt
e
d
i
n
t
h
i
s
s
t
udy
de
m
o
ns
t
ra
t
e
t
h
e
m
od
e
l
'
s
pra
c
t
i
c
a
l
v
i
a
bi
l
i
t
y
a
nd
hi
ghl
i
gh
t
t
he
i
m
por
t
a
nc
e
of
d
a
t
a
qua
l
i
t
y,
s
pe
c
i
e
s
d
i
s
t
i
nc
t
i
v
e
ne
s
s
,
a
nd
c
on
t
e
x
t
u
a
l
a
ug
m
e
nt
a
t
i
on
a
s
ke
y
c
ons
i
d
e
ra
t
i
o
ns
i
n
a
dva
nc
i
ng
s
e
a
g
ra
s
s
c
l
a
s
s
i
fi
c
a
t
i
on
e
ffor
t
s
us
i
ng
d
e
e
p
l
e
a
r
ni
ng
.
4.
C
O
N
C
LU
S
I
O
N
T
hi
s
s
t
udy
p
re
s
e
nt
e
d
t
h
e
d
e
v
e
l
op
m
e
nt
a
nd
pe
r
form
a
n
c
e
e
va
l
ua
t
i
on
o
f
a
C
N
N
m
od
e
l
fo
r
t
h
e
c
l
a
s
s
i
fi
c
a
t
i
on
of
t
hre
e
s
e
a
g
ra
s
s
s
p
e
c
i
e
s
—
H
.
uni
ne
r
v
i
s
,
S
.
i
s
oe
t
i
f
ol
i
um
,
a
nd
T
.
he
m
pr
i
c
h
i
i
—
i
n
Bi
nt
a
n,
Indone
s
i
a
.
T
h
e
m
od
e
l
w
a
s
t
r
a
i
ne
d
us
i
ng
va
ry
i
ng
t
r
a
i
n
-
va
l
i
da
t
i
on
s
pl
i
t
s
(60:
4
0,
70
:
30
,
80:
20
,
a
n
d
90:
10
),
de
m
o
ns
t
ra
t
i
n
g
rob
us
t
c
onv
e
rg
e
nc
e
a
nd
h
i
gh
c
l
a
s
s
fi
f
i
c
a
t
i
on
a
c
c
ur
a
c
y
a
c
ros
s
a
l
l
c
on
fi
gu
ra
t
i
ons
.
T
he
90:
1
0
c
onfi
gura
t
i
o
n
y
i
e
l
de
d
t
he
h
i
gh
e
s
t
va
l
i
d
a
t
i
on
a
c
c
ur
a
c
y
(9
8.
53%
)
a
nd
l
ow
e
s
t
va
l
i
d
a
t
i
on
l
os
s
(0.
08881)
,
i
ndi
c
a
t
i
ng
s
t
ro
ng
m
od
e
l
pe
r
form
a
n
c
e
duri
ng
t
ra
i
ni
ng.
Inf
e
re
n
c
e
on
uns
e
e
n
t
e
s
t
i
m
a
g
e
s
c
onf
i
rm
e
d
t
h
e
m
od
e
l
’s
ge
ne
r
a
l
i
z
a
t
i
on
c
a
pa
b
i
l
i
t
y
,
a
c
h
i
e
v
i
ng
up
t
o
8
6.
7
%.
a
l
t
hou
gh
t
he
80
:
20
a
nd
90:
1
0
s
pl
i
t
s
e
x
c
e
l
l
e
d
dur
i
ng
t
h
e
va
l
i
da
t
i
on
p
ha
s
e
,
t
he
60
:
40
s
pl
i
t
pr
oduc
e
d
t
he
m
os
t
s
t
a
b
l
e
a
nd
re
l
i
a
bl
e
pe
r
form
a
n
c
e
dur
i
ng
i
nfe
r
e
n
c
e
,
s
ugge
s
t
i
ng
b
e
t
t
e
r
ge
n
e
ra
l
i
z
a
t
i
o
n
t
o
re
a
l
-
w
o
rl
d
da
t
a
.
T
hi
s
h
i
gh
l
i
gh
t
s
t
h
e
i
m
po
rt
a
nc
e
of
not
onl
y
da
t
a
s
e
t
s
i
z
e
but
a
l
s
o
t
h
e
r
e
pr
e
s
e
n
t
a
t
i
on
a
l
qu
a
l
i
t
y
a
nd
di
s
t
r
i
bu
t
i
on
of
t
ra
i
ni
ng
s
a
m
pl
e
s
.
H
ow
e
v
e
r
,
r
e
c
urri
n
g
m
i
s
c
l
a
s
s
i
fi
c
a
t
i
ons
be
t
w
e
e
n
H.
un
i
n
e
r
v
i
s
a
n
d
S.
i
s
o
e
t
i
f
o
l
i
um
–
s
pe
c
i
e
s
w
i
t
h
s
u
bl
t
e
m
orp
hol
o
gi
c
a
l
d
i
ff
e
re
n
c
e
s
–
poi
nt
s
t
o
t
h
e
l
i
m
i
t
a
t
i
ons
of
s
i
ngl
e
-
v
i
e
w
i
m
a
g
e
i
npu
t
.
In
c
ont
r
a
s
t
,
T
.
h
e
m
pr
i
c
hi
i
w
a
s
c
ons
i
s
t
e
n
t
l
y
c
l
a
s
s
i
f
i
e
d
w
i
t
h
hi
g
h
pre
c
i
s
i
on
,
a
ffi
r
m
i
ng
t
he
m
od
e
l
’s
e
ff
i
c
a
c
y
i
n
i
d
e
nt
i
fy
i
ng
m
orpho
l
og
i
c
a
l
l
y
d
i
s
t
i
n
c
t
s
pe
c
i
e
s
.
