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
.
4
,
A
u
g
u
s
t
20
25
,
p
p
.
3
7
5
9
~
3
7
6
8
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
4
.
pp
3
7
5
9
-
3
7
6
8
3759
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Deep
feature
repr
esenta
tion
for a
ut
o
ma
ted
pla
nt
sp
e
cies
cla
ss
ificatio
n f
ro
m lea
f
ima
g
es
Nik
hil
I
na
m
da
r
1
,
M
a
njuna
t
h M
a
na
g
uli
1
,
Ut
t
a
m
P
a
t
il
2
1
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
C
o
m
mu
n
i
c
a
t
i
o
s
E
n
g
i
n
e
e
r
i
n
g
,
K
LS G
o
g
t
e
I
n
st
i
t
u
t
e
o
f
T
e
c
h
n
o
l
o
g
y
B
e
l
a
g
a
v
i
a
n
d
A
f
f
i
l
i
a
t
e
d
t
o
V
i
sw
e
r
a
v
a
y
a
Te
c
h
n
o
l
o
g
i
c
a
l
U
n
i
v
e
r
s
i
t
y
,
B
e
l
a
g
a
v
i
,
I
n
d
i
a
2
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
E
n
g
i
n
e
e
r
i
n
g
,
Ja
i
n
C
o
l
l
e
g
e
o
f
En
g
i
n
e
e
r
i
n
g
,
B
e
l
a
g
a
v
i
a
n
d
A
f
f
i
l
i
a
t
e
d
t
o
V
i
sw
e
r
a
v
a
y
a
T
e
c
h
n
o
l
o
g
i
c
a
l
U
n
i
v
e
r
si
t
y
,
B
e
l
a
g
a
vi
,
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
Au
g
2
3
,
2
0
2
4
R
ev
is
ed
Ap
r
1
5
,
2
0
2
5
Acc
ep
ted
Ma
y
2
4
,
2
0
2
5
Au
to
m
a
ted
p
lan
t
sp
e
c
ies
c
las
sifica
ti
o
n
u
sin
g
lea
f
ima
g
e
s
h
o
ld
s
imm
e
n
se
p
o
ten
ti
a
l
fo
r
a
d
v
a
n
c
in
g
a
g
ricu
lt
u
ra
l
re
se
a
rc
h
,
b
io
d
i
v
e
rsity
c
o
n
se
rv
a
ti
o
n
,
a
n
d
e
c
o
lo
g
ica
l
m
o
n
it
o
ri
n
g
.
Th
is
stu
d
y
in
tr
o
d
u
c
e
s
a
n
o
v
e
l
a
p
p
ro
a
c
h
l
e
v
e
ra
g
in
g
d
e
e
p
fe
a
tu
re
re
p
re
se
n
tatio
n
to
a
c
h
iev
e
a
c
c
u
ra
te
a
n
d
e
fficie
n
t
c
la
ss
ifi
c
a
ti
o
n
b
a
se
d
o
n
lea
f
m
o
r
p
h
o
lo
g
y
.
Co
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
two
rk
s
(CNN
s),
in
c
lu
d
in
g
VG
G
1
6
,
Re
sN
e
t5
0
,
De
n
se
Ne
t1
,
In
c
e
p
ti
o
n
,
a
n
d
Xc
e
p
ti
o
n
,
a
re
e
m
p
lo
y
e
d
t
o
e
x
trac
t
h
i
g
h
-
lev
e
l
fe
a
tu
re
s
fro
m
lea
f
ima
g
e
s,
c
a
p
tu
rin
g
in
tri
c
a
te
p
a
tt
e
rn
s e
ss
e
n
ti
a
l
fo
r
sp
e
c
ies
d
iffe
re
n
ti
a
ti
o
n
.
T
o
m
a
n
a
g
e
t
h
e
e
x
ten
si
v
e
fe
a
tu
re
se
t
e
x
trac
ted
b
y
t
h
e
se
m
o
d
e
ls,
o
p
ti
m
iza
ti
o
n
tec
h
n
i
q
u
e
s
s
u
c
h
a
s
p
ri
n
c
ip
a
l
c
o
m
p
o
n
e
n
t
a
n
a
ly
sis
(P
CA),
v
a
r
ian
c
e
th
re
sh
o
l
d
in
g
,
a
n
d
re
c
u
rsiv
e
fe
a
tu
re
e
li
m
in
a
ti
o
n
(RF
E)
a
re
a
p
p
l
ied
.
Th
e
se
m
e
th
o
d
s
stre
a
m
li
n
e
th
e
fe
a
tu
re
se
t,
m
a
k
in
g
th
e
c
las
sifica
ti
o
n
p
ro
c
e
ss
m
o
re
e
fficie
n
t
.
T
h
e
o
p
ti
m
ize
d
fe
a
tu
re
s
a
re
th
e
n
trai
n
e
d
u
sin
g
c
las
sifiers
li
k
e
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
(S
VM)
,
k
-
n
e
a
re
st
n
e
ig
h
b
o
rs
(K
-
NN
),
d
e
c
isio
n
tree
s
(DT),
a
n
d
n
a
i
v
e
Ba
y
e
s
(NB),
a
c
h
ie
v
in
g
a
v
e
ra
g
e
a
c
c
u
ra
c
ie
s
o
f
9
8
.
6
%
,
9
6
.
6
%
,
9
9
.
6
%
,
a
n
d
9
9
.
7
%
,
re
sp
e
c
ti
v
e
ly
,
a
c
ro
ss
v
a
rio
u
s
c
r
o
ss
-
v
a
li
d
a
ti
o
n
m
e
th
o
d
s.
Ex
p
e
rime
n
tal
re
su
lt
s
o
n
b
e
n
c
h
m
a
rk
d
a
tas
e
ts
d
e
m
o
n
stra
te
th
e
e
ffe
c
ti
v
e
n
e
ss
o
f
th
is
a
p
p
ro
a
c
h
,
a
c
h
iev
i
n
g
sta
te
-
of
-
th
e
-
a
rt
p
e
rfo
rm
a
n
c
e
in
p
lan
t
sp
e
c
ies
c
las
sific
a
ti
o
n
.
T
h
is
wo
rk
u
n
d
e
rsc
o
re
s
th
e
p
o
te
n
ti
a
l
o
f
d
e
e
p
fe
a
tu
re
re
p
re
se
n
tatio
n
i
n
a
u
t
o
m
a
ted
p
la
n
t
sp
e
c
ies
c
las
sifica
ti
o
n
,
o
ffe
ri
n
g
v
a
lu
a
b
le
in
sig
h
ts
fo
r
a
p
p
li
c
a
ti
o
n
s
in
a
g
ricu
lt
u
re
,
e
c
o
lo
g
y
,
a
n
d
e
n
v
iro
n
m
e
n
tal
sc
ien
c
e
.
K
ey
w
o
r
d
s
:
C
o
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
Plan
t sp
ec
ies cla
s
s
if
icatio
n
Dee
p
lear
n
in
g
Featu
r
e
r
ep
r
esen
tatio
n
L
ea
f
im
ag
es
T
r
an
s
f
er
lear
n
i
n
g
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
:
Nik
h
il I
n
am
d
a
r
Dep
ar
tm
en
t o
f
E
lectr
o
n
ics an
d
C
o
m
m
u
n
icatio
n
E
n
g
in
ee
r
i
n
g
,
KL
S Go
g
te
I
n
s
titu
te
o
f
T
ec
h
n
o
lo
g
y
B
elag
av
i
an
d
Af
f
iliated
to
Vis
wer
av
ay
a
T
ec
h
n
o
lo
g
ical
Un
iv
er
s
ity
J
n
ash
o
d
h
a
C
am
p
u
s
Ud
y
a
m
b
h
ag
,
B
elag
av
i I
n
d
ia
E
m
ail: n
ik
h
il0
8
7
0
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Dete
ctin
g
p
lan
t
d
is
ea
s
es
th
r
o
u
g
h
im
a
g
e
an
aly
s
is
is
ess
en
ti
al
in
p
r
ec
is
io
n
a
g
r
icu
ltu
r
e
.
T
r
ad
itio
n
ally
,
d
is
ea
s
e
s
ev
er
ity
h
as
b
ee
n
a
s
s
es
s
ed
v
is
u
ally
b
y
ex
p
er
ts
[
1
]
.
Alth
o
u
g
h
th
e
u
s
e
o
f
d
i
g
ital
ca
m
er
as
an
d
ad
v
an
ce
m
e
n
ts
in
in
f
o
r
m
atio
n
tech
n
o
lo
g
y
h
as
in
cr
ea
s
ed
th
e
ad
o
p
tio
n
o
f
ex
p
e
r
t
s
y
s
tem
s
in
ag
r
icu
ltu
r
e,
en
h
an
cin
g
p
lan
t
p
r
o
d
u
ctio
n
[
2
]
,
th
ese
s
y
s
tem
s
s
till
f
ac
e
s
ig
n
i
f
ican
t c
h
allen
g
es.
Var
io
u
s
a
r
ti
f
i
cia
l
i
n
t
ell
ig
e
n
ce
(
A
I
)
tec
h
n
i
q
u
es
,
in
cl
u
d
i
n
g
K
-
n
ea
r
est
n
ei
g
h
b
o
r
s
(
K
-
NN
)
,
l
o
g
is
t
ic
r
e
g
r
ess
i
o
n
,
d
ec
is
io
n
tr
ee
s
,
s
u
p
p
o
r
t
v
ec
to
r
m
a
ch
in
es
(
SVM
)
[
3
]
,
an
d
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
r
al
n
etw
o
r
k
s
(
C
NN
)
,
a
r
e
ex
t
e
n
s
i
v
el
y
u
s
e
d
f
o
r
p
la
n
t
d
is
e
ase
d
ete
cti
o
n
an
d
class
i
f
ic
ati
o
n
.
T
h
es
e
m
et
h
o
d
s
o
f
t
en
in
v
o
l
v
e
i
m
a
g
e
p
r
e
p
r
o
c
ess
i
n
g
t
o
e
n
h
a
n
c
e
f
e
atu
r
e
e
x
t
r
ac
ti
o
n
.
K
-
NN
,
a
s
u
p
e
r
v
is
e
d
al
g
o
r
i
th
m
,
c
lass
i
f
ies
b
y
co
m
p
a
r
i
n
g
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
.
4
,
Au
g
u
s
t
20
25
:
3
7
5
9
-
3768
3760
s
im
il
ar
iti
es
b
e
twe
e
n
u
n
la
b
ele
d
a
n
d
la
b
el
e
d
d
a
ta
[
4
]
.
D
ec
is
io
n
t
r
ee
s
,
s
t
r
u
ct
u
r
e
d
li
k
e
a
f
l
o
wc
h
ar
t
,
m
ak
e
d
ec
is
io
n
s
b
as
ed
o
n
a
tt
r
i
b
u
tes
at
e
ac
h
n
o
d
e
b
u
t
ca
n
f
ac
e
c
h
all
en
g
es s
u
c
h
as
o
v
e
r
la
p
p
i
n
g
n
o
d
es
a
n
d
o
v
er
f
itt
in
g
[
5
]
.
S
h
a
r
i
f
e
t
a
l
.
[
6
]
I
n
t
r
o
d
u
c
e
d
a
t
e
c
h
n
i
q
u
e
f
o
r
d
e
t
e
c
t
i
n
g
d
i
s
e
ase
s
i
n
c
i
t
r
u
s
p
l
a
n
t
s
u
s
i
n
g
t
e
x
t
u
r
e
f
e
a
t
u
r
e
s
,
c
o
m
b
i
n
i
n
g
p
r
i
n
c
i
p
a
l
c
o
m
p
o
n
e
n
t
a
n
a
l
y
s
is
w
it
h
f
e
a
t
u
r
e
s
t
a
ti
s
t
ic
s
f
o
r
h
y
b
r
i
d
f
e
a
t
u
r
e
s
e
l
ec
t
i
o
n
.
P
a
t
il
a
n
d
K
u
m
a
r
[
7
]
d
e
v
e
l
o
p
e
d
a
s
y
s
t
e
m
t
o
d
e
t
e
ct
d
i
s
e
as
es
i
n
s
o
y
b
e
a
n
l
e
a
v
e
s
b
y
an
a
l
y
z
i
n
g
c
o
l
o
r
,
s
h
a
p
e
,
a
n
d
t
e
x
t
u
r
e
f
e
a
t
u
r
e
s
u
s
i
n
g
a
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
a
p
p
r
o
a
c
h
.
S
a
n
d
i
k
a
e
t
a
l
.
[
8
]
p
r
o
p
o
s
e
d
a
m
e
t
h
o
d
f
o
r
i
d
e
n
t
i
f
y
i
n
g
g
r
a
p
e
l
e
a
f
d
i
s
e
a
s
e
s
i
n
c
o
m
p
l
e
x
b
a
c
k
g
r
o
u
n
d
s
,
u
t
i
li
z
i
n
g
f
e
a
t
u
r
es
li
k
e
l
o
c
al
b
i
n
a
r
y
p
a
t
t
e
r
n
s
(
L
B
P
)
a
n
d
R
GB
co
l
o
r
s
t
a
ti
s
t
i
cs
,
a
n
d
a
p
p
l
y
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
m
s
s
u
c
h
as
s
u
p
p
o
r
t
v
e
c
t
o
r
m
a
c
h
i
n
e
s
(
SV
M
)
a
n
d
r
a
n
d
o
m
f
o
r
e
s
t
.
