T
E
L
K
O
M
NIKA
T
elec
o
mm
un
ica
t
io
n Co
m
pu
t
i
ng
E
lect
ro
nics
a
nd
Co
ntr
o
l
Vo
l.
23
,
No
.
5
,
Octo
b
er
20
25
,
p
p
.
1
353
~1
3
6
2
I
SS
N:
1
6
9
3
-
6
9
3
0
,
DOI
:
1
0
.
1
2
9
2
8
/
T
E
L
KOM
NI
K
A
.
v
23
i
5
.
26887
1353
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//jo
u
r
n
a
l.u
a
d
.
a
c.
id
/in
d
ex
.
p
h
p
/TELK
OM
N
I
K
A
Deep
trans
fer
learning
bas
ed disea
se dete
ction a
nd
c
la
ss
ificatio
n
o
f
to
m
a
to lea
v
es
-
a
co
m
pa
ra
tive
ana
ly
sis
M
un
ira
Ak
t
er
L
a
t
a
1
,
M
a
rj
i
a
Su
lt
a
na
2
,
I
f
f
a
t
Ara
B
a
dh
a
n
3
,
M
a
stu
r
a
J
a
h
a
n
M
a
r
ia
1
,
Fa
r
i
h
a
Ta
s
n
i
m
N
u
h
a
1
1
D
e
p
a
r
t
me
n
t
o
f
E
d
u
c
a
t
i
o
n
a
l
T
e
c
h
n
o
l
o
g
y
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
D
i
g
i
t
a
l
T
r
a
n
sf
o
r
mat
i
o
n
En
g
i
n
e
e
r
i
n
g
,
U
n
i
v
e
r
si
t
y
o
f
F
r
o
n
t
i
e
r
T
e
c
h
n
o
l
o
g
y
,
B
a
n
g
l
a
d
e
sh
,
K
a
l
i
a
k
a
i
r
,
G
a
z
i
p
u
r
,
B
a
n
g
l
a
d
e
sh
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
,
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
B
e
g
u
m R
o
k
e
y
a
U
n
i
v
e
r
si
t
y
,
R
a
n
g
p
u
r
,
R
a
n
g
p
u
r
,
B
a
n
g
l
a
d
e
sh
3
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
a
n
d
El
e
c
t
r
o
n
i
c
E
n
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
B
e
g
u
m R
o
k
e
y
a
U
n
i
v
e
r
si
t
y
,
R
a
n
g
p
u
r
,
R
a
n
g
p
u
r
,
B
a
n
g
l
a
d
e
sh
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
3
1
,
2
0
2
4
R
ev
i
s
ed
J
u
l 1
9
,
2
0
2
5
A
cc
ep
ted
Sep
1
0
,
2
0
2
5
A
w
id
e
v
a
riet
y
o
f
d
ise
a
se
s
h
a
v
e
a
sig
n
if
ica
n
t
im
p
a
c
t
o
n
to
m
a
to
p
lan
ts.
T
o
a
v
o
id
c
ro
p
q
u
a
li
ty
issu
e
s,
a
p
ro
m
p
t
a
n
d
p
re
c
ise
d
iag
n
o
sis
i
s
c
ru
c
ial.
Clas
sify
in
g
p
lan
t
d
ise
a
se
s
is
o
n
e
o
f
th
e
n
u
m
e
ro
u
s
a
p
p
li
c
a
ti
o
n
s
w
h
e
re
d
e
e
p
tran
sf
e
r
lea
rn
in
g
m
o
d
e
ls
h
a
v
e
re
c
e
n
tl
y
p
ro
d
u
c
e
d
re
m
a
rk
a
b
le
re
su
lt
s.
T
h
is
stu
d
y
d
e
a
lt
w
it
h
f
in
e
-
tu
n
in
g
b
y
c
o
n
tras
ti
n
g
th
e
m
o
st
a
d
v
a
n
c
e
d
a
rc
h
it
e
c
tu
re
s,
in
c
lu
d
in
g
I
n
c
e
p
ti
o
n
V
3
,
Re
s
Ne
t
-
1
8
,
Re
sN
e
t
-
5
0
,
V
GG
-
1
6
,
V
GG
-
1
9
,
G
o
o
g
LeN
e
t,
a
n
d
A
le
x
Ne
t.
In
th
e
e
n
d
,
a
c
o
m
p
a
riso
n
e
v
a
lu
a
ti
o
n
is
c
o
n
d
u
c
ted
.
Nin
e
d
is
ti
n
c
t
to
m
a
to
d
ise
a
se
c
las
se
s
a
n
d
o
n
e
h
e
a
lt
h
y
c
las
s
f
ro
m
P
la
n
tVil
lag
e
m
a
k
e
u
p
t
h
e
d
a
tas
e
t
u
se
d
i
n
t
h
is
stu
d
y
.
P
re
c
isi
o
n
,
re
c
a
ll
,
F
1
-
s
c
o
re
,
a
n
d
a
c
c
u
ra
c
y
w
e
re
th
e
b
a
sis
f
o
r
a
m
u
lt
icla
ss
sta
ti
stica
l
a
n
a
l
y
sis
th
a
t
a
s
se
ss
e
d
th
e
m
o
d
e
ls.
T
h
e
Re
sN
e
t
-
5
0
a
p
p
r
o
a
c
h
y
ie
ld
e
d
sig
n
if
ica
n
t
re
su
lt
s
w
it
h
p
re
c
isio
n
:
8
2
%
,
re
c
a
ll
:
8
1
%
,
F
1
-
sc
o
re
:
8
1
%
,
a
n
d
a
c
c
u
ra
c
y
:
8
5
%
.
