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
3
3
6
~
5
3
4
6
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
6
.
pp
5
3
3
6
-
5
3
4
6
5336
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Plant
disea
se
dete
ction a
nd clas
sifi
c
a
tion: bas
ed on m
a
chine
lea
rning
a
nd Eig
(
H
ess
)
-
co
-
o
ccurr
e
nce histog
ra
ms
of
oriented
g
ra
dients
E
l
Aro
us
s
i
El
M
eh
di
1
,
B
a
ra
k
a
t
L
a
t
if
a
2
,
Sil
ka
n
H
a
s
s
a
n
3
1
ELI
TES La
b
o
r
a
t
o
r
y
,
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
M
a
t
h
e
mat
i
c
s
,
H
i
g
h
e
r
S
c
h
o
o
l
o
f
T
e
c
h
n
o
l
o
g
y
,
C
h
o
u
a
i
b
D
o
u
k
k
a
l
i
U
n
i
v
e
r
s
i
t
y
,
El
J
a
d
i
d
a
,
M
o
r
o
c
c
o
2
M
a
n
a
g
e
me
n
t
o
f
S
u
st
a
i
n
a
b
l
e
A
g
r
i
c
u
l
t
u
r
e
L
a
b
o
r
a
t
o
r
y
,
H
i
g
h
e
r
S
c
h
o
o
l
o
f
Te
c
h
n
o
l
o
g
y
,
C
h
o
u
a
i
b
D
o
u
k
k
a
l
i
U
n
i
v
e
r
si
t
y
,
El
J
a
d
i
d
a
,
M
o
r
o
c
c
o
3
La
R
O
S
E
R
I
La
b
o
r
a
t
o
r
y
,
D
e
p
a
r
t
m
e
n
t
o
f
c
o
m
p
u
t
e
r
s
c
i
e
n
c
e
s
,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
s,
C
h
o
u
a
i
b
D
o
u
k
k
a
l
i
U
n
i
v
e
r
si
t
y
,
E
l
Ja
d
i
d
a
,
M
o
r
o
c
c
o
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
an
3
0
,
2
0
2
5
R
ev
is
ed
J
u
l 1
5
,
2
0
2
5
Acc
ep
ted
Sep
1
4
,
2
0
2
5
Ag
ricu
lt
u
ra
l
d
istri
c
ts
p
ro
v
id
e
h
ig
h
-
q
u
a
li
ty
f
o
o
d
a
n
d
c
o
n
tri
b
u
te
su
b
sta
n
ti
a
ll
y
to
e
c
o
n
o
m
ic
g
ro
wt
h
a
n
d
p
o
p
u
la
ti
o
n
s
u
p
p
o
rt
.
Ho
we
v
e
r,
p
lan
t
d
ise
a
se
s
c
a
n
d
irec
tl
y
re
d
u
c
e
fo
o
d
p
r
o
d
u
c
ti
o
n
a
n
d
th
re
a
ten
sp
e
c
ies
d
iv
e
rsity
.
T
h
e
u
se
o
f
p
re
c
ise
,
a
u
to
m
a
ted
d
e
tec
ti
o
n
tec
h
n
i
q
u
e
s
f
o
r
e
a
rly
d
ise
a
se
id
e
n
ti
fi
c
a
ti
o
n
c
a
n
imp
ro
v
e
f
o
o
d
q
u
a
li
t
y
a
n
d
m
it
ig
a
te
e
c
o
n
o
m
ic
lo
ss
e
s.
Ov
e
r
t
h
e
p
a
st
d
e
c
a
d
e
,
n
u
m
e
ro
u
s
m
e
th
o
d
s
h
a
v
e
b
e
e
n
p
r
o
p
o
se
d
fo
r
p
lan
t
d
ise
a
se
c
las
sifica
ti
o
n
,
a
n
d
in
re
c
e
n
t
y
e
a
rs
th
e
f
o
c
u
s
h
a
s
sh
ift
e
d
t
o
wa
rd
d
e
e
p
lea
rn
i
n
g
a
p
p
r
o
a
c
h
e
s
b
e
c
a
u
se
o
f
t
h
e
ir
o
u
tstan
d
i
n
g
p
e
rfo
rm
a
n
c
e
.
In
th
is
stu
d
y
,
we
e
m
p
lo
y
th
e
Ei
g
(He
ss
)
-
co
-
o
c
c
u
rre
n
c
e
h
ist
o
g
ra
m
s
o
f
o
rien
ted
g
ra
d
ien
ts
(
Co
HO
G
)
d
e
sc
rip
to
r
a
l
o
n
g
sid
e
p
re
-
train
e
d
m
a
c
h
in
e
-
lea
rn
in
g
m
o
d
e
ls
t
o
a
c
c
u
ra
tely
id
e
n
ti
f
y
v
a
ri
o
u
s
p
lan
t
d
ise
a
se
s.
We
a
p
p
ly
p
rin
c
i
p
a
l
c
o
m
p
o
n
e
n
t
a
n
a
ly
sis
(P
CA)
fo
r
d
ime
n
sio
n
a
li
ty
re
d
u
c
ti
o
n
,
t
h
e
re
b
y
e
n
h
a
n
c
in
g
c
o
m
p
u
tatio
n
a
l
e
fficie
n
c
y
a
n
d
o
v
e
ra
ll
m
o
d
e
l
p
e
rfo
rm
a
n
c
e
.
Ou
r
e
x
p
e
rime
n
ts
we
re
c
o
n
d
u
c
ted
o
n
th
e
p
o
p
u
lar
P
lan
tVi
ll
a
g
e
d
a
tab
a
se
,
wh
ich
c
o
n
tai
n
s
5
4
,
3
0
5
ima
g
e
s
a
c
ro
ss
3
8
d
ise
a
se
c
las
s
e
s.
We
e
v
a
lu
a
te
m
o
d
e
l
p
e
rfo
rm
a
n
c
e
u
sin
g
c
la
ss
ifi
c
a
ti
o
n
a
c
c
u
ra
c
y
,
se
n
siti
v
it
y
,
sp
e
c
ifi
c
it
y
,
a
n
d
F
1
-
sc
o
re
,
a
n
d
we
p
e
rfo
rm
a
c
o
m
p
a
ra
ti
v
e
a
n
a
l
y
sis
a
g
a
i
n
st
sta
t
e
-
of
-
th
e
-
a
rt
m
e
th
o
d
s.
T
h
e
fin
d
in
g
s
in
d
ica
te
th
a
t
th
e
a
p
p
ro
a
c
h
we
p
ro
p
o
s
e
d
a
c
h
iev
e
s
u
p
t
o
9
9
.
8
3
%
a
c
c
u
ra
c
y
,
o
u
t
p
e
rfo
rm
in
g
e
x
isti
n
g
m
o
d
e
ls.
Ad
d
it
i
o
n
a
ll
y
,
we
tes
t
t
h
e
ro
b
u
st
n
e
ss
o
f
o
u
r
m
e
th
o
d
u
n
d
e
r
v
a
ri
o
u
s
c
o
n
d
it
io
n
s
to
h
ig
h
li
g
h
t
it
s
p
o
ten
ti
a
l
fo
r
r
e
a
l
-
wo
rld
a
g
ricu
lt
u
ra
l
a
p
p
l
ica
ti
o
n
s.
K
ey
w
o
r
d
s
:
Ag
r
icu
ltu
r
al
Alex
Net
E
ig
(
Hess
)
-
co
-
o
cc
u
r
r
en
ce
h
is
to
g
r
am
s
o
f
o
r
ien
ted
g
r
ad
ien
ts
f
ea
tu
r
es
Plan
t d
is
ea
s
es r
ec
o
g
n
itio
n
Prin
cip
al
co
m
p
o
n
en
t a
n
al
y
s
is
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
:
E
l
Ar
o
u
s
s
i
E
l
Me
h
d
i
Dep
ar
tem
en
t o
f
C
o
m
p
u
ter
Sci
en
ce
an
d
Ma
th
e
m
atics Hig
h
er
Sch
o
o
l o
f
T
ec
h
n
o
lo
g
y
C
h
o
u
ai
b
Do
u
k
k
ali
Un
iv
er
s
ity
3
5
4
Salam
s
u
b
d
iv
is
io
n
2
4
0
0
0
E
l Jad
id
a,
Mo
r
o
cc
o
E
m
ail:
elar
o
u
s
s
i.e
@
u
cd
.
ac
.
m
a
1.
I
NT
RO
D
UCT
I
O
N
Ag
r
icu
ltu
r
e
h
as
b
ee
n
in
d
is
p
en
s
ab
le
to
th
e
s
m
o
o
t
h
f
u
n
ctio
n
in
g
o
f
th
e
g
lo
b
al
ec
o
n
o
m
y
,
p
r
o
v
i
d
in
g
b
asic
f
o
o
d
,
em
p
lo
y
m
e
n
t,
an
d
in
co
m
e.
I
t
r
e
p
r
esen
ts
ab
o
u
t
1
2
%
o
f
Mo
r
o
cc
o
’
s
n
atio
n
al
in
co
m
e
an
d
em
p
lo
y
s
ap
p
r
o
x
im
ately
3
1
%
o
f
th
e
p
o
p
u
latio
n
,
as
is
th
e
ca
s
e
in
v
ar
i
o
u
s
lo
w
an
d
m
id
d
le
-
in
co
m
e
c
o
u
n
tr
ies
th
at
d
ep
en
d
o
n
ag
r
icu
ltu
r
e
[
1
]
.
I
n
a
d
d
itio
n
to
its
ec
o
n
o
m
ic
im
p
o
r
tan
ce
,
ag
r
icu
ltu
r
e
is
ess
en
tial
f
o
r
en
s
u
r
in
g
f
o
o
d
s
tab
ilit
y
,
s
u
p
p
o
r
tin
g
r
u
r
al
c
o
m
m
u
n
ities
,
an
d
p
r
o
m
o
tin
g
s
o
cio
-
ec
o
n
o
m
ic
p
r
o
g
r
ess
.
I
ts
i
n
f
lu
en
ce
ex
ten
d
s
t
o
allied
in
d
u
s
tr
ies
s
u
ch
as
a
g
r
o
-
p
r
o
ce
s
s
in
g
an
d
s
u
p
p
ly
ch
ain
s
,
am
p
lify
in
g
t
h
eir
o
v
er
all
im
p
ac
t
o
n
n
atio
n
al
g
r
o
wth
.
Ov
er
th
e
p
ast
f
iv
e
y
ea
r
s
,
th
e
ag
r
icu
ltu
r
al
s
ec
to
r
’
s
co
n
tr
ib
u
tio
n
to
g
r
o
s
s
v
alu
e
ad
d
ed
(
GV
A)
to
th
e
Mo
r
o
cc
an
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
P
la
n
t d
is
ea
s
e
d
etec
tio
n
a
n
d
c
la
s
s
ifica
tio
n
:
b
a
s
ed
o
n
ma
ch
in
e
lea
r
n
in
g
a
n
d
…
(
E
l
A
r
o
u
s
s
i
El
M
eh
d
i
)
5337
th
e
ec
o
n
o
m
y
h
as
g
r
o
wn
f
r
o
m
1
1
.
1
6
%
[
1
]
,
u
n
d
er
s
co
r
in
g
its
p
iv
o
tal
r
o
le
in
d
r
iv
in
g
ec
o
n
o
m
ic
g
r
o
wth
an
d
en
s
u
r
in
g
f
o
o
d
s
ec
u
r
ity
.
Ho
w
ev
er
,
p
lan
t
d
is
ea
s
es
an
d
p
est
in
f
estatio
n
s
co
n
ti
n
u
e
t
o
p
o
s
e
m
ajo
r
ch
allen
g
es,
th
r
ea
ten
in
g
b
o
th
th
e
q
u
ality
a
n
d
q
u
a
n
tity
o
f
g
lo
b
al
f
o
o
d
p
r
o
d
u
ctio
n
a
n
d
d
estab
ilizin
g
ec
o
n
o
m
ies
r
elian
t
o
n
ag
r
icu
ltu
r
al
o
u
tp
u
ts
.
Pr
o
p
h
y
lactic
tr
ea
tm
en
ts
alo
n
e
a
r
e
in
ad
eq
u
ate,
as
th
e
y
d
o
n
o
t
f
u
lly
p
r
ev
en
t
ag
r
icu
ltu
r
a
l
ep
id
em
ics
an
d
en
d
em
ics.
C
o
m
b
in
ed
s
y
s
tem
s
f
o
r
ea
r
ly
i
d
en
tific
atio
n
,
ac
cu
r
ate
d
iag
n
o
s
is
,
an
d
r
ea
l
-
tim
e
m
an
ag
em
en
t
o
f
c
r
o
p
d
is
ea
s
es
esp
ec
ially
wh
en
co
m
b
in
ed
with
s
u
s
tain
ab
le
an
d
m
o
d
e
r
n
cr
o
p
p
r
o
tectio
n
p
r
ac
tices
h
av
e
t
h
e
p
o
ten
tial
t
o
en
s
u
r
e
n
o
t
o
n
ly
s
tab
le
p
r
o
d
u
ctio
n
an
d
p
r
ices,
b
u
t
also
lo
n
g
-
ter
m
ag
r
icu
ltu
r
al
r
esil
ien
ce
an
d
p
r
o
f
itab
ilit
y
.
C
las
s
if
icatio
n
is
a
cr
itical
co
m
p
o
n
e
n
t
o
f
au
to
m
ate
d
p
lan
t
d
is
ea
s
e
d
etec
tio
n
,
an
d
ac
cu
r
a
te,
tim
ely
r
ec
o
g
n
itio
n
o
f
illn
ess
is
es
s
e
n
tial
f
o
r
in
f
o
r
m
ed
d
ec
is
io
n
-
m
ak
in
g
in
ag
r
icu
ltu
r
al
p
r
o
d
u
cti
o
n
.
I
n
f
ec
ted
p
lan
ts
o
f
ten
d
is
p
lay
s
y
m
p
to
m
s
s
u
c
h
as
s
o
o
ty
m
ar
k
s
o
n
s
tem
s
,
f
r
u
it,
f
o
liag
e
,
o
r
f
l
o
wer
s
[
2
]
.
T
h
ese
d
is
tin
ctiv
e
m
ar
k
in
g
s
ca
n
ass
is
t
in
id
en
tify
in
g
ab
n
o
r
m
alities
.
Ho
wev
e
r
,
d
iag
n
o
s
in
g
p
lan
t
d
is
ea
s
es
ac
cu
r
ately
r
eq
u
i
r
es
s
p
ec
ialized
k
n
o
wled
g
e
an
d
co
n
s
id
er
ab
le
h
u
m
an
r
eso
u
r
ce
s
,
an
d
m
a
n
u
al
an
al
y
s
is
ca
n
b
e
s
u
b
jectiv
e
an
d
tim
e
-
co
n
s
u
m
in
g
.
Misd
iag
n
o
s
es
b
y
f
ar
m
er
s
o
r
s
p
ec
ialis
ts
ca
n
lead
to
in
ap
p
r
o
p
r
iate
tr
ea
tm
en
ts
,
wh
ich
m
a
y
d
am
ag
e
cr
o
p
q
u
ality
an
d
y
ield
an
d
e
v
en
co
n
tam
in
ate
th
e
en
v
ir
o
n
m
en
t
if
in
co
r
r
ec
t
ch
e
m
icals
ar
e
ap
p
lied
[
3
]
.
R
ec
en
t
ad
v
an
ce
s
in
m
ac
h
in
e
lear
n
i
n
g
an
d
d
ee
p
lear
n
in
g
h
av
e
r
ev
o
lu
tio
n
ized
d
iag
n
o
s
tic
tech
n
iq
u
e
f
o
r
p
la
n
t d
is
ea
s
es.
A
u
t
o
m
a
t
e
d
d
i
g
es
t
i
o
n
a
n
d
f
e
a
tu
r
e
e
x
t
r
a
c
t
i
o
n
h
a
v
e
b
e
c
o
m
e
m
o
r
e
a
c
c
e
s
s
i
b
l
e
,
e
n
a
b
li
n
g
m
o
r
e
a
c
c
u
r
a
te
i
m
a
g
e
-
b
a
s
e
d
r
e
p
r
es
e
n
t
at
i
o
n
s
o
f
d
i
s
e
as
e
s
y
m
p
t
o
m
s
.
T
h
e
r
e
c
en
t
a
c
c
es
s
i
b
i
li
t
y
o
f
l
a
r
g
e
i
m
a
g
e
d
a
t
a
b
a
s
e
s
,
p
o
w
e
r
f
u
l
G
P
Us
,
a
n
d
a
d
v
a
n
c
e
d
c
o
m
p
u
tin
g
s
o
f
t
w
a
r
e
t
h
a
t
a
r
e
c
o
m
p
u
t
a
ti
o
n
a
l
l
y
l
ess
d
e
m
a
n
d
i
n
g
h
a
s
p
r
o
m
p
t
e
d
a
t
r
a
n
s
i
t
i
o
n
f
r
o
m
c
l
a
s
s
ic
m
e
t
h
o
d
s
t
o
m
o
d
e
r
n
d
e
e
p
l
e
a
r
n
i
n
g
-
b
a
s
e
d
f
r
a
m
e
w
o
r
k
s
.
W
h
il
e
c
u
r
r
e
n
t
m
o
d
e
l
s
s
h
o
w
e
n
c
o
u
r
a
g
i
n
g
r
e
s
u
l
ts
o
n
s
p
e
c
i
f
i
c
d
a
ta
s
et
s
[
4
]
,
t
h
e
y
a
r
e
t
y
p
i
c
al
l
y
t
r
ai
n
e
d
o
n
i
m
a
g
e
s
wi
t
h
s
i
m
p
l
e
r
b
a
c
k
g
r
o
u
n
d
s
l
i
m
it
i
n
g
t
h
ei
r
a
p
p
l
i
c
a
t
i
o
n
i
n
r
e
a
l
a
g
r
i
c
u
lt
u
r
a
l
s
e
tt
i
n
g
s
w
h
e
r
e
t
h
e
d
i
v
e
r
s
i
t
y
o
f
i
m
a
g
e
s
a
n
d
c
o
m
p
l
e
x
it
y
o
f
b
a
c
k
g
r
o
u
n
d
s
i
s
h
i
g
h
.
D
i
v
e
r
s
i
f
i
c
a
ti
o
n
a
n
d
r
e
a
li
s
m
o
f
a
m
o
d
e
l'
s
t
r
ai
n
i
n
g
d
a
t
as
e
ts
i
s
n
e
c
e
s
s
a
r
y
t
o
i
m
p
r
o
v
e
g
e
n
e
r
a
l
iz
at
i
o
n
a
n
d
c
o
n
f
i
d
e
n
c
e
i
n
p
r
e
d
i
c
t
i
o
n
s
[
5
]
–
[
7
]
.
T
h
e
n
o
v
e
l
c
o
n
t
r
i
b
u
t
i
o
n
o
f
t
h
i
s
m
a
n
u
s
c
r
i
p
t
i
s
t
h
e
c
r
ea
t
i
o
n
o
f
a
n
o
v
e
l
h
y
b
r
i
d
m
o
d
e
l
f
o
r
d
e
t
e
c
t
i
o
n
a
n
d
c
l
a
s
s
i
f
i
c
a
ti
o
n
o
f
p
l
a
n
t
d
i
s
e
as
es
.
