I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
11
,
No
.
3
,
J
u
n
e
2021
,
p
p
.
2508
~
2
5
1
5
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
11
i
3
.
p
p
2
5
0
8
-
2
5
1
5
2508
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Fruit
t
re
e disea
se
c
la
ss
ificatio
n
sy
st
e
m
us
ing
g
enerati
v
e
a
dv
ersa
ria
l net
w
o
rk
s
Cha
ng
s
u K
i
m
1
,
H
y
eso
o
L
ee
2
,
H
o
ek
y
un
g
J
un
g
3
1
,3
De
p
a
rtm
e
n
t
o
f
Co
m
p
u
ter E
n
g
in
e
e
rin
g
,
P
a
iC
h
a
i
Un
iv
e
rsity
,
Da
e
je
o
n
,
Re
p
u
b
li
c
o
f
Ko
re
a
2
Ko
re
a
In
stit
u
te
o
f
S
c
ien
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
In
f
o
rm
a
ti
o
n
,
Da
e
jeo
n
,
Re
p
u
b
li
c
o
f
Ko
re
a
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l 3
,
2
0
2
0
R
ev
i
s
ed
Dec
1
0
,
2
0
2
0
A
cc
ep
ted
Dec
2
6
,
2
0
2
0
S
m
a
rt
f
a
r
m
re
f
e
rs
to
a
f
a
r
m
th
a
t
c
a
n
re
m
o
tely
a
n
d
a
u
t
o
m
a
ti
c
a
ll
y
m
a
in
tain
p
ro
p
e
r
g
ro
w
th
a
n
d
m
a
n
a
g
e
m
e
n
t
o
f
c
ro
p
s
a
n
d
li
v
e
sto
c
k
b
y
in
teg
ra
ti
n
g
tec
h
n
o
l
o
g
y
w
it
h
a
g
ricu
lt
u
re
.
Cu
r
re
n
tl
y
,
s
m
a
rt
f
a
r
m
s
a
re
c
o
n
c
e
n
tr
a
ted
in
th
e
f
ield
o
f
s
m
a
rt
h
o
rti
c
u
lt
u
re
,
a
n
d
a
lt
h
o
u
g
h
sp
re
a
d
i
n
g
re
se
a
rc
h
is
b
e
in
g
c
o
n
d
u
c
ted
in
li
m
it
e
d
sp
a
c
e
s.
I
n
a
d
d
it
i
o
n
,
it
is
d
if
f
icu
lt
to
o
b
tain
a
su
f
f
icie
n
t
a
m
o
u
n
t
o
f
d
a
ta
to
b
e
u
se
d
f
o
r
l
e
a
rn
in
g
,
a
n
d
th
e
re
is
a
p
ro
b
lem
th
a
t
d
a
ta
im
b
a
lan
c
e
o
c
c
u
rs
b
e
c
a
u
se
it
is
d
iff
icu
lt
to
o
b
tai
n
a
sim
il
a
r
a
m
o
u
n
t
f
o
r
e
a
c
h
c
las
s.
In
t
h
is
p
a
p
e
r,
w
e
p
ro
p
o
s
e
a
m
e
th
o
d
t
o
a
m
p
li
fy
a
s
m
a
ll
a
m
o
u
n
t
o
f
d
a
ta
a
n
d
to
s
o
lv
e
th
e
p
ro
b
lem
s
o
f
i
m
b
a
lan
c
e
d
a
ta
b
y
u
sin
g
a
f
e
a
tu
re
th
a
t
c
a
n
lea
rn
to
m
im
ic
th
e
d
a
ta
o
f
a
g
e
n
e
ra
ti
v
e
a
d
v
e
rsa
rial
n
e
tw
o
rk
.
T
h
e
p
ro
p
o
s
e
d
m
e
th
o
d
c
a
n
c
re
a
te
d
a
tas
e
t
o
f
v
a
rio
u
s
c
ro
p
s
a
n
d
a
ls
o
sh
o
w
h
ig
h
h
it
ra
t
e
.
Da
tas
e
t
g
e
n
e
ra
ted
f
ro
m
c
ro
p
s
w
o
u
ld
b
e
u
se
d
to
so
lv
e
p
r
o
b
lem
s
o
f
d
a
ta
imb
a
lan
c
e
b
y
lea
rn
in
g
.
K
ey
w
o
r
d
s
:
C
las
s
i
f
icatio
n
s
y
s
te
m
GAN
Ma
ch
i
n
e
lear
n
i
n
g
S
m
ar
t
f
ar
m
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
:
Ho
ek
y
u
n
g
J
u
n
g
Dep
a
rt
m
en
t o
f
C
o
m
p
u
ter
E
n
g
i
n
ee
r
in
g
P
aich
ai
Un
i
v
er
s
it
y
1
55
-
40
B
ae
j
ae
-
ro
,
Seo
g
u
,
Dae
J
eo
n
,
R
ep
u
b
lic
o
f
Ko
r
ea
E
m
ail:
h
k
j
u
n
g
@
p
c
u
.
ac
.
k
r
1.
I
NT
RO
D
UCT
I
O
N
S
m
ar
t
f
ar
m
i
n
g
is
e
v
o
lv
i
n
g
w
i
t
h
th
e
ap
p
licatio
n
o
f
th
e
4
th
in
d
u
s
tr
ial
r
ev
o
l
u
tio
n
tec
h
n
o
lo
g
y
.
I
t
r
ef
er
s
to
a
tech
n
o
lo
g
y
t
h
at
en
ab
le
s
r
e
m
o
te
co
n
tr
o
l
as
w
ell
as
s
e
r
v
ices
s
u
c
h
as
au
to
m
atio
n
an
d
in
tell
ig
e
n
ce
b
y
in
te
g
r
atin
g
n
e
w
tech
n
o
lo
g
ie
s
s
u
c
h
as
I
C
T
,
I
o
T
,
b
ig
d
ata,
clo
u
d
,
an
d
A
I
i
n
to
t
h
e
g
r
o
w
t
h
a
n
d
en
v
ir
o
n
m
e
n
t
o
f
cr
o
p
s
o
r
liv
esto
ck
[1
-
3
]
.
T
h
e
n
ee
d
f
o
r
s
m
ar
t
f
ar
m
s
i
s
d
ee
p
ly
li
n
k
ed
to
g
lo
b
al
cli
m
ate
ch
a
n
g
e
an
d
f
o
o
d
s
h
o
r
tag
e
s
as
p
o
p
u
latio
n
s
g
r
o
w.
Fo
o
d
d
em
a
n
d
is
o
n
t
h
e
r
is
e
d
u
e
to
an
in
cr
ea
s
e
in
th
e
w
o
r
l
d
'
s
p
o
p
u
latio
n
,
b
u
t
th
er
e
is
a
g
r
o
w
in
g
s
h
o
r
ta
g
e
o
f
p
eo
p
le
w
h
o
w
il
l
g
r
o
w
b
ec
au
s
e
o
f
th
e
s
h
r
in
k
i
n
g
a
n
d
ag
in
g
p
o
p
u
latio
n
.
I
n
o
t
h
er
w
o
r
d
s
,
as
th
e
p
o
p
u
latio
n
in
cr
ea
s
es,
u
r
b
an
izatio
n
r
ed
u
ce
s
th
e
ar
ea
o
f
cr
o
p
cu
ltiv
atio
n
,
a
n
d
f
ar
m
er
s
o
n
th
e
p
r
o
d
u
ctio
n
s
ite
ar
e
ag
in
g
.
E
v
e
n
i
n
Ko
r
ea
,
as
o
f
2
0
1
7
,
th
e
av
er
ag
e
ag
e
o
f
f
ar
m
er
s
'
c
h
ie
f
e
x
ec
u
ti
v
es
i
s
6
7
,
an
d
th
is
a
v
er
ag
e
ag
e
i
s
o
n
l
y
i
n
cr
e
asin
g
w
it
h
ti
m
e.
As
a
r
esu
lt,
s
m
ar
t
f
ar
m
s
ar
e
b
ec
o
m
i
n
g
in
cr
ea
s
in
g
l
y
i
m
p
o
r
tan
t.
Ho
w
e
v
er
,
s
m
ar
t
f
ar
m
s
ar
e
co
n
ce
n
tr
ated
i
n
t
h
e
f
ie
ld
o
f
s
m
a
r
t
h
o
r
ticu
lt
u
r
e,
an
d
alt
h
o
u
g
h
th
e
y
ar
e
r
ec
en
tl
y
s
p
r
ea
d
in
g
,
r
esear
ch
is
b
ein
g
c
o
n
d
u
cted
in
li
m
ited
s
p
ac
es,
s
u
ch
as
f
ac
ilit
ie
s
.
I
n
ad
d
itio
n
,
it
i
s
d
if
f
ic
u
lt
to
o
b
tain
a
s
u
f
f
icie
n
t
a
m
o
u
n
t
o
f
d
ata
to
b
e
u
s
ed
f
o
r
lear
n
in
g
,
a
n
d
th
er
e
is
a
p
r
o
b
le
m
t
h
at
d
ata
i
m
b
a
l
a
n
ce
o
cc
u
r
s
b
ec
a
u
s
e
it is
d
if
f
ic
u
lt to
o
b
tain
a
s
i
m
ila
r
am
o
u
n
t f
o
r
ea
ch
clas
s
[
2
]
.
I
n
t
h
is
p
ap
er
,
w
e
p
r
o
p
o
s
e
a
m
et
h
o
d
to
a
m
p
l
if
y
t
h
e
a
m
o
u
n
t
o
f
d
ata
t
h
r
o
u
g
h
g
en
er
a
tiv
e
n
e
u
r
al
n
et
w
o
r
k
s
an
d
to
s
o
lv
e
d
ata
im
b
a
lan
ce
b
et
w
ee
n
class
e
s
b
y
g
en
er
ati
n
g
d
ata
o
f
v
ar
io
u
s
class
es.
T
h
e
p
r
o
p
o
s
ed
tech
n
iq
u
e
i
m
p
r
o
v
es
t
h
e
s
p
ee
d
b
y
r
ed
u
ci
n
g
th
e
a
m
o
u
n
t
o
f
co
m
p
u
ta
tio
n
b
y
p
r
ep
r
o
ce
s
s
i
n
g
th
e
d
ata
a
n
d
h
ig
h
li
g
h
ts
t
h
e
f
ea
tu
r
es
to
h
elp
lear
n
in
g
.