T
h
e
p
r
im
ar
y
r
esear
ch
p
r
o
b
lem
ad
d
r
ess
ed
in
th
is
s
tu
d
y
is
t
h
e
ac
cu
r
ate
an
d
ef
f
icien
t
clas
s
if
icatio
n
o
f
p
lan
t
s
p
ec
ies
u
s
in
g
leaf
im
ag
e
s
,
o
v
er
co
m
i
n
g
th
e
lim
itatio
n
s
o
f
tr
ad
itio
n
al
m
eth
o
d
s
th
at
r
el
y
o
n
m
an
u
al
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
co
n
v
en
ti
o
n
al
m
ac
h
in
e
lear
n
i
n
g
tec
h
n
iq
u
es.
Ou
r
u
n
iq
u
e
ap
p
r
o
ac
h
lev
er
ag
es
d
ee
p
f
ea
tu
r
e
r
ep
r
esen
tatio
n
b
y
in
teg
r
atin
g
m
u
ltip
le
ad
v
an
ce
d
C
NN
ar
ch
itectu
r
es
(
VGG1
6
,
R
esNet
5
0
,
Den
s
eNe
t1
2
1
,
I
n
ce
p
tio
n
,
an
d
Xce
p
tio
n
)
to
c
ap
tu
r
e
i
n
tr
icate
p
atter
n
s
ess
en
tial
f
o
r
s
p
ec
ies
d
if
f
e
r
en
tiatio
n
.
Ad
d
itio
n
all
y
,
we
em
p
lo
y
s
o
p
h
is
ticated
f
ea
t
u
r
e
o
p
tim
izatio
n
tec
h
n
iq
u
es
s
u
ch
as
p
r
in
cip
al
co
m
p
o
n
en
t
an
aly
s
is
(
PC
A
)
,
v
ar
ian
c
e
th
r
esh
o
ld
in
g
,
an
d
r
ec
u
r
s
iv
e
f
e
atu
r
e
elim
in
atio
n
(
R
FE)
t
o
s
tr
ea
m
lin
e
th
e
f
ea
tu
r
e
s
et,
r
e
d
u
c
in
g
co
m
p
u
tatio
n
al
co
m
p
lex
ity
an
d
im
p
r
o
v
in
g
ef
f
icien
cy
.
T
h
is
co
m
b
in
atio
n
o
f
d
ee
p
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
o
p
tim
izatio
n
,
alo
n
g
with
th
e
u
s
e
o
f
d
iv
er
s
e
class
if
ier
s
(
SVM,
K
-
NN,
d
ec
i
s
io
n
tr
ee
s
,
n
aiv
e
B
ay
es),
r
esu
lts
in
s
tate
-
of
-
th
e
-
ar
t
p
er
f
o
r
m
an
ce
with
ac
cu
r
ac
ies u
p
to
9
9
.
7
%.
2.
RE
L
AT
E
D
S
T
UDY
Dee
p
lear
n
in
g
ar
c
h
itectu
r
es
h
av
e
r
ec
en
tly
s
h
o
w
n
g
r
e
at
ef
f
ec
tiv
en
ess
in
task
s
li
k
e
o
b
ject
id
en
tific
atio
n
,
class
if
icatio
n
,
an
d
s
eg
m
e
n
tatio
n
,
with
C
NNs
b
ein
g
p
ar
ticu
lar
ly
p
r
o
m
in
e
n
t
[
9
]
.
Fo
r
ex
am
p
le,
Selv
ar
aj
et
a
l.
[
1
0
]
d
ev
elo
p
ed
a
d
ataset
o
f
b
an
an
a
p
lan
t
s
am
p
les
f
r
o
m
Af
r
ica
an
d
So
u
th
e
r
n
I
n
d
ia,
co
v
er
in
g
1
7
d
is
ea
s
ed
class
e
s
an
d
o
n
e
h
ea
lth
y
class
.
Usi
n
g
C
NN
a
r
ch
itectu
r
es
lik
e
R
esNet5
0
,
I
n
ce
p
tio
n
V
2
,
an
d
Mo
b
ileNetV1
,
th
ey
ac
h
iev
ed
a
9
0
%
ac
cu
r
ac
y
r
ate.
L
u
et
a
l.
[1
1]
ex
p
lo
r
es
v
ar
io
u
s
class
if
i
er
s
,
in
clu
d
in
g
SVM,
KNN,
an
d
R
an
d
o
m
Fo
r
est,
f
o
r
class
if
icatio
n
.
T
h
e
s
tu
d
y
[
1
2
]
u
n
d
er
s
co
r
es
th
e
ad
ap
tab
i
lity
o
f
C
NN
-
b
ased
f
ea
tu
r
e
ex
tr
ac
tio
n
a
n
d
th
e
p
o
te
n
tial
o
f
d
if
f
er
e
n
t
class
if
ier
s
in
ac
cu
r
ately
class
if
y
in
g
ag
r
icu
lt
u
r
al
d
ata.
B
ar
b
ed
o
[
1
3
]
ex
p
l
o
r
es
7
9
d
is
ea
s
e
d
ete
ctio
n
o
f
1
4
d
if
f
er
e
n
t
p
lan
t
s
p
ec
ies.
[
1
4
]
in
tr
o
d
u
ce
s
a
n
o
p
tim
al
f
ea
tu
r
e
s
et
f
o
r
ac
h
iev
in
g
h
ig
h
er
class
if
icatio
n
ac
cu
r
ac
y
in
ag
r
icu
ltu
r
al
cr
o
p
s
p
ec
ies
class
if
icatio
n
.
B
y
co
m
b
in
i
n
g
v
ar
io
u
s
f
ea
tu
r
es
an
d
d
atasets
,
th
e
s
tu
d
y
aim
s
to
f
u
r
th
er
o
p
tim
ize
cla
s
s
if
ica
tio
n
ac
cu
r
ac
y
an
d
ex
p
lo
r
e
d
if
f
er
e
n
t
f
ea
tu
r
e
co
m
b
in
atio
n
s
t
o
en
h
an
ce
p
er
f
o
r
m
an
ce
.
[
1
5
]
p
r
o
p
o
s
es
an
in
tellig
en
t
s
y
s
tem
f
o
r
r
ea
l
-
tim
e
id
en
tific
atio
n
o
f
I
n
d
ian
m
e
d
icin
al
h
er
b
s
b
ased
o
n
leaf
im
ag
es,
u
tili
zin
g
R
asp
b
er
r
y
Pi.
T
h
e
d
ev
elo
p
ed
m
ac
h
in
e
lear
n
in
g
m
o
d
els
ac
h
iev
e
im
p
r
ess
iv
e
ac
cu
r
ac
y
r
ates,
d
em
o
n
s
tr
atin
g
th
e
f
ea
s
i
b
ilit
y
o
f
u
s
in
g
lo
w
-
co
s
t
h
a
r
d
war
e
f
o
r
r
ea
l
-
wo
r
ld
ap
p
licatio
n
s
in
p
lan
t i
d
en
tific
a
tio
n
.
I
n
tr
o
d
u
cin
g
a
C
NN
b
as
ed
m
et
h
o
d
ca
lled
D
-
L
ea
f
f
o
r
leaf
class
if
icatio
n
,
[
1
6
]
co
m
p
a
r
es d
if
f
e
r
en
t CNN
m
o
d
els
b
ased
o
n
th
eir
f
ea
tu
r
e
ex
tr
ac
tio
n
ca
p
ab
ilit
ies.
T
h
e
D
-
L
ea
f
m
o
d
el
ac
h
ie
v
es
co
m
p
etitiv
e
test
in
g
ac
cu
r
ac
y
co
m
p
a
r
ed
to
p
r
e
-
tr
ain
ed
C
NN
m
o
d
els.
C
h
u
an
l
ei
et
a
l.
[
1
7
]
ac
h
iev
e
d
9
7
%
ac
cu
r
ac
y
in
th
ei
r
ex
p
er
im
en
ts
o
n
wh
ea
t
p
lan
ts
,
u
s
in
g
a
d
ataset
wi
th
s
ix
d
is
ea
s
ed
class
e
s
an
d
o
n
e
h
ea
lth
y
class
.
S
im
ilar
ly
,
Sin
g
h
et
a
l.
[
1
8
]
d
e
v
elo
p
e
d
a
s
y
s
tem
f
o
r
d
etec
tin
g
tea
lea
f
d
i
s
ea
s
es
with
a
m
o
d
if
ied
d
ee
p
c
o
n
v
o
lu
ti
o
n
al
n
eu
r
a
l
n
etwo
r
k
,
ac
h
iev
i
n
g
an
av
er
a
g
e
ac
cu
r
ac
y
o
f
9
2
%.
C
h
ak
r
a
b
o
r
ty
et
a
l.
[
1
9
]
co
n
d
u
cted
ex
p
er
im
en
ts
o
n
7
9
d
if
f
er
en
t
d
is
ea
s
es
ac
r
o
s
s
1
4
p
lan
t
ty
p
es
u
s
in
g
th
e
Go
o
g
L
eN
et
ar
ch
itectu
r
e,
with
ac
cu
r
ac
y
s
co
r
es
co
n
s
is
ten
tly
ab
o
v
e
7
5
%.
Kr
izh
e
v
s
k
y
et
a
l.
[
2
0
]
ex
p
l
o
r
ed
v
ar
io
u
s
C
NN
a
r
ch
itectu
r
es,
ac
h
ie
v
in
g
u
p
to
9
9
%
ac
cu
r
ac
y
,
with
VGG
r
ea
ch
in
g
8
1
%
ac
r
o
s
s
m
u
ltip
le
p
lan
t
ty
p
es.
Gee
th
a
r
am
an
i
an
d
Pan
d
ian
[
3
]
wo
r
k
ed
with
a
d
ataset
co
n
tain
in
g
3
8
class
es
f
r
o
m
1
4
p
lan
t
ty
p
es,
attain
in
g
a
9
6
%
ac
cu
r
ac
y
r
ate.
T
r
a
d
itio
n
a
l
m
ac
h
in
e
lear
n
in
g
ap
p
r
o
ac
h
es,
th
o
u
g
h
ef
f
ec
tiv
e
in
p
lan
t
d
is
ea
s
e
id
en
tific
atio
n
,
ar
e
lim
ited
b
y
th
e
s
eq
u
en
ti
al
n
atu
r
e
o
f
im
a
g
e
s
eg
m
en
tatio
n
[
2
1
]
,
f
ea
tu
r
e
e
x
tr
ac
tio
n
[
2
2
]
,
an
d
p
atter
n
r
ec
o
g
n
itio
n
[
2
3
]
.
W
h
ile
b
asic
C
NN
m
o
d
els
lik
e
Alex
Net,
VGGN
et,
Go
o
g
L
e
Net,
Den
s
eNe
t,
an
d
R
e
s
Net
h
av
e
b
ee
n
ex
ten
s
iv
ely
u
tili
ze
d
f
o
r
p
lan
t
d
is
ea
s
e
class
if
icatio
n
,
th
ey
co
m
e
with
d
r
aw
b
ac
k
s
s
u
ch
as
h
ig
h
p
a
r
am
eter
d
e
m
an
d
s
a
n
d
s
lo
w
c
o
m
p
u
tatio
n
s
p
ee
d
s
.
Desp
ite
th
eir
s
tr
en
g
th
in
ca
p
tu
r
in
g
b
o
t
h
h
ig
h
-
a
n
d
lo
w
-
le
v
el
f
ea
tu
r
es,
th
ese
m
o
d
els
o
f
ten
s
tr
u
g
g
le
with
co
n
s
is
ten
tly
d
escr
ib
in
g
lo
ca
l
s
p
atial
ch
ar
ac
ter
is
tics
[
2
4
]
.
[
2
4
]
I
m
p
lem
en
te
d
r
esid
u
al
lear
n
in
g
f
r
am
ewo
r
k
to
ea
s
e
th
e
tr
ain
in
g
m
ec
h
an
is
m
t
h
er
e
is
2
8
% r
elativ
e
im
p
r
o
v
e
m
en
t in
C
OC
O
o
b
ject
d
etec
tio
n
d
ataset
.
3.
M
E
T
H
O
DO
L
O
G
Y
B
lo
ck
s
ch
em
atic
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
is
s
h
o
wn
in
Fig
u
r
e
1
.
I
t
co
n
s
is
ts
o
f
f
o
u
r
s
tag
e
s
,
n
am
ely
,
p
r
e
-
p
r
o
ce
s
s
in
g
,
d
ee
p
f
ea
tu
r
e
ex
tr
ac
tio
n
,
f
ea
tu
r
e
o
p
tim
izati
o
n
a
n
d
class
if
icatio
n
.