U
n
l
i
k
e
c
o
n
v
e
n
ti
o
n
a
l
a
p
p
r
o
a
c
h
e
s
t
h
a
t
r
e
l
y
s
o
l
el
y
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
(
C
N
Ns
)
,
o
u
r
m
e
t
h
o
d
c
o
m
b
i
n
e
s
d
e
e
p
f
e
at
u
r
es
e
x
t
r
ac
t
e
d
v
i
a
a
n
i
m
p
r
o
v
e
d
Al
e
x
N
e
t
a
r
c
h
i
t
ec
t
u
r
e
w
i
t
h
h
a
n
d
c
r
a
f
t
e
d
d
e
s
c
r
i
p
t
o
r
s
o
b
t
a
i
n
e
d
f
r
o
m
t
h
e
E
i
g
(
H
e
s
s
)
-
co
-
o
c
c
u
r
r
e
n
c
e
h
i
s
t
o
g
r
a
m
o
f
o
r
i
e
n
t
e
d
g
r
a
d
i
e
n
t
s
(
C
o
H
OG
)
a
l
g
o
r
i
t
h
m
,
w
h
i
c
h
c
a
p
t
u
r
e
s
f
i
n
e
g
e
o
m
et
r
i
c
a
n
d
t
e
x
t
u
r
a
l
c
h
a
r
a
ct
e
r
is
t
i
cs
.
T
o
b
e
t
t
e
r
c
o
m
p
u
t
a
ti
o
n
a
l
e
f
f
i
c
a
c
y
,
we
c
a
r
r
y
o
u
t
p
r
i
n
c
i
p
a
l
c
o
m
p
o
n
e
n
t
a
n
al
y
s
is
(
PC
A
)
f
o
r
d
i
m
e
n
s
i
o
n
a
l
i
t
y
r
e
d
u
c
t
i
o
n
,
r
e
t
a
i
n
i
n
g
e
s
s
e
n
ti
a
l
i
n
f
o
r
m
a
t
i
o
n
w
h
i
l
e
m
i
n
i
m
i
z
i
n
g
r
e
d
u
n
d
a
n
c
y
.
T
h
i
s
f
u
s
i
o
n
o
f
d
e
e
p
a
n
d
h
a
n
d
c
r
a
f
t
e
d
f
e
a
t
u
r
e
s
i
n
to
a
c
o
m
p
a
c
t
v
e
c
t
o
r
e
n
h
a
n
c
e
s
t
h
e
m
o
d
e
l
’
s
r
o
b
u
s
t
n
es
s
a
n
d
g
e
n
e
r
a
l
i
z
at
i
o
n
c
a
p
a
b
i
l
ity
,
p
a
r
t
i
c
u
l
a
r
l
y
o
n
i
m
a
g
e
s
c
a
p
t
u
r
e
d
u
n
d
e
r
r
e
a
l
-
w
o
r
l
d
c
o
n
d
i
t
i
o
n
s
.
F
u
r
t
h
e
r
m
o
r
e
,
we
v
a
l
i
d
a
t
e
o
u
r
m
o
d
e
l
o
n
t
w
o
co
n
t
r
a
s
t
i
n
g
d
a
t
as
e
ts
Pl
a
n
t
Vi
l
lag
e
a
n
d
P
l
a
n
t
D
is
e
as
e
d
e
m
o
n
s
t
r
a
t
i
n
g
p
e
r
f
o
r
m
a
n
c
e
s
u
p
e
r
i
o
r
t
o
t
h
e
s
ta
t
e
o
f
t
h
e
a
r
t
,
w
i
t
h
a
c
c
u
r
a
c
y
r
e
ac
h
i
n
g
9
9
.
8
3
%
,
w
h
i
le
m
a
i
n
t
ai
n
i
n
g
s
t
a
b
il
i
t
y
a
c
r
o
s
s
v
a
r
y
i
n
g
a
c
q
u
i
s
it
i
o
n
c
o
n
d
i
t
i
o
n
s
.
T
h
e
r
e
s
t
o
f
t
h
is
p
a
p
e
r
i
s
s
t
r
u
c
t
u
r
e
d
as
f
o
ll
o
ws:
s
e
c
t
i
o
n
2
p
r
e
v
i
o
u
s
s
t
u
d
i
es
w
o
r
k
i
n
t
h
e
r
a
n
g
e
o
f
c
o
n
t
e
n
t
-
b
a
s
e
d
i
m
a
g
e
r
e
t
r
i
e
v
al
(
C
B
I
R
)
.
S
e
c
ti
o
n
3
o
u
t
l
i
n
e
s
t
h
e
p
r
o
p
o
s
e
d
m
e
t
h
o
d
.
S
e
c
t
i
o
n
4
s
h
o
w
s
t
h
e
f
i
n
d
i
n
g
s
f
r
o
m
t
h
e
e
x
p
e
r
i
m
e
n
t
,
w
h
i
c
h
a
r
e
c
o
m
p
a
r
e
d
t
o
e
x
i
s
t
i
n
g
a
p
p
r
o
a
c
h
e
s
i
n
s
e
c
ti
o
n
5
.
A
t
l
a
s
t
,
s
e
ct
i
o
n
6
c
o
n
c
l
u
d
e
s
t
h
e
p
a
p
e
r
a
n
d
d
i
s
c
u
s
s
e
s
p
r
o
s
p
e
c
t
i
v
e
s
t
u
d
i
e
s
.
2.
RE
L
AT
E
D
WO
RK
T
r
ad
itio
n
al
m
eth
o
d
s
to
d
iag
n
o
s
e
p
lan
t
d
is
ea
s
es
b
ased
o
n
v
is
u
al
ev
alu
atio
n
a
r
e
s
u
b
jectiv
e,
ex
p
en
s
iv
e,
tim
e
co
n
s
u
m
in
g
an
d
r
eq
u
i
r
e
co
n
s
id
er
ab
le
m
an
u
al
lab
o
u
r
.
T
h
ese
lim
itatio
n
s
h
av
e
d
r
iv
en
r
esear
ch
er
s
to
s
ee
k
m
o
r
e
ef
f
e
ctiv
e
s
o
lu
tio
n
s
.
Ma
n
y
s
tu
d
ies
n
o
w
r
ely
o
n
m
ac
h
in
e
lear
n
in
g
m
eth
o
d
s
to
ac
h
i
ev
e
h
ig
h
ac
cu
r
ac
y
,
r
ed
u
ce
d
c
o
s
ts
,
an
d
g
r
ea
ter
o
b
jectiv
ity
.
I
n
th
is
s
ec
tio
n
,
we
r
e
v
iew
k
ey
wo
r
k
s
in
th
is
ar
ea
a
n
d
h
ig
h
lig
h
t
r
ec
e
n
t
d
ev
elo
p
m
e
n
ts
.
E
ar
ly
ef
f
o
r
ts
a
p
p
lied
s
h
allo
w
C
NNs
with
f
o
u
r
to
s
ix
lay
e
r
s
to
v
ar
io
u
s
cr
o
p
s
,
ex
p
lo
itin
g
th
eir
f
lex
ib
ilit
y
an
d
r
o
b
u
s
tn
ess
.
Fo
r
ex
am
p
le,
Mo
h
an
ty
et
a
l.
[
8
]
p
r
o
p
o
s
ed
an
alg
o
r
ith
m
f
o
r
b
o
t
h
class
if
icatio
n
an
d
p
r
ep
r
o
ce
s
s
in
g
o
f
r
ice
p
lan
t
im
ag
es,
an
d
Fer
en
tin
o
s
[
9
]
id
en
tifie
d
d
is
ea
s
es
in
r
ice
f
ield
s
.
C
h
en
et
a
l.
[
1
0
]
im
p
r
o
v
e
d
u
p
o
n
th
ese
m
eth
o
d
s
b
y
co
m
b
in
in
g
C
NNs
with
tr
an
s
f
er
lear
n
in
g
s
tr
ateg
ies
u
s
in
g
ar
ch
itectu
r
es
s
u
ch
as M
o
b
ileNet
an
d
I
n
ce
p
tio
n
V
3
f
o
r
e
n
h
an
ce
d
class
if
icatio
n
,
r
ec
o
g
n
itio
n
,
an
d
s
eg
m
en
tatio
n
o
f
p
lan
t
d
is
ea
s
es.
B
as
e
d
o
n
s
e
l
e
ct
e
d
g
r
a
p
e
v
i
n
e
an
d
t
o
m
a
t
o
d
is
e
a
s
e
s
f
r
o
m
t
h
e
Pl
a
n
t
V
il
l
a
g
e
d
at
a
b
a
s
e
[
1
1
]
,
K
a
r
t
h
i
k
e
t
a
l
.
[
1
2
]
e
m
p
l
o
y
e
d
a
m
u
l
t
i
-
cl
a
s
s
cla
s
s
i
f
i
c
at
i
o
n
a
p
p
r
o
a
c
h
u
s
i
n
g
VG
G
1
6
w
it
h
t
r
a
n
s
f
e
r
l
ea
r
n
i
n
g
.
B
a
r
b
e
d
o
[
1
3
]
s
u
r
v
e
y
e
d
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
m
e
t
h
o
d
s
f
o
r
c
r
o
p
d
i
s
ea
s
e
c
l
as
s
i
f
i
c
at
i
o
n
,
ex
a
m
i
n
i
n
g
s
u
p
e
r
v
is
e
d
a
p
p
r
o
a
ch
e
s
i
n
cl
u
d
i
n
g
n
a
ï
v
e
B
a
y
e
s
(
NB
)
,
d
i
s
c
r
i
m
i
n
a
n
t
a
n
a
l
y
s
is
(
D
A
)
,
k
-
n
e
a
r
es
t
n
ei
g
h
b
o
r
s
(k
-
NN)
,
s
u
p
p
o
r
t
v
e
c
t
o
r
m
a
c
h
i
n
e
s
(
S
V
Ms
)
,
d
e
c
i
s
i
o
n
t
r
ees
(
D
T
s
)
,
a
n
d
r
an
d
o
m
f
o
r
e
s
t
s
(
R
Fs
)
.
V
a
r
d
h
i
n
i
e
t
a
l
.
[
1
4
]
c
o
m
p
a
r
e
d
t
h
es
e
w
i
t
h
u
n
s
u
p
e
r
v
is
e
d
t
e
c
h
n
i
q
u
e
s
K
-
m
ea
n
s
c
l
u
s
t
e
r
i
n
g
,
f
u
z
z
y
c
l
u
s
t
e
r
i
n
g
,
a
n
d
G
a
u
s
s
i
an
m
i
x
t
u
r
e
m
o
d
e
l
s
a
s
w
e
ll
a
s
p
r
e
-
t
r
a
i
n
e
d
C
N
Ns
l
i
k
e
A
l
e
x
N
e
t
a
n
d
G
o
o
g
l
e
N
e
t
f
o
r
s
o
y
b
e
a
n
d
i
s
e
a
s
e
cl
a
s
s
i
f
i
c
a
ti
o
n
.
T
h
e
e
m
e
r
g
e
n
c
e
o
f
g
e
n
e
r
a
t
i
v
e
n
e
t
w
o
r
k
s
h
a
s
f
u
r
t
h
er
a
c
c
e
l
e
r
a
te
d
d
a
t
a
a
c
q
u
i
s
i
t
i
o
n
b
y
o
v
e
r
c
o
m
i
n
g
t
r
a
d
i
t
i
o
n
a
l
c
o
s
t
an
d
t
i
m
e
c
o
n
s
t
r
a
i
n
ts
.
M
o
n
d
a
l
e
t
a
l
.
[
1
5
]
p
r
e
s
e
n
t
e
d
a
f
r
a
m
e
w
o
r
k
c
o
m
b
i
n
i
n
g
m
o
d
e
l
-
a
g
n
o
s
t
i
c
m
e
t
a
-
l
e
a
r
n
i
n
g
(
M
A
ML
)
w
i
t
h
D
e
e
p
L
a
b
V
3
f
o
r
l
e
a
f
d
i
s
e
as
e
a
n
a
l
y
s
is
a
n
d
e
s
t
i
m
a
t
i
o
n
,
i
m
p
r
o
v
i
n
g
g
e
n
e
r
a
l
iz
a
t
i
o
n
o
n
l
i
m
it
e
d
d
a
ta
s
et
s
.
A
co
n
d
i
t
i
o
n
a
l
g
e
n
e
r
a
ti
v
e
a
d
v
e
r
s
a
r
ia
l
n
e
tw
o
r
k
(
c
GA
N
)
w
a
s
p
r
o
p
o
s
e
d
b
y
A
b
b
a
s
et
a
l
.
[
1
6
]
t
o
e
n
h
a
n
c
e
m
u
l
t
i
-
l
e
a
f
c
la
s
s
if
i
c
a
t
i
o
n
b
a
s
e
d
o
n
r
e
f
e
r
e
n
c
e
d
a
ta
.
I
n
a
n
o
t
h
e
r
s
t
u
d
y
,
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
3
3
6
-
5
3
4
6
5338
a
p
r
e
-
t
r
a
i
n
e
d
M
o
b
i
l
e
N
et
C
N
N
a
c
h
i
e
v
e
d
9
6
.
5
8
%
c
l
a
s
s
i
f
i
c
at
i
o
n
a
c
c
u
r
a
c
y
f
o
r
t
o
m
a
t
o
l
e
a
f
d
i
s
e
a
s
es
.
H
u
g
h
e
s
a
n
d
S
a
l
a
t
h
é
[
1
7
]
d
e
m
o
n
s
t
r
a
t
e
d
t
h
a
t
i
n
c
o
r
p
o
r
a
t
i
n
g
a
w
ei
g
h
t
e
d
a
v
er
a
g
e
f
u
s
i
o
n
o
f
g
a
m
m
a
c
o
r
r
e
ct
io
n
a
n
d
m
u
l
t
i
-
s
c
al
e
r
e
t
i
n
e
x
w
it
h
c
o
l
o
r
r
e
s
t
o
r
a
t
i
o
n
(
M
S
R
C
R
)
p
r
e
p
r
o
c
e
s
s
i
n
g
i
m
p
r
o
v
e
d
I
n
c
e
p
t
i
o
n
V
3
’
s
r
e
c
o
g
n
i
ti
o
n
r
a
t
e
b
y
2
9
.
7
1
%
.
R
e
h
m
a
n
et
a
l
.
[
1
8
]
i
n
t
r
o
d
u
c
e
d
f
u
s
i
o
n
t
e
c
h
n
i
q
u
es
a
d
a
p
ti
n
g
D
en
s
e
N
e
t
-
121
[
1
9
]
t
o
a
c
c
e
p
t b
o
t
h
r
a
w
a
n
d
s
e
g
m
e
n
t
e
d
l
e
a
f
i
m
a
g
es
.
P
a
l
a
n
d
K
u
m
a
r
[
2
0
]
d
e
v
e
l
o
p
e
d
A
g
r
i
D
e
t
,
a
d
e
e
p
l
e
a
r
n
i
n
g
–
b
a
s
e
d
d
e
t
e
c
tio
n
f
r
a
m
e
w
o
r
k
f
o
r
q
u
a
n
t
i
f
y
i
n
g
d
i
s
e
a
s
e
s
e
v
e
r
it
y
o
n
p
l
a
n
t
l
e
a
v
e
s
.
S
u
n
il
e
t
a
l
.
[
2
1
]
p
r
o
p
o
s
e
d
a
m
u
l
t
i
-
l
e
v
e
l
f
e
a
tu
r
e
f
u
s
i
o
n
n
e
t
w
o
r
k
b
a
s
e
d
o
n
R
es
N
et
-
5
0
w
it
h
a
n
a
d
a
p
t
i
v
e
at
t
e
n
ti
o
n
m
e
c
h
a
n
is
m
,
e
n
a
b
l
i
n
g
d
e
e
p
e
r
i
n
s
i
g
h
t
s
a
t
ch
a
n
n
e
l
,
s
p
at
i
a
l
,
a
n
d
p
i
x
e
l
l
e
v
e
ls
.
F
i
n
a
ll
y
,
A
n
h
a
n
d
D
u
c
[
2
2
]
r
e
v
i
e
w
e
d
1
6
0
s
t
u
d
i
es
a
c
r
o
s
s
5
0
p
l
a
n
t
l
e
a
f
d
at
a
s
et
s
,
a
n
a
l
y
z
i
n
g
m
o
d
e
r
n
d
e
e
p
l
e
a
r
n
i
n
g
m
o
d
e
l
s
(
A
l
e
x
Ne
t
[
2
3
]
,
G
o
o
g
l
e
N
e
t
[
2
4
]
,
R
e
s
N
e
t
[
2
5
]
,
M
o
b
i
l
e
N
e
t
[
2
6
]
,
S
q
u
e
e
z
e
N
e
t
[
2
7
]
,
a
n
d
Y
O
L
O
[
2
8
]
)
,
w
h
o
s
e
c
o
n
t
r
i
b
u
t
i
o
n
s
h
a
v
e
b
e
e
n
s
i
g
n
i
f
i
c
a
n
t
in
p
r
o
v
i
d
i
n
g
r
e
l
i
a
b
l
e
a
n
d
s
c
a
la
b
l
e
p
l
a
n
t
d
i
s
e
a
s
e
d
i
a
g
n
o
s
t
i
cs
.
3.
M
AT
E
R
I
AL
S AN
D
M
E
T
H
O
DS
3
.
1
.
Appro
a
ches a
nd
r
eso
urce
s
T
h
is
r
esear
ch
p
r
o
p
o
s
es
an
ef
f
icien
t
co
n
ce
p
tu
al
f
r
a
m
ewo
r
k
w
ith
in
an
ex
p
er
im
e
n
tal
lear
n
in
g
p
ar
ad
ig
m
to
en
ab
le
r
eliab
le,
ad
ap
tab
le
d
ep
lo
y
m
e
n
t
o
f
p
lan
t
d
is
ea
s
e
–
id
en
tific
atio
n
m
o
d
els
in
r
ea
l
-
wo
r
ld
s
ettin
g
s
.
Ou
r
wo
r
k
lev
er
ag
es
d
ata
f
r
o
m
d
i
v
er
s
e
s
o
u
r
ce
s
r
an
g
in
g
f
r
o
m
c
u
r
ated
,
h
ig
h
-
q
u
ality
im
ag
es
to
r
aw,
f
ield
-
ca
p
tu
r
e
d
s
am
p
les
to
ass
es
s
an
d
en
h
a
n
ce
m
o
d
el
g
en
er
aliza
tio
n
.
W
e
ap
p
ly
a
s
u
ite
o
f
p
r
e
p
r
o
ce
s
s
in
g
tech
n
iq
u
es
(
n
o
r
m
aliza
tio
n
,
r
esizin
g
,
an
d
au
g
m
en
tatio
n
v
ia
f
lip
p
in
g
,
r
o
t
atio
n
,
an
d
b
r
ig
h
tn
ess
ad
j
u
s
tm
en
t)
to
s
tan
d
ar
d
ize
th
e
in
p
u
t
d
ata,
r
ed
u
ce
d
o
m
ain
in
co
n
s
is
ten
cies,
an
d
in
cr
ea
s
e
t
h
e
m
o
d
el’
s
r
o
b
u
s
tn
ess
ac
r
o
s
s
v
ar
y
in
g
co
n
d
itio
n
s
.
