I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
15
,
No
.
6
,
Decem
b
er
20
25
,
p
p
.
5
8
2
7
~
5
8
3
6
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
6
.
pp
5
8
2
7
-
5
8
3
6
5827
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Identifica
tion typ
es o
f
pla
nt
using
c
o
nv
o
lutiona
l neural network
Ra
dity
o
H
endra
t
mo
j
o
J
a
t
i N
o
t
o
neg
o
r
o
,
H
us
t
ina
wa
t
y
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
m
a
t
i
o
n
S
y
st
e
m
M
a
n
a
g
e
me
n
t
,
U
n
i
v
e
r
si
t
a
s
G
u
n
a
d
a
r
m
a
,
D
e
p
o
k
,
I
n
d
o
n
e
si
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Sep
2
5
,
2
0
2
4
R
ev
is
ed
J
u
l 1
4
,
2
0
2
5
Acc
ep
ted
Sep
1
4
,
2
0
2
5
Art
if
ic
ial
i
n
tel
l
i
g
e
n
c
e
c
a
n
b
e
im
p
lem
e
n
t
e
d
i
n
f
ie
ld
s
t
h
a
t
re
la
te
d
t
o
e
n
v
ir
o
n
m
e
n
t
a
l
e
d
u
c
a
ti
o
n
b
y
p
r
o
v
i
d
i
n
g
k
n
o
wle
d
g
e
f
o
r
ta
x
o
n
o
m
y
w
h
ic
h
re
c
o
g
n
ize
a
n
d
i
d
e
n
t
if
y
p
la
n
t
s
p
e
c
i
e
s b
a
s
e
d
o
n
it
s
fe
a
t
u
re
s
.
T
h
e
v
a
rie
ty
o
f
p
la
n
t
sp
e
c
ies
t
h
a
t
i
n
h
a
b
i
t
i
n
a
c
e
r
tai
n
a
re
a
a
ll
o
w
s
m
a
n
y
p
la
n
t
s
p
e
c
ies
t
o
b
e
f
o
u
n
d
th
a
t
l
o
o
k
s
imi
lar
s
o
t
h
a
t
d
i
ff
ic
u
l
t
t
o
d
is
ti
n
g
u
is
h
a
n
d
re
c
o
g
n
ize
a
p
a
r
ti
c
u
lar
p
l
a
n
t
.
C
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
r
k
(CNN
)
o
f
te
n
u
se
d
i
n
o
b
jec
t
d
e
tec
t
i
o
n
,
y
o
u
o
n
l
y
l
o
o
k
o
n
c
e
(YO
LO
),
o
n
e
o
f
C
NN
’s
o
b
jec
t
d
e
tec
ti
o
n
s
,
c
o
u
l
d
i
d
e
n
t
if
y
o
b
j
e
c
t
i
n
re
a
l
t
ime
a
n
d
o
b
ta
i
n
e
d
g
o
o
d
p
e
rf
o
rm
a
n
c
e
a
n
d
a
c
c
u
ra
c
y
in
se
v
e
ra
l
re
se
a
r
c
h
e
d
.
H
o
we
v
e
r
,
n
o
st
u
d
ies
h
a
v
e
e
v
e
r
i
d
e
n
ti
fie
d
a
p
la
n
t
fr
o
m
i
t
s
fl
o
we
rs
,
lea
v
e
s
,
a
n
d
f
r
u
i
ts
.
T
h
e
re
f
o
re
,
t
h
e
m
a
i
n
o
b
jec
t
o
f
t
h
i
s
p
a
p
e
r
is
id
e
n
t
i
fie
d
t
y
p
e
s
o
f
p
la
n
t
w
it
h
CNN
(Y
OLO
v
8
).
T
h
e
YO
L
O
v
8
m
o
d
e
l
wi
t
h
0
.
0
1
le
a
r
n
i
n
g
ra
te,
3
2
b
a
tc
h
s
ize
,
s
t
o
c
h
a
s
ti
c
g
r
a
d
ie
n
t
d
e
sc
e
n
t
(S
G
D)
o
p
t
im
ize
r
o
b
ta
i
n
e
d
h
i
g
h
e
s
t
p
re
c
is
i
o
n
o
f
6
9
.
6
2
%
a
n
d
F
1
sc
o
r
e
o
f
6
1
.
2
2
%
,
re
c
a
ll
o
f
5
4
.
7
3
%
,
m
AP
5
0
a
n
d
m
AP
5
0
–
9
0
o
n
t
h
e
tr
a
i
n
i
n
g
d
a
ta
o
f
5
7
.
6
1
%
a
n
d
4
2
.
4
9
%
.
K
ey
w
o
r
d
s
:
C
o
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
I
d
en
tific
atio
n
Plan
t
R
ea
l tim
e
Yo
u
o
n
ly
lo
o
k
o
n
ce
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
:
R
ad
ity
o
Hen
d
r
atm
o
j
o
J
ati
No
to
n
eg
o
r
o
Dep
ar
tm
en
t
o
f
I
n
f
o
r
m
atio
n
Sy
s
tem
Ma
n
ag
em
en
t,
Un
iv
e
r
s
itas
Gu
n
ad
ar
m
a
Kar
tik
a
J
ay
a
G
4
4
,
C
ip
ay
u
n
g
,
J
ak
ar
ta,
I
n
d
o
n
esia
E
m
ail: r
ad
itjati0
6
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Ar
tific
ial
in
tellig
en
ce
(
AI
)
c
an
s
ig
n
if
ican
tly
r
ed
u
ce
h
u
m
a
n
er
r
o
r
s
an
d
s
av
e
tim
e
a
n
d
en
er
g
y
b
y
h
an
d
lin
g
s
im
p
le
task
s
ef
f
icien
tly
.
I
n
th
e
f
ield
o
f
e
n
v
ir
o
n
m
en
tal
ed
u
ca
tio
n
,
AI
ca
n
also
p
lay
a
k
ey
r
o
le
—
p
ar
ticu
lar
ly
in
ta
x
o
n
o
m
y
—
b
y
h
elp
in
g
to
id
e
n
tify
an
d
class
if
y
p
la
n
t
s
p
ec
ies
b
ased
o
n
th
eir
ch
ar
ac
ter
is
tics
.
Giv
en
th
e
wid
e
v
a
r
iety
o
f
p
lan
t
s
p
ec
ies
in
a
s
p
ec
if
ic
r
eg
i
o
n
,
m
an
y
o
f
w
h
ich
a
p
p
ea
r
q
u
ite
s
im
ilar
,
AI
ca
n
ass
is
t
in
ac
cu
r
ately
d
is
tin
g
u
is
h
in
g
b
etwe
en
th
em
.
T
h
is
ca
p
ab
ilit
y
s
u
p
p
o
r
ts
th
e
co
n
s
er
v
atio
n
o
f
en
d
an
g
e
r
ed
s
p
ec
ies
an
d
e
n
h
an
ce
s
o
u
r
u
n
d
er
s
tan
d
i
n
g
o
f
u
n
iq
u
e,
en
d
em
ic
p
lan
ts
th
at
ex
is
t
o
n
ly
i
n
ce
r
tain
ar
ea
s
.
Ad
d
itio
n
ally
,
A
I
ca
n
h
elp
id
en
tify
e
d
ib
le
an
d
n
o
n
-
ed
ib
le
p
la
n
ts
,
as
well
a
s
m
ed
icin
al
p
lan
ts
an
d
ec
o
n
o
m
ically
v
alu
ab
le
f
o
r
est
s
p
ec
ies.
T
h
is
k
n
o
wled
g
e
ca
n
co
n
tr
ib
u
te
to
s
u
s
tain
ab
le
r
eso
u
r
ce
m
an
ag
e
m
en
t
an
d
im
p
r
o
v
e
th
e
co
m
m
er
cial
v
alu
e
o
f
th
ese
p
la
n
ts
.
I
n
ag
r
ic
u
ltu
r
e,
AI
is
u
s
ef
u
l
f
o
r
m
o
n
it
o
r
in
g
cr
o
p
g
r
o
wth
at
ev
e
r
y
s
tag
e,
en
ab
lin
g
m
o
r
e
s
u
s
tain
ab
le
an
d
o
p
tim
ized
h
ar
v
ests
.
Ob
ject
d
etec
tio
n
is
a
n
im
a
g
e
class
if
icatio
n
-
b
ased
task
th
at
r
eq
u
ir
es
b
o
u
n
d
in
g
b
o
x
es
as
m
ar
k
er
s
an
d
id
en
tifie
s
in
p
u
t
im
a
g
es
in
to
a
p
p
r
o
p
r
iate
ca
teg
o
r
ies
[
1
]
.
T
h
e
u
s
e
o
f
o
b
ject
d
etec
tio
n
to
id
e
n
tify
ce
r
tain
p
lan
ts
in
r
ea
l
tim
e
ca
n
m
ak
e
it
ea
s
ier
t
o
r
ec
o
g
n
ize
a
ty
p
e
o
f
p
lan
t.
C
o
n
v
o
lu
ti
o
n
al
n
e
u
r
al
n
etwo
r
k
(
C
NN)
o
f
ten
u
s
ed
to
o
b
ject
d
etec
tio
n
,
th
er
e
a
r
e
t
wo
ty
p
es
o
f
d
etec
tio
n
alg
o
r
ith
m
,
two
s
tag
es
an
d
o
n
e
s
ta
g
e.
T
wo
s
tag
es
h
as
d
ed
icate
d
m
o
d
u
le
f
o
r
g
en
er
a
tin
g
r
e
g
io
n
p
r
o
p
o
s
als,
m
o
d
u
l
e
f
ir
s
tly
id
e
n
tifie
s
a
v
ar
iab
le
n
u
m
b
er
o
f
o
b
ject
p
r
o
p
o
s
als
with
in
an
im
ag
e,
th
en
th
e
m
o
d
u
le
class
if
y
in
g
an
d
lo
ca
lizin
g
th
o
s
e
p
r
o
p
o
s
als
o
n
s
ec
o
n
d
s
tag
e,
lik
e
r
eg
io
n
s
with
C
NN
(
R
C
NN)
,
wh
ile
o
n
e
s
tag
e,
lik
e
y
o
u
o
n
ly
lo
o
k
o
n
ce
(
YOL
O)
,
p
r
o
v
id
es
o
b
ject
class
if
icatio
n
an
d
b
o
u
n
d
in
g
b
o
x
es
d
ir
ec
tl
y
u
s
in
g
a
s
in
g
le
f
ee
d
f
o
r
war
d
f
u
ll
y
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
etwo
r
k
an
d
k
e
y
p
o
in
ts
o
f
d
if
f
er
en
t scale
s
an
d
asp
ec
t r
ati
o
s
to
id
en
tify
o
b
jects.
T
h
is
d
esig
n
o
f
f
e
r
s
ad
v
an
ta
g
es o
v
er
tw
o
-
s
tag
e
d
etec
to
r
s
in
ter
m
s
o
f
r
ea
l
-
tim
e
p
e
r
f
o
r
m
an
c
e
an
d
s
im
p
licity
[
2
]
–
[
4
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
6
,
Decem
b
e
r
20
25
:
5
8
2
7
-
5
8
3
6
5828
YOL
O
m
o
d
els
co
u
ld
im
p
lem
en
ted
in
m
an
y
th
in
g
s
:
r
ea
l
tim
e
d
etec
tio
n
f
o
r
g
i
n
g
er
s
h
o
o
t
an
d
s
ee
d
s
[
5
]
,
d
etec
tio
n
r
o
b
o
ts
f
o
r
p
ic
k
in
g
m
a
n
g
o
s
f
r
u
its
in
r
ea
l
tim
e
[
6
]
an
d
to
m
at
o
es
f
o
r
d
etec
tin
g
r
ip
e
(
m
atu
r
e)
,
u
n
r
ip
ed
(
im
m
atu
r
e
)
,
s
em
i
m
at
u
r
e,
an
d
d
is
ea
s
ed
f
r
u
its
[
7
]
,
[
8
]
,
d
is
ea
s
e
d
etec
tio
n
o
n
to
m
ato
leaf
[
9
]
,
[
1
0
]
,
ju
te
(
C
o
r
ch
o
r
u
s
o
lito
r
iu
s
o
r
C
o
r
ch
o
r
u
s
ca
p
s
u
la
r
is
)
d
is
ea
s
e
f
r
o
m
leav
es
a
n
d
s
tem
a
n
d
p
ests
(
Ju
te
Ha
ir
y
C
a
terp
illa
r
an
d
C
o
mo
p
h
ila
s
a
b
u
lifer
s
)
[
1
1
]
,
tea
leaf
d
is
ea
s
e
(
b
lig
h
t)
an
d
p
est
(
A
p
o
ly
g
u
s
lu
co
r
u
m
)
[
1
2
]
,
d
eter
m
in
e
wh
ic
h
p
la
n
t
is
h
ea
lt
h
y
an
d
h
as
d
is
ea
s
e
o
n
s
o
y
b
ea
n
(
Glyc
in
e
m
a
x
)
,
o
k
r
a
(
A
b
elmo
s
ch
u
s
escu
len
tu
s
)
,
an
d
m
aize
(
Zea
ma
ys
)
leav
es
[
1
3
]
,
wh
ite
g
r
ap
e
f
r
u
it
r
ea
l
tim
e
co
u
n
tin
g
an
d
b
u
n
ch
d
etec
tio
n
f
o
r
g
r
ap
e
y
ield
d
ec
r
ea
s
e
tim
e
esti
m
atio
n
[
1
4
]
,
C
o
u
n
tin
g
leav
es
o
f
Ar
ab
id
o
p
s
is
p
lan
t
(
A
r
a
b
id
o
p
s
is
th
a
lia
n
a
)
[
1
5
]
.