O
v
e
ra
l
l
,
t
he
fi
ndi
n
gs
s
u
pport
t
h
e
s
ui
t
a
bi
l
i
t
y
of
CN
N
-
ba
s
e
d
a
pp
roa
c
he
s
for
a
u
t
om
a
t
e
d
s
e
a
gr
a
s
s
c
l
a
s
s
i
fi
c
a
t
i
on
i
n
m
a
r
i
ne
m
oni
t
or
i
ng
a
ppl
i
c
a
t
i
ons
.
F
u
t
ure
i
m
prov
e
m
e
nt
s
m
a
y
foc
us
o
n
i
nc
or
pora
t
i
n
g
ri
c
he
r
i
npu
t
m
oda
l
i
t
i
e
s
,
i
nc
r
e
a
s
i
ng
da
t
a
s
e
t
di
v
e
rs
i
t
y,
a
nd
e
x
pl
or
i
ng
m
od
e
l
e
ns
e
m
b
l
e
s
t
o
fur
t
h
e
r
i
m
pro
ve
a
c
c
ur
a
c
y
a
nd
re
l
i
a
bi
l
i
t
y
i
n
c
om
p
l
e
x
unde
rw
a
t
e
r
e
nvi
ron
m
e
n
t
s
.
A
C
K
N
O
WL
ED
G
M
EN
TS
T
he
a
u
t
hor
gra
t
e
fu
l
l
y
a
c
know
l
e
dg
e
s
F
a
r
i
d
a
for
a
s
s
i
s
t
a
nc
e
du
ri
ng
f
i
e
l
dw
ork
i
n
Bi
n
t
a
n
,
p
a
rt
i
c
u
l
a
r
l
y
i
n
s
e
a
gr
a
s
s
s
a
m
pl
e
c
ol
l
e
c
t
i
on
a
nd
org
a
ni
z
a
t
i
on
.
A
p
pre
c
i
a
t
i
on
i
s
a
l
s
o
e
x
t
e
nde
d
t
o
l
oc
a
l
a
u
t
hori
t
i
e
s
a
nd
c
om
m
un
i
t
y
m
e
m
be
rs
for
t
h
e
i
r
c
oope
ra
t
i
on
a
nd
a
c
c
e
s
s
t
o
re
s
e
a
rc
h
s
i
t
e
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
7
,
N
o
.
1
,
M
a
rc
h
20
26
:
20
-
29
28
F
U
N
D
I
N
G
I
N
F
O
R
M
A
TI
O
N
T
he
a
ut
hor
d
e
c
l
a
r
e
s
t
h
a
t
t
hi
s
r
e
s
e
a
r
c
h
w
a
s
c
ond
uc
t
e
d
i
nd
e
p
e
nde
nt
l
y
a
nd
re
c
e
i
ve
d
no
s
pe
c
i
f
i
c
g
ra
n
t
from
a
ny
fu
ndi
ng
a
ge
n
c
y
i
n
t
h
e
publ
i
c
,
c
o
m
m
e
r
c
i
a
l
,
or
no
t
-
f
or
-
profi
t
s
e
c
t
ors
.
A
U
TH
O
R
C
O
N
TR
I
BU
TI
O
N
S
S
TA
T
EM
EN
T
T
hi
s
j
ou
rna
l
us
e
s
t
h
e
Cont
r
i
bu
t
or
Rol
e
s
T
a
xono
m
y
(C
Re
di
T
)
t
o
r
e
c
o
gni
z
e
i
ndi
v
i
du
a
l
a
u
t
hor
c
ont
r
i
bu
t
i
ons
,
r
e
du
c
e
a
ut
hors
h
i
p
di
s
pu
t
e
s
,
a
nd
fa
c
i
l
i
t
a
t
e
c
o
l
l
a
bora
t
i
on
.
N
ame
of
A
u
th
o
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
N
urul
H
a
ya
t
y
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
H
ol
l
a
nda
A
ri
e
f
K
us
um
a
✓
✓
✓
✓
✓
✓
✓
✓
✓
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
ft
w
a
re
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
rm
a
l
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
s
o
u
rc
e
s
D
:
D
a
t
a
Cu
ra
t
i
o
n
O
:
W
ri
t
i
n
g
-
O
ri
g
i
n
a
l
D
ra
ft
E
:
W
ri
t
i
n
g
-
Re
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
s
u
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
rv
i
s
i
o
n
P
:
P
ro
j
e
c
t
a
d
m
i
n
i
s
t
ra
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
s
i
t
i
o
n
C
O
N
F
LI
C
T
O
F
I
N
T
ER
ES
T
S
TA
T
EM
EN
T
T
he
a
u
t
hors
s
t
a
t
e
no
c
onfl
i
c
t
o
f
i
nt
e
re
s
t
.
D
A
TA
A
V
A
I
LA
BI
LI
TY
T
he
c
od
e
,
t
ra
i
ni
ng
d
a
t
a
,
t
e
s
t
da
t
a
,
t
r
a
i
n
i
ng
a
nd
v
a
l
i
da
t
i
o
n
s
a
m
pl
e
ge
ne
r
a
t
i
ons
,
pre
d
i
c
t
i
o
n
re
s
u
l
t
s
,
a
nd
fi
gur
e
s
fo
r
l
os
s
a
nd
a
c
c
ura
c
y
us
e
d
i
n
t
h
i
s
s
t
udy
a
re
publ
i
c
l
y
a
va
i
l
a
bl
e
i
n
t
he
G
i
t
H
ub
r
e
pos
i
t
ory:
ht
t
ps
:
/
/
g
i
t
h
ub.
c
o
m
/
ho
l
l
a
nd
a
kus
u
m
a
/
L
a
m
unM
a
c
hi
n
e
L
e
a
r
ni
ng
CN
N
.