T
o
p
r
o
m
o
t
e
ac
cu
r
ate
f
ea
t
u
r
e
e
x
t
r
a
cti
o
n
a
n
d
cl
ass
if
ic
ati
o
n
,
w
e
o
p
ti
m
iz
e
o
u
r
d
ee
p
le
ar
n
i
n
g
a
r
c
h
it
ec
t
u
r
e
an
en
h
an
ce
d
A
le
x
Ne
t
C
NN
w
ith
r
e
d
u
c
e
d
c
o
m
p
l
e
x
it
y
a
n
d
f
a
s
ter
c
o
m
p
u
ta
ti
o
n
a
n
d
c
o
m
p
l
e
m
e
n
t
i
t
wit
h
a
n
ew
h
a
n
d
c
r
a
f
t
ed
d
esc
r
i
p
t
o
r
,
E
i
g
(
H
ess
)
-
C
o
HOG
,
w
h
ic
h
ca
p
t
u
r
es
r
i
ch
g
eo
m
e
tr
ic
an
d
te
x
t
u
r
al
in
f
o
r
m
a
ti
o
n
f
r
o
m
g
r
a
y
s
ca
le
i
m
a
g
es.
W
e
t
h
e
n
a
p
p
l
y
PC
A
t
o
c
o
m
p
r
ess
t
h
e
h
i
g
h
-
d
i
m
e
n
s
i
o
n
al
E
i
g
(
Hess
)
-
C
o
HOG
v
ec
to
r
i
n
t
o
a
1
×
5
9
r
e
p
r
es
en
tati
o
n
,
p
r
ese
r
v
i
n
g
its
d
is
cr
im
i
n
ati
v
e
p
o
we
r
w
h
i
le
r
e
d
u
ci
n
g
r
ed
u
n
d
a
n
c
y
.
T
h
is
PC
A
-
r
e
d
u
ce
d
v
ec
t
o
r
is
c
o
n
ca
te
n
a
te
d
w
it
h
t
h
e
1
×
6
4
d
e
e
p
f
ea
tu
r
e
v
e
ct
o
r
,
p
r
o
d
u
c
in
g
a
c
o
m
p
ac
t
1
×
1
2
8
f
ea
tu
r
e
r
ep
r
es
e
n
tat
io
n
t
h
at
in
t
eg
r
ates
l
o
w
-
le
v
e
l
h
a
n
d
c
r
a
f
t
ed
f
ea
t
u
r
es
wit
h
h
ig
h
-
l
ev
el
le
ar
n
e
d
f
ea
t
u
r
es
.
T
h
is
h
y
b
r
i
d
a
p
p
r
o
a
c
h
o
v
e
r
c
o
m
es
C
NNs’
t
y
p
i
ca
l
i
n
a
b
il
it
y
t
o
ca
p
t
u
r
e
f
i
n
e
s
tr
u
c
tu
r
al
p
a
tte
r
n
s
w
it
h
o
u
t
s
ac
r
i
f
i
ci
n
g
t
h
e
e
f
f
ici
en
c
y
o
f
d
ee
p
ar
ch
ite
ct
u
r
es.
W
e
e
v
al
u
at
e
th
e
m
o
d
el
u
s
i
n
g
c
o
m
p
r
e
h
e
n
s
i
v
e
m
et
r
i
cs
ac
cu
r
a
c
y
,
p
r
ec
is
io
n
,
r
ec
a
ll,
F
1
-
s
c
o
r
e,
an
d
AUC
-
R
OC
ac
r
o
s
s
b
o
t
h
c
o
n
tr
o
lle
d
(
Pla
n
t
Vill
a
g
e
)
a
n
d
r
e
al
w
o
r
ld
(
P
la
n
tD
is
ea
s
e
)
d
at
asets
.
E
x
p
e
r
im
e
n
ta
l
r
es
u
lts
d
e
m
o
n
s
tr
ate
h
i
g
h
a
cc
u
r
a
cy
a
n
d
s
ta
b
i
lit
y
u
n
d
er
d
i
v
er
s
e
c
o
n
d
iti
o
n
s
,
c
o
n
f
i
r
m
i
n
g
th
e
f
r
a
m
ew
o
r
k
’
s
ec
o
n
o
m
i
c
v
ia
b
i
lit
y
an
d
p
r
a
ctic
al
a
p
p
lic
ab
ilit
y
i
n
ag
r
i
cu
lt
u
r
e.
C
o
n
c
ep
tu
al
ly
,
t
h
is
h
y
b
r
i
d
ar
c
h
it
ec
t
u
r
e
p
r
o
v
i
d
es
a
n
o
v
el,
le
a
n
,
an
d
i
n
te
r
p
r
eta
b
l
e
d
i
r
ec
ti
o
n
f
o
r
f
u
tu
r
e
r
es
ea
r
c
h
in
s
m
a
r
t
a
g
r
ic
u
ltu
r
e
a
n
d
im
ag
e
-
b
ase
d
p
l
a
n
t
d
i
s
ea
s
e
d
ia
g
n
o
s
is
.
3
.
2
.
Descript
io
n o
f
d
a
t
a
s
et
W
e
ev
alu
ated
o
u
r
m
eth
o
d
s
o
n
two
d
atasets
,
Plan
tVillag
e
an
d
Plan
tDis
ea
s
e.
T
h
e
m
o
s
t
im
p
o
r
tan
t
d
if
f
er
en
ce
b
etwe
en
th
e
two
d
a
tasets
i
s
in
th
e
tex
tu
r
e
an
d
co
m
p
lex
ity
o
f
th
e
im
a
g
es.
T
h
e
Plan
tVillag
e
d
atase
t
co
n
tain
s
im
ag
es
tak
e
n
u
n
d
er
l
ab
o
r
ato
r
y
co
n
d
itio
n
s
with
u
n
if
o
r
m
s
m
o
o
th
g
r
ey
b
ac
k
g
r
o
u
n
d
s
an
d
v
e
r
y
m
in
im
al
n
o
is
e,
th
u
s
it
is
q
u
ite
a
"p
e
r
f
ec
t"
d
ataset
f
o
r
e
v
alu
atin
g
m
o
d
els,
w
h
er
ea
s
Plan
tDis
ea
s
e
co
n
tain
s
im
ag
es
co
llected
in
t
h
e
wild
,
wh
er
e
lig
h
tin
g
c
o
n
d
itio
n
s
an
d
n
atu
r
al
b
ac
k
g
r
o
u
n
d
s
ar
e
v
ar
ia
b
le,
th
u
s
it
r
ep
r
esen
ts
a
m
o
r
e
c
o
m
p
lex
d
ataset.
T
h
is
al
lo
ws
u
s
to
ev
alu
ate
th
e
p
r
o
p
o
s
ed
m
o
d
el'
s
r
eliab
ilit
y
an
d
ab
ili
ty
to
g
en
er
alize
i
n
b
o
th
id
ea
l c
o
n
d
itio
n
s
an
d
r
ea
li
s
tic
f
ield
-
lik
e
co
n
d
itio
n
s
.
3
.
2
.
1
.
P
la
ntV
illa
g
e
da
t
a
s
et
T
h
e
Plan
tVillag
e
d
ataset
is
a
p
r
im
ar
y
o
p
en
-
s
o
u
r
ce
d
ataset
f
o
r
p
lan
t
d
is
ea
s
e
id
en
tific
atio
n
.
T
h
e
d
ata
co
n
tain
s
6
1
,
4
8
6
p
ictu
r
es
ac
r
o
s
s
3
9
s
p
ec
ies,
in
clu
d
in
g
v
ar
i
o
u
s
p
lan
t
leaf
d
is
ea
s
es
an
d
b
ac
k
g
r
o
u
n
d
im
ag
es.
I
n
to
tal,
th
er
e
ar
e
5
4
,
3
0
6
im
ag
es
ac
r
o
s
s
3
8
class
es
f
o
r
i
n
d
i
v
id
u
al
d
is
ea
s
es
af
ter
ex
clu
d
in
g
th
e
b
ac
k
g
r
o
u
n
d
im
ag
es.
E
ac
h
o
f
th
ese
im
ag
es
co
n
tain
s
an
in
d
iv
i
d
u
al
p
lan
t
leaf
o
n
a
u
n
if
o
r
m
g
r
ey
b
ac
k
g
r
o
u
n
d
,
d
esig
n
e
d
to
m
im
ic
id
ea
l
in
p
u
t
f
o
r
a
u
to
m
a
ted
an
aly
tical
s
y
s
tem
s
.
T
h
is
u
n
if
o
r
m
b
ac
k
g
r
o
u
n
d
im
itates
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
s
tep
,
in
wh
ich
i
n
d
iv
id
u
al
leav
es
ar
e
m
ask
ed
f
r
o
m
a
lar
g
e
r
c
an
o
p
y
im
ag
e
b
y
a
s
m
ar
t
m
o
n
ito
r
in
g
s
y
s
tem
an
d
ex
tr
ac
ted
.
T
h
e
o
r
g
an
izatio
n
o
f
th
e
d
ataset
an
d
th
e
m
an
y
class
e
s
h
av
e
m
ad
e
it
a
cr
iti
ca
l
b
en
ch
m
ar
k
f
o
r
d
ev
elo
p
in
g
an
d
v
alid
atin
g
m
a
ch
in
e
lear
n
in
g
m
o
d
els f
o
r
a
g
r
i
cu
ltu
r
al
d
is
ea
s
e
id
en
tific
atio
n
.
3
.
2
.
2
.
P
la
nt
dis
ea
s
e
da
t
a
s
et
T
h
e
p
lan
t
d
is
ea
s
e
d
ataset
is
a
n
im
p
o
r
tan
t
co
m
p
o
n
en
t
o
f
ass
ess
in
g
th
e
r
o
b
u
s
tn
ess
an
d
g
en
er
aliza
tio
n
o
f
th
e
m
o
d
els
we
p
r
esen
ted
.
I
t
p
r
o
v
id
es
a
b
en
c
h
m
ar
k
f
o
r
ass
ess
in
g
p
er
f
o
r
m
a
n
ce
wh
e
n
n
o
p
r
e
p
r
o
ce
s
s
in
g
tech
n
iq
u
es
ar
e
u
s
ed
,
in
cl
u
d
in
g
o
b
ject
m
ask
in
g
o
r
n
o
is
e
r
ed
u
ctio
n
;
th
is
d
ataset
was
au
g
m
en
ted
o
f
f
lin
e
b
ase
d
o
n
th
e
o
r
ig
in
al
d
ataset
wh
ich
we
p
r
o
v
id
ed
as
a
lin
k
in
th
is
GitHu
b
.
Alth
o
u
g
h
o
u
r
d
ataset
d
o
es
n
o
t
h
av
e
an
y
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
P
la
n
t d
is
ea
s
e
d
etec
tio
n
a
n
d
c
la
s
s
ifica
tio
n
:
b
a
s
ed
o
n
ma
ch
in
e
lea
r
n
in
g
a
n
d
…
(
E
l
A
r
o
u
s
s
i
El
M
eh
d
i
)
5339
n
o
is
e
r
ed
u
ctio
n
it
co
n
tain
s
al
m
o
s
t
8
7
,
0
0
0
R
GB
im
ag
es
o
f
s
o
u
n
d
a
n
d
d
is
ea
s
ed
cr
o
p
leav
es
ac
r
o
s
s
3
8
d
if
f
e
r
en
t
ca
teg
o
r
ies
.
T
h
e
co
m
p
lete
d
at
aset
was
d
iv
id
ed
with
in
tr
ain
in
g
an
d
v
alid
atio
n
s
ets
in
an
8
0
/2
0
r
atio
,
wh
ile
p
r
eser
v
in
g
t
h
e
o
r
i
g
in
al
d
ir
ec
to
r
y
s
tr
u
ctu
r
e.
3
.
3
.
E
x
perim
ent
a
l
des
ig
n a
n
d c
o
nfig
ura
t
io
n
T
h
e
ex
p
er
im
en
tal
m
o
d
els
we
r
e
test
ed
o
n
a
d
ataset
s
p
ec
if
ically
ch
o
s
en
to
ev
alu
ate
s
tab
ilit
y
an
d
g
en
er
aliza
tio
n
at
f
air
ly
co
m
p
r
o
m
is
ed
d
ata
co
n
d
itio
n
s
,
with
o
u
t
s
tep
p
ed
o
r
o
th
er
p
r
e
-
p
r
o
c
ess
es,
n
o
is
e
r
em
o
v
al
o
r
m
ask
s
eg
m
en
tatio
n
.
W
e
u
s
ed
av
er
a
g
e
p
r
ec
is
io
n
a
n
d
a
v
er
ag
e
r
ec
all
to
m
ea
s
u
r
e
p
er
f
o
r
m
an
ce
.
T
h
e
co
d
e
was
wr
itten
in
Py
th
o
n
in
An
ac
o
n
d
a
en
v
ir
o
n
m
en
t
u
s
in
g
W
in
d
o
w
s
1
0
,
r
u
n
n
in
g
o
n
h
ar
d
war
e
th
a
t
h
ad
2
.
2
GHz
I
n
tel
C
o
r
e
i5
C
PU,
1
2
GB
o
f
1
6
0
0
MH
z
DDR
R
AM
,
I
n
tel
HD
Gr
ap
h
ics
5
0
0
0
GPU
to
1
5
3
6
MB
o
f
R
OM
.
T
h
e
s
o
f
twar
e
in
clu
d
e
d
lib
r
a
r
ies
s
u
ch
as
Ker
as,
Op
e
n
C
V,
Nu
m
Py
,
C
u
DNN,
T
h
ea
n
o
,
u
tili
zin
g
An
ac
o
n
d
a3
w
h
ich
in
clu
d
ed
NVI
DI
A'
s
C
u
DNN
as
well
as
C
u
Me
M
wi
th
th
e
b
ac
k
en
d
o
f
T
h
ea
n
o
,
f
o
r
o
p
ti
m
ized
m
em
o
r
y
an
d
p
er
f
o
r
m
an
ce
wh
en
r
u
n
n
i
n
g
d
ee
p
lear
n
in
g
,
as
well
as
ex
ce
llen
t
im
ag
e
-
p
r
o
ce
s
s
in
g
ac
r
o
s
s
m
u
ltip
le
p
latf
o
r
m
s
(
L
in
u
x
,
W
in
d
o
ws,
Ma
c,
iOS,
An
d
r
o
id
,
an
d
J
av
a
)
.
T
h
ese
el
em
en
ts
to
g
eth
er
e
n
s
u
r
e
th
e
al
g
o
r
ith
m
ac
h
iev
es
a
h
ig
h
-
lev
el
o
f
p
r
ec
is
io
n
an
d
s
ca
lab
ilit
y
f
o
r
r
ea
l
-
w
o
r
ld
ap
p
licatio
n
s
.
Fig
u
r
e
1
s
h
o
ws th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
.
I
n
p
u
t
I
m
ag
es
wer
e
i
n
itially
r
e
s
ized
v
ia
b
icu
b
ic
in
ter
p
o
latio
n
in
MA
T
L
AB
,
t
o
2
2
7
×
2
2
7
×
3
,
an
d
th
en
r
u
n
th
r
o
u
g
h
th
e
al
g
o
r
ith
m
in
two
m
ain
s
tep
s
:
i
)
an
im
p
r
o
v
ed
Alex
Net
u
s
in
g
C
NN
an
d
ii
)
th
e
E
ig
(
Hess
)
-
C
o
HOG
alg
o
r
ith
m
.
T
h
e
en
h
a
n
ce
d
Alex
Net
ex
tr
ac
ts
d
ee
p
f
ea
tu
r
es,
y
ield
in
g
a
1
×6
4
f
ea
tu
r
e
v
ec
to
r
,
wh
ile
E
ig
(
Hess
)
-
C
o
HOG
p
r
o
d
u
ce
s
co
m
p
lem
en
tar
y
h
an
d
cr
af
ted
f
ea
tu
r
es.
T
o
r
ed
u
ce
d
im
en
s
io
n
ality
,
we
ap
p
ly
PC
A
to
th
e
E
i
g
(
Hess
)
-
C
o
HOG
v
ec
to
r
,
co
m
p
r
ess
in
g
it
t
o
1
×
5
9
with
o
u
t
s
ig
n
if
ican
t
in
f
o
r
m
atio
n
lo
s
s
.
T
h
ese
PC
A
-
r
ed
u
ce
d
f
ea
tu
r
es
ar
e
co
n
ca
te
n
ated
with
th
e
Alex
Net
f
ea
t
u
r
es
to
f
o
r
m
a
co
m
p
o
s
ite
1
×1
2
8
f
ea
tu
r
e
v
ec
to
r
,
co
m
b
in
in
g
th
e
s
tr
en
g
t
h
s
o
f
b
o
th
m
eth
o
d
s
.
T
h
e
f
o
llo
win
g
s
ec
tio
n
s
o
f
f
er
c
o
m
p
r
e
h
en
s
iv
e
ex
p
lan
atio
n
s
o
f
th
e
Alex
Net
C
NN,
th
e
E
ig
(
Hess
)
-
C
o
HOG
alg
o
r
ith
m
,
an
d
PC
A
.
Fig
u
r
e
1
.
Ou
r
p
r
o
p
o
s
ed
d
escr
ip
to
r
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
3
3
6
-
5
3
4
6
5340
3
.
3
.
1
.
P
CA
I
n
th
is
ar
ticle,
PC
A
[
2
9
]
was
u
tili
ze
d
as
a
k
ey
d
im
en
s
io
n
a
lity
r
ed
u
ctio
n
tech
n
iq
u
e
to
en
h
an
ce
th
e
p
er
f
o
r
m
an
ce
an
d
ef
f
icie
n
cy
o
f
th
e
p
r
o
p
o
s
ed
p
lan
t
d
is
ea
s
e
class
if
ica
tio
n
m
o
d
el.
T
h
e
h
an
d
cr
a
f
ted
f
ea
tu
r
es
ex
tr
ac
ted
u
s
in
g
th
e
E
i
g
(
Hess
)
-
C
o
HOG
d
escr
ip
to
r
,
alth
o
u
g
h
r
ich
in
g
e
o
m
etr
ic
an
d
te
x
tu
r
al
in
f
o
r
m
atio
n
,
in
itially
p
r
o
d
u
ce
d
a
h
i
g
h
-
d
im
en
s
io
n
al
f
ea
tu
r
e
v
ec
to
r
.
Su
c
h
h
ig
h
d
im
e
n
s
io
n
ality
o
f
ten
in
tr
o
d
u
ce
s
r
ed
u
n
d
a
n
cy
,
in
cr
ea
s
es
co
m
p
u
tatio
n
al
lo
a
d
,
an
d
m
a
y
d
e
g
r
ad
e
m
o
d
el
g
en
er
aliza
tio
n
d
u
e
t
o
o
v
er
f
itti
n
g
.