T
h
e
r
e
s
ea
r
ch
f
o
c
u
s
es
o
n
class
if
icatio
n
o
f
p
lan
t
leaf
im
ag
es,
m
an
y
r
esear
ch
e
r
s
h
av
e
wo
r
k
ed
o
n
id
en
tific
atio
n
o
f
l
ea
f
d
is
ea
s
es
r
ath
er
th
an
o
n
id
e
n
tific
atio
n
lea
f
ty
p
e
s
s
p
ec
if
ically
g
r
o
win
g
in
th
is
p
ar
t
o
f
th
e
co
u
n
tr
y
an
d
also
p
r
o
v
ed
to
h
av
e
r
ich
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
Dee
p
fea
tu
r
e
r
ep
r
esen
ta
tio
n
f
o
r
a
u
to
ma
ted
p
la
n
t sp
ec
ies cla
s
s
ifica
tio
n
fr
o
m
lea
f ima
g
es
(
N
ikh
il I
n
a
md
a
r
)
3761
b
u
s
in
ess
v
alu
e.
T
h
er
e
ex
is
t
m
an
y
v
ar
ieties
o
f
p
lan
ts
;
th
e
ar
ticle
p
r
o
p
o
s
es
1
4
p
lan
t
leaf
class
if
icatio
n
m
eth
o
d
o
l
o
g
y
u
s
in
g
d
ee
p
f
ea
t
u
r
es.
Data
s
ets o
f
1
4
p
lan
ts
ar
e
co
llected
f
r
o
m
d
i
f
f
er
en
t r
esear
c
h
wo
r
k
cited
in
th
e
liter
atu
r
e
s
u
r
v
ey
s
ec
tio
n
.
C
o
m
m
o
n
ly
an
d
wid
ely
u
s
ed
C
NN
ar
e
u
s
ed
f
o
r
d
ee
p
f
ea
tu
r
e
ex
tr
ac
tio
n
,
n
am
ely
,
VGG1
6
,
R
esn
et5
0
,
I
n
ce
p
tio
n
V3
,
Xce
p
tio
n
an
d
Den
s
en
et1
2
1
.
Mo
d
els
s
u
ch
as
VGG1
6
,
p
r
e
-
tr
ain
ed
o
n
lar
g
e
d
atasets
lik
e
I
m
ag
eNe
t,
o
f
f
er
a
tr
an
s
f
er
ab
le
s
et
o
f
f
ea
tu
r
es
t
h
at
ca
p
tu
r
e
r
ich
v
is
u
al
in
f
o
r
m
atio
n
f
r
o
m
im
ag
es.
R
esNet
en
ab
les
ef
f
ec
tiv
e
g
r
a
d
ien
t
f
lo
w
a
n
d
f
ac
ilit
ates
th
e
lear
n
in
g
o
f
h
ig
h
l
y
ab
s
tr
ac
t
f
ea
tu
r
es
th
r
o
u
g
h
o
u
t
th
e
n
etwo
r
k
lay
er
s
.
I
n
ce
p
tio
n
V3
a
n
d
its
v
ar
ian
ts
lev
er
ag
e
in
ce
p
t
io
n
m
o
d
u
les,
wh
ich
u
s
e
m
u
lti
p
le
co
n
v
o
l
u
tio
n
s
o
f
d
if
f
er
en
t
k
er
n
el
s
izes
with
in
t
h
e
s
am
e
lay
er
.
Xce
p
tio
n
,
i
n
s
p
ir
ed
b
y
I
n
ce
p
tio
n
ar
c
h
itectu
r
e,
r
ep
lace
s
tr
ad
itio
n
al
co
n
v
o
l
u
tio
n
s
with
d
ep
t
h
wis
e
s
ep
ar
ab
le
co
n
v
o
lu
tio
n
s
.
T
h
is
m
o
d
if
icatio
n
d
ec
o
u
p
les
s
p
atial
an
d
ch
a
n
n
el
-
wis
e
co
r
r
elatio
n
s
in
f
ea
t
u
r
e
m
a
p
s
,
lead
in
g
to
im
p
r
o
v
ed
f
ea
t
u
r
e
r
e
p
r
esen
tatio
n
with
f
ewe
r
p
ar
am
eter
s
.
Fig
u
r
e
2
s
h
o
ws
th
e
u
s
e
o
f
VGG1
6
m
o
d
el
f
o
r
d
ee
p
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
its
r
e
lev
an
t
a
p
p
r
o
x
im
ate
b
r
ea
k
d
o
wn
o
f
th
e
n
u
m
b
er
o
f
f
ea
tu
r
es e
x
tr
ac
ted
at
ea
ch
lay
er
:
a.
I
n
p
u
t la
y
e
r
:
th
is
lay
er
d
o
esn
'
t p
r
o
d
u
ce
f
ea
tu
r
es
d
ir
ec
tly
,
b
u
t it
ac
ce
p
ts
in
p
u
t im
ag
es o
f
s
ize
(
2
2
4
,
2
2
4
,
3
)
.
b.
C
o
n
v
o
lu
tio
n
al
lay
e
r
s
: V
GG1
6
h
as a
to
tal
o
f
1
3
co
n
v
o
l
u
tio
n
a
l la
y
er
s
(
in
clu
d
in
g
p
o
o
lin
g
lay
er
s
)
.
c.
E
ac
h
co
n
v
o
lu
tio
n
al
lay
er
ty
p
ically
o
u
tp
u
ts
f
ea
tu
r
e
m
ap
s
o
f
v
ar
y
in
g
s
izes,
g
r
a
d
u
ally
r
ed
u
cin
g
s
p
atial
d
im
en
s
io
n
s
wh
ile
in
cr
ea
s
in
g
d
ep
th
(
n
u
m
b
er
o
f
f
ilter
s
)
.
d.
Fu
lly
co
n
n
ec
ted
lay
e
r
s
:
af
ter
f
latten
in
g
,
VGG1
6
in
clu
d
es
th
r
ee
f
u
lly
co
n
n
ec
ted
lay
e
r
s
wit
h
d
ec
r
ea
s
in
g
n
u
m
b
er
s
o
f
n
eu
r
o
n
s
:
4
0
9
6
,
4
0
9
6
,
an
d
1
0
0
0
(
f
o
r
I
m
ag
eNe
t
'
s
1
0
0
0
class
es).
2
5
,
0
8
8
r
e
p
r
e
s
en
t
th
e
f
ea
tu
r
e
v
ec
to
r
e
x
tr
ac
ted
f
r
o
m
th
e
last
co
n
v
o
lu
tio
n
al
lay
e
r
b
e
f
o
r
e
f
ee
d
i
n
g
i
n
to
f
u
lly
c
o
n
n
e
cted
lay
er
s
o
r
class
if
icatio
n
.
Hav
in
g
v
ar
io
u
s
C
NN
m
o
d
els
av
ail
ab
le
in
th
e
liter
atu
r
e,
an
ex
p
e
r
im
en
t
wa
s
co
n
d
u
cted
to
an
aly
ze
th
e
b
eh
av
io
r
o
f
d
if
f
er
en
t
C
NN
m
o
d
els
an
d
th
e
n
u
m
b
er
o
f
f
ea
tu
r
es
th
ey
ex
t
r
ac
t.
T
ab
le
1
p
r
o
v
i
d
es
d
etailed
p
er
f
o
r
m
an
ce
m
etr
ics f
o
r
th
ese
m
o
d
els b
o
th
with
an
d
with
o
u
t f
ea
tu
r
e
o
p
tim
izatio
n
.
Fi
g
u
r
e
1
.
B
lo
ck
s
ch
em
atic
o
f
t
h
e
p
r
o
p
o
s
ed
s
tu
d
y
3.
1
.
P
rincipa
l c
o
m
po
nent
a
na
ly
s
is
(
P
CA)
I
t
is
a
m
eth
o
d
u
s
ed
f
o
r
d
im
en
s
io
n
ality
r
e
d
u
ctio
n
,
wid
ely
ap
p
lied
in
f
ield
s
lik
e
im
ag
e
p
r
o
ce
s
s
in
g
,
d
ata
v
is
u
aliza
tio
n
,
an
d
m
ac
h
in
e
l
ea
r
n
in
g
.
Ma
th
e
m
atica
lly
,
PC
A
b
eg
in
s
b
y
s
tan
d
ar
d
izin
g
t
h
e
d
ata
to
en
s
u
r
e
u
n
if
o
r
m
ity
ac
r
o
s
s
v
ar
iab
les.
T
h
en
,
it c
o
m
p
u
tes th
e
co
v
ar
ian
c
e
m
atr
ix
o
f
th
e
s
tan
d
ar
d
ized
d
ata:
=
1
(
−
1
)
(
)
(
1
)
Her
e,
is
th
e
×
m
atr
ix
o
f
th
e
s
ta
n
d
ar
d
ized
d
ata.
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
.
4
,
Au
g
u
s
t
20
25
:
3
7
5
9
-
3768
3762
P
C
A
t
h
e
n
d
e
r
i
v
es
t
h
e
ei
g
e
n
v
ec
t
o
r
s
1
,
2
,
…
,
a
n
d
t
h
e
i
r
c
o
r
r
e
s
p
o
n
d
i
n
g
e
i
g
e
n
v
a
l
u
e
s
λ
1
,
λ
2
,
…,
λ
p
a
r
r
a
n
g
e
d
i
n
d
e
s
c
e
n
d
i
n
g
o
r
d
e
r
.
T
h
e
s
e
r
e
p
r
e
s
e
n
t
t
h
e
d
i
r
e
c
ti
o
n
s
o
f
m
a
x
i
m
u
m
v
a
r
i
a
n
c
e
a
n
d
t
h
e
i
r
m
a
g
n
i
t
u
d
e
s
,
r
e
s
p
e
c
t
i
v
el
y
.
Fig
u
r
e
2
.
Dee
p
f
ea
tu
r
e
e
x
tr
ac
ti
o
n
m
eth
o
d
o
lo
g
y
T
ab
le
1
.
Nu
m
b
er
o
f
f
ea
tu
r
es e
x
tr
ac
ted
f
r
o
m
d
if
f
er
en
t CNN m
o
d
els co
n
s
id
er
ed
V
G
G
1
6
R
ESN
ET
5
0
I
N
C
EPT
I
O
N
V
3
D
EN
S
EN
ET1
2
1
X
C
EPTI
O
N
O
r
i
g
i
n
a
l
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s
2
5
0
8
8
1
0
0
3
5
2
1
3
1
0
7
2
5
0
1
7
6
2
0
4
8
0
0
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
P
C
A
1
7
9
1
8
5
1
8
3
1
7
4
1
7
2
A
c
c
u
r
a
c
y
a
f
t
e
r
P
C
A
:
0
.
7
8
5
7
0
.
8
9
2
8
0
.
8
9
2
8
0
.
8
7
5
0
.
9
2
8
5
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
v
a
r
i
a
n
c
e
t
h
r
e
s
h
o
l
d
i
n
g
2
1
0
1
6
5
9
9
0
0
7
0
4
8
0
2
1
1
8
0
2
9
0
8
3
A
c
c
u
r
a
c
y
a
f
t
e
r
v
a
r
i
a
n
c
e
t
h
r
e
s
h
o
l
d
i
n
g
0
.
8
9
2
8
0
.
9
1
0
7
0
.
8
2
1
4
0
.
9
2
8
5
0
.
9
1
0
7
N
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
p
r
e
l
i
m
i
n
a
r
y
P
C
A
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
R
F
E
50
50
50
50
50
A
c
c
u
r
a
c
y
a
f
t
e
r
R
F
E:
0
.
8
9
2
8
0
.
8
9
2
8
0
.
8
5
7
1
0
.
9
1
0
7
0
.
9
2
8
5
A
c
c
u
r
a
c
y
w
i
t
h
o
u
t
o
p
t
i
mi
z
a
t
i
o
n
:
0
.
9
1
0
7
0
.
9
4
6
4
0
.
8
3
9
2
0
.
8
9
2
8
0
.
9
2
8
5
3.
1
.
1
.
Va
ria
nce
t
hresh
o
ldi
ng
Var
ian
ce
th
r
esh
o
ld
i
n
g
is
em
p
lo
y
ed
as
a
f
ea
tu
r
e
s
elec
tio
n
te
ch
n
iq
u
e
to
en
h
an
ce
t
h
e
p
e
r
f
o
r
m
an
ce
o
f
th
e
r
an
d
o
m
f
o
r
est
class
if
ier
b
y
elim
in
atin
g
f
ea
tu
r
es
with
lo
w
v
ar
ian
ce
th
at
ar
e
d
ee
m
ed
l
ess
in
f
o
r
m
ativ
e
f
o
r
class
if
icatio
n
task
s
.
T
h
e
m
et
h
o
d
h
i
n
g
es
o
n
th
e
p
r
em
is
e
th
at
f
ea
tu
r
es
ex
h
ib
itin
g
m
in
i
m
al
v
ar
ian
ce
ac
r
o
s
s
s
am
p
les
ar
e
less
lik
ely
to
co
n
tr
ib
u
te
m
ea
n
in
g
f
u
lly
to
d
i
s
tin
g
u
is
h
in
g
b
etwe
en
d
if
f
e
r
e
n
t
class
es.