P
y
t
hon
s
c
r
i
pt
s
a
nd
J
up
yt
e
r
not
e
books
(.
i
py
nb)
a
r
e
a
l
s
o
prov
i
de
d
i
n
t
h
e
r
e
pos
i
t
ory
t
o
f
a
c
i
l
i
t
a
t
e
re
pro
duc
i
b
i
l
i
t
y
a
nd
f
urt
h
e
r
a
n
a
l
ys
i
s
.
R
EF
ER
EN
C
ES
[1
]
L
.
C
.
C.
-
U
n
s
w
o
rt
h
a
n
d
R.
U
n
s
w
o
rt
h
,
“
A
c
a
l
l
fo
r
s
e
a
g
ra
s
s
p
ro
t
e
c
t
i
o
n
,
”
S
c
i
e
n
c
e
,
v
o
l
.
3
6
1
,
n
o
.
6
4
0
1
,
p
p
.
4
4
6
–
4
4
8
,
2
0
1
8
.
[2
]
M
.
G
ro
s
s
,
“
S
a
v
e
o
u
r
s
e
a
g
ra
s
s
e
s
,
”
Cu
r
r
e
n
t
B
i
o
l
o
g
y
,
v
o
l
.
3
0
,
n
o
.
1
6
,
p
p
.
R9
0
5
–
R9
0
7
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
u
b
.
2
0
2
0
.
0
7
.
0
8
1
.
[3
]
C.
J
.
M
i
l
l
e
r,
S
.
J
.
Ca
m
p
b
e
l
l
,
a
n
d
S
.
S
c
u
d
d
s
,
“
S
p
a
t
i
a
l
v
a
ri
a
t
i
o
n
o
f
Z
o
s
t
e
ra
t
a
s
m
a
n
i
c
a
m
o
rp
h
o
l
o
g
y
a
n
d
s
t
ru
c
t
u
re
a
c
r
o
s
s
a
n
e
n
v
i
ro
n
m
e
n
t
a
l
g
ra
d
i
e
n
t
,
”
M
a
r
i
n
e
E
c
o
l
o
g
y
P
r
o
g
r
e
s
s
S
e
r
i
e
s
,
v
o
l
.
3
0
4
,
p
p
.
4
5
–
5
3
,
2
0
0
5
,
d
o
i
:
1
0
.
3
3
5
4
/
m
e
p
s
3
0
4
0
4
5
.
[4
]
N
.
T
.
H
a
,
M
.
M
.
-
H
a
rri
s
,
T
.
D
.
P
h
a
m
,
a
n
d
I.
H
a
w
e
s
,
“
A
c
o
m
p
a
ra
t
i
v
e
a
s
s
e
s
s
m
e
n
t
o
f
e
n
s
e
m
b
l
e
-
b
a
s
e
d
m
a
c
h
i
n
e
l
e
a
rn
i
n
g
a
n
d
m
a
x
i
m
u
m
l
i
k
e
l
i
h
o
o
d
m
e
t
h
o
d
s
fo
r
m
a
p
p
i
n
g
s
e
a
g
r
a
s
s
u
s
i
n
g
s
e
n
t
i
n
e
l
-
2
i
m
a
g
e
ry
i
n
T
a
u
ra
n
g
a
H
a
rb
o
r,
N
e
w
Z
e
a
l
a
n
d
,
”
R
e
m
o
t
e
S
e
n
s
i
n
g
,
v
o
l
.
1
2
,
n
o
.
3
,
2
0
2
0
,
d
o
i
:
1
0
.
3
3
9
0
/
rs
1
2
0
3
0
3
5
5
.
[5
]
S
.
D
a
h
i
y
a
,
R.
T
y
a
g
i
,
a
n
d
N
.
G
a
b
a
,
“
Co
m
p
a
ri
s
o
n
o
f
M
L
c
l
a
s
s
i
f
i
e
rs
fo
r
i
m
a
g
e
d
a
t
a
,
”
E
a
s
y
Ch
a
i
r
P
r
e
p
r
i
n
t
No
3
8
1
5
,
2
0
2
0
,
[O
n
l
i
n
e
].
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
e
a
s
y
c
h
a
i
r.
o
rg
/
p
u
b
l
i
c
a
t
i
o
n
s
/
p
re
p
ri
n
t
/
K
n
C4
.
[6
]
M
.
M
o
n
i
ru
z
z
a
m
a
n
,
S
.
M
.
S
.
Is
l
a
m
,
M
.
Be
n
n
a
m
o
u
n
,
a
n
d
P
.
L
a
v
e
ry
,
“
D
e
e
p
l
e
a
rn
i
n
g
o
n
u
n
d
e
rw
a
t
e
r
m
a
ri
n
e
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
:
a
s
u
rv
e
y
,
”
i
n
A
d
v
a
n
c
e
d
Co
n
c
e
p
t
s
f
o
r
In
t
e
l
l
i
g
e
n
t
V
i
s
i
o
n
S
y
s
t
e
m
s
(
A
CIV
S
2
0
1
7
)
,
Ch
a
m
,
S
w
i
t
z
e
rl
a
n
d
:
S
p
ri
n
g
e
r,
2
0
1
7
,
p
p
.
1
5
0
–
1
6
0
,
doi
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
319
-
7
0
3
5
3
-
4
_
1
3
.
[7
]
M
.
S
.
H
o
s
s
a
i
n
,
J
.
S
.
Bu
j
a
n
g
,
M
.
H
.
Z
a
k
a
ri
a
,
a
n
d
M
.