W
it
h
th
is
h
ig
h
su
c
c
e
ss
ra
te,
it
is
re
a
so
n
a
b
le
to
sa
y
th
a
t
m
o
b
il
e
a
p
p
li
c
a
ti
o
n
s
o
r
Io
T
-
c
o
m
p
a
ti
b
le
g
a
d
g
e
ts
i
m
p
le
m
e
n
ted
w
it
h
th
e
Re
sN
e
t
-
5
0
m
o
d
e
l
c
a
n
a
ss
ist
f
a
r
m
e
rs
in
id
e
n
ti
f
y
in
g
a
n
d
sa
fe
g
u
a
rd
in
g
to
m
a
to
e
s a
g
a
in
st t
h
e
a
f
o
re
m
e
n
ti
o
n
e
d
d
ise
a
se
s.
K
ey
w
o
r
d
s
:
C
las
s
i
f
icatio
n
Dee
p
tr
an
s
f
er
lear
n
in
g
I
m
ag
e
p
r
o
ce
s
s
i
n
g
Occ
lu
s
io
n
s
e
n
s
i
tiv
it
y
T
o
m
ato
leaf
d
is
ea
s
e
s
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
:
Ma
r
j
ia
Su
ltan
a
Dep
ar
t
m
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
an
d
E
n
g
i
n
ee
r
in
g
,
Fac
u
lt
y
o
f
E
n
g
in
ee
r
i
n
g
an
d
T
ec
h
n
o
lo
g
y
B
eg
u
m
R
o
k
e
y
a
U
n
i
v
er
s
it
y
,
R
a
n
g
p
u
r
A
ca
d
e
m
ic
B
u
ild
i
n
g
-
2
,
Mo
r
d
an
Mo
r
R
o
ad
,
R
an
g
p
u
r
-
540
4
,
B
an
g
lad
e
s
h
E
m
ail:
m
ar
j
ia.
cse@
b
r
u
r
.
ac
.
b
d
1.
I
NT
RO
D
UCT
I
O
N
On
e
o
f
t
h
e
m
o
s
t
s
ig
n
i
f
ica
n
t
a
n
d
w
id
el
y
co
n
s
u
m
ed
v
e
g
etab
l
es
in
B
an
g
lad
es
h
is
th
e
to
m
at
o
(
S
o
la
n
u
m
lyco
p
ers
icu
m
)
.
I
t
h
as
an
tio
x
i
d
an
t
co
m
p
o
n
en
ts
i
n
cl
u
d
in
g
l
y
co
p
en
e,
w
h
ic
h
p
r
o
tects
ca
n
ce
r
,
an
d
is
a
s
tr
o
n
g
s
o
u
r
ce
o
f
v
ita
m
in
s
A
an
d
C
[
1
]
.
I
t
co
m
es
in
th
ir
d
p
lace
in
ter
m
s
o
f
ar
ea
an
d
f
o
u
r
th
in
ter
m
s
o
f
p
r
o
d
u
ctio
n
[
2
]
.
Sin
ce
d
is
ea
s
es
h
a
v
e
a
s
i
g
n
i
f
i
ca
n
t
i
m
p
ac
t
o
n
h
ea
lt
h
y
p
lan
t
s
an
d
ca
u
s
e
s
ig
n
i
f
ica
n
t
lo
s
s
es
in
t
h
e
a
g
r
icu
lt
u
r
al
s
ec
to
r
,
it
is
n
ec
e
s
s
ar
y
to
p
r
o
tect
th
e
m
f
r
o
m
d
is
ea
s
e
in
o
r
d
er
to
en
s
u
r
e
t
h
e
q
u
al
it
y
a
n
d
q
u
an
tit
y
o
f
cr
o
p
s
.
Dis
ea
s
e
s
ca
u
s
ed
ab
o
u
t
1
2
.
4
5
%
o
f
to
m
ato
es
to
b
e
lo
s
t
at
t
h
e
f
ar
m
lev
e
l
a
f
ter
h
ar
v
est,
w
it
h
8
.
8
6
%
o
f
t
h
o
s
e
lo
s
s
es
b
ei
n
g
d
u
e
to
f
u
ll
d
es
tr
u
ctio
n
,
ac
co
r
d
in
g
to
a
s
tu
d
y
w
a
s
o
u
t
i
n
t
h
e
d
is
tr
ict
s
o
f
J
a
m
alp
u
r
an
d
R
a
n
g
p
u
r
o
f
B
an
g
lad
es
h
[
3
]
.
T
h
is
m
ea
n
s
t
h
at
f
o
r
ev
er
y
d
ec
i
m
al
o
f
to
m
a
to
cu
lti
v
atio
n
,
B
an
g
lad
es
h
i
T
ak
a
(
B
DT
)
1
5
2
.
4
5
is
lo
s
t.
T
h
er
ef
o
r
e,
it
s
h
o
u
ld
b
e
e
m
p
h
a
s
ized
th
at
ea
r
l
y
m
o
n
ito
r
i
n
g
is
cr
u
c
ial
f
o
r
s
elec
ti
n
g
t
h
e
b
est
co
u
r
s
e
o
f
ac
tio
n
an
d
h
alti
n
g
t
h
e
s
p
r
ea
d
o
f
th
e
d
is
ea
s
es
in
b
o
th
to
m
ato
es
an
d
p
lan
ts
o
f
t
h
e
to
m
ato
es.
H
o
w
e
v
er
,
in
o
r
d
er
to
co
n
tr
o
l
d
is
ea
s
e,
f
ar
m
er
s
c
an
h
ar
d
ly
a
f
f
o
r
d
to
k
ee
p
a
clo
s
e
e
y
e
o
n
th
eir
to
m
ato
cr
o
p
s
.
Giv
en
th
e
d
if
f
ic
u
lt
y
o
f
ac
q
u
ir
in
g
ag
r
icu
l
tu
r
al
k
n
o
w
le
d
g
e
in
r
em
o
te
ar
ea
s
d
u
e
to
li
m
ited
ac
ce
s
s
to
s
u
ch
ex
p
er
tis
e
,
m
a
n
u
al
m
o
n
ito
r
i
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
23
,
No
.