B
ased
o
n
r
esear
ch
b
y
Kh
an
et
a
l.
[
1
6
]
wo
r
k
o
n
r
ea
l
tim
e
wee
d
s
d
ete
ctio
n
in
p
o
tato
(
S
o
la
n
u
m
tu
b
e
r
o
s
u
m
)
cr
o
p
s
u
s
in
g
YOL
Ov
4
-
tin
y
,
th
e
a
d
o
p
ted
m
o
d
el
g
et
4
9
.
4
%
ac
cu
r
ac
y
o
n
v
er
y
lim
ited
d
ataset.
Ab
o
za
r
et
a
l.
[
1
7
]
d
etec
t
th
e
d
am
ag
ed
o
f
th
e
s
u
g
ar
b
ee
t
(
B
eta
vu
lg
a
r
is
)
r
o
o
ts
b
y
m
ec
h
a
n
ical
s
tr
es
s
d
u
r
in
g
h
ar
v
esti
n
g
u
s
in
g
YOL
Ov
4
,
th
e
m
eth
o
d
b
e
ab
le
to
d
etec
t th
e
d
am
ag
e
with
r
ec
all
9
2
%,
p
r
ec
is
io
n
9
4
%,
an
d
F1
s
co
r
e
9
3
% (
b
etter
p
er
f
o
r
m
a
n
ce
)
.
R
esear
ch
f
r
o
m
Yao
et
a
l.
[
1
8
]
d
etec
t
th
e
d
e
f
ec
t
in
k
iwif
r
u
it
u
s
in
g
YOL
Ov
5
,
th
e
m
o
d
el
r
ea
ch
ed
9
4
.
7
%
o
f
m
AP5
0
.
An
o
th
e
r
d
is
ea
s
e
d
ete
cto
r
f
o
r
to
m
ato
f
r
u
it
with
co
m
p
ar
in
g
m
eth
o
d
u
s
in
g
o
b
tain
e
d
m
AP
Fas
ter
R
C
NN
(
8
0
.
8
%),
SS
D
(
7
6
.
7
%),
a
n
d
Y
OL
O
v
er
s
io
n
4
,
5
,
7
,
8
(
8
8
.
4
%,
9
1
.
2
%,
9
1
.
6
%,
9
1
.
9
%)
[
1
9
]
,
th
e
r
ice
an
d
c
o
tto
n
d
is
ea
s
e
u
s
in
g
Fas
t
R
C
NN,
Y
OL
O
v
7
,
YOL
Ov
8
g
et
m
AP
v
alu
es
(
4
9
.
3
3
%,
6
1
.
8
0
%,
6
6
.
4
7
%)
f
o
r
r
ice
an
d
(
7
6
.
8
8
%,
7
8
.
3
6
%,
7
9
.
5
6
%)
f
o
r
co
tto
n
[
2
0
]
.
YOL
O
v
er
s
io
n
8
(
YOL
Ov
8
)
was
ch
o
s
en
b
ec
au
s
e
it
h
as
ad
v
an
tag
es:
n
o
t
u
s
in
g
an
ch
o
r
b
o
x
es,
r
ed
u
cin
g
t
h
e
n
u
m
b
er
o
f
p
r
ed
ic
tio
n
b
o
x
es,
an
d
ac
ce
ler
atin
g
n
o
n
m
ax
im
u
m
im
p
r
ess
io
n
[
2
1
]
.
T
h
is
v
er
s
io
n
o
f
th
e
YOL
O
m
o
d
el
is
co
n
s
id
er
ed
m
o
r
e
ef
f
ec
tiv
e
b
ec
a
u
s
e
it
h
as
a
n
u
p
d
ate
d
f
ea
t
u
r
e
m
ap
a
n
d
c
o
n
v
o
lu
tio
n
al
n
etwo
r
k
[
2
2
]
,
u
s
es
a
task
alig
n
ed
ass
ig
n
er
th
at
c
o
m
p
u
tes
a
task
alig
n
m
en
t
task
m
atr
ic
u
s
in
g
r
eg
r
ess
io
n
co
o
r
d
in
ates
a
n
d
th
e
class
if
icatio
n
s
co
r
es,
co
m
b
in
e
with
th
e
v
al
u
e
o
f
in
ter
s
ec
tio
n
o
v
er
u
n
i
o
n
(
I
o
U)
,
allo
ws
lo
ca
lizatio
n
an
d
class
if
icatio
n
o
p
tim
izatio
n
s
i
m
u
ltan
eo
u
s
ly
wh
ile
s
u
p
p
r
ess
in
g
p
r
e
d
ictio
n
b
o
x
es
wh
ich
h
a
v
e
lo
w
q
u
ality
[
2
3
]
.
Ma
n
y
r
esear
c
h
in
id
en
tif
y
in
g
p
lan
t
u
s
in
g
YOL
O
m
o
d
el
im
p
lem
en
ted
f
o
r
d
etec
t
an
d
i
d
en
ti
f
y
th
e
d
is
ea
s
es
an
d
cr
o
p
d
a
m
ag
es in
r
ea
l tim
e,
b
u
t
im
p
lem
en
tatio
n
YOL
O
m
o
d
e
l w
ith
v
er
s
io
n
8
f
o
r
id
e
n
tify
in
g
p
lan
t b
ased
o
n
th
e
f
lo
wer
s
,
leav
es,
an
d
f
r
u
its
h
as
n
o
t
b
ee
n
ca
r
r
ied
o
u
t
in
th
e
r
ec
o
g
n
itio
n
o
f
an
o
b
ject.
T
h
e
o
b
j
ec
tiv
e
o
f
t
h
is
s
tu
d
y
is
id
en
tify
in
g
ty
p
es o
f
p
la
n
t w
ith
YOL
Ov
8
m
o
d
el
b
ased
o
n
its
f
ea
tu
r
es.
2.
M
E
T
H
O
D
I
n
th
is
r
esear
ch
,
th
e
p
r
o
ce
s
s
o
f
cr
ea
tin
g
t
h
e
m
o
d
el
f
o
r
p
la
n
t
s
p
ec
ies
id
en
tific
atio
n
in
v
o
l
v
es
s
ev
er
al
k
ey
s
tep
s
,
as
o
u
tlin
ed
in
th
e
m
ain
f
r
am
ewo
r
k
s
h
o
wn
in
Fig
u
r
e
1
.
T
h
e
f
ir
s
t
s
tep
is
to
co
l
lect
a
d
iv
er
s
e
s
et
o
f
im
ag
e
d
ata
t
h
at
ac
cu
r
atel
y
r
e
p
r
esen
ts
th
e
p
lan
t
s
p
ec
ies
u
n
d
e
r
s
tu
d
y
.
On
ce
th
e
d
ata
is
g
ath
er
ed
,
th
e
n
ex
t
s
tag
e
in
v
o
lv
es
p
r
ep
r
o
ce
s
s
in
g
,
wh
ic
h
in
clu
d
es
r
esizin
g
th
e
im
a
g
e
s
to
a
u
n
i
f
o
r
m
s
ize,
an
n
o
tatin
g
th
em
t
o
lab
el
th
e
p
lan
t
s
p
ec
ies,
an
d
ap
p
ly
in
g
d
ata
au
g
m
en
tatio
n
tech
n
iq
u
es
to
in
cr
ea
s
e
th
e
d
ataset
's
v
ar
iab
ilit
y
.
Af
ter
p
r
ep
r
o
ce
s
s
in
g
,
th
e
d
ata
is
s
p
lit
in
to
th
r
ee
d
is
tin
ct
s
ets:
a
tr
ain
in
g
s
et,
a
v
alid
atio
n
s
et,
an
d
a
test
s
et.
T
h
e
tr
ain
in
g
an
d
v
alid
atio
n
s
ets
ar
e
u
s
ed
to
tr
ain
an
d
f
in
e
-
tu
n
e
t
h
e
m
o
d
el,
wh
ile
th
e
test
s
et
is
u
s
ed
to
ev
al
u
ate
its
p
er
f
o
r
m
an
ce
.
R
o
b
o
f
lo
w
is
em
p
lo
y
ed
t
o
m
an
a
g
e
th
e
d
ata
an
d
ass
is
t
in
m
o
d
el
d
ev
elo
p
m
en
t,
an
d
th
e
tr
ain
in
g
an
d
ev
alu
atio
n
o
f
th
e
id
en
ti
f
icatio
n
m
o
d
el
ar
e
ca
r
r
ie
d
o
u
t
u
s
in
g
Go
o
g
le
C
o
llab
,
p
r
o
v
id
in
g
a
f
lex
i
b
le
en
v
ir
o
n
m
en
t f
o
r
d
ee
p
lear
n
in
g
task
s
.
Fig
u
r
e
1
.
Ma
in
w
o
r
k
f
l
o
w
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
d
en
tifi
ca
tio
n
ty
p
es o
f p
la
n
t
u
s
in
g
co
n
vo
lu
tio
n
a
l n
e
u
r
a
l n
et
w
o
r
k
(
R
a
d
ityo
Hen
d
r
a
tmo
jo
J
a
ti N
o
to
n
eg
o
r
o
)
5829
2
.
1
.
Da
t
a
c
o
llect
io
n
T
h
e
im
ag
e
d
ata
u
s
ed
as
tr
ain
in
g
,
v
alid
atio
n
,
an
d
v
a
lid
atio
n
was
co
llected
f
r
o
m
Kag
g
le
(
h
ttp
s
:
//w
w
w
.
ka
g
g
le.
co
m
)
,
Gl
o
b
al
B
io
d
iv
er
s
ity
I
n
f
o
r
m
atio
n
Facility
(
GB
I
F
)
(
h
ttp
s
:
//w
w
w
.
g
b
if.o
r
g
)
,
an
d
R
o
b
o
f
lo
w
u
n
iv
er
s
e
(
h
ttp
s
:
//u
n
ivers
e.
r
o
b
o
flo
w
.
co
m
)
web
s
ites
as
m
an
y
as
2
,
5
2
6
im
ag
es
wit
h
s
p
ec
if
icatio
n
s
o
f
640
×
6
4
0
p
i
x
els
in
J
PEG
an
d
J
PG
f
o
r
m
ats.
Data
co
llected
in
t
h
e
f
o
r
m
o
f
p
ictu
r
es
o
f
f
lo
wer
s
,
f
r
u
its
,
o
r
lea
v
es
f
r
o
m
2
5
d
if
f
e
r
en
t
p
lan
t
s
p
ec
i
es,
s
u
ch
as:
ad
am
h
awa
u
n
g
u
(
Tr
a
d
esca
n
tia
p
a
llid
a
)
,
an
g
g
r
ek
(
Orch
id
a
ce
a
e
s
p
p
.
)
,
h
elico
n
ia
(
Helico
n
ia
la
tis
p
a
th
a
)
,
je
n
g
g
e
r
a
y
am
(
C
elo
s
ia
a
r
g
en
tea
)
,
k
em
b
a
n
g
s
ep
atu
(
Hib
is
cu
s
r
o
s
a
s
in
en
s
i
)
,
k
en
ca
n
a
u
n
g
u
(
R
u
ell
ia
s
imp
lex
)
,
m
ar
i
g
o
ld
(
Ta
g
ete
s
s
p
p
.
)
,
b
u
n
g
a
m
ata
h
ar
i
(
Heli
a
n
th
u
s
s
p
p
.
)
,
k
an
a
(
C
a
n
n
a
s
p
p
.
)
,
th
u
n
b
er
g
ia
(
Th
u
n
b
erg
ia
l
a
u
r
ifo
lia
)
,
b
u
n
g
a
telan
g
(
C
lito
r
ia
tern
a
tea
)
,
b
u
n
g
a
p
atr
ak
o
m
ala
(
C
a
esa
lp
in
ia
p
u
lch
err
ima
)
,
ca
b
ai
(
C
a
p
s
iu
m
a
n
n
u
m
)
,
d
au
n
b
awa
n
g
(
A
lliu
m
fis
tu
lo
s
u
m
)
,
k
em
b
an
g
b
o
k
o
r
(
Hyd
r
a
n
g
ea
ma
cro
p
h
ylla
)
,
k
e
m
u
n
in
g
(
Mu
r
r
a
ya
p
a
n
icu
la
ta
)
,
lid
ah
m
er
tu
a
(
S
a
n
s
ev
ieri
a
tr
ifa
s
cia
ta
)
,
m
ian
a
(
C
o
leu
s
s
cu
tella
r
io
id
es),
p
ac
ar
air
(
I
mp
a
tien
s
b
a
ls
a
min
a
)
,
p
u
cu
k
m
er
a
h
(
S
yzyg
iu
m
myrtif
o
liu
m
)
,
p
u
r
in
g
(
C
o
d
ia
eu
m
va
r
ieg
a
tu
m
)
,
s
awi
(
B
r
a
s
s
ica
s
p
p
.