T
h
e
p
r
ep
r
o
ce
s
s
ed
d
ata
u
s
es
a
g
en
er
ati
v
e
n
e
u
r
al
n
et
w
o
r
k
to
g
e
n
er
at
e
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
-
8708
F
r
u
it tr
ee
d
is
ea
s
e
c
la
s
s
ifica
tio
n
s
ystem
u
s
in
g
g
en
era
tive
a
d
v
ers
a
r
ia
l n
etw
o
r
k
s
(
C
h
a
n
g
s
u
K
im
)
2509
n
e
w
d
ata
an
d
f
ilter
s
o
u
t
t
h
e
lo
w
q
u
ali
t
y
d
ata
th
r
o
u
g
h
f
ilter
i
n
g
f
o
r
d
ata
in
teg
r
it
y
.
I
n
ca
s
e
o
f
u
s
i
n
g
th
e
p
r
o
p
o
s
ed
tech
n
iq
u
e,
it
is
p
o
s
s
ib
le
to
a
m
p
li
f
y
a
s
m
all
a
m
o
u
n
t
o
f
d
ata
to
g
e
n
er
ate
a
s
u
f
f
icie
n
t
a
m
o
u
n
t
o
f
d
ata
f
o
r
lear
n
in
g
,
a
n
d
to
ap
p
ly
it
to
an
ex
is
t
in
g
s
y
s
te
m
a
s
it i
s
.
B
ased
o
n
th
e
s
e
ad
v
a
n
tag
e
s
,
th
e
p
r
o
p
o
s
ed
m
eth
o
d
ca
n
b
e
u
s
ed
to
co
m
p
en
s
ate
f
o
r
th
e
li
m
itatio
n
s
o
f
t
h
e
r
ec
en
t
s
m
ar
t f
ar
m
.
2.
RE
L
AT
E
D
R
E
SE
ARCH
I
n
t
h
is
c
h
ap
ter
,
w
e
a
n
al
y
ze
th
e
r
e
g
io
n
o
f
i
n
ter
est
ex
tr
ac
tio
n
m
eth
o
d
a
n
d
t
h
e
g
en
er
ati
v
e
n
eu
r
a
l
n
et
w
o
r
k
u
s
ed
in
p
r
ep
r
o
ce
s
s
in
g
an
d
th
e
p
r
o
b
lem
s
th
a
t c
an
b
e
ca
u
s
ed
b
y
d
ata
i
m
b
ala
n
ce
.
2
.
1
.
G
ener
a
t
iv
e
a
dv
er
s
a
ri
a
l
net
w
o
rk
s
(
G
A
N)
GAN
is
a
p
ap
er
w
r
it
ten
b
y
I
an
Go
o
d
f
ello
w
i
n
2
0
1
4
an
d
v
ar
io
u
s
s
t
u
d
ies
ar
e
b
ein
g
co
n
d
u
cted
b
ased
o
n
it
[
4
-
7
]
.
T
h
e
g
en
er
ati
v
e
h
o
s
tile
n
e
u
r
al
n
e
t
w
o
r
k
lear
n
s
an
d
p
r
o
d
u
ce
s
th
e
r
es
u
lt
t
h
r
o
u
g
h
th
e
co
m
p
etitio
n
o
f
t
w
o
n
e
u
r
al
n
et
w
o
r
k
s
as
th
e
n
a
m
e
o
f
th
e
h
o
s
tile
n
e
u
r
al
n
et
w
o
r
k
.
In
(
1
)
r
ep
r
esen
ts
th
e
f
o
r
m
u
la
o
f
th
e
g
en
er
ati
v
e
h
o
s
t n
e
u
r
al
n
e
t
w
o
r
k
(
,
)
=
~
(
)
[
l
og
(
)
]
+
~
(
)
[
l
og
(
1
−
(
(
)
)
]
]
(
1
)
I
n
th
e
ab
o
v
e
f
o
r
m
u
la,
x
~
Pd
ata
(
x
)
r
ep
r
esen
ts
an
im
ag
e
g
en
er
ated
b
y
th
e
p
r
o
b
ab
ility
d
is
tr
ib
u
ti
o
n
o
f
th
e
actu
al
im
ag
e,
an
d
x
~
Px
(
z)
r
e
p
r
esen
ts
an
im
ag
e
g
en
er
ated
u
s
in
g
n
o
is
e.
D
(
x
)
is
a
d
is
cr
im
in
ato
r
,
wit
h
a
v
alu
e
b
etween
0
an
d
1
in
d
icatin
g
th
e
p
r
o
b
ab
ility
th
at
th
e
im
a
g
e
is
r
e
al.
I
n
th
e
case
o
f
D
(
G
(
z)
)
,
th
e
im
ag
e
g
en
er
ated
b
y
th
e
cr
eato
r
is
d
is
tin
g
u
is
h
ed
t
h
r
o
u
g
h
th
e
d
i
s
cr
im
in
ato
r
,
wh
ich
al
s
o
h
as
a
p
r
o
b
a
b
ilit
y
b
etween
0
an
d
1
th
at
th
e
im
ag
e
is
r
eal.
T
o
m
a
x
im
ize
th
e
ab
o
v
e
eq
u
atio
n
,
th
e
v
alu
e
o
f
D
(
G
(
z)
)
s
h
o
u
ld
b
e
clo
s
e
to
1
an
d
t
h
e
v
alu
e
o
f
D
(
x
)
s
h
o
u
ld
b
e clo
s
e to
1
.
I
n
th
is
wa
y
,
co
n
s
tr
u
cto
r
s
an
d
d
is
cr
im
in
ato
r
s
lear
n
b
y
s
o
lv
in
g
th
e Min
m
a
x
p
r
o
b
lem
.
2
.
2
.
Cro
p r
e
g
io
n o
f
inte
re
s
t
(
Ro
I
)
T
h
e
r
eg
io
n
o
f
in
ter
e
s
t
m
ea
n
s
an
ar
ea
o
f
in
ter
est
o
n
th
e
i
m
ag
e.
W
h
e
n
t
h
e
o
b
j
ec
t
is
d
etec
ted
o
r
d
etec
ted
b
y
p
r
o
ce
s
s
in
g
an
i
m
ag
e
[
8
-
10
]
,
th
e
d
etec
ted
ar
ea
m
a
y
b
e
r
ef
er
r
ed
to
as
a
r
eg
io
n
o
f
in
ter
e
s
t
[
11
].
T
h
er
e
ar
e
th
r
ee
m
ain
r
ea
s
o
n
s
f
o
r
s
p
ec
if
y
i
n
g
a
R
e
g
io
n
o
f
I
n
t
er
est.
T
h
e
f
ir
s
t
i
s
to
r
e
m
o
v
e
u
n
n
ec
e
s
s
ar
y
i
m
ag
e
s
o
f
th
e
ar
ea
ar
o
u
n
d
t
h
e
o
b
j
ec
t,
an
d
th
e
s
ec
o
n
d
is
to
r
ed
u
ce
th
e
a
m
o
u
n
t
o
f
co
m
p
u
tatio
n
an
d
r
eso
u
r
ce
s
.
Fi
n
all
y
,
th
e
ac
cu
r
a
c
y
o
f
lear
n
i
n
g
ca
n
b
e
im
p
r
o
v
ed
.
B
y
eli
m
in
at
in
g
u
n
n
ec
es
s
ar
y
in
f
o
r
m
at
io
n
in
ad
v
an
ce
,
o
n
l
y
n
ec
es
s
ar
y
p
ar
ts
ar
e
u
s
ed
f
o
r
lear
n
in
g
,
w
h
ich
i
n
cr
ea
s
es a
cc
u
r
a
c
y
.
T
h
e
m
o
s
t
r
ep
r
esen
tati
v
e
o
n
es
in
cr
o
p
r
eg
io
n
o
f
in
ter
est
m
a
y
b
e
r
ef
er
r
ed
to
as
R
GB
s
ep
a
r
atio
n
an
d
co
n
to
u
r
ex
tr
ac
tio
n
.
I
n
t
h
e
ca
s
e
o
f
R
GB
s
ep
ar
atio
n
,
a
r
e
g
io
n
o
f
in
ter
est
ca
n
b
e
o
b
tain
ed
b
y
s
p
ec
if
y
i
n
g
a
r
an
g
e
o
f
ch
a
n
n
els
a
n
d
o
u
tp
u
tti
n
g
p
i
x
el
v
alu
e
s
w
it
h
i
n
a
r
a
n
g
e
b
y
u
s
i
n
g
p
r
o
p
er
ties
h
av
in
g
d
i
f
f
er
en
t
v
alu
e
s
f
o
r
ea
ch
ch
an
n
el
o
f
an
i
m
a
g
e.
C
o
n
to
u
r
ex
tr
ac
tio
n
m
ea
n
s
ex
tr
ac
ti
n
g
a
p
o
in
t
w
h
er
e
th
e
b
r
ig
h
tn
e
s
s
o
f
th
e
i
m
a
g
e
ch
a
n
g
es
f
r
o
m
a
lo
w
v
al
u
e
to
a
h
ig
h
v
alu
e
o
r
v
ice
v
er
s
a.
C
o
n
to
u
r
d
etec
tio
n
is
a
tech
n
iq
u
e
to
f
in
d
th
e
p
ix
el
s
co
r
r
esp
o
n
d
in
g
to
th
e
co
n
to
u
r
,
an
d
ca
lcu
lates
t
h
e
s
lo
p
e
b
as
ed
o
n
th
e
ca
lcu
latio
n
o
f
t
h
e
p
ar
tial
d
if
f
er
en
tial
o
p
er
ato
r
[
12
-
13
]
.
2
.
3
.
Da
t
a
im
ba
la
nce
I
f
th
er
e
is
a
lar
g
e
d
if
f
er
en
ce
in
th
e
a
m
o
u
n
t
o
f
d
ata
th
at
ea
ch
class
h
as
in
th
e
d
ata,
it
is
s
aid
th
at
th
er
e
is
a
cla
s
s
i
m
b
ala
n
ce
.
Mo
s
t
o
f
t
h
e
d
ata
i
n
t
h
e
r
ea
l
w
o
r
ld
h
as
t
h
is
d
ata
i
m
b
ala
n
ce
p
r
o
b
le
m
,
a
n
d
it
r
eb
alan
ce
s
t
h
e
class
es
b
y
s
a
m
p
lin
g
a
g
ai
n
b
ef
o
r
e
tr
ain
in
g
th
e
m
o
d
el.
Un
d
er
s
a
m
p
lin
g
is
s
elec
ti
n
g
p
ar
ts
o
f
a
lo
t
o
f
d
ata,
s
u
ch
as
th
e
i
m
a
g
e
o
n
t
h
e
le
f
t,
a
n
d
ali
g
n
i
n
g
t
h
e
m
to
w
ar
d
th
e
s
m
al
le
r
d
ata.