5
,
Octo
b
e
r
20
25
:
1
3
5
3
-
1
362
1354
is
ti
m
e
-
co
n
s
u
m
in
g
a
n
d
lab
o
r
-
in
te
n
s
i
v
e.
Fo
r
th
e
s
e
r
ea
s
o
n
s
,
in
d
iv
i
d
u
al
f
ar
m
er
s
f
i
n
d
it
d
if
f
ic
u
lt
to
p
r
o
m
p
tl
y
d
iag
n
o
s
e
a
n
d
tr
ea
t
d
is
ea
s
es
in
o
r
d
er
t
o
en
s
u
r
e
th
e
b
est
p
o
s
s
i
b
le
q
u
alit
y
.
I
n
th
e
p
ast,
all
d
is
ea
s
es
an
d
p
r
o
b
lem
s
w
er
e
id
en
tifie
d
b
y
v
i
s
u
a
l
ex
a
m
i
n
atio
n
b
y
s
k
illed
p
eo
p
le
w
h
o
m
i
g
h
t
h
a
v
e
aid
ed
th
e
ir
an
al
y
s
i
s
w
it
h
f
ea
t
u
r
es
lik
e
co
lo
r
,
tex
t
u
r
e,
an
d
s
h
ap
e.
B
u
t
lo
w
ef
f
icie
n
c
y
an
d
e
x
ce
s
s
iv
e
e
x
p
en
s
e
s
w
er
e
t
h
e
o
u
tco
m
es
o
f
t
h
is
s
tr
ate
g
y
.
T
h
is
s
tu
d
y
co
n
s
id
er
s
t
h
ese
p
r
o
b
lem
s
as
ch
al
len
g
es
a
n
d
attem
p
ts
to
u
s
e
d
ee
p
tr
an
s
f
er
lear
n
in
g
tec
h
n
iq
u
es
t
o
p
r
o
p
o
s
e
a
tech
n
ical
s
o
lu
tio
n
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
v
ar
io
u
s
d
ee
p
tr
an
s
f
er
lear
n
in
g
ar
ch
ite
ctu
r
es
is
co
m
p
ar
ed
in
th
i
s
s
tu
d
y
to
h
elp
ch
o
o
s
e
an
a
u
to
m
a
ted
s
y
s
te
m
t
h
at
allo
w
s
its
ap
p
licatio
n
s
to
b
e
ex
p
an
d
ed
in
t
h
e
a
g
r
icu
ltu
r
al
d
o
m
ai
n
.
T
h
e
m
ai
n
ac
h
iev
e
m
e
n
t
o
f
t
h
i
s
s
t
u
d
y
is
a
co
m
p
r
eh
e
n
s
i
v
e
s
u
m
m
ar
y
o
f
t
h
e
w
o
r
k
i
n
g
s
o
f
ea
c
h
d
ee
p
tr
an
s
f
er
lear
n
in
g
tech
n
iq
u
e,
i
n
ad
d
itio
n
to
ap
p
l
y
i
n
g
ea
c
h
m
o
d
el
i
n
a
d
atab
ase
m
ad
e
u
p
o
f
a
n
u
m
b
er
o
f
p
h
o
to
s
r
elate
d
to
u
n
h
ea
lt
h
y
an
d
h
ea
lth
y
to
m
ato
leav
es
.
Fo
r
p
o
s
s
ib
le
f
u
t
u
r
e
ap
p
licatio
n
s
,
t
h
is
e
n
ab
les a
n
u
n
b
ia
s
ed
co
m
p
ar
is
o
n
o
f
th
e
b
eh
a
v
io
r
o
f
th
e
s
ev
er
al
d
ee
p
lear
n
in
g
b
ased
tr
an
s
f
er
lear
n
in
g
m
o
d
el
s
.
Fi
n
d
in
g
th
e
ar
ch
itectu
r
e
t
h
a
t
ef
f
ec
tiv
e
l
y
ca
p
t
u
r
es
t
h
e
is
s
u
e,
s
u
cc
es
s
f
u
l
l
y
cla
s
s
i
f
ie
s
to
m
a
to
p
lan
t
ill
n
es
s
es,
an
d
v
alid
ates
it
u
s
in
g
a
r
an
g
e
o
f
s
tatis
t
ical
m
ea
s
u
r
es
is
t
h
e
p
u
r
p
o
s
e.
Ag
r
icu
ltu
r
al
tec
h
n
ici
an
s
a
n
d
s
p
ec
ialis
t
s
m
a
y
f
i
n
d
th
e
d
ee
p
tr
an
s
f
er
lear
n
in
g
m
o
d
els
u
s
e
f
u
l
as
an
au
to
m
ated
s
y
s
te
m
f
o
r
id
en
ti
f
y
in
g
p
lan
t
ill
n
e
s
s
es.
Far
m
er
s
ca
n
p
r
o
v
id
e
s
u
ited
tr
ea
t
m
e
n
ts
,
c
u
t
d
o
w
n
o
n
n
ee
d
less
p
esti
cid
e
u
s
e,
in
cr
ea
s
e
cr
o
p
y
ield
s
,
an
d
s
a
v
e
p
r
o
d
u
ctio
n
e
x
p
en
s
es
b
y
e
m
p
lo
y
i
n
g
t
h
i
s
ar
ch
itect
u
r
e.
T
h
e
f
o
llo
w
i
n
g
ar
e
th
e
s
t
u
d
y
’
s
m
ain
co
n
tr
ib
u
tio
n
s
:
−
T
o
ac
cu
r
atel
y
id
en
ti
f
y
an
d
ca
t
eg
o
r
ize
d
is
ea
s
es o
f
to
m
ato
lea
v
es.