T
o
m
itig
ate
th
ese
is
s
u
es,
P
C
A
was
ap
p
lied
to
tr
an
s
f
o
r
m
th
e
co
r
r
elate
d
f
ea
t
u
r
es
in
to
a
n
ew
s
et
o
f
u
n
co
r
r
elate
d
,
o
r
th
o
g
o
n
al
co
m
p
o
n
en
ts
r
an
k
e
d
b
y
th
e
a
m
o
u
n
t
o
f
v
a
r
ian
ce
th
e
y
ca
p
t
u
r
e
f
r
o
m
th
e
o
r
ig
in
al
d
ata.
B
y
s
elec
tin
g
th
e
to
p
co
m
p
o
n
en
ts
th
at
p
r
eser
v
ed
m
o
r
e
th
an
9
5
%
o
f
th
e
to
tal
v
a
r
ian
ce
,
th
e
f
ea
tu
r
e
v
ec
to
r
was
r
ed
u
ce
d
to
a
m
o
r
e
m
an
ag
ea
b
le
s
ize
o
f
1
×5
9
wit
h
o
u
t
s
ig
n
if
ican
t
lo
s
s
o
f
d
is
cr
i
m
in
ativ
e
p
o
wer
.
T
h
is
r
ed
u
ce
d
v
ec
to
r
r
etain
e
d
th
e
m
o
s
t
m
ea
n
in
g
f
u
l
s
tr
u
ctu
r
al
a
n
d
tex
tu
r
al
cu
es
f
r
o
m
th
e
leaf
im
ag
es.
Su
b
s
eq
u
en
tly
,
th
e
PC
A
-
co
m
p
r
ess
ed
f
ea
tu
r
e
v
ec
to
r
was
co
n
ca
te
n
ated
with
a
1
×
6
4
d
ee
p
f
ea
tu
r
e
v
ec
to
r
ex
tr
ac
te
d
f
r
o
m
th
e
im
p
r
o
v
ed
Alex
Net
C
NN
.
T
h
e
r
esu
ltin
g
c
o
m
p
o
s
ite
f
ea
tu
r
e
v
ec
to
r
o
f
s
ize
1
×
1
2
8
ef
f
ec
ti
v
ely
co
m
b
in
es
lo
w
-
lev
el
h
an
d
cr
af
ted
d
escr
ip
to
r
s
with
h
ig
h
-
lev
el
s
em
an
tic
f
ea
t
u
r
es,
en
ab
lin
g
a
r
o
b
u
s
t
an
d
co
m
p
u
tatio
n
ally
ef
f
icien
t
clas
s
if
icatio
n
p
ip
elin
e.
Ov
er
all,
PC
A
p
lay
ed
a
c
r
u
cia
l
r
o
le
in
b
alan
cin
g
m
o
d
el
co
m
p
lex
ity
an
d
ac
cu
r
ac
y
,
i
m
p
r
o
v
in
g
tr
ain
i
n
g
s
p
ee
d
,
r
ed
u
cin
g
m
em
o
r
y
u
s
ag
e,
an
d
e
n
h
an
cin
g
th
e
m
o
d
el'
s
ab
ilit
y
to
g
en
er
alize
ac
r
o
s
s
d
iv
er
s
e
d
at
asets
.
3
.
3
.
2
.
I
m
pro
v
ed
Alex
Net
a
rc
hite
ct
ure
T
h
e
Alex
Net
ar
ch
itectu
r
e
is
a
s
tr
o
n
g
n
eu
r
al
n
etwo
r
k
wh
e
n
it
co
m
es
to
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
is
co
m
p
o
s
ed
o
f
2
5
lay
e
r
s
.
W
e
p
r
o
p
o
s
e
an
im
p
r
o
v
ed
Alex
Net
a
r
ch
itectu
r
e
in
th
is
r
esear
ch
to
o
p
tim
ize
ef
f
icien
c
y
,
an
d
p
er
f
o
r
m
a
n
ce
.
T
h
e
r
esu
ltin
g
task
was to
o
p
tim
ize
th
e
Alex
Net
ar
ch
itectu
r
e
b
y
r
em
o
v
i
n
g
th
e
last
th
r
ee
f
u
lly
co
n
n
ec
ted
lay
er
s
;
th
ey
ar
e
also
ty
p
ically
th
e
c
o
s
tlies
t
lay
er
s
to
d
ev
elo
p
.
I
n
s
tead
,
we
r
ep
l
ac
e
th
em
with
o
n
e
f
u
lly
co
n
n
ec
te
d
lay
er
o
f
s
ize
1
×6
4
.
I
n
th
eo
r
y
,
o
p
tim
izatio
n
r
ed
u
ce
s
th
e
to
tal
n
u
m
b
er
o
f
p
ar
am
eter
s
s
ig
n
if
ican
tly
wh
ile
r
e
d
u
cin
g
o
u
r
co
s
t,
b
u
t
we
ca
n
m
ain
tain
o
r
im
p
r
o
v
e
p
er
f
o
r
m
an
ce
in
r
ela
tio
n
to
th
e
o
r
i
g
in
al
d
esig
n
.
I
n
p
r
ac
tice,
o
u
r
u
p
d
at
ed
ar
c
h
itectu
r
e
m
ain
tain
ed
s
im
ilar
r
ep
r
esen
tatio
n
s
as
p
r
io
r
to
o
p
tim
izatio
n
with
im
p
r
o
v
e
d
p
er
f
o
r
m
an
c
e.
T
h
e
f
u
lly
co
n
n
ec
te
d
lay
er
s
ar
e
ty
p
i
ca
lly
u
s
ed
f
o
r
id
en
tif
y
in
g
a
n
d
class
if
y
in
g
th
e
m
o
s
t
r
elev
an
t,
h
i
g
h
-
lev
el
f
ea
tu
r
es;
an
d
c
o
n
s
eq
u
en
tly
,
th
e
n
ee
d
f
o
r
m
an
u
ally
d
ev
elo
p
ed
f
ea
tu
r
es
is
n
o
lo
n
g
er
ap
p
licab
le.
E
s
s
en
tially
,
th
e
d
e
s
ig
n
is
well
b
alan
ce
d
,
allo
win
g
m
ax
im
u
m
p
er
f
o
r
m
an
ce
wh
i
le
r
em
ain
in
g
q
u
ite
s
im
p
le.
B
ec
au
s
e
o
p
tim
izatio
n
n
ee
d
ed
to
b
e
s
im
p
le
e
n
o
u
g
h
to
d
ev
elo
p
a
n
ef
f
ec
tiv
e,
s
m
ar
t
n
eu
r
al
n
etwo
r
k
t
o
b
e
ab
le
to
ex
tr
ac
t f
ea
t
u
r
es e
f
f
icie
n
tly
an
d
a
p
p
r
o
p
r
iately
.
3
.
3
.
3
.
T
he
E
ig
(
H
ess
)
-
Co
H
O
G
des
cr
ipto
r
T
h
e
E
ig
(
Hess
)
-
C
o
HOG
alg
o
r
ith
m
[
3
0
]
ca
n
b
e
class
if
ied
as
a
s
h
ap
e
-
b
ased
im
ag
e
d
escr
ip
to
r
.
T
h
e
E
ig
(
Hess
)
-
C
o
HOG
alg
o
r
ith
m
allo
ws
f
o
r
th
e
e
x
tr
ac
tio
n
o
f
s
tr
u
ctu
r
al
a
n
d
tex
tu
r
al
in
f
o
r
m
atio
n
;
t
h
is
is
ac
co
m
p
lis
h
ed
b
y
co
m
b
in
in
g
Hess
ian
-
b
ased
cu
r
v
atu
r
e
with
a
C
o
HOG.
I
n
th
is
ca
s
e,
th
e
r
esu
ltin
g
f
ea
tu
r
e
r
ep
r
esen
tatio
n
is
v
er
y
d
is
cr
im
in
ativ
e,
wh
ich
is
ad
v
a
n
tag
eo
u
s
f
o
r
u
s
e
in
ag
r
icu
ltu
r
al
ap
p
licatio
n
s
,
s
in
ce
p
lan
t
d
is
ea
s
es
u
s
u
ally
r
ep
r
esen
t
d
ev
iatio
n
s
in
leaf
o
u
tlin
es,
leaf
te
x
tu
r
es,
an
d
leaf
v
ein
p
atter
n
s
.
T
h
e
alg
o
r
ith
m
f
ir
s
t
co
m
p
u
tes th
e
Hess
ian
m
atr
ix
(
,
)
f
o
r
ev
e
r
y
p
i
x
el
o
f
a
g
r
ay
s
ca
le
i
m
ag
e
(
,
)
d
ef
in
ed
as:
(
,
)
=
[
2
2
2
2
2
2
]
(
1
)
w
ith
2
2
an
d
2
2
ar
e
th
e
s
ec
o
n
d
p
ar
ti
al
d
er
iv
ativ
es with
r
esp
ec
t to
x
an
d
y
r
esp
ec
tiv
ely
;
2
is
th
e
cr
o
s
s
p
ar
tial d
er
iv
ativ
e;
T
h
e
lo
ca
l
c
u
r
v
atu
r
e
o
f
th
e
m
atr
ix
,
b
y
m
ea
n
s
o
f
its
eig
en
v
alu
es,
p
r
o
v
id
es
a
m
eth
o
d
to
d
etec
t
g
eo
m
etr
ically
s
ig
n
i
f
ican
t
ar
ea
s
lik
e
co
r
n
er
s
,
r
i
d
g
es,
o
r
b
lo
b
s
.
T
h
e
f
e
atu
r
es
ar
e
s
tab
le
to
i
m
ag
e
r
o
tatio
n
a
n
d
r
ea
s
o
n
ab
ly
r
o
b
u
s
t
to
illu
m
in
at
io
n
ch
an
g
es.
At
th
e
s
am
e
tim
e
,
C
o
HOG
is
u
s
ed
f
o
r
en
co
d
in
g
th
e
o
r
ien
tatio
n
o
f
g
r
ad
ien
ts
.
W
h
er
ea
s
HOG
o
n
ly
ca
p
tu
r
es
lo
ca
l
e
d
g
e
o
r
i
en
tatio
n
in
f
o
r
m
atio
n
,
C
o
HOG
ca
p
tu
r
es
s
p
atial
d
is
tr
ib
u
tio
n
in
f
o
r
m
atio
n
p
er
ta
in
in
g
to
g
r
ad
ie
n
t
o
r
ie
n
tatio
n
p
air
s
an
d
t
h
er
ef
o
r
e
p
r
o
v
id
es
a
r
ich
er
m
ea
n
s
o
f
tex
tu
r
al
an
d
s
tr
u
ct
u
r
al
co
n
tex
t.
E
ig
(
Hess
)
-
C
o
HOG
d
escr
ip
to
r
is
b
u
ilt as.
−
S
t
e
p
1
:
C
a
l
c
u
l
a
t
e
t
h
e
e
i
g
e
n
v
a
l
u
e
s
o
f
t
h
e
H
ess
i
a
n
m
a
t
r
i
x
a
t
e
v
er
y
p
i
x
e
l
a
s
a
wa
y
t
o
i
d
e
n
t
i
f
y
c
u
r
v
a
t
u
r
e
f
e
a
t
u
r
e
s
.
−
Step
2:
F
o
r
ea
c
h
im
p
o
r
tan
t
e
ig
en
v
alu
e
ar
ea
,
ca
lcu
late
C
o
HOG
f
ea
tu
r
es
b
y
u
s
in
g
lo
c
al
g
r
ad
ie
n
t
co
-
o
cc
u
r
r
e
n
ce
at
d
if
f
e
r
en
t d
is
tan
c
es a
n
d
an
g
les.
−
Step
3:
C
o
m
b
in
e
b
o
t
h
th
e
cu
r
v
atu
r
e
-
b
ased
an
d
c
o
-
o
cc
u
r
r
e
n
ce
f
ea
tu
r
es
in
to
o
n
e
d
escr
i
p
to
r
v
ec
to
r
th
at
s
h
o
ws b
o
th
s
h
ap
e
a
n
d
tex
tu
r
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
P
la
n
t d
is
ea
s
e
d
etec
tio
n
a
n
d
c
la
s
s
ifica
tio
n
:
b
a
s
ed
o
n
ma
ch
in
e
lea
r
n
in
g
a
n
d
…
(
E
l
A
r
o
u
s
s
i
El
M
eh
d
i
)
5341
3
.
3
.
4
.
Ca
t
eg
o
rica
l
cr
o
s
s
-
ent
r
o
py
lo
s
s
Fo
r
m
o
d
el
p
er
f
o
r
m
a
n
ce
o
p
tim
izatio
n
,
th
e
ca
teg
o
r
ical
cr
o
s
s
-
en
tr
o
p
y
l
o
s
s
f
u
n
ctio
n
was
im
p
lem
en
ted
,
wh
ich
is
ty
p
ical
f
o
r
m
u
lti
-
c
lass
m
u
lti
-
lab
el
class
if
icat
io
n
p
r
o
b
lem
s
.
T
h
is
lo
s
s
f
u
n
ct
io
n
m
ea
s
u
r
es
h
o
w
d
if
f
er
en
t
th
e
p
r
ed
icted
p
r
o
b
ab
i
lity
d
is
tr
ib
u
tio
n
is
to
th
e
tr
u
e
lab
el
d
is
tr
ib
u
tio
n
f
o
r
a
g
iv
en
s
am
p
le.
T
h
e
lo
s
s
o
n
a
s
in
g
le
tr
ain
in
g
e
x
am
p
le
let
b
e:
ℒ
=
−
∑
=
1
.
l
og
(
̂
)
(
2
)
w
h
er
e
ℒ
:
T
o
tal
lo
s
s
co
m
p
u
ted
f
o
r
o
n
e
tr
ain
in
g
in
s
tan
ce
,
C
:
t
o
tal
n
u
m
b
er
o
f
class
,
y
i
:
a
ctu
al
lab
el
f
o
r
class
I
(
1
co
r
r
ec
t c
lass
,
0
all
o
th
er
s
)
,
̂
:
t
h
e
So
f
tMa
x
lay
er
'
s
o
u
tp
u
t r
e
p
r
esen
tin
g
th
e
p
r
ed
icted
p
r
o
b
a
b
ilit
y
f
o
r
class
i
.
T
h
e
lo
s
s
was
r
ed
u
ce
d
with
th
e
Ad
am
o
p
tim
izer
lear
n
in
g
r
a
te
o
f
0
.
0
0
1
.
Du
r
in
g
tr
ain
in
g
,
th
e
m
o
d
el
f
in
e
-
tu
n
e
d
weig
h
ts
iter
ativ
ely
with
b
ac
k
-
p
r
o
p
ag
atio
n
a
n
d
a
s
to
ch
asti
c
g
r
ad
ien
t
d
escen
t,
to
o
b
tain
o
p
tim
al
p
ar
am
eter
s
th
at
m
in
im
ized
th
e
tr
ain
in
g
lo
s
s
.
T
h
e
m
o
d
el
was
tr
ain
ed
f
o
r
3
0
ep
o
c
h
s
,
an
d
t
h
e
lo
s
s
v
alu
es
wer
e
tr
ac
k
ed
an
d
p
lo
tted
at
ea
ch
ep
o
ch
f
o
r
b
o
t
h
tr
ain
in
g
an
d
v
alid
atio
n
s
ets
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
ese
m
etr
ics
to
g
eth
er
estab
l
is
h
a
co
m
p
r
eh
en
s
iv
e
f
r
am
ew
o
r
k
f
o
r
ev
al
u
atin
g
th
e
ef
f
ec
ti
v
en
ess
an
d
r
eliab
ilit
y
o
f
th
e
p
r
o
p
o
s
ed
f
as
h
io
n
.
T
h
ey
illu
s
tr
ate
h
o
w
t
h
e
m
o
d
el
b
alan
ce
s
ac
c
u
r
ac
y
a
g
ain
s
t
d
if
f
er
en
t
ty
p
es
o
f
er
r
o
r
s
an
d
g
u
id
e
r
esear
ch
e
r
s
an
d
p
r
ac
titi
o
n
er
s
in
r
e
f
in
in
g
th
e
ap
p
r
o
ac
h
f
o
r
r
ea
l
-
wo
r
ld
ap
p
licatio
n
s
.
W
e
te
s
ted
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
n
tw
o
d
atasets
Plan
tVil
lag
e
an
d
Plan
tDis
ea
s
e
ac
r
o
s
s
m
u
ltip
le
r
u
n
s
.
T
h
e
r
esu
lts
,
s
u
m
m
ar
ized
in
T
ab
les
1
(
in
A
,
2
an
d
3
,
d
em
o
n
s
tr
ate
t
h
e
m
eth
o
d
’
s
e
f
f
icien
cy
u
n
d
e
r
v
ar
i
o
u
s
tr
ain
i
n
g
co
n
d
itio
n
s
.
T
h
e
m
o
d
el'
s
r
esu
lts
f
o
r
ea
ch
p
la
n
t
d
is
ea
s
e
class
o
n
b
o
th
th
e
Plan
tVillag
e
an
d
Plan
tDis
ea
s
e
d
atasets
ar
e
d
etailed
in
T
ab
le
1
.
Pre
cisi
o
n
(
%),
F
-
s
co
r
e
(
%)
,
an
d
ac
c
u
r
ac
y
(
%)
wer
e
r
ep
o
r
ted
f
o
r
ea
ch
class
d
u
r
in
g
tr
ain
i
n
g
(
1
0
an
d
3
0
iter
atio
n
s
)
an
d
v
alid
atio
n
(
a
f
ter
1
0
an
d
3
0
iter
atio
n
s
)
.
On
Plan
tVillag
e,
th
e
m
o
d
el
ac
h
iev
ed
v
er
y
h
i
g
h
t
r
ain
in
g
a
n
d
v
alid
atio
n
p
r
ec
is
io
n
s
,
as
w
ell
as
F
-
s
co
r
es
f
o
r
m
o
s
t
class
es.
Fo
r
e
x
am
p
le,
in
th
e
"Ap
p
le
-
Ap
p
le
Scab
"
c
lass
,
th
e
tr
ain
in
g
p
r
ec
is
io
n
i
n
cr
ea
s
ed
f
r
o
m
9
4
.
6
%
(
1
0
iter
atio
n
s
)
to
9
7
.
5
%
(
3
0
iter
atio
n
s
)
,
a
n
d
th
e
F
-
s
co
r
e
r
ea
ch
ed
9
3
%
af
ter
3
0
iter
atio
n
s
.
C
o
n
tin
u
in
g
with
ex
am
p
le
s
f
r
o
m
o
u
r
p
r
e
v
io
u
s
S
crip
ts
&
F
ile
O
r
g
a
n
iz
a
tio
n
ch
ap
ter
,
th
e
m
o
d
el,
a
f
ter
3
0
iter
atio
n
s
,
ac
h
iev
ed
v
alid
at
io
n
p
r
ec
is
io
n
s
an
d
F
-
s
co
r
es
clo
s
e
to
1
0
0
%
f
o
r
s
o
m
e
class
es
s
u
ch
as
"Ap
p
le
-
Hea
lth
y
,
"
"Bl
u
eb
e
r
r
y
-
H
ea
lth
y
,
"
an
d
"Co
r
n
(
Ma
ize)
-
Hea
lth
y
.
"
T
h
is
s
u
g
g
ests
th
at
th
e
m
o
d
el
p
er
f
o
r
m
ed
q
u
ite
well
o
n
th
e
Plan
tVillag
e
d
ataset.
