I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Ro
bo
t
ics a
nd
Aut
o
m
a
t
io
n
(
I
J
RA)
Vo
l.
1
4
,
No
.
1
,
Ma
r
ch
20
2
5
,
p
p
.
38
~
46
I
SS
N:
2722
-
2
5
8
6
,
DOI
:
1
0
.
1
1
5
9
1
/i
jr
a
.
v
1
4
i
1
.
pp
38
-
46
38
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
r
a
.
ia
esco
r
e.
co
m
Vo
tTo
mNet:
Voti
ng
-
ba
sed tom
a
to
disea
se dia
g
no
sis
wit
h
trans
fer
l
ea
rning
Sh
ra
dh
a
J
o
s
hi
-
B
a
g
1,
2
,
Wa
ni V.
P
a
t
il
1
,
Sh
rik
a
nt
Cha
v
a
t
e
1
1
El
e
c
t
r
o
n
i
c
s
a
n
d
Te
l
e
c
o
m
mu
n
i
c
a
t
i
o
n
En
g
i
n
e
e
r
i
n
g
D
e
p
a
r
t
m
e
n
t
,
G
.
H
.
R
a
i
s
o
n
i
U
n
i
v
e
r
si
t
y
,
A
mr
a
v
a
t
i
,
M
a
h
a
r
a
s
h
t
r
a
,
I
n
d
i
a
2
El
e
c
t
r
o
n
i
c
s
a
n
d
Te
l
e
c
o
m
mu
n
i
c
a
t
i
o
n
En
g
i
n
e
e
r
i
n
g
D
e
p
a
r
t
m
e
n
t
,
N
K
O
r
c
h
i
d
C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
a
n
d
Te
c
h
n
o
l
o
g
y
,
S
o
l
a
p
u
r
,
M
a
h
a
r
a
s
h
t
r
a
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l 2
8
,
2
0
2
4
R
ev
is
ed
Sep
1
8
,
2
0
2
4
Acc
ep
ted
Oct
6
,
2
0
2
4
Th
e
re
se
a
rc
h
p
re
se
n
ts
a
n
a
d
v
a
n
c
e
d
a
u
to
m
a
ti
o
n
sy
ste
m
,
term
e
d
V
o
tT
o
m
Ne
t,
d
e
sig
n
e
d
f
o
r
d
iag
n
o
sin
g
to
m
a
to
lea
f
d
ise
a
se
s u
sin
g
tran
sfe
r
lea
rn
in
g
,
a
n
d
so
ft
a
n
d
h
a
rd
v
o
ti
n
g
e
n
se
m
b
le
tec
h
n
iq
u
e
s
.
By
lev
e
ra
g
i
n
g
si
x
p
re
-
trai
n
e
d
d
e
e
p
lea
rn
in
g
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
two
rk
s
—
VG
G
1
6
,
In
c
e
p
ti
o
n
Ne
t,
Re
sN
e
t,
M
o
b
i
leN
e
t,
Eff
icie
n
tNe
t,
a
n
d
De
n
se
Ne
t
—
th
e
sy
ste
m
a
c
h
iev
e
d
a
n
i
m
p
re
ss
iv
e
a
c
c
u
ra
c
y
o
f
9
9
.
2
%
.
Th
e
se
m
o
d
e
l
s
we
re
m
e
ti
c
u
lo
u
sly
fin
e
-
t
u
n
e
d
t
o
d
ia
g
n
o
se
m
u
lt
ip
le
ty
p
e
s
o
f
to
m
a
to
d
ise
a
se
s
with
h
e
ig
h
ten
e
d
p
re
c
isio
n
.
Th
e
i
n
teg
ra
ti
o
n
o
f
a
so
ft
a
n
d
h
a
rd
v
o
ti
n
g
m
e
c
h
a
n
ism
fu
rth
e
r
e
n
h
a
n
c
e
d
th
e
o
v
e
ra
ll
d
iag
n
o
stic
a
c
c
u
ra
c
y
b
y
c
o
m
b
in
i
n
g
t
h
e
stre
n
g
th
s
o
f
th
e
se
d
i
v
e
rse
m
o
d
e
ls
in
to
a
p
o
we
rfu
l
e
n
se
m
b
le.
Th
e
fin
d
i
n
g
s
u
n
d
e
rsc
o
re
th
e
ro
b
u
st
n
e
ss
,
re
li
a
b
il
it
y
,
a
n
d
e
ffe
c
ti
v
e
n
e
ss
o
f
th
is
e
n
se
m
b
le
tec
h
n
i
q
u
e
,
m
a
rk
i
n
g
a
sig
n
ifi
c
a
n
t
a
d
v
a
n
c
e
m
e
n
t
in
p
re
c
isio
n
a
g
ricu
lt
u
re
a
n
d
c
r
o
p
h
e
a
lt
h
a
ss
e
ss
m
e
n
t.
By
o
u
tp
e
rfo
rm
in
g
trad
it
io
n
a
l
m
e
th
o
d
s,
th
is
a
p
p
ro
a
c
h
o
ffe
rs
a
m
o
re
p
ra
c
ti
c
a
l
a
n
d
e
fficie
n
t
so
lu
ti
o
n
f
o
r
lar
g
e
-
sc
a
le
a
g
ricu
lt
u
ra
l
a
p
p
li
c
a
ti
o
n
s,
e
n
a
b
li
n
g
c
o
m
p
re
h
e
n
siv
e
c
ro
p
m
a
n
a
g
e
m
e
n
t
a
n
d
imp
r
o
v
e
d
y
ield
.
In
c
o
n
c
l
u
sio
n
,
t
h
is
re
se
a
rc
h
lay
s
a
stro
n
g
f
o
u
n
d
a
ti
o
n
fo
r
f
u
tu
re
in
n
o
v
a
ti
o
n
s
in
a
u
t
o
m
a
ted
p
lan
t
d
ise
a
se
d
iag
n
o
sis
a
n
d
a
g
ricu
lt
u
ra
l
tec
h
n
o
lo
g
y
.
I
ts
c
o
n
tri
b
u
ti
o
n
s
h
a
v
e
th
e
p
o
ten
ti
a
l
to
re
v
o
lu
ti
o
n
ize
d
ise
a
se
m
a
n
a
g
e
m
e
n
t,
re
d
u
c
e
c
ro
p
lo
ss
e
s,
a
n
d
u
lt
ima
tely
e
n
h
a
n
c
e
fo
o
d
se
c
u
rit
y
o
n
a
g
l
o
b
a
l
sc
a
le.
K
ey
w
o
r
d
s
:
E
n
s
em
b
le
lear
n
in
g
Har
d
v
o
tin
g
So
f
t v
o
tin
g
T
o
m
ato
leaf
d
is
ea
s
e
T
r
an
s
f
er
lear
n
i
n
g
Vo
tTo
m
Net
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
:
Sh
r
ad
h
a
J
o
s
h
i
-
B
ag
E
lectr
o
n
ics an
d
T
elec
o
m
m
u
n
i
ca
tio
n
E
n
g
in
ee
r
in
g
De
p
ar
tm
en
t,
G.
H.
R
aiso
n
i U
n
iv
er
s
ity
Am
r
av
ati,
Ma
h
ar
ash
tr
a,
I
n
d
ia
E
m
ail:
b
ag
s
h
r
ad
h
a@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
u
s
e
o
f
ar
tific
ial
in
tellig
en
ce
in
ag
r
icu
ltu
r
e
h
as seen
s
ig
n
if
ican
t g
r
o
wth
,
with
co
n
v
o
l
u
tio
n
al
n
eu
r
a
l
n
etwo
r
k
s
(
C
NNs)
b
ec
o
m
in
g
e
s
s
en
tial
in
p
lan
t
p
ath
o
l
o
g
y
.
T
h
is
r
esear
ch
p
ap
er
f
o
cu
s
es
o
n
d
etec
tin
g
cr
o
p
lea
f
d
is
ea
s
es.
T
h
e
p
r
o
p
o
s
ed
s
tr
ateg
y
co
m
b
i
n
es
tr
an
s
f
er
lear
n
i
n
g
an
d
e
n
s
em
b
le
lear
n
in
g
t
o
o
p
tim
ize
d
is
ea
s
e
class
if
icatio
n
,
a
cr
itical
asp
ec
t
o
f
ea
r
ly
d
is
ea
s
e
d
etec
tio
n
a
n
d
cr
o
p
m
a
n
ag
em
e
n
t.
T
r
a
n
s
f
er
lear
n
in
g
lev
er
a
g
es
p
r
e
-
tr
ain
ed
m
o
d
els
—
s
u
ch
as
VGG1
6
,
I
n
ce
p
tio
n
Net,
R
esNet,
Mo
b
ileNet,
E
f
f
icien
tNet,
an
d
Den
s
eNe
t
—
th
at
h
av
e
b
ee
n
tr
ai
n
ed
o
n
lar
g
e,
d
i
v
er
s
e
d
atasets
.
T
h
ese
m
o
d
els,
ca
p
ab
le
o
f
lear
n
i
n
g
co
m
p
lex
f
ea
tu
r
es,
ca
n
th
en
b
e
ad
ap
ted
to
wo
r
k
o
n
s
m
aller
,
s
p
ec
if
ic
d
atasets
.
I
n
o
u
r
r
esear
c
h
,
we
u
tili
ze
th
ese
p
r
e
-
tr
ain
e
d
m
o
d
els
to
en
h
a
n
ce
C
NN
p
er
f
o
r
m
an
ce
in
d
is
tin
g
u
is
h
in
g
b
etwe
en
h
ea
lth
y
an
d
d
is
ea
s
ed
to
m
ato
leav
es.
T
h
e
r
ich
f
ea
tu
r
es
lear
n
e
d
b
y
th
ese
m
o
d
els
aim
to
d
eli
v
er
h
ig
h
ac
cu
r
ac
y
an
d
r
eliab
le
d
is
ea
s
e
clas
s
if
icatio
n
.
I
n
a
d
d
itio
n
to
tr
a
n
s
f
er
lear
n
in
g
,
we
in
co
r
p
o
r
ate
e
n
s
em
b
le
lear
n
in
g
,
p
ar
ticu
lar
l
y
s
o
f
t
an
d
h
ar
d
v
o
tin
g
tech
n
iq
u
es,
to
im
p
r
o
v
e
class
if
icatio
n
ac
cu
r
ac
y
.
E
n
s
em
b
le
lear
n
in
g
in
v
o
lv
es
g
en
er
atin
g
v
ar
io
u
s
C
NN
ar
ch
itectu
r
es
an
d
in
teg
r
atin
g
th
eir
o
u
tp
u
ts
to
ac
h
iev
e
b
etter
d
iag
n
o
s
tic
r
esu
lts
.
T
h
e
m
ajo
r
ity
v
o
tin
g
a
p
p
r
o
ac
h
at
th
e
c
o
r
e
o
f
th
e
p
r
o
p
o
s
ed
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
V
o
tTo
mN
et:
V
o
tin
g
-
b
a
s
ed
t
o
ma
to
d
is
ea
s
e
d
ia
g
n
o
s
is
w
ith
tr
a
n
s
fer lea
r
n
in
g
(
S
h
r
a
d
h
a
J
o
s
h
i
-
B
a
g
)
39
en
s
em
b
le
m
eth
o
d
b
en
ef
its
f
r
o
m
th
e
s
tr
en
g
th
s
o
f
in
d
iv
id
u
al
m
o
d
els
wh
ile
m
in
im
izin
g
th
eir
b
iases
.
T
h
is
tech
n
iq
u
e
s
ig
n
if
ican
tly
en
h
a
n
ce
s
d
iag
n
o
s
tic
ac
cu
r
ac
y
b
y
h
a
r
n
ess
in
g
th
e
d
iv
er
s
e
s
tr
en
g
th
s
an
d
wea
k
n
ess
es
o
f
m
u
ltip
le
C
NN
s
tr
u
ctu
r
es,
im
p
r
o
v
in
g
th
e
m
o
d
el
’
s
ab
ilit
y
t
o
d
etec
t
to
m
ato
leaf
d
is
ea
s
es.
T
h
e
f
in
d
in
g
s
o
f
t
h
is
s
tu
d
y
d
em
o
n
s
tr
ate
th
at
en
s
em
b
le
lear
n
in
g
o
f
f
er
s
a
s
u
b
s
tan
tial
im
p
r
o
v
em
e
n
t
in
ag
r
icu
ltu
r
a
l
d
iag
n
o
s
tics
,
with
n
o
tab
ly
h
i
g
h
er
d
is
ea
s
e
d
etec
tio
n
r
ates
co
m
p
ar
e
d
to
u
s
in
g
in
d
iv
id
u
al
C
NN
tech
n
iq
u
e
s
.