−
T
o
co
m
p
ar
e
d
ee
p
tr
an
s
f
er
lear
n
in
g
-
b
ased
ap
p
r
o
ac
h
es f
o
r
d
etec
tin
g
an
d
clas
s
i
f
y
i
n
g
to
m
ato
l
ea
f
d
is
ea
s
e
s
.
−
T
o
d
eter
m
i
n
e
th
e
m
o
s
t e
f
f
ec
ti
v
e
d
ee
p
tr
an
s
f
er
lear
n
in
g
m
o
d
el
f
o
r
id
en
ti
f
y
i
n
g
to
m
ato
lea
f
d
is
ea
s
es.
T
o
en
s
u
r
e
a
r
esp
ec
tab
le
cr
o
p
o
u
tp
u
t,
n
u
m
er
o
u
s
r
esear
ch
er
s
h
av
e
co
n
ce
n
tr
ated
o
n
d
ee
p
tr
an
s
f
er
lear
n
in
g
-
b
ased
s
y
s
te
m
s
to
au
t
o
m
a
te
tas
k
s
i
n
th
e
a
g
r
icu
l
tu
r
e
in
d
u
s
tr
y
,
i
n
clu
d
i
n
g
f
ield
m
o
n
it
o
r
in
g
,
p
lan
t
d
is
ea
s
e
d
iag
n
o
s
t
ics,
an
d
p
r
ed
ictio
n
.
T
h
an
g
ar
aj
et
a
l.
[
4
]
in
v
esti
g
a
ted
a
d
ee
p
co
n
v
o
lu
tio
n
a
l
n
e
u
r
al
n
et
w
o
r
k
(
C
NN
)
m
o
d
el
b
ased
o
n
tr
an
s
f
er
lear
n
in
g
to
id
e
n
ti
f
y
to
m
a
to
leaf
d
i
s
ea
s
e.
T
h
e
m
o
d
el
d
etec
ts
ill
n
es
s
in
to
m
ato
p
lan
ts
b
y
u
s
in
g
b
o
th
r
ea
l
-
ti
m
e
a
n
d
s
to
r
ed
p
ictu
r
es.
Fu
r
t
h
er
m
o
r
e,
r
o
o
t
m
ea
n
s
q
u
ar
e
p
r
o
p
ag
atio
n
(
R
MSp
r
o
p
)
o
p
tim
izer
s
,
s
to
ch
a
s
tic
g
r
ad
ie
n
t
d
escen
t
(
SGD)
,
an
d
ad
ap
tiv
e
m
o
m
e
n
t
esti
m
at
io
n
(
A
d
a
m
)
ar
e
em
p
lo
y
ed
to
ev
alu
a
te
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
t
h
e
p
r
o
p
o
s
ed
m
o
d
el.
T
h
e
ex
p
e
r
i
m
en
t
’
s
f
in
d
i
n
g
s
d
e
m
o
n
s
tr
ate
th
at
t
h
e
p
r
o
p
o
s
ed
m
o
d
el,
w
h
ich
m
ak
e
s
u
s
e
o
f
t
h
e
tr
an
s
f
er
lear
n
in
g
tech
n
iq
u
e,
ca
n
s
u
cc
ess
f
u
ll
y
cla
s
s
i
f
y
t
o
m
a
to
leaf
d
is
ea
s
es
au
to
m
at
icall
y
.
T
h
e
ac
cu
r
ac
y
o
f
th
e
A
d
a
m
o
p
ti
m
izer
is
h
i
g
h
er
th
a
n
th
at
o
f
S
GD
an
d
R
MSp
r
o
p
.
A
ttallah
[
5
]
p
r
esen
ted
a
m
et
h
o
d
f
o
r
th
e
au
to
m
at
ic
d
etec
tio
n
o
f
to
m
a
to
d
is
ea
s
es
f
r
o
m
leaf
i
m
a
g
es
u
s
in
g
t
h
r
ee
d
if
f
er
e
n
t
C
NNs
(
R
e
s
Net
-
1
8
,
Sh
u
f
f
leNe
t
,
an
d
Mo
b
ileNet)
.
Naiv
e
B
ay
e
s
(
NB
)
,
K
-
n
ea
r
est
n
eig
h
b
o
r
(
KNN)
,
d
ec
is
io
n
tr
ee
(
DT
)
,
lin
ea
r
d
is
cr
i
m
in
a
n
t
c
la
s
s
i
f
ier
(
L
D
A
)
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
,
an
d
q
u
ad
r
atic
d
is
cr
i
m
i
n
a
n
t
an
al
y
s
is
(
QD
A
)
ar
e
t
h
e
s
ix
cl
ass
i
f
ier
s
u
s
ed
in
to
m
ato
lea
f
d
is
ea
s
e
id
e
n
ti
f
icatio
n
.
T
h
e
r
esu
lts
d
e
m
o
n
s
tr
ate
th
a
t
th
e
KNN
a
n
d
SVM
o
b
tain
ed
t
h
e
h
i
g
h
est
ac
c
u
r
ac
y
o
f
9
9
.
9
2
%
an
d
9
9
.
9
0
%,
r
esp
ec
tiv
el
y
,
u
s
i
n
g
o
n
l
y
2
2
an
d
2
4
f
ea
t
u
r
es.
Kh
asa
w
n
e
h
et
a
l.
[
6
]
co
n
d
u
cted
an
u
p
d
ate
an
d
r
etr
ain
i
n
g
o
f
ele
v
en
d
ee
p
lear
n
in
g
m
o
d
els
to
id
en
ti
f
y
n
in
e
t
y
p
es
o
f
to
m
ato
d
is
ea
s
es
alo
n
g
w
it
h
h
ea
l
th
y
p
lan
t
s
.