On
th
e
o
th
er
h
an
d
,
o
v
er
all
p
er
f
o
r
m
a
n
ce
o
n
th
e
Plan
tDis
ea
s
e
d
ataset
was
g
en
er
ally
lo
wer
t
h
a
n
o
n
Plan
tVillag
e;
h
o
wev
er
,
it wa
s
s
till
m
ea
n
in
g
f
u
l.
Fo
r
"Ap
p
le
-
Ap
p
le
Scab
"
o
n
th
e
Plan
tDis
ea
s
e
d
ataset,
th
e
tr
ain
in
g
p
r
ec
is
io
n
was
9
2
.
2
%
(
1
0
iter
atio
n
s
)
an
d
9
5
.
8
%
(
3
0
iter
atio
n
s
)
,
with
a
n
F
-
s
co
r
e
o
f
8
6
%
at
3
0
iter
atio
n
s
.
T
h
e
m
o
d
el
also
p
er
f
o
r
m
ed
q
u
ite
well
o
n
s
o
m
e
class
es
lik
e
"Or
an
g
e
-
Hu
an
g
lo
n
g
b
i
n
g
(
C
itru
s
Gr
ee
n
in
g
)
"
an
d
"Gr
ap
e
-
E
s
ca
(
B
lack
Me
asles
)
,
"
with
F
-
s
co
r
es
ar
o
u
n
d
8
0
%
–
8
2
%.
T
h
is
s
u
g
g
ests
th
at
th
ese
two
clas
s
es
m
ay
b
e
p
ar
ticu
lar
l
y
im
p
o
r
tan
t
to
d
is
tin
g
u
is
h
with
i
n
th
e
Plan
tDis
ea
s
e
d
ataset.
Fo
r
m
o
s
t
o
th
er
class
es,
in
cr
ea
s
in
g
th
e
n
u
m
b
er
o
f
iter
atio
n
s
f
r
o
m
1
0
to
3
0
led
to
im
p
r
o
v
e
d
p
r
ec
is
io
n
a
n
d
F
-
s
co
r
es
.
T
ab
le
2
p
r
o
v
id
es
a
n
o
v
e
r
all
p
er
f
o
r
m
an
ce
s
u
m
m
a
r
y
o
f
th
e
p
r
o
p
o
s
ed
f
ash
io
n
o
n
th
e
Plan
tVillag
e
d
ataset,
f
o
cu
s
in
g
o
n
p
r
ec
is
io
n
an
d
lo
s
s
f
o
r
tr
ain
in
g
an
d
v
alid
atio
n
af
ter
1
0
a
n
d
3
0
tr
ai
n
in
g
iter
atio
n
s
.
W
e
f
o
cu
s
ed
o
n
ea
c
h
m
etr
ic:
tr
ain
in
g
p
r
ec
is
io
n
(
%),
v
alid
atio
n
p
r
ec
is
io
n
(
%),
tr
ain
in
g
lo
s
s
,
an
d
v
alid
atio
n
lo
s
s
.
af
ter
1
0
tr
ain
in
g
iter
atio
n
s
,
th
e
m
eth
o
d
attain
s
a
tr
ai
n
in
g
p
r
ec
is
io
n
o
f
9
4
.
9
6
%
an
d
a
v
ali
d
atio
n
p
r
ec
is
io
n
o
f
9
2
.
3
%,
alo
n
g
with
th
e
tr
ain
in
g
an
d
v
alid
atio
n
lo
s
s
es
o
f
0
.
1
8
7
5
an
d
0
.
2
0
1
6
,
r
esp
ec
tiv
ely
.
Af
ter
3
0
tr
ain
in
g
iter
atio
n
s
,
th
e
tr
ai
n
in
g
p
r
ec
is
io
n
im
p
r
o
v
e
d
to
9
8
.
6
4
%,
an
d
th
e
v
alid
atio
n
p
r
ec
is
io
n
in
cr
ea
s
ed
to
9
3
.
5
%.
Mo
r
eo
v
er
,
th
e
tr
ain
in
g
lo
s
s
s
ig
n
if
ican
tly
d
ec
r
ea
s
ed
to
0
.
0
6
2
3
,
wh
ile
th
e
v
alid
atio
n
lo
s
s
f
o
r
3
0
iter
atio
n
s
was
n
o
t
s
h
o
wn
in
th
e
tab
le.
As
n
o
ted
,
th
e
m
o
d
el’
s
tr
ain
in
g
p
e
r
f
o
r
m
an
ce
im
p
r
o
v
ed
s
ig
n
if
ican
tly
,
an
d
th
e
r
esu
lts
f
r
o
m
tr
ain
i
n
g
o
n
th
e
Plan
tVil
lag
e
d
ataset
in
d
icate
th
at
ad
d
itio
n
al
tr
ain
in
g
lead
s
to
b
ette
r
co
n
v
e
r
g
en
ce
an
d
o
v
er
all
im
p
r
o
v
e
d
p
er
f
o
r
m
an
ce
f
o
r
th
e
p
r
o
p
o
s
ed
m
o
d
el.
T
ab
l
e
3
o
f
f
er
e
d
th
e
ac
cu
r
ac
y
a
n
d
lo
s
s
o
f
th
e
p
r
o
p
o
s
e
d
f
ash
io
n
o
n
th
e
Plan
tDis
ea
s
e
d
ataset
af
ter
1
0
an
d
3
0
iter
at
io
n
s
o
f
tr
ain
in
g
.
Similar
to
T
ab
le
2
,
it
i
n
clu
d
es
tr
ain
in
g
ac
c
u
r
ac
y
,
v
alid
atio
n
ac
cu
r
ac
y
,
t
r
ain
in
g
lo
s
s
,
an
d
v
alid
atio
n
l
o
s
s
.
Af
ter
1
0
iter
atio
n
s
o
n
th
e
Plan
tDis
ea
s
e
d
ataset,
th
e
tr
ain
in
g
ac
c
u
r
ac
y
was
8
2
.
2
5
%
an
d
th
e
v
alid
atio
n
ac
cu
r
ac
y
was
8
3
.
1
1
%,
with
tr
ain
in
g
an
d
v
alid
atio
n
lo
s
s
es
o
f
0
.
3
2
4
7
an
d
0
.
2
8
0
3
,
r
esp
e
ctiv
ely
.
Af
ter
3
0
iter
atio
n
s
,
th
e
tr
ain
in
g
ac
cu
r
ac
y
im
p
r
o
v
e
d
to
8
5
.
4
3
%
an
d
th
e
v
alid
atio
n
ac
cu
r
ac
y
to
8
7
.
2
6
%,
wh
ile
th
e
tr
ain
in
g
lo
s
s
d
e
cr
ea
s
ed
to
0
.
1
5
4
3
.
Ho
wev
er
,
th
e
v
alid
atio
n
lo
s
s
f
o
r
3
0
iter
atio
n
s
is
n
o
t
r
ep
o
r
te
d
.
T
h
ese
r
esu
lts
in
d
icate
co
n
s
i
s
ten
t
im
p
r
o
v
em
e
n
t
with
ad
d
itio
n
al
tr
ain
in
g
o
n
th
e
Plan
tDis
ea
s
e
d
ataset,
alth
o
u
g
h
th
e
p
r
ec
is
io
n
s
co
r
es
f
o
r
Pl
an
tDis
ea
s
e
r
em
ain
g
en
er
ally
lo
wer
th
an
th
o
s
e
f
o
r
th
e
Plan
tVillag
e
d
ataset.
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
3
3
6
-
5
3
4
6
5342
T
ab
le
1
.
C
lass
-
wis
e
p
er
f
o
r
m
an
ce
o
n
Plan
tDis
ea
s
e
d
ataset
an
d
Plan
tVillag
e
d
ataset
P
l
a
n
t
V
i
l
l
a
g
e
d
a
t
a
se
t
P
l
a
n
t
D
i
s
e
a
s
e
d
a
t
a
se
t
Tr
a
i
n
i
n
g
p
r
e
c
i
si
o
n
%
V
a
l
i
d
a
t
i
o
n
p
r
e
c
i
si
o
n
%
F1
-
sco
r
e
(
%)
Tr
a
i
n
i
n
g
p
r
e
c
i
si
o
n
%
V
a
l
i
d
a
t
i
o
n
p
r
e
c
i
si
o
n
%
F1
-
sco
r
e
(
%)
1
0
I
t
e
r
a
t
i
o
n
s
3
0
I
t
e
r
a
t
i
o
n
s
1
0
I
t
e
r
a
t
i
o
n
s
3
0
I
t
e
r
a
t
i
o
n
s
1
0
I
t
e
r
a
t
i
o
n
s
3
0
I
t
e
r
a
t
i
o
n
s
1
0
I
t
e
r
a
t
i
o
n
s
3
0
I
t
e
r
a
t
i
o
n
s
A
p
p
l
e
-
A
p
p
l
e
S
c
a
b
94
.
6
97
.
5
93
92
.
2
95
.
8
82
.
6
86
81
.
2
84
.
4
83
.
55
A
p
p
l
e
-
B
l
a
c
k
R
o
t
94
.
2
97
92
92
95
.
4
81
.
5
84
.
4
80
.
5
83
.
3
82
.
4
A
p
p
l
e
-
C
e
d
a
r
A
p
p
l
e
R
u
s
t
96
97
.
8
93
.
5
92
.
8
97
83
.
5
87
.
7
83
85
.
3
84
.
8
A
p
p
l
e
–
H
e
a
l
t
h
y
97
.
3
98
.
8
93
93
.
8
99
87
.
5
91
.
4
86
89
.
3
88
B
l
u
e
b
e
r
r
y
–
H
e
a
l
t
h
y
97
.
1
1
0
0
92
95
98
.
7
86
90
.
9
85
.
5
87
.
8
87
C
h
e
r
r
y
(
S
o
u
r
)
-
P
o
w
d
e
r
y
M
i
l
d
e
w
93
.
5
95
.
6
89
90
95
81
85
80
.
3
82
.
8
82
.
2
C
h
e
r
r
y
(
S
o
u
r
)
–
H
e
a
l
t
h
y
97
.
2
97
.
5
91
92
.
6
99
86
.
3
91
85
88
.
1
87
.
6
C
o
r
n
(
M
a
i
z
e
)
-
C
e
r
c
o
s
p
o
r
a
Le
a
f
S
p
o
t
94
.
1
95
.
4
92
90
.
3
95
.
4
82
85
.
8
81
83
.
8
83
.
15
C
o
r
n
(
M
a
i
z
e
)
-
C
o
m
m
o
n
R
u
s
t
95
.
4
98
.
5
93
93
.
4
97
83
.
4
87
.
5
82
.
8
85
.
2
84
.
7
C
o
r
n
(
M
a
i
z
e
)
-
N
o
r
t
h
e
r
n
Le
a
f
B
l
i
g
h
t
94
.
9
1
0
0
91
95
96
.
4
83
86
.
9
82
.
3
84
.
8
8
4
,
2
5
C
o
r
n
(
M
a
i
z
e
)
–
H
e
a
l
t
h
y
98
.
1
1
0
0
94
95
1
0
0
88
93
.
1
87
89
.
8
89
.
4
G
r
a
p
e
-
B
l
a
c
k
R
o
t
93
.
8
97
.
3
91
.
3
92
.
3
95
.
3
81
84
.
8
80
82
.
8
82
.
15
G
r
a
p
e
-
Esc
a
(
B
l
a
c
k
M
e
a
sl
e
s)
92
.
2
96
.
1
90
91
94
79
.
6
83
.
7
78
.
7
81
.
4
8
0
,
8
5
G
r
a
p
e
-
Le
a
f
B
l
i
g
h
t
93
.
4
97
.
8
91
92
.
6
95
81
85
.
3
80
.
4
82
.
8
82
.
5
G
r
a
p
e
–
H
e
a
l
t
h
y
96
.
5
1
0
0
94
.
2
95
98
.
6
85
.
6
90
84
.
7
87
.
4
87
O
r
a
n
g
e
-
H
u
a
n
g
l
o
n
g
b
i
n
g
(
C
i
t
r
u
s Gr
e
e
n
i
n
g
)
90
.
5
96
.
8
88
.
2
91
.
5
91
.
7
80
82
.
7
77
.
9
81
.
8
8
0
,
6
P
e
a
c
h
-
B
a
c
t
e
r
i
a
l
S
p
o
t
91
.
3
97
.
5
89
92
.
5
93
.
2
81
83
.
8
79
.
3
82
.
8
8
1
,
7
P
e
a
c
h
–
H
e
a
l
t
h
y
97
1
0
0
94
95
98
.
6
86
89
.
9
84
.
7
87
.
8
8
7
,
1
P
e
p
p
e
r
(
B
e
l
l
)
-
B
a
c
t
e
r
i
a
l
S
p
o
t
92
.
4
97
.
5
91
92
.
5
94
.
3
80
.
4
84
.
2
79
.
8
82
.
2
81
.
65
P
e
p
p
e
r
(
B
e
l
l
)
–
H
e
a
l
t
h
y
97
.
8
1
0
0
94
.
7
95
99
.
5
87
91
.
7
86
.
5
88
.
8
88
.
5
P
o
t
a
t
o
-
E
a
r
l
y
B
l
i
g
h
t
95
1
0
0
93
95
96
.
3
82
85
.
5
81
83
.
8
83
P
o
t
a
t
o
-
L
a
t
e
B
l
i
g
h
t
95
.
6
1
0
0
9
2
,
5
95
97
83
87
82
.
4
84
.
8
84
.
3
P
o
t
a
t
o
–
H
e
a
l
t
h
y
97
.
3
1
0
0
95
95
99
86
.
5
91
.
3
85
.
8
88
.
3
88
R
a
s
p
b
e
r
r
y
–
H
e
a
l
t
h
y
96
.
7
1
0
0
94
95
98
.
3
85
89
.
8
84
.
6
86
.
8
86
.
55
S
o
y
b
e
a
n
–
H
e
a
l
t
h
y
97
.
4
1
0
0
95
.
3
95
99
.
1
87
91
.
6
86
.
2
88
.
8
88
.
4
S
q
u
a
sh
-
P
o
w
d
e
r
y
M
i
l
d
e
w
93
.
2
98
91
.
1
93
94
81
84
.
9
79
.
5
82
.
8
82
.
05
S
t
r
a
w
b
e
r
r
y
-
Le
a
f
S
c
o
r
c
h
92
.
4
97
.
5
90
92
.
5
93
80
83
.
7
78
.
7
81
.
8
81
.
05
S
t
r
a
w
b
e
r
r
y
–
H
e
a
l
t
h
y
96
.
8
1
0
0
94
.
7
95
98
.
4
85
.
4
90
.
1
84
.
8
87
.
2
86
.
8
To
ma
t
o
-
B
a
c
t
e
r
i
a
l
S
p
o
t
93
.
5
97
91
92
94
80
.
5
84
.
6
79
82
.
3
81
.
6
To
ma
t
o
-
E
a
r
l
y
B
l
i
g
h
t
94
.
8
98
91
.
7
93
96
.
1
82
86
81
.
2
83
.
8
83
.
25
To
ma
t
o
-
L
a
t
e
B
l
i
g
h
t
95
1
0
0
92
95
96
.
5
82
.
7
86
.
7
82
.
1
84
.
5
84
To
ma
t
o
-
L
e
a
f
M
o
l
d
93
.
2
98
91
.
5
93
95
81
85
80
.
4
82
.
8
82
.
3
To
ma
t
o
-
S
e
p
t
o
r
i
a
Le
a
f
S
p
o
t
94
.
3
1
0
0
92
.
2
95
95
.
7
81
.
5
85
.
5
80
.
9
83
.
3
82
.
8
To
ma
t
o
-
S
p
i
d
e
r
M
i
t
e
s
91
.
7
96
.
6
90
.
6
91
.
6
93
79
.
4
83
.
4
78
.
6
81
.
2
80
.
65
To
ma
t
o
-
T
a
r
g
e
t
S
p
o
t
92
.
8
97
.
7
91
.
7
92
.
7
94
80
.
3
84
79
.
7
82
.
1
81
.
5
To
ma
t
o
-
Y
e
l
l
o
w
L
e
a
f
C
u
r
l
V
i
r
u
s
97
.
4
1
0
0
95
.
5
95
99
85
.
6
90
.
3
85
87
.
4
87
To
ma
t
o
-
M
o
sai
c
V
i
r
u
s
96
.
7
1
0
0
94
.
1
95
98
84
.
4
89
83
86
.
2
85
.
65
To
ma
t
o
–
H
e
a
l
t
h
y
97
1
0
0
95
95
99
.
2
87
91
86
88
.
8
88
.
2
T
ab
le
2
.
T
h
e
p
r
o
p
o
s
ed
a
p
p
r
o
a
ch
’
s
p
r
ec
is
io
n
an
d
lo
s
s
af
ter
3
0
tr
ain
in
g
e
p
o
ch
s
o
n
th
e
Plan
t
Villag
e
d
ataset
I
t
e
r
a
t
i
o
n
s
Tr
a
i
n
i
n
g
p
r
e
c
i
si
o
n
%
V
a
l
i
d
a
t
i
o
n
p
r
e
c
i
si
o
n
%
Tr
a
i
n
i
n
g
l
o
ss
Lo
ss
o
f
v
a
l
i
d
a
t
i
o
n
P
l
a
n
t
v
i
l
l
a
g
e
D
a
t
a
b
a
s
e
10
I
t
e
r
a
t
i
o
n
s
9
4
.
9
6
9
2
.
3
0
.
1
8
7
5
0
.
2
0
1
6
30
I
t
e
r
a
t
i
o
n
s
9
8
.
6
4
9
3
.
5
0
.
0
6
2
3
T
ab
le
3
.
T
h
e
p
r
o
p
o
s
ed
a
p
p
r
o
a
ch
’
s
p
r
ec
is
io
n
an
d
lo
s
s
af
ter
3
0
tr
ain
in
g
e
p
o
ch
s
o
n
th
e
Plan
t
Dis
ea
s
e
d
ataset
I
t
e
r
a
t
i
o
n
s
Tr
a
i
n
i
n
g
p
r
e
c
i
si
o
n
%
V
a
l
i
d
a
t
i
o
n
p
r
e
c
i
si
o
n
%
Tr
a
i
n
i
n
g
l
o
ss
Lo
ss
o
f
v
a
l
i
d
a
t
i
o
n
P
l
a
n
t
D
i
s
e
a
s
e
D
a
t
a
s
e
t
10
I
t
e
r
a
t
i
o
n
s
8
2
.
2
5
8
3
.
1
1
0
.
3
2
4
7
0
.
2
8
0
3
3
0
I
t
e
r
a
t
i
o
n
s
8
5
.
4
3
8
7
.
2
6
0
.
1
5
4
3
5.