Fo
r
ea
ch
f
ea
tu
r
e
f
in
th
e
d
ataset,
th
e
v
ar
ian
ce
2
is
co
m
p
u
ted
as th
e
av
er
a
g
e
s
q
u
ar
ed
d
ev
iatio
n
f
r
o
m
th
e
m
ea
n
:
2
=
1
∑
(
,
=
1
−
̅
)
2
(
2
)
Her
e,
,
r
ep
r
esen
ts
th
e
v
alu
e
o
f
f
ea
tu
r
e
in
s
am
p
le
,
(
̅
)
d
en
o
tes
th
e
m
ea
n
o
f
f
ea
t
u
r
e
f
ac
r
o
s
s
all
s
am
p
les,
an
d
n
s
ig
n
if
ies th
e
t
o
tal
n
u
m
b
er
o
f
s
am
p
les.
3.
1
.
2
.
Rec
urs
iv
e
f
ea
t
ure
elimin
a
t
io
n
(
RF
E
)
R
FE
i
s
a
f
ea
tu
r
e
s
e
lectio
n
t
ec
h
n
iq
u
e
th
at
s
y
s
tem
atica
lly
r
em
o
v
es
attr
ib
u
tes
to
im
p
r
o
v
e
m
o
d
el
p
er
f
o
r
m
an
ce
an
d
in
ter
p
r
etab
ilit
y
.
I
n
R
FE,
a
m
ac
h
in
e
lea
r
n
in
g
m
o
d
el
is
tr
ain
ed
o
n
th
e
d
ata
s
et,
an
d
f
ea
tu
r
es
ar
e
r
ec
u
r
s
iv
ely
p
r
u
n
ed
b
ased
o
n
t
h
eir
im
p
o
r
tan
ce
u
n
til
th
e
o
p
ti
m
al
s
u
b
s
et
is
d
eter
m
in
ed
.
T
h
e
p
r
o
ce
s
s
b
eg
in
s
b
y
f
itti
n
g
a
m
ac
h
in
e
lear
n
in
g
m
o
d
el
(
in
o
u
r
ca
s
e,
a
r
an
d
o
m
f
o
r
est
clas
s
if
ier
)
to
th
e
d
ataset
w
ith
an
in
itial
s
et
o
f
f
ea
tu
r
es,
wh
ich
m
a
y
b
e
r
ed
u
c
ed
u
s
in
g
p
r
in
cip
al
c
o
m
p
o
n
en
t
an
aly
s
is
(
PC
A)
to
m
an
ag
e
h
ig
h
-
d
im
en
s
io
n
al
d
ata
ef
f
ec
tiv
ely
:
=
(
,
=
100
)
Her
e,
r
ep
r
esen
ts
th
e
tr
ain
i
n
g
d
ata,
an
d
d
en
o
tes
th
e
PC
A
-
tr
an
s
f
o
r
m
ed
f
ea
t
u
r
e
s
et
with
r
ed
u
ce
d
d
im
en
s
io
n
s
.
3.
2
.
O
pti
m
ized
f
ea
t
ure
s
elec
t
io
n
B
u
ild
in
g
u
p
o
n
th
e
m
ath
em
atic
al
f
o
u
n
d
atio
n
o
f
o
p
tim
iz
atio
n
tech
n
iq
u
es d
is
cu
s
s
ed
ea
r
lier
,
th
is
s
ec
tio
n
o
u
tlin
es
th
e
s
p
ec
if
ic
ap
p
r
o
ac
h
u
s
ed
f
o
r
f
ea
tu
r
e
o
p
tim
izatio
n
.
T
h
e
c
h
o
s
en
m
eth
o
d
o
l
o
g
y
is
tailo
r
ed
to
e
n
h
an
ce
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
Dee
p
fea
tu
r
e
r
ep
r
esen
ta
tio
n
f
o
r
a
u
to
ma
ted
p
la
n
t sp
ec
ies cla
s
s
ifica
tio
n
fr
o
m
lea
f ima
g
es
(
N
ikh
il I
n
a
md
a
r
)
3763
th
e
s
elec
tio
n
o
f
r
elev
a
n
t
f
ea
tu
r
es.
E
ac
h
s
tep
is
d
esig
n
ed
to
en
s
u
r
e
im
p
r
o
v
e
d
m
o
d
el
p
er
f
o
r
m
an
ce
an
d
co
m
p
u
tatio
n
al
e
f
f
icien
cy
.
3.
3
.
P
rincipa
l c
o
m
po
nent
a
na
ly
s
is
(
P
CA)
Prin
cip
al
co
m
p
o
n
e
n
t
an
aly
s
is
(
PC
A)
is
em
p
lo
y
ed
to
r
ed
u
ce
th
e
d
im
en
s
io
n
ality
o
f
th
e
ex
tr
ac
ted
d
ee
p
f
ea
tu
r
es
wh
ile
p
r
eser
v
in
g
9
5
%
o
f
th
eir
o
r
ig
in
al
v
a
r
ian
ce
b
y
s
ettin
g
to
0
.
9
5
.
T
h
is
d
im
en
s
io
n
ality
r
ed
u
ctio
n
p
la
y
s
a
k
ey
r
o
le
in
ad
d
r
ess
in
g
th
e
cu
r
s
e
o
f
d
im
en
s
io
n
ality
,
wh
ich
ca
n
n
eg
ati
v
ely
im
p
ac
t
m
o
d
el
ac
cu
r
ac
y
an
d
ef
f
icien
c
y
.
A
d
d
itio
n
ally
,
it
s
ig
n
if
ican
tly
d
ec
r
ea
s
es
co
m
p
u
tatio
n
al
lo
a
d
,
en
a
b
lin
g
f
aster
p
r
o
ce
s
s
in
g
with
o
u
t
a
n
o
tab
le
lo
s
s
in
m
o
d
el
im
p
lem
en
tatio
n
.
PC
A
i
s
ap
p
lied
to
r
ed
u
ce
th
e
d
im
en
s
io
n
ality
o
f
th
e
ex
tr
ac
ted
d
ee
p
f
ea
tu
r
es
w
h
ile
r
etain
in
g
9
5
%
o
f
th
ei
r
v
a
r
ian
ce
(
=
0
.
9
5
)
.
T
h
is
r
ed
u
ctio
n
h
elp
s
m
itig
ate
th
e
c
u
r
s
e
o
f
d
im
en
s
io
n
a
lity
an
d
s
p
ee
d
s
u
p
co
m
p
u
tatio
n
with
o
u
t
s
ig
n
if
ican
tly
s
ac
r
if
icin
g
m
o
d
el
p
er
f
o
r
m
an
ce
.
Af
ter
PC
A
tr
an
s
f
o
r
m
atio
n
(
an
d
)
,
a
r
an
d
o
m
f
o
r
est
class
if
ier
(
)
is
tr
ain
ed
an
d
e
v
alu
ated
o
n
th
e
r
ed
u
ce
d
f
ea
tu
r
e
s
et
to
ass
ess
it
s
class
if
icatio
n
ac
cu
r
ac
y
(
)
.
Settin
g
PC
A
(
=
0
.
5
5
)
to
r
etain
5
5
%
o
f
v
ar
ian
ce
is
a
s
tr
ateg
ic
c
h
o
ice
th
at
b
alan
ce
s
d
im
e
n
s
io
n
ality
r
ed
u
ctio
n
with
th
e
p
r
eser
v
ati
o
n
o
f
ess
en
tial
in
f
o
r
m
atio
n
.
B
y
r
etain
in
g
5
5
%
o
f
th
e
v
ar
i
an
ce
,
th
e
n
u
m
b
e
r
o
f
f
ea
tu
r
es
is
s
ig
n
if
ican
tly
r
ed
u
c
ed
,
wh
ich
h
elp
s
allev
iate
th
e
cu
r
s
e
o
f
d
im
en
s
io
n
ality
an
d
m
itig
ate
o
v
er
f
itti
n
g
.
T
h
is
r
ed
u
ctio
n
s
im
p
lifie
s
th
e
m
o
d
el
an
d
d
ec
r
ea
s
es
co
m
p
u
ta
tio
n
al
co
m
p
lex
ity
,
wh
ile
s
till
m
ain
tain
in
g
e
n
o
u
g
h
v
ar
ian
ce
to
en
s
u
r
e
th
at
c
r
u
cia
l
in
f
o
r
m
ati
o
n
is
p
r
eser
v
ed
f
o
r
ac
cu
r
ate
p
r
ed
ictio
n
s
.
T
h
is
ap
p
r
o
ac
h
ef
f
ec
tiv
ely
ad
d
r
ess
es
th
e
tr
ad
e
-
o
f
f
b
etwe
en
r
ed
u
cin
g
f
ea
tu
r
e
s
p
ac
e
a
n
d
r
etain
in
g
s
ig
n
if
ican
t
d
ata
ch
ar
ac
ter
is
tics
,
th
u
s
o
p
tim
izin
g
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
m
o
d
el
.
Als
o
,
T
ab
le
2
g
i
v
es
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
m
o
d
el
f
o
r
d
if
f
er
e
n
t
p
er
ce
n
tag
e
o
f
v
a
r
ian
ce
.
3.
3
.
1
.
Va
ria
nce
t
hresh
o
ldi
ng
Var
ian
ce
th
r
esh
o
ld
in
g
(
ℎ
ℎ
)
is
u
s
ed
to
r
em
o
v
e
f
ea
t
u
r
es
with
lo
w
v
ar
ian
ce
.
T
h
e
th
r
esh
o
ld
is
s
et
d
y
n
am
ically
b
ased
o
n
th
e
v
ar
ian
ce
o
f
ea
ch
f
ea
tu
r
e
(
ℎ
ℎ
=
(
0
.
8
∗
(
1
−
0
.
8
)
)
)
.
T
h
is
tech
n
iq
u
e
is
b
en
e
f
icial
f
o
r
eli
m
in
atin
g
f
ea
tu
r
es
th
at
d
o
n
o
t
v
ar
y
m
u
ch
with
in
th
e
d
ataset,
p
o
ten
tially
r
e
d
u
cin
g
n
o
is
e
an
d
im
p
r
o
v
i
n
g
m
o
d
el
r
o
b
u
s
tn
ess
.
T
h
e
r
an
d
o
m
f
o
r
est
class
if
i
er
(
_
)
is
tr
ain
ed
an
d
ev
alu
ate
d
o
n
th
e
s
elec
ted
f
ea
tu
r
es
(
_
_
an
d
_
_
)
,
an
d
its
ac
cu
r
ac
y
(
_
)
is
co
m
p
u
ted
to
co
m
p
ar
e
with
PC
A.
T
h
e
u
s
e
o
f
ℎ
ℎ
(
ℎ
ℎ
=
(
4
∗
(
1
−
0
.
8
)
)
)
with
a
th
r
esh
o
ld
o
f
0
.
8
(
o
r
4
×
0
.
2
)
is
d
esig
n
ed
to
f
ilter
o
u
t
f
ea
t
u
r
es
with
lo
w
v
ar
ian
ce
,
wh
ich
a
r
e
less
lik
ely
to
co
n
tr
ib
u
te
m
ea
n
in
g
f
u
l
in
f
o
r
m
atio
n
to
th
e
m
o
d
el.
T
h
e
th
r
esh
o
ld
v
alu
e
o
f
0
.
8
is
ch
o
s
en
to
r
em
o
v
e
f
ea
tu
r
es
th
at
h
av
e
v
er
y
lo
w
v
ar
i
an
ce
—
s
p
ec
if
ically
,
th
o
s
e
with
v
ar
ian
ce
less
th
an
2
0
%
o
f
th
e
o
v
e
r
all
v
ar
ian
ce
.
T
h
is
ap
p
r
o
ac
h
en
h
a
n
ce
s
m
o
d
el
ef
f
icien
cy
b
y
f
o
cu
s
in
g
o
n
m
o
r
e
in
f
o
r
m
ativ
e
f
ea
tu
r
es
w
h
ile
d
is
ca
r
d
in
g
t
h
o
s
e
th
at
co
n
t
r
ib
u
te
litt
le
to
p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
.
T
ab
le
3
p
r
o
v
id
es th
e
v
ar
iatio
n
o
f
th
e
m
o
d
el
p
e
r
f
o
r
m
an
ce
f
o
r
d
if
f
er
en
t
v
ar
ian
ce
th
r
esh
o
ld
v
a
lu
es.
T
ab
le
2
.
Per
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
f
o
r
d
if
f
e
r
en
t v
a
r
ian
ce
d
u
r
in
g
o
p
tim
izatio
n
V
a
r
i
a
n
c
e
0
.
9
5
0
.
7
5
0
.
5
5
0
.
4
5
0
.
3
5
O
r
i
g
i
n
a
l
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s:
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
P
C
A
:
9
0
1
7
9
90
43
27
15
A
c
c
u
r
a
c
y
a
f
t
e
r
P
C
A
:
0
.
8
2
1
4
2
8
5
7
1
4
2
8
5
7
1
4
0
.