I
t
h
ig
h
lig
h
ts
th
e
p
o
ten
tial
o
f
AI
s
o
lu
tio
n
s
to
en
h
an
ce
cr
o
p
m
an
ag
e
m
en
t
p
r
ac
t
ices,
r
ein
f
o
r
cin
g
th
e
s
tab
ilit
y
a
n
d
s
u
s
tain
ab
ilit
y
o
f
th
e
g
lo
b
al
f
o
o
d
s
u
p
p
ly
s
y
s
tem
am
id
ag
r
icu
ltu
r
e
’
s
in
h
er
e
n
t
u
n
ce
r
tain
ties
.
I
n
co
n
clu
s
io
n
,
th
e
p
r
o
p
o
s
ed
s
tr
ateg
y
p
r
o
v
id
es
a
p
r
o
ac
tiv
e
f
r
am
ew
o
r
k
f
o
r
ea
r
l
y
d
is
ea
s
e
d
etec
ti
o
n
an
d
ef
f
ec
tiv
e
cr
o
p
p
r
o
tectio
n
,
ess
en
tial
f
o
r
s
u
s
tain
ab
le
ag
r
icu
ltu
r
e
an
d
m
a
in
tain
in
g
a
h
ea
lth
y
g
lo
b
al
f
o
o
d
ch
ain
.
Ap
p
licatio
n
s
o
f
t
h
e
co
n
v
o
l
u
tio
n
al
n
e
u
r
al
n
etwo
r
k
ar
e
m
an
y
an
d
t
h
ey
i
n
clu
d
e
MRI
im
ag
e
class
if
icatio
n
[
1
]
,
v
id
eo
s
h
o
t
b
o
u
n
d
ar
y
d
etec
tio
n
[
2
]
,
an
d
o
b
ject
d
etec
tio
n
[
3
]
.
On
its
p
ar
t,
Hase
et.
a
l.
[
4
]
an
d
Alg
an
i
et
a
l.
[
5
]
ex
am
i
n
e
d
ee
p
lear
n
in
g
tec
h
n
o
lo
g
i
es
f
o
r
p
lan
t
d
is
ea
s
e
id
en
tific
atio
n
an
d
d
is
cu
s
s
th
e
m
o
d
e
r
n
t
r
en
d
s
a
n
d
u
s
es
o
f
d
iag
n
o
s
is
in
th
e
a
g
r
icu
ltu
r
al
f
ield
.
A
co
m
b
in
ed
in
n
o
v
ativ
e
m
eth
o
d
c
o
n
ce
r
n
in
g
d
ee
p
lear
n
in
g
f
o
r
th
e
id
en
tific
atio
n
o
f
to
m
ato
leaf
d
is
ea
s
es
a
n
d
th
eir
class
if
icatio
n
is
d
e
s
ig
n
ed
an
d
estab
lis
h
ed
b
y
T
r
iv
ed
i
et
a
l.
[
6
]
u
s
in
g
m
o
r
e
th
a
n
o
n
e
n
eu
r
al
n
etwo
r
k
ap
p
r
o
ac
h
.
T
h
u
s
,
t
h
e
T
o
L
eD
m
o
d
el
is
in
t
r
o
d
u
ce
d
th
at
u
tili
ze
s
C
NN
to
d
etec
t
to
m
ato
leaf
d
is
ea
s
es
an
d
in
d
icate
s
an
ef
f
ec
tiv
e
u
s
e
o
f
d
ee
p
lear
n
in
g
to
im
p
r
o
v
e
ag
r
icu
ltu
r
al
d
is
ea
s
e
co
n
tr
o
l
Ag
ar
wal
et
a
l.
[
7
]
.
T
h
is
p
ap
er
af
f
ir
m
s
th
at
ap
p
lied
d
ee
p
lear
n
in
g
b
y
Am
ar
a
et
a
l
.
[
8
]
,
wh
er
e
b
an
a
n
a
leaf
d
is
ea
s
es
an
d
o
th
e
r
p
la
n
t
d
is
ea
s
es
ca
n
ea
s
ily
b
e
d
iag
n
o
s
ed
in
r
ea
l
f
ar
m
i
n
g
ar
ea
s
an
d
th
is
h
as
b
ee
n
m
a
d
e
p
o
s
s
ib
le
b
y
th
e
ap
p
licatio
n
o
f
n
eu
r
al
n
etwo
r
k
s
wh
ic
h
is
v
er
y
u
s
ef
u
l
in
in
cr
ea
s
in
g
th
e
y
ield
o
f
b
a
n
an
a
p
r
o
d
u
ctio
n
.
B
ar
b
e
d
o
[
9
]
d
is
cu
s
s
ed
d
if
f
er
en
t
f
ac
to
r
s
af
f
ec
t
in
g
th
e
u
tili
za
tio
n
o
f
d
ee
p
lear
n
in
g
f
o
r
cr
o
p
d
is
ea
s
e
d
iag
n
o
s
is
f
ac
ilit
ies
n
am
ely
d
ata
q
u
ality
,
m
o
d
el
s
tr
u
ctu
r
es,
an
d
en
v
ir
o
n
m
en
tal
p
ar
a
m
eter
s
.
A
s
y
s
tem
i
s
im
p
lem
en
ted
u
s
in
g
d
ee
p
lear
n
in
g
f
o
r
th
e
d
et
ec
tio
n
o
f
to
m
ato
leaf
d
is
ea
s
es
an
d
th
e
id
en
tific
atio
n
o
f
th
e
s
y
m
p
to
m
s
th
at
m
ay
b
e
p
r
esen
t,
wh
ic
h
estab
lis
h
es
th
e
ap
p
r
o
ac
h
’
s
e
f
f
i
cien
cy
in
r
e
v
ea
lin
g
co
m
p
r
eh
e
n
s
iv
e
ch
a
r
ac
ter
is
tics
o
f
d
is
ea
s
es
[
1
0
]
.
T
h
is
is
p
r
o
v
ed
ef
f
ec
tiv
ely
b
y
C
h
e
n
et
a
l.
[
1
1
]
,
wh
o
ex
p
er
im
en
ted
with
th
e
C
NNs
an
d
tr
an
s
f
er
lear
n
in
g
f
o
r
cr
o
p
d
is
ea
s
e
d
etec
tio
n
wh
er
e
th
e
y
s
ee
th
at
tr
an
s
f
er
lear
n
in
g
h
as
a
m
ajo
r
im
p
r
o
v
e
m
en
t
o
n
th
e
m
o
d
el
p
er
f
o
r
m
an
ce
ev
en
w
h
er
e
tr
ain
i
n
g
d
ata
is
v
er
y
lim
ited
.
Oth
e
r
r
esear
ch
er
s
[
1
2
]
h
av
e
also
p
r
o
p
o
s
ed
th
e
in
teg
r
atio
n
o
f
th
e
b
ac
ter
ial
s
ca
v
en
g
in
g
tech
n
i
q
u
e
in
a
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
f
o
r
en
h
a
n
cin
g
th
e
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
o
n
p
lan
t
leaf
d
is
ea
s
e
id
en
tific
atio
n
.
Dee
p
n
eu
r
al
n
etwo
r
k
-
b
ased
m
o
d
els
f
o
r
d
et
ec
tin
g
d
is
ea
s
es
in
m
illet
cr
o
p
s
wer
e
s
tu
d
ied
b
y
s
ev
er
al
r
esear
ch
er
s
,
wh
o
u
s
ed
tr
an
s
f
er
lear
n
i
n
g
to
e
n
h
an
ce
th
e
ac
cu
r
ac
y
an
d
e
f
f
icien
cy
o
f
t
h
e
d
iag
n
o
s
es
[
1
3
]
.
R
esear
ch
s
tu
d
ies,
f
o
r
in
s
tan
ce
,
o
n
DC
NN
f
o
r
th
e
p
r
o
g
n
o
s
is
o
f
cr
o
p
leaf
ailm
en
ts
ar
e
cr
itiq
u
ed
to
estab
lis
h
an
d
d
em
o
n
s
tr
ate
th
e
ad
v
an
tag
es
an
d
p
itfa
lls
o
f
d
if
f
er
in
g
C
NN
co
n
f
ig
u
r
atio
n
s
an
d
th
eir
u
s
ag
e
in
p
lan
t p
at
h
o
lo
g
y
[
1
4
]
.
An
aly
zin
g
d
if
f
e
r
en
t
m
o
d
els
o
f
d
ee
p
lear
n
in
g
f
o
r
cr
o
p
d
is
ea
s
e
d
etec
tio
n
,
it
is
n
o
ted
th
at
th
ey
d
em
o
n
s
tr
ate
h
ig
h
ac
cu
r
ac
y
an
d
ca
n
s
ig
n
if
ican
tly
tr
an
s
f
o
r
m
th
e
ag
r
ic
u
ltu
r
e
i
n
d
u
s
tr
y
th
r
o
u
g
h
th
e
in
tr
o
d
u
ctio
n
o
f
n
ew
r
eliab
le
m
eth
o
d
s
o
f
d
is
ea
s
e
id
en
tific
atio
n
[
1
5
]
.
T
h
e
r
esear
ch
er
s
illu
s
tr
ated
th
e
a
p
p
licatio
n
o
f
d
ee
p
lear
n
in
g
in
to
m
ato
cr
o
p
d
is
ea
s
es
an
d
p
est
’
d
etec
tio
n
,
a
n
d
d
e
m
o
n
s
tr
ated
th
e
p
r
ac
tical
f
ea
s
i
b
ilit
y
o
f
em
p
lo
y
in
g
th
e
tech
n
iq
u
es
f
o
r
m
o
n
it
o
r
in
g
an
d
m
an
a
g
in
g
c
r
o
p
s
u
r
v
ei
llan
ce
[
1
6
]
,
[
1
7
]
.
An
n
o
tated
im
ag
e
d
ia
g
n
o
s
tic
m
eth
o
d
s
o
f
p
lan
t
h
ea
lth
d
is
o
r
d
er
s
wer
e
cr
ea
ted
an
d
in
t
r
o
d
u
ce
d
in
clu
d
in
g
a
n
u
m
b
er
o
f
m
ed
ia
p
r
o
ce
s
s
in
g
alg
o
r
ith
m
s
u
s
in
g
a
r
a
n
g
e
o
f
AI
m
eth
o
d
o
lo
g
ies
to
in
cr
ea
s
e
th
e
r
eliab
ilit
y
o
f
th
e
d
iag
n
o
s
is
[
1
8
]
–
[
2
0
]
.
So
m
e
r
ec
en
t
s
tu
d
ies
co
m
p
ar
i
n
g
d
if
f
er
en
t
alg
o
r
ith
m
ic
p
r
o
ce
d
u
r
es
o
f
n
eu
r
al
n
etwo
r
k
s
f
o
r
p
lan
t
leaf
d
is
ea
s
e
class
if
icatio
n
h
av
e
p
o
in
ted
o
u
t
th
e
ad
v
an
tag
es
a
n
d
t
h
e
d
is
ad
v
an
tag
es
o
f
ea
ch
o
f
t
h
ese
m
eth
o
d
s
[
2
1
]
,
[
2
2
]
.
J
i
et
a
l.
[
2
3
]
p
r
o
v
i
d
e
an
o
u
tlin
e
o
f
th
e
m
o
d
el,
m
u
ltip
le
C
NNs
in
th
e
id
en
tific
atio
n
o
f
g
r
ap
e
lea
f
d
is
ea
s
es,
an
d
th
e
s
ig
n
if
ican
ce
o
f
u
s
in
g
m
u
ltip
le
n
eu
r
al
n
etwo
r
k
s
f
o
r
b
etter
r
esu
lts
.
A
co
m
p
r
e
h
en
s
iv
e
s
u
r
v
ey
o
f
t
h
e
ap
p
licatio
n
s
o
f
DL
in
f
ar
m
in
g
[
2
4
]
h
ig
h
lig
h
ts
its
ef
f
ec
tiv
en
ess
in
ar
ea
s
s
u
ch
as
cr
o
p
an
d
s
o
il
m
an
ag
em
en
t,
d
is
ea
s
e
d
etec
tio
n
,
an
d
p
r
ec
is
io
n
f
ar
m
in
g
.
A
d
ee
p
co
n
v
o
lu
tio
n
a
l
n
eu
r
al
n
etwo
r
k
with
an
atten
tio
n
m
ec
h
an
is
m
is
u
s
ed
b
y
W
an
g
et
a
l
.
[
2
5
]
f
o
r
th
e
id
en
tifi
ca
tio
n
o
f
ap
p
le
leaf
d
is
ea
s
es
wh
ich
h
as
g
i
v
en
s
atis
f
ac
to
r
y
ac
c
u
r
ac
y
.
Ma
n
y
r
ese
ar
ch
wo
r
k
s
[
2
6
]
–
[
2
8
]
a
r
e
u
tili
zin
g
d
ee
p
C
NN
f
o
r
r
ice
d
is
ea
s
e
id
en
tific
atio
n
,
d
e
m
o
n
s
tr
atin
g
th
e
p
o
te
n
tial
o
f
C
NNs
in
ac
cu
r
ately
d
ia
g
n
o
s
in
g
p
lan
t
d
is
ea
s
es.