CO
M
P
ARA
T
I
V
E
ANA
L
YS
I
S O
F
DE
E
P
L
E
ARNI
NG
AND
H
YB
RI
D
M
E
T
H
O
DS
T
o
th
o
r
o
u
g
h
ly
test
th
e
ef
f
ica
cy
o
f
th
e
p
r
o
p
o
s
ed
c
o
m
b
in
e
d
m
o
d
el
,
co
m
p
ar
ativ
e
ex
p
er
im
en
tal
test
s
wer
e
co
n
d
u
cted
ac
r
o
s
s
f
iv
e
p
o
p
u
lar
p
r
e
-
t
r
ain
ed
C
NN
ar
ch
i
tectu
r
es:
Den
s
eNe
t
-
2
0
1
,
VGGN
et
-
1
9
,
I
n
ce
p
tio
n
V3
,
R
esNet
-
5
0
,
an
d
a
h
y
b
r
i
d
m
o
d
el
b
ased
o
n
th
e
ar
c
h
itectu
r
e
o
f
I
n
ce
p
tio
n
a
n
d
VGGN
et
ca
lled
I
NC
-
VGGN
.
T
h
ese
f
iv
e
m
o
d
els
wer
e
s
elec
t
ed
b
ec
a
u
s
e
o
f
th
eir
p
o
p
u
lar
ity
an
d
k
n
o
wn
ef
f
ec
tiv
e
n
ess
in
v
ar
io
u
s
v
is
io
n
task
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
P
la
n
t d
is
ea
s
e
d
etec
tio
n
a
n
d
c
la
s
s
ifica
tio
n
:
b
a
s
ed
o
n
ma
ch
in
e
lea
r
n
in
g
a
n
d
…
(
E
l
A
r
o
u
s
s
i
El
M
eh
d
i
)
5343
Mo
d
el
p
er
f
o
r
m
an
ce
was
ev
a
lu
ated
ac
r
o
s
s
k
ey
m
etr
ics
in
clu
d
in
g
tr
ain
in
g
p
r
ec
is
io
n
,
v
alid
atio
n
p
r
ec
is
io
n
,
tr
ain
in
g
lo
s
s
,
an
d
v
alid
atio
n
l
o
s
s
,
af
ter
1
0
an
d
3
0
tr
ain
i
n
g
iter
atio
n
s
.
As
in
d
icate
d
in
T
ab
le
4
,
th
e
p
r
o
p
o
s
ed
m
o
d
el
o
u
tp
er
f
o
r
m
e
d
th
e
o
th
e
r
m
eth
o
d
s
o
n
all
m
etr
ics
ac
r
o
s
s
all
iter
atio
n
s
.
Af
ter
o
n
ly
1
0
tr
ain
in
g
iter
atio
n
s
,
th
e
p
r
o
p
o
s
ed
m
o
d
el
ac
h
ie
v
ed
a
tr
ain
i
n
g
p
r
ec
is
io
n
o
f
9
4
.
9
6
%
an
d
a
v
alid
atio
n
p
r
ec
is
io
n
o
f
9
2
.
3
%,
with
a
tr
ain
in
g
lo
s
s
v
alu
e
o
f
0
.
1
8
7
5
alr
ea
d
y
o
u
tp
e
r
f
o
r
m
in
g
m
o
s
t
co
m
p
etin
g
m
o
d
els
ev
en
af
ter
3
0
iter
atio
n
s
.
Fo
llo
win
g
3
0
iter
atio
n
s
,
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
f
u
r
t
h
er
in
cr
ea
s
ed
its
tr
ain
in
g
p
r
ec
is
io
n
to
9
8
.
6
4
%
an
d
v
alid
atio
n
p
r
ec
is
io
n
to
9
3
.
5
%,
ac
h
iev
in
g
a
tr
ain
in
g
lo
s
s
o
f
0
.
0
6
2
3
an
d
a
v
alid
atio
n
lo
s
s
o
f
0
.
2
0
1
6
.
T
h
ese
r
esu
lts
in
d
icate
n
o
t o
n
l
y
r
ap
id
co
n
v
er
g
e
n
ce
b
u
t a
ls
o
th
e
s
tr
o
n
g
g
e
n
er
aliza
tio
n
ab
ilit
y
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
as d
esig
n
ed
.
T
ab
le
4
.
Acc
u
r
ac
y
an
d
er
r
o
r
r
a
tes o
f
d
if
f
er
e
n
t te
ch
n
iq
u
e
af
ter
3
0
tr
ain
in
g
iter
atio
n
s
1
0
I
t
e
r
a
t
i
o
n
s
3
0
I
t
e
r
a
t
i
o
n
s
M
e
t
h
o
d
s
Tr
a
i
n
i
n
g
p
r
e
c
i
si
o
n
%
V
a
l
i
d
a
t
i
o
n
p
r
e
c
i
si
o
n
%
Tr
a
i
n
i
n
g
l
o
ss
Tr
a
i
n
i
n
g
p
r
e
c
i
si
o
n
%
V
a
l
i
d
a
t
i
o
n
p
r
e
c
i
si
o
n
%
Tr
a
i
n
i
n
g
l
o
ss
Lo
ss
o
f
v
a
l
i
d
a
t
i
o
n
D
e
n
seN
e
t
2
0
1
8
0
.
2
7
7
6
.
3
0
0
.
5
7
3
8
4
.
2
0
79
.
00
0
.
4
4
5
0
.
4
9
9
R
e
sN
e
t
5
0
6
5
.
2
0
6
4
.
7
0
1
.
0
0
3
7
0
.
4
0
6
9
.
7
0
0
.
8
3
4
0
,
8
4
4
I
n
c
e
p
t
i
o
n
V
3
8
5
.
6
0
8
2
.
3
0
0
.
4
0
9
9
2
.
1
0
8
5
.
0
0
0
.
2
5
8
0
.
3
7
2
V
G
G
N
e
t
1
9
6
5
.
2
0
6
6
.
7
0
1
.
1
6
4
7
4
.
2
0
7
4
.
8
0
0
.
9
1
6
0
.
9
0
3
I
N
C
-
V
G
G
N
9
3
.
9
0
9
0
.
2
0
0
.
2
1
2
9
7
.
6
0
9
1
.
8
0
0
.
0
8
6
0
.
2
4
1
P
r
o
p
o
se
d
t
e
c
h
n
i
q
u
e
9
4
.
9
6
9
2
.
3
0
0
.
1
8
7
9
8
.
6
4
9
3
.
5
0
0
.
0
6
2
0
.
2
0
2
I
n
co
n
tr
ast,
Den
s
eNe
t
-
2
0
1
,
wh
ile
ca
p
ab
le
o
f
ac
h
iev
in
g
r
elativ
ely
h
ig
h
p
r
ec
is
io
n
,
u
ltima
tely
u
n
d
er
p
er
f
o
r
m
ed
wh
en
co
m
p
a
r
ed
with
th
e
p
r
o
p
o
s
ed
m
eth
o
d
.
Af
ter
3
0
tr
ain
in
g
iter
atio
n
s
,
it
ac
h
iev
ed
a
tr
ain
in
g
p
r
ec
is
io
n
o
f
8
4
.
2
%
an
d
a
v
al
id
atio
n
p
r
ec
is
io
n
o
f
7
9
.
0
%
b
u
t
also
r
ep
o
r
te
d
a
tr
ain
in
g
lo
s
s
o
f
0
.
4
4
5
1
an
d
a
v
alid
atio
n
lo
s
s
o
f
0
.
4
9
8
7
.
Alth
o
u
g
h
th
ese
m
etr
ics
s
u
g
g
est
th
at
Den
s
eNe
t
-
2
0
1
lear
n
ed
s
o
m
e
f
ea
tu
r
es
ef
f
ec
tiv
ely
,
it
is
clea
r
th
er
e
is
a
lack
o
f
p
er
f
o
r
m
an
ce
in
f
ea
tu
r
e
d
is
cr
im
in
atio
n
co
m
p
ar
e
d
to
th
e
h
y
b
r
id
m
eth
o
d
.
T
h
e
o
v
er
all
p
e
r
f
o
r
m
an
ce
o
f
R
esNet
-
5
0
was
lo
wer
t
h
an
t
h
at
o
f
th
e
o
th
e
r
ex
a
m
in
ed
ar
ch
ite
ctu
r
es.
Af
ter
t
h
ir
ty
iter
atio
n
s
,
its
tr
ain
in
g
p
r
ec
is
io
n
was
o
n
ly
7
0
.
4
%
an
d
v
ali
d
atio
n
p
r
ec
is
io
n
6
9
.
7
%,
wh
il
e
th
e
tr
ain
in
g
an
d
v
alid
atio
n
lo
s
s
es
r
em
ain
ed
h
ig
h
at
0
.
8
3
3
8
an
d
0
.
8
4
4
2
,
r
esp
ec
tiv
ely
.
T
h
ese
r
esu
lts
in
d
icate
p
o
o
r
o
v
e
r
all
co
n
v
er
g
en
ce
o
n
th
e
p
la
n
t
d
is
ea
s
e
d
ataset.
T
h
e
r
elativ
ely
lo
w
p
er
f
o
r
m
an
ce
o
f
R
esNet
-
5
0
m
ay
b
e
d
u
e
t
o
its
s
en
s
itiv
ity
to
in
itial
izatio
n
an
d
lear
n
in
g
r
ate
f
ac
to
r
s
th
at
ar
e
p
ar
ticu
lar
ly
p
r
o
b
lem
atic
f
o
r
h
eter
o
g
en
e
o
u
s
an
d
n
o
is
y
d
o
m
ain
s
lik
e
p
lan
t
p
a
th
o
lo
g
y
im
ag
es.
I
n
ce
p
tio
n
V3
p
e
r
f
o
r
m
ed
b
etter
,
ac
h
iev
in
g
9
2
.
1
%
tr
ai
n
in
g
p
r
ec
is
io
n
an
d
8
5
.
0
%
v
alid
atio
n
p
r
ec
is
io
n
af
te
r
3
0
iter
atio
n
s
,
alo
n
g
with
a
tr
ain
in
g
lo
s
s
o
f
0
.
2
5
7
6
a
n
d
a
v
alid
atio
n
l
o
s
s
o
f
0
.
3
7
1
7
.
W
h
ile
th
ese
f
in
d
in
g
s
a
r
e
r
esp
ec
ta
b
le,
th
e
y
s
till
lag
g
e
d
b
eh
in
d
t
h
e
p
r
o
p
o
s
ed
m
o
d
el,
p
ar
ticu
lar
ly
i
n
v
alid
atio
n
ac
c
u
r
ac
y
,
wh
e
r
e
f
i
n
e
-
g
r
ain
ed
d
is
ea
s
e
class
if
icatio
n
r
em
ain
s
a
c
h
allen
g
e.
VGGN
et
-
1
9
d
eliv
er
ed
d
ec
e
n
t
r
esu
lts
d
u
e
to
its
d
ep
th
an
d
s
tr
aig
h
tf
o
r
war
d
ar
c
h
itectu
r
e.
Af
ter
3
0
tr
ain
in
g
iter
atio
n
s
,
it
ac
h
iev
ed
a
tr
ain
in
g
p
r
ec
is
io
n
o
f
7
4
.
2
%
an
d
v
alid
ati
o
n
p
r
ec
i
s
io
n
o
f
7
4
.
8
%,
with
r
elativ
ely
h
ig
h
tr
ain
in
g
an
d
v
alid
atio
n
lo
s
s
es
o
f
0
.
9
1
6
2
an
d
0
.
9
0
2
6
,
r
esp
ec
tiv
el
y
.
T
h
ese
r
esu
lts
in
d
icate
th
at
id
e
n
tify
in
g
an
o
p
tim
al
f
ea
tu
r
e
r
ep
r
esen
tatio
n
f
o
r
g
e
n
er
aliza
tio
n
is
d
if
f
icu
lt with
co
m
p
lex
p
lan
t d
is
ea
s
e
d
atasets
.
T
h
e
h
y
b
r
i
d
I
N
C
-
V
G
G
N
m
o
d
e
l
p
e
r
f
o
r
m
e
d
b
e
t
t
e
r
t
h
a
n
t
h
e
s
t
a
n
d
a
r
d
C
NN
m
o
d
e
l
s
,
a
c
h
ie
v
i
n
g
9
7
.
6
%
t
r
a
i
n
i
n
g
p
r
e
c
i
s
i
o
n
a
n
d
9
1
.
8
%
v
a
l
i
d
a
t
i
o
n
p
r
e
c
is
i
o
n
a
f
t
e
r
3
0
i
t
e
r
a
t
i
o
n
s
,
w
it
h
l
o
w
t
r
a
i
n
i
n
g
a
n
d
v
a
l
i
d
a
t
i
o
n
l
o
s
s
es
o
f
0
.
0
8
5
6
a
n
d
0
.
2
4
0
9
,
r
e
s
p
e
c
t
i
v
e
ly
.
H
o
w
e
v
e
r
,
d
e
s
p
i
t
e
t
h
e
s
e
s
t
r
o
n
g
r
e
s
u
l
ts
,
i
t
s
t
i
ll
f
a
i
le
d
t
o
o
u
t
p
e
r
f
o
r
m
t
h
e
p
r
o
p
o
s
e
d
m
e
t
h
o
d
i
n
t
e
r
m
s
o
f
t
r
a
i
n
i
n
g
s
p
ee
d
a
n
d
o
v
e
r
a
l
l
p
e
r
f
o
r
m
a
n
c
e
.
T
h
e
m
a
r
k
e
d
i
m
p
r
o
v
e
m
e
n
t
d
e
l
i
v
er
e
d
b
y
t
h
e
p
r
o
p
o
s
e
d
m
o
d
e
l
i
s
d
u
e
t
o
i
t
s
h
y
b
r
i
d
n
a
t
u
r
e
,
w
h
e
r
e
i
n
f
e
a
t
u
r
es
d
e
r
i
v
e
d
f
r
o
m
A
l
e
x
N
e
t
w
e
r
e
c
o
m
b
i
n
ed
w
i
t
h
h
a
n
d
c
r
a
f
t
e
d
d
e
s
c
r
i
p
t
o
r
s
f
r
o
m
Hi
l
b
e
r
t
s
p
a
c
e
-
b
a
s
e
d
(
H
e
s
s
)
-
C
o
H
O
G
.
T
h
e
h
y
b
r
i
d
m
o
d
e
l
u
t
i
l
i
z
e
d
b
o
t
h
t
y
p
e
s
o
f
r
e
p
r
e
s
e
n
t
at
i
o
n
t
h
e
a
b
s
t
r
a
c
t
,
h
i
g
h
-
l
e
v
el
h
i
e
r
a
r
c
h
i
c
a
l
l
e
a
r
n
i
n
g
p
r
o
v
i
d
e
d
b
y
C
N
Ns
a
n
d
t
h
e
t
e
x
t
u
r
e
-
s
e
n
s
it
iv
e
,
r
o
t
a
t
i
o
n
-
i
n
v
a
r
i
a
n
t
q
u
a
l
i
t
i
es
o
f
h
a
n
d
c
r
a
f
t
e
d
d
es
c
r
ip
t
o
r
s
.
T
o
d
e
c
r
e
as
e
t
h
e
s
i
z
e
o
f
t
h
e
r
e
p
r
e
s
e
n
ta
t
i
o
n
t
o
a
c
o
m
p
u
t
a
t
i
o
n
a
l
l
y
f
r
ie
n
d
l
y
1
×
1
2
8
v
e
c
t
o
r
w
h
i
l
e
r
e
t
ai
n
i
n
g
d
i
s
c
r
i
m
i
n
a
t
i
v
e
i
n
f
o
r
m
a
t
i
o
n
,
t
h
e
f
e
a
t
u
r
e
s
we
r
e
c
o
m
p
r
e
s
s
e
d
u
s
i
n
g
P
C
A
.
T
h
is
b
alan
ce
b
etwe
en
f
ea
tu
r
e
r
ich
n
ess
an
d
r
ed
u
ce
d
d
i
m
en
s
io
n
ality
co
n
tr
ib
u
ted
to
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
ac
r
o
s
s
all
m
etr
ic
s
.
Ov
er
all,
th
e
co
m
p
ar
ativ
e
r
e
s
u
lts
s
u
b
s
tan
tiate
th
at
th
e
h
y
b
r
id
m
o
d
el
is
r
o
b
u
s
t
an
d
o
u
t
p
er
f
o
r
m
s
s
tate
-
of
-
th
e
-
ar
t
C
NN
ar
ch
itectu
r
es
test
ed
in
th
is
s
tu
d
y
,
p
ar
ticu
lar
l
y
af
ter
tr
ain
in
g
f
o
r
3
0
ep
o
ch
s
.
I
t
r
ep
r
esen
ts
a
v
iab
le
an
d
p
r
o
m
is
in
g
ap
p
r
o
ac
h
f
o
r
lar
g
e
-
s
ca
le,
r
ea
l
-
wo
r
ld
p
la
n
t
d
is
ea
s
e
r
ec
o
g
n
itio
n
wh
er
e
ac
cu
r
ac
y
a
n
d
ef
f
icien
c
y
ar
e
ess
en
tial.
T
ab
le
5
p
r
o
v
id
es
a
s
im
ilar
an
aly
s
is
o
f
th
e
m
ajo
r
d
ee
p
lear
n
in
g
ar
ch
itectu
r
es
(
an
d
t
h
e
p
r
o
p
o
s
e
d
h
y
b
r
id
m
o
d
el)
to
b
etter
ex
p
l
ain
th
e
r
elativ
e
ad
v
an
tag
es
an
d
d
is
ad
v
an
tag
es
o
f
ea
ch
m
eth
o
d
an
al
y
ze
d
in
th
is
s
tu
d
y
.
E
ac
h
m
eth
o
d
was
ass
ess
ed
b
ased
o
n
its
co
m
p
u
tatio
n
al
ch
ar
ac
ter
is
tics
,
lear
n
in
g
ab
ilit
y
,
an
d
ap
p
licab
il
ity
to
p
lan
t
d
is
ea
s
e
d
etec
tio
n
.
C
o
n
v
en
tio
n
al
C
NN
-
b
ased
m
eth
o
d
s
lik
e
Alex
Net,
VGGNet
-
1
9
,
R
esNet
-
5
0
,
I
n
c
ep
tio
n
V3
,
a
n
d
Den
s
eNe
t
-
2
0
1
ea
ch
o
f
f
er
s
p
ec
if
ic
a
d
v
an
t
ag
es
s
u
ch
as
f
ast
tr
ain
in
g
,
d
ee
p
f
ea
tu
r
e
ex
t
r
ac
tio
n
,
o
r
s
tr
o
n
g
g
r
a
d
ien
t
f
lo
w
b
u
t
also
p
r
esen
t
d
r
awb
ac
k
s
lik
e
s
en
s
itiv
ity
to
in
itializatio
n
,
h
ig
h
m
e
m
o
r
y
r
eq
u
ir
em
en
ts
,
an
d
lo
n
g
tr
ain
i
n
g
tim
es.
T
h
e
I
NC
-
VGGN
h
y
b
r
id
m
o
d
el,
wh
ich
f
u
s
es
in
ter
m
ed
iate
an
d
v
is
u
al
s
ty
les
f
r
o
m
d
if
f
er
en
t
a
r
c
h
itectu
r
es,
o
f
f
er
s
s
o
m
e
tr
ain
in
g
ad
v
a
n
tag
es
b
u
t
in
tr
o
d
u
ce
s
g
r
ea
ter
co
m
p
lex
ity
an
d
h
ig
h
er
tr
ain
i
n
g
c
o
s
ts
.