)
,
s
elad
a
(
La
ctu
ca
s
a
tiva
)
,
s
r
i
r
ez
ek
i
(
A
g
la
o
n
ema
s
p
p
.
)
,
to
m
at
(
S
o
la
n
u
m
lyco
p
ers
icu
m
)
.
On
e
o
f
s
am
p
le
p
ictu
r
e
r
ep
r
esen
tin
g
ea
ch
ty
p
e
o
f
p
la
n
t c
o
llected
in
F
ig
u
r
e
2.
A
d
a
m
H
a
w
a
U
n
g
u
A
n
g
g
r
e
k
C
a
b
a
i
K
a
n
a
D
a
u
n
B
a
w
a
n
g
A
g
l
a
o
n
e
m
a
Te
l
a
n
g
Th
u
n
b
e
r
g
i
a
To
ma
t
H
y
d
r
a
n
g
e
a
Jen
g
g
e
r
A
y
a
m
K
e
m
u
n
i
n
g
K
e
m
b
a
n
g
S
e
p
a
t
u
Li
d
a
h
M
e
r
t
u
a
M
a
r
i
g
o
l
d
P
a
c
a
r
A
i
r
M
a
t
a
h
a
r
i
P
u
c
u
k
M
e
r
a
h
P
u
r
i
n
g
H
e
l
i
c
o
n
i
a
P
a
t
r
a
k
o
m
a
l
a
S
e
l
a
d
a
R
u
e
l
l
i
a
S
a
w
i
M
i
a
n
a
Fig
u
r
e
2
.
T
y
p
es o
f
p
lan
t sam
p
l
e
2
.
2
.
P
re
pro
ce
s
s
ing
T
h
e
co
llected
im
ag
e
d
ata,
wh
i
ch
in
itially
v
ar
ies
in
p
ix
el
s
iz
es,
is
r
esized
to
a
s
tan
d
ar
d
r
es
o
lu
tio
n
o
f
640
×
6
4
0
p
ix
els
to
en
s
u
r
e
co
n
s
is
ten
cy
ac
r
o
s
s
th
e
d
ataset.
Af
te
r
r
esizin
g
,
th
e
im
ag
es
ar
e
lab
eled
an
d
an
n
o
tate
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
6
,
Decem
b
e
r
20
25
:
5
8
2
7
-
5
8
3
6
5830
with
b
o
u
n
d
b
o
x
ar
o
u
n
d
th
e
o
b
ject
ac
co
r
d
in
g
to
s
p
ec
if
ic
p
l
an
t
f
ea
tu
r
es
s
u
ch
as
th
e
f
lo
wer
,
leaf
,
an
d
f
r
u
it
to
f
ac
ilit
ate
ac
cu
r
ate
class
if
icatio
n
,
as
s
h
o
wn
in
Fig
u
r
e
3
.
T
o
en
h
an
ce
th
e
r
o
b
u
s
tn
ess
o
f
th
e
m
o
d
el
an
d
im
p
r
o
v
e
its
g
en
er
aliza
tio
n
ab
ilit
y
,
d
ata
au
g
m
en
tatio
n
tech
n
iq
u
es
ar
e
ap
p
lied
.
T
h
ese
tech
n
iq
u
es
in
clu
d
e
h
o
r
iz
o
n
tal
an
d
v
er
tical
f
lip
p
in
g
,
9
0
º
r
o
tatio
n
s
in
b
o
th
cl
o
ck
wis
e
an
d
c
o
u
n
ter
-
clo
ck
wis
e
d
ir
ec
tio
n
s
,
as
w
ell
as
u
p
s
id
e
-
d
o
wn
r
o
tatio
n
s
.
Ad
d
itio
n
al
au
g
m
e
n
t
atio
n
s
s
u
ch
as
v
ar
y
in
g
th
e
s
atu
r
atio
n
b
etwe
en
-
5
0
%
an
d
+5
0
%,
ad
ju
s
tin
g
th
e
b
r
ig
h
tn
ess
b
etwe
en
-
2
0
%
an
d
+2
0
%,
an
d
m
o
d
if
y
in
g
ex
p
o
s
u
r
e
b
etwe
en
-
1
5
%
an
d
+1
5
%
h
elp
s
im
u
late
d
if
f
er
en
t
en
v
ir
o
n
m
en
tal
co
n
d
itio
n
s
.
Fu
r
th
er
m
o
r
e,
a
b
lu
r
ef
f
e
ct
o
f
u
p
to
1
.
5
p
x
is
ap
p
lied
to
r
an
d
o
m
im
a
g
es
in
th
e
tr
ain
in
g
s
et.
T
h
ese
au
g
m
en
tatio
n
m
eth
o
d
s
s
ig
n
if
ican
tly
in
cr
ea
s
e
th
e
d
ataset
s
ize
b
y
g
en
er
atin
g
u
p
to
4
,
2
0
6
im
ag
es,
as
s
h
o
wn
in
F
ig
u
r
e
s
4
(
a)
to
4
(
m
)
,
r
ed
u
cin
g
th
e
r
is
k
o
f
o
v
e
r
f
itti
n
g
b
y
en
s
u
r
in
g
th
e
m
o
d
el
is
ex
p
o
s
ed
t
o
a
d
iv
er
s
e
r
an
g
e
o
f
im
ag
e
v
ar
iatio
n
s
[
2
4
]
.
T
h
is
a
ls
o
h
elp
s
ad
d
r
ess
class
im
b
alan
ce
s
b
y
ar
tific
ially
in
cr
ea
s
in
g
th
e
r
ep
r
esen
tatio
n
o
f
u
n
d
e
r
r
ep
r
esen
ted
class
es
i
n
th
e
d
ataset,
im
p
r
o
v
in
g
th
e
m
o
d
el’
s
ab
ilit
y
to
g
en
er
alize
ac
r
o
s
s
d
if
f
e
r
en
t p
la
n
t ty
p
es
[
2
5
]
.
Fig
u
r
e
3
.
An
n
o
tated
a
n
d
lab
el
ed
th
e
k
em
u
n
in
g
(
Mu
r
r
a
ya
p
a
n
icu
la
ta
)
with
b
o
u
n
d
in
g
b
o
x
(
a)
(
b
)
(
c)
(
d
)
(
e)
(f)
(
g
)
(
h
)
(
i)
(
j)
(
k
)
(
l)
(m)
Fig
u
r
e
4
.
Au
g
m
en
ted
im
ag
e,
(
a
)
o
r
ig
in
al,
(
b
)
f
lip
(
v
er
tical
)
,
(
c)
f
lip
(
h
o
r
izo
n
tal
)
,
(
d
)
9
0
º
r
o
t
atio
n
s
(
clo
ck
wis
e)
,
(
e
)
9
0
º
r
o
tatio
n
s
(
c
o
u
n
te
r
clo
c
k
wis
e)
,
(
f
)
9
0
º
r
o
tatio
n
s
(
u
p
s
id
e
d
o
wn
)
,
(
g
)
s
atu
r
atio
n
(
5
0
%),
(
h
)
s
atu
r
atio
n
(
+5
0
%),
(
i
)
b
r
ig
h
tn
ess
(
d
ar
k
e
n
:
-
2
0
%),
(
j
)
b
r
ig
h
tn
ess
(
b
r
ig
h
te
n
: +
2
0
%),
(
k
)
ex
p
o
s
u
r
e
(
-
1
5
%),
(
l
)
ex
p
o
s
u
r
e
(
+1
5
%),
an
d
(
m
)
b
lu
r
1
.
5
p
x
2
.
3
.
T
ra
in
T
h
e
d
ataset
co
n
tain
s
2
5
class
es:
ad
am
h
awa
u
n
g
u
(
1
0
1
im
ag
es),
an
g
g
r
ek
(
1
0
0
im
ag
es),
h
elico
n
ia
(
1
0
0
im
ag
es),
jen
g
g
e
r
a
y
am
(
1
0
2
im
a
g
es),
k
e
m
b
an
g
s
e
p
atu
(
1
0
0
im
ag
es),
k
en
ca
n
a
u
n
g
u
(
9
8
im
ag
es),
m
ar
ig
o
ld
(
9
8
im
a
g
es),
b
u
n
g
a
m
atah
ar
i
(
1
0
0
im
ag
es),
k
an
a
(
1
0
5
im
a
g
es),
th
u
n
b
er
g
ia
(
1
0
1
im
ag
es),
b
u
n
g
a
telan
g
(
1
1
4
im
a
g
es),
b
u
n
g
a
p
atr
ak
o
m
ala
(
1
0
2
im
ag
es),
c
ab
ai
(
1
0
2
im
ag
es),
d
au
n
b
a
wan
g
(
1
0
9
im
ag
es),
k
em
b
an
g
b
o
k
o
r
(
1
0
0
im
ag
es),
k
em
u
n
in
g
(
1
0
1
im
ag
es),
lid
a
h
m
er
tu
a
(
9
9
im
ag
es),
m
ian
a
(
9
7
im
ag
es)
,
p
ac
a
r
air
(
1
0
2
im
ag
es),
p
u
cu
k
m
e
r
a
h
(
8
8
im
ag
es),
p
u
r
in
g
(
1
0
0
im
ag
es),
s
awi
(
1
0
3
im
ag
es),
s
elad
a
(
1
0
4
im
a
g
es),
s
r
i
r
ez
ek
i
(
1
0
0
im
ag
es),
t
o
m
at
(
1
0
0
im
ag
es),
a
n
d
allo
ca
ted
as
s
h
o
wn
in
T
a
b
le
1
s
p
lit
in
to
:
8
0
%
o
f
th
e
d
ata
is
f
o
r
tr
ain
in
g
th
e
m
o
d
el,
1
0
%
is
f
o
r
v
alid
atin
g
t
h
e
m
o
d
el
a
n
d
th
e
r
em
ain
in
g
1
0
%
is
f
o
r
test
in
g
th
e
m
o
d
el.
T
o
en
h
an
ce
th
e
r
o
b
u
s
tn
ess
an
d
g
en
er
aliza
tio
n
ca
p
a
b
ilit
y
o
f
t
h
e
m
o
d
el,
d
ata
au
g
m
e
n
tatio
n
tech
n
iq
u
es
wer
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
-
8
7
0
8
I
d
en
tifi
ca
tio
n
ty
p
es o
f p
la
n
t
u
s
in
g
co
n
vo
lu
tio
n
a
l n
e
u
r
a
l n
et
w
o
r
k
(
R
a
d
ityo
Hen
d
r
a
tmo
jo
J
a
ti N
o
to
n
eg
o
r
o
)
5831
ap
p
lied
p
r
io
r
to
th
e
tr
ain
in
g
p
h
ase.
T
h
ese
au
g
m
en
ted
an
d
p
r
e
-
p
r
o
ce
s
s
ed
im
a
g
es
wer
e
th
e
n
u
s
ed
t
o
tr
ain
th
e
YOL
Ov
8
m
o
d
el
u
n
d
er
v
a
r
io
u
s
ex
p
er
im
en
tal
s
ettin
g
s
,
u
s
in
g
co
m
b
i
n
atio
n
s
o
f
t
h
e
s
to
ch
ast
ic
g
r
ad
ie
n
t
d
escen
t
(
SGD)
an
d
Ad
am
o
p
tim
izer
s
,
b
atch
s
izes
o
f
1
6
an
d
3
2
,
lear
n
in
g
r
ates
o
f
0
.
0
1
an
d
0
.
0
0
1
,
an
d
tr
ai
n
in
g
d
u
r
atio
n
s
o
f
2
5
,
5
0
,
7
5
,
an
d
1
0
0
ep
o
ch
s
.
T
h
ese
v
ar
iatio
n
s
wer
e
d
esig
n
ed
to
ex
p
l
o
r
e
an
d
id
en
tify
th
e
m
o
s
t
o
p
tim
al
s
et
o
f
p
ar
am
eter
s
f
o
r
ac
cu
r
ate
an
d
e
f
f
icien
t
p
lan
t
cl
ass
if
icatio
n
.
T
h
e
tr
ai
n
in
g
p
r
o
c
ess
was
co
n
d
u
cted
o
n
Go
o
g
le
C
o
lab
,
u
tili
zin
g
a
c
lo
u
d
-
b
ased
e
n
v
ir
o
n
m
en
t
e
q
u
ip
p
ed
with
an
NVI
DI
A
T
esla
T
4
GPU,
r
u
n
n
in
g
o
n
C
UDA
v
er
s
io
n
1
2
.
0
,
a
n
d
s
u
p
p
o
r
ted
b
y
1
6
GB
o
f
m
e
m
o
r
y
.