T
h
e
o
th
er
r
esear
ch
wo
r
k
s
[
2
9
]
,
[
3
0
]
ex
p
lo
r
ed
r
ea
l
-
tim
e
p
la
n
t
d
is
ea
s
e
r
ec
o
g
n
itio
n
u
s
in
g
tr
an
s
f
er
lear
n
in
g
,
s
h
o
wca
s
in
g
th
e
p
r
ac
tical
ap
p
l
icatio
n
o
f
AI
in
r
ea
l
-
tim
e
a
g
r
i
cu
ltu
r
al
m
o
n
ito
r
in
g
.
Ma
ch
i
n
e
l
ea
r
n
in
g
is
u
s
ed
to
m
ea
s
u
r
e
th
e
ca
s
es
o
f
cr
o
p
d
is
ea
s
e
an
d
th
e
p
er
ce
n
tag
e
o
f
in
f
ec
tio
n
f
r
o
m
th
e
im
ag
es
o
f
leav
es,
o
f
f
er
i
n
g
v
alu
ab
le
in
s
ig
h
ts
in
to
au
to
m
ated
p
lan
t
h
ea
lth
ass
ess
m
en
t
[
3
1
]
,
[
3
2
]
.
T
h
e
s
y
s
tem
u
s
es
th
e
d
i
s
ea
s
ed
an
d
h
ea
lth
y
im
ag
es
f
o
r
tr
ain
in
g
a
n
d
C
NN
f
etch
es
th
e
v
a
r
io
u
s
f
ea
tu
r
es
d
u
r
in
g
tr
ain
in
g
a
n
d
lear
n
s
.
T
h
e
lear
n
ed
alg
o
r
ith
m
ac
h
iev
es
v
er
y
h
ig
h
ac
c
u
r
ac
y
.
Fro
m
th
e
ab
o
v
e
s
tu
d
y
o
f
th
e
l
iter
atu
r
e,
th
e
f
o
llo
win
g
s
cien
ti
f
ic
q
u
esti
o
n
s
ar
is
e.
C
an
tr
an
s
f
er
lear
n
in
g
b
e
u
s
ed
f
o
r
id
en
tify
i
n
g
leaf
d
is
ea
s
es?
Usi
n
g
th
e
m
ajo
r
ity
v
o
tin
g
tec
h
n
iq
u
e,
is
it
p
o
s
s
ib
le
to
im
p
r
o
v
e
th
e
ac
cu
r
ac
y
o
f
th
e
class
if
icatio
n
alg
o
r
ith
m
?
C
an
we
im
p
r
o
v
e
th
e
y
ield
o
f
c
r
o
p
s
b
y
ea
r
l
y
d
etec
tio
n
o
f
leaf
d
is
ea
s
es u
s
in
g
tr
an
s
f
er
lear
n
in
g
an
d
s
o
f
t v
o
tin
g
?
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
4
,
No
.
1
,
Ma
r
ch
2
0
2
5
:
38
-
46
40
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
T
h
is
s
ec
tio
n
co
n
tain
s
a
d
etailed
in
tr
o
d
u
ctio
n
to
t
h
e
d
ataset,
th
e
p
r
o
p
o
s
ed
m
o
d
el,
an
d
th
e
a
p
p
licatio
n
o
f
en
s
em
b
le
u
s
in
g
s
o
f
t a
n
d
h
ar
d
v
o
tin
g
.
2
.
1
.
Da
t
a
s
et
T
h
e
p
lan
t
v
illag
e
d
ataset
is
a
co
m
p
r
eh
e
n
s
iv
e
co
llectio
n
o
f
im
ag
es
d
esig
n
ed
f
o
r
th
e
id
e
n
tific
atio
n
an
d
class
if
icatio
n
o
f
cr
o
p
d
is
ea
s
es.
W
h
ile
th
is
s
tu
d
y
f
o
c
u
s
es
s
p
ec
if
ically
o
n
to
m
ato
im
a
g
es,
th
e
d
ataset
s
p
an
s
a
v
ar
iety
o
f
cr
o
p
s
,
in
clu
d
in
g
p
o
tato
es,
g
r
ap
es,
ap
p
les,
co
r
n
,
b
lu
eb
er
r
ies,
r
asp
b
er
r
ies,
s
o
y
b
ea
n
s
,
s
q
u
ash
,
an
d
s
tr
awb
er
r
ies.
I
t
in
clu
d
es
class
if
icatio
n
s
f
o
r
b
o
th
d
is
ea
s
ed
an
d
h
ea
lth
y
p
lan
ts
.
Fo
r
to
m
ato
es
,
th
e
d
ataset
co
v
er
s
s
ev
er
al
d
is
ea
s
es
s
u
ch
as
b
ac
te
r
ial
s
p
o
ts
,
m
o
s
aic
v
ir
u
s
es,
s
p
id
er
m
ites
,
ea
r
ly
b
lig
h
t,
late
b
lig
h
t,
leaf
m
o
l
d
,
an
d
s
ep
to
r
ia
leaf
s
p
o
ts
.
E
ac
h
d
is
e
ase
ca
teg
o
r
y
c
o
n
tain
s
ap
p
r
o
x
im
ately
1
,
5
0
0
im
a
g
es,
m
an
y
o
f
wh
ich
ar
e
u
tili
ze
d
f
o
r
ex
p
er
im
en
tatio
n
an
d
a
n
aly
s
is
in
th
is
s
tu
d
y
.
2
.
2
.
Vo
t
T
o
m
Net
:
T
he
p
ro
po
s
ed
m
o
del
T
h
e
d
esig
n
o
f
th
e
Vo
tTo
m
Net
s
y
s
tem
,
wh
ich
co
m
b
in
es
en
s
e
m
b
le
s
o
f
t
an
d
h
ar
d
v
o
tin
g
with
tr
an
s
f
er
lear
n
in
g
to
en
h
an
ce
cr
o
p
leaf
d
is
ea
s
e
d
etec
tio
n
,
esp
ec
ially
ab
o
u
t
to
m
ato
leav
es,
is
illu
s
tr
ated
in
Fig
u
r
e
1
.
T
h
e
im
ag
e
d
ataset
—
s
p
ec
if
ically
,
t
h
e
to
m
at
o
leaf
d
ataset
—
co
m
e
s
f
r
o
m
Plan
t
Villag
e.
T
o
g
et
r
ea
d
y
f
o
r
m
ac
h
in
e
lear
n
in
g
m
o
d
el
t
r
ain
in
g
,
r
aw
i
m
ag
e
d
ata
m
u
s
t
g
o
th
r
o
u
g
h
n
e
ce
s
s
ar
y
ch
an
g
es
in
clu
d
in
g
s
ca
lin
g
,
n
o
r
m
aliza
tio
n
,
an
d
n
o
is
e
r
ed
u
ctio
n
d
u
r
in
g
th
e
im
ag
e
p
r
ep
r
o
ce
s
s
in
g
s
tag
e.
Fu
r
th
er
m
o
r
e
,
im
ag
e
au
g
m
e
n
ta
tio
n
tech
n
iq
u
es
ar
e
u
s
ed
to
im
p
r
o
v
e
m
o
d
el
r
o
b
u
s
tn
ess
b
y
in
cr
ea
s
in
g
d
ataset
v
ar
iety
th
r
o
u
g
h
z
o
o
m
s
,
f
lip
s
,
tr
an
s
latio
n
s
,
an
d
r
o
tatio
n
s
.
Af
ter
th
at,
th
e
d
at
aset
is
d
iv
id
ed
in
to
tr
ain
in
g
an
d
v
alid
atio
n
s
ets,
as
wel
l
as
a
tes
t
s
et
f
o
r
ass
es
s
m
en
t.
Pre
-
tr
ain
ed
C
NN
m
o
d
els,
s
u
ch
as
VGG1
6
,
I
n
ce
p
tio
n
Net,
R
esNet,
Mo
b
ileNet,
E
f
f
icien
tNet,
an
d
Den
s
eNe
t,
ar
e
u
s
ed
f
o
r
th
e
im
ag
e
class
if
icatio
n
task
d
u
r
in
g
t
h
e
m
o
d
el
lear
n
in
g
p
h
ase.
T
h
e
to
m
ato
leaf
d
is
ea
s
e
d
ataset
is
u
s
ed
to
r
ef
in
e
th
ese
m
o
d
els
o
n
ce
t
h
ey
h
a
v
e
b
ee
n
p
r
e
-
tr
ain
ed
o
n
s
izab
le
d
atasets
.
E
v
er
y
m
o
d
el
g
ai
n
s
p
r
o
f
icien
c
y
in
class
if
y
in
g
d
iv
er
s
e
cr
o
p
leaf
d
is
ea
s
es,
en
h
a
n
cin
g
its
ca
p
ac
ity
t
o
r
ec
o
g
n
i
ze
an
d
d
if
f
er
en
tiate
b
etwe
en
d
iv
er
s
e
co
n
d
itio
n
s
im
p
ac
tin
g
to
m
ato
p
lan
ts
.
W
eig
h
ts
th
at
h
av
e
b
ee
n
p
r
e
-
tr
ain
ed
o
n
th
e
I
m
ag
eNe
t
d
ataset
ar
e
in
itialized
f
o
r
ea
ch
m
o
d
el.
T
h
r
ee
R
GB
co
lo
r
ch
an
n
els
an
d
2
2
4
x
2
2
4
p
ix
els
ar
e
ty
p
ical
f
o
r
in
p
u
t
im
ag
e
s
izes.
Fig
u
r
e
1
.
T
h
e
s
u
g
g
ested
m
o
d
e
l
’
s
s
y
s
tem
ar
ch
itectu
r
e
A
lay
er
o
f
Glo
b
alAv
er
ag
eP
o
o
lin
g
2
D
b
y
tak
i
n
g
th
e
a
v
e
r
ag
e
o
f
ev
er
y
elem
en
t
in
e
ac
h
m
ap
,
Vo
tTo
m
Net
elim
in
ates
th
e
r
eq
u
ir
em
en
t
f
o
r
a
co
m
p
letel
y
co
n
n
ec
ted
lay
e
r
,
h
en
ce
r
e
d
u
cin
g
th
e
s
p
atial
d
im
en
s
io
n
s
o
f
f
ea
tu
r
e
m
a
p
s
.
Un
d
er
s
tan
d
in
g
co
m
p
le
x
p
atter
n
s
is
m
ad
e
p
o
s
s
ib
le
b
y
a
f
u
lly
co
n
n
ec
ted
la
y
er
th
at
co
m
es
n
ex
t,
f
ea
tu
r
in
g
1
,
0
2
4
u
n
its
an
d
R
eL
U
ac
tiv
atio
n
.
T
h
e
f
in
al
o
u
tp
u
t
lay
er
u
s
es
a
So
f
tMa
x
ac
tiv
atio
n
f
u
n
ctio
n
a
n
d
is
d
esig
n
ed
f
o
r
t
h
e
m
u
lti
-
class
class
if
icat
io
n
o
f
to
m
ato
leaf
d
is
ea
s
es,
wh
ich
in
clu
d
e
1
0
class
es.
T
o
m
in
im
ize
l
o
s
s
,
th
e
s
tep
s
ize
d
u
r
in
g
tr
ain
i
n
g
e
p
o
ch
s
is
d
et
er
m
in
ed
b
y
th
e
lear
n
in
g
r
ate,
wh
ich
in
o
u
r
m
o
d
el
is
s
et
at
0
.
0
0
0
0
1
.
W
h
ile
e
x
ac
t
co
n
v
e
r
g
en
ce
is
en
c
o
u
r
ag
e
d
b
y
a
d
ec
r
ea
s
ed
lear
n
i
n
g
r
ate,
tr
ain
in
g
tim
es
m
ay
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
V
o
tTo
mN
et:
V
o
tin
g
-
b
a
s
ed
t
o
ma
to
d
is
ea
s
e
d
ia
g
n
o
s
is
w
ith
tr
a
n
s
fer lea
r
n
in
g
(
S
h
r
a
d
h
a
J
o
s
h
i
-
B
a
g
)
41
in
cr
ea
s
e.
T
h
e
ef
f
ec
tiv
en
ess
o
f
ea
ch
p
r
e
-
tr
ain
ed
m
o
d
el
—
I
n
c
ep
tio
n
Net,
R
esNet,
Mo
b
ileNet,
E
f
f
icien
tNet,
an
d
Den
s
eNe
t
—
in
id
en
tify
in
g
ag
r
icu
ltu
r
al
leaf
d
is
ea
s
es
is
ass
es
s
ed
s
ep
ar
ately
.
T
r
an
s
f
er
lear
n
in
g
is
th
e
p
r
o
ce
s
s
o
f
u
s
in
g
p
r
e
-
tr
ain
ed
m
o
d
els
to
i
m
p
r
o
v
e
o
v
er
all
p
e
r
f
o
r
m
an
ce
i
n
im
ag
e
ca
teg
o
r
izatio
n
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
t
h
e
VGG1
6
,
I
n
ce
p
tio
n
Net,
R
esN
et,
Mo
b
ileNet,
E
f
f
icien
tNet,
an
d
Den
s
eNe
t
d
u
r
in
g
tr
ain
in
g
an
d
v
alid
atio
n
is
ex
am
in
ed
.