7
8
0
.
8
2
1
4
0
.
8
5
0
.
8
2
0
.
8
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
V
a
r
i
a
n
c
e
T
h
r
e
s
h
o
l
d
i
n
g
:
2
1
0
1
6
2
1
0
1
6
2
1
0
1
6
2
1
0
1
6
2
1
0
1
6
2
1
0
1
6
A
c
c
u
r
a
c
y
a
f
t
e
r
V
a
r
i
a
n
c
e
T
h
r
e
s
h
o
l
d
i
n
g
:
0
.
8
9
2
8
5
7
1
4
2
8
5
7
1
4
2
9
0
.
8
9
0
.
8
9
2
8
0
.
8
9
0
.
8
9
2
0
.
8
9
2
N
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
p
r
e
l
i
m
i
n
a
r
y
P
C
A
:
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
R
F
E:
5
0
50
50
50
50
50
A
c
c
u
r
a
c
y
a
f
t
e
r
R
F
E:
0
.
8
2
1
4
2
8
5
7
1
4
2
8
5
7
1
4
0
.
8
5
0
.
8
2
1
4
0
.
8
2
1
4
0
.
8
2
1
4
0
.
8
2
1
4
T
ab
le
3
.
Per
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
f
o
r
d
if
f
e
r
en
t th
r
esh
o
ld
d
u
r
in
g
o
p
tim
izatio
n
Th
r
e
s
h
o
l
d
0
.
8
0
.
9
1
1
.
2
2
4
8
16
N
u
mb
e
r
o
f
i
m
a
g
e
s
2
8
0
2
8
0
2
8
0
2
8
0
2
8
0
2
8
0
2
8
0
2
8
0
O
r
i
g
i
n
a
l
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
P
C
A
43
43
43
43
43
43
43
43
A
c
c
u
r
a
c
y
a
f
t
e
r
P
C
A
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
R
e
d
a
t
u
r
e
s
a
f
t
e
r
V
a
r
i
a
n
c
e
T
h
r
e
s
h
o
l
d
i
n
g
2
1
0
1
6
2
0
8
6
3
2
0
7
3
4
2
0
4
9
0
1
9
6
8
8
1
8
1
0
2
1
5
9
5
9
1
3
1
4
1
A
c
c
u
r
a
c
y
a
f
t
e
r
V
a
r
i
a
n
c
e
T
h
r
e
s
h
o
l
d
i
n
g
0
.
8
9
0
.
8
9
0
.
8
3
0
.
8
7
5
0
.
9
1
0
.
9
1
0
.
8
7
5
0
.
8
5
7
N
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
p
r
e
l
i
m
i
n
a
r
y
P
C
A
2
0
0
2
0
0
2
0
0
2
0
0
2
0
0
2
0
0
2
0
0
2
0
0
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
R
F
E
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
A
c
c
u
r
a
c
y
a
f
t
e
r
R
F
E
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
0
.
8
5
A
c
c
u
r
a
c
y
w
i
t
h
o
u
t
o
p
t
i
mi
z
a
t
i
o
n
0
.
9
1
0
.
9
1
0
.
9
1
0
.
9
1
0
.
9
1
0
.
9
1
0
.
9
1
0
.
9
1
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
.
4
,
Au
g
u
s
t
20
25
:
3
7
5
9
-
3768
3764
3.
3
.
2
.
Rec
urs
iv
e
f
ea
t
ure
elimin
a
t
io
n
(
RF
E
)
R
FE
i
s
ap
p
li
ed
af
ter
an
i
n
i
ti
a
l
P
C
A
r
e
d
u
c
tio
n
(
P
C
A
(
_
=
1
0
0
)
)
to
f
u
r
th
er
s
el
ec
t
th
e
m
o
s
t
in
f
o
r
m
at
iv
e
f
ea
t
u
r
e
s
(
_
_
_
=
5
0
)
.
R
FE
i
ter
at
iv
e
ly
r
em
o
v
e
s
le
s
s
im
p
o
r
tan
t
f
ea
tu
r
es
b
as
ed
o
n
th
eir
co
n
tr
ib
u
tio
n
to
m
o
d
el
a
cc
u
r
ac
y
,
u
s
in
g
a
R
a
n
d
o
m
Fo
r
e
s
t
e
s
ti
m
a
to
r
(
_
)
.
I
n
th
i
s
ap
p
r
o
a
ch
,
a
r
an
d
o
m
f
o
r
e
s
t
c
l
as
s
if
i
er
is
em
p
lo
y
ed
to
r
an
k
f
ea
tu
r
e
im
p
o
r
tan
ce
d
u
e
to
it
s
ca
p
a
ci
ty
t
o
m
an
ag
e
lar
g
e
f
e
atu
r
e
s
e
t
s
an
d
g
en
e
r
a
t
e
r
e
li
ab
l
e
i
m
p
o
r
t
an
c
e
s
co
r
e
s
.
T
h
e
R
FE
p
r
o
ce
s
s
is
co
n
f
ig
u
r
ed
to
s
el
ec
t
th
e
to
p
1
0
0
f
ea
tu
r
e
s
,
s
y
s
tem
at
ic
al
ly
r
em
o
v
in
g
5
f
ea
tu
r
e
s
at
e
ac
h
it
e
r
at
io
n
.
R
an
d
o
m
f
o
r
es
t
c
la
s
s
i
f
i
er
s
ar
e
u
s
ed
to
r
an
k
f
ea
tu
r
e
im
p
o
r
t
an
c
e,
lev
er
ag
in
g
i
ts
ab
il
i
ty
to
h
a
n
d
l
e
lar
g
e
f
e
atu
r
e
s
e
t
s
a
n
d
p
r
o
v
id
e
r
o
b
u
s
t
im
p
o
r
tan
ce
s
co
r
e
s
.
T
h
e
R
FE
c
o
n
f
ig
u
r
a
tio
n
ai
m
s
to
s
e
lec
t
1
0
0
f
ea
tu
r
es
,
r
e
m
o
v
i
n
g
5
f
ea
tu
r
es
in
at
ea
ch
i
ter
at
io
n
.
T
h
i
s
ap
p
r
o
a
ch
in
v
o
lv
e
s
tr
ain
in
g
th
e
m
o
d
e
l
o
n
t
h
e
f
u
ll
f
e
at
u
r
e
s
e
t,
r
an
k
i
n
g
f
e
at
u
r
e
s
b
a
s
e
d
o
n
th
e
ir
i
m
p
o
r
t
an
ce
,
an
d
th
en
el
im
in
at
in
g
th
e
lea
s
t
s
ig
n
i
f
i
c
an
t
o
n
e
s
.
T
h
is
p
r
o
ce
s
s
is
r
e
p
ea
t
ed
u
n
t
il
th
e
d
es
ir
ed
n
u
m
b
er
o
f
f
ea
t
u
r
e
s
i
s
r
ea
ch
ed
,
en
s
u
r
in
g
th
a
t
o
n
ly
t
h
e
m
o
s
t
i
m
p
a
ctf
u
l
f
ea
tu
r
es
a
r
e
r
e
ta
in
ed
f
o
r
en
h
an
c
ed
m
o
d
el
ef
f
ic
ien
cy
an
d
p
er
f
o
r
m
an
c
e.
T
ab
le
4
p
r
o
v
id
es
t
h
e
v
ar
ia
ti
o
n
o
f
th
e
m
o
d
el
p
er
f
o
r
m
an
ce
f
o
r
d
if
f
e
r
en
t
e
s
tim
ato
r
to
r
ed
u
ce
d
f
ea
tu
r
e
s
v
alu
e
s
.
T
ab
le
4
.
Per
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
f
o
r
d
if
f
e
r
en
t r
atio
s
o
f
f
ea
t
u
r
es to
iter
atio
n
R
e
d
u
c
e
d
f
e
a
t
u
r
e
s
1
5
0
/
1
0
0
2
0
0
/
1
0
0
1
0
0
/
1
0
0
O
r
i
g
i
n
a
l
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s:
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
2
5
0
8
8
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
P
C
A
:
4
3
43
43
43
A
c
c
u
r
a
c
y
a
f
t
e
r
P
C
A
:
0
.
8
5
7
1
4
2
8
5
7
1
4
2
8
5
7
1
0
.
8
5
0
.
8
5
0
.
8
5
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
V
a
r
i
a
n
c
e
T
h
r
e
s
h
o
l
d
i
n
g
:
2
1
0
1
6
2
1
0
1
6
2
1
0
1
6
2
1
0
1
6
A
c
c
u
r
a
c
y
a
f
t
e
r
V
a
r
i
a
n
c
e
T
h
r
e
s
h
o
l
d
i
n
g
:
0
.
8
9
2
8
5
7
1
4
2
8
5
7
1
4
2
9
0
.
8
9
0
.
8
9
0
.
8
9
N
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
p
r
e
l
i
m
i
n
a
r
y
P
C
A
:
1
5
0
1
5
0
2
0
0
1
0
0
R
e
d
u
c
e
d
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s a
f
t
e
r
R
F
E:
1
0
0
1
0
0
1
0
0
1
0
0
A
c
c
u
r
a
c
y
a
f
t
e
r
R
F
E:
0
.
8
5
7
1
4
2
8
5
7
1
4
2
8
5
7
1
0
.
8
2
0
.
8
5
0
.
8
3
3.
3
.
3.
Cla
s
s
if
ica
t
io
n
T
h
is
s
tu
d
y
ev
alu
ates
f
o
u
r
class
if
ier
s
:
r
an
d
o
m
f
o
r
est,
K
-
NN
,
SVM,
an
d
n
aïv
e
B
ay
es.
R
an
d
o
m
f
o
r
est,
an
en
s
em
b
le
m
eth
o
d
,
co
n
s
tr
u
c
ts
m
u
ltip
le
d
ec
is
io
n
tr
ee
s
,
o
f
f
e
r
in
g
r
o
b
u
s
tn
ess
an
d
h
ig
h
ac
c
u
r
ac
y
with
c
o
m
p
lex
d
atasets
.
K
-
NN
class
if
ies
d
ata
b
ased
o
n
th
e
n
ea
r
est
n
eig
h
b
o
r
s
;
it
is
s
im
p
le
b
u
t
ca
n
b
e
co
m
p
u
tatio
n
ally
in
ten
s
iv
e
with
lar
g
e
d
atasets
.
SVM
is
a
p
o
wer
f
u
l
s
u
p
er
v
is
e
d
alg
o
r
ith
m
th
at
f
in
d
s
th
e
o
p
ti
m
al
h
y
p
er
p
la
n
e
f
o
r
class
s
ep
ar
atio
n
in
h
ig
h
-
d
im
en
s
io
n
al
s
p
ac
es b
u
t r
eq
u
ir
es c
ar
e
f
u
l tu
n
in
g
.
Naïv
e
B
ay
es,
a
p
r
o
b
ab
ilis
tic
clas
s
if
ier
b
ased
o
n
B
ay
es'
th
eo
r
em
,
a
s
s
u
m
es
f
ea
tu
r
e
in
d
ep
e
n
d
en
c
e
an
d
is
ef
f
ec
tiv
e
f
o
r
s
p
ec
if
ic
task
s
lik
e
tex
t
class
if
icatio
n
.
T
h
ese
cla
s
s
if
i
er
s
wer
e
s
elec
ted
f
o
r
th
eir
u
n
iq
u
e
ap
p
r
o
ac
h
es
an
d
s
tr
en
g
th
s
,
o
f
f
er
i
n
g
a
co
m
p
r
eh
e
n
s
iv
e
co
m
p
ar
is
o
n
o
f
th
eir
p
er
f
o
r
m
a
n
ce
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
th
ese
m
o
d
els
is
d
e
p
icted
in
Fig
u
r
e
3
,
with
co
r
r
esp
o
n
d
in
g
clas
s
if
icatio
n
m
etr
ics f
o
r
d
if
f
e
r
en
t CNN m
o
d
els p
r
o
v
id
ed
i
n
T
ab
le
5
.
Fig
u
r
e
3
.
Per
f
o
r
m
an
c
e
o
f
class
if
ier
s
f
o
r
d
if
f
er
en
t CNN m
o
d
el
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
Dee
p
fea
tu
r
e
r
ep
r
esen
ta
tio
n
f
o
r
a
u
to
ma
ted
p
la
n
t sp
ec
ies cla
s
s
ifica
tio
n
fr
o
m
lea
f ima
g
es
(
N
ikh
il I
n
a
md
a
r
)
3765
T
ab
le
5
.