T
h
is
h
ar
d
war
e
s
etu
p
en
s
u
r
ed
s
u
f
f
icien
t
co
m
p
u
tatio
n
al
p
o
wer
f
o
r
h
a
n
d
lin
g
h
i
g
h
-
r
eso
lu
ti
o
n
im
a
g
e
d
ata
an
d
lar
g
e
m
o
d
el
ar
ch
ite
ctu
r
es,
wh
ile
also
ac
ce
ler
atin
g
tr
ain
in
g
tim
es
ac
r
o
s
s
all
ex
p
er
im
en
tal
co
n
f
i
g
u
r
atio
n
s
.
T
h
e
d
etails
o
f
th
e
d
a
taset
allo
ca
tio
n
f
o
r
tr
ain
in
g
,
v
alid
atio
n
,
an
d
test
in
g
p
u
r
p
o
s
es
ar
e
co
m
p
r
eh
e
n
s
iv
ely
p
r
esen
ted
in
T
ab
le
1
,
wh
ich
s
er
v
es
as
a
r
ef
er
en
ce
f
o
r
u
n
d
e
r
s
tan
d
in
g
th
e
d
ata
d
is
tr
ib
u
tio
n
u
s
ed
th
r
o
u
g
h
o
u
t th
e
m
o
d
el
d
ev
elo
p
m
en
t p
r
o
ce
s
s
.
T
ab
le
1
.
Data
s
et
allo
ca
tio
n
D
a
t
a
s
e
t
Tr
a
i
n
V
a
l
i
d
a
t
i
o
n
Te
st
i
n
g
Ty
p
e
s
o
f
p
l
a
n
t
3
6
6
0
4
2
3
4
2
3
2
.
4
.
E
v
a
lua
t
e
T
h
e
r
esu
lts
o
f
th
e
m
o
d
el
tr
ain
in
g
g
et
p
r
ec
is
io
n
,
r
ec
all
,
as
w
ell
as
th
e
ac
cu
r
ac
y
o
f
t
h
e
m
ea
n
av
er
a
g
e
p
r
ec
is
io
n
(
m
AP)
5
0
a
n
d
m
AP5
0
-
9
0
.
T
h
e
p
r
ec
is
io
n
v
alu
e
m
ea
s
u
r
es
th
e
YOL
Ov
8
m
o
d
el'
s
p
r
ed
ictio
n
o
f
p
o
s
itiv
e
in
s
tan
ce
s
co
r
r
ec
tly
,
r
ec
all
m
ea
s
u
r
es
th
e
id
en
tific
atio
n
o
f
th
e
YOL
Ov
8
m
o
d
e
l
ag
ain
s
t
p
o
s
itiv
e
in
s
tan
ce
s
,
th
e
F1
s
co
r
e
is
th
e
av
er
ag
e
b
alan
ce
b
etwe
en
p
r
ec
is
io
n
an
d
s
en
s
itiv
ity
,
th
e
F1
s
co
r
e
is
u
s
ef
u
l
wh
e
n
th
e
m
eth
o
d
h
as
lo
w
s
en
s
itiv
ity
b
u
t
h
ig
h
p
r
ec
is
io
n
o
r
h
ig
h
s
en
s
itiv
ity
b
u
t
lo
w
p
r
ec
is
io
n
[
2
6
]
,
m
ea
s
u
r
e
th
e
class
if
ier
p
er
f
o
r
m
an
ce
co
m
p
r
e
h
en
s
iv
ely
[
2
7
]
,
an
d
a
h
ig
h
F1
s
co
r
e
in
d
icate
s
th
e
m
o
d
el
m
o
r
e
r
o
b
u
s
t
[
2
8
]
,
m
AP
ca
lcu
lates
av
er
ag
e
p
r
ec
is
io
n
a
g
ain
s
t
s
en
s
itiv
ity
v
alu
es
in
th
e
r
an
g
e
0
-
1
[
2
9
]
,
co
m
p
a
r
es
p
e
r
f
o
r
m
a
n
ce
b
etwe
en
d
etec
to
r
s
[
3
]
,
ass
ess
e
s
th
e
m
o
d
els
o
f
o
b
ject
d
etec
tio
n
p
er
f
o
r
m
an
ce
ac
r
o
s
s
m
u
ltip
le
ca
t
eg
o
r
ies
[
3
0
]
,
an
d
p
r
o
v
id
es
m
o
d
el
s
u
m
m
ar
y
[
3
1
]
.
m
AP5
0
e
x
p
r
ess
es
av
er
ag
e
p
r
ec
is
io
n
at
th
e
I
o
U
th
r
es
h
o
ld
o
f
5
0
%
an
d
m
AP5
0
-
9
0
ex
p
r
ess
es
th
e
av
e
r
ag
e
p
r
ec
is
io
n
at
th
e
I
o
U
th
r
esh
o
ld
o
f
5
0
%
to
9
0
%
[
1
0
]
.
I
o
U
ca
lcu
late
th
e
q
u
an
tific
atio
n
s
im
ilar
ity
o
f
p
r
e
d
icted
b
o
u
n
d
in
g
b
o
x
(
)
an
d
g
r
o
u
n
d
tr
u
t
h
b
o
u
n
d
in
g
b
o
x
(
)
[
3
2
]
,
I
o
U
v
alu
es
th
at
ex
ce
ed
a
ce
r
tain
th
r
esh
o
l
d
,
ca
n
b
e
co
n
s
id
er
ed
to
p
r
o
d
u
ce
tr
u
e
p
o
s
itiv
e
d
etec
tio
n
r
esu
lts
[
3
3
]
,
an
d
o
b
jects
th
at
ex
ce
ed
I
o
U
v
al
u
e
o
f
5
0
%
ca
n
b
e
class
if
ied
as
d
etec
ted
[
3
4
]
.
Av
er
a
g
e
p
r
ed
ictio
n
(
AP)
v
alu
e
is
n
ee
d
ed
t
o
co
m
p
u
te
m
AP
[
3
5
]
.
T
h
e
e
q
u
at
io
n
s
o
f
p
r
ec
is
io
n
[
3
6
]
,
r
ec
all
[
3
7
]
,
F1
s
co
r
e
[
3
8
]
a
n
d
m
AP
ar
e
f
o
r
m
u
lated
in
(
1
)
th
r
o
u
g
h
(
6
)
.
(
)
=
+
(
1
)
(
)
=
+
(
2
)
1
=
2
×
×
+
(
3
)
=
(
∩
)
(
∪
)
(
4
)
=
∫
(
)
1
0
(
5
)
=
1
∑
A
=
1
(
6
)
I
n
th
e
e
v
alu
atio
n
o
f
th
e
YOL
Ov
8
m
o
d
el'
s
p
er
f
o
r
m
an
ce
,
s
ev
er
al
k
ey
m
etr
ics
wer
e
u
s
ed
t
o
in
ter
p
r
et
th
e
ac
cu
r
ac
y
an
d
r
elia
b
ilit
y
o
f
o
b
ject
d
etec
tio
n
an
d
class
if
icatio
n
.
T
r
u
e
p
o
s
itiv
e
(
T
P)
r
ef
er
s
to
in
s
tan
ce
s
wh
er
e
th
e
m
o
d
el
co
r
r
ec
tly
id
e
n
tifie
s
an
d
class
if
ies
a
s
p
ec
if
ic
p
lan
t
o
b
ject,
an
d
th
e
p
r
e
d
icted
b
o
u
n
d
in
g
b
o
x
o
v
e
r
lap
s
s
ig
n
if
ican
tly
with
th
e
g
r
o
u
n
d
tr
u
th
.
Fals
e
p
o
s
itiv
e
(
FP
)
o
cc
u
r
s
wh
en
th
e
m
o
d
el
co
r
r
ec
tly
class
if
ies
an
o
b
ject
b
u
t
th
e
p
r
ed
icted
b
o
u
n
d
in
g
b
o
x
d
o
es
n
o
t
co
r
r
esp
o
n
d
to
an
y
ac
tu
al
o
b
ject,
lead
in
g
to
a
m
is
m
atch
.
C
o
n
v
er
s
ely
,
Fals
e
n
eg
ativ
e
(
FN)
in
d
icate
s
th
at
th
e
m
o
d
el
f
ails
to
d
etec
t
o
r
co
r
r
ec
tly
class
if
y
a
p
lan
t
o
b
ject
th
at
is
p
r
esen
t
in
th
e
im
ag
e.
N
d
en
o
tes
th
e
n
u
m
b
er
o
f
o
b
ject
ca
teg
o
r
ies
b
ein
g
d
etec
ted
,
wh
ile
APᵢ
r
ep
r
esen
ts
th
e
av
er
ag
e
p
r
ec
is
io
n
(
AP)
f
o
r
th
e
i
-
th
ca
teg
o
r
y
,
r
ef
lectin
g
h
o
w
well
th
e
m
o
d
el
p
er
f
o
r
m
s
p
er
class
.
T
h
e
p
er
f
o
r
m
an
ce
o
u
tco
m
es
ar
e
f
u
r
th
er
v
is
u
alize
d
u
s
in
g
a
s
er
ies
o
f
ev
alu
atio
n
g
r
ap
h
s
.
T
h
e
b
o
x
lo
s
s
m
etr
ic
ev
alu
ates
h
o
w
ac
c
u
r
at
ely
th
e
p
r
ed
icted
b
o
u
n
d
in
g
b
o
x
es
alig
n
with
th
e
tr
u
e
lo
ca
t
io
n
s
o
f
th
e
o
b
jects,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
6
,
Decem
b
e
r
20
25
:
5
8
2
7
-
5
8
3
6
5832
s
er
v
in
g
as
an
in
d
icato
r
o
f
lo
ca
lizatio
n
p
er
f
o
r
m
an
ce
.
C
lass
if
icatio
n
lo
s
s
(
C
L
S
lo
s
s
)
r
ef
l
ec
ts
h
o
w
well
th
e
m
o
d
el
d
is
tin
g
u
is
h
es
b
etwe
en
d
if
f
e
r
en
t
p
lan
t
ca
teg
o
r
ies,
h
ig
h
lig
h
tin
g
its
class
if
icatio
n
ac
cu
r
ac
y
.
T
h
e
d
is
tr
ib
u
tio
n
f
o
ca
l
lo
s
s
(
DFL
l
o
s
s
)
is
p
ar
ticu
lar
ly
u
s
ef
u
l
in
s
ce
n
ar
io
s
in
v
o
lv
in
g
class
im
b
alan
ce
,
as
it
h
elp
s
r
ef
in
e
p
r
ed
ictio
n
s
f
o
r
ca
teg
o
r
ies
th
at
ar
e
u
n
d
e
r
r
ep
r
esen
ted
in
th
e
d
ataset.
Ad
d
itio
n
al
m
etr
ic
s
in
clu
d
e
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
an
d
r
ec
all,
wh
ich
co
llectiv
ely
d
escr
ib
e
th
e
m
o
d
el'
s
o
v
er
all
co
r
r
ec
tn
ess
an
d
co
m
p
leten
ess
in
d
etec
tio
n
.
T
h
e
m
ea
n
a
v
er
ag
e
p
r
ec
is
io
n
(
m
AP
)
is
r
ep
o
r
ted
a
t
b
o
th
m
AP5
0
)
an
d
m
AP5
0
-
9
0
,
p
r
o
v
id
in
g
a
m
o
r
e
n
u
an
ce
d
u
n
d
er
s
tan
d
in
g
o
f
th
e
m
o
d
el'
s
r
o
b
u
s
tn
ess
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
YOL
Ov
8
m
o
d
el
was
tr
ain
ed
th
r
o
u
g
h
eig
h
t
d
if
f
er
e
n
t
ex
p
er
im
en
ts
,
ea
ch
u
s
in
g
a
u
n
iq
u
e
co
m
b
in
atio
n
o
f
h
y
p
er
p
ar
a
m
eter
s
s
u
ch
as
lear
n
in
g
r
ate,
b
atc
h
s
ize,
an
d
o
p
tim
izer
s
ettin
g
s
,
all
co
n
d
u
cted
o
v
er
1
0
0
e
p
o
ch
s
.
T
h
e
o
u
tco
m
es
o
f
th
ese
ex
p
e
r
im
en
ts
wer
e
e
v
a
lu
ated
u
s
in
g
k
ey
p
er
f
o
r
m
an
c
e
m
etr
ics
in
clu
d
in
g
p
r
ec
is
io
n
,
r
ec
all,
F1
s
co
r
e,
m
ea
n
av
er
ag
e
p
r
ec
is
io
n
at
I
o
U
th
r
esh
o
ld
s
o
f
0
.
5
0
(
m
AP5
0
)
,
an
d
0
.
5
0
to
0
.
9
5
(
m
AP5
0
-
9
0
)
,
with
th
e
s
u
m
m
a
r
ized
r
esu
lts
p
r
esen
ted
in
T
ab
l
e
2
.
I
n
ad
d
itio
n
to
th
e
tab
u
late
d
m
etr
ics,
T
ab
le
2
also
in
clu
d
es
ev
alu
atio
n
g
r
a
p
h
s
illu
s
tr
atin
g
th
e
tr
en
d
s
o
f
b
o
x
lo
s
s
,
class
if
icat
io
n
lo
s
s
(
C
L
S
lo
s
s
)
,
an
d
d
is
tr
ib
u
tio
n
f
o
ca
l
lo
s
s
(
DFL
lo
s
s
)
,
as
well
a
s
cu
r
v
es
d
ep
ictin
g
th
e
ev
o
lu
tio
n
o
f
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
b
o
th
m
AP5
0
an
d
m
AP5
0
-
9
0
o
v
er
th
e
co
u
r
s
e
o
f
th
e
tr
ain
in
g
ep
o
ch
s
.