T
h
e
n
etwo
r
k
s
wer
e
tr
ain
ed
u
s
in
g
d
ata
o
n
t
o
m
ato
l
ea
f
d
is
ea
s
e.
T
o
b
e
u
s
ed
at
a
lat
er
tim
e,
th
e
tr
ain
ed
an
d
v
er
i
f
ied
m
o
d
els ar
e
p
r
eser
v
ed
o
n
d
is
k
.
2
.
3
.
Appl
y
ing
ens
em
bli
ng
wit
h so
f
t
a
nd
ha
rd
v
o
t
ing
f
o
r
e
nh
a
ncing
cla
s
s
if
ica
t
io
n a
cc
ura
cy
I
n
o
r
d
er
to
in
cr
ea
s
e
class
if
icatio
n
ac
cu
r
ac
y
,
en
s
em
b
le
s
o
f
t
v
o
tin
g
an
d
h
ar
d
v
o
tin
g
ar
e
em
p
lo
y
ed
.
Alg
o
r
ith
m
1
is
a
wr
itten
an
d
p
r
in
te
d
v
er
s
io
n
o
f
th
e
en
s
em
b
le
lear
n
in
g
u
tili
zin
g
s
o
f
t
an
d
h
a
r
d
v
o
tin
g
(
Vo
tTo
m
Net)
alg
o
r
ith
m
s
.
T
h
e
m
eth
o
d
u
s
es
s
ix
p
r
e
-
tr
ai
n
e
d
im
ag
e
class
if
icatio
n
m
o
d
el
s
to
s
o
f
t
v
o
te
to
d
eter
m
in
e
th
e
f
in
al
p
r
e
d
icted
c
lass
n
am
e.
I
t d
is
p
lay
s
th
e
p
s
eu
d
o
-
co
d
e
f
o
r
th
is
p
r
o
ce
s
s
.
Alg
o
r
it
h
m
:
E
n
s
e
m
b
le
le
ar
n
i
n
g
u
s
i
n
g
s
o
f
t
an
d
h
a
r
d
v
o
ti
n
g
(
V
o
tT
o
m
N
et
)
In
p
ut
:
D
is
e
as
ed
C
r
op
L
e
af
I
n
st
a
nc
es
Ou
t
pu
t
:
Cl
a
ss
n
am
e
o
f
d
is
ea
s
eC
V
o
t
T
o
m
N
e
t
1.
Lo
a
d
th
e
p
ic
kl
e
d
mo
d
el
s:
VG
G
16
,
D
en
se
Ne
t
,
In
c
ep
ti
o
nN
e
t,
M
o
bi
le
Ne
t
,
Re
s
Ne
t,
Ef
f
ic
ie
n
tN
et
2.
Pr
e
pr
oc
e
ss
t
he
in
pu
t
t
es
t
in
st
an
c
e
im
ag
e
s
3.
In
i
ti
al
i
ze
a
n
a
rr
ay
to
s
t
or
e
p
re
d
ic
te
d
p
ro
ba
b
il
it
i
es
fo
r
e
ac
h
mo
d
el
4.
Fo
r
e
ac
h
t
es
t
i
ns
ta
n
ce
do
:
a.
Ob
t
ai
n
p
re
di
ct
e
d
pr
o
ba
bi
l
it
i
es
f
r
om
V
GG
1
6,
D
e
ns
eN
e
t,
In
ce
p
ti
on
Ne
t
,
Mo
b
il
eN
e
t,
Re
sN
e
t,
Ef
f
ic
i
en
tN
e
t
b.
St
o
re
t
h
e
pr
ed
i
ct
ed
pr
ob
a
bi
l
it
ie
s
i
n
th
e
a
rr
a
y
En
d
f
o
r
5.
Ca
l
cu
la
t
e
av
er
a
ge
p
r
ed
ic
t
ed
pr
ob
a
bi
li
ti
e
s
fo
r
e
ac
h
c
l
as
s
a
cr
os
s
t
he
m
o
de
ls
:
-
I
ni
t
ia
li
z
e
an
a
r
ra
y
t
o
st
o
re
av
er
a
ge
p
ro
b
ab
il
i
ti
es
-
fo
r
e
a
ch
c
l
as
sd
o
Ca
l
cu
l
at
e
t
he
a
ve
r
ag
e
p
ro
ba
b
il
i
ty
b
y
a
ve
ra
g
in
g
c
or
re
s
po
n
di
ng
pr
ob
ab
i
li
ti
e
s
fr
o
m
a
ll
m
o
de
ls
En
d
f
o
r
6.
Se
l
ec
t
t
he
c
la
s
s
la
b
el
w
h
ic
h
i
s
h
av
in
g
m
ax
im
u
m
av
e
ra
g
e
pr
o
ba
bi
li
t
y
as
th
e
f
in
a
l
pr
e
di
ct
ed
cl
a
ss
7.
Ou
t
pu
t
t
he
f
in
a
l
pr
e
di
ct
e
d
c
la
ss
na
me
C
V
o
t
T
o
m
N
e
t
Usi
n
g
an
e
n
s
em
b
le
ap
p
r
o
ac
h
ca
lled
s
o
f
t
an
d
h
ar
d
v
o
tin
g
,
th
e
p
r
o
jecte
d
p
r
o
b
ab
ilit
ies
o
f
ea
ch
m
o
d
el
ar
e
av
er
ag
ed
,
an
d
th
e
class
with
th
e
h
ig
h
est
av
er
ag
e
p
r
o
b
a
b
ilit
y
is
id
en
tifie
d
as
th
e
f
in
al
p
r
ed
icted
class
.
T
h
e
Vo
tTo
m
Net
alg
o
r
ith
m
’
s
im
p
le
m
en
tatio
n
u
tili
ze
s
(
1
)
a
n
d
(
2
)
,
wh
ich
ar
e
d
e
r
iv
ed
m
at
h
em
atica
lly
.
=
(
+
+
+
+
+
)
6
(
1
)
w
h
er
e
is
av
er
ag
e
o
f
p
r
o
b
ab
ilit
ies p
r
e
d
icted
f
o
r
ea
ch
class
f
o
r
s
o
f
t a
n
d
h
ar
d
v
o
tin
g
is
p
r
ed
icted
p
r
o
b
ab
ilit
ies b
y
VGG1
6
is
p
r
ed
icted
p
r
o
b
ab
ilit
ies b
y
I
n
ce
p
tio
n
is
p
r
ed
icted
p
r
o
b
ab
ilit
ies b
y
Den
s
eNe
t
is
p
r
ed
icted
p
r
o
b
ab
ilit
ies b
y
Mo
b
ile
N
et
is
p
r
ed
icted
p
r
o
b
ab
ilit
ies b
y
R
es
N
et
is
p
r
ed
icted
p
r
o
b
ab
ilit
ies b
y
E
f
f
icien
t
N
et
=ma
x
(
)
(
2
)
w
h
er
e
is
av
er
ag
e
o
f
p
r
o
b
a
b
ilit
ies p
r
ed
icted
f
o
r
ea
ch
class
f
o
r
s
o
f
t
v
o
tin
g
is
class
p
r
ed
icted
b
y
p
r
o
p
o
s
ed
m
o
d
el
Vo
tTo
m
Net
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
tr
ain
e
d
en
s
em
b
le
d
m
o
d
el
is
test
ed
th
r
o
u
g
h
a
d
esk
t
o
p
a
p
p
licatio
n
d
ev
elo
p
e
d
wh
ic
h
will
ask
to
in
p
u
t
th
e
im
ag
e
o
f
a
leaf
an
d
d
is
p
lay
th
e
d
etec
ted
class
o
f
d
is
ea
s
e
f
o
r
th
e
leaf
,
as
s
h
o
wn
in
Fig
u
r
e
2
.
T
h
e
d
esk
to
p
a
p
p
licatio
n
s
cr
ee
n
s
h
o
t
f
o
r
h
ea
lth
y
leaf
p
r
e
d
ictio
n
i
s
s
h
o
wn
in
Fig
u
r
e
2
(
a
)
a
n
d
t
h
e
im
ag
e
with
leaf
d
is
ea
s
e
an
d
th
e
n
am
e
o
f
th
e
d
i
s
ea
s
e
i.e
.
to
m
ato
m
o
s
aic
v
ir
u
s
is
s
h
o
wn
in
Fig
u
r
e
2
(
b
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
4
,
No
.
1
,
Ma
r
ch
2
0
2
5
:
38
-
46
42
(
a)
(
b
)
Fig
u
r
e
2
.
T
h
e
d
esk
to
p
ap
p
licat
io
n
s
cr
ee
n
s
h
o
t f
o
r
Vo
tTo
m
Net
with
(
a)
h
ea
lth
y
an
d
(
b
)
d
is
ea
s
e
leaf
p
r
ed
ictio
n
All
th
e
m
o
d
els
wer
e
tr
ai
n
ed
u
s
in
g
a
d
ataset
o
f
to
m
ato
le
af
d
is
ea
s
es,
an
d
th
eir
p
er
f
o
r
m
an
ce
was
ass
es
s
ed
u
s
in
g
ac
cu
r
ac
y
m
ea
s
u
r
es
th
r
o
u
g
h
o
u
t
all
ep
o
ch
s
.
A
s
am
p
le
class
if
icatio
n
r
ep
o
r
t
o
f
VGG1
6
is
s
h
o
wn
in
T
ab
le
1
.
T
h
e
ac
cu
r
ac
y
tr
e
n
d
s
o
v
e
r
e
p
o
ch
s
a
r
e
s
h
o
wn
i
n
Fig
u
r
e
3
.
A
co
n
s
is
ten
t
r
is
in
g
tr
ajec
to
r
y
i
n
th
e
tr
ain
in
g
ac
c
u
r
ac
y
s
h
o
wed
t
h
a
t
th
e
s
y
s
tem
was
co
n
tin
u
o
u
s
l
y
lear
n
in
g
f
r
o
m
t
h
e
tr
ain
i
n
g
s
et.
Simu
ltan
eo
u
s
ly
,
th
e
v
alid
atio
n
ac
cu
r
ac
y
also
r
o
s
e,
th
o
u
g
h
s
p
o
r
a
d
ically
co
m
m
o
n
o
cc
u
r
r
en
ce
in
d
ee
p
lear
n
i
n
g
tr
ai
n
in
g
s
h
o
win
g
s
tr
o
n
g
g
en
e
r
aliza
tio
n
to
p
r
e
v
io
u
s
ly
u
n
s
ee
n
d
ata.
Fig
u
r
e
3
s
h
o
ws
h
o
w
th
e
tr
ain
in
g
lo
s
s
f
o
r
I
n
ce
p
tio
n
Net,
Den
s
eNe
t,
an
d
VGG1
6
s
tead
ily
d
r
o
p
p
e
d
as
th
e
ep
o
ch
s
wen
t
o
n
.
T
h
is
d
ec
r
ea
s
e
in
d
icate
s
th
at
tr
ain
in
g
was
s
u
cc
ess
f
u
l
in
m
in
im
izin
g
er
r
o
r
s
an
d
p
r
o
m
o
tin
g
lear
n
i
n
g
.
De
n
s
eNe
t
an
d
I
n
ce
p
tio
n
Net
also
ex
h
ib
it
co
m
p
a
r
ab
le
p
er
f
o
r
m
an
ce
m
etr
ics,
s
u
ch
as tr
ain
in
g
an
d
v
alid
atio
n
ac
cu
r
ac
y
,
as we
ll a
s
tr
ain
in
g
an
d
v
alid
atio
n
lo
s
s
.
T
ab
le
1
.
C
lass
if
icatio
n
r
ep
o
r
t
o
f
V
G
G
1
6
D
i
sea
s
e
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
S
u
p
p
o
r
t
To
ma
t
o
B
a
c
t
e
r
i
a
l
sp
o
t
0
.
9
3
0
.
9
7
0
.
9
5
3
0
1
To
ma
t
o
E
a
r
l
y
b
l
i
g
h
t
0
.
9
9
0
.
9
0
.
9
4
2
9
8
To
ma
t
o
H
e
a
l
t
h
y
0
.
9
6
1
0
.
9
8
3
3
4
To
ma
t
o
L
a
t
e
b
l
i
g
h
t
0
.
8
8
0
.
9
6
0
.
9
2
3
0
2
To
ma
t
o
L
e
a
f
M
o
l
d
0
.
9
9
0
.
8
0
.
8
9
2
7
2
To
ma
t
o
M
o
s
a
i
c
v
i
r
u
s
0
.
9
9
0
.
9
8
0
.
9
9
3
0
2
To
ma
t
o
S
e
p
t
o
r
i
a
l
e
a
f
sp
o
t
0
.
8
8
0
.
9
3
0
.
9
1
2
9
4
To
ma
t
o
S
p
i
d
e
r
m
i
t
e
s
0
.
8
5
1
0
.