H
a
s
h
i
m
,
“
T
h
e
a
p
p
l
i
c
a
t
i
o
n
o
f
re
m
o
t
e
s
e
n
s
i
n
g
t
o
s
e
a
g
ra
s
s
e
c
o
s
y
s
t
e
m
s
:
a
n
o
v
e
rv
i
e
w
a
n
d
fu
t
u
re
re
s
e
a
rc
h
p
ro
s
p
e
c
t
s
,
”
In
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
R
e
m
o
t
e
S
e
n
s
i
n
g
,
v
o
l
.
3
6
,
n
o
.
1
,
p
p
.
6
1
–
1
1
4
,
2
0
1
5
,
d
o
i
:
1
0
.
1
0
8
0
/
0
1
4
3
1
1
6
1
.
2
0
1
4
.
9
9
0
6
4
9
.
[8
]
M
.
U
.
G
u
m
u
s
a
y
,
T
.
Ba
k
i
rm
a
n
,
I.
T
.
K
i
z
i
l
k
a
y
a
,
a
n
d
N
.
O
.
A
y
k
u
t
,
“
A
re
v
i
e
w
o
f
s
e
a
g
ra
s
s
d
e
t
e
c
t
i
o
n
,
m
a
p
p
i
n
g
a
n
d
m
o
n
i
t
o
ri
n
g
a
p
p
l
i
c
a
t
i
o
n
s
u
s
i
n
g
a
c
o
u
s
t
i
c
s
y
s
t
e
m
s
,
”
E
u
r
o
p
e
a
n
J
o
u
r
n
a
l
o
f
R
e
m
o
t
e
S
e
n
s
i
n
g
,
v
o
l
.
5
2
,
n
o
.
1
,
p
p
.
1
–
2
9
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
8
0
/
2
2
7
9
7
2
5
4
.
2
0
1
8
.
1
5
4
4
8
3
8
.
[9
]
Y
.
S
.
D
e
s
h
m
u
k
h
,
R.
G
.
T
a
m
b
e
,
R.
D
.
Ch
i
n
t
a
m
a
n
i
,
S
.
S
.
Bh
o
s
a
l
e
,
a
n
d
S
.
M
u
t
h
u
ra
j
,
“
S
e
a
g
ra
s
s
d
e
t
e
c
t
i
o
n
a
n
d
c
l
a
s
s
i
fi
c
a
t
i
o
n
fro
m
u
d
e
rw
a
t
e
r
i
m
a
g
e
s
u
s
i
n
g
d
e
e
p
t
ra
n
s
fe
r
l
e
a
r
n
i
n
g
,
”
A
d
v
a
n
c
e
s
i
n
No
n
l
i
n
e
a
r
V
a
r
i
a
t
i
o
n
a
l
In
e
q
u
a
l
i
t
i
e
s
,
v
o
l
.
2
8
,
n
o
.
1
S
,
p
p
.
1
2
9
–
1
3
6
,
2
0
2
5
,
d
o
i
:
1
0
.
5
2
7
8
3
/
a
n
v
i
.
v
2
8
.
2
2
3
3
.
[1
0
]
H
.
M
o
h
a
m
e
d
,
K
.
N
a
d
a
o
k
a
,
a
n
d
T
.
N
a
k
a
m
u
ra
,
“
S
e
m
i
a
u
t
o
m
a
t
e
d
m
a
p
p
i
n
g
o
f
b
e
n
t
h
i
c
h
a
b
i
t
a
t
s
a
n
d
s
e
a
g
r
a
s
s
s
p
e
c
i
e
s
u
s
i
n
g
a
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
ra
l
n
e
t
w
o
rk
fra
m
e
w
o
r
k
i
n
s
h
a
l
l
o
w
w
a
t
e
r
e
n
v
i
ro
n
m
e
n
t
s
,
”
R
e
m
o
t
e
S
e
n
s
i
n
g
,
v
o
l
.
1
2
,
n
o
.
2
3
,
p
p
.
1
–
1
8
,
2
0
2
0
,
d
o
i
:
1
0
.
3
3
9
0
/
rs
1
2
2
3
4
0
0
2
.
[1
1
]
D
.
P
e
re
z
e
t
a
l
.
,
“
Q
u
a
n
t
i
fy
i
n
g
s
e
a
g
ra
s
s
d
i
s
t
ri
b
u
t
i
o
n
i
n
c
o
a
s
t
a
l
w
a
t
e
r
w
i
t
h
d
e
e
p
l
e
a
rn
i
n
g
m
o
d
e
l
s
,
”
R
e
m
o
t
e
S
e
n
s
i
n
g
,
v
o
l
.
1
2
,
n
o
.
1
0
,
2
0
2
0
,
d
o
i
:
1
0
.
3
3
9
0
/
rs
1
2
1
0
1
5
8
1
.
[1
2
]
M
.
K
.
N
o
m
a
n
,
S
.
M
.
J
.
J
a
l
a
l
i
,
S
.
M
.
S
.
Is
l
a
m
,
a
n
d
P
.
L
a
v
e
ry
,
“
O
fd
a
-
C
NN
:
a
n
o
v
e
l
m
e
t
a
h
e
u
ri
s
t
i
c
a
l
g
o
ri
t
h
m
-
b
a
s
e
d
d
e
e
p
C
NN
fo
r
m
u
l
t
i
-
s
p
e
c
i
e
s
s
e
a
g
ra
s
s
c
l
a
s
s
i
fi
c
a
t
i
o
n
,
”
S
S
R
N
,
2
0
2
3
,
d
o
i
:
1
0
.
2
1
3
9
/
s
s
rn
.
4
3
4
8
7
9
3
.
[1
3
]
M
.
A
.
A
.
O
z
a
e
t
a
,
A
.
C.
F
a
j
a
rd
o
,
F
.
P
.
Bra
z
a
s
,
a
n
d
J
.
A
.
M
.