C
lass
if
icatio
n
m
etr
ics f
o
r
d
i
f
f
er
en
t CNN m
o
d
els
C
N
N
C
l
a
s
si
f
i
e
r
A
c
c
u
r
a
c
y
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
S
c
o
r
e
V
G
G
1
6
RF
95
97
95
95
S
V
M
91
95
91
91
NB
95
96
95
95
K
-
NN
91
92
91
91
R
ESN
ET
5
0
RF
95
96
95
95
S
V
M
84
91
84
84
NB
96
97
96
97
K
-
NN
86
90
86
85
I
N
C
EPT
I
O
N
V
3
RF
88
90
88
88
S
V
M
84
82
84
85
NB
89
90
89
89
K
-
NN
88
91
88
87
X
C
EPTI
O
N
RF
95
97
95
95
S
V
M
86
92
86
87
NB
95
96
95
95
K
-
NN
88
91
88
87
D
EN
S
EN
ET1
2
1
RF
95
97
95
95
S
V
M
91
95
91
91
NB
95
96
95
95
K
-
NN
91
92
91
91
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
p
er
f
o
r
m
a
n
ce
e
v
alu
atio
n
o
f
v
a
r
io
u
s
C
NN
m
o
d
els,
in
cl
u
d
in
g
VGG1
6
,
R
esNet5
0
,
I
n
ce
p
tio
n
V3
,
Den
s
eNe
t1
2
1
,
an
d
Xce
p
tio
n
,
b
ef
o
r
e
an
d
af
ter
ap
p
ly
in
g
f
ea
tu
r
e
o
p
tim
izatio
n
tech
n
i
q
u
es,
p
r
o
v
id
es
v
alu
a
b
le
in
s
ig
h
ts
in
to
th
e
ef
f
ec
tiv
en
es
s
o
f
th
ese
m
o
d
els
an
d
th
e
i
m
p
ac
t
o
f
o
p
tim
izatio
n
o
n
class
if
icatio
n
ac
cu
r
ac
y
.
I
n
itially
,
m
o
d
els
lik
e
R
esNet5
0
an
d
Xce
p
tio
n
p
er
f
o
r
m
ed
well
ev
en
with
o
u
t
o
p
tim
izatio
n
,
in
d
icatin
g
t
h
eir
in
h
er
en
t
ca
p
a
b
ilit
y
to
ca
p
tu
r
e
an
d
r
ep
r
esen
t
in
tr
icate
p
atter
n
s
in
leaf
im
ag
es.
Ho
wev
er
,
f
e
atu
r
e
o
p
tim
izati
o
n
tech
n
iq
u
es
s
u
ch
as
PC
A,
v
ar
i
an
ce
th
r
esh
o
ld
in
g
,
an
d
R
FE
s
i
g
n
if
ican
tly
en
h
a
n
ce
d
th
e
m
o
d
els
'
p
er
f
o
r
m
an
ce
b
y
r
ed
u
cin
g
d
im
en
s
io
n
ality
,
f
ilte
r
in
g
o
u
t
less
in
f
o
r
m
ativ
e
f
ea
t
u
r
es,
an
d
s
y
s
tem
atica
lly
r
em
o
v
in
g
less
im
p
o
r
ta
n
t
f
ea
tu
r
es.
T
h
is
led
to
im
p
r
o
v
ed
co
m
p
u
tat
io
n
al
ef
f
icien
c
y
an
d
ac
cu
r
ac
y
,
with
o
p
tim
ized
f
ea
t
u
r
e
s
ets p
r
o
v
id
in
g
a
m
o
r
e
e
f
f
icien
t
f
o
u
n
d
atio
n
f
o
r
f
u
tu
r
e
m
o
d
el
d
e
v
elo
p
m
e
n
t.
T
a
b
le
6
p
r
o
v
id
es
t
h
e
d
etails
o
f
t
h
e
h
y
p
er
p
ar
am
eter
s
u
s
ed
to
s
et
th
e
class
if
ier
s
.
T
h
e
co
m
p
a
r
ativ
e
p
er
f
o
r
m
an
ce
o
f
th
e
m
o
d
els
h
ig
h
lig
h
ts
th
e
im
p
o
r
tan
ce
o
f
f
ea
tu
r
e
o
p
tim
izatio
n
in
ac
h
iev
in
g
s
ta
te
-
of
-
th
e
-
a
r
t
r
esu
lts
.
Alth
o
u
g
h
s
o
m
e
m
o
d
els
ac
h
iev
e
d
h
i
g
h
ac
cu
r
ac
y
with
o
u
t
o
p
tim
izatio
n
,
th
e
ap
p
licatio
n
o
f
PC
A,
Var
ian
ce
T
h
r
esh
o
ld
i
n
g
,
an
d
R
FE
p
r
o
v
id
ed
ad
d
itio
n
al
b
en
ef
its
in
ter
m
s
o
f
ef
f
ic
ien
c
y
an
d
r
o
b
u
s
tn
ess
.
Fig
u
r
e
4
s
h
o
ws
th
e
co
n
f
u
s
io
n
m
atr
ix
an
d
R
OC
cu
r
v
e
f
o
r
v
ar
io
u
s
C
NN
m
o
d
els
with
s
elec
ted
class
if
ier
s
,
d
em
o
n
s
tr
atin
g
s
tr
o
n
g
class
if
icatio
n
ac
cu
r
ac
y
as
d
etailed
in
T
ab
le
5
.
T
h
e
c
o
n
f
u
s
io
n
m
atr
ix
an
d
R
OC
cu
r
v
e
an
aly
s
es f
u
r
th
er
illu
s
tr
ate
th
e
m
o
d
els
'
ef
f
ec
tiv
en
ess
in
d
is
tin
g
u
is
h
in
g
b
etwe
en
d
if
f
e
r
en
t
p
lan
t
s
p
ec
ies,
with
h
ig
h
e
r
AUC
v
alu
es
r
ef
lectin
g
b
ett
er
class
d
is
cr
im
in
atio
n
.
T
h
i
s
in
-
d
ep
th
an
al
y
s
is
u
n
d
er
s
co
r
es
th
e
s
ig
n
if
ican
ce
o
f
in
teg
r
atin
g
ad
v
a
n
ce
d
C
NN
ar
ch
itectu
r
es
with
f
ea
tu
r
e
o
p
ti
m
izatio
n
tech
n
iq
u
es,
m
ak
in
g
th
e
p
r
o
p
o
s
ed
m
o
d
el
a
v
alu
ab
le
to
o
l
f
o
r
p
r
ec
is
i
o
n
ag
r
ic
u
ltu
r
e,
b
io
d
iv
e
r
s
ity
co
n
s
er
v
atio
n
,
an
d
ec
o
lo
g
ical
m
o
n
ito
r
in
g
.
4
.
1
.
Co
m
pa
riso
n wit
h e
x
is
t
ing
m
et
ho
ds
T
h
e
d
ataset
u
n
d
e
r
co
n
s
id
er
ati
o
n
co
n
s
is
ts
o
f
2
0
,
3
5
7
im
ag
es
r
ep
r
esen
tin
g
f
o
u
r
tee
n
class
es
o
f
p
lan
t
leav
es.
A
co
m
p
ar
ativ
e
an
aly
s
is
,
h
ig
h
lig
h
ts
th
e
p
er
f
o
r
m
a
n
ce
d
if
f
er
en
ce
s
b
etwe
en
h
a
n
d
cr
af
ted
an
d
d
ee
p
f
ea
tu
r
es.
W
h
ile
d
ee
p
f
ea
t
u
r
es
ex
h
ib
it
s
u
p
e
r
io
r
p
er
f
o
r
m
an
ce
,
th
ey
also
p
r
esen
t
c
h
allen
g
es,
s
u
ch
as
t
h
e
n
ee
d
f
o
r
s
u
b
s
tan
tial
d
atasets
an
d
s
i
g
n
if
ican
t
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
.
Fu
r
th
er
m
o
r
e,
T
ab
le
6
illu
s
tr
ates
th
at
th
is
wo
r
k
s
u
r
p
ass
es c
o
n
v
en
tio
n
al
r
esear
c
h
m
eth
o
d
o
lo
g
ies,
p
a
r
ticu
lar
ly
i
n
h
an
d
li
n
g
a
lar
g
er
n
u
m
b
er
o
f
class
es.
T
ab
le
6
.
C
o
m
p
a
r
is
o
n
with
r
ela
ted
wo
r
k
S
l
n
o
R
e
f
e
r
e
n
c
e
n
o
M
e
t
h
o
d
N
o
o
f
c
l
a
sse
s
A
c
c
u
r
a
c
y
1
[
2
5
]
V
G
G
1
6
4
9
0
.
4
0
2
[
2
0
]
M
-
S
V
M
4
9
7
.
2
0
3
[
2
1
]
D
W
T,
C
O
LO
R
H
I
S
TO
G
R
A
M
3
9
8
.
6
3
4
[
2
2
]
S
H
U
F
F
LEN
ETV
1
4
9
7
.
7
9
5
[
2
3
]
D
EEP FEA
TU
R
E
+
LB
P
4
9
8
.
8
0
6
P
r
o
p
o
se
d
D
e
e
p
f
e
a
t
u
r
e
s
14
9
9
.
7
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
.
4
,
Au
g
u
s
t
20
25
:
3
7
5
9
-
3768
3766
V
G
G
1
6
m
o
d
e
l
w
i
t
h
R
F
c
l
a
ssi
f
i
e
r
D
e
n
seN
e
t
mo
d
e
l
w
i
t
h
N
B
c
l
a
ss
i
f
i
e
r
I
n
c
e
p
t
i
o
n
V
3
m
o
d
e
l
w
i
t
h
N
B
c
l
a
ssi
f
i
e
r
Fig
u
r
e
4
.
C
o
n
f
u
s
io
n
m
atr
i
x
an
d
R
OC
cu
r
v
e
o
f
C
NN
m
o
d
els
5.
CO
NCLU
SI
O
N
T
h
e
s
tu
d
y
d
em
o
n
s
tr
ates
th
e
ef
f
icac
y
o
f
d
ee
p
f
ea
tu
r
e
r
ep
r
ese
n
tatio
n
in
th
e
au
to
m
ated
class
if
icatio
n
o
f
p
lan
t
s
p
ec
ies
u
s
in
g
leaf
im
a
g
es.
B
y
em
p
lo
y
in
g
C
NN
m
o
d
els
s
u
ch
as
VGG1
6
,
R
esNet5
0
,
Den
s
eNe
t1
2
1
,
I
n
ce
p
tio
n
,
an
d
Xce
p
tio
n
,
we
s
u
cc
ess
f
u
lly
ca
p
tu
r
ed
h
ig
h
-
le
v
e
l,
d
is
cr
im
in
ativ
e
f
ea
tu
r
es
ess
en
tial
f
o
r
ac
cu
r
ate
s
p
ec
ies
d
if
f
er
en
tiatio
n
.
T
h
e
a
p
p
licatio
n
o
f
o
p
tim
izatio
n
te
ch
n
iq
u
es
lik
e
PC
A,
Var
ian
ce
T
h
r
esh
o
ld
in
g
,
a
n
d
R
FE
f
u
r
th
er
en
h
an
ce
d
th
e
e
f
f
icien
cy
o
f
th
e
f
ea
tu
r
e
s
et,
lead
in
g
to
h
ig
h
class
if
icatio
n
ac
cu
r
ac
ies
wh
e
n
co
m
b
in
ed
with
class
if
ier
s
s
u
c
h
as
SVM,
K
-
NN,
DT
,
an
d
N
B
.
T
h
e
ac
h
iev
ed
r
esu
lts
,
with
ac
cu
r
ac
ies
r
ea
ch
in
g
u
p
to
9
9
.
7
%,
u
n
d
er
s
co
r
e
th
e
p
o
ten
tial
o
f
th
is
ap
p
r
o
ac
h
i
n
ad
v
an
cin
g
ag
r
icu
ltu
r
al
r
esear
ch
,
b
io
d
i
v
er
s
ity
co
n
s
er
v
atio
n
,
an
d
ec
o
lo
g
ical
m
o
n
ito
r
in
g
.
Fu
tu
r
e
wo
r
k
will
f
o
cu
s
o
n
ex
p
an
d
in
g
th
e
d
at
aset
an
d
ex
p
lo
r
in
g
ad
d
itio
n
al
o
p
tim
izatio
n
s
tr
ateg
ies to
f
u
r
th
er
r
ef
in
e
th
e
class
if
i
ca
tio
n
p
r
o
ce
s
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
Dee
p
fea
tu
r
e
r
ep
r
esen
ta
tio
n
f
o
r
a
u
to
ma
ted
p
la
n
t sp
ec
ies cla
s
s
ifica
tio
n
fr
o
m
lea
f ima
g
es
(
N
ikh
il I
n
a
md
a
r
)
3767
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
is
r
esear
ch
d
id
n
o
t
r
ec
eiv
e
an
y
s
p
ec
if
ic
g
r
an
t
f
r
o
m
f
u
n
d
i
n
g
ag
en
cies
in
th
e
p
u
b
lic,
co
m
m
er
cial,
o
r
not
-
f
o
r
-
p
r
o
f
it secto
r
s
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
Nik
h
il
I
n
am
d
ar
co
n
ce
iv
e
d
th
e
s
tu
d
y
,
d
esig
n
ed
th
e
m
eth
o
d
o
lo
g
y
,
a
n
d
p
er
f
o
r
m
ed
th
e
e
x
p
er
im
en
ts
.
Ma
n
ju
n
ath
Ma
n
g
u
li
co
n
t
r
ib
u
t
ed
to
d
ata
an
aly
s
is
,
in
ter
p
r
etatio
n
,
a
n
d
tech
n
ical
v
alid
atio
n
.