9
2
3
0
5
To
ma
t
o
T
a
r
g
e
t
S
p
o
t
0
.
8
5
0
.
8
5
0
.
8
5
2
9
7
To
ma
t
o
Y
e
l
l
o
w
L
e
a
f
C
u
r
l
V
i
r
u
s
0
.
9
8
0
.
9
4
0
.
9
6
2
9
5
M
e
t
r
i
c
V
a
l
u
e
A
c
c
u
r
a
c
y
9
2
.
6
M
a
c
r
o
a
v
e
r
a
g
e
9
2
.
9
W
e
i
g
h
t
e
d
a
v
e
r
a
g
e
9
3
.
1
No
tab
le
f
in
d
in
g
s
f
r
o
m
t
h
e
class
if
icatio
n
r
ep
o
r
t
in
clu
d
e
h
ig
h
p
r
ec
is
io
n
s
co
r
es
f
o
r
d
is
ea
s
es
in
clu
d
in
g
to
m
ato
m
o
s
aic
v
ir
u
s
(
0
.
9
9
)
,
to
m
ato
leaf
m
o
ld
(
0
.
9
9
)
,
a
n
d
to
m
ato
ea
r
ly
b
lig
h
t
(
0
.
9
9
)
,
wh
ic
h
s
h
o
w
a
lo
w
r
ate
o
f
f
alse
p
o
s
itiv
es.
W
ith
r
ec
all
s
co
r
es
o
f
1
.
0
0
,
T
o
m
ato
Hea
lth
y
an
d
T
o
m
ato
Sp
id
er
m
it
es
wer
e
th
e
m
o
s
t
s
u
cc
ess
f
u
l,
in
d
icatin
g
th
at
alm
o
s
t
all
r
ea
l
ca
s
es
wer
e
ac
cu
r
at
ely
id
en
tifie
d
.
Pre
cisi
o
n
,
r
ec
all,
an
d
F1
-
s
co
r
es
f
o
r
to
m
ato
b
ac
ter
ial
s
p
o
t
(
0
.
9
5
)
a
n
d
to
m
at
o
y
ello
w
leaf
cu
r
l
v
ir
u
s
(
0
.
9
6
)
wer
e
f
o
u
n
d
to
b
e
in
b
alan
ce
.
W
ith
a
to
tal
ac
cu
r
ac
y
o
f
9
2
.
6
%,
th
e
m
o
d
el
was sh
o
wn
to
b
e
ac
cu
r
ate
in
cl
ass
if
y
in
g
9
2
.
6
% o
f
to
m
ato
d
is
ea
s
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
V
o
tTo
mN
et:
V
o
tin
g
-
b
a
s
ed
t
o
ma
to
d
is
ea
s
e
d
ia
g
n
o
s
is
w
ith
tr
a
n
s
fer lea
r
n
in
g
(
S
h
r
a
d
h
a
J
o
s
h
i
-
B
a
g
)
43
V
alid
atio
n
ac
cu
r
ac
y
v
s
n
u
m
b
e
r
o
f
ep
o
ch
s
an
d
tr
ain
in
g
V
alid
atio
n
lo
s
s
v
s
n
u
m
b
er
o
f
e
p
o
ch
s
Fig
u
r
e
3
.
Gr
a
p
h
o
f
VGG1
6
f
o
r
tr
ain
in
g
Pre
-
tr
ain
ed
m
o
d
els
s
u
ch
as
De
n
s
eNe
t,
I
n
ce
p
tio
n
Net,
Mo
b
ile
Net,
R
esNet,
an
d
E
f
f
icien
tNet
wer
e
also
tr
ain
ed
an
d
ass
ess
ed
.
W
ith
an
ac
cu
r
ac
y
o
f
9
7
.
8
3
%,
Den
s
eNe
t
wa
s
th
e
m
o
s
t
ac
cu
r
ate,
f
o
llo
wed
b
y
I
n
ce
p
tio
n
Net
with
9
5
.
6
1
%.
T
h
e
ac
cu
r
ac
y
v
alu
es
o
f
th
e
o
t
h
er
p
r
e
-
tr
ain
ed
m
o
d
els:
R
es
Net
Mo
b
ileNet
an
d
E
f
f
icien
tNet
ar
e
9
5
.
1
2
%,
9
4
.
3
%,
an
d
9
6
.
7
%
,
r
esp
ec
tiv
ely
.
A
co
m
p
ar
is
o
n
g
r
a
p
h
s
h
o
win
g
th
ese
p
r
e
-
tr
ain
ed
m
o
d
els
’
class
if
icatio
n
ac
cu
r
a
cies
is
s
h
o
wn
in
Fig
u
r
e
4
.
T
h
e
“
tr
ain
in
g
ac
c
u
r
ac
y
a
n
d
v
a
lid
atio
n
ac
cu
r
ac
y
”
g
r
ap
h
illu
s
tr
ates
h
o
w
d
if
f
er
en
t
d
ee
p
lear
n
in
g
m
o
d
els
b
eh
av
e
in
ter
m
s
o
f
ac
c
u
r
ac
y
o
n
tr
ai
n
in
g
a
n
d
v
alid
atio
n
d
atasets
.
Fig
u
r
e
4
d
is
p
lay
s
a
9
2
%
tr
ain
in
g
ac
cu
r
ac
y
a
n
d
a
9
0
%
v
alid
atio
n
ac
cu
r
ac
y
f
o
r
th
e
VGG1
6
.
Po
ten
tial
o
v
er
f
itti
n
g
is
in
d
icate
d
b
y
th
e
m
o
d
el
’
s
m
ar
g
in
ally
s
u
p
er
i
o
r
p
er
f
o
r
m
an
ce
o
n
tr
ain
in
g
d
ata
as
o
p
p
o
s
ed
to
v
alid
atio
n
d
ata.
De
n
s
eNe
t
h
as
a
9
8
%
tr
ain
in
g
ac
cu
r
ac
y
an
d
a
9
6
%
v
alid
atio
n
ac
cu
r
ac
y
.
Den
s
eNe
t
h
as
g
o
o
d
p
er
f
o
r
m
an
ce
with
o
n
ly
a
s
m
a
ll
am
o
u
n
t
o
f
o
v
er
f
itti
n
g
,
with
v
er
y
h
ig
h
tr
ain
in
g
ac
cu
r
ac
y
an
d
s
lig
h
tly
lo
we
r
v
alid
atio
n
ac
cu
r
ac
y
.
C
o
m
p
ar
a
b
ly
,
th
e
Den
s
eNe
t
an
d
I
n
ce
p
t
io
n
Net
m
o
d
els,
w
h
ich
h
a
v
e
r
esp
ec
tiv
e
ac
cu
r
ac
y
p
er
ce
n
tag
es
o
f
9
6
%
an
d
9
5
% o
n
tr
ain
in
g
an
d
v
alid
atio
n
s
ets,
s
h
o
w
s
tr
o
n
g
g
e
n
er
aliza
tio
n
a
n
d
litt
le
o
v
er
f
itti
n
g
.
T
h
e
v
alid
atio
n
a
cc
u
r
ac
y
was 9
2
% a
n
d
th
e
tr
ai
n
in
g
ac
c
u
r
ac
y
o
f
Mo
b
ileNet
was
9
4
%.
Fig
u
r
e
4
.
C
o
m
p
a
r
is
o
n
g
r
ap
h
o
f
tr
ain
in
g
a
n
d
v
alid
atio
n
ac
cu
r
ac
y
f
o
r
m
o
d
els VGG
1
6
,
I
n
ce
p
tio
n
Net,
R
esNet,
Mo
b
ileNet,
E
f
f
icien
tNet
an
d
Den
s
eNe
t
T
h
e
m
o
d
er
ate
d
is
cr
ep
an
cy
in
ac
cu
r
ac
y
b
etwe
en
tr
ain
i
n
g
a
n
d
v
alid
atio
n
f
o
r
Mo
b
ileNet
in
d
icate
s
s
o
m
e
o
v
er
f
itti
n
g
b
u
t
o
v
e
r
all
s
tr
o
n
g
p
er
f
o
r
m
an
ce
,
with
a
9
4
%
v
alid
atio
n
ac
cu
r
ac
y
an
d
9
5
%
tr
ain
in
g
ac
cu
r
ac
y
f
o
r
R
esNet.
R
esNet
ex
h
ib
its
g
o
o
d
g
en
e
r
aliza
tio
n
with
a
s
m
all
g
ap
an
d
h
i
g
h
ac
c
u
r
ac
y
f
o
r
b
o
th
tr
ain
in
g
an
d
v
alid
atio
n
s
ets.
W
ith
E
f
f
icien
tNet,
9
7
%
o
f
tr
ain
in
g
an
d
9
5
%
o
f
v
alid
atio
n
ac
cu
r
ac
y
ar
e
ac
h
ie
v
ed
.
Ad
d
itio
n
ally
,
E
f
f
icien
tNet
ex
h
ib
its
ex
ce
llen
t
p
er
f
o
r
m
a
n
ce
an
d
o
u
ts
tan
d
in
g
g
en
e
r
aliza
tio
n
with
v
er
y
h
ig
h
tr
ain
in
g
ac
cu
r
ac
y
an
d
s
o
m
ewh
at
lo
wer
v
alid
atio
n
ac
c
u
r
ac
y
w
ith
a
tin
y
g
a
p
.
T
o
in
cr
ea
s
e
th
e
ac
c
u
r
ac
y
o
f
t
o
m
ato
leaf
d
is
ea
s
e
id
en
tific
atio
n
in
th
e
ex
p
e
r
im
en
t,
we
e
n
s
em
b
led
s
ix
p
r
e
-
tr
ain
ed
d
ee
p
lear
n
in
g
m
o
d
els,
in
clu
d
in
g
VGG1
6
,
I
n
ce
p
tio
n
Net,
R
esNet,
Mo
b
ileNet,
E
f
f
icien
tNet,
an
d
Den
s
eNe
t.
T
h
e
m
o
d
els we
r
e
p
ick
led
f
o
r
later
u
s
e
af
ter
b
ein
g
in
d
ep
en
d
e
n
tly
tr
ain
ed
an
d
v
al
id
ated
o
n
th
e
1
,
5
0
0
p
h
o
to
s
f
o
r
1
0
d
if
f
er
e
n
t
class
es
in
th
e
to
m
at
o
leaf
d
is
ea
s
e
d
ataset.
Nex
t,
we
u
s
ed
t
h
e
e
n
s
em
b
le
s
o
f
t
v
o
tin
g
m
eth
o
d
.
T
ab
le
2
p
r
in
ts
th
e
class
if
icatio
n
r
ep
o
r
t
f
o
r
th
e
Vo
tTo
m
Net
m
o
d
el.
T
ab
le
3
d
is
p
lay
s
th
e
in
d
iv
id
u
al
87
89
91
93
95
97
99
VG
G
16
D
e
ns
e
N
e
t
Inc
e
pt
i
o
nN
e
t
M
o
bi
l
e
N
e
t
R
e
sN
e
t
E
f
f
i
c
i
e
nt
N
e
t
M
ode
l
A
ccur
acy
i
n %
M
ach
i
n
e
L
e
ar
n
i
n
g
M
od
e
l
s
T
r
ai
ni
ng
A
c
c
ur
ac
y
Val
i
da
t
i
o
n
A
c
c
ur
a
c
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
4
,
No
.
1
,
Ma
r
ch
2
0
2
5
:
38
-
46
44
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
b
o
th
b
ef
o
r
e
an
d
a
f
ter
th
e
en
s
em
b
l
e.
So
f
t
v
o
tin
g
was
u
s
ed
to
en
s
em
b
le
VGG1
6
,
I
n
ce
p
tio
n
Net,
R
esNet,
Mo
b
ile
Net,
E
f
f
icien
tNet,
a
n
d
Den
s
e
Net,
wh
ich
i
n
cr
ea
s
ed
t
o
m
ato
l
ea
f
d
is
ea
s
e
d
etec
tio
n
ac
cu
r
ac
y
an
d
r
o
b
u
s
tn
ess
.
T
h
e
g
r
ap
h
i
n
Fig
u
r
e
5
illu
s
tr
ates
th
e
s
lig
h
t
ad
v
an
tag
e
in
v
alid
atio
n
ac
cu
r
ac
y
b
etwe
en
th
e
two
tech
n
iq
u
es: b
ef
o
r
e
en
s
em
b
le
a
n
d
a
f
ter
en
s
e
m
b
le
u
s
in
g
s
o
f
t a
n
d
h
a
r
d
v
o
tin
g
.
T
ab
le
2
.
T
h
e
r
ep
o
r
t o
f
f
o
r
Vo
t
T
o
m
Net
m
o
d
el
D
i
sea
s
e
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
S
u
p
p
o
r
t
To
ma
t
o
B
a
c
t
e
r
i
a
l
sp
o
t
0
.
9
9
0
.
9
9
0
.
9
9
3
0
1
To
ma
t
o
E
a
r
l
y
b
l
i
g
h
t
0
.
9
9
2
0
.
9
2
0
.