Ca
n
t
a
l
,
“
D
e
e
p
l
e
a
rn
i
n
g
a
p
p
ro
a
c
h
fo
r
s
e
a
g
ra
s
s
s
p
e
c
i
e
s
c
l
a
s
s
i
fi
c
a
t
i
o
n
,
”
2
0
2
4
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
o
n
G
r
e
e
n
E
n
e
r
g
y
,
Co
m
p
u
t
i
n
g
a
n
d
S
u
s
t
a
i
n
a
b
l
e
T
e
c
h
n
o
l
o
g
y
(G
E
CO
S
T
)
,
p
p
.
2
5
0
–
2
5
4
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
G
E
CO
S
T
6
0
9
0
2
.
2
0
2
4
.
1
0
4
7
4
9
8
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
D
e
v
e
l
opm
e
nt
and
p
e
r
f
or
m
an
c
e
e
v
al
u
at
i
on
of
a
c
nn
m
ode
l
f
or
s
e
agr
as
s
s
p
e
c
i
e
s
…
(
Nur
ul
H
ay
a
t
y
)
29
[1
4
]
G
.
Re
u
s
e
t
a
l
.
,
“
L
o
o
k
i
n
g
fo
r
s
e
a
g
ra
s
s
:
d
e
e
p
l
e
a
rn
i
n
g
fo
r
v
i
s
u
a
l
c
o
v
e
ra
g
e
e
s
t
i
m
a
t
i
o
n
,
”
2
0
1
8
O
CE
A
NS
-
M
T
S
/
IE
E
E
Ko
b
e
T
e
c
h
n
o
-
O
c
e
a
n
s
(O
T
O
),
Ko
b
e
,
J
a
p
a
n
,
p
p
.
1
–
6
,
2
0
1
8
,
d
o
i
:
1
0
.
1
1
0
9
/
O
CE
A
N
S
K
O
B
E
.
2
0
1
8
.
8
5
5
9
3
0
2
.
[1
5
]
D
.
P
.
Ri
n
i
a
n
d
W
.
K
u
rn
i
a
S
a
ri
,
“
O
p
t
i
m
i
z
i
n
g
h
y
p
e
rp
a
ra
m
e
t
e
rs
o
f
CN
N
a
n
d
D
N
N
fo
r
e
m
o
t
i
o
n
c
l
a
s
s
i
fi
c
a
t
i
o
n
b
a
s
e
d
o
n
E
E
G
s
i
g
n
a
l
s
,
”
In
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
n
In
f
o
r
m
a
t
i
o
n
a
n
d
Co
m
m
u
n
i
c
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
(IJ
o
ICT
)
,
v
o
l
.
1
0
,
n
o
.
1
,
p
p
.
1
–
1
2
,
2
0
2
4
,
d
o
i
:
1
0
.
2
1
1
0
8
/
i
j
o
i
c
t
.
v
1
0
i
1
.
8
5
7
.
[1
6
]
Y
.
L
e
c
u
n
,
Y
.
Be
n
g
i
o
,
a
n
d
G
.
H
i
n
t
o
n
,
“
D
e
e
p
l
e
a
rn
i
n
g
,
”
Na
t
u
r
e
,
v
o
l
.
5
2
1
,
n
o
.
7
5
5
3
,
p
p
.
4
3
6
–
4
4
4
,
2
0
1
5
,
d
o
i
:
1
0
.
1
0
3
8
/
n
a
t
u
re
1
4
5
3
9
.
[1
7
]
M
.
A
b
a
d
i
e
t
a
l
.
,
“
T
e
n
s
o
rF
l
o
w
:
a
s
y
s
t
e
m
fo
r
l
a
rg
e
-
s
c
a
l
e
m
a
c
h
i
n
e
l
e
a
rn
i
n
g
,
”
O
S
D
I’1
6
:
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
1
2
t
h
U
S
E
NI
X
c
o
n
f
e
r
e
n
c
e
on
O
p
e
r
a
t
i
n
g
S
y
s
t
e
m
s
D
e
s
i
g
n
a
n
d
Im
p
l
e
m
e
n
t
a
t
i
o
n
,
p
p
.
2
6
5
–
2
8
3
,
2
0
1
6
.
[1
8
]
L
.
P
re
c
h
e
l
t
,
“
E
a
rl
y
s
t
o
p
p
i
n
g
—
b
u
t
w
h
e
n
?
,
”
i
n
Ne
u
r
a
l
Ne
t
wo
r
k
s
:
T
r
i
c
k
s
o
f
t
h
e
T
r
a
d
e
,
Be
rl
i
n
,
G
e
rm
a
n
y
:
S
p
ri
n
g
e
r,
p
p
.
5
3
–
67
,
2012,
doi
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
6
4
2
-
3
5
2
8
9
-
8
_
5
.
[1
9
]
R.
Ca
ru
a
n
a
,
S
.
L
a
w
re
n
c
e
,
a
n
d
L
.
G
i
l
e
s
,
“
O
v
e
rfi
t
t
i
n
g
i
n
n
e
u
ra
l
n
e
t
s
:
b
a
c
k
p
ro
p
a
g
a
t
i
o
n
,
c
o
n
j
u
g
a
t
e
g
ra
d
i
e
n
t
,
a
n
d
e
a
rl
y
s
t
o
p
p
i
n
g
,
”
A
d
v
a
n
c
e
s
i
n
Ne
u
r
a
l
In
f
o
r
m
a
t
i
o
n
P
r
o
c
e
s
s
i
n
g
S
y
s
t
e
m
s
,
2
0
0
1
.
[2
0
]
G
.
H
i
n
t
o
n
,
A
.
K
r
i
z
h
e
v
s
k
y
,
I.
S
u
t
s
k
e
v
e
r,
R.
S
a
l
a
k
h
u
t
d
i
n
o
v
,
a
n
d
N
.