Fu
r
th
er
m
o
r
e,
th
e
tab
le
p
r
o
v
id
es
in
f
o
r
m
atio
n
o
n
th
e
to
tal
tim
e
r
eq
u
ir
ed
f
o
r
ea
ch
ex
p
er
im
en
t,
e
n
ab
lin
g
a
co
m
p
r
eh
en
s
iv
e
c
o
m
p
ar
is
o
n
o
f
tr
ain
in
g
ef
f
icien
cy
an
d
m
o
d
el
p
er
f
o
r
m
an
ce
u
n
d
er
d
if
f
er
en
t
p
ar
am
ete
r
s
ettin
g
s
.
T
ab
le
2
.
Pre
cisi
o
n
,
r
ec
all,
F1
s
co
r
e,
m
AP5
0
,
an
d
m
AP5
0
-
9
0
v
alu
es
T
h
e
r
esu
lt
in
Fig
u
r
e
5
s
h
o
ws
th
e
tr
ain
b
o
x
lo
s
s
d
ec
r
ea
s
in
g
tr
en
d
o
v
e
r
th
e
tr
ain
in
g
ep
o
c
h
s
,
in
d
icatin
g
th
at
th
e
m
o
d
el
is
lear
n
in
g
to
p
r
ed
ict
m
o
r
e
ac
cu
r
ate
b
o
u
n
d
in
g
b
o
x
es
as
tr
ain
in
g
p
r
o
g
r
ess
es.
I
t
s
ee
m
s
to
p
latea
u
to
war
d
s
th
e
en
d
,
s
u
g
g
esti
n
g
c
o
n
v
er
g
en
ce
.
Als
o
,
th
e
tr
ain
cl
ass
lo
s
s
ex
h
ib
its
a
d
ec
r
ea
s
in
g
tr
en
d
,
im
p
l
y
in
g
t
h
at
th
e
m
o
d
el
is
im
p
r
o
v
in
g
its
ab
ilit
y
to
class
if
y
o
b
jects
c
o
r
r
ec
tly
d
u
r
i
n
g
tr
ain
in
g
.
I
t
also
ap
p
ea
r
s
to
b
e
co
n
v
er
g
in
g
.
Fo
llo
ws
a
s
im
ilar
d
ec
r
ea
s
in
g
p
atter
n
,
th
e
tr
ain
DFL
lo
s
s
s
u
g
g
esti
n
g
th
at
th
e
m
o
d
el
is
b
ec
o
m
in
g
b
etter
at
p
r
ed
ictin
g
th
e
p
r
ec
is
e
d
is
tr
ib
u
tio
n
o
f
b
o
u
n
d
in
g
b
o
x
co
o
r
d
in
ates.
Gen
er
ally
,
in
cr
ea
s
es
o
v
er
tr
ain
in
g
,
th
e
tr
ain
p
r
ec
is
io
n
in
d
icatin
g
th
at
th
e
m
o
d
el
is
m
ak
in
g
f
ew
er
f
alse
p
o
s
itiv
e
b
o
u
n
d
in
g
b
o
x
p
r
e
d
ictio
n
s
as
it
lear
n
s
.
I
t
f
lu
ctu
ates,
wh
ich
is
co
m
m
o
n
d
u
r
in
g
tr
ai
n
in
g
.
Sh
o
ws
an
in
cr
ea
s
in
g
tr
en
d
in
iti
ally
o
n
tr
ain
r
ec
all,
m
ea
n
in
g
th
e
m
o
d
el
is
lear
n
in
g
to
d
etec
t
m
o
r
e
o
f
th
e
ac
t
u
a
l
o
b
jects
p
r
esen
t.
I
t
s
ee
m
s
to
p
latea
u
o
r
s
lig
h
tly
d
ec
r
ea
s
e
to
war
d
s
th
e
e
n
d
,
w
h
ich
co
u
ld
b
e
a
s
ig
n
o
f
o
v
e
r
f
itti
n
g
if
th
e
v
alid
atio
n
r
ec
all
d
o
esn
'
t
f
o
llo
w
th
e
s
am
e
tr
en
d
.
T
h
e
v
alid
atio
n
b
o
x
lo
s
s
d
ec
r
ea
s
es
in
itially
o
n
b
u
t
t
h
en
s
ee
m
s
to
s
tab
ilize
an
d
m
ig
h
t
ev
en
s
lig
h
tly
in
cr
ea
s
e
o
r
f
lu
ctu
ate
in
th
e
lat
er
ep
o
ch
s
.
T
h
is
s
u
g
g
ests
th
at
th
e
m
o
d
el'
s
ab
ilit
y
to
g
en
er
alize
its
b
o
u
n
d
in
g
b
o
x
p
r
ed
ictio
n
s
o
n
u
n
s
ee
n
d
ata
m
i
g
h
t
h
av
e
p
latea
u
e
d
o
r
s
tar
ted
t
o
s
lig
h
tly
d
eg
r
ad
e
.
Dec
r
ea
s
es
in
th
e
ea
r
ly
s
tag
es
b
u
t
th
en
p
latea
u
s
an
d
s
h
o
ws
s
o
m
e
f
lu
ctu
atio
n
s
o
n
v
alid
atio
n
class
lo
s
s
.
T
h
is
in
d
icate
s
th
at
th
e
class
if
icatio
n
p
er
f
o
r
m
an
ce
o
n
u
n
s
ee
n
d
ata
is
n
o
lo
n
g
er
s
ig
n
if
ica
n
tly
im
p
r
o
v
in
g
.
Similar
to
th
e
b
o
x
lo
s
s
,
th
e
v
alid
atio
n
DFL
lo
s
s
s
h
o
ws
an
in
itial
d
ec
r
ea
s
e
f
o
llo
wed
b
y
s
tab
ilizatio
n
a
n
d
s
o
m
e
f
lu
ctu
atio
n
s
.
I
n
c
r
ea
s
es
s
ig
n
if
ican
tly
in
th
e
b
eg
in
n
in
g
a
n
d
t
h
en
s
tar
ts
to
p
latea
u
.
T
h
is
s
u
g
g
ests
th
at
t
h
e
o
v
er
all
d
etec
tio
n
p
e
r
f
o
r
m
an
ce
at
a
5
0
%
I
o
U
th
r
esh
o
ld
o
n
th
e
v
alid
atio
n
s
et
h
as
r
ea
ch
ed
a
ce
r
tain
lev
el
a
n
d
is
n
o
lo
n
g
er
im
p
r
o
v
i
n
g
m
u
c
h
.
Sh
o
ws
a
s
im
ilar
tr
en
d
to
m
AP5
0
b
u
t
with
g
en
er
ally
lo
wer
v
alu
es,
as
ex
p
ec
ted
d
u
e
to
th
e
s
tr
icter
I
o
U
th
r
esh
o
ld
s
.
T
h
e
p
latea
u
in
g
in
d
icate
s
th
at
th
e
m
o
d
el'
s
ab
ilit
y
to
p
r
ec
is
ely
lo
ca
lize
o
b
jects
o
n
u
n
s
ee
n
d
ata
i
s
also
n
o
t
im
p
r
o
v
in
g
s
ig
n
if
ican
tly
.
Ov
er
all,
th
e
m
o
d
el
o
n
tr
ain
in
g
p
h
ase
lo
s
s
es
is
g
en
er
ally
d
ec
r
e
asin
g
,
in
d
icati
n
g
th
at
th
e
m
o
d
el
is
lear
n
in
g
o
n
th
e
tr
ain
i
n
g
d
at
a.
Ho
wev
er
,
th
e
v
alid
atio
n
p
h
ase
lo
s
s
es
an
d
m
AP
m
etr
ics
h
av
e
p
latea
u
ed
,
s
u
g
g
esti
n
g
th
at
th
e
m
o
d
el
m
ig
h
t
h
av
e
r
ea
c
h
ed
its
o
p
tim
al
p
er
f
o
r
m
an
ce
o
n
th
e
u
n
s
ee
n
d
ata
o
r
is
s
tar
tin
g
to
o
v
er
f
it to
th
e
tr
ain
in
g
d
ata.
Fi
g
u
r
e
6
s
h
o
ws th
e
co
n
f
u
s
io
n
m
atr
ix
o
f
0
.
0
1
lear
n
in
g
r
ate,
3
2
b
atch
s
ize,
an
d
SGD
o
p
tim
izer
.
Le
a
r
n
i
n
g
r
a
t
e
B
a
t
c
h
S
i
z
e
O
p
t
i
mi
z
e
r
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F
1
(
%)
mA
P
5
0
(
%)
mA
P
5
0
-
9
0
(
%)
Ti
me
0
.
0
1
16
S
G
D
6
3
.
1
1
5
3
.
4
4
5
7
.
8
8
5
6
.
3
5
4
1
.
1
4
1
2
5
m
2
3
s
0
.
0
1
16
A
d
a
m
6
5
.
1
7
5
1
.
5
3
5
7
.
4
8
5
3
.
4
4
3
7
.
2
2
1
2
5
m
3
1
s
0
.
0
1
32
S
G
D
6
9
.
6
2
5
4
.
7
3
6
1
.
2
2
5
7
.
6
1
4
2
.
4
9
1
2
3
m
4
0
s
0
.
0
1
32
A
d
a
m
5
6
.
2
3
5
2
.
7
6
5
4
.
3
0
5
2
.
8
6
3
7
.
2
0
1
2
5
m
3
4
s
0
.
0
0
1
16
S
G
D
5
5
.
6
1
5
7
.
1
7
5
6
.
4
4
5
5
.
4
9
3
9
.
7
1
1
1
3
m
8
s
0
.
0
0
1
16
A
d
a
m
6
0
.
0
7
5
8
.
5
9
5
9
.
2
8
5
8
.
9
7
4
3
.
5
5
1
1
4
m
5
9
s
0
.
0
0
1
32
S
G
D
6
2
.
2
4
5
8
.
9
9
6
0
.
6
4
5
5
.
9
4
3
9
.
8
5
1
1
0
m
4
4
s
0
.
0
0
1
32
A
d
a
m
6
3
.
8
9
5
3
.
8
5
5
8
.
5
2
5
7
.
5
3
4
2
.
0
5
1
1
1
m
1
2
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
d
en
tifi
ca
tio
n
ty
p
es o
f p
la
n
t
u
s
in
g
co
n
vo
lu
tio
n
a
l n
e
u
r
a
l n
et
w
o
r
k
(
R
a
d
ityo
Hen
d
r
a
tmo
jo
J
a
ti N
o
to
n
eg
o
r
o
)
5833
Fig
u
r
e
5
.
Mo
d
el
r
esu
lt o
f
0
.
0
1
lear
n
in
g
r
ate,
3
2
b
atch
s
ize,
an
d
SGD
o
p
tim
izer
Fig
u
r
e
6
.
T
h
e
co
n
f
u
s
io
n
m
atr
ix
o
f
0
.
0
1
lear
n
in
g
r
ate,
3
2
b
atc
h
s
ize,
an
d
SGD
o
p
tim
izer
4.
CO
NCLU
SI
O
N
T
h
e
YOL
Ov
8
m
o
d
el
with
0
.
0
1
lear
n
in
g
r
ate,
3
2
b
atch
s
ize,
SGD
o
p
tim
izer
o
b
tain
e
d
h
ig
h
est
p
r
ec
is
io
n
o
f
6
9
.
6
2
%
a
n
d
F1
s
c
o
r
e
o
f
6
1
.
2
2
%,
r
ec
all
o
f
5
4
.
7
3
%,
m
AP5
0
an
d
m
AP5
0
–
9
0
o
n
th
e
tr
ain
in
g
d
ata
o
f
5
7
.
6
1
%
an
d
4
2
.
4
9
%
f
o
r
1
2
3
m
4
0
s
.
T
h
e
co
n
f
u
s
io
n
m
atr
i
x
in
d
icate
s
th
e
p
r
o
p
o
r
tio
n
o
f
p
r
ed
ictio
n
s
f
o
r
ea
ch
ac
tu
al
class
.
T
h
e
n
u
m
b
er
s
o
u
t
s
id
e
th
e
m
ain
d
iag
o
n
al
clea
r
ly
in
d
icate
class
if
icatio
n
er
r
o
r
s
.
T
h
e
m
ain
er
r
o
r
s
o
f
th
e
m
o
d
el
lie
in
th
e
d
if
f
icu
lt
y
o
f
d
is
tin
g
u
is
h
in
g
b
etwe
en
p
lan
ts
an
d
th
e
b
ac
k
g
r
o
u
n
d
,
a
s
well
as
co
n
f
u
s
io
n
b
etwe
en
ce
r
tain
p
lan
t
class
es,
th
is
b
ec
au
s
e
m
o
s
t
o
f
im
ag
es
c
o
llected
in
d
ataset
h
av
e
m
an
y
o
b
jects
co
m
p
le
x
ity
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
6
,
Decem
b
e
r
20
25
:
5
8
2
7
-
5
8
3
6
5834
at
b
ac
k
g
r
o
u
n
d
.
th
e
h
ig
h
v
al
u
ed
class
es,
Sawi
an
d
Sela
d
a,
th
o
s
e
im
ag
es
d
o
n
o
t
h
av
e
m
u
ch
o
b
ject
at
b
ac
k
g
r
o
u
n
d
s
o
th
e
m
o
d
el
co
u
ld
ea
s
ily
r
ec
o
g
n
ize
th
e
o
b
je
ct.