T
h
e
r
esu
l
tin
g
te
n
class
e
s
w
er
e
ch
ar
ac
ter
ized
w
it
h
m
ea
n
v
al
u
e
s
o
f
9
9
.
4
%,
9
9
.
2
%,
9
9
.
1
%,
an
d
9
9
.
3
%
f
o
r
ac
cu
r
ac
y
,
F1
-
s
co
r
e,
r
ec
all
,
an
d
p
r
ec
is
io
n
,
r
esp
ec
ti
v
el
y
.
San
id
a
et
a
l.
[
7
]
s
u
g
g
ested
a
VGGN
et
-
b
ased
m
o
d
el
th
at
co
n
s
i
s
ts
o
f
t
w
o
i
n
ce
p
tio
n
b
lo
ck
s
an
d
I
m
ag
eNe
t
p
r
e
-
tr
ai
n
ed
o
n
it.
A
d
d
itio
n
all
y
,
t
h
e
m
o
d
el
tr
ain
in
g
p
r
o
ce
s
s
w
as
ex
ten
d
ed
to
in
clu
d
e
th
e
en
h
a
n
ce
d
ca
teg
o
r
ical
cr
o
s
s
-
e
n
tr
o
p
y
lo
s
s
f
u
n
ctio
n
f
o
r
th
e
m
u
lti
-
attr
ib
u
te
id
e
n
ti
f
i
ca
tio
n
p
r
o
b
lem
a
n
d
t
w
o
-
s
tag
e
tr
an
s
f
er
lear
n
i
n
g
.
A
b
b
as
et
a
l.
[
8
]
d
em
o
n
s
tr
ated
a
d
ee
p
lear
n
in
g
–
b
ased
m
eth
o
d
f
o
r
d
iag
n
o
s
i
n
g
to
m
ato
d
is
ea
s
es
b
y
g
en
er
at
in
g
s
y
n
t
h
etic
i
m
a
g
es
o
f
to
m
ato
leav
es
u
s
in
g
a
co
n
d
itio
n
al
g
en
er
ati
v
e
ad
v
er
s
ar
ial
n
et
w
o
r
k
(
C
-
G
A
N)
.
A
Den
s
eNe
t1
2
1
m
o
d
el,
w
h
ic
h
h
as
b
ee
n
tr
ai
n
ed
o
n
b
o
th
g
en
e
r
ated
an
d
r
ea
l
i
m
ag
e
s
u
s
i
n
g
t
r
an
s
f
er
lear
n
i
n
g
,
i
s
th
en
u
s
ed
to
clas
s
i
f
y
th
e
to
m
at
o
leaf
p
h
o
to
g
r
ap
h
s
i
n
to
te
n
d
is
ea
s
e
ca
teg
o
r
ies.
Alza
h
r
a
n
i
et
a
l.
[
9
]
in
v
esti
g
ated
th
e
ef
f
ec
t
iv
e
n
es
s
o
f
t
h
r
ee
d
ee
p
lear
n
in
g
–
b
ased
m
o
d
els
D
en
s
eNe
t1
6
9
,
R
esNe
t5
0
V2
,
a
n
d
th
e
tr
a
n
s
f
o
r
m
er
m
o
d
el
ViT
f
o
r
th
e
clas
s
i
f
icati
o
n
o
f
h
ea
lth
y
a
n
d
d
is
ea
s
ed
to
m
ato
p
lan
t
s
.
T
h
e
b
est
-
p
er
f
o
r
m
i
n
g
m
o
d
el
w
as
th
e
Den
s
eNe
t1
2
1
,
w
h
ic
h
ac
h
iev
ed
test
in
g
ac
c
u
r
ac
y
o
f
9
9
.
0
0
% a
n
d
tr
ain
i
n
g
ac
c
u
r
ac
y
o
f
9
9
.
8
8
%.
P
attn
aik
et
a
l.
[
1
0
]
h
av
e
d
ev
elo
p
ed
a
d
ee
p
C
NN
-
b
ased
s
y
s
t
e
m
f
o
r
to
m
ato
p
lan
t
p
est
clas
s
if
ica
tio
n
th
at
u
s
es
tr
an
s
f
er
lear
n
i
n
g
o
f
p
r
ev
io
u
s
l
y
lear
n
ed
d
ata.
T
h
e
s
tu
d
y
’
s
d
ataset,
w
h
ich
co
n
s
is
t
s
o
f
8
5
9
p
h
o
to
s
d
iv
id
ed
in
to
1
0
class
if
ica
tio
n
s
,
w
as
g
at
h
er
ed
f
r
o
m
in
te
r
n
et
s
o
u
r
ce
s
.
A
t
h
o
r
o
u
g
h
a
s
s
es
s
m
en
t
o
f
th
e
class
i
f
icatio
n
p
er
f
o
r
m
a
n
ce
o
f
1
5
p
r
e
-
tr
ain
ed
d
ee
p
C
NN
m
o
d
el
s
is
al
s
o
ca
r
r
ied
o
u
t
in
th
is
w
o
r
k
.
T
h
e
ex
p
er
i
m
e
n
tal
r
es
u
lts
s
h
o
w
ed
th
at
th
e
De
n
s
eNe
t1
6
9
o
u
tp
er
f
o
r
m
ed
w
it
h
ac
cu
r
ac
y
8
8
.
8
3
%.
D
is
ea
s
ed
leav
e
s
of
t
w
o
cr
o
p
s
(
g
r
ap
es
a
n
d
to
m
ato
es)
w
er
e
g
at
h
er
ed
an
d
p
r
o
d
u
ce
d
in
to
a
d
ataset
f
o
r
th
e
s
t
u
d
y
[
1
1
]
.