S
ri
v
a
s
t
a
v
a
,
“
D
ro
p
o
u
t
:
a
s
i
m
p
l
e
w
a
y
t
o
p
re
v
e
n
t
n
e
u
ra
l
n
e
t
w
o
rk
s
fro
m
o
v
e
rfi
t
t
i
n
g
,
”
J
o
u
r
n
a
l
o
f
M
a
c
h
i
n
e
L
e
a
r
n
i
n
g
R
e
s
e
a
r
c
h
,
v
o
l
.
1
5
,
n
o
.
1
,
p
p
.
1
9
2
9
–
1
9
5
8
,
2
0
1
4
.
[2
1
]
I.
G
o
o
d
fe
l
l
o
w
,
Y
.
Be
n
g
i
o
,
a
n
d
A
.
Co
u
rv
i
l
l
e
,
D
e
e
p
l
e
a
r
n
i
n
g
.
Ca
m
b
ri
d
g
e
,
M
a
s
s
a
c
h
u
s
e
t
t
s
:
T
h
e
M
IT
P
re
s
s
,
2
0
1
6
.
[2
2
]
Z
.
Ca
o
,
J
.
C.
P
r
i
n
c
i
p
e
,
B.
O
u
y
a
n
g
,
F
.
D
a
l
g
l
e
i
s
h
,
a
n
d
A
.
V
u
o
re
n
k
o
s
k
i
,
“
M
a
ri
n
e
a
n
i
m
a
l
c
l
a
s
s
i
fi
c
a
t
i
o
n
u
s
i
n
g
c
o
m
b
i
n
e
d
CN
N
a
n
d
h
a
n
d
-
d
e
s
i
g
n
e
d
i
m
a
g
e
fe
a
t
u
re
s
,
”
O
CE
A
NS
2
0
1
5
-
M
T
S
/
IE
E
E
W
a
s
h
i
n
g
t
o
n
,
p
p
.
1
–
6
,
2
0
1
6
,
d
o
i
:
1
0
.
2
3
9
1
9
/
o
c
e
a
n
s
.
2
0
1
5
.
7
4
0
4
3
7
5
.
[2
3
]
E
.
L
.
W
h
i
t
e
,
H
.
K
l
i
n
c
k
,
J
.
M
.
Bu
l
l
,
P
.
R.
W
h
i
t
e
,
a
n
d
D
.
R
i
s
c
h
,
“
O
n
e
s
i
z
e
fi
t
s
a
l
l
?
a
d
a
p
t
a
t
i
o
n
o
f
t
ra
i
n
e
d
CN
N
s
t
o
n
e
w
m
a
ri
n
e
a
c
o
u
s
t
i
c
e
n
v
i
ro
n
m
e
n
t
s
,
”
E
c
o
l
o
g
i
c
a
l
In
f
o
r
m
a
t
i
c
s
,
v
o
l
.
7
8
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
c
o
i
n
f.
2
0
2
3
.
1
0
2
3
6
3
.
[2
4
]
J
.
F
re
e
m
a
n
,
“
Co
n
t
e
n
t
s
e
a
rc
h
w
i
t
h
i
n
l
a
rg
e
e
n
v
i
ro
n
m
e
n
t
a
l
d
a
t
a
s
e
t
s
u
s
i
n
g
a
c
o
n
v
o
l
u
t
i
o
n
n
e
u
ra
l
n
e
t
w
o
rk
,
”
Co
m
p
u
t
e
r
s
a
n
d
G
e
o
s
c
i
e
n
c
e
s
,
v
o
l
.
1
3
9
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
a
g
e
o
.
2
0
2
0
.
1
0
4
4
7
9
.
[2
5
]
E
.
C.
O
re
n
s
t
e
i
n
a
n
d
O
.
Be
i
j
b
o
m
,
“
T
ra
n
s
fe
r
l
e
a
rn
i
n
g
a
n
d
d
e
e
p
fe
a
t
u
re
e
x
t
ra
c
t
i
o
n
fo
r
p
l
a
n
k
t
o
n
i
c
i
m
a
g
e
d
a
t
a
s
e
t
s
,
”
2
0
1
7
IE
E
E
W
i
n
t
e
r
Co
n
f
e
r
e
n
c
e
o
n
A
p
p
l
i
c
a
t
i
o
n
s
o
f
Co
m
p
u
t
e
r
V
i
s
i
o
n
(W
A
CV
)
,
v
o
l
.
6
,
n
o
.
4
,
p
p
.
1
0
8
2
–
1
0
8
8
,
2
0
1
7
,
d
o
i
:
1
0
.
1
1
0
9
/
W
A
CV
.
2
0
1
7
.
1
2
5
.
[2
6
]
J
.
E
l
s
ä
ß
e
r,
L
.
W
e
i
h
l
,
V
.
Ch
e
p
l
y
g
i
n
a
,
a
n
d
L
.
T
.
N
i
e
l
s
e
n
,
“
S
e
a
g
ra
s
s
F
i
n
d
e
r:
d
e
e
p
l
e
a
rn
i
n
g
fo
r
e
e
l
g
ra
s
s
d
e
t
e
c
t
i
o
n
a
n
d
c
o
v
e
r
a
g
e
e
s
t
i
m
a
t
i
o
n
i
n
t
h
e
w
i
l
d
,
”
E
c
o
l
o
g
i
c
a
l
In
f
o
r
m
a
t
i
c
s
,
v
o
l
.
9
0
,
2
0
2
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
c
o
i
n
f.
2
0
2
5
.
1
0
3
2
0
0
.
[2
7
]
S
.
Ra
i
n
e
,
R
.
M
a
rc
h
a
n
t
,
P
.
M
o
g
h
a
d
a
m
,
F
.
M
a
i
re
,
B.
K
e
t
t
l
e
,
a
n
d
B.