Un
d
er
s
a
m
p
li
n
g
ca
n
r
e
d
u
ce
ex
ec
u
tio
n
ti
m
e
b
y
r
ed
u
ci
n
g
th
e
s
ize
o
f
t
h
e
d
at
a
s
et,
b
u
t
it
i
s
al
s
o
p
o
s
s
ib
le
t
h
at
th
e
u
s
ef
u
l
d
ata
is
n
o
t
e
x
tr
ac
ted
o
r
b
iased
to
o
n
e
s
id
e
[
14
-
1
7
]
.
Ov
er
s
a
m
p
lin
g
i
n
v
o
l
v
es
co
p
y
i
n
g
less
d
ata
a
n
d
f
itt
in
g
it
to
w
ar
d
m
o
r
e
d
ata.
I
n
th
e
ca
s
e
o
f
o
v
er
s
a
m
p
li
n
g
,
th
e
d
ata
s
et
h
a
s
m
o
r
e
o
v
er
s
a
m
p
li
n
g
,
s
o
lear
n
in
g
i
s
b
etter
.
I
n
o
t
h
er
w
o
r
d
s
,
it
p
er
f
o
r
m
s
b
etter
th
an
u
n
d
er
s
a
m
p
lin
g
.
Ho
w
e
v
er
,
th
er
e
is
a
p
o
s
s
i
b
ilit
y
o
f
o
v
er
f
itti
n
g
w
h
e
n
t
h
e
s
a
m
e
d
ata
i
s
e
x
tr
ac
ted
r
ep
ea
ted
l
y
[
1
8
-
19
].
2
.
4
.
P
r
o
ble
m
s
a
nd
s
o
lutio
n
s
P
r
o
b
lem
s
b
et
w
ee
n
s
m
ar
t
f
ar
m
s
an
d
d
ata
in
cl
u
d
e:
I
t
is
d
i
f
f
ic
u
lt
to
o
b
tain
e
n
o
u
g
h
d
ata
to
b
e
u
s
ed
f
o
r
lear
n
in
g
,
an
d
ev
e
n
t
h
e
p
r
o
v
id
ed
d
ata
h
as
an
ad
v
er
s
e
ef
f
e
ct
o
n
lear
n
in
g
d
u
e
to
u
n
b
ala
n
ce
d
d
ata
p
er
class
[
20
-
23
]
.
B
ased
o
n
th
ese
p
r
o
b
l
e
m
s
,
t
h
is
p
ap
er
p
r
o
p
o
s
es
a
m
e
th
o
d
o
f
a
m
p
li
f
y
in
g
t
h
e
a
m
o
u
n
t
o
f
d
ata
th
r
o
u
g
h
a
g
en
er
ati
v
e
a
n
ta
g
o
n
i
s
tic
n
eu
r
al
n
et
w
o
r
k
a
n
d
g
en
er
ati
n
g
d
ata
o
f
v
ar
io
u
s
cla
s
s
e
s
to
s
o
l
v
e
t
h
e
d
ata
i
m
b
a
lan
ce
b
et
w
ee
n
cla
s
s
es.
Usi
n
g
a
g
e
n
er
ativ
e
ad
v
er
s
ar
ial
n
et
w
o
r
k
n
o
t
o
n
l
y
a
m
p
li
f
ie
s
a
s
m
all
a
m
o
u
n
t
o
f
d
ata
to
g
en
er
ate
en
o
u
g
h
d
ata
f
o
r
lear
n
in
g
,
b
u
t
also
h
a
s
th
e
ad
v
a
n
ta
g
e
o
f
u
s
in
g
o
v
er
s
a
m
p
li
n
g
w
it
h
less
r
ed
u
n
d
an
c
y
o
f
th
e
d
ata
u
s
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
3
,
J
u
n
e
2021
:
2
5
0
8
-
2515
2510
3.
SYST
E
M
DE
SI
G
N
3
.
1
.
O
v
er
a
ll sy
s
t
e
m
des
ig
n
T
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
ca
n
b
e
d
iv
id
ed
in
to
t
h
e
d
ata
p
r
o
ce
s
s
in
g
s
ta
g
e
t
h
at
m
a
n
a
g
es
d
ata
p
r
ep
r
o
ce
s
s
in
g
an
d
p
o
s
tp
r
o
ce
s
s
in
g
,
an
d
th
e
n
et
w
o
r
k
s
tag
e
t
h
at
d
ef
in
e
s
th
e
m
o
d
el
f
o
r
g
en
er
ati
n
g
d
ata
an
d
class
i
f
y
in
g
i
m
ag
e
s
.
Fig
u
r
e
1
s
h
o
w
s
t
h
i
s
s
tr
u
ct
u
r
e.
I
n
th
e
d
ata
p
r
o
ce
s
s
i
n
g
s
tag
e,
it
i
s
d
iv
id
ed
in
to
p
r
ep
r
o
ce
s
s
in
g
an
d
p
o
s
tp
r
o
ce
s
s
in
g
.
I
n
t
h
e
p
r
ep
r
o
c
ess
i
n
g
s
tep
,
i
m
a
g
e
s
ize
ad
j
u
s
t
m
en
t,
co
n
tr
ast
ad
j
u
s
t
m
e
n
t,
a
n
d
r
eg
io
n
o
f
i
n
ter
est
ex
tr
ac
tio
n
ar
e
p
er
f
o
r
m
ed
.
I
m
a
g
e
r
esizi
n
g
r
ed
u
ce
s
t
h
e
s
ize
o
f
th
e
i
m
ag
e,
lo
w
er
i
n
g
t
h
e
m
e
m
o
r
y
f
o
o
tp
r
in
t
o
f
t
h
e
s
y
s
te
m
a
n
d
r
ed
u
cin
g
co
m
p
u
ta
tio
n
.
T
h
e
s
ize
o
f
th
e
i
m
a
g
e
is
ch
an
g
ed
to
th
e
n
p
o
w
er
o
f
2
to
ad
j
u
s
t
th
e
s
ize
o
f
th
e
i
m
ag
e
s
o
t
h
at
it
d
o
es
n
o
t
g
iv
e
a
r
ea
l
v
al
u
e
w
h
e
n
tr
ain
in
g
th
e
m
o
d
el.
R
e
g
io
n
o
f
i
n
ter
est
ex
tr
ac
tio
n
r
ed
u
ce
s
th
e
a
m
o
u
n
t
o
f
co
m
p
u
tat
io
n
b
y
r
ed
u
cin
g
d
ata
o
u
ts
id
e
th
e
r
e
g
i
o
n
o
f
i
n
ter
est.
C
o
n
tr
ast
m
a
k
es
th
e
c
h
ar
ac
ter
is
tic
s
o
f
t
h
e
d
ata
s
tan
d
o
u
t
b
y
m
ak
i
n
g
th
e
co
n
tr
as
t
a
n
d
h
u
e
clea
r
e
r
.
P
o
s
t
-
p
r
o
ce
s
s
in
g
i
n
v
o
lv
es
a
d
ata
f
ilter
th
at
eit
h
er
s
to
r
es
th
e
r
es
u
lt
s
o
r
f
ilter
s
t
h
e
d
a
ta.
T
h
e
d
ata
f
ilter
f
ilter
s
o
u
t
lo
w
q
u
alit
y
d
ata
f
r
o
m
i
m
ag
es
g
en
er
ated
b
y
g
en
et
ic
ad
v
er
s
ar
ial
n
e
t
w
o
r
k
s
to
en
s
u
r
e
t
h
e
q
u
a
lit
y
o
f
th
e
g
en
er
ated
i
m
a
g
e.
Da
ta
s
to
r
ag
e
is
r
esp
o
n
s
ib
le
f
o
r
s
en
d
i
n
g
d
ata
to
th
e
cla
s
s
i
f
icati
o
n
s
y
s
te
m
w
h
e
n
t
h
e
u
s
er
g
at
h
e
r
s
m
o
r
e
th
a
n
a
p
r
ed
eter
m
i
n
ed
a
m
o
u
n
t.
Fig
u
r
e
1
.
S
y
s
te
m
ar
ch
itect
u
r
e
d
iag
r
a
m
A
t
t
h
e
n
et
w
o
r
k
le
v
el,
th
e
g
e
n
er
ato
r
is
d
iv
id
ed
in
to
a
g
en
er
ato
r
,
a
d
is
cr
im
i
n
ato
r
n
et
w
o
r
k
,
an
d
a
class
i
f
icatio
n
n
et
w
o
r
k
th
at
d
i
v
id
es
th
e
i
m
a
g
e.
Gen
et
ic
ad
v
er
s
ar
ial
n
et
w
o
r
k
s
ta
k
e
p
r
ep
r
o
ce
s
s
ed
d
ata
as
in
p
u
t
an
d
g
en
er
ate
n
e
w
i
m
ag
e
s
b
ased
o
n
it.
I
n
th
e
class
i
f
icatio
n
s
y
s
te
m
,
t
h
e
g
en
er
ated
i
m
a
g
e
s
ar
e
lear
n
ed
an
d
th
e
ac
tu
al
i
m
a
g
es
ar
e
d
eter
m
i
n
ed
b
ased
o
n
th
e
g
en
er
ated
im
a
g
e
s
.
I
n
th
e
im
a
g
e
d
is
cr
i
m
in
at
io
n
s
tep
,
th
e
v
alid
it
y
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
is
v
er
i
f
ie
d
b
y
co
m
p
ar
in
g
t
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
w
it
h
th
e
e
x
is
t
in
g
s
y
s
te
m
.
Fig
u
r
e
2
s
h
o
w
s
th
e
f
lo
w
c
h
ar
t.
T
h
e
f
lo
w
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
p
r
o
ce
ed
s
f
r
o
m
th
e
s
ize
co
n
tr
o
l
o
f
th
e
in
p
u
t
d
ata.
B
y
ad
j
u
s
tin
g
t
h
e
s
ize,
it
p
la
y
s
a
r
o
le
in
s
p
ee
d
i
n
g
u
p
th
e
f
u
tu
r
e
w
o
r
k
.
T
h
e
r
esized
i
m
ag
e
i
s
th
e
n
ex
tr
ac
ted
f
r
o
m
t
h
e
r
eg
io
n
o
f
i
n
ter
est.
T
h
is
p
la
y
s
a
r
o
le
o
f
i
m
p
r
o
v
in
g
t
h
e
s
p
ee
d
o
f
lear
n
i
n
g
b
y
r
ed
u
ci
n
g
t
h
e
a
m
o
u
n
t
o
f
co
m
p
u
tatio
n
later
.
T
h
e
f
in
al
p
r
ep
r
o
ce
s
s
in
g
is
t
o
g
iv
e
th
e
i
m
ag
e
a
co
n
tr
as
t.