T
h
e
C
NN
-
b
ased
VGG1
6
m
o
d
el
is
s
u
b
j
ec
ted
to
tr
ain
in
g
,
te
s
ti
n
g
,
d
ataset
p
r
e
-
p
r
o
ce
s
s
in
g
,
an
d
d
ata
au
g
m
en
ta
tio
n
p
r
o
ce
d
u
r
es.
Saee
d
et
a
l.
[
1
2
]
h
av
e
clas
s
i
f
ied
i
m
a
g
es
o
f
h
ea
lth
y
a
n
d
d
is
ea
s
ed
to
m
ato
lea
v
es
u
s
in
g
t
w
o
p
r
e
-
tr
ain
ed
C
NN
s
,
I
n
ce
p
tio
n
V3
an
d
I
n
ce
p
tio
n
R
esNe
t
V2
,
in
o
r
d
er
to
d
iag
n
o
s
e
to
m
ato
lea
f
illn
e
s
s
e
s
.
T
h
e
t
w
o
m
o
d
el
s
w
er
e
tr
ain
ed
u
s
i
n
g
5
2
2
5
f
ield
-
r
ec
o
r
d
ed
im
a
g
es
a
n
d
an
o
p
en
-
s
o
u
r
ce
d
atab
ase
n
a
m
ed
P
lan
tVilla
g
e.
W
ith
an
ac
c
u
r
a
c
y
o
f
9
9
.
2
2
%
an
d
a
lo
s
s
o
f
0
.
0
3
,
th
e
m
o
s
t
n
o
te
w
o
r
th
y
o
u
tco
m
es
I
n
ce
p
tio
n
V3
an
d
I
n
ce
p
tio
n
R
esNet
V2
m
o
d
els
p
er
f
o
r
m
ed
th
e
b
est
w
i
th
d
r
o
p
o
u
t
r
ates
o
f
5
0
%
an
d
1
5
%
r
esp
ec
tiv
el
y
.
Hass
a
n
et
a
l.
[
1
3
]
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
Dee
p
tr
a
n
s
fer lea
r
n
in
g
b
a
s
ed
d
is
ea
s
e
d
etec
tio
n
a
n
d
cla
s
s
ifica
tio
n
o
f to
m
a
to
lea
ve
s
…
(
Mu
n
ir
a
A
kter La
ta
)
1355
e
m
p
lo
y
ed
f
o
u
r
d
ee
p
lear
n
i
n
g
m
o
d
els,
n
a
m
el
y
I
n
ce
p
ti
o
n
V3
,
I
n
ce
p
tio
n
R
esNetV2
,
Mo
b
ileNetV2
,
an
d
E
f
f
icien
tNetB
0
,
f
o
r
th
e
id
e
n
ti
f
icatio
n
o
f
p
la
n
t
ill
n
e
s
s
es
u
s
i
n
g
i
m
a
g
e
s
o
f
h
ea
lt
h
y
an
d
d
is
ea
s
ed
leav
es.
T
h
e
y
tr
ain
ed
an
d
ass
e
s
s
ed
t
h
e
m
o
d
el
u
s
i
n
g
th
e
5
3
,
4
0
7
im
a
g
e
s
-
all
s
h
o
t
i
n
a
lab
-
f
r
o
m
t
h
e
s
t
an
d
ar
d
P
lan
tVillag
e
d
ataset.
T
h
is
co
llectio
n
co
n
tain
s
i
m
ag
e
s
o
f
1
4
d
if
f
er
en
t
s
p
ec
ies
in
3
8
d
if
f
er
en
t
clas
s
es
o
f
b
o
th
h
ea
lth
y
an
d
d
is
ea
s
ed
leav
es.
Ag
ar
w
al
et
a
l.
[
1
4
]
em
p
lo
y
ed
a
C
NN
-
b
ase
d
ap
p
r
o
ac
h
f
o
r
th
e
id
en
tific
at
io
n
o
f
to
m
ato
leaf
d
is
ea
s
e.
W
ith
v
ar
y
i
n
g
n
u
m
b
er
s
o
f
f
i
lter
s
,
th
is
m
o
d
el
co
n
s
is
t
s
o
f
th
r
ee
co
n
v
o
l
u
tio
n
an
d
m
a
x
p
o
o
lin
g
la
y
er
s
.
T
h
e
n
o
v
elt
y
o
f
o
u
r
s
tu
d
y
i
s
th
e
in
teg
r
atio
n
o
f
d
ee
p
tr
an
s
f
er
lear
n
in
g
an
d
o
cc
lu
s
io
n
s
en
s
iti
v
it
y
.
T
h
is
f
r
a
m
e
w
o
r
k
’
s
in
ter
p
r
etab
ilit
y
a
n
d
r
o
b
u
s
tn
e
s
s
ar
e
i
m
p
r
o
v
ed
b
y
in
c
lu
d
i
n
g
o
cc
l
u
s
io
n
s
e
n
s
iti
v
it
y
,
p
ar
t
icu
lar
l
y
in
in
tr
icate
a
g
r
icu
lt
u
r
al
en
v
ir
o
n
m
en
ts
w
h
er
e
ill
n
ess
s
i
g
n
s
ten
d
t
o
b
e
o
b
s
cu
r
ed
o
r
s
u
b
tle.
2.
M
E
T
H
O
D
2
.1
.
P
r
o
po
s
ed
s
y
s
t
e
m
T
h
e
s
tu
d
y
co
n
s
i
s
ted
o
f
f
i
v
e
m
o
d
u
les:
i
m
a
g
e
ac
q
u
is
it
io
n
,
im
a
g
e
p
r
e
-
p
r
o
ce
s
s
i
n
g
,
s
e
g
m
en
tatio
n
,
co
n
ta
m
i
n
ated
lea
f
id
e
n
ti
f
icatio
n
,
an
d
d
i
s
ea
s
e
ca
te
g
o
r
izatio
n
.