K
u
s
y
,
“
M
u
l
t
i
-
s
p
e
c
i
e
s
s
e
a
g
ra
s
s
d
e
t
e
c
t
i
o
n
a
n
d
c
l
a
s
s
i
fi
c
a
t
i
o
n
fro
m
u
n
d
e
rw
a
t
e
r
i
m
a
g
e
s
,
”
2
0
2
0
D
i
g
i
t
a
l
Im
a
g
e
Co
m
p
u
t
i
n
g
:
T
e
c
h
n
i
q
u
e
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
(D
ICT
A
)
,
p
p
.
1
–
8
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/
D
ICT
A
5
1
2
2
7
.
2
0
2
0
.
9
3
6
3
3
7
1
.
[2
8
]
I.
F
a
w
w
a
z
,
Y
.
Y
e
n
n
i
m
a
r,
N
.
P
.
D
h
a
rs
i
n
n
i
,
a
n
d
B.
A
.
W
i
j
a
y
a
,
“
T
h
e
o
p
t
i
m
i
z
a
t
i
o
n
o
f
CN
N
a
l
g
o
ri
t
h
m
u
s
i
n
g
t
ra
n
s
fe
r
l
e
a
r
n
i
n
g
fo
r
m
a
ri
n
e
fa
u
n
a
c
l
a
s
s
i
fi
c
a
t
i
o
n
,
”
s
i
n
k
r
o
n
,
v
o
l
.
8
,
n
o
.
4
,
p
p
.
2
2
3
6
–
2
2
4
5
,
O
c
t
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
3
9
5
/
s
i
n
k
ro
n
.
v
8
i
4
.
1
2
8
9
3
.
[2
9
]
T
.
E
e
ro
l
a
e
t
a
l
.
,
“
S
u
rv
e
y
o
f
a
u
t
o
m
a
t
i
c
p
l
a
n
k
t
o
n
i
m
a
g
e
re
c
o
g
n
i
t
i
o
n
:
c
h
a
l
l
e
n
g
e
s
,
e
x
i
s
t
i
n
g
s
o
l
u
t
i
o
n
s
a
n
d
fu
t
u
re
p
e
rs
p
e
c
t
i
v
e
s
,
”
A
r
t
i
f
i
c
i
a
l
In
t
e
l
l
i
g
e
n
c
e
R
e
v
i
e
w
,
v
o
l
.
5
7
,
n
o
.
5
,
A
p
r.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
4
6
2
-
024
-
1
0
7
4
5
-
y.
[3
0
]
J
.
Y
a
n
g
,
M
.
Ca
i
,
X
.
Y
a
n
g
,
a
n
d
Z
.
Z
h
o
u
,
“
U
n
d
e
rw
a
t
e
r
i
m
a
g
e
c
l
a
s
s
i
fi
c
a
t
i
o
n
a
l
g
o
ri
t
h
m
b
a
s
e
d
o
n
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
ra
l
n
e
t
w
o
r
k
a
n
d
o
p
t
i
m
i
z
e
d
e
x
t
re
m
e
l
e
a
r
n
i
n
g
m
a
c
h
i
n
e
,
”
J
o
u
r
n
a
l
o
f
M
a
r
i
n
e
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
1
0
,
n
o
.
1
2
,
D
e
c
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
j
m
s
e
1
0
1
2
1
8
4
1
.
[3
1
]
N
.
A
b
i
d
e
t
a
l
.
,
“
S
e
a
g
ra
s
s
c
l
a
s
s
i
fi
c
a
t
i
o
n
u
s
i
n
g
u
n
s
u
p
e
rv
i
s
e
d
c
u
rri
c
u
l
u
m
l
e
a
rn
i
n
g
(U
CL
),
”
E
c
o
l
o
g
i
c
a
l
In
f
o
r
m
a
t
i
c
s
,
v
o
l
.
8
3
,
N
o
v
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
c
o
i
n
f.
2
0
2
4
.
1
0
2
8
0
4
.
BI
O
G
R
A
P
H
I
ES
O
F
A
U
T
H
O
R
S
N
u
r
u
l
H
ay
at
y
i
s
a
l
e
c
t
ur
e
r
i
n
t
he
D
e
pa
r
t
m
e
nt
of
I
nf
or
m
a
t
i
c
s
E
ng
i
ne
e
r
i
ng
a
t
U
ni
ve
r
s
i
t
a
s
M
a
r
i
t
i
m
R
a
j
a
A
l
i
H
a
j
i
(
U
M
R
A
H
)
,
T
a
n
j
ung
pi
na
ng
,
I
n
done
s
i
a
.
S
he
ho
l
d
s
a
B
a
c
he
l
o
r
'
s
de
gr
e
e
i
n
C
o
m
pu
t
e
r
E
n
gi
ne
e
r
i
ng
a
nd
a
M
a
s
t
e
r
'
s
de
gr
e
e
i
n
C
o
m
put
e
r
S
c
i
e
n
c
e
(
M
.
C
s
)
.
H
e
r
r
e
s
e
a
r
c
h
i
nt
e
r
e
s
t
s
e
nc
o
m
pa
s
s
a
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
ge
nc
e
,
i
m
a
g
e
pr
o
c
e
s
s
i
ng
,
m
a
c
h
i
ne
l
e
a
r
n
i
ng
,
a
nd
d
a
t
a
m
i
n
i
ng
,
w
i
t
h
a
p
a
r
t
i
c
u
l
a
r
f
oc
u
s
on
a
pp
l
i
c
a
t
i
ons
i
n
e
nv
i
r
o
nm
e
n
t
a
l
i
nf
o
r
m
a
t
i
c
s
a
nd
s
m
a
r
t
s
ys
t
e
m
s
.