T
h
e
im
a
g
e
is
th
e
n
g
en
e
r
ated
t
h
r
o
u
g
h
g
e
n
etic
ad
v
er
s
ar
ial
n
et
w
o
r
k
s
.
T
h
e
i
m
a
g
e
y
o
u
cr
ea
te
i
s
b
ased
o
n
a
n
u
n
i
n
f
ec
ted
i
m
ag
e,
w
h
ic
h
allo
w
s
y
o
u
to
cr
ea
te
a
n
i
m
a
g
e
s
i
m
ilar
to
th
e
d
is
ea
s
ed
o
n
e
f
r
o
m
th
e
o
r
ig
i
n
al.
T
h
e
g
en
er
ated
i
m
ag
e
i
s
s
av
ed
o
r
d
is
ca
r
d
ed
ac
co
r
d
in
g
to
its
q
u
alit
y
a
n
d
is
r
ep
ea
ted
u
n
t
il
t
h
e
i
m
ag
e
r
ea
c
h
es
a
ce
r
t
ain
n
u
m
b
er
.
W
h
en
m
o
r
e
t
h
a
n
a
ce
r
tain
n
u
m
b
er
o
f
i
m
ag
e
s
ar
e
cr
ea
ted
,
th
e
g
en
er
atio
n
s
to
p
s
an
d
t
h
e
class
i
f
icat
i
o
n
m
o
d
el
is
u
s
ed
to
tr
ain
th
e
cla
s
s
i
f
icat
io
n
m
o
d
el.
3.
2
.
Cro
p r
e
g
io
n o
f
inte
re
s
t
des
ig
n
T
h
e
r
eg
io
n
o
f
in
ter
est
w
as
e
x
tr
ac
ted
u
s
i
n
g
s
o
m
e
So
b
el
m
ask
m
eth
o
d
b
ased
o
n
R
GB
ex
tr
ac
tio
n
.
Fig
u
r
e
3
s
h
o
w
s
th
e
f
lo
w
c
h
ar
t
.
First,
th
e
i
m
a
g
e
is
s
ea
r
c
h
ed
an
d
th
e
av
er
a
g
e
v
al
u
e
is
ca
lcu
lated
.
T
h
e
av
er
ag
e
v
alu
e
is
ca
lcu
la
ted
u
s
in
g
a
So
b
el
m
as
k
.
F
in
d
t
h
e
a
v
er
ag
e
a
n
d
s
tar
t
ch
ec
k
i
n
g
ev
er
y
p
ix
e
l.
I
f
t
h
e
a
v
er
ag
e
v
al
u
e
is
n
o
t
e
x
ce
ed
ed
,
th
e
n
e
x
t
p
i
x
el
is
ch
ec
k
ed
.
I
f
t
h
e
av
er
a
g
e
v
a
l
u
e
is
e
x
ce
ed
ed
,
th
e
n
ex
t
p
i
x
el
is
ch
ec
k
ed
to
s
ee
i
f
th
e
p
ix
el
is
s
to
r
ed
.
I
f
n
o
t,
th
e
p
ix
el
is
s
to
r
ed
an
d
m
o
v
ed
to
th
e
p
ix
el.
A
t
th
is
ti
m
e,
th
e
i
n
s
p
ec
tio
n
d
ir
ec
tio
n
o
f
th
e
p
ix
el
p
r
o
ce
ed
s
co
n
s
ta
n
tl
y
.
I
f
th
e
p
ix
el
i
s
s
to
r
ed
,
th
e
s
ea
r
ch
en
d
s
an
d
a
m
a
s
k
i
s
cr
ea
t
ed
f
r
o
m
t
h
e
s
to
r
ed
p
ix
els to
ex
tr
ac
t t
h
e
p
o
in
t.
3
.
4
.
G
AN
des
ig
n
T
h
e
d
esig
n
ed
g
e
n
er
ati
v
e
ad
v
e
r
s
ar
ial
n
et
w
o
r
k
s
is
d
es
ig
n
ed
b
ased
o
n
D
C
G
A
N.
T
h
e
co
n
s
tr
u
cto
r
u
s
es
th
e
ex
i
s
ti
n
g
m
o
d
el
o
f
g
e
n
er
ati
v
e
ad
v
er
s
ar
ial
n
e
t
w
o
r
k
s
,
b
u
t
r
ec
eiv
es
6
4
x
6
4
x
3
v
ar
iab
les
w
h
ich
ar
e
th
e
s
ize
o
f
th
e
i
m
a
g
e
to
b
e
g
e
n
er
ated
as
in
p
u
t
v
al
u
es
[
24
-
25
].
Fig
u
r
e
4
s
h
o
w
s
t
h
e
m
o
d
el
o
f
t
h
e
d
is
cr
i
m
i
n
ato
r
n
et
w
o
r
k
.
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
-
8708
F
r
u
it tr
ee
d
is
ea
s
e
c
la
s
s
ifica
tio
n
s
ystem
u
s
in
g
g
en
era
tive
a
d
v
ers
a
r
ia
l n
etw
o
r
k
s
(
C
h
a
n
g
s
u
K
im
)
2511
T
h
e
d
is
cr
i
m
i
n
ato
r
w
as
m
o
d
if
i
ed
ac
co
r
d
in
g
to
t
h
e
i
m
a
g
e
g
iv
en
as
in
p
u
t,
w
h
ic
h
h
as
a
ce
r
tain
f
o
r
m
o
f
lea
f
,
t
h
e
d
is
ea
s
e
o
cc
u
r
r
in
g
in
t
h
e
lea
f
,
an
d
th
e
s
ize
o
f
th
e
i
m
a
g
e
is
6
4
x
6
4
x
3
.
T
h
e
s
ize
o
f
th
e
co
n
v
o
lu
tio
n
al
la
y
er
w
as
ch
an
g
ed
f
r
o
m
5
x
5
to
3
x
3
to
r
ed
u
ce
th
e
a
m
o
u
n
t
o
f
co
m
p
u
ta
tio
n
an
d
s
a
v
e
m
e
m
o
r
y
.
I
n
ad
d
itio
n
,
th
e
ac
ti
v
atio
n
f
u
n
ctio
n
is
d
esi
g
n
ed
to
g
en
er
at
e
i
m
ag
e
s
o
f
v
ar
io
u
s
cla
s
s
e
s
u
s
i
n
g
s
o
f
t
m
a
x
r
ath
er
th
a
n
s
ig
m
o
i
d
.
Fig
u
r
e
2
.
S
y
s
te
m
f
lo
w
ch
ar
t
Fig
u
r
e
3
.
C
r
o
p
r
eg
io
n
o
f
in
ter
e
s
t f
lo
w
c
h
ar
t
Fig
u
r
e
4
.
Stru
ct
u
r
e
o
f
d
is
cr
i
m
i
n
ato
r
3
.
5
.
Cla
s
s
if
ica
t
io
n
s
y
s
t
e
m
de
s
ig
n
T
h
e
d
esig
n
ed
m
o
d
el
w
as
d
es
i
g
n
ed
b
ased
o
n
C
NN.
First,
i
n
th
e
i
n
p
u
t
p
h
a
s
e,
th
e
i
m
a
g
e
is
ch
an
g
ed
to
co
d
e
an
d
th
e
d
ata
f
o
r
ea
ch
class
is
ad
j
u
s
ted
r
eg
u
lar
l
y
.
T
h
er
e
ar
e
5
co
n
v
o
lu
tio
n
a
l
la
y
er
s
,
w
h
ic
h
ch
a
n
g
e
ac
co
r
d
in
g
to
t
h
e
s
ize
o
f
t
h
e
i
n
p
u
t
i
m
ag
e.
T
h
e
s
m
a
ller
th
e
s
iz
e
o
f
t
h
e
i
m
ag
e,
th
e
s
m
aller
t
h
e
s
ize
o
f
th
e
f
ea
tu
r
e,
s
o
th
er
e
is
n
o
n
ee
d
to
b
u
ild
a
d
ee
p
er
c
o
n
v
o
lu
t
io
n
al
la
y
er
.
T
h
er
ef
o
r
e,
th
e
i
m
ag
e
i
n
t
h
is
p
a
p
er
is
6
4
x
6
4
x
3
,
s
o
5
co
n
v
o
lu
tio
n
al
la
y
er
s
a
n
d
1
f
u
ll
y
-
co
n
n
ec
ted
la
y
er
ar
e
u
s
ed
.
T
h
e
co
n
v
o
lu
t
io
n
al
la
y
er
is
a
f
ilter
t
h
at
i
s
ch
ar
ac
ter
ized
b
y
a
3
x
3
f
ilter
an
d
a
Ma
x
P
o
o
lin
g
la
y
er
th
at
ig
n
o
r
es
m
i
n
o
r
ch
an
g
es.
Ma
x
P
o
o
lin
g
cr
ea
tes
a
s
m
al
l
o
u
tp
u
t
i
m
a
g
e
b
y
ex
tr
ac
t
i
n
g
o
n
l
y
t
h
e
m
ain
v
al
u
es
f
r
o
m
th
e
f
ilter
ed
i
m
a
g
e.
T
h
e
Ma
x
P
o
o
lin
g
s
ize
i
s
u
s
ed
to
r
ed
u
ce
th
e
s
ize
o
f
th
e
i
m
ag
e
b
y
h
al
f
e
v
er
y
2
x
2
.
I
n
t
h
e
f
u
l
l
y
-
co
n
n
ec
ted
la
y
er
,
th
e
i
m
a
g
e
w
h
ic
h
h
a
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
3
,
J
u
n
e
2021
:
2
5
0
8
-
2515
2512
r
ep
ea
ted
ly
p
ass
ed
th
e
co
n
v
o
l
u
tio
n
al
la
y
er
is
ex
tr
ac
ted
o
n
ly
f
r
o
m
th
e
f
ea
tu
r
e,
an
d
it
ch
an
g
e
s
th
e
i
m
a
g
e
in
t
o
o
n
e
d
i
m
en
s
io
n
a
n
d
d
is
ti
n
g
u
i
s
h
es th
e
i
m
a
g
e
in
o
r
d
er
to
tr
an
s
f
er
th
e
ex
tr
ac
ted
f
ea
t
u
r
e
to
th
e
p
r
ec
o
m
b
in
ed
la
y
er
.
4.