T
h
e
ev
alu
at
io
n
a
n
d
co
m
p
ar
is
o
n
o
f
clas
s
if
ier
s
ar
e
also
p
er
f
o
r
m
ed
i
n
th
is
s
t
u
d
y
.
F
ig
u
r
e
1
s
h
o
w
s
t
h
e
ca
teg
o
r
izati
o
n
s
ch
e
m
e
f
o
r
to
m
ato
leaf
d
is
ea
s
es.
Fig
u
r
e
1
.
C
ateg
o
r
izatio
n
s
ch
e
m
e
f
o
r
to
m
ato
lea
f
d
is
ea
s
e
s
2
.
1
.
1
.
I
m
a
g
e
a
cquis
it
io
n
I
n
itiall
y
,
to
m
a
to
leaf
i
m
a
g
es
a
r
e
o
b
tain
ed
f
r
o
m
t
h
e
P
lan
tVil
l
ag
e
d
ataset
f
o
r
th
i
s
s
t
u
d
y
.
I
n
t
o
tal,
th
er
e
ar
e
1
8
,
1
6
0
to
m
ato
leaf
i
m
ag
e
s
in
th
i
s
d
ataset
[
1
5
]
.
T
h
is
im
ag
e
d
ataset
h
as
te
n
class
e
s
an
d
is
s
ep
ar
ated
in
to
t
w
o
p
ar
ts
:
d
is
ea
s
ed
(
9
class
e
s
)
an
d
h
ea
lt
h
y
(
1
class
)
.
T
h
e
s
a
m
p
le
i
m
a
g
es
f
o
r
ea
ch
ca
teg
o
r
y
ar
e
s
h
o
w
n
i
n
Fig
u
r
e
s
2
(
a)
to
(
j
)
,
w
h
er
e
Fi
g
u
r
es
2
(
a)
is
h
ea
lt
h
y
,
(
b
)
is
b
a
cter
ial
s
p
o
t,
(
c)
is
ea
r
l
y
b
li
g
h
t,
(
d
)
is
late
b
li
g
h
t,
(
e)
is
leaf
m
o
ld
,
(
f
)
is
s
ep
to
r
ia
leaf
s
p
o
t,
(
g
)
is
s
p
id
er
m
ite
s
(
h
)
is
tar
g
et
s
p
o
t,
(
i)
is
to
m
ato
m
o
s
a
ic
v
ir
u
s
,
an
d
(
j
)
is
y
el
lo
w
lea
f
cu
r
l
v
ir
u
s
.
T
h
e
d
escr
ip
tio
n
o
f
th
e
d
ataset
i
s
s
h
o
w
n
in
T
ab
le
1
.
T
h
e
tr
ain
in
g
d
ataset
co
n
ta
in
s
8
0
% o
f
th
e
ac
q
u
ir
ed
d
ata,
w
h
il
e
th
e
test
i
n
g
d
ataset
f
o
r
class
i
f
i
ca
tio
n
tas
k
s
co
n
tain
s
t
h
e
r
e
m
a
in
i
n
g
2
0
%.
(
a)
(
b
)
(
c)
(
d
)
(
e)
(
f
)
(
g
)
(
h
)
(
i)
(
j
)
Fig
u
r
e
2
.
Sa
m
p
le
i
m
ag
e
s
(
u
p
p
er
r
o
w
,
f
r
o
m
le
f
t to
r
ig
h
t)
; (
a)
h
ea
lt
h
y
,
(
b
)
b
ac
ter
ial
s
p
o
t,
(
c)
ea
r
l
y
b
lig
h
t,
(
d
)
late
b
lig
h
t,
(
e)
leaf
m
o
ld
,
(
f
)
s
ep
to
r
ia
leaf
s
p
o
t,
(
g
)
s
p
id
er
m
ite
s
(
h
)
tar
g
et
s
p
o
t,
(
i)
to
m
at
o
m
o
s
a
ic
v
ir
u
s
,
an
d
(
j
)
y
ello
w
lea
f
c
u
r
l v
ir
u
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
23
,
No
.
5
,
Octo
b
e
r
20
25
:
1
3
5
3
-
1
362
1356
T
ab
le
1
.
Data
s
et
d
escr
ip
tio
n
I
mag
e
t
y
p
e
A
mo
u
n
t
H
e
a
l
t
h
y
1
,
5
9
1
B
a
c
t
e
r
i
a
l
s
p
o
t
2
,
1
2
7
Ea
r
l
y
b
l
i
g
h
t
1
,
0
0
0
L
a
t
e
b
l
i
g
h
t
1
,
9
0
9
L
e
a
f
mo
l
d
9
5
2
S
e
p
t
o
r
i
a
l
e
a
f
sp
o
t
1
,
7
7
1
S
p
i
d
e
r
mi
t
e
s
1
,
6
7
6
T
a
r
g
e
t
sp
o
t
1
,
4
0
4
T
o
mat
o
mo
sai
c
v
i
r
u
s
3
7
3
Y
e
l
l
o
w
l
e
a
f
c
u
r
l
v
i
r
u
s
5
,
3
5
7
Ov
e
r
a
l
l
1
8
,
1
6
0
2
.
1
.
2
.
P
re
-
pro
ce
s
s
ing
T
h
e
in
p
u
t
i
m
ag
e
is
p
r
ep
r
o
ce
s
s
ed
b
y
co
n
v
er
ti
n
g
it
f
r
o
m
th
e
r
ed
-
g
r
ee
n
-
b
l
u
e
(
R
GB
)
co
lo
r
s
p
ac
e
to
a
s
in
g
le
-
ch
a
n
n
el
g
r
a
y
s
ca
le
i
m
a
g
e
in
o
r
d
er
to
s
im
p
li
f
y
ca
lcu
la
t
io
n
an
d
m
a
k
e
f
u
r
t
h
er
an
al
y
s
i
s
ea
s
ier
.