S
he
h
a
s
c
ont
r
i
bu
t
e
d
t
o
va
r
i
o
us
s
t
ud
i
e
s
,
i
nc
l
ud
i
ng
r
a
i
nf
a
l
l
p
r
e
d
i
c
t
i
o
n
us
i
ng
su
ppo
r
t
v
e
c
t
or
m
a
c
hi
ne
s
a
nd
ne
u
r
a
l
ne
t
w
or
k
m
od
e
l
s
f
o
r
m
a
ngr
ove
s
pe
c
i
e
s
c
l
a
s
s
i
f
i
c
a
t
i
on
.
I
n
t
he
c
ur
r
e
n
t
s
t
u
dy,
s
h
e
l
e
d
t
he
c
on
c
e
p
t
u
a
l
i
z
a
t
i
on
,
d
a
t
a
c
u
r
a
t
i
on,
m
od
e
l
de
v
e
l
op
m
e
n
t
,
a
nd
m
a
n
us
c
r
i
pt
dr
a
f
t
i
ng
p
r
o
c
e
s
s
e
s
.
H
e
r
w
or
k
c
o
nt
i
nu
e
s
t
o
br
i
dg
e
c
o
m
pu
t
a
t
i
ona
l
m
e
t
ho
d
s
w
i
t
h
pr
a
c
t
i
c
a
l
a
pp
l
i
c
a
t
i
o
ns
i
n
e
nv
i
r
on
m
e
n
t
a
l
m
o
ni
t
or
i
n
g
a
n
d
d
e
c
i
s
i
on
s
uppo
r
t
s
ys
t
e
m
s
.
S
he
c
a
n
be
c
ont
a
c
t
e
d
a
t
e
m
a
i
l
:
nur
ul
.
h
a
ya
t
y@
u
m
r
a
h.
a
c
.
i
d
.
H
o
l
l
an
d
a
A
r
i
e
f
K
u
s
u
m
a
i
s
a
l
e
c
t
u
r
e
r
a
t
t
he
D
e
p
a
r
t
m
e
nt
o
f
E
l
e
c
t
r
i
c
a
l
E
ngi
n
e
e
r
i
ng,
U
n
i
ve
r
s
i
t
a
s
M
a
r
i
t
i
m
R
a
j
a
A
l
i
H
a
j
i
(
U
M
R
A
H
)
,
T
a
n
j
ung
pi
n
a
ng
,
I
nd
one
s
i
a
.
H
e
hol
d
s
a
B
a
c
he
l
o
r
’
s
a
nd
M
a
s
t
e
r
’
s
de
gr
e
e
f
r
om
I
P
B
U
n
i
ve
r
s
i
t
y
(
I
ns
t
i
t
u
t
P
e
r
t
a
n
i
a
n
B
o
gor
)
,
s
pe
c
i
a
l
i
z
i
ng
i
n
M
a
r
i
n
e
S
c
i
e
nc
e
a
nd
T
e
c
h
nol
ogy
.
H
i
s
a
c
a
de
m
i
c
a
nd
r
e
s
e
a
r
c
h
i
n
t
e
r
e
s
t
s
l
i
e
i
n
m
a
r
i
ne
t
e
c
hn
ol
ogy
,
ge
o
s
pa
t
i
a
l
i
n
f
or
m
a
t
i
on
s
ys
t
e
m
s
,
und
e
r
w
a
t
e
r
r
o
bot
i
c
s
,
a
n
d
r
e
m
o
t
e
s
e
n
s
i
n
g
a
ppl
i
c
a
t
i
on
s
f
or
c
oa
s
t
a
l
e
nv
i
r
o
nm
e
n
t
s
.
H
e
i
s
a
c
t
i
v
e
l
y
i
n
vol
ve
d
i
n
t
he
de
ve
l
o
pm
e
n
t
of
ope
n
-
s
our
c
e
o
c
e
a
n
ogr
a
ph
i
c
i
n
s
t
r
um
e
n
t
s
a
n
d
t
h
e
i
n
t
e
g
r
a
t
i
on
of
i
n
t
e
r
ne
t
o
f
t
h
i
ng
s
(
I
oT
)
t
e
c
hno
l
og
i
e
s
f
or
r
e
a
l
-
t
i
m
e
c
oa
s
t
a
l
m
oni
t
o
r
i
ng.
O
ve
r
t
he
p
a
s
t
f
e
w
y
e
a
r
s
,
he
h
a
s
l
e
d
a
nd
pa
r
t
i
c
i
pa
t
e
d
i
n
num
e
r
o
us
r
e
s
e
a
r
c
h
p
r
o
j
e
c
t
s
f
oc
u
s
i
n
g
o
n
t
i
d
a
l
pa
t
t
e
r
n
a
na
l
ys
i
s
,
w
a
t
e
r
qua
l
i
t
y
a
s
s
e
s
s
m
e
n
t
,
a
nd
en
vi
r
on
m
e
nt
a
l
da
t
a
c
ol
l
e
c
t
i
on
u
s
i
ng
e
m
b
e
dd
e
d
s
y
s
t
e
m
s
.
I
n
t
h
i
s
s
t
u
d
y,
he
c
on
t
r
i
b
ut
e
d
t
o
pr
oj
e
c
t
s
upe
r
v
i
s
i
on
,
m
e
t
h
odo
l
ogy
de
ve
l
op
m
e
nt
,
da
t
a
va
l
i
d
a
t
i
on
,
a
nd
m
a
n
us
c
r
i
p
t
r
e
vi
e
w
a
nd
e
di
t
i
ng
.
H
e
c
a
n
be
c
o
nt
a
c
t
e
d
a
t
e
m
a
i
l
:
h
ol
l
a
nda
k
us
u
m
a
@u
m
r
a
h
.
a
c
.
i
d
.
Evaluation Warning : The document was created with Spire.PDF for Python.