B
esid
e
s
,
th
e
lab
el
al
s
o
u
s
i
n
g
o
th
er
p
ar
ts
lik
e,
f
lo
wer
,
leav
es,
an
d
f
r
u
its
o
f
th
e
p
lan
ts
an
d
n
o
t
all
o
f
th
e
ty
p
e
in
d
ataset
h
av
e
th
o
s
e
p
ar
ts
.
I
n
f
u
t
u
r
e,
th
e
co
llected
d
ata
s
h
o
u
ld
r
ed
u
ce
th
e
im
ag
e
b
ac
k
g
r
o
u
n
d
to
av
o
id
d
ata
co
m
p
lex
ity
,
g
at
h
er
m
o
r
e
im
ag
e
d
ata
an
d
au
g
m
en
tatio
n
to
av
o
id
o
v
er
f
itt
in
g
,
an
d
ad
ju
s
tin
g
m
o
d
el
h
y
p
e
r
p
ar
am
eter
s
u
c
h
as
lear
n
in
g
r
a
te
an
d
o
p
tim
izer
to
p
r
ev
en
t
u
n
d
e
r
f
itti
n
g
an
d
in
s
tab
ilit
y
in
tr
ain
in
g
p
r
o
ce
s
s
.
ACK
NO
WL
E
DG
E
M
E
NT
S
T
h
e
au
th
o
r
wo
u
ld
lik
e
to
ex
p
r
ess
s
in
ce
r
e
ap
p
r
ec
iatio
n
to
t
h
e
s
ec
o
n
d
a
u
th
o
r
,
Hu
s
tin
awa
ty
,
f
o
r
h
er
v
alu
ab
le
g
u
i
d
an
ce
th
r
o
u
g
h
o
u
t
th
is
r
esear
ch
.
T
h
e
au
th
o
r
also
g
r
atef
u
lly
ac
k
n
o
wled
g
es
th
e
Facu
lty
o
f
I
n
f
o
r
m
atio
n
Sy
s
tem
s
an
d
Ma
n
ag
em
en
t,
U
n
iv
er
s
itas
Gu
n
ad
ar
m
a,
I
n
d
o
n
esia,
f
o
r
th
eir
s
u
p
p
o
r
t
in
f
ac
ilit
atin
g
th
is
s
tu
d
y
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT
)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
e
d
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
a
n
d
f
ac
ilit
ate
co
ll
ab
o
r
atio
n
.
R
ad
ity
o
Hen
d
r
atm
o
jo
J
ati
No
to
n
e
g
o
r
o
co
n
tr
ib
u
ted
to
m
eth
o
d
o
l
o
g
y
,
s
o
f
twar
e
d
ev
el
o
p
m
en
t,
v
alid
ati
o
n
,
f
o
r
m
al
a
n
aly
s
is
,
in
v
esti
g
atio
n
,
r
eso
u
r
ce
s
,
d
ata
cu
r
atio
n
,
wr
itin
g
–
o
r
i
g
in
al
d
r
af
t
p
r
ep
a
r
atio
n
,
wr
itin
g
–
r
ev
iew
an
d
ed
itin
g
,
v
is
u
aliz
atio
n
,
an
d
p
r
o
ject
ad
m
in
is
tr
atio
n
.
Hu
s
tin
awa
ty
co
n
tr
ib
u
ted
to
co
n
ce
p
tu
aliza
tio
n
,
wr
itin
g
–
r
ev
iew
an
d
ed
itin
g
,
s
u
p
er
v
is
io
n
,
an
d
f
u
n
d
in
g
ac
q
u
is
itio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
R
ad
ity
o
Hen
d
r
atm
o
j
o
J
ati
No
to
n
eg
o
r
o
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Hu
s
tin
awa
ty
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
in
g
s
o
f
th
is
s
tu
d
y
a
r
e
o
p
en
ly
av
ailab
le
f
r
o
m
m
u
ltip
le
p
u
b
lic
r
ep
o
s
ito
r
ies,
in
clu
d
in
g
Kag
g
le,
GB
I
F,
an
d
R
o
b
o
f
lo
w
U
n
iv
er
s
e.
Sp
ec
if
ic
d
atasets
w
er
e
o
b
tain
ed
f
r
o
m
th
ese
s
o
u
r
ce
s
an
d
c
o
m
b
in
e
d
f
o
r
th
e
p
u
r
p
o
s
e
o
f
th
is
r
e
s
ea
r
ch
.
T
h
e
d
atasets
ca
n
b
e
ac
ce
s
s
ed
v
ia
th
e
f
o
llo
win
g
p
latf
o
r
m
s
:
Kag
g
le
(
h
ttp
s
:
//w
w
w
.
ka
g
g
le.
co
m
)
GB
I
F
(
h
ttp
s
:
//w
w
w
.
g
b
if.o
r
g
)
R
o
b
o
f
lo
w
Un
iv
e
r
s
e
(
h
ttp
s
:
//u
n
ivers
e.
r
o
b
o
flo
w
.
co
m
).
RE
F
E
R
E
NC
E
S
[
1
]
Z.
L
i
,
F
.
Li
u
,
W
.
Y
a
n
g
,
S
.
P
e
n
g
,
a
n
d
J.
Z
h
o
u
,
“
A
s
u
r
v
e
y
o
f
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
:
a
n
a
l
y
si
s,
a
p
p
l
i
c
a
t
i
o
n
s,
a
n
d
p
r
o
sp
e
c
t
s,”
I
EEE
T
ra
n
s
a
c
t
i
o
n
s
o
n
N
e
u
r
a
l
N
e
t
w
o
rks
a
n
d
L
e
a
r
n
i
n
g
S
y
s
t
e
m
s
,
v
o
l
.
3
3
,
n
o
.
1
2
,
p
p
.
6
9
9
9
–
7
0
1
9
,
D
e
c
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
TN
N
LS.
2
0
2
1
.
3
0
8
4
8
2
7
.
[
2
]
M
.
C
a
r
r
a
n
z
a
-
G
a
r
c
í
a
,
J.
T
o
r
r
e
s
-
M
a
t
e
o
,
P
.
La
r
a
-
B
e
n
í
t
e
z
,
a
n
d
J
.
G
a
r
c
í
a
-
G
u
t
i
é
r
r
e
z
,
“
O
n
t
h
e
p
e
r
f
o
r
ma
n
c
e
o
f
o
n
e
st
a
g
e
a
n
d
t
w
o
s
t
a
g
e
o
b
j
e
c
t
d
e
t
e
c
t
o
r
s
i
n
a
u
t
o
n
o
m
o
u
s
v
e
h
i
c
l
e
s
u
s
i
n
g
c
a
m
e
r
a
d
a
t
a
,
”
Re
m
o
t
e
S
e
n
si
n
g
,
v
o
l
.
1
3
,
n
o
.
1
,
p
.
8
9
,
D
e
c
.
2
0
2
0
,
d
o
i
:
1
0
.
3
3
9
0
/
r
s1
3
0
1
0
0
8
9
.
[
3
]
S
.
S
.
A
.
Za
i
d
i
,
M
.
S
.
A
n
sar
i
,
A
.
A
sl
a
m,
N
.
K
a
n
w
a
l
,
M
.
A
sg
h
a
r
,
a
n
d
B
.
Le
e
,
“
A
su
r
v
e
y
o
f
m
o
d
e
r
n
d
e
e
p
l
e
a
r
n
i
n
g
b
a
se
d
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
m
o
d
e
l
s
,
”
D
i
g
i
t
a
l
S
i
g
n
a
l
Pr
o
c
e
ssi
n
g
:
A
Re
v
i
e
w
J
o
u
r
n
a
l
,
v
o
l
.
1
2
6
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
d
s
p
.
2
0
2
2
.
1
0
3
5
1
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
d
en
tifi
ca
tio
n
ty
p
es o
f p
la
n
t
u
s
in
g
co
n
vo
lu
tio
n
a
l n
e
u
r
a
l n
et
w
o
r
k
(
R
a
d
ityo
Hen
d
r
a
tmo
jo
J
a
ti N
o
to
n
eg
o
r
o
)
5835
[
4
]
X
.
Y
u
e
,
H
.
Li
,
M
.
S
h
i
m
i
z
u
,
S
.
K
a
w
a
mu
r
a
,
a
n
d
L
.
M
e
n
g
,
“
Y
O
LO
-
G
D
:
a
d
e
e
p
l
e
a
r
n
i
n
g
b
a
s
e
d
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
a
l
g
o
r
i
t
h
m
f
o
r
e
mp
t
y
d
i
s
h
r
e
c
y
c
l
i
n
g
s
r
o
b
o
t
s,
”
M
a
c
h
i
n
e
s
,
v
o
l
.
1
0
,
n
o
.
5
,
p
.
2
9
4
,
A
p
r
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
mac
h
i
n
e
s1
0
0
5
0
2
9
4
.
[
5
]
L.
F
a
n
g
e
t
a
l
.
,
“
U
s
i
n
g
c
h
a
n
n
e
l
a
n
d
n
e
t
w
o
r
k
l
a
y
e
r
p
r
u
n
i
n
g
b
a
se
d
o
n
d
e
e
p
l
e
a
r
n
i
n
g
r
e
a
l
t
i
me
d
e
t
e
c
t
i
o
n
o
f
g
i
n
g
e
r
i
ma
g
e
s
,
”
Ag
ri
c
u
l
t
u
r
e
,
v
o
l
.
1
1
,
n
o
.
1
2
,
p
.
1
1
9
0
,
N
o
v
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
i
c
u
l
t
u
r
e
1
1
1
2
1
1
9
0
.
[
6
]
Z.
C
a
o
a
n
d
R
.
Y
u
a
n
,
“
R
e
a
l
-
t
i
me
d
e
t
e
c
t
i
o
n
o
f
m
a
n
g
o
b
a
se
d
o
n
i
m
p
r
o
v
e
d
Y
O
LO
v
4
,
”
El
e
c
t
r
o
n
i
c
s
,
v
o
l
.
1
1
,
n
o
.
2
3
,
p
.
3
8
5
3
,
N
o
v
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
e
l
e
c
t
r
o
n
i
c
s
1
1
2
3
3
8
5
3
.
[
7
]
G
.
Y
a
n
g
,
J.
W
a
n
g
,
Z.
N
i
e
,
H
.
Y
a
n
g
,
a
n
d
S
.
Y
u
,
“
A
l
i
g
h
t
w
e
i
g
h
t
Y
O
LO
v
8
t
o
m
a
t
o
d
e
t
e
c
t
i
o
n
a
l
g
o
r
i
t
h
m
c
o
mb
i
n
i
n
g
f
e
a
t
u
r
e
e
n
h
a
n
c
e
me
n
t
a
n
d
a
t
t
e
n
t
i
o
n
,
”
Ag
r
o
n
o
m
y
,
v
o
l
.
1
3
,
n
o
.
7
,
p
.
1
8
2
4
,
J
u
l
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
o
n
o
m
y
1
3
0
7
1
8
2
4
.
[
8
]
P
.
L
i
,
J.
Z
h
e
n
g
,
P
.
Li
,
H
.
L
o
n
g
,
M
.
Li
,
a
n
d
L.
G
a
o
,
“
T
o
m
a
t
o
ma
t
u
r
i
t
y
d
e
t
e
c
t
i
o
n
a
n
d
c
o
u
n
t
i
n
g
m
o
d
e
l
b
a
se
d
o
n
M
H
S
A
-
Y
O
LO
v
8
,
”
S
e
n
so
rs
,
v
o
l
.
2
3
,
n
o
.
1
5
,
p
.
6
7
0
1
,
J
u
l
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
3
1
5
6
7
0
1
.
[
9
]
A
.
M
.
R
o
y
,
R
.
B
o
se,
a
n
d
J.
B
h
a
d
u
r
i
,
“
A
f
a
st
a
c
c
u
r
a
t
e
f
i
n
e
-
g
r
a
i
n
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
mo
d
e
l
b
a
s
e
d
o
n
Y
O
LO
v
4
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
N
e
u
ra
l
C
o
m
p
u
t
i
n
g
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
3
4
,
n
o
.
5
,
p
p
.
3
8
9
5
–
3
9
2
1
,
M
a
r
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
5
2
1
-
0
2
1
-
0
6
6
5
1
-
x.
[
1
0
]
S
.
Zh
a
o
,
Y
.
P
e
n
g
,
J.
L
i
u
,
a
n
d
S
.