I
n
c
o
n
tr
ast,
th
e
p
r
o
p
o
s
ed
m
o
d
el
ef
f
ec
tiv
ely
ad
d
r
ess
es
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
3
3
6
-
5
3
4
6
5344
th
ese
d
is
ad
v
an
tag
es
b
y
in
teg
r
atin
g
d
ee
p
lear
n
in
g
with
h
a
n
d
cr
af
ted
f
ea
tu
r
es
an
d
d
im
en
s
io
n
ality
r
ed
u
ctio
n
.
I
ts
two
p
r
im
ar
y
a
d
v
an
ta
g
es
ar
e
th
eir
r
o
b
u
s
tn
ess
to
n
o
is
e
an
d
r
o
tatio
n
,
an
d
its
h
ig
h
d
eg
r
e
e
o
f
ac
cu
r
ac
y
i
n
a
co
m
p
ac
t
r
ep
r
esen
tatio
n
d
e
m
o
n
s
tr
ates
its
ap
p
licab
ilit
y
.
T
h
er
ef
o
r
e
,
th
e
p
r
o
p
o
s
ed
h
y
b
r
id
ar
ch
itectu
r
e
h
as
en
o
u
g
h
th
eo
r
etica
l
b
ac
k
i
n
g
f
o
r
f
u
tu
r
e
c
o
n
s
id
er
atio
n
a
n
d
p
o
ten
tial
ap
p
licatio
n
in
r
ea
l
-
w
o
r
ld
d
ep
l
o
y
m
en
t
o
n
f
ar
m
s
.
T
h
is
co
m
p
ar
ativ
e
in
v
esti
g
atio
n
n
o
t
o
n
l
y
v
alid
ates
th
e
r
esear
ch
r
elev
an
ce
o
f
th
e
p
r
o
p
o
s
ed
h
y
b
r
i
d
m
o
d
e
l
but
also
p
r
o
v
id
es
d
ir
ec
tio
n
to
o
th
er
in
v
esti
g
ato
r
s
an
d
p
r
ac
titi
o
n
er
s
lo
o
k
in
g
to
s
elec
t
o
r
alter
th
eir
d
ee
p
lear
n
in
g
m
o
d
els as p
er
co
n
tex
t
u
al
lim
itatio
n
s
.
T
ab
le
5
.
Key
s
tr
en
g
th
s
an
d
lim
itatio
n
s
o
f
d
if
f
e
r
en
t d
ee
p
lea
r
n
in
g
tech
n
iq
u
es
M
e
t
h
o
d
P
o
si
t
i
v
e
a
s
p
e
c
t
s
N
e
g
a
t
i
v
e
a
s
p
e
c
t
s
A
l
e
x
N
e
t
-
F
a
st
t
r
a
i
n
i
n
g
-
S
i
mp
l
e
a
r
c
h
i
t
e
c
t
u
r
e
-
Ef
f
e
c
t
i
v
e
f
o
r
b
a
si
c
i
ma
g
e
c
l
a
ssi
f
i
c
a
t
i
o
n
-
S
h
a
l
l
o
w
d
e
p
t
h
-
Li
mi
t
e
d
g
e
n
e
r
a
l
i
z
a
t
i
o
n
o
n
c
o
m
p
l
e
x
d
a
t
a
se
t
s
V
G
G
N
e
t
-
19
-
D
e
e
p
n
e
t
w
o
r
k
w
i
t
h
g
o
o
d
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
-
P
e
r
f
o
r
ms w
e
l
l
w
i
t
h
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
-
H
i
g
h
n
u
mb
e
r
o
f
p
a
r
a
me
t
e
r
s
-
R
e
q
u
i
r
e
s
mo
r
e
mem
o
r
y
a
n
d
c
o
m
p
u
t
a
t
i
o
n
a
l
r
e
so
u
r
c
e
s
R
e
sN
e
t
-
50
-
R
e
s
i
d
u
a
l
c
o
n
n
e
c
t
i
o
n
s e
n
a
b
l
e
d
e
e
p
t
r
a
i
n
i
n
g
-
M
i
t
i
g
a
t
e
s
v
a
n
i
s
h
i
n
g
g
r
a
d
i
e
n
t
p
r
o
b
l
e
m
-
S
e
n
s
i
t
i
v
e
t
o
l
e
a
r
n
i
n
g
r
a
t
e
a
n
d
i
n
i
t
i
a
l
i
z
a
t
i
o
n
-
Lo
n
g
e
r
t
r
a
i
n
i
n
g
t
i
me
I
n
c
e
p
t
i
o
n
V
3
-
Ef
f
i
c
i
e
n
t
m
u
l
t
i
-
s
c
a
l
e
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
-
G
o
o
d
t
r
a
d
e
-
o
f
f
b
e
t
w
e
e
n
d
e
p
t
h
a
n
d
s
p
e
e
d
-
C
o
m
p
l
e
x
a
r
c
h
i
t
e
c
t
u
r
e
-
R
e
q
u
i
r
e
s
i
ma
g
e
i
n
p
u
t
s
t
a
n
d
a
r
d
i
z
a
t
i
o
n
D
e
n
seN
e
t
-
2
0
1
-
S
t
r
o
n
g
g
r
a
d
i
e
n
t
f
l
o
w
-
F
e
a
t
u
r
e
r
e
u
s
e
i
mp
r
o
v
e
s
e
f
f
i
c
i
e
n
c
y
a
n
d
a
c
c
u
r
a
c
y
-
S
l
o
w
e
r
t
r
a
i
n
i
n
g
d
u
e
t
o
d
e
n
se
c
o
n
n
e
c
t
i
o
n
s
-
M
e
m
o
r
y
-
i
n
t
e
n
s
i
v
e
-
I
n
c
r
e
a
se
d
a
r
c
h
i
t
e
c
t
u
r
a
l
c
o
m
p
l
e
x
i
t
y
-
H
i
g
h
e
r
t
r
a
i
n
i
n
g
a
n
d
i
n
f
e
r
e
n
c
e
t
i
me
I
N
C
-
V
G
G
N
-
H
y
b
r
i
d
a
r
c
h
i
t
e
c
t
u
r
e
l
e
v
e
r
a
g
e
s VG
G
a
n
d
I
n
c
e
p
t
i
o
n
st
r
e
n
g
t
h
s
-
H
i
g
h
c
l
a
s
si
f
i
c
a
t
i
o
n
a
c
c
u
r
a
c
y
P
r
o
p
o
se
d
m
o
d
e
l
-
C
o
m
b
i
n
e
s
d
e
e
p
a
n
d
h
a
n
d
c
r
a
f
t
e
d
f
e
a
t
u
r
e
s
-
R
o
b
u
st
t
o
n
o
i
se
a
n
d
r
o
t
a
t
i
o
n
-
H
i
g
h
a
c
c
u
r
a
c
y
w
i
t
h
l
o
w
d
i
me
n
s
i
o
n
(
1
×
1
2
8
v
e
c
t
o
r
)
-
P
C
A
ma
y
r
e
d
u
c
e
s
u
b
t
l
e
b
u
t
u
sef
u
l
f
e
a
t
u
r
e
s
-
R
e
q
u
i
r
e
s
p
a
r
a
me
t
e
r
t
u
n
i
n
g
f
o
r
f
u
s
i
o
n
5.
CO
NCLU
SI
O
N
Fo
r
m
an
y
y
ea
r
s
,
p
lan
t
d
is
ea
s
es
h
av
e
b
ee
n
an
im
p
o
r
ta
n
t
ar
ea
o
f
co
n
ce
r
n
i
n
ag
r
icu
ltu
r
e
a
n
d
r
ep
r
esen
t
a
co
n
s
id
er
ab
le
th
r
ea
t
to
f
o
o
d
m
an
u
f
ac
tu
r
i
n
g
s
ec
u
r
ity
.
I
n
ex
tr
em
e
ca
s
es
th
ese
d
is
ea
s
e
s
ca
n
lead
to
to
tal
cr
o
p
lo
s
s
es,
wi
th
s
p
illo
v
er
ef
f
ec
ts
th
at
ca
n
in
clu
d
e
ec
o
n
o
m
ic
s
h
o
ck
s
,
f
o
o
d
s
ca
r
city
an
d
b
r
ea
k
s
in
s
u
p
p
ly
ch
ain
s
.
Su
ch
ch
allen
g
es a
r
e
p
ar
ticu
lar
ly
p
r
o
n
o
u
n
ce
d
in
d
e
v
elo
p
in
g
g
eo
g
r
ap
h
ies,
wh
er
e
f
ar
m
er
s
lack
ac
ce
s
s
to
m
o
d
er
n
ag
r
icu
ltu
r
al
tech
n
o
lo
g
ies
an
d
th
e
ec
o
n
o
m
ic
well
-
b
ein
g
o
f
s
m
allh
o
ld
er
f
a
r
m
er
s
ar
e
d
ir
ec
tly
lin
k
ed
to
cr
o
p
y
ield
s
.
Plan
t
d
is
ea
s
es
h
av
e
s
o
cial
an
d
e
n
v
ir
o
n
m
en
tal
im
p
ac
ts
in
ad
d
itio
n
to
th
eir
ec
o
n
o
m
ic
ef
f
ec
ts
,
f
u
elin
g
p
o
v
er
ty
,
d
im
in
is
h
in
g
b
io
d
iv
er
s
ity
,
an
d
lead
in
g
to
u
n
s
u
s
tain
ab
le
f
ar
m
in
g
p
r
ac
tices.
W
it
h
in
cr
ea
s
in
g
g
lo
b
a
l
p
o
p
u
latio
n
g
r
o
wth
,
a
n
d
th
e
c
o
m
in
g
th
r
ea
ts
o
f
clim
ate
ch
a
n
g
e,
th
e
d
r
iv
e
f
o
r
m
o
r
e
in
n
o
v
ativ
e
an
d
ef
f
icien
t
way
s
to
co
m
b
at
p
lan
t
d
is
ea
s
es
is
cr
itical.
Am
o
n
g
th
e
in
f
o
r
m
atio
n
in
a
g
r
icu
ltu
r
e,
th
er
e
is
a
g
r
ea
t
d
em
an
d
f
o
r
th
e
ea
r
ly
id
e
n
tific
atio
n
o
f
p
la
n
t
d
is
ea
s
e.
E
ar
ly
d
etec
tio
n
o
f
p
lan
t
d
is
ea
s
es
is
cr
itical
ly
im
p
o
r
tan
t
in
th
e
f
ield
o
f
ag
r
icu
ltu
r
al
i
n
f
o
r
m
atio
n
.
Utilizin
g
ea
r
l
y
d
etec
tio
n
th
r
o
u
g
h
d
is
ea
s
e
m
an
ag
em
en
t
en
ab
les
tim
ely
in
ter
v
e
n
tio
n
an
d
th
e
a
p
p
licatio
n
o
f
tar
g
eted
tr
ea
tm
en
ts
th
at
m
ay
d
r
asti
ca
lly
r
ed
u
ce
cr
o
p
lo
s
s
es in
th
e
lo
n
g
r
u
n
.
I
t
also
p
r
o
v
id
es
a
m
ea
n
s
o
f
d
ec
r
ea
s
in
g
r
elian
ce
o
n
ch
e
m
ical
p
esti
cid
es,
wh
ich
f
r
eq
u
en
tly
ca
u
s
e
s
o
il
d
eg
r
ad
atio
n
,
wate
r
c
o
n
tam
in
a
tio
n
,
an
d
d
etr
im
e
n
tal
ef
f
ec
ts
o
n
n
o
n
-
tar
g
et
o
r
g
an
is
m
s
lik
e
p
o
llin
ato
r
s
.
E
ar
ly
d
etec
tio
n
tech
n
o
l
o
g
ies
h
av
e
t
h
e
p
o
ten
tial
to
m
ak
e
a
s
ig
n
i
f
ican
t
im
p
ac
t
if
th
ey
ca
n
s
h
o
w
th
at
th
ey
ca
n
h
elp
m
ee
t
th
e
p
r
in
cip
les
o
f
s
u
s
tain
ab
le
ag
r
icu
ltu
r
e.
C
u
r
r
en
tly
d
ee
p
lear
n
in
g
m
eth
o
d
s
,
esp
ec
ially
C
NNs,
h
av
e
s
h
o
wn
o
u
ts
tan
d
in
g
p
e
r
f
o
r
m
a
n
ce
o
n
m
a
n
y
p
r
o
b
lem
s
in
v
o
l
v
ed
in
th
e
f
ield
o
f
d
is
ea
s
e
d
iag
n
o
s
is
.
No
t
o
n
ly
wo
u
ld
th
ese
m
eth
o
d
s
p
er
f
o
r
m
ex
ce
p
tio
n
ally
o
n
h
ig
h
-
d
im
en
s
io
n
al
im
ag
e
d
ata,
b
u
t
th
ey
al
s
o
en
ab
le
r
etr
ain
in
g
o
n
u
p
d
ated
d
atasets
,
allo
win
g
th
em
to
b
e
s
ea
m
less
ly
ad
ap
ted
to
ch
a
n
g
in
g
ag
r
icu
ltu
r
al
co
n
d
itio
n
s
o
v
e
r
tim
e.
T
h
is
s
ec
tio
n
p
r
esen
ted
an
in
n
o
v
ativ
e
m
et
h
o
d
th
at
c
o
m
b
in
e
s
C
NN
,
s
p
ec
if
ically
Alex
Net,
with
th
e
Hess
ian
m
atr
ix
to
ca
lcu
late
t
h
e
eig
e
n
v
alu
es
o
f
th
e
im
ag
e
s
u
r
f
ac
e.
B
y
in
co
r
p
o
r
atin
g
th
e
Hess
ian
m
atr
ix
,
o
u
r
a
p
p
r
o
ac
h
en
h
an
ce
s
th
e
m
o
d
el'
s
ab
ilit
y
to
id
en
tify
s
u
b
tle
tex
tu
r
al
f
ea
tu
r
es
an
d
v
ar
iatio
n
s
in
p
lan
t
im
ag
es
th
at
m
ay
o
th
er
wis
e
g
o
u
n
d
etec
ted
.
Fu
r
t
h
er
m
o
r
e
,
we
u
s
ed
th
e
PC
A
alg
o
r
ith
m
h
ar
n
ess
ed
f
o
r
m
in
im
i
ze
s
s
ize,
en
ab
lin
g
th
e
ex
tr
ac
tio
n
o
f
ess
en
tial
in
f
o
r
m
atio
n
wh
ile
d
ec
r
ea
s
in
g
c
o
m
p
u
tatio
n
al
o
v
e
r
h
ea
d
.
T
h
is
n
esti
n
g
ad
d
r
ess
es
v
ar
io
u
s
p
r
ac
tical
p
u
r
v
iews
,
s
u
ch
as
p
r
o
ce
s
s
in
g
tim
e
an
d
s
ca
lab
ilit
y
,
wh
ich
ar
e
c
r
itical
f
o
r
wid
esp
r
ea
d
ad
o
p
tio
n
.
T
h
e
r
esu
lts
o
f
o
u
r
e
x
p
er
im
en
ts
d
e
m
o
n
s
tr
ate
th
e
e
f
f
ec
tiv
en
ess
o
f
our
p
r
o
p
o
s
ed
m
eth
o
d
ac
h
iev
in
g
a
n
o
tab
le
s
cr
ee
n
in
g
ac
cu
r
ac
y
o
f
9
1
.
8
3
%
o
n
th
e
Plan
tDis
ea
s
e
d
ataset
an
d
an
im
p
o
s
in
g
9
3
.
6
7
%
ac
cu
r
ac
y
o
n
t
h
e
Plan
tVillag
e
d
ataset.
T
h
e
r
es
u
lts
o
b
tain
ed
s
h
o
w
m
o
r
e
th
a
n
th
e
r
eliab
ilit
y
an
d
r
o
b
u
s
tn
e
s
s
o
f
o
u
r
alg
o
r
ith
m
p
r
o
p
o
s
e
th
ey
also
s
h
o
w
also
h
ig
h
lig
h
t
its
p
o
ten
tial
to
h
as
ex
ce
ed
ex
is
tin
g
al
g
o
r
ith
m
s
in
th
e
f
ield
o
f
p
lan
t
d
is
ea
s
e
class
if
icatio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
P
la
n
t d
is
ea
s
e
d
etec
tio
n
a
n
d
c
la
s
s
ifica
tio
n
:
b
a
s
ed
o
n
ma
ch
in
e
lea
r
n
in
g
a
n
d
…
(
E
l
A
r
o
u
s
s
i
El
M
eh
d
i
)
5345
T
h
e
p
er
ce
n
ta
g
e
o
f
p
r
ec
is
io
n
ac
h
iev
ed
co
u
ld
h
a
v
e
tr
an
s
f
o
r
m
ativ
e
im
p
licatio
n
s
f
o
r
a
g
r
icu
ltu
r
al
p
r
ac
tices,
f
ac
ilit
atin
g
clar
ity
o
f
d
ec
is
io
n
s
f
o
r
f
ar
m
e
r
s
an
d
th
e
u
s
e
o
f
tec
h
n
o
lo
g
y
in
t
r
ad
itio
n
al
f
a
r
m
in
g
s
y
s
tem
s
.
Fu
r
th
er
m
o
r
e,
th
e
r
es
u
lts
o
b
tain
ed
s
h
o
w
th
e
a
b
ilit
y
o
f
ar
tific
ial
in
tellig
en
ce
to
b
r
i
d
g
e
th
e
g
ap
b
etwe
en
cu
ttin
g
-
ed
g
e
r
esear
ch
a
n
d
p
r
ac
tical
ap
p
licatio
n
s
in
ag
r
icu
ltu
r
e.
I
n
o
u
r
f
u
tu
r
e
p
r
o
jects,
we
will
p
r
o
p
o
s
e
d
ep
lo
y
m
e
n
t
ap
p
r
o
ac
h
es
o
n
m
o
b
ile
d
ev
ices
u
s
in
g
c
o
n
n
ec
te
d
o
b
jects
to
f
ac
ilit
ate
th
e
m
o
n
it
o
r
in
g
a
n
d
au
t
o
m
atic
r
ec
o
g
n
itio
n
o
f
a
b
r
o
ad
e
r
s
p
ec
tr
u
m
o
f
d
is
ea
s
es
at
v
ar
io
u
s
lo
ca
tio
n
s
in
th
e
p
lan
t
at
v
ar
io
u
s
s
tag
es
o
f
d
is
ea
s
e
p
r
o
g
r
ess
io
n
.
T
h
e
in
teg
r
atio
n
o
f
th
is
tech
n
o
l
o
g
y
with
m
o
b
il
e
d
ev
ices
co
u
ld
d
em
o
c
r
atize
ac
ce
s
s
to
ad
v
an
ce
d
d
iag
n
o
s
tic
to
o
ls
,
p
ar
ticu
lar
l
y
i
n
r
u
r
al
a
n
d
u
n
d
e
r
s
er
v
ed
ar
ea
s
.
B
y
em
p
o
wer
in
g
f
a
r
m
er
s
wit
h
u
s
er
-
f
r
ien
d
l
y
an
d
co
s
t
-
ef
f
ec
tiv
e
s
o
lu
tio
n
s
,
s
u
ch
tech
n
o
lo
g
ies
h
av
e
th
e
p
o
ten
tial
to
tr
an
s
f
o
r
m
p
r
ec
is
io
n
ag
r
icu
ltu
r
e,
r
e
d
u
cin
g
y
ield
g
ap
s
an
d
f
o
s
ter
in
g
r
esil
ien
ce
ag
ain
s
t e
n
v
ir
o
n
m
e
n
tal
s
tr
ess
o
r
s
.