Uttam
Patil
ass
is
ted
with
m
an
u
s
cr
ip
t
d
r
af
tin
g
,
cr
it
ical
r
ev
is
io
n
s
,
an
d
f
in
al
ap
p
r
o
v
al
o
f
th
e
v
er
s
io
n
to
b
e
p
u
b
lis
h
ed
.
All
au
th
o
r
s
h
av
e
r
ea
d
an
d
a
g
r
ee
d
to
th
e
p
u
b
lis
h
ed
v
er
s
io
n
o
f
th
e
m
an
u
s
cr
ip
t.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Nik
h
il I
n
am
d
a
r
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Ma
n
ju
n
ath
Ma
n
a
g
u
li
✓
✓
Uttam
Patil
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
d
ec
lar
e
th
at
t
h
er
e
is
n
o
co
n
f
lict o
f
in
ter
est
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
u
s
ed
in
th
is
s
tu
d
y
a
r
e
p
u
b
licly
a
v
ailab
le
an
d
ca
n
b
e
ac
ce
s
s
ed
f
r
o
m
o
p
en
s
o
u
r
c
es
as
cited
with
in
th
e
m
an
u
s
cr
ip
t
o
r
ca
n
b
e
co
n
tacte
d
to
a
u
th
o
r
RE
F
E
R
E
NC
E
S
[
1
]
A
.
A
.
B
h
a
r
a
t
e
a
n
d
M
.
S
.
S
h
i
r
d
h
o
n
k
a
r
,
“
A
r
e
v
i
e
w
o
n
p
l
a
n
t
d
i
se
a
se
d
e
t
e
c
t
i
o
n
u
s
i
n
g
i
ma
g
e
p
r
o
c
e
ss
i
n
g
,
”
i
n
2
0
1
7
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
I
n
t
e
l
l
i
g
e
n
t
S
u
s
t
a
i
n
a
b
l
e
S
y
s
t
e
m
s
(
I
C
I
S
S
)
,
D
e
c
.
2
0
1
7
,
p
p
.
1
0
3
–
1
0
9
.
d
o
i
:
1
0
.
1
1
0
9
/
i
ss1
.
2
0
1
7
.
8
3
8
9
3
2
6
.
[
2
]
P
.
Zh
a
o
,
G
.
Li
u
,
M
.
L
i
,
a
n
d
D
.
L
i
,
“
M
a
n
a
g
e
m
e
n
t
i
n
f
o
r
ma
t
i
o
n
sy
st
e
m
f
o
r
a
p
p
l
e
d
i
s
e
a
ses
a
n
d
i
n
se
c
t
p
e
st
s
b
a
se
d
o
n
G
I
S
,
”
N
o
n
g
y
e
G
o
n
g
c
h
e
n
g
X
u
e
b
a
o
/
T
ra
n
s
a
c
t
i
o
n
s
o
f
t
h
e
C
h
i
n
e
se
S
o
c
i
e
t
y
o
f
A
g
r
i
c
u
l
t
u
r
a
l
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
2
2
,
n
o
.
1
2
,
p
p
.
1
5
0
–
1
5
4
,
2
0
0
6
,
d
o
i
:
1
0
.
3
9
6
9
/
j
.
i
ss
n
.
1
0
0
2
-
6
8
1
9
.
2
0
0
6
.
1
2
.
0
2
9
.
[
3
]
G
.
G
e
e
t
h
a
r
a
m
a
n
i
a
n
d
J.
A
.
P
a
n
d
i
a
n
,
“
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
se
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
& E
l
e
c
t
r
i
c
a
l
E
n
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
.
[
4
]
E.
M
.
F
.
El
H
o
u
b
y
,
“
A
su
r
v
e
y
o
n
a
p
p
l
y
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s
f
o
r
ma
n
a
g
e
m
e
n
t
o
f
d
i
sea
s
e
s,
”
J
o
u
rn
a
l
o
f
A
p
p
l
i
e
d
Bi
o
m
e
d
i
c
i
n
e
,
v
o
l
.
1
6
,
n
o
.
3
,
p
p
.
1
6
5
–
1
7
4
,
A
u
g
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
a
b
.
2
0
1
8
.
0
1
.
0
0
2
.
[
5
]
C.
-
C
.
Y
a
n
g
e
t
a
l
.
,
“
A
p
p
l
i
c
a
t
i
o
n
o
f
d
e
c
i
s
i
o
n
t
r
e
e
t
e
c
h
n
o
l
o
g
y
f
o
r
i
m
a
g
e
c
l
a
ssi
f
i
c
a
t
i
o
n
u
si
n
g
r
e
m
o
t
e
s
e
n
s
i
n
g
d
a
t
a
,
”
A
g
ri
c
u
l
t
u
r
a
l
S
y
s
t
e
m
s
,
v
o
l
.
7
6
,
n
o
.
3
,
p
p
.
1
1
0
1
–
1
1
1
7
,
J
u
n
.
2
0
0
3
,
d
o
i
:
1
0
.
1
0
1
6
/
s
0
3
0
8
-
5
2
1
x
(
0
2
)
0
0
0
5
1
-
3.
[
6
]
M
.
S
h
a
r
i
f
,
M
.
A
.
K
h
a
n
,
Z.
I
q
b
a
l
,
M
.
F
.
A
z
a
m,
M
.
I
.
U
.
La
l
i
,
a
n
d
M
.
Y
.
Ja
v
e
d
,
“
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
c
i
t
r
u
s
d
i
sea
ses
i
n
a
g
r
i
c
u
l
t
u
r
e
b
a
s
e
d
o
n
o
p
t
i
m
i
z
e
d
w
e
i
g
h
t
e
d
s
e
g
m
e
n
t
a
t
i
o
n
a
n
d
f
e
a
t
u
r
e
s
e
l
e
c
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
A
g
r
i
c
u
l
t
u
r
e
,
v
o
l
.
1
5
0
,
p
p
.
2
2
0
–
2
3
4
,
J
u
l
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
a
g
.
2
0
1
8
.
0
4
.
0
2
3
.
[
7
]
J.
K
.
P
a
t
i
l
a
n
d
R
.
K
u
mar,
“
A
n
a
l
y
si
s
o
f
c
o
n
t
e
n
t
b
a
se
d
i
m
a
g
e
r
e
t
r
i
e
v
a
l
f
o
r
p
l
a
n
t
l
e
a
f
d
i
se
a
ses
u
si
n
g
c
o
l
o
r
,
s
h
a
p
e
a
n
d
t
e
x
t
u
r
e
f
e
a
t
u
r
e
s,
”
E
n
g
i
n
e
e
r
i
n
g
i
n
A
g
r
i
c
u
l
t
u
re,
En
v
i
r
o
n
m
e
n
t
a
n
d
F
o
o
d
,
v
o
l
.
1
0
,
n
o
.
2
,
p
p
.
6
9
–
7
8
,
A
p
r
.
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
a
e
f
.
2
0
1
6
.
1
1
.
0
0
4
.
[
8
]
B
.
S
a
n
d
i
k
a
,
S
.
A
v
i
l
,
S
.
S
a
n
a
t
,
a
n
d
P
.
S
r
i
n
i
v
a
s
u
,
“
R
a
n
d
o
m
f
o
r
e
s
t
b
a
se
d
c
l
a
s
si
f
i
c
a
t
i
o
n
o
f
d
i
se
a
ses
i
n
g
r
a
p
e
s
f
r
o
m
i
mag
e
s
c
a
p
t
u
r
e
d
i
n
u
n
c
o
n
t
r
o
l
l
e
d
e
n
v
i
r
o
n
m
e
n
t
s,
”
i
n
2
0
1
6
I
EE
E
1
3
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
S
i
g
n
a
l
Pr
o
c
e
ss
i
n
g
(
I
C
S
P)
,
N
o
v
.
2
0
1
6
,
p
p
.
1
7
7
5
–
1
7
8
0
.
d
o
i
:
1
0
.
1
1
0
9
/
i
c
s
p
.
2
0
1
6
.
7
8
7
8
1
3
3
.
[
9
]
S
.
U
ğ
u
z
a
n
d
N
.
U
y
s
a
l
,
“
C
l
a
s
si
f
i
c
a
t
i
o
n
o
f
o
l
i
v
e
l
e
a
f
d
i
s
e
a
ses
u
s
i
n
g
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s,
”
N
e
u
r
a
l
C
o
m
p
u
t
i
n
g
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
3
3
,
n
o
.
9
,
p
p
.
4
1
3
3
–
4
1
4
9
,
A
u
g
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
5
2
1
-
0
2
0
-
0
5
2
3
5
-
5.
[
1
0
]
M
.
G
.
S
e
l
v
a
r
a
j
e
t
a
l
.
,
“
A
I
-
p
o
w
e
r
e
d
b
a
n
a
n
a
d
i
sea
s
e
s
a
n
d
p
e
st
d
e
t
e
c
t
i
o
n
,
”
Pl
a
n
t
Me
t
h
o
d
s
,
v
o
l
.
1
5
,
n
o
.
1
,
A
u
g
.
2
0
1
9
,
d
o
i
:
1
0
.
1
1
8
6
/
s
1
3
0
0
7
-
0
1
9
-
0
4
7
5
-
z.
[
1
1
]
J.
L
u
,
J.
H
u
,
G
.
Z
h
a
o
,
F
.
M
e
i
,
a
n
d
C
.
Zh
a
n
g
,
“
A
n
i
n
-
f
i
e
l
d
a
u
t
o
ma
t
i
c
w
h
e
a
t
d
i
s
e
a
se
d
i
a
g
n
o
si
s
s
y
st
e
m,”
C
o
m
p
u
t
e
r
s
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
4
2
,
p
p
.
3
6
9
–
3
7
9
,
N
o
v
.
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
a
g
.
2
0
1
7
.
0
9
.
0
1
2
.
[
1
2
]
G
.
H
u
,
X
.
Y
a
n
g
,
Y
.
Zh
a
n
g
,
a
n
d
M
.
W
a
n
,
“
I
d
e
n
t
i
f
i
c
a
t
i
o
n
o
f
t
e
a
l
e
a
f
d
i
se
a
ses
b
y
u
si
n
g
a
n
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
u
st
a
i
n
a
b
l
e
C
o
m
p
u
t
i
n
g
:
I
n
f
o
rm
a
t
i
c
s
a
n
d
S
y
s
t
e
m
s
,
v
o
l
.
2
4
,
p
.
1
0
0
3
5
3
,
D
e
c
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
su
s
c
o
m.2
0
1
9
.
1
0
0
3
5
3
.
[
1
3
]
J.
G
.
A
r
n
a
l
B
a
r
b
e
d
o
,
“
P
l
a
n
t
d
i
se
a
se
i
d
e
n
t
i
f
i
c
a
t
i
o
n
f
r
o
m i
n
d
i
v
i
d
u
a
l
l
e
s
i
o
n
s
a
n
d
sp
o
t
s u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
,
”
Bi
o
sys
t
e
m
s E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
8
0
,
p
p
.
9
6
–
1
0
7
,
A
p
r
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
i
o
s
y
s
t
e
ms
e
n
g
.
2
0
1
9
.
0
2
.
0
0
2
.
[
1
4
]
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
,
J
u
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
.
0
3
.
0
3
2
.
[
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
,
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
.
4
,
Au
g
u
s
t
20
25
:
3
7
5
9
-
3768
3768
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
]
B
.
Li
u
,
Y
.
Zh
a
n
g
,
D
.
H
e
,
a
n
d
Y
.
Li
,
“
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
ses
b
a
s
e
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,
”
S
y
m
m
e
t
ry
,
v
o
l
.
1
0
,
n
o
.
1
,
p
.
1
1
,
D
e
c
.
2
0
1
7
,
d
o
i
:
1
0
.
3
3
9
0
/
s
y
m
1
0
0
1
0
0
1
1
.
[
1
7
]
Z.
C
h
u
a
n
l
e
i
,
Z.
S
h
a
n
w
e
n
,
Y
.
J
u
c
h
e
n
g
,
S
.
Y
a
n
c
u
i
,
a
n
d
C
.
J
i
a
,
“
A
p
p
l
e
l
e
a
f
d
i
sea
s
e
i
d
e
n
t
i
f
i
c
a
t
i
o
n
u
si
n
g
g
e
n
e
t
i
c
a
l
g
o
r
i
t
h
m
a
n
d
c
o
r
r
e
l
a
t
i
o
n
-
b
a
s
e
d
f
e
a
t
u
r
e
se
l
e
c
t
i
o
n
m
e
t
h
o
d
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
A
g
ri
c
u
l
t
u
ra
l
a
n
d
Bi
o
l
o
g
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
1
0
,
n
o
.
2
,
p
p
.
74
–
8
3
,
2
0
1
7
,
d
o
i
:
1
0
.
3
9
6
5
/
j
.
i
j
a
b
e
.
2
0
1
7
1
0
0
2
.
2
1
6
6
.
[
1
8
]
S
.
S
i
n
g
h
,
S
.