9
5
5
2
9
8
To
ma
t
o
H
e
a
l
t
h
y
1
0
.
9
8
8
0
.
9
9
4
3
3
4
To
ma
t
o
L
a
t
e
b
l
i
g
h
t
0
.
9
8
5
0
.
9
8
0
.
9
8
2
5
3
0
2
To
ma
t
o
L
e
a
f
M
o
l
d
0
.
9
8
9
0
.
9
8
5
0
.
9
8
7
2
7
2
To
ma
t
o
M
o
s
a
i
c
v
i
r
u
s
0
.
9
8
7
0
.
9
9
3
0
.
9
9
3
0
2
To
ma
t
o
S
e
p
t
o
r
i
a
l
e
a
f
sp
o
t
0
.
9
6
8
0
.
9
8
3
0
.
9
7
5
2
9
4
To
ma
t
o
S
p
i
d
e
r
m
i
t
e
s
0
.
9
4
1
0
.
9
6
9
3
0
5
To
ma
t
o
T
a
r
g
e
t
S
p
o
t
0
.
9
6
0
.
9
7
3
0
.
9
6
6
5
2
9
7
To
ma
t
o
Y
e
l
l
o
w
L
e
a
f
C
u
r
l
V
i
r
u
s
0
.
9
9
3
0
.
9
9
3
0
.
9
9
3
2
9
5
M
e
t
r
i
c
V
a
l
u
e
A
c
c
u
r
a
c
y
-
-
0
.
9
8
6
1
3
0
0
0
M
a
c
r
o
a
v
e
r
a
g
e
0
.
9
8
7
2
0
.
9
8
6
2
0
.
9
8
6
5
3
0
0
0
W
e
i
g
h
t
e
d
a
v
e
r
a
g
e
0
.
9
8
6
3
0
.
9
8
6
4
0
.
9
8
6
5
3
0
0
0
T
ab
le
3
.
Acc
u
r
ac
y
v
al
u
es in
v
ar
io
u
s
s
ce
n
ar
io
s
Pre
-
t
r
a
i
n
e
d
M
o
d
e
l
s
B
e
f
o
r
e
E
n
sem
b
l
i
n
g
A
f
t
e
r
E
n
sem
b
l
i
n
g
a
c
c
u
r
a
c
y
Tr
a
i
n
i
n
g
a
c
c
u
r
a
c
y
V
a
l
i
d
a
t
i
o
n
a
c
c
u
r
a
c
y
V
G
G
1
6
9
2
.
6
1
9
1
.
8
8
9
9
.
2
1
D
e
n
seN
e
t
9
7
.
8
3
9
7
.
1
2
I
n
c
e
p
t
i
o
n
N
e
t
9
5
.
6
1
9
4
.
8
8
M
o
b
i
l
e
N
e
t
9
4
.
3
9
4
.
5
4
R
e
sN
e
t
9
5
.
1
2
9
5
.
4
3
Ef
f
i
c
i
e
n
t
N
e
t
9
6
.
7
9
6
.
2
Fig
u
r
e
5
.
C
o
m
p
a
r
is
o
n
g
r
ap
h
o
f
ac
cu
r
ac
ies b
ef
o
r
e
an
d
af
ter
e
n
s
em
b
le
f
o
r
m
o
d
els VGG
1
6
,
I
n
ce
p
tio
n
Net,
R
esNet,
Mo
b
ileNet,
E
f
f
icien
tNet,
an
d
Den
s
eNe
t w
ith
Vo
tTo
m
Net
4.
CO
NCLU
SI
O
N
I
n
o
r
d
er
to
a
u
to
m
ate
t
h
e
d
ete
ctio
n
o
f
leaf
d
is
ea
s
es
in
to
m
a
to
cr
o
p
s
,
th
e
r
esear
ch
p
r
esen
t
s
a
n
o
v
el
m
o
d
el
ca
lled
Vo
tTo
m
Net
th
at
m
ak
es
u
s
e
o
f
C
NN
an
d
tr
an
s
f
er
lear
n
in
g
tech
n
iq
u
es.
T
h
e
s
tu
d
y
g
r
ea
tly
im
p
r
o
v
es
class
if
icatio
n
ac
c
u
r
a
cy
in
d
if
f
er
en
tiatin
g
b
etwe
en
d
am
ag
ed
an
d
h
ea
lth
y
to
m
ato
leav
es
b
y
u
tili
zin
g
p
r
e
-
tr
ain
ed
m
o
d
els
s
u
ch
as
VGG1
6
,
I
n
ce
p
tio
n
Net,
R
esNet,
Mo
b
ileNet,
E
f
f
icien
tNet,
an
d
Den
s
eNe
t.
B
y
co
m
b
in
in
g
th
e
ad
v
an
ta
g
es
o
f
d
is
tin
ct
m
o
d
els,
th
e
s
u
g
g
ested
m
o
d
el
u
s
es
en
s
em
b
le
lear
n
in
g
v
ia
s
o
f
t
an
d
h
ar
d
v
o
tin
g
to
in
cr
ea
s
e
p
r
ed
ictio
n
r
esil
ien
ce
an
d
r
eliab
ilit
y
.
B
y
allo
win
g
th
e
C
NN
m
o
d
els
t
o
lear
n
f
r
o
m
la
r
g
e
d
atasets
lik
e
I
m
ag
eNe
t,
t
r
an
s
f
er
lear
n
in
g
p
lay
s
a
c
r
itical
r
o
l
e
in
im
p
r
o
v
in
g
th
e
m
o
d
els
’
c
ap
ac
ity
to
p
r
ec
is
ely
class
if
y
leaf
d
is
ea
s
e
s
.
B
y
co
m
b
in
in
g
th
e
r
esu
lts
o
f
m
u
ltip
le
m
o
d
els,
th
is
m
eth
o
d
r
ed
u
ce
s
b
iases
an
d
er
r
o
r
s
th
at
9
2
,
6
1
9
7
,
8
3
9
5
,
6
1
9
4
,
3
9
5
,
1
2
9
6
,
7
9
1
,
8
8
9
7
,
1
2
9
4
,
8
8
9
4
,
5
4
9
5
,
4
3
9
6
,
2
9
9
,
2
1
70
75
80
85
90
95
100
VG
G
1
6
D
en
s
eN
et
I
n
c
ep
t
i
o
n
N
et
M
o
b
i
l
eN
et
R
es
N
et
E
f
f
i
c
i
en
t
N
et
E
n
s
em
b
l
e b
y
S
o
f
t
an
d
Hard
V
o
t
i
n
g
M
o
d
e
l
Ac
c
u
r
ac
y
i
n
%
M
ac
h
i
n
e
Lear
n
i
n
g
M
o
d
e
l
s
T
r
ai
ni
ng
A
c
c
ur
ac
y
Val
i
da
t
i
o
n
A
c
c
ur
a
c
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
V
o
tTo
mN
et:
V
o
tin
g
-
b
a
s
ed
t
o
ma
to
d
is
ea
s
e
d
ia
g
n
o
s
is
w
ith
tr
a
n
s
fer lea
r
n
in
g
(
S
h
r
a
d
h
a
J
o
s
h
i
-
B
a
g
)
45
ar
e
co
m
m
o
n
in
s
o
lo
m
o
d
els.
W
ith
Vo
tTo
m
Net,
clas
s
if
icati
o
n
ac
cu
r
ac
y
in
c
r
ea
s
ed
s
ig
n
if
ican
tly
as
ev
id
en
ce
d
b
y
th
e
f
in
d
i
n
g
s
,
wh
ich
s
h
o
w
an
asto
u
n
d
in
g
9
9
.
2
%
ac
cu
r
ac
y
th
r
o
u
g
h
s
o
f
t
a
n
d
h
ar
d
v
o
tin
g
.
Hig
h
r
ec
all
an
d
p
r
ec
is
io
n
r
ates
ar
e
s
h
o
wn
b
y
VGG1
6
,
I
n
ce
p
tio
n
Net,
R
esNet,
Mo
b
ileNet,
E
f
f
icien
tNet,
an
d
Den
s
eNe
t
f
o
r
a
v
ar
iety
o
f
t
o
m
ato
illn
ess
es.
Vo
tTo
m
Net,
a
r
ec
en
tly
cr
ea
t
ed
m
o
d
el,
h
as
p
r
ac
tical
im
p
li
ca
tio
n
s
th
at
in
clu
d
e
tar
g
eted
th
er
a
p
ies
f
o
r
h
ig
h
er
ag
r
icu
ltu
r
al
y
ield
an
d
ea
r
l
y
d
is
ea
s
e
id
en
tific
atio
n
to
m
in
i
m
ize
cr
o
p
lo
s
s
.
I
ts
ca
p
ac
ity
to
ad
ap
t
to
d
if
f
e
r
en
t
cr
o
p
s
in
c
r
ea
s
es
its
u
s
ef
u
ln
ess
in
a
wid
e
r
r
an
g
e
o
f
ag
r
icu
ltu
r
al
ap
p
licatio
n
s
a
n
d
g
r
ea
tly
in
cr
ea
s
es a
g
r
icu
ltu
r
al
p
r
o
d
u
ctiv
ity
.
ACK
NO
WL
E
DG
E
M
E
NT
S
T
h
e
au
th
o
r
s
wo
u
ld
lik
e
to
th
a
n
k
G.
H.
R
aiso
n
i
Un
iv
er
s
ity
,
Am
r
av
ati,
Ma
h
ar
ash
tr
a,
I
n
d
ia
an
d
N
K
Or
ch
id
C
o
lleg
e
o
f
E
n
g
in
ee
r
in
g
an
d
T
ec
h
n
o
lo
g
y
,
So
lap
u
r
,
Ma
h
ar
ash
tr
a,
I
n
d
ia
f
o
r
th
eir
s
u
p
p
o
r
t
an
d
h
elp
in
ex
p
er
im
en
tatio
n
th
r
o
u
g
h
o
u
t th
is
r
esear
ch
.
RE
F
E
R
E
NC
E
S
[
1
]
V
.
V
.
B
a
g
,
M
.
B
.
P
a
t
i
l
,
a
n
d
S
.
N
.
K
e
n
d
r
e
,
“
F
r
e
q
u
e
n
t
C
N
N
b
a
se
d
e
n
s
e
m
b
l
i
n
g
f
o
r
M
R
I
c
l
a
s
si
f
i
c
a
t
i
o
n
f
o
r
a
b
n
o
r
mal
b
r
a
i
n
g
r
o
w
t
h
d
e
t
e
c
t
i
o
n
,
”
J
o
u
rn
a
l
o
f
I
n
t
e
g
r
a
t
e
d
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
2
,
n
o
.
3
,
Ja
n
.
2
0
2
4
,
d
o
i
:
1
0
.
6
2
1
1
0
/
sc
i
e
n
c
e
i
n
.
j
i
st
.
2
0
2
4
.
v
1
2
.
7
8
5
.
[
2
]
S
.
C
h
a
v
a
t
e
,
R
.
M
i
sh
r
a
,
S
.
K
.
S
i
n
g
h
,
a
n
d
D
.
S
h
a
r
ma,
“
H
y
b
r
i
d
i
z
e
d
n
e
u
r
a
l
n
e
t
w
o
r
k
-
b
a
s
e
d
a
p
p
r
o
a
c
h
e
s
u
se
d
f
o
r
v
i
d
e
o
s
h
o
t
b
o
u
n
d
a
r
y
d
e
t
e
c
t
i
o
n
,
”
I
n
d
i
a
n
J
o
u
rn
a
l
O
f
S
c
i
e
n
c
e
A
n
d
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
6
,
n
o
.
3
3
,
p
p
.
2
6
7
0
–
2
6
8
0
,
S
e
p
.
2
0
2
3
,
d
o
i
:
1
0
.
1
7
4
8
5
/
I
JS
T/
v
1
6
i
3
3
.
1
6
7
1
.
[
3
]
A
.
R
.
P
a
t
h
a
k
,
M
.
P
a
n
d
e
y
,
a
n
d
S
.
R
a
u
t
a
r
a
y
,
“
A
p
p
l
i
c
a
t
i
o
n
o
f
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
,
”
Pr
o
c
e
d
i
a
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
1
3
2
,
p
p
.
1
7
0
6
–
1
7
1
7
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
r
o
c
s
.
2
0
1
8
.
0
5
.
1
4
4
.
[
4
]
A
.
K
.
H
a
s
e
,
P
.
S
.
A
h
e
r
,
a
n
d
S
.
K
.
H
a
s
e
,
“
D
e
t
e
c
t
i
o
n
,
c
a
t
e
g
o
r
i
z
a
t
i
o
n
a
n
d
su
g
g
e
s
t
i
o
n
t
o
c
u
r
e
i
n
f
e
c
t
e
d
p
l
a
n
t
s
o
f
t
o
m
a
t
o
a
n
d
g
r
a
p
e
s
b
y
u
si
n
g
O
p
e
n
C
V
f
r
a
m
e
w
o
r
k
f
o
r
a
n
d
r
i
o
d
e
n
v
i
r
o
n
me
n
t
,
”
i
n
2
0
1
7
2
n
d
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
f
o
r
C
o
n
v
e
rg
e
n
c
e
i
n
T
e
c
h
n
o
l
o
g
y
(
I
2
C
T
)
,
A
p
r
.