W
u
,
“
To
ma
t
o
l
e
a
f
d
i
se
a
se
d
i
a
g
n
o
si
s
b
a
s
e
d
o
n
i
m
p
r
o
v
e
d
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
b
y
a
t
t
e
n
t
i
o
n
mo
d
u
l
e
,
”
A
g
r
i
c
u
l
t
u
r
e
,
v
o
l
.
1
1
,
n
o
.
7
,
p
.
6
5
1
,
J
u
l
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
i
c
u
l
t
u
r
e
1
1
0
7
0
6
5
1
.
[
1
1
]
D
.
Li
,
F
.
A
h
me
d
,
N
.
W
u
,
a
n
d
A
.
I
.
S
e
t
h
i
,
“
Y
O
LO
-
JD
:
a
d
e
e
p
l
e
a
r
n
i
n
g
n
e
t
w
o
r
k
f
o
r
j
u
t
e
d
i
se
a
ses
a
n
d
p
e
st
d
e
t
e
c
t
i
o
n
f
r
o
m
i
ma
g
e
s,”
Pl
a
n
t
s
,
v
o
l
.
1
1
,
n
o
.
7
,
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
p
l
a
n
t
s
1
1
0
7
0
9
3
7
.
[
1
2
]
Z.
X
u
e
,
R
.
X
u
,
D
.
B
a
i
,
a
n
d
H
.
L
i
n
,
“
Y
O
LO
-
Te
a
:
A
t
e
a
d
i
s
e
a
s
e
d
e
t
e
c
t
i
o
n
mo
d
e
l
i
m
p
r
o
v
e
d
b
y
Y
O
LO
v
5
,
”
Fo
r
e
st
s
,
v
o
l
.
1
4
,
n
o
.
2
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
f
1
4
0
2
0
4
1
5
.
[
1
3
]
M
.
K
h
a
l
i
d
,
M
.
S
a
r
f
r
a
z
,
U
.
I
q
b
a
l
,
M
.
A
f
t
a
b
,
G
.
N
i
e
d
b
a
ł
a
,
a
n
d
H
.
R
a
u
f
,
“
R
e
a
l
t
i
me
p
l
a
n
t
h
e
a
l
t
h
d
e
t
e
c
t
i
o
n
u
s
i
n
g
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
Ag
r
i
c
u
l
t
u
re
,
v
o
l
.
1
3
,
n
o
.
2
,
p
.
5
1
0
,
F
e
b
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
i
c
u
l
t
u
r
e
1
3
0
2
0
5
1
0
.
[
1
4
]
M
.
S
o
z
z
i
,
S
.
C
a
n
t
a
l
a
mess
a
,
A
.
C
o
g
a
t
o
,
A
.
K
a
y
a
d
,
a
n
d
F
.
M
a
r
i
n
e
l
l
o
,
“
A
u
t
o
ma
t
i
c
b
u
n
c
h
d
e
t
e
c
t
i
o
n
i
n
w
h
i
t
e
g
r
a
p
e
s
v
a
r
i
e
t
i
e
s
u
si
n
g
Y
O
LO
v
3
,
Y
O
LO
v
4
,
a
n
d
Y
O
LO
v
5
d
e
e
p
l
e
a
r
n
i
n
g
s
a
l
g
o
r
i
t
h
ms
,
”
Ag
r
o
n
o
m
y
,
v
o
l
.
1
2
,
n
o
.
2
,
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
o
n
o
m
y
1
2
0
2
0
3
1
9
.
[
1
5
]
M
.
B
u
z
z
y
,
V
.
T
h
e
sm
a
,
M
.
D
a
v
o
o
d
i
,
a
n
d
J
.
M
.
V
e
l
n
i
,
“
R
e
a
l
-
t
i
me
p
l
a
n
t
l
e
a
f
c
o
u
n
t
i
n
g
u
s
i
n
g
d
e
e
p
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
n
e
t
w
o
r
k
s,”
S
e
n
so
rs (S
w
i
t
zer
l
a
n
d
)
,
v
o
l
.
2
0
,
n
o
.
2
3
,
p
p
.
1
–
1
4
,
2
0
2
0
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
0
2
3
6
8
9
6
.
[
1
6
]
F
.
K
h
a
n
,
N
.
Z
a
f
a
r
,
M
.
N
.
T
a
h
i
r
,
M
.
A
q
i
b
,
S
.
S
a
l
e
e
m,
a
n
d
Z
.
H
a
r
o
o
n
,
“
D
e
e
p
l
e
a
r
n
i
n
g
b
a
s
e
d
a
p
p
r
o
a
c
h
f
o
r
w
e
e
d
d
e
t
e
c
t
i
o
n
i
n
p
o
t
a
t
o
c
r
o
p
s,”
i
n
T
h
e
1
st
I
n
t
e
r
n
a
t
i
o
n
a
l
Pr
e
c
i
si
o
n
A
g
r
i
c
u
l
t
u
r
e
P
a
k
i
s
t
a
n
C
o
n
f
e
re
n
c
e
2
0
2
2
(
PAP
C
2
0
2
2
)
&m
d
a
sh
;
C
h
a
n
g
e
t
h
e
C
u
l
t
u
r
e
o
f
Ag
ri
c
u
l
t
u
r
e
,
N
o
v
.
2
0
2
2
,
p
.
6
,
d
o
i
:
1
0
.
3
3
9
0
/
e
n
v
i
r
o
n
sci
p
r
o
c
2
0
2
2
0
2
3
0
0
6
.
[
1
7
]
A
.
N
a
si
r
a
h
ma
d
i
,
U
.
W
i
l
c
z
e
k
,
a
n
d
O
.
H
e
n
se
l
,
“
S
u
g
a
r
b
e
e
t
d
a
ma
g
e
d
e
t
e
c
t
i
o
n
d
u
r
i
n
g
h
a
r
v
e
st
i
n
g
u
si
n
g
d
i
f
f
e
r
e
n
t
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
m
o
d
e
l
s
,
”
Ag
r
i
c
u
l
t
u
re
,
v
o
l
.
1
1
,
n
o
.
1
1
,
p
.
1
1
1
1
,
N
o
v
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
i
c
u
l
t
u
r
e
1
1
1
1
1
1
1
1
.
[
1
8
]
J.
Y
a
o
,
J.
Q
i
,
J
.
Zh
a
n
g
,
H
.
S
h
a
o
,
J
.
Y
a
n
g
,
a
n
d
X
.
Li
,
“
A
r
e
a
l
t
i
me
d
e
t
e
c
t
i
o
n
a
l
g
o
r
i
t
h
m
f
o
r
k
i
w
i
f
r
u
i
t
d
e
f
e
c
t
s
b
a
sed
o
n
Y
O
LO
v
5
,
”
El
e
c
t
r
o
n
i
c
s
,
v
o
l
.
1
0
,
n
o
.
1
4
,
p
.
1
7
1
1
,
J
u
l
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
e
l
e
c
t
r
o
n
i
c
s1
0
1
4
1
7
1
1
.
[
1
9
]
G
.
Y
a
n
g
,
J.
W
a
n
g
,
Z.
N
i
e
,
H
.
Y
a
n
g
,
a
n
d
S
.
Y
u
,
“
A
l
i
g
h
t
w
e
i
g
h
t
Y
O
LO
v
8
t
o
m
a
t
o
d
e
t
e
c
t
i
o
n
a
l
g
o
r
i
t
h
m
c
o
mb
i
n
i
n
g
f
e
a
t
u
r
e
e
n
h
a
n
c
e
me
n
t
a
n
d
a
t
t
e
n
t
i
o
n
,
”
Ag
r
o
n
o
m
y
,
v
o
l
.
1
3
,
n
o
.
7
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
o
n
o
m
y
1
3
0
7
1
8
2
4
.
[
2
0
]
Y
.
W
a
n
g
,
C
.
Y
i
,
T.
H
u
a
n
g
,
a
n
d
J.
L
i
u
,
“
R
e
sea
r
c
h
o
n
i
n
t
e
l
l
i
g
e
n
t
r
e
c
o
g
n
i
t
i
o
n
f
o
r
p
l
a
n
t
p
e
st
a
n
d
d
i
se
a
se
b
a
s
e
d
o
n
i
m
p
r
o
v
e
d
Y
O
LO
v
8
mo
d
e
l
,
”
Ap
p
l
i
e
d
S
c
i
e
n
c
e
s (
S
w
i
t
z
e
rl
a
n
d
)
,
v
o
l
.
1
4
,
n
o
.
1
2
,
2
0
2
4
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
1
4
1
2
5
3
5
3
.
[
2
1
]
J.
Te
r
v
e
n
a
n
d
D
.
C
o
r
d
o
v
a
-
Esp
a
r
z
a
,
“
A
c
o
m
p
r
e
h
e
n
s
i
v
e
r
e
v
i
e
w
o
f
Y
O
LO
:
f
r
o
m
Y
O
LO
v
1
a
n
d
b
e
y
o
n
d
,
”
C
o
m
p
u
t
e
r
V
i
si
o
n
a
n
d
Pa
t
t
e
r
n
Re
c
o
g
n
i
t
i
o
n
,
p
p
.
1
–
3
4
,
2
0
2
3
,
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s/
2
3
0
4
.
0
0
5
0
1
.
[
2
2
]
A
.
A
b
o
a
h
,
B
.
W
a
n
g
,
U
.
B
a
g
c
i
,
a
n
d
Y
.
A
d
u
-
G
y
a
mf
i
,
“
R
e
a
l
t
i
me
m
u
l
t
i
c
l
a
ss h
e
l
me
t
v
i
o
l
a
t
i
o
n
d
e
t
e
c
t
i
o
n
u
s
i
n
g
f
e
w
sh
o
t
d
a
t
a
sam
p
l
i
n
g
t
e
c
h
n
i
q
u
e
a
n
d
Y
O
LO
v
8
,
”
I
E
EE
C
o
m
p
u
t
e
r
S
o
c
i
e
t
y
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
s
i
o
n
a
n
d
P
a
t
t
e
r
n
R
e
c
o
g
n
i
t
i
o
n
Wo
rk
sh
o
p
s
,
v
o
l
.
2
0
2
3
-
Ju
n
e
,
p
p
.
5
3
5
0
–
5
3
5
8
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
C
V
P
R
W
5
9
2
2
8
.
2
0
2
3
.
0
0
5
6
4
.
[
2
3
]
Y
.
Eg
i
,
M
.
H
a
j
y
z
a
d
e
h
,
a
n
d
E
.
E
y
c
e
y
u
r
t
,
“
D
r
o
n
e
c
o
mp
u
t
e
r
c
o
mm
u
n
i
c
a
t
i
o
n
b
a
s
e
d
t
o
ma
t
o
g
e
n
e
r
a
t
i
v
e
o
r
g
a
n
c
o
u
n
t
i
n
g
m
o
d
e
l
u
s
i
n
g
Y
O
LO
v
5
a
n
d
d
e
e
p
s
o
r
t
,
”
Ag
r
i
c
u
l
t
u
re
,
v
o
l
.
1
2
,
n
o
.
9
,
p
.
1
2
9
0
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
i
c
u
l
t
u
r
e
1
2
0
9
1
2
9
0
.
[
2
4
]
L.
C
h
e
n
,
S
.
Li
,
Q
.
B
a
i
,
J
.
Y
a
n
g
,
S
.
Ji
a
n
g
,
a
n
d
Y
.
M
i
a
o
,
“
R
e
v
i
e
w
o
f
i
ma
g
e
c
l
a
ssi
f
i
c
a
t
i
o
n
a
l
g
o
r
i
t
h
ms
b
a
s
e
d
o
n
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s,”
Re
m
o
t
e
S
e
n
si
n
g
,
v
o
l
.
1
3
,
n
o
.
2
2
,
p
.
4
7
1
2
,
N
o
v
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
r
s
1
3
2
2
4
7
1
2
.
[
2
5
]
T.
M
.
A
l
a
m
e
t
a
l
.
,
“
A
n
e
f
f
i
c
i
e
n
t
d
e
e
p
l
e
a
r
n
i
n
g
b
a
se
d
s
k
i
n
c
a
n
c
e
r
c
l
a
ss
i
f
i
e
r
f
o
r
a
n
i
mb
a
l
a
n
c
e
d
d
a
t
a
se
t
,
”
D
i
a
g
n
o
s
t
i
c
s
,
v
o
l
.
1
2
,
n
o
.
9
,
p
.
2
1
1
5
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
d
i
a
g
n
o
s
t
i
c
s1
2
0
9
2
1
1
5
.
[
2
6
]
M
.
A
.
A
s’
a
r
i
a
n
d
N
.
I
.
A
.
M
a
n
a
p
,
“
C
o
v
i
d
-
1
9
d
e
t
e
c
t
i
o
n
f
r
o
m
c
h
e
s
t
x
-
r
a
y
i
ma
g
e
s
:
c
o
m
p
a
r
i
so
n
o
f
w
e
l
l
-
e
s
t
a
b
l
i
s
h
e
d
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
mo
d
e
l
s,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
A
d
v
a
n
c
e
s
i
n
I
n
t
e
l
l
i
g
e
n
t
I
n
f
o
rm
a
t
i
c
s
,
v
o
l
.
8
,
n
o
.
2
,
p
.
2
2
4
,
Ju
l
.
2
0
2
2
,
d
o
i
:
1
0
.