G
u
p
t
a
,
A
.
T
a
n
t
a
,
a
n
d
R
.
G
u
p
t
a
,
“
Ex
t
r
a
c
t
i
o
n
o
f
mu
l
t
i
p
l
e
d
i
se
a
s
e
s
i
n
a
p
p
l
e
l
e
a
f
u
si
n
g
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
m
a
g
e
a
n
d
G
r
a
p
h
i
c
s
,
v
o
l
.
2
2
,
n
o
.
0
3
,
F
e
b
.
2
0
2
1
,
d
o
i
:
1
0
.
1
1
4
2
/
s
0
2
1
9
4
6
7
8
2
1
4
0
0
0
9
x
.
[
1
9
]
S
.
C
h
a
k
r
a
b
o
r
t
y
,
S
.
P
a
u
l
,
a
n
d
M
.
R
a
h
a
t
-
uz
-
Z
a
ma
n
,
“
P
r
e
d
i
c
t
i
o
n
o
f
A
p
p
l
e
l
e
a
f
d
i
se
a
ses
u
si
n
g
mu
l
t
i
c
l
a
ss
s
u
p
p
o
r
t
v
e
c
t
o
r
ma
c
h
i
n
e
,
”
i
n
2
0
2
1
2
n
d
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
Ro
b
o
t
i
c
s,
E
l
e
c
t
r
i
c
a
l
a
n
d
S
i
g
n
a
l
Pro
c
e
ssi
n
g
T
e
c
h
n
i
q
u
e
s
(
I
C
RE
S
T
)
,
Ja
n
.
2
0
2
1
,
p
p
.
1
4
7
–
1
5
1
.
d
o
i
:
1
0
.
1
1
0
9
/
i
c
r
e
st
5
1
5
5
5
.
2
0
2
1
.
9
3
3
1
1
3
2
.
[
2
0
]
A
.
K
r
i
z
h
e
v
s
k
y
,
I
.
S
u
t
s
k
e
v
e
r
,
a
n
d
G
.
E.
H
i
n
t
o
n
,
“
I
mag
e
N
e
t
c
l
a
ssi
f
i
c
a
t
i
o
n
w
i
t
h
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
,
”
C
o
m
m
u
n
i
c
a
t
i
o
n
s
o
f
t
h
e
A
C
M
,
v
o
l
.
6
0
,
n
o
.
6
,
p
p
.
8
4
–
9
0
,
M
a
y
2
0
1
7
,
d
o
i
:
1
0
.
1
1
4
5
/
3
0
6
5
3
8
6
.
[
2
1
]
K
.
S
i
m
o
n
y
a
n
a
n
d
A
.
Z
i
sserm
a
n
,
“
V
e
r
y
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
t
w
o
r
k
s
f
o
r
l
a
r
g
e
-
s
c
a
l
e
i
ma
g
e
r
e
c
o
g
n
i
t
i
o
n
,
”
a
rXi
v
:
1
4
0
9
.
1
5
5
6
,
S
e
p
.
2
0
1
4
.
[
2
2
]
C
.
S
z
e
g
e
d
y
e
t
a
l
.
,
“
G
o
i
n
g
d
e
e
p
e
r
w
i
t
h
c
o
n
v
o
l
u
t
i
o
n
s,
”
i
n
2
0
1
5
I
EEE
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
s
i
o
n
a
n
d
P
a
t
t
e
r
n
R
e
c
o
g
n
i
t
i
o
n
(
C
VPR)
,
Ju
n
.
2
0
1
5
,
p
p
.
1
–
9
.
d
o
i
:
1
0
.
1
1
0
9
/
c
v
p
r
.
2
0
1
5
.
7
2
9
8
5
9
4
.
[
2
3
]
F
.
N
.
I
a
n
d
o
l
a
,
M
.
W
.
M
o
s
k
e
w
i
c
z
,
S
.
K
a
r
a
y
e
v
,
R
.
G
i
r
s
h
i
c
k
,
T
.
D
a
r
r
e
l
l
,
a
n
d
K
.
K
e
u
t
z
e
r
,
“
D
e
n
s
e
N
e
t
:
I
mp
l
e
men
t
i
n
g
e
f
f
i
c
i
e
n
t
c
o
n
v
n
e
t
d
e
s
c
r
i
p
t
o
r
p
y
r
a
mi
d
s,
”
a
rXi
v
p
rep
r
i
n
t
a
rX
i
v
:
1
4
0
4
.
1
8
6
9
,
2
0
1
4
.
[
2
4
]
K
.
H
e
,
X
.
Z
h
a
n
g
,
S
.
R
e
n
,
a
n
d
J
.
S
u
n
,
“
D
e
e
p
r
e
s
i
d
u
a
l
l
e
a
r
n
i
n
g
f
o
r
i
ma
g
e
r
e
c
o
g
n
i
t
i
o
n
,
”
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
I
EEE
C
o
m
p
u
t
e
r
S
o
c
i
e
t
y
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
s
i
o
n
a
n
d
Pa
t
t
e
r
n
Re
c
o
g
n
i
t
i
o
n
,
p
p
.
7
7
0
–
7
7
8
,
2
0
1
6
,
d
o
i
:
1
0
.
1
1
0
9
/
C
V
P
R
.
2
0
1
6
.
9
0
.
[
25]
L.
C
a
i
,
J.
Z
h
u
,
H
.
Ze
n
g
,
J
.
C
h
e
n
,
C
.
C
a
i
,
a
n
d
K
.
-
K
.
M
a
,
“
H
O
G
-
a
ssi
s
t
e
d
d
e
e
p
f
e
a
t
u
r
e
l
e
a
r
n
i
n
g
f
o
r
p
e
d
e
st
r
i
a
n
g
e
n
d
e
r
r
e
c
o
g
n
i
t
i
o
n
,
”
J
o
u
rn
a
l
o
f
t
h
e
Fra
n
k
l
i
n
I
n
st
i
t
u
t
e
,
v
o
l
.
3
5
5
,
n
o
.
4
,
p
p
.
1
9
9
1
–
2
0
0
8
,
M
a
r
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
f
r
a
n
k
l
i
n
.
2
0
1
7
.
0
9
.
0
0
3
.
BIO
G
RAP
H
I
E
S O
F
AUTH
O
RS
Nik
h
il
I
n
a
m
d
a
r
a
ss
istan
t
p
ro
fe
ss
o
r
a
t
KL
S
G
o
g
te
In
stit
u
te
o
f
Tec
h
n
o
lo
g
y
,
wh
ich
is
a
ffil
iate
d
wit
h
Vis
v
e
sv
a
ra
y
a
Tec
h
n
o
lo
g
ica
l
Un
i
v
e
rsity
,
l
o
c
a
ted
in
Be
lag
a
v
i
,
In
d
ia
.
Ho
ld
s
a
M
a
ste
r'
s
d
e
g
re
e
(M
.
Tec
h
.
)
i
n
In
d
u
strial
El
e
c
tro
n
ics
fr
o
m
Visv
e
sv
a
ra
y
a
Un
iv
e
rsit
y
,
Be
lag
a
v
i,
sh
o
wc
a
sin
g
h
is
e
x
p
e
rti
se
in
t
h
e
fiel
d
.
C
u
rre
n
tl
y
p
u
rs
u
in
g
a
P
h
.
D.
,
d
e
m
o
n
stra
ti
n
g
h
is
c
o
m
m
it
m
e
n
t
t
o
f
u
rt
h
e
rin
g
h
is
k
n
o
wle
d
g
e
a
n
d
c
o
n
tri
b
u
ti
n
g
to
a
c
a
d
e
m
ia.
In
h
is
a
c
a
d
e
m
ic
jo
u
r
n
e
y
,
h
e
h
a
s
a
k
e
e
n
i
n
tere
st
i
n
a
re
a
s
su
c
h
a
s
a
rti
ficia
l
in
telli
g
e
n
c
e
(AI),
m
a
c
h
in
e
lea
rn
in
g
(M
L),
d
e
e
p
lea
rn
i
n
g
(DL)
,
a
n
d
e
m
b
e
d
d
e
d
sy
ste
m
s
.
Th
e
se
a
re
a
s
o
f
in
tere
st
re
flec
t
p
a
ss
io
n
fo
r
c
u
tt
in
g
-
e
d
g
e
tec
h
n
o
l
o
g
ies
a
n
d
t
h
e
ir
a
p
p
li
c
a
ti
o
n
in
v
a
rio
u
s
field
s.
As
a
n
e
d
u
c
a
to
r
a
n
d
re
se
a
rc
h
e
r,
striv
e
s
to
b
rid
g
e
t
h
e
g
a
p
b
e
twe
e
n
t
h
e
o
r
y
a
n
d
p
ra
c
ti
c
e
,
a
imin
g
to
m
a
k
e
sig
n
ifi
c
a
n
t
c
o
n
tri
b
u
ti
o
n
s
to
t
h
e
a
d
v
a
n
c
e
m
e
n
t
o
f
t
h
e
se
field
s.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
n
ik
h
il
0
8
7
0
@g
m
a
il
.
c
o
m
a
n
d
n
ji
n
a
m
d
a
r@g
it
.
e
d
u
.
Ma
n
ju
n
a
th
M
a
n
a
g
u
li
is
a
n
a
c
c
o
m
p
li
sh
e
d
a
ss
o
c
iate
p
ro
fe
ss
o
r
with
a
d
e
c
a
d
e
o
f
d
e
d
ica
ted
se
rv
ice
in
a
c
a
d
e
m
ia.
Cu
rre
n
tl
y
a
ffil
iate
d
wi
th
t
h
e
e
ste
e
m
e
d
KLS
G
o
g
te
In
stit
u
te
o
f
Tec
h
n
o
l
o
g
y
,
Dr.
M
a
n
a
g
u
li
h
a
s
e
sta
b
li
sh
e
d
h
ims
e
lf
a
s
a
d
y
n
a
m
ic
e
d
u
c
a
to
r
a
n
d
a
c
o
m
m
it
te
d
re
se
a
rc
h
e
r.
Wi
th
a
p
a
ss
io
n
fo
r
tea
c
h
in
g
,
Dr.
M
a
n
a
g
u
l
i
h
a
s
b
e
e
n
a
c
t
iv
e
ly
in
v
o
l
v
e
d
i
n
sh
a
p
in
g
th
e
a
c
a
d
e
m
ic
jo
u
rn
e
y
o
f
stu
d
e
n
t
s,
imp
a
rti
n
g
k
n
o
wle
d
g
e
in
[m
e
n
ti
o
n
t
h
e
sp
e
c
ifi
c
field
o
r
su
b
jec
t]
.
His
e
n
g
a
g
in
g
tea
c
h
i
n
g
m
e
th
o
d
s
a
n
d
a
b
il
it
y
t
o
c
o
n
n
e
c
t
w
it
h
stu
d
e
n
ts
h
a
v
e
g
a
r
n
e
re
d
a
d
m
iratio
n
a
n
d
re
sp
e
c
t
with
in
th
e
a
c
a
d
e
m
ic
c
o
m
m
u
n
it
y
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
m
a
n
ju
n
a
th
m
@g
it
.
e
d
u
.
Utta
m
P
a
til
is
a
n
a
c
c
o
m
p
li
sh
e
d
p
ro
fe
ss
o
r
a
n
d
HO
D
Co
m
p
u
ter
S
c
ien
c
e
De
p
a
rtme
n
t
with
a
m
o
re
th
a
n
d
e
c
a
d
e
o
f
d
e
d
ica
ted
se
rv
ice
in
a
c
a
d
e
m
ia.
Cu
rre
n
tl
y
a
ffil
iate
d
with
th
e
e
ste
e
m
e
d
Ja
in
Co
ll
e
g
e
o
f
En
g
i
n
e
e
rin
g
,
Dr.
Uttam
P
a
ti
l
h
a
s
e
sta
b
li
sh
e
d
h
ims
e
lf
a
s
a
d
y
n
a
m
ic
e
d
u
c
a
to
r
a
n
d
a
c
o
m
m
it
ted
re
se
a
rc
h
e
r.
Wi
th
a
p
a
ss
io
n
fo
r
tea
c
h
in
g
,
Dr.
Ut
tam
P
a
ti
l
ha
s
b
e
e
n
a
c
ti
v
e
ly
in
v
o
l
v
e
d
i
n
sh
a
p
in
g
t
h
e
a
c
a
d
e
m
ic
jo
u
rn
e
y
o
f
st
u
d
e
n
ts,
imp
a
rti
n
g
k
n
o
wle
d
g
e
.
His
e
n
g
a
g
in
g
tea
c
h
in
g
m
e
th
o
d
s
a
n
d
a
b
i
li
ty
t
o
c
o
n
n
e
c
t
wit
h
st
u
d
e
n
ts
h
a
v
e
g
a
rn
e
re
d
a
d
m
iratio
n
a
n
d
re
sp
e
c
t
with
in
t
h
e
a
c
a
d
e
m
ic
c
o
m
m
u
n
it
y
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
u
tt
a
m
p
a
ti
l@jain
b
g
m
.
i
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.