SYST
E
M
I
M
P
L
E
M
E
NT
A
T
I
O
N
T
o
im
p
le
m
e
n
t
t
h
e
s
y
s
te
m
,
t
h
e
W
in
d
o
w
s
1
0
P
r
o
o
p
er
ati
n
g
s
y
s
te
m
w
as
i
n
s
talled
o
n
a
co
m
p
u
ter
eq
u
ip
p
ed
w
it
h
an
i5
-
6
5
0
0
p
r
o
ce
s
s
o
r
,
1
6
G
m
e
m
o
r
y
,
a
n
d
a
GeFo
r
ce
GT
X
1
0
6
0
g
r
ap
h
ics
ca
r
d
.
W
e
also
in
s
ta
lled
C
UD
A
1
0
.
0
to
u
s
e
th
e
GP
U
an
d
i
m
p
le
m
e
n
ted
th
e
s
y
s
te
m
u
s
i
n
g
P
y
th
o
n
3
.
5
in
J
u
p
y
ter
No
teb
o
o
k
.
4
.
1
.
I
m
ple
m
e
nt
pre
pro
ce
s
s
T
h
e
cr
o
p
r
eg
io
n
o
f
I
n
ter
est
is
r
esp
o
n
s
ib
le
f
o
r
f
i
n
d
in
g
k
e
y
p
ar
ts
o
f
th
e
i
m
ag
e
b
e
f
o
r
e
th
e
g
en
er
ati
v
e
h
o
s
tile
n
e
u
r
al
n
e
t
w
o
r
k
p
r
o
ce
ed
s
.
First,
w
e
s
ea
r
ch
t
h
e
e
n
tire
p
ix
el
to
f
i
n
d
th
e
a
v
er
ag
e
v
al
u
e,
an
d
t
h
e
n
s
ea
r
c
h
th
r
o
u
g
h
th
e
p
ix
el
v
al
u
es
o
f
t
h
e
i
m
a
g
e
a
s
d
esi
g
n
ed
.
Af
ter
s
e
ar
ch
in
g
,
it
p
r
in
ts
an
d
s
av
e
s
o
n
l
y
t
h
e
i
m
a
g
e
i
n
s
id
e
th
e
n
e
w
l
y
cr
ea
ted
m
a
s
k
.
F
ig
u
r
e
5
s
h
o
w
s
th
e
r
es
u
lt
s
o
f
t
h
e
r
eg
io
n
o
f
i
n
ter
es
t
.
I
n
F
ig
u
r
e
5
,
f
r
o
m
lef
t
to
r
ig
h
t,
th
e
o
r
ig
in
al
,
cr
o
p
o
f
in
ter
est
ex
tr
ac
tio
n
1
ti
m
e
an
d
in
ter
est
r
eg
io
n
ex
tr
ac
tio
n
2
t
i
m
e
s
.
As
ca
n
b
e
s
ee
n
f
r
o
m
th
e
F
i
g
u
r
e
5
,
it
is
n
ec
ess
ar
y
to
p
r
o
ce
ed
t
w
o
ti
m
es
in
o
r
d
er
to
s
ee
t
h
e
ex
tr
ac
tio
n
o
f
t
h
e
r
eg
io
n
o
f
in
ter
est
.
T
h
is
o
cc
u
r
s
w
h
e
n
th
e
av
er
ag
e
v
a
lu
e
is
ca
lc
u
lated
.
T
h
e
f
ir
s
t
ti
m
e,
t
h
e
r
eg
io
n
o
f
i
n
ter
est
is
e
x
tr
ac
ted
b
y
ca
lc
u
lati
n
g
th
e
a
v
er
ag
e
v
al
u
e
o
f
th
e
b
ac
k
g
r
o
u
n
d
,
th
e
s
h
a
d
o
w
,
an
d
th
e
leav
e
s
.
Get
it.
I
n
ad
d
itio
n
,
th
e
s
h
ad
o
w
is
ca
u
s
ed
in
t
h
e
p
ictu
r
e,
b
u
t
th
is
p
r
o
b
lem
o
cc
u
r
s
w
h
e
n
th
e
f
o
cu
s
is
b
lu
r
r
ed
at
th
e
ed
g
e
o
f
t
h
e
lea
f
.
T
h
er
ef
o
r
e,
in
th
is
p
ap
er
,
w
e
i
m
p
le
m
en
ted
th
e
r
eg
io
n
o
f
i
n
ter
est
t
w
ice
u
s
i
n
g
th
e
cr
o
p
r
eg
io
n
o
f
i
n
ter
est
to
co
n
ce
n
tr
at
e
th
e
r
eg
io
n
o
f
in
ter
e
s
t.
Fi
g
u
r
e
6
s
h
o
w
s
th
e
p
r
ep
r
o
ce
s
s
ed
i
m
a
g
e.
I
n
F
ig
u
r
e
6
,
Fro
m
t
h
e
lef
t,
t
h
e
o
r
ig
in
al,
th
e
ex
tr
ac
ted
r
eg
io
n
o
f
in
ter
est,
a
n
d
th
e
co
n
tr
asted
i
m
a
g
e.
I
n
th
e
la
s
t
i
m
ag
e,
y
o
u
ca
n
s
ee
t
h
at
t
h
e
i
m
a
g
e
o
f
t
h
e
a
f
f
ec
ted
ar
ea
is
m
o
r
e
i
n
te
n
s
e
i
n
co
lo
r
.
T
h
e
i
m
a
g
e
is
th
e
n
g
en
er
ated
t
h
r
o
u
g
h
a
g
en
er
ati
v
e
h
o
s
tile
n
e
u
r
al
n
et
w
o
r
k
.
T
h
e
i
m
ag
e
y
o
u
cr
ea
te
i
s
b
ase
d
o
n
an
u
n
i
n
f
ec
ted
i
m
a
g
e,
w
h
ich
allo
w
s
y
o
u
to
cr
ea
te
an
i
m
a
g
e
s
i
m
ilar
to
th
e
d
is
ea
s
ed
o
n
e
f
r
o
m
th
e
o
r
i
g
i
n
al.
T
h
e
g
e
n
er
ated
i
m
a
g
e
is
s
av
ed
o
r
d
is
ca
r
d
ed
ac
co
r
d
in
g
to
its
q
u
alit
y
a
n
d
is
r
ep
ea
ted
u
n
til
th
e
i
m
a
g
e
r
ea
ch
e
s
a
ce
r
tain
n
u
m
b
er
.
W
h
en
m
o
r
e
th
a
n
a
ce
r
ta
in
n
u
m
b
er
o
f
i
m
ag
e
s
ar
e
cr
ea
ted
,
th
e
g
e
n
er
atio
n
i
s
s
to
p
p
ed
an
d
th
e
clas
s
i
f
icatio
n
m
o
d
el
i
s
u
s
ed
to
tr
ain
th
e
cla
s
s
if
icatio
n
m
o
d
el.
Fig
u
r
e
5
.
C
r
o
p
p
ed
r
eg
io
n
o
f
in
ter
est i
m
a
g
es
Fig
u
r
e
6
.
P
r
ep
r
o
ce
s
s
ed
im
a
g
es
4
.
2
.
I
m
ple
m
e
nt
G
A
N
Gen
etic
ad
v
er
s
ar
ial
n
et
w
o
r
k
s
p
lay
a
r
o
le
i
n
lear
n
i
n
g
p
r
ep
r
o
ce
s
s
ed
d
ata
a
n
d
g
e
n
er
ati
n
g
n
e
w
i
m
a
g
e
s
.
T
h
e
g
en
er
ato
r
n
et
w
o
r
k
h
a
s
f
o
u
r
co
n
v
o
lu
t
io
n
la
y
er
s
to
n
o
r
m
alize
th
e
o
u
tp
u
t
an
d
ac
tiv
ati
o
n
v
alu
e
s
v
ia
b
atch
n
o
r
m
aliza
t
io
n
.
T
h
e
d
is
cr
im
i
n
ato
r
is
im
p
le
m
e
n
ted
w
it
h
th
r
ee
co
n
v
o
lu
tio
n
a
l
la
y
er
s
an
d
o
n
e
f
u
ll
y
-
co
n
n
ec
ted
la
y
er
,
an
d
s
i
g
m
o
id
is
u
s
ed
as
an
ac
ti
v
atio
n
f
u
n
ct
io
n
to
d
is
cr
i
m
i
n
ate
i
m
a
g
es
o
f
v
ar
io
u
s
cl
a
s
s
es.
F
ig
u
r
e
7
is
a
n
i
m
a
g
e
cr
ea
ted
b
y
ap
p
ly
i
n
g
t
h
e
d
is
ea
s
e
ca
lled
B
la
ck
R
o
t
to
th
e
b
lu
eb
er
r
y
leav
e
s
.
I
n
F
ig
u
r
e
7
,
th
e
b
lack
r
o
t
leav
es a
r
e
w
ell
f
o
r
m
ed
.
I
n
ad
d
itio
n
,
b
y
li
m
iti
n
g
th
e
r
e
g
io
n
o
f
in
ter
est d
u
r
i
n
g
th
e
p
r
etr
ea
t
m
e
n
t p
r
o
ce
s
s
it c
an
b
e
s
ee
n
t
h
at
t
h
e
lo
ca
tio
n
o
f
t
h
e
d
is
ea
s
e
is
p
r
o
p
er
ly
s
ea
ted
o
n
th
e
leav
es.
Fig
u
r
e
7
.
Gen
er
ated
b
lu
eb
er
r
y
b
lack
r
o
t i
m
ag
e
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
-
8708
F
r
u
it tr
ee
d
is
ea
s
e
c
la
s
s
ifica
tio
n
s
ystem
u
s
in
g
g
en
era
tive
a
d
v
ers
a
r
ia
l n
etw
o
r
k
s
(
C
h
a
n
g
s
u
K
im
)
2513
4
.
3
.
I
m
ple
m
e
nt
pro
po
s
ed
s
y
s
t
e
m
W
e
im
p
le
m
en
ted
an
d
co
m
p
ar
ed
th
e
d
esig
n
ed
s
y
s
te
m
w
it
h
th
e
ex
is
t
in
g
s
y
s
te
m
.
Fig
u
r
e
8
s
h
o
w
s
t
h
e
class
i
f
icatio
n
p
er
ep
o
ch
w
h
en
s
o
r
tin
g
w
it
h
o
u
t
p
r
ep
r
o
ce
s
s
in
g
(
s
y
s
te
m
1
)
an
d
o
v
er
s
a
m
p
lin
g
b
y
in
cr
ea
s
in
g
d
ata
th
r
o
u
g
h
r
o
tatio
n
(
s
y
s
te
m
2
)
.
T
h
is
g
r
ap
h
s
h
o
w
s
t
h
e
ac
cu
r
ac
y
co
m
p
ar
is
o
n
.
T
h
e
ax
is
,
ep
o
ch
,
r
ep
r
esen
t
s
1
0
ti
m
es.