2
6
5
5
5
/
i
j
a
i
n
.
v
8
i
2
.
8
0
7
.
[
2
7
]
T.
Ji
a
n
g
,
T.
L
i
u
,
W
.
D
o
n
g
,
Y
.
Li
u
,
a
n
d
Q
.
Z
h
a
n
g
,
“
S
e
c
u
r
i
t
y
r
i
sk
l
e
v
e
l
p
r
e
d
i
c
t
i
o
n
o
f
c
a
r
b
o
f
u
r
a
n
p
e
st
i
c
i
d
e
r
e
s
i
d
u
e
s
i
n
c
h
i
n
e
s
e
v
e
g
e
t
a
b
l
e
s
b
a
se
d
o
n
d
e
e
p
l
e
a
r
n
i
n
g
,
”
F
o
o
d
s
,
v
o
l
.
1
1
,
n
o
.
7
,
p
.
1
0
6
1
,
A
p
r
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
f
o
o
d
s1
1
0
7
1
0
6
1
.
[
2
8
]
J.
-
W
.
C
h
e
n
,
W
.
-
J
.
L
i
n
,
H
.
-
J
.
C
h
e
n
g
,
C
.
-
L
.
H
u
n
g
,
C
.
-
Y
.
L
i
n
,
a
n
d
S
.
-
P
.
C
h
e
n
,
“
A
s
m
a
r
t
p
h
o
n
e
b
a
s
e
d
a
p
p
l
i
c
a
t
i
o
n
f
o
r
s
c
a
l
e
p
e
s
t
d
e
t
e
c
t
i
o
n
u
s
i
n
g
m
u
l
t
i
p
l
e
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
m
e
t
h
o
d
s
,
”
E
l
e
c
t
r
o
n
i
c
s
,
v
o
l
.
1
0
,
n
o
.
4
,
p
.
3
7
2
,
F
e
b
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
e
l
e
c
t
r
o
n
i
c
s
1
0
0
4
0
3
7
2
.
[
2
9
]
V
.
P
o
n
n
u
s
a
my
,
A
.
C
o
u
mara
n
,
A
.
S
.
S
h
u
n
mu
g
a
m,
K
.
R
a
j
a
r
a
m
,
a
n
d
S
.
S
e
n
t
h
i
l
v
e
l
a
v
a
n
,
“
S
mar
t
g
l
a
ss:
R
e
a
l
t
i
me
l
e
a
f
d
e
sea
s
e
d
e
t
e
c
t
i
o
n
u
s
i
n
g
Y
O
LO
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
2
0
2
0
I
E
EE
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
o
m
m
u
n
i
c
a
t
i
o
n
a
n
d
S
i
g
n
a
l
Pr
o
c
e
ssi
n
g
,
I
C
C
S
P
2
0
2
0
,
2
0
2
0
,
p
p
.
1
1
5
0
–
1
1
5
4
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
C
S
P
4
8
5
6
8
.
2
0
2
0
.
9
1
8
2
1
4
6
.
[
3
0
]
H
.
C
h
e
n
,
G
.
Z
h
o
u
,
a
n
d
H
.
J
i
a
n
g
,
“
S
t
u
d
e
n
t
b
e
h
a
v
i
o
r
d
e
t
e
c
t
i
o
n
i
n
t
h
e
c
l
a
ssr
o
o
m
b
a
s
e
d
o
n
i
m
p
r
o
v
e
d
Y
O
LO
v
8
,
”
S
e
n
so
rs
,
v
o
l
.
2
3
,
n
o
.
2
0
,
p
.
8
3
8
5
,
O
c
t
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
3
2
0
8
3
8
5
.
[
3
1
]
G
.
O
h
a
n
d
S
.
Li
m
,
“
O
n
e
st
a
g
e
b
r
e
a
k
l
i
g
h
t
s
t
a
t
u
s
d
e
t
e
c
t
i
o
n
b
a
se
d
o
n
Y
O
LO
v
8
,
”
S
e
n
so
rs
,
v
o
l
.
2
3
,
n
o
.
1
7
,
p
.
7
4
3
6
,
A
u
g
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
3
1
7
7
4
3
6
.
[
3
2
]
M
.
H
u
ss
a
i
n
,
“
Y
O
LO
-
v
1
t
o
Y
O
LO
-
v
8
,
t
h
e
r
i
s
e
o
f
Y
O
LO
a
n
d
i
t
s
c
o
m
p
l
e
me
n
t
a
r
y
n
a
t
u
r
e
t
o
w
a
r
d
d
i
g
i
t
a
l
m
a
n
u
f
a
c
t
u
r
i
n
g
a
n
d
i
n
d
u
st
r
i
a
l
d
e
f
e
c
t
d
e
t
e
c
t
i
o
n
,
”
M
a
c
h
i
n
e
s
,
v
o
l
.
1
1
,
n
o
.
7
,
p
p
.
6
7
7
–
6
8
9
,
2
0
2
3
.
[
3
3
]
C
.
J
i
a
n
g
e
t
a
l
.
,
“
O
b
j
e
c
t
d
e
t
e
c
t
i
o
n
f
r
o
m
U
A
V
t
h
e
r
ma
l
i
n
f
r
a
r
e
d
i
m
a
g
e
s
a
n
d
v
i
d
e
o
s
u
s
i
n
g
Y
O
LO
mo
d
e
l
s,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
Ap
p
l
i
e
d
E
a
rt
h
O
b
ser
v
a
t
i
o
n
a
n
d
G
e
o
i
n
f
o
rm
a
t
i
o
n
,
v
o
l
.
1
1
2
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
a
g
.
2
0
2
2
.
1
0
2
9
1
2
.
[
3
4
]
X
.
W
a
n
g
,
H
.
G
a
o
,
Z.
J
i
a
,
a
n
d
Z.
Li
,
“
B
L
-
Y
O
LO
v
8
:
a
n
i
m
p
r
o
v
e
d
r
o
a
d
d
e
f
e
c
t
d
e
t
e
c
t
i
o
n
m
o
d
e
l
b
a
s
e
d
o
n
Y
O
LO
v
8
,
”
S
e
n
s
o
rs
(
B
a
se
l
,
S
w
i
t
z
e
rl
a
n
d
)
,
v
o
l
.
2
3
,
n
o
.
2
0
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
3
2
0
8
3
6
1
.
[
3
5
]
D
.
G
.
S
t
u
p
a
r
u
,
R
.
I
.
C
i
o
b
a
n
u
,
a
n
d
C
.
D
o
b
r
e
,
“
V
e
h
i
c
l
e
d
e
t
e
c
t
i
o
n
i
n
o
v
e
r
h
e
a
d
s
a
t
e
l
l
i
t
e
i
m
a
g
e
s
u
si
n
g
a
o
n
e
-
st
a
g
e
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
mo
d
e
l
,
”
S
e
n
s
o
rs (
S
w
i
t
z
e
rl
a
n
d
)
,
v
o
l
.
2
0
,
n
o
.
2
2
,
p
p
.
1
–
1
8
,
2
0
2
0
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
0
2
2
6
4
8
5
.
[
3
6
]
W
.
M
.
El
m
e
sser
y
e
t
a
l
.
,
“
Y
O
LO
b
a
sed
mo
d
e
l
f
o
r
a
u
t
o
ma
t
i
c
d
e
t
e
c
t
i
o
n
o
f
b
r
o
i
l
e
r
p
a
t
h
o
l
o
g
i
c
a
l
p
h
e
n
o
m
e
n
a
t
h
r
o
u
g
h
v
i
s
u
a
l
a
n
d
t
h
e
r
mal
i
m
a
g
e
s
i
n
i
n
t
e
r
n
s
i
v
e
p
o
u
l
t
r
y
h
o
u
s
e
s,
”
A
g
ri
c
u
l
t
u
r
e
,
v
o
l
.
1
3
,
n
o
.
8
,
p
.
1
5
2
7
,
J
u
l
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
i
c
u
l
t
u
r
e
1
3
0
8
1
5
2
7
.
[
3
7
]
F
.
Zh
o
u
,
H
.
D
e
n
g
,
Q
.
X
u
,
a
n
d
X
.
La
n
,
“
C
N
TR
-
Y
O
LO
:
i
m
p
r
o
v
e
d
Y
O
LO
v
5
b
a
se
d
o
n
c
o
n
v
n
e
x
t
a
n
d
t
r
a
n
sf
o
r
mer
f
o
r
a
i
r
c
r
a
f
t
d
e
t
e
c
t
i
o
n
i
n
r
e
mo
t
e
se
n
si
n
g
i
ma
g
e
s,
”
El
e
c
t
r
o
n
i
c
s
,
v
o
l
.
1
2
,
n
o
.
1
2
,
p
.
2
6
7
1
,
J
u
n
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
e
l
e
c
t
r
o
n
i
c
s
1
2
1
2
2
6
7
1
.
[
3
8
]
D
.
K
o
l
o
s
o
v
,
V
.
K
e
l
e
f
o
u
r
a
s,
P
.
K
o
u
r
t
e
ssi
s,
a
n
d
I
.
M
p
o
r
a
s,
“
A
n
a
t
o
m
y
of
d
e
e
p
l
e
a
r
n
i
n
g
i
ma
g
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
a
n
d
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
o
n
c
o
mm
e
r
c
i
a
l
e
d
g
e
d
e
v
i
c
e
s
:
a
c
a
se
s
t
u
d
y
o
n
f
a
c
e
mas
k
d
e
t
e
c
t
i
o
n
,
”
I
E
E
E
Ac
c
e
ss
,
v
o
l
.
1
0
,
p
p
.
1
0
9
1
6
7
–
1
0
9
1
8
6
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
2
.
3
2
1
4
2
1
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
6
,
Decem
b
e
r
20
25
:
5
8
2
7
-
5
8
3
6
5836
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Ra
d
ity
o
H
e
n
d
r
a
tm
o
jo
J
a
ti
No
t
o
n
e
g
o
r
o
h
o
ld
s
a
b
a
c
h
e
lo
r'
s
d
e
g
re
e
in
c
o
m
p
u
ter
sc
ien
c
e
(S
.
Ko
m
.
)
with
a
fo
c
u
s
o
n
in
f
o
rm
a
ti
o
n
sy
ste
m
s
a
n
d
is
c
u
rre
n
tl
y
p
u
rsu
in
g
a
m
a
ste
r’s
d
e
g
re
e
in
in
fo
rm
a
ti
o
n
sy
ste
m
s
m
a
n
a
g
e
m
e
n
t
(M
M
S
I),
sp
e
c
ializi
n
g
i
n
so
ftwa
re
in
f
o
rm
a
ti
o
n
sy
ste
m
s.
In
a
d
d
it
io
n
to
h
is
a
c
a
d
e
m
ic
b
a
c
k
g
ro
u
n
d
,
h
e
h
a
s
o
b
tain
e
d
v
a
rio
u
s
c
e
rti
fica
ti
o
n
s
a
n
d
p
ro
fe
ss
io
n
a
l
sk
il
ls.
He
is
c
u
rre
n
tl
y
e
m
p
lo
y
e
d
a
s
a
sta
ff
m
e
m
b
e
r
a
t
G
u
n
a
d
a
rm
a
Un
iv
e
rsity
.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
a
rti
ficia
l
in
telli
g
e
n
c
e
,
ima
g
e
p
r
o
c
e
ss
in
g
,
a
n
d
d
a
ta
sc
ien
c
e
.
H
e
c
a
n
b
e
re
a
c
h
e
d
v
ia em
a
il
:
ra
d
it
h
jn
@s
taff.
g
u
n
a
d
a
rm
a
.
a
c
.
id
.
H
u
stin
a
wa
t
y
e
a
rn
e
d
h
e
r
Do
c
to
ra
te
in
in
f
o
rm
a
ti
o
n
tec
h
n
o
lo
g
y
fro
m
G
u
n
a
d
a
rm
a
Un
iv
e
rsity
,
In
d
o
n
e
sia
,
in
2
0
1
2
,
a
fter
c
o
m
p
letin
g
h
e
r
M
M
S
I
d
e
g
re
e
a
t
th
e
sa
m
e
u
n
iv
e
rsity
in
1
9
9
4
.
Cu
rre
n
tl
y
,
sh
e
h
o
ld
s
th
e
p
o
siti
o
n
o
f
He
a
d
o
f
t
h
e
M
a
st
e
r
o
f
I
n
fo
rm
a
ti
o
n
S
y
ste
m
s
M
a
n
a
g
e
m
e
n
t
P
ro
g
ra
m
a
t
G
u
n
a
d
a
rm
a
Un
iv
e
rsity
.
He
r
re
se
a
rc
h
fo
c
u
se
s
o
n
in
f
o
rm
a
ti
o
n
sy
ste
m
s
m
a
n
a
g
e
m
e
n
t
a
n
d
tec
h
n
o
l
o
g
y
.
S
h
e
c
a
n
b
e
re
a
c
h
e
d
v
ia em
a
il
a
t
h
u
sti
n
a
@s
taff.g
u
n
a
d
a
rm
a
.
a
c
.
id
.
Evaluation Warning : The document was created with Spire.PDF for Python.