2
0
1
7
,
p
p
.
9
5
6
–
9
5
9
,
d
o
i
:
1
0
.
1
1
0
9
/
I
2
C
T.
2
0
1
7
.
8
2
2
6
2
7
0
.
[
5
]
Y
.
M
.
A
b
d
A
l
g
a
n
i
,
O
.
J.
M
.
C
a
r
o
,
L.
M
.
R
.
B
r
a
v
o
,
C
.
K
a
u
r
,
M
.
S
.
A
l
A
n
sari
,
a
n
d
B
.
K
.
B
a
l
a
,
“
Le
a
f
d
i
s
e
a
se
i
d
e
n
t
i
f
i
c
a
t
i
o
n
a
n
d
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
s
i
n
g
o
p
t
i
mi
z
e
d
d
e
e
p
l
e
a
r
n
i
n
g
,
”
M
e
a
s
u
rem
e
n
t
:
S
e
n
so
r
s
,
v
o
l
.
2
5
,
F
e
b
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
m
e
a
s
e
n
.
2
0
2
2
.
1
0
0
6
4
3
.
[
6
]
N
.
K
.
Tr
i
v
e
d
i
e
t
a
l
.
,
“
E
a
r
l
y
d
e
t
e
c
t
i
o
n
a
n
d
c
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
t
o
m
a
t
o
l
e
a
f
d
i
sea
s
e
u
s
i
n
g
h
i
g
h
-
p
e
r
f
o
r
m
a
n
c
e
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
S
e
n
so
rs
,
v
o
l
.
2
1
,
n
o
.
2
3
,
N
o
v
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
1
2
3
7
9
8
7
.
[
7
]
M
.
A
g
a
r
w
a
l
,
A
.
S
i
n
g
h
,
S
.
A
r
j
a
r
i
a
,
A
.
S
i
n
h
a
,
a
n
d
S
.
G
u
p
t
a
,
“
To
Le
D
:
t
o
ma
t
o
l
e
a
f
d
i
se
a
se
d
e
t
e
c
t
i
o
n
u
si
n
g
c
o
n
v
o
l
u
t
i
o
n
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
Pr
o
c
e
d
i
a
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
1
6
7
,
p
p
.
2
9
3
–
3
0
1
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
r
o
c
s.
2
0
2
0
.
0
3
.
2
2
5
.
[
8
]
J.
A
m
a
r
a
,
B
.
B
o
u
a
z
i
z
,
a
n
d
A
.
A
l
g
e
r
g
a
w
y
,
“
A
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
b
a
n
a
n
a
l
e
a
f
d
i
se
a
ses
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
i
n
L
e
c
t
u
re
N
o
t
e
s
i
n
I
n
f
o
rm
a
t
i
c
s (L
N
I
)
,
Pro
c
e
e
d
i
n
g
s
-
S
e
r
i
e
s
o
f
t
h
e
G
e
s
e
l
l
sc
h
a
f
t
f
u
r I
n
f
o
rm
a
t
i
k
(
G
I
)
,
2
0
1
7
,
v
o
l
.
2
6
6
,
p
p
.
7
9
–
8
8
.
[
9
]
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
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
i
o
s
y
st
e
ms
e
n
g
.
2
0
1
8
.
0
5
.
0
1
3
.
[
1
0
]
M
.
B
r
a
h
i
mi
,
K
.
B
o
u
k
h
a
l
f
a
,
a
n
d
A
.
M
o
u
ss
a
o
u
i
,
“
D
e
e
p
l
e
a
r
n
i
n
g
f
o
r
t
o
m
a
t
o
d
i
se
a
ses
:
c
l
a
ssi
f
i
c
a
t
i
o
n
a
n
d
s
y
m
p
t
o
ms
v
i
s
u
a
l
i
z
a
t
i
o
n
,
”
Ap
p
l
i
e
d
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
3
1
,
n
o
.
4
,
p
p
.
2
9
9
–
3
1
5
,
A
p
r
.
2
0
1
7
,
d
o
i
:
1
0
.
1
0
8
0
/
0
8
8
3
9
5
1
4
.
2
0
1
7
.
1
3
1
5
5
1
6
.
[
1
1
]
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
,
Ju
n
.
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
3
9
3
.
[
1
2
]
S
.
S
.
C
h
o
u
h
a
n
,
A
.
K
a
u
l
,
U
.
P
.
S
i
n
g
h
,
a
n
d
S
.
Ja
i
n
,
“
B
a
c
t
e
r
i
a
l
f
o
r
a
g
i
n
g
o
p
t
i
mi
z
a
t
i
o
n
b
a
se
d
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
n
e
u
r
a
l
n
e
t
w
o
r
k
(
B
R
B
F
N
N
)
f
o
r
i
d
e
n
t
i
f
i
c
a
t
i
o
n
a
n
d
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
p
l
a
n
t
l
e
a
f
d
i
sea
s
e
s:
a
n
a
u
t
o
ma
t
i
c
a
p
p
r
o
a
c
h
t
o
w
a
r
d
s
p
l
a
n
t
p
a
t
h
o
l
o
g
y
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
6
,
p
p
.
8
8
5
2
–
8
8
6
3
,
2
0
1
8
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
1
8
.
2
8
0
0
6
8
5
.
[
1
3
]
S
.
C
o
u
l
i
b
a
l
y
,
B
.
K
a
ms
u
-
F
o
g
u
e
m,
D
.
K
a
m
i
ss
o
k
o
,
a
n
d
D
.
Tr
a
o
r
e
,
“
D
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s
w
i
t
h
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
i
n
mi
l
l
e
t
c
r
o
p
i
ma
g
e
s,
”
C
o
m
p
u
t
e
rs i
n
I
n
d
u
s
t
ry
,
v
o
l
.
1
0
8
,
p
p
.
1
1
5
–
1
2
0
,
J
u
n
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
i
n
d
.
2
0
1
9
.
0
2
.
0
0
3
.
[
1
4
]
V
.
S
.
D
h
a
k
a
e
t
a
l
.
,
“
A
s
u
r
v
e
y
o
f
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
s
a
p
p
l
i
e
d
f
o
r
p
r
e
d
i
c
t
i
o
n
o
f
p
l
a
n
t
l
e
a
f
d
i
se
a
s
e
s,
”
S
e
n
s
o
rs
,
v
o
l
.
2
1
,
n
o
.
1
4
,
p
.
4
7
4
9
,
Ju
l
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
1
1
4
4
7
4
9
.
[
1
5
]
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
6
]
A
.
F
u
e
n
t
e
s,
S
.
Y
o
o
n
,
S
.
K
i
m,
a
n
d
D
.
P
a
r
k
,
“
A
r
o
b
u
st
d
e
e
p
-
l
e
a
r
n
i
n
g
-
b
a
s
e
d
d
e
t
e
c
t
o
r
f
o
r
r
e
a
l
-
t
i
me
t
o
m
a
t
o
p
l
a
n
t
d
i
sea
s
e
s
a
n
d
p
e
s
t
s
r
e
c
o
g
n
i
t
i
o
n
,
”
S
e
n
s
o
rs
,
v
o
l
.
1
7
,
n
o
.
9
,
S
e
p
.
2
0
1
7
,
d
o
i
:
1
0
.
3
3
9
0
/
s
1
7
0
9
2
0
2
2
.
[
1
7
]
K
.
Ti
a
n
,
J.
Z
e
n
g
,
T.
S
o
n
g
,
Z.
L
i
,
A
.
Ev
a
n
s,
a
n
d
J
.
L
i
,
“
To
m
a
t
o
l
e
a
f
d
i
s
e
a
ses
r
e
c
o
g
n
i
t
i
o
n
b
a
se
d
o
n
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
s,
”
J
o
u
r
n
a
l
o
f
A
g
r
i
c
u
l
t
u
r
a
l
E
n
g
i
n
e
e
ri
n
g
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
4
0
8
1
/
j
a
e
.
2
0
2
2
.
1
4
3
2
.
[
1
8
]
R
.
S
h
a
r
ma
,
M
.
M
i
t
t
a
l
,
V
.
G
u
p
t
a
,
a
n
d
D
.
V
a
s
d
e
v
,
“
D
e
t
e
c
t
i
o
n
o
f
p
l
a
n
t
l
e
a
f
d
i
s
e
a
s
e
u
si
n
g
a
d
v
a
n
c
e
d
d
e
e
p
l
e
a
r
n
i
n
g
a
r
c
h
i
t
e
c
t
u
r
e
s
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
I
n
f
o
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
6
,
n
o
.
6
,
p
p
.
3
4
7
5
–
3
4
9
2
,
A
u
g
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
4
1
8
7
0
-
0
2
4
-
0
1
9
3
7
-
4.
[
1
9
]
A
.
V
.
P
a
n
c
h
a
l
,
S
.
C
.
P
a
t
e
l
,
K
.
B
a
g
y
a
l
a
k
sh
m
i
,
P
.
K
u
mar
,
I
.
R
.
K
h
a
n
,
a
n
d
M
.
S
o
n
i
,
“
I
mag
e
-
b
a
se
d
p
l
a
n
t
d
i
sea
se
s
d
e
t
e
c
t
i
o
n
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
,
”
Ma
t
e
r
i
a
l
s T
o
d
a
y
:
Pr
o
c
e
e
d
i
n
g
s
,
v
o
l
.
8
0
,
p
p
.
3
5
0
0
–
3
5
0
6
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
ma
t
p
r
.
2
0
2
1
.
0
7
.
2
8
1
.
[
2
0
]
C
.
R
.
R
a
h
m
a
n
e
t
a
l
.
,
“
I
d
e
n
t
i
f
i
c
a
t
i
o
n
a
n
d
r
e
c
o
g
n
i
t
i
o
n
o
f
r
i
c
e
d
i
s
e
a
ses
a
n
d
p
e
st
s
u
s
i
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,
”
Bi
o
s
y
st
e
m
s
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
9
4
,
p
p
.
1
1
2
–
1
2
0
,
J
u
n
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
i
o
s
y
s
t
e
ms
e
n
g
.
2
0
2
0
.
0
3
.
0
2
0
.
[
2
1
]
E.
C
.
T
o
o
,
L.
Y
u
j
i
a
n
,
S
.
N
j
u
k
i
,
a
n
d
L.
Y
i
n
g
c
h
u
n
,
“
A
c
o
m
p
a
r
a
t
i
v
e
s
t
u
d
y
o
f
f
i
n
e
-
t
u
n
i
n
g
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
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
6
1
,
p
p
.
2
7
2
–
2
7
9
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
a
g
.
2
0
1
8
.
0
3
.
0
3
2
.
[
2
2
]
A
.
Jafar
,
N
.
B
i
b
i
,
R
.
A
.
N
a
q
v
i
,
A
.
S
a
d
e
g
h
i
-
N
i
a
r
a
k
i
,
a
n
d
D
.
Je
o
n
g
,
“
R
e
v
o
l
u
t
i
o
n
i
z
i
n
g
a
g
r
i
c
u
l
t
u
r
e
w
i
t
h
a
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
:
p
l
a
n
t
d
i
s
e
a
se
d
e
t
e
c
t
i
o
n
me
t
h
o
d
s,
a
p
p
l
i
c
a
t
i
o
n
s,
a
n
d
t
h
e
i
r
l
i
m
i
t
a
t
i
o
n
s,
”
F
ro
n
t
i
e
rs
i
n
Pl
a
n
t
S
c
i
e
n
c
e
,
v
o
l
.
1
5
,
2
0
2
4
,
d
o
i
:
1
0
.
3
3
8
9
/
f
p
l
s.
2
0
2
4
.
1
3
5
6
2
6
0
.
[
2
3
]
M
.
Ji
,
L.
Z
h
a
n
g
,
a
n
d
Q
.
W
u
,
“
A
u
t
o
mat
i
c
g
r
a
p
e
l
e
a
f
d
i
s
e
a
s
e
s
i
d
e
n
t
i
f
i
c
a
t
i
o
n
v
i
a
U
n
i
t
e
d
M
o
d
e
l
b
a
se
d
o
n
m
u
l
t
i
p
l
e
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
,
”
I
n
f
o
rm
a
t
i
o
n
P
ro
c
e
s
si
n
g
i
n
A
g
ri
c
u
l
t
u
re
,
v
o
l
.
7
,
n
o
.
3
,
p
p
.
4
1
8
–
4
2
6
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
n
p
a
.
2
0
1
9
.
1
0
.
0
0
3
.
[
2
4
]
A
.
K
a
m
i
l
a
r
i
s
a
n
d
F
.
X
.