I
n
Fi
g
u
r
e
s
8
a
n
d
9
,
s
y
s
te
m
1
to
o
k
a
lo
n
g
t
i
m
e
to
m
a
x
i
m
ize
ac
cu
r
ac
y
d
u
e
to
lack
o
f
d
ata.
S
y
s
te
m
2
w
a
s
m
a
x
i
m
ized
in
t
h
e
ep
o
ch
s
i
m
i
lar
to
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
,
b
u
t
t
h
e
ac
cu
r
ac
y
w
as
co
n
f
ir
m
ed
to
b
e
lo
w
.
I
n
co
n
tr
ast,
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
r
ea
ch
ed
its
m
a
x
i
m
u
m
at
a
f
aster
r
ate
th
an
s
y
s
te
m
1
a
n
d
s
h
o
w
ed
an
a
v
er
ag
e
2
% a
cc
u
r
ac
y
i
m
p
r
o
v
e
m
e
n
t
o
v
er
s
y
s
te
m
2
.
T
ab
le
1
s
h
o
w
s
t
h
e
co
m
p
ar
is
o
n
o
f
ac
c
u
r
ac
y
w
h
e
n
clas
s
i
f
ied
u
s
i
n
g
t
h
e
test
s
et.
E
x
p
er
i
m
en
tal
r
esu
l
ts
s
h
o
w
th
at
s
y
s
te
m
1
h
as
a
7
7
.
9
9
%
p
r
o
b
ab
ilit
y
o
f
lack
in
g
a
d
ata
s
et,
w
h
ile
s
y
s
te
m
2
h
as
a
9
1
.
1
4
%
p
r
o
b
ab
ilit
y
th
at
s
o
m
e
d
ata
s
ets
ar
e
s
m
all.
On
t
h
e
o
th
er
h
a
n
d
,
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
i
n
cr
ea
s
ed
th
e
a
m
o
u
n
t
o
f
d
ata
u
s
ed
f
o
r
lear
n
in
g
u
s
i
n
g
g
e
n
er
ativ
e
a
n
ta
g
o
n
i
s
tic
n
e
u
r
al
n
et
w
o
r
k
s
,
a
n
d
p
r
o
v
id
ed
th
e
s
a
m
e
a
m
o
u
n
t
o
f
d
ata
f
o
r
ea
ch
class
,
s
h
o
w
i
n
g
h
i
g
h
ac
c
u
r
ac
y
o
f
9
8
.
1
7
%
.
Fig
u
r
e
8
.
Gr
ap
h
co
m
p
ar
i
n
g
ac
cu
r
ac
y
Fig
u
r
e
9
.
Gen
er
ated
b
lu
eb
er
r
y
b
lack
r
o
t i
m
ag
e
s
T
ab
le
1
.
T
h
e
n
u
m
b
er
o
f
p
r
o
d
u
ct
r
ev
ie
w
s
co
llected
b
y
p
r
o
d
u
ct
attr
ib
u
te
S
y
st
e
m
A
c
c
1
A
c
c
2
A
c
c
3
A
c
c
4
A
c
c
5
A
v
g
A
c
c
S
y
st
e
m1
7
6
.
8
5
7
8
.
3
3
7
8
.
1
2
7
7
.
3
0
7
9
.
3
9
7
7
.
9
9
S
y
st
e
m2
8
9
.
6
7
9
1
.
3
3
9
0
.
6
5
9
2
.
4
7
9
1
.
6
1
9
1
.
1
4
P
r
o
p
o
se
d
sy
st
e
m
9
8
.
0
8
9
7
.
2
9
9
8
.
7
8
9
8
.
5
4
9
8
.
1
7
9
8
.
1
7
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
3
,
J
u
n
e
2021
:
2
5
0
8
-
2515
2514
5.
CO
NCLU
SI
O
N
S
m
ar
t
f
ar
m
r
e
f
er
s
to
a
f
ar
m
th
at
ca
n
m
ai
n
tai
n
an
d
m
an
a
g
e
cr
o
p
an
d
liv
esto
ck
g
r
o
w
t
h
co
n
d
itio
n
s
ap
p
r
o
p
r
iately
r
e
m
o
tel
y
a
n
d
au
to
m
atica
ll
y
b
y
u
s
i
n
g
I
C
T
in
a
g
r
icu
l
tu
r
e.
I
n
ad
d
itio
n
,
it
w
a
s
p
o
s
s
ib
le
to
i
m
p
r
o
v
e
th
e
p
r
o
d
u
ctiv
it
y
an
d
q
u
al
it
y
o
f
ag
r
icu
ltu
r
al
p
r
o
d
u
cts
b
y
cr
ea
tin
g
an
o
p
ti
m
al
g
r
o
w
th
e
n
v
ir
o
n
m
e
n
t
b
ased
o
n
d
ata
o
n
cr
o
p
g
r
o
w
t
h
i
n
f
o
r
m
at
io
n
an
d
e
n
v
ir
o
n
m
en
ta
l
in
f
o
r
m
atio
n
.
Ho
w
ev
er
,
cr
o
p
s
to
w
h
i
ch
s
m
ar
t
f
ar
m
s
ar
e
cu
r
r
en
tl
y
ap
p
lied
ar
e
li
m
ited
to
h
o
r
ticu
ltu
r
e
o
r
h
o
u
s
e
cr
o
p
s
,
an
d
th
er
e
is
a
p
r
o
b
lem
i
n
th
at
a
lar
g
e
co
s
t
is
r
eq
u
ir
ed
to
in
tr
o
d
u
ce
s
m
ar
t
f
a
r
m
s
.
I
n
ad
d
itio
n
,
t
h
e
p
r
o
v
id
ed
d
ata
m
a
y
s
h
o
w
a
b
i
g
d
if
f
er
e
n
ce
f
r
o
m
th
e
ac
tu
al
d
ata
,
an
d
s
in
ce
th
e
a
m
o
u
n
t
o
f
d
ata
is
d
if
f
er
en
t,
th
er
e
is
a
p
r
o
b
lem
th
a
t
th
e
d
ata
i
m
b
alan
c
e
p
r
o
b
lem
m
u
s
t
b
e
s
o
lv
ed
in
o
r
d
er
to
lear
n
.
I
n
o
r
d
er
to
s
o
lv
e
th
is
p
r
o
b
lem
,
T
h
is
p
ap
er
p
r
o
p
o
s
es
a
m
e
th
o
d
th
at
lear
n
s
th
e
d
ata
th
r
o
u
g
h
g
e
n
etic
ad
v
er
s
ar
ial
n
et
w
o
r
k
s
a
n
d
g
e
n
er
ates
d
ata
s
i
m
ilar
to
t
h
e
ac
t
u
al
d
ata
b
y
u
s
i
n
g
th
e
in
p
u
t
d
at
a.
I
n
o
r
d
er
to
r
ed
u
ce
th
e
r
eso
u
r
ce
s
u
s
ed
w
h
e
n
lear
n
in
g
d
ata,
th
is
s
y
s
te
m
p
r
o
p
o
s
ed
a
p
r
ep
r
o
ce
s
s
in
g
m
et
h
o
d
to
r
e
d
u
ce
th
e
a
m
o
u
n
t
o
f
co
m
p
u
tatio
n
an
d
s
p
ee
d
it
u
p
.
W
e
th
en
u
s
ed
g
e
n
etic
ad
v
er
s
a
r
ial
n
et
w
o
r
k
s
to
i
n
cr
ea
s
e
t
h
e
a
m
o
u
n
t
o
f
d
ata
u
s
ed
f
o
r
lear
n
in
g
an
d
to
r
eso
lv
e
d
ata
i
m
b
alan
ce
s
b
y
class
.
T
h
is
en
ab
led
th
e
m
to
g
en
er
ate
d
ata
f
o
r
s
m
ar
t
f
ar
m
s
ac
r
o
s
s
a
w
id
e
r
an
g
e
o
f
cr
o
p
s
,
in
cr
ea
s
in
g
th
e
ac
cu
r
ac
y
o
f
lear
n
in
g
b
y
9
8
p
er
ce
n
t.
Fu
tu
r
e
r
esear
ch
s
h
o
u
ld
g
en
er
a
te
d
ata
th
at
ca
n
b
e
u
s
ed
n
o
t
o
n
l
y
in
th
e
clas
s
i
f
icatio
n
s
y
s
te
m
u
s
ed
in
th
is
p
ap
er
b
u
t
also
in
lo
g
is
tic
s
an
d
au
to
m
at
io
n
s
y
s
te
m
s
.
A
s
a
r
es
u
lt,
it is
co
n
s
id
er
ed
th
at
v
ar
io
u
s
s
m
ar
t
f
ar
m
s
ca
n
b
e
co
n
s
tr
u
cte
d
.
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
is
w
o
r
k
w
a
s
s
u
p
p
o
r
ted
b
y
th
e
r
es
ea
r
ch
g
r
an
t o
f
P
aiC
h
ar
Un
i
v
er
s
it
y
in
2
0
2
1
.
RE
F
E
R
E
NC
E
S
[1
]
N.
S
u
m
a
,
S
.
R.
S
a
m
so
n
,
S
.
S
a
ra
n
y
a
,
G
.
S
h
a
n
m
u
g
a
p
ri
y
a
,
a
n
d
R.
S
u
b
h
a
sh
r
i,
"
IOT
Ba
se
d
S
m
a
rt
A
g
ricu
lt
u
re
M
o
n
i
to
ri
n
g
S
y
ste
m
,
"
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
n
Rec
e
n
t
a
n
d
In
n
o
v
a
ti
o
n
T
re
n
d
s
i
n
C
o
mp
u
ti
n
g
a
n
d
C
o
mm
u
n
ic
a
ti
o
n
,
v
o
l.
5
,
p
p
.
9
9
-
1
0
9
,
F
e
b
.
2
0
1
7
.
[2
]
C.
J.
Kim
,
J.
H.
J
e
o
n
g
,
C.
W
.
J
o
,
J.
K.
Yo
o
,
“
A
P
e
rf
o
r
m
a
n
c
e
E
v
a
lu
a
ti
o
n
A
n
a
l
y
sis
o
f
P
ro
d
u
c
t
R
e
c
o
m
m
e
n
d
a
ti
o
n
T
e
c
h
n
iq
u
e
s
,
”
J
o
u
rn
a
l
o
f
K
n
o
wle
d
g
e
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
lo
g
y
a
n
d
S
y
ste
m,
v
o
l.
1
4
,
n
o
.
5
,
p
p
.
5
1
5
-
5
2
5
,
Oc
t.
2
0
1
9
.
[3
]
S
.
H.
L
e
e
a
n
d
J
.
Y.