I
n
p
ar
allel,
we
g
o
al
to
p
r
o
l
o
n
g
th
e
r
eq
u
est
o
f
o
u
r
alg
o
r
ith
m
to
r
ea
l
-
w
o
r
ld
s
itu
atio
n
s
,
in
clu
s
iv
e
co
m
p
u
ter
-
aid
ed
d
iag
n
o
s
is
,
th
e
r
eb
y
co
n
d
u
ci
v
e
to
a
d
v
an
ce
m
e
n
ts
in
ag
r
icu
ltu
r
al
in
f
o
r
m
atio
n
s
y
s
tem
s
.
B
ey
o
n
d
d
is
ea
s
e
class
if
icatio
n
,
o
u
r
s
tr
u
ctu
r
e
is
allo
wed
t
o
b
e
ex
p
an
d
ed
to
in
clu
d
e
p
r
ed
ictiv
e
d
esig
n
,
e
n
ab
lin
g
s
tak
eh
o
ld
er
s
to
an
ticip
ate
d
is
e
ase
o
u
tb
r
ea
k
s
b
ased
o
n
en
v
ir
o
n
m
en
tal
an
d
clim
atic
d
ata.
T
h
i
s
in
teg
r
atio
n
wo
u
ld
n
o
t
o
n
ly
im
p
r
o
v
e
cr
o
p
m
an
ag
em
en
t
s
tr
ateg
ies
b
u
t
also
in
f
o
r
m
r
e
g
io
n
al
an
d
n
atio
n
al
p
o
licy
d
ec
is
io
n
s
,
co
n
tr
ib
u
tin
g
to
a
m
o
r
e
s
u
s
tain
ab
le
an
d
s
ec
u
r
e
a
g
r
icu
ltu
r
al
ec
o
s
y
s
tem
.
I
n
ad
d
itio
n
,
th
e
a
p
p
r
o
ac
h
co
u
l
d
s
u
p
p
o
r
t
th
e
d
ev
elo
p
m
e
n
t
o
f
au
to
m
ated
ag
r
icu
ltu
r
al
r
o
b
o
ts
ca
p
ab
le
o
f
p
er
f
o
r
m
in
g
task
s
s
u
ch
as
tar
g
eted
s
p
r
ay
in
g
,
h
ar
v
esti
n
g
,
o
r
p
r
u
n
in
g
b
ased
o
n
r
ea
l
-
tim
e
d
is
ea
s
e
d
etec
tio
n
.
Mo
r
e
o
v
er
,
th
e
p
o
ten
tial
o
f
th
is
tech
n
o
lo
g
y
ex
ten
d
s
to
in
ter
d
is
cip
lin
ar
y
a
p
p
licatio
n
s
.
Fo
r
ex
am
p
le,
its
in
teg
r
atio
n
in
to
f
o
o
d
s
u
p
p
l
y
ch
ain
s
y
s
tem
s
co
u
ld
en
s
u
r
e
q
u
ality
co
n
tr
o
l
b
y
id
e
n
tify
in
g
d
is
ea
s
ed
cr
o
p
s
d
u
r
in
g
p
r
o
ce
s
s
in
g
o
r
s
to
r
a
g
e,
th
e
r
e
b
y
m
in
im
izin
g
f
o
o
d
waste.
W
e
will
co
llab
o
r
ate
with
ec
o
lo
g
ical
m
o
n
ito
r
in
g
s
y
s
tem
s
to
f
u
r
th
er
ex
p
lo
it
ec
o
lo
g
ical
m
o
n
ito
r
in
g
tech
n
o
lo
g
y
to
ass
ess
th
e
h
ea
lth
o
f
n
atu
r
al
p
lan
t
p
o
p
u
latio
n
s
,
as
well
as
f
o
r
co
n
s
er
v
atio
n
ef
f
o
r
ts
.
W
e
b
eliev
e
th
at
th
r
o
u
g
h
th
ese
in
ter
d
is
cip
lin
ar
y
o
p
p
o
r
t
u
n
ities
,
o
u
r
ap
p
r
o
a
ch
co
u
ld
s
er
v
e
as
a
cr
itical
to
o
l
to
ad
d
r
ess
g
lo
b
al
ch
allen
g
es,
f
r
o
m
f
o
o
d
s
ec
u
r
ity
to
en
v
ir
o
n
m
en
tal
s
u
s
tain
ab
ilit
y
.
RE
F
E
R
E
NC
E
S
[
1
]
W
o
r
l
d
P
e
r
s
p
e
c
t
i
v
e
,
“
S
t
a
t
i
st
i
c
s
,
”
S
c
h
o
o
l
o
f
A
p
p
l
i
e
d
Po
l
i
t
i
c
s
F
a
c
u
l
t
y
o
f
A
rt
s
a
n
d
H
u
m
a
n
i
t
i
e
s
U
n
i
v
e
rsi
t
y
o
f
S
h
e
r
b
ro
o
k
e
.
h
t
t
p
s
:
/
/
p
e
r
sp
e
c
t
i
v
e
.
u
s
h
e
r
b
r
o
o
k
e
.
c
a
/
b
i
l
a
n
/
serv
l
e
t
/
B
M
Te
n
d
a
n
c
e
S
t
a
t
P
a
y
s (a
c
c
e
sse
d
Ja
n
.
3
0
,
2
0
2
5
)
.
[
2
]
S
.
S
a
n
k
a
r
a
n
,
A
.
M
i
s
h
r
a
,
R
.
E
h
sa
n
i
,
a
n
d
C
.
D
a
v
i
s,
“
A
r
e
v
i
e
w
o
f
a
d
v
a
n
c
e
d
t
e
c
h
n
i
q
u
e
s
f
o
r
d
e
t
e
c
t
i
n
g
p
l
a
n
t
d
i
se
a
ses,
”
C
o
m
p
u
t
e
rs
a
n
d
El
e
c
t
r
o
n
i
c
s
i
n
A
g
r
i
c
u
l
t
u
r
e
,
v
o
l
.
7
2
,
n
o
.
1
,
p
p
.
1
–
1
3
,
J
u
n
.
2
0
1
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
a
g
.
2
0
1
0
.
0
2
.
0
0
7
.
[
3
]
R
.
G
.
D
a
w
o
d
a
n
d
C
.
D
o
b
r
e
,
“
U
p
p
e
r
a
n
d
l
o
w
e
r
l
e
a
f
si
d
e
d
e
t
e
c
t
i
o
n
w
i
t
h
mac
h
i
n
e
l
e
a
r
n
i
n
g
me
t
h
o
d
s,”
S
e
n
so
rs
,
v
o
l
.
2
2
,
n
o
.
7
,
p
.
2
6
9
6
,
M
a
r
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
2
0
7
2
6
9
6
.
[
4
]
K
.
P
.
P
a
n
i
g
r
a
h
i
,
A
.
K
.
S
a
h
o
o
,
a
n
d
H
.
D
a
s,
“
A
C
N
N
a
p
p
r
o
a
c
h
f
o
r
c
o
r
n
l
e
a
v
e
s
d
i
s
e
a
s
e
d
e
t
e
c
t
i
o
n
t
o
s
u
p
p
o
r
t
d
i
g
i
t
a
l
a
g
r
i
c
u
l
t
u
r
a
l
sy
st
e
m,
”
i
n
2
0
2
0
4
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
T
re
n
d
s
i
n
El
e
c
t
r
o
n
i
c
s
a
n
d
I
n
f
o
rm
a
t
i
c
s
(
I
C
O
EI
)
(
4
8
1
8
4
)
,
Ju
n
.
2
0
2
0
,
p
p
.
6
7
8
–
6
8
3
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
O
EI
4
8
1
8
4
.
2
0
2
0
.
9
1
4
2
8
7
1
.
[
5
]
A
.
R
a
mc
h
a
r
a
n
e
t
a
l
.
,
“
A
m
o
b
i
l
e
-
b
a
se
d
d
e
e
p
l
e
a
r
n
i
n
g
m
o
d
e
l
f
o
r
c
a
ss
a
v
a
d
i
sea
se
d
i
a
g
n
o
s
i
s,”
Fr
o
n
t
i
e
rs
i
n
Pl
a
n
t
S
c
i
e
n
c
e
,
v
o
l
.
1
0
,
M
a
r
.
2
0
1
9
,
d
o
i
:
1
0
.
3
3
8
9
/
f
p
l
s
.
2
0
1
9
.
0
0
2
7
2
.
[
6
]
Y
.
L
i
a
n
d
J
.
Y
a
n
g
,
“
F
e
w
-
sh
o
t
c
o
t
t
o
n
p
e
s
t
r
e
c
o
g
n
i
t
i
o
n
a
n
d
t
e
r
mi
n
a
l
r
e
a
l
i
z
a
t
i
o
n
,
”
C
o
m
p
u
t
e
rs
a
n
d
El
e
c
t
ro
n
i
c
s
i
n
Ag
r
i
c
u
l
t
u
re
,
v
o
l
.
1
6
9
,
p
.
1
0
5
2
4
0
,
F
e
b
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
a
g
.
2
0
2
0
.
1
0
5
2
4
0
.
[
7
]
R
.
G
a
j
j
a
r
,
N
.
G
a
j
j
a
r
,
V
.
J
.
Th
a
k
o
r
,
N
.
P
.
P
a
t
e
l
,
a
n
d
S
.
R
u
p
a
r
e
l
i
a
,
“
R
e
a
l
-
t
i
m
e
d
e
t
e
c
t
i
o
n
a
n
d
i
d
e
n
t
i
f
i
c
a
t
i
o
n
o
f
p
l
a
n
t
l
e
a
f
d
i
s
e
a
s
e
s
u
si
n
g
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
o
n
a
n
e
m
b
e
d
d
e
d
p
l
a
t
f
o
r
m,
”
T
h
e
V
i
s
u
a
l
C
o
m
p
u
t
e
r
,
v
o
l
.
3
8
,
n
o
.
8
,
p
p
.
2
9
2
3
–
2
9
3
8
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
3
7
1
-
021
-
0
2
1
6
4
-
9.
[
8
]
S
.
P
.
M
o
h
a
n
t
y
,
D
.
P
.
H
u
g
h
e
s,
a
n
d
M
.
S
a
l
a
t
h
é
,
“
U
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
i
m
a
g
e
-
b
a
s
e
d
p
l
a
n
t
d
i
se
a
se
d
e
t
e
c
t
i
o
n
,
”
F
ro
n
t
i
e
rs
i
n
Pl
a
n
t
S
c
i
e
n
c
e
,
v
o
l
.
7
,
S
e
p
.
2
0
1
6
,
d
o
i
:
1
0
.
3
3
8
9
/
f
p
l
s.
2
0
1
6
.
0
1
4
1
9
.
[
9
]
K
.
P
.
F
e
r
e
n
t
i
n
o
s
,
“
D
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s
f
o
r
p
l
a
n
t
d
i
se
a
se
d
e
t
e
c
t
i
o
n
a
n
d
d
i
a
g
n
o
si
s,
”
C
o
m
p
u
t
e
rs
a
n
d
El
e
c
t
ro
n
i
c
s
i
n
A
g
r
i
c
u
l
t
u
r
e
,
v
o
l
.
1
4
5
,
p
p
.
3
1
1
–
3
1
8
,
F
e
b
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
a
g
.
2
0
1
8
.
0
1
.
0
0
9
.
[
1
0
]
J.
C
h
e
n
,
J
.
C
h
e
n
,
D
.
Z
h
a
n
g
,
Y
.
S
u
n
,
a
n
d
Y
.
A
.
N
a
n
e
h
k
a
r
a
n
,
“
U
s
i
n
g
d
e
e
p
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
f
o
r
i
ma
g
e
-
b
a
sed
p
l
a
n
t
d
i
s
e
a
s
e
i
d
e
n
t
i
f
i
c
a
t
i
o
n
,
”
C
o
m
p
u
t
e
rs
a
n
d
El
e
c
t
r
o
n
i
c
s
i
n
A
g
r
i
c
u
l
t
u
r
e
,
v
o
l
.
1
7
3
,
p
.
1
0
5
3
9
3
,
J
u
n
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
a
g
.
2
0
2
0
.
1
0
5
3
9
3
.
[
1
1
]
G
.
G
.
a
n
d
A
.
P
.
J
.
,
“
I
d
e
n
t
i
f
i
c
a
t
i
o
n
o
f
p
l
a
n
t
l
e
a
f
d
i
s
e
a
s
e
s
u
si
n
g
a
n
i
n
e
-
l
a
y
e
r
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
,
”
C
o
m
p
u
t
e
rs
&
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
r
i
n
g
,
v
o
l
.
7
6
,
p
p
.
3
2
3
–
3
3
8
,
Ju
n
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
e
l
e
c
e
n
g
.
2
0
1
9
.
0
4
.
0
1
1
.
[
1
2
]
K
.
R
.
,
H
.
M
.
,
S
.
A
n
a
n
d
,
P
.
M
a
t
h
i
k
s
h
a
r
a
,
A
.
J
o
h
n
s
o
n
,
a
n
d
M
.
R
.
,
“
A
t
t
e
n
t
i
o
n
e
m
b
e
d
d
e
d
r
e
si
d
u
a
l
C
N
N
f
o
r
d
i
s
e
a
se
d
e
t
e
c
t
i
o
n
i
n
t
o
m
a
t
o
l
e
a
v
e
s,”
A
p
p
l
i
e
d
S
o
f
t
C
o
m
p
u
t
i
n
g
,
v
o
l
.
8
6
,
p
.
1
0
5
9
3
3
,
J
a
n
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
s
o
c
.
2
0
1
9
.
1
0
5
9
3
3
.
[
1
3
]
J.
G
.
A
.
B
a
r
b
e
d
o
,
“
F
a
c
t
o
r
s
i
n
f
l
u
e
n
c
i
n
g
t
h
e
u
s
e
o
f
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
p
l
a
n
t
d
i
sea
s
e
r
e
c
o
g
n
i
t
i
o
n
,
”
Bi
o
sys
t
e
m
s
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
7
2
,
p
p
.
8
4
–
9
1
,
A
u
g
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
i
o
s
y
st
e
mse
n
g
.
2
0
1
8
.
0
5
.
0
1
3
.
[
1
4
]
P
.
A
.
H
.
V
a
r
d
h
i
n
i
,
S
.
A
sr
i
t
h
a
,
a
n
d
Y
.
S
.
D
e
v
i
,
“
Ef
f
i
c
i
e
n
t
d
i
sea
se
d
e
t
e
c
t
i
o
n
o
f
p
a
d
d
y
c
r
o
p
u
si
n
g
C
N
N
,
”
i
n
2
0
2
0
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
S
m
a
r
t
T
e
c
h
n
o
l
o
g
i
e
s
i
n
C
o
m
p
u
t
i
n
g
,
El
e
c
t
r
i
c
a
l
a
n
d
El
e
c
t
r
o
n
i
c
s
(
I
C
S
T
C
EE)
,
O
c
t
.
2
0
2
0
,
p
p
.
1
1
6
–
1
1
9
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
S
TC
EE
4
9
6
3
7
.
2
0
2
0
.
9
2
7
6
7
7
5
.
[
1
5
]
J.
Li
,
Q
.
F
e
n
g
,
J
.
Z
h
a
n
g
,
a
n
d
S
.
Y
a
n
g
,
“
E
M
S
A
M
:
e
n
h
a
n
c
e
d
m
u
l
t
i
-
s
c
a
l
e
s
e
g
m
e
n
t
a
n
y
t
h
i
n
g
mo
d
e
l
f
o
r
l
e
a
f
d
i
se
a
se
se
g
me
n
t
a
t
i
o
n
,
”
Fro
n
t
i
e
rs
i
n
P
l
a
n
t
S
c
i
e
n
c
e
,
v
o
l
.
1
6
,
p
.
1
5
6
4
0
7
9
,
M
a
r
.
2
0
2
5
,
d
o
i
:
1
0
.
3
3
8
9
/
f
p
l
s.2
0
2
5
.
1
5
6
4
0
7
9
.
[
1
6
]
J.
J.
M
o
n
d
a
l
,
M
.
F
.
I
sl
a
m,
S
.
Za
b
e
e
n
,
A
.
B
.
M
.
A
.
A
l
I
sl
a
m
,
a
n
d
J.
N
o
o
r
,
“
N
o
t
e
:
p
l
a
n
t
l
e
a
f
d
i
s
e
a
s
e
n
e
t
w
o
r
k
(
P
Le
a
D
-
N
e
t
)
:
i
d
e
n
t
i
f
y
i
n
g
p
l
a
n
t
l
e
a
f
d
i
s
e
a
s
e
s
t
h
r
o
u
g
h
l
e
v
e
r
a
g
i
n
g
l
i
mi
t
e
d
-
r
e
so
u
r
c
e
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
,
”
i
n
AC
M
S
I
G
C
A
S
/
S
I
G
C
H
I
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
i
n
g
a
n
d
S
u
s
t
a
i
n
a
b
l
e
S
o
c
i
e
t
i
e
s
(
C
O
MP
AS
S
)
,
Ju
n
.
2
0
2
2
,
p
p
.
6
6
8
–
6
7
3
,
d
o
i
:
1
0
.
1
1
4
5
/
3
5
3
0
1
9
0
.
3
5
3
4
8
4
4
.
[
1
7
]
D
.
H
u
g
h
e
s,
M
.
S
a
l
a
t
h
é
,
a
n
d
O
t
h
e
r
s
,
“
A
n
o
p
e
n
a
c
c
e
ss
r
e
p
o
si
t
o
r
y
o
f
i
m
a
g
e
s
o
n
p
l
a
n
t
h
e
a
l
t
h
t
o
e
n
a
b
l
e
t
h
e
d
e
v
e
l
o
p
m
e
n
t
o
f
mo
b
i
l
e
d
i
s
e
a
se
d
i
a
g
n
o
s
t
i
c
s,
”
a
rXi
v
p
re
p
ri
n
t
a
rXi
v
:
1
5
1
1
.
0
8
0
6
0
,
2
0
1
5
.
[
1
8
]
T.
U
.
R
e
h
ma
n
,
M
.
S
.
M
a
h
mu
d
,
Y
.
K
.
C
h
a
n
g
,
J.
J
i
n
,
a
n
d
J.
S
h
i
n
,
“
C
u
r
r
e
n
t
a
n
d
f
u
t
u
r
e
a
p
p
l
i
c
a
t
i
o
n
s
o
f
s
t
a
t
i
s
t
i
c
a
l
mac
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms
f
o
r
a
g
r
i
c
u
l
t
u
r
a
l
ma
c
h
i
n
e
v
i
si
o
n
s
y
st
e
ms,”
C
o
m
p
u
t
e
rs
a
n
d
El
e
c
t
r
o
n
i
c
s i
n
Ag
r
i
c
u
l
t
u
re
,
v
o
l
.
1
5
6
,
p
p
.
5
8
5
–
6
0
5
,
Ja
n
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
a
g
.
2
0
1
8
.
1
2
.
0
0
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.