P
r
e
n
a
f
e
t
a
-
B
o
l
d
ú
,
“
D
e
e
p
l
e
a
r
n
i
n
g
i
n
a
g
r
i
c
u
l
t
u
r
e
:
a
s
u
r
v
e
y
,
”
C
o
m
p
u
t
e
rs
a
n
d
E
l
e
c
t
r
o
n
i
c
s
i
n
A
g
ri
c
u
l
t
u
re
,
v
o
l
.
1
4
7
,
p
p
.
7
0
–
9
0
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
a
g
.
2
0
1
8
.
0
2
.
0
1
6
.
[
2
5
]
P
.
W
a
n
g
,
T
.
N
i
u
,
Y
.
M
a
o
,
Z
.
Zh
a
n
g
,
B
.
Li
u
,
a
n
d
D
.
H
e
,
“
I
d
e
n
t
i
f
i
c
a
t
i
o
n
o
f
a
p
p
l
e
l
e
a
f
d
i
s
e
a
s
e
s
b
y
i
m
p
r
o
v
e
d
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
s
w
i
t
h
a
n
a
t
t
e
n
t
i
o
n
me
c
h
a
n
i
sm
,
”
Fro
n
t
i
e
rs i
n
Pl
a
n
t
S
c
i
e
n
c
e
,
v
o
l
.
1
2
,
S
e
p
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
8
9
/
f
p
l
s.
2
0
2
1
.
7
2
3
2
9
4
.
[
2
6
]
Y
.
L
u
,
S
.
Y
i
,
N
.
Z
e
n
g
,
Y
.
L
i
u
,
a
n
d
Y
.
Zh
a
n
g
,
“
I
d
e
n
t
i
f
i
c
a
t
i
o
n
o
f
r
i
c
e
d
i
s
e
a
se
s
u
si
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
s,
”
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
4
,
No
.
1
,
Ma
r
ch
2
0
2
5
:
38
-
46
46
N
e
u
ro
c
o
m
p
u
t
i
n
g
,
v
o
l
.
2
6
7
,
p
p
.
3
7
8
–
3
8
4
,
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
n
e
u
c
o
m
.
2
0
1
7
.
0
6
.
0
2
3
.
[
2
7
]
V
.
K
.
S
h
r
i
v
a
st
a
v
a
,
M
.
K
.
P
r
a
d
h
a
n
,
S
.
M
i
n
z
,
a
n
d
M
.
P
.
T
h
a
k
u
r
,
“
R
i
c
e
p
l
a
n
t
d
i
s
e
a
s
e
c
l
a
ssi
f
i
c
a
t
i
o
n
u
si
n
g
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
o
f
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
Arc
h
i
v
e
s
o
f
t
h
e
Ph
o
t
o
g
r
a
m
m
e
t
ry,
Re
m
o
t
e
S
e
n
s
i
n
g
a
n
d
S
p
a
t
i
a
l
I
n
f
o
r
m
a
t
i
o
n
S
c
i
e
n
c
e
s
-
I
S
PR
S
Ar
c
h
i
v
e
s
,
v
o
l
.
4
2
,
n
o
.
3
/
W
6
,
p
p
.
6
3
1
–
6
3
5
,
2
0
1
9
,
d
o
i
:
1
0
.
5
1
9
4
/
i
s
p
r
s
-
a
r
c
h
i
v
e
s
-
X
LI
I
-
3
-
W6
-
6
3
1
-
2
0
1
9
.
[
2
8
]
C
.
K
.
R
a
i
a
n
d
R
.
P
a
h
u
j
a
,
“
D
e
t
e
c
t
i
o
n
a
n
d
se
g
m
e
n
t
a
t
i
o
n
o
f
r
i
c
e
d
i
sea
s
e
s
u
si
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
s,
”
S
N
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
4
,
n
o
.
5
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s4
2
9
7
9
-
023
-
0
2
0
1
4
-
6.
[
2
9
]
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
,
n
o
.
S
e
p
t
e
m
b
e
r
,
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
.
[
3
0
]
S
.
M
.
H
a
ssa
n
a
n
d
A
.
K
.
M
a
j
i
,
“
D
e
e
p
f
e
a
t
u
r
e
-
b
a
se
d
p
l
a
n
t
d
i
se
a
se
i
d
e
n
t
i
f
i
c
a
t
i
o
n
u
s
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
c
l
a
ss
i
f
i
e
r
,
”
I
n
n
o
v
a
t
i
o
n
s
i
n
S
y
s
t
e
m
s
a
n
d
S
o
f
t
w
a
r
e
E
n
g
i
n
e
e
r
i
n
g
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
3
3
4
-
0
2
2
-
0
0
5
1
3
-
y.
[
3
1
]
G
.
O
w
o
mu
g
i
sh
a
a
n
d
E.
M
w
e
b
a
z
e
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
f
o
r
p
l
a
n
t
d
i
s
e
a
s
e
i
n
c
i
d
e
n
c
e
a
n
d
se
v
e
r
i
t
y
me
a
su
r
e
me
n
t
s fr
o
m l
e
a
f
i
ma
g
e
s
,
”
i
n
Pro
c
e
e
d
i
n
g
s
-
2
0
1
6
1
5
t
h
I
EE
E
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
M
a
c
h
i
n
e
L
e
a
rn
i
n
g
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s,
I
C
ML
A
2
0
1
6
,
2
0
1
7
,
p
p
.
1
5
8
–
1
6
3
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
M
LA
.
2
0
1
6
.
1
2
6
.
[
3
2
]
S
.
X
i
a
n
g
,
Q
.
Li
a
n
g
,
W
.
S
u
n
,
D
.
Zh
a
n
g
,
a
n
d
Y
.
W
a
n
g
,
“
L
-
C
S
M
S
:
n
o
v
e
l
l
i
g
h
t
w
e
i
g
h
t
n
e
t
w
o
r
k
f
o
r
p
l
a
n
t
d
i
s
e
a
s
e
se
v
e
r
i
t
y
r
e
c
o
g
n
i
t
i
o
n
,
”
J
o
u
r
n
a
l
o
f
P
l
a
n
t
D
i
s
e
a
s
e
s a
n
d
Pr
o
t
e
c
t
i
o
n
,
v
o
l
.
1
2
8
,
n
o
.
2
,
p
p
.
5
5
7
–
5
6
9
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
0
7
/
s
4
1
3
4
8
-
0
2
0
-
0
0
4
2
3
-
w.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
S
h
r
a
d
h
a
J
o
shi
-
B
a
g
c
o
m
p
lete
d
h
e
r
M
.
Tec
h
.
i
n
El
e
c
tro
n
ics
Tec
h
n
o
l
o
g
y
fro
m
S
h
iv
a
ji
Un
i
v
e
rsity
,
Ko
l
h
a
p
u
r
.
S
h
e
is
p
u
rs
u
in
g
a
P
h
.
D.
fro
m
Ra
iso
n
i
Un
i
v
e
rsity
,
Am
ra
v
a
ti
.
S
h
e
h
a
s
a
tea
c
h
in
g
e
x
p
e
rien
c
e
o
f
1
8
y
e
a
rs.
P
re
se
n
tl
y
sh
e
is
w
o
rk
i
n
g
a
s
a
n
a
ss
istan
t
p
ro
fe
ss
o
r
a
t
N.K.
Orc
h
id
Co
ll
e
g
e
o
f
En
g
i
n
e
e
rin
g
a
n
d
Tec
h
n
o
lo
g
y
,
S
h
o
lap
u
r.
S
h
e
is
a
li
fe
m
e
m
b
e
r
o
f
th
e
In
d
ian
S
o
c
iety
f
o
r
Tec
h
n
ica
l
Ed
u
c
a
ti
o
n
.
He
r
a
re
a
s o
f
in
tere
st i
n
c
lu
d
e
e
m
b
e
d
d
e
d
sy
ste
m
d
e
sig
n
,
VLS
I
d
e
sig
n
,
I
o
T,
a
n
d
P
y
t
h
o
n
p
r
o
g
ra
m
m
in
g
.
S
h
e
h
a
s
a
tt
e
n
d
e
d
2
0
+
Wo
rk
sh
o
p
s
a
n
d
s
h
o
rt
-
term
train
in
g
p
ro
g
ra
m
s
o
rg
a
n
ize
d
b
y
AICTE
a
n
d
IS
T
E.
S
h
e
h
a
s
p
u
b
li
s
h
e
d
6
p
a
p
e
rs
i
n
I
n
tern
a
ti
o
n
a
l
Jo
u
rn
a
ls
a
n
d
o
n
e
P
a
ten
t.
S
h
e
w
a
s
a
wa
rd
e
d
Be
st
Tea
c
h
in
g
F
a
c
u
lt
y
M
e
m
b
e
r
a
t
NK
OCET,
S
o
lap
u
r
i
n
2
0
1
5
-
2
0
1
6
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
b
a
g
sh
ra
d
h
a
@g
m
a
il
.
c
o
m
.
Wa
n
i
V.
Pa
t
il
c
o
m
p
lete
d
h
e
r
Ba
c
h
e
lo
r
o
f
En
g
in
e
e
rin
g
in
E
lec
tro
n
ics
En
g
in
e
e
rin
g
fro
m
RTM
Un
iv
e
rsit
y
Na
g
p
u
r
,
M
a
ste
r
o
f
Tec
h
n
o
lo
g
y
i
n
VLS
I
fro
m
VN
IT,
Na
g
p
u
r,
a
n
d
P
h
.
D.
in
El
e
c
tro
n
ics
En
g
in
e
e
rin
g
fr
o
m
RTM
Na
g
p
u
r.
S
h
e
h
a
s
p
u
b
l
ish
e
d
5
2
re
se
a
rc
h
p
a
p
e
rs
i
n
re
p
u
ted
In
ter
n
a
ti
o
n
a
l
C
o
n
fe
re
n
c
e
a
n
d
Jo
u
rn
a
ls.
He
r
re
se
a
rc
h
i
n
tere
sts
a
re
d
i
g
it
a
l
ima
g
e
p
ro
c
e
ss
in
g
a
n
d
b
io
m
e
d
ica
l
e
n
g
i
n
e
e
rin
g
.
S
h
e
h
a
s
a
lso
p
u
b
l
ish
e
d
b
o
o
k
c
h
a
p
ters
a
n
d
b
o
o
k
s
re
late
d
to
ima
g
e
p
ro
c
e
ss
in
g
a
n
d
b
io
m
e
d
ica
l
e
n
g
i
n
e
e
rin
g
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
wa
n
ime
sh
ra
m
@g
m
a
il
.
c
o
m
.
S
h
r
ik
a
n
t
Ch
a
v
a
te
is
th
e
fa
c
u
lt
y
o
f
d
isc
ip
l
in
e
El
e
c
tro
n
ics
a
n
d
Te
lec
o
m
m
u
n
ica
ti
o
n
En
g
i
n
e
e
rin
g
a
t
G
.
H.
Ra
iso
n
i
Un
iv
e
rsity
,
Am
ra
v
a
ti
(M
a
h
a
ra
sh
tra)
4
4
4
7
0
1
,
I
n
d
ia.
He
re
c
e
iv
e
d
a
B.
E.
d
e
g
re
e
fro
m
S
G
BAU
,
A
m
ra
v
a
ti
in
2
0
0
9
.
He
re
c
e
iv
e
d
a
n
M
.
E.
d
e
g
re
e
fro
m
RG
P
V
Bh
o
p
a
l
in
2
0
1
3
.
He
se
rv
e
d
a
s
a
n
a
ss
istan
t
p
ro
fe
ss
o
r
a
t
G
HRCEM
Am
ra
v
a
ti
fro
m
De
c
2
0
1
1
to
Ju
n
e
2
0
1
8
.
He
se
rv
e
d
a
s
a
n
a
ss
istan
t
p
ro
fe
ss
o
r
a
n
d
He
a
d
(EXT
C)
a
t
G
.
H.
Ra
is
o
n
i
Un
i
v
e
rsity
,
Am
ra
v
a
ti
fro
m
Ju
n
e
2
0
1
8
to
Ju
n
e
2
0
2
4
.
He
h
a
s
a
u
t
h
o
re
d
/c
o
-
a
u
t
h
o
re
d
se
v
e
ra
l
re
se
a
rc
h
p
a
p
e
rs
in
v
a
ri
o
u
s
re
p
u
te
d
i
n
tern
a
ti
o
n
a
l
p
u
b
li
sh
e
rs
’
j
o
u
r
n
a
ls/co
n
fe
re
n
c
e
s.
G
ra
n
ted
with
a
p
a
ten
t
a
n
d
se
v
e
ra
l
c
o
p
y
rig
h
ts.
His
re
se
a
rc
h
in
tere
sts
in
c
l
u
d
e
v
i
d
e
o
sh
o
t
b
o
u
n
d
a
ry
d
e
tec
ti
o
n
,
e
m
b
e
d
d
e
d
sy
ste
m
s
,
a
n
d
a
rti
fi
c
ial
i
n
telli
g
e
n
c
e
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
S
h
rik
a
n
t.
c
h
a
v
a
te@
g
h
ru
a
.
e
d
u
.
in
.
Evaluation Warning : The document was created with Spire.PDF for Python.