Ba
e
,
“
P
r
e
d
ictio
n
Cr
o
p
P
ro
d
u
c
ti
o
n
f
o
r
Ag
ricu
lt
u
ra
l
C
o
n
su
lt
a
ti
o
n
S
e
rv
ice
,
"
J
o
ru
n
a
l
o
f
In
fo
rm
a
t
io
n
a
n
d
Co
mm
u
n
ica
ti
o
n
Co
n
v
e
rg
e
n
c
e
En
g
in
e
e
rin
g
,
v
o
l
.
1
7
,
n
o
.
1
,
p
p
.
8
-
1
3
,
M
a
r.
2
0
1
9
.
[4
]
H.
S
.
Kim
a
n
d
H.
S
.
L
e
e
,
"
Ge
n
e
ra
ti
v
e
A
d
v
e
rs
a
rial
Ne
t
w
o
rk
s
b
a
s
e
d
Da
ta
G
e
n
e
ra
ti
o
n
F
ra
m
e
w
o
rk
fo
r
Ov
e
rc
o
m
in
g
Im
b
a
lan
c
e
d
M
a
n
u
f
a
c
tu
rin
g
P
ro
c
e
ss
Da
ta,"
J
o
u
rn
a
l
o
f
Ko
re
a
n
I
n
stit
u
te
o
f
In
tel
li
g
e
n
t
S
y
ste
ms
,
v
o
l
.
2
9
,
n
o
.
1
,
p
p
.
1
-
8
,
F
e
b
.
2
0
1
9
.
[5
]
I.
J.
G
o
o
d
f
e
ll
o
w
,
e
t
a
l.
,
“
G
e
n
e
ra
ti
v
e
A
d
v
e
rsa
ri
a
l
Ne
t
w
o
rk
s
,
”
Co
rn
e
ll
Un
ive
rs
it
y
,
2
0
1
6
.
[6
]
P
.
H.
Hu
y
n
h
,
V.
H.
Ng
u
y
e
n
,
a
n
d
T
.
N.
Do
,
“
En
h
a
n
c
i
n
g
G
e
n
e
Ex
p
re
ss
io
n
Clas
sif
ica
ti
o
n
o
f
S
u
p
p
o
rt
V
e
c
to
r
M
a
c
h
in
e
s
w
it
h
G
e
n
e
ra
ti
v
e
A
d
v
e
rsa
rial
Ne
t
w
o
rk
s,"
J
o
ru
n
a
l
o
f
In
fo
rm
a
t
io
n
a
n
d
Co
mm
u
n
ica
ti
o
n
C
o
n
v
e
rg
e
n
c
e
En
g
i
n
e
e
rin
g
,
v
o
l
.
1
7
,
n
o
.
1
,
p
p
.
14
-
2
0
,
M
a
r.
2
0
1
9
.
[7
]
W
.
W
a
n
a
n
d
H.
J.
L
e
e
,
“
P
e
rc
e
p
tu
a
l
G
e
n
e
ra
ti
v
e
A
d
v
e
rsa
rial
Ne
tw
o
rk
f
o
r
S
in
g
le
Im
a
g
e
De
-
S
n
o
w
in
g
,
”
KT
S
DE,
v
o
l.
8
,
n
o
.
1
0
,
p
p
.
4
0
3
-
4
1
0
,
A
u
g
.
2
0
1
9
.
[8
]
G
.
Y.
He
o
,
"
ROI
Ex
trac
ti
o
n
f
o
r
A
u
to
m
a
ti
c
P
lac
a
rd
Re
c
o
g
n
it
i
o
n
,
"
J
o
u
rn
a
l
o
f
th
e
K
o
re
a
In
st
it
u
te
o
f
In
fo
rm
a
t
io
n
a
n
d
Co
mm
u
n
ica
ti
o
n
E
n
g
i
n
e
e
rin
g
,
v
o
l.
2
3
,
n
o
.
4
,
p
p
.
3
7
4
-
3
8
0
,
A
p
r.
2
0
1
9
.
[9
]
K.
J.
Ch
a
e
,
Y.
R.
L
e
e
,
Y.
J.
Ch
o
,
a
n
d
J.
H
.
P
a
rk
,
“
De
v
e
lo
p
m
e
n
t
o
f
a
G
a
n
g
w
o
n
P
ro
v
i
n
c
e
F
o
re
st
F
ire
P
re
d
ictio
n
M
o
d
e
l
u
si
n
g
M
a
c
h
in
e
L
e
a
rn
in
g
a
n
d
S
a
m
p
li
n
g
,
”
T
h
e
K
o
e
a
J
o
u
r
n
a
l
o
f
Bi
g
d
a
t
a
,
v
o
l.
3
,
n
o
.
2
,
p
p
.
7
1
-
7
8
,
F
e
b
.
2
0
1
8
,
d
o
i:
1
0
.
3
6
4
9
8
/k
b
ig
d
t.
2
0
1
8
.
3
.
2
.
7
1
.
[1
0
]
G
.
Wan
g
a
n
d
S
.
Y.
S
h
in
,
“
A
n
Im
p
ro
v
e
d
T
e
x
t
Clas
si
f
ica
ti
o
n
M
e
th
o
d
f
o
r
S
e
n
ti
m
e
n
t
Clas
sif
ica
ti
o
n
,
"
J
o
ru
n
a
l
o
f
In
fo
rm
a
t
io
n
a
n
d
Co
mm
u
n
ica
ti
o
n
Co
n
v
e
rg
e
n
c
e
En
g
in
e
e
rin
g
,
v
o
l
.
1
7
,
n
o
.
1
,
p
p
.
41
-
4
8
,
M
a
r.
2
0
1
9
.
[1
1
]
Z.
Zh
u
,
D
.
L
ian
g
,
S
.
Zh
a
n
g
,
X
,
Hu
a
n
g
,
B.
L
i,
a
n
d
S
.
Hu
,
“
T
ra
ff
ic
-
S
ig
n
De
tec
ti
o
n
a
n
d
Clas
sif
ica
ti
o
n
i
n
t
h
e
W
il
d
,
”
Pro
c
e
e
d
in
g
s o
f
th
e
2
0
1
6
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
mp
u
ter
Vi
si
o
n
a
n
d
Pa
tt
e
rn
Rec
o
g
n
it
i
o
n
,
2
0
1
6
,
p
p
.
2
1
1
0
-
2
1
1
8
,
d
o
i
:
1
0
.
1
1
0
9
/C
VP
R
.
2
0
1
6
.
2
3
2
.
[1
2
]
J.
H.
S
h
in
,
K.
H.
C
h
u
n
g
,
a
n
d
K
.
H.
Ch
o
i,
“
De
stru
c
ti
v
e
T
e
st
o
f
a
BL
DC
M
o
to
r
C
o
n
tr
o
ll
e
r
Util
izin
g
a
M
o
d
if
ied
Clas
sif
ic
a
ti
o
n
T
re
e
M
e
th
o
d
,
”
KT
S
DE,
v
o
l.
3
,
n
o
.
6
,
p
p
.
2
0
1
-
2
1
4
,
M
a
r.
2
0
1
4
.
[1
3
]
S
.
T
.
Oh
a
n
d
B.
H.
Ju
n
,
"
Co
n
to
u
r
Ex
trac
ti
o
n
M
e
th
o
d
u
sin
g
p
-
S
n
a
k
e
w
it
h
P
ro
t
o
ty
p
e
En
e
rg
y
,
"
J
o
u
rn
a
l
o
f
th
e
In
stit
u
te
o
f
El
e
c
tro
n
ics
a
n
d
I
n
fo
rm
a
ti
o
n
E
n
g
i
n
e
e
rs
,
v
o
l.
5
1
,
n
o
.
4
,
p
p
.
1
0
1
-
1
0
9
,
A
p
r.
2
0
1
4
.
[1
4
]
J.
H.
Oh
a
n
d
J.
G
.
P
a
ik
,
"
Clu
ste
rin
g
-
b
a
se
d
Un
d
e
rsa
m
p
li
n
g
f
o
r
I
m
b
a
lan
c
e
d
Da
ta
Clas
si
f
ic
a
ti
o
n
,
"
J
o
u
rn
a
l
o
f
K
o
re
a
n
In
stit
u
te
o
f
I
n
d
u
stria
l
En
g
i
n
e
e
rs
,
p
p
.
1
9
1
0
-
1
9
1
6
,
No
v
.
2
0
1
7
.
[1
5
]
S
.
R.
S
a
l
u
ti
,
“
A
su
rv
e
y
o
f
b
ig
d
a
ta
a
n
d
m
a
c
h
in
e
lea
rn
in
g
,
”
I
n
t
e
rn
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
En
g
i
n
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
1
0
,
n
o
.
1
,
p
p
.
5
7
5
-
5
8
0
,
F
e
b
.
2
0
2
0
.
[1
6
]
H.
J.
S
h
in
a
n
d
C.
H
.
Oh
,
"
M
a
c
h
in
e
L
e
a
rn
in
g
b
a
se
d
o
n
A
p
p
r
o
a
c
h
f
o
r
Clas
sif
ica
ti
o
n
o
f
A
b
n
o
rm
a
l
Da
ta
in
S
h
o
p
-
f
lo
o
r,
"
J
o
u
rn
a
l
o
f
th
e
Ko
re
a
In
sti
tu
te
o
f
I
n
fo
rm
a
t
io
n
a
n
d
Co
mm
u
n
ica
t
io
n
E
n
g
in
e
e
rin
g
,
v
o
l.
2
1
,
n
o
.
1
1
,
p
p
.
2
0
3
7
-
2
0
4
2
,
No
v
.
2
0
1
7
.
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
-
8708
F
r
u
it tr
ee
d
is
ea
s
e
c
la
s
s
ifica
tio
n
s
ystem
u
s
in
g
g
en
era
tive
a
d
v
ers
a
r
ia
l n
etw
o
r
k
s
(
C
h
a
n
g
s
u
K
im
)
2515
[1
7
]
S
.
T
.
Na
m
,
S
.
Y.
S
h
in
a
n
d
C.
Y.
Jin
,
"
A
Re
c
o
n
stru
c
ti
o
n
o
f
Clas
sif
i
c
a
ti
o
n
f
o
r
Iris
S
p
e
c
ies
Us
in
g
Eu
c
li
d
e
a
n
Dista
n
c
e
Ba
se
d
o
n
a
M
a
c
h
in
e
L
e
a
rn
in
g
,
"
J
o
u
rn
a
l
o
f
th
e
Ko
re
a
In
st
it
u
te
o
f
In
f
o
rm
a
ti
o
n
a
n
d
C
o
mm
u
n
ic
a
ti
o
n
E
n
g
in
e
e
rin
g
,
v
o
l.
2
4
,
n
o
.
2
,
p
p
.
2
2
5
-
2
3
0
,
F
e
b
.
2
0
2
0
.
[1
8
]
D.
Oh
,
B
.
T
.
Oh
a
n
d
J.
S
h
i
n
,
“
Up
-
sa
m
p
li
n
g
m
e
th
o
d
o
f
d
e
p
th
m
a
p
u
sin
g
w
e
ig
h
ted
jo
i
n
t
b
il
a
tera
l
f
il
ter,”
T
h
e
J
o
u
rn
a
l
o
f
Ko
re
a
n
I
n
stit
u
te o
f
C
o
mm
u
n
ica
ti
o
n
s a
n
d
In
f
o
rm
a
ti
o
n
S
c
ien
c
e
s
,
v
o
l.
4
0
,
n
o
.
6
,
p
p
.
1
1
7
5
-
11
8
4
,
Ju
n
.
2
0
1
5
.
[1
9
]
S
.
A
ich
,
S
.
C
h
a
k
ra
b
o
rty
a
n
d
H.
C.
Kim
,
“
Co
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
-
b
a
se
d
m
o
d
e
l
f
o
r
w
e
b
-
b
a
se
d
tex
t
c
las
si
f
ica
ti
o
n
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
i
n
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
9
,
n
o
.
6
,
p
p
.
5
1
8
5
-
5
1
9
1
,
De
c
.
2
0
1
9
.
[2
0
]
X
.
T
.
Ya
n
g
,
J.
S
.
L
e
e
a
n
d
H.
K.
Ju
n
g
,
"
F
a
u
lt
Dia
g
n
o
sis
M
a
n
a
g
e
m
e
n
t
M
o
d
e
l
u
si
n
g
M
a
c
h
i
n
e
L
e
a
rn
in
g
,
"
J
o
ru
n
a
l
o
f
In
fo
rm
a
t
io
n
a
n
d
Co
mm
u
n
ica
ti
o
n
Co
n
v
e
rg
e
n
c
e
En
g
in
e
e
rin
g
,
v
o
l
.
1
7
,
n
o
.
2
,
p
p
.
1
2
8
-
1
3
4
,
J
u
n
.
2
0
1
9
.
[2
1
]
H.
K.
Kim
,
S
.
O.
L
e
e
,
a
n
d
H.
K.
Ju
n
g
,
“
Hu
m
a
n
a
c
ti
v
it
y
re
c
o
g
n
it
io
n
b
y
u
sin
g
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
,
”
In
ter
n
a
t
io
n
a
l
J
o
u
r
n
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
9
,
n
o
.
6
,
p
p
.
5
2
7
0
-
5
2
7
6
,
De
c
.
2
0
1
9
,
d
o
i:
1
0
.
1
1
5
9
1
/
ij
e
c
e
.
v
9
i6
.
p
p
5
2
7
0
-
5
2
7
6
.
[2
2
]
J.
K.
Rh
e
e
,
"
P
re
d
ictio
n
f
o
r
P
e
rio
d
o
n
tal
Dise
a
se
u
sin
g
G
e
n
e
Ex
p
re
ss
io
n
P
r
o
f
il
e
Da
ta
b
a
se
d
o
n
M
a
c
h
in
e
L
e
a
rn
in
g
,
"
J
o
u
rn
a
l
o
f
th
e
K
o
re
a
I
n
stit
u
te
o
f
In
f
o
rm
a
ti
o
n
a
n
d
Co
mm
u
n
ica
t
io
n
E
n
g
i
n
e
e
rin
g
,
v
o
l
.
2
3
,
n
o
.
8
,
p
p
.
9
0
3
-
9
0
8
,
Oc
t.
2
0
1
9
.
[2
3
]
S
.
Kh
a
n
,
K.
V
.
Kir
u
b
a
n
a
n
d
“C
o
m
p
a
rin
g
m
a
c
h
in
e
lea
rn
in
g
a
n
d
e
n
se
m
b
le
lea
rn
in
g
in
th
e
f
ield
o
f
f
o
o
tb
a
ll
,
”
In
ter
n
a
t
io
n
a
l
J
o
u
r
n
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
E
n
g
i
n
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
9
,
n
o
.
5
,
p
p
.
4
3
2
1
-
4
3
2
5
,
Oc
t.
2
0
1
9
,
d
o
i:
1
0
.
1
1
5
9
1
/
ij
e
c
e
.
v
9
i5
.
p
p
4
3
2
1
-
4
3
2
5
.
[2
4
]
A
.
Ra
d
f
o
rd
,
L
.
M
e
tz
a
n
d
S
.
C
h
in
t
a
la,
“
Un
su
p
e
rv
ise
d
Re
p
re
se
n
tatio
n
L
e
a
rn
in
g
w
it
h
De
e
p
Co
n
v
o
lu
ti
o
n
a
l
G
e
n
e
ra
ti
v
e
A
d
v
e
r
sa
rial
Ne
t
w
o
rk
s,”
Co
rn
e
ll
Un
ive
rs
it
y
,
2
0
1
6
.
[2
5
]
S
.
A
ich
,
S
.
C
h
a
k
ra
b
o
rty
a
n
d
H.
C.
Kim
,
“
Co
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
-
b
a
se
d
m
o
d
e
l
f
o
r
w
e
b
-
b
a
se
d
tex
t
c
las
si
f
ica
ti
o
n
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
i
n
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
9
,
n
o
.
6
,
p
p
.
5
1
8
5
-
5
1
9
1
,
De
c
.
2
0
1
9
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
R
S
Ch
a
n
g
s
u
K
i
m
He
r
ec
eiv
ed
th
e
B
.
S.,
M.
S.,
an
d
P
h
.
D.
d
eg
r
ee
s
f
r
o
m
t
h
e
Dep
ar
t
m
en
t
o
f
C
o
m
p
u
ter
E
n
g
i
n
ee
r
in
g
o
f
P
aich
ai
Un
i
v
er
s
i
t
y
,
Ko
r
ea
,
in
1
9
9
6
,
1
9
9
8
,
an
d
2
0
0
2
,
r
esp
ec
tiv
el
y
.
Fro
m
2
0
0
5
to
2
0
1
2
,
h
e
w
o
r
k
ed
f
o
r
th
e
Dep
ar
t
m
en
t
o
f
I
n
ter
n
et
at
C
h
u
n
g
w
o
o
n
U
n
iv
er
s
it
y
as a
p
r
o
f
ess
o
r
.
Si
n
ce
2
0
1
3
,
h
e
h
a
s
wo
r
k
ed
in
t
h
e
Dep
ar
t
m
en
t
o
f
C
o
m
p
u
ter
E
n
g
in
ee
r
i
n
g
at
P
aic
h
ai
U
n
i
v
er
s
it
y
,
w
h
er
e
h
e
n
o
w
w
o
r
k
s
as
a
p
r
o
f
e
s
s
o
r
.
His
cu
r
r
en
t
r
esear
ch
in
ter
e
s
ts
in
c
l
u
d
e
m
u
l
ti
m
ed
ia
d
o
cu
m
en
t
ar
ch
itect
u
r
e
m
o
d
elin
g
,
Data
m
i
n
in
g
,
an
d
t
h
e
s
e
m
a
n
tic
w
eb
.
H
y
e
so
o
Le
e
Sh
e
r
ec
eiv
ed
t
h
e
B
.
S
d
eg
r
ee
in
2
0
1
4
an
d
M.
S
d
eg
r
ee
in
2
0
1
6
f
r
o
m
t
h
e
Dep
ar
t
m
en
t
o
f
C
o
m
p
u
ter
E
n
g
in
ee
r
i
n
g
o
f
P
aich
ai
Un
iv
er
s
it
y
,
Ko
r
ea
.
Fro
m
2
0
1
3
to
2
0
1
4
,
s
h
e
co
m
p
leted
h
er
in
te
r
n
s
h
ip
at
Natio
n
al
Sec
u
r
it
y
R
esear
ch
I
n
s
ti
tu
te
a
n
d
s
h
e
w
o
r
k
s
f
o
r
Su
p
er
co
m
p
u
ti
n
g
I
n
f
r
astr
u
ctu
r
e
C
e
n
ter
o
f
Div
.
o
f
Natio
n
al
Su
p
er
co
m
p
u
ti
n
g
in
Ko
r
ea
I
n
s
tit
u
te
o
f
Scien
ce
an
d
T
ec
h
n
o
lo
g
y
in
f
o
r
m
at
io
n
(
KI
ST
I
)
.
Sh
e
is
in
ch
ar
g
e
o
f
co
n
tr
ac
t
o
f
u
s
a
g
e
an
d
ac
co
u
n
t
m
a
n
a
g
e
m
en
t
f
o
r
s
u
p
er
co
m
p
u
ter
,
an
d
p
r
o
v
id
es
tech
n
ical
s
u
p
p
o
r
t to
a
u
s
er
o
f
s
u
p
er
co
m
p
u
ter
.
H
o
e
k
y
u
n
g
J
u
n
g
He
r
ec
eiv
ed
t
h
e
M.
S.
d
eg
r
ee
i
n
1
9
8
7
an
d
P
h
.
D.
d
eg
r
ee
in
1
9
9
3
f
r
o
m
th
e
Dep
ar
t
m
en
t
o
f
C
o
m
p
u
ter
E
n
g
in
ee
r
i
n
g
o
f
K
w
a
n
g
w
o
o
n
Un
iv
er
s
i
t
y
,
Ko
r
ea
.
Fro
m
1
9
9
4
to
1
9
9
5
,
h
e
w
o
r
k
ed
f
o
r
E
T
R
I
as
a
r
esear
ch
er
.
Si
n
ce
1
9
9
4
,
h
e
h
as
w
o
r
k
ed
in
th
e
Dep
ar
t
m
en
t
o
f
C
o
m
p
u
ter
E
n
g
in
ee
r
in
g
at
P
aich
ai
U
n
iv
er
s
it
y
,
w
h
er
e
h
e
n
o
w
w
o
r
k
s
as
a
p
r
o
f
ess
o
r
.
His
c
u
r
r
en
t
r
esear
ch
i
n
ter
est
s
i
n
clu
d
e
m
u
lti
m
e
d
ia
d
o
cu
m
en
t
ar
ch
i
tectu
r
e
m
o
d
eli
n
g
,
i
n
f
o
r
m
atio
n
p
r
o
ce
s
s
in
g
,
e
m
b
ed
d
ed
s
y
s
te
m
,
m
ac
h
in
e
lear
n
in
g
,
b
i
g
d
ata,
an
d
I
o
T
.
Evaluation Warning : The document was created with Spire.PDF for Python.