I
n
o
r
d
er
to
i
m
p
r
o
v
e
i
m
a
g
e
co
n
tr
as
t,
th
i
s
s
tu
d
y
u
s
e
s
ad
ap
tiv
e
h
i
s
to
g
r
a
m
eq
u
aliza
tio
n
(
AHE
)
an
d
m
ed
ian
f
ilter
i
n
g
.
I
n
i
m
a
g
e
p
r
e
-
p
r
o
ce
s
s
in
g
,
A
HE
a
n
d
m
ed
ian
f
ilter
ca
n
o
f
f
er
a
n
u
m
b
er
o
f
ad
v
a
n
ta
g
es
t
h
at
i
m
p
r
o
v
e
i
m
ag
e
q
u
alit
y
f
o
r
s
u
b
s
eq
u
e
n
t
a
n
al
y
s
is
.
W
h
e
n
co
m
b
i
n
ed
,
th
e
y
p
r
o
v
id
e
a
p
o
ten
t
m
eth
o
d
f
o
r
lo
ca
l
co
n
tr
a
s
t
en
h
a
n
ce
m
en
t
an
d
n
o
is
e
r
ed
u
ctio
n
,
g
u
ar
a
n
teei
n
g
th
at
th
e
i
m
a
g
e
is
clea
r
an
d
co
n
tain
s
d
is
ti
n
ct
f
ea
tu
r
es,
w
h
ich
m
a
k
es
i
t
b
etter
s
u
ited
f
o
r
f
u
r
th
er
p
r
o
ce
s
s
i
n
g
task
s
li
k
e
s
e
g
m
e
n
tat
io
n
a
n
d
im
ag
e
c
lass
if
ica
tio
n
.
Fi
g
u
r
e
3
s
h
o
w
s
t
h
e
o
r
ig
i
n
a
l
i
m
a
g
e
an
d
p
r
e
-
p
r
o
ce
s
s
ed
i
m
a
g
es
,
w
h
er
e
Fi
g
u
r
es
3
(
a)
is
o
r
ig
i
n
al
i
m
ag
e,
(
b
)
is
R
GB
to
g
r
a
y
i
m
a
g
e,
(
c)
is
A
HE
i
m
a
g
e,
an
d
(
d
)
is
A
HE
+
m
ed
ian
f
i
lter
i
m
a
g
e
.
(
a
)
(
b
)
(
c
)
(
d
)
Fig
u
r
e
3
.
T
h
e
o
r
ig
in
al
i
m
a
g
e
a
n
d
p
r
e
-
p
r
o
ce
s
s
ed
i
m
ag
e
s
;
(
a
)
o
r
ig
in
a
l i
m
a
g
e,
(
b
)
R
GB
to
g
r
ay
i
m
a
g
e,
(
c
)
AHE
i
m
a
g
e,
an
d
(
d
)
A
HE
+
m
ed
ia
n
f
ilter
i
m
ag
e
2
.
1
.
3
.
Seg
m
ent
a
t
io
n
I
n
th
is
s
tu
d
y
,
o
cc
lu
s
io
n
s
en
s
it
iv
it
y
is
e
m
p
lo
y
ed
as
a
s
eg
m
e
n
tatio
n
tec
h
n
iq
u
e.
A
p
er
tu
r
b
atio
n
-
b
ased
in
ter
p
r
etab
ilit
y
tech
n
iq
u
e
ca
lled
o
cc
lu
s
io
n
s
en
s
i
tiv
it
y
is
u
s
ed
to
d
eter
m
in
e
w
h
ich
ar
ea
s
o
f
an
in
p
u
t
i
m
ag
e
h
a
v
e
th
e
m
o
s
t
i
m
p
ac
t
o
n
a
d
ee
p
n
e
u
r
al
n
e
t
w
o
r
k
’
s
cla
s
s
i
f
ica
tio
n
d
ec
is
io
n
.
Us
in
g
t
h
i
s
m
eth
o
d
,
v
ar
io
u
s
ar
ea
s
o
f
th
e
in
p
u
t
i
m
a
g
e
ar
e
co
v
er
ed
w
it
h
an
o
cc
l
u
s
io
n
m
a
s
k
,
a
n
d
t
h
e
ch
an
g
e
in
th
e
m
o
d
el
’
s
o
u
tp
u
t
s
co
r
e
is
r
ec
o
r
d
ed
.
T
h
is
ap
p
r
o
ac
h
p
r
o
d
u
ce
s
a
s
en
s
iti
v
it
y
m
ap
th
at
s
h
o
w
s
w
h
ich
p
ar
ts
o
f
th
e
i
m
ag
e
ar
e
m
o
s
t
i
m
p
o
r
tan
t
f
o
r
class
i
f
icatio
n
.
T
h
e
d
ef
in
itio
n
o
f
th
e
o
cc
l
u
s
io
n
s
e
n
s
iti
v
it
y
m
ap
is
as (
1
)
:
∅
=
(
)
−
(
[
=
̅
]
)
(
1
)
w
h
er
e,
is
t
h
e
u
n
n
o
r
m
alize
d
cl
ass
s
co
r
e,
is
t
h
e
r
ep
lace
m
en
t
o
f
o
n
e
f
ea
t
u
r
e
w
it
h
b
aseli
n
e
̅
[
1
6
]
.
Fig
u
r
e
4
s
h
o
w
s
t
h
e
s
a
m
p
le
o
cc
lu
s
io
n
s
e
n
s
it
iv
i
t
y
m
ap
.
Fig
u
r
e
4
.
Sa
m
p
le
o
cc
lu
s
io
n
s
e
n
s
it
iv
i
t
y
m
ap
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
Dee
p
tr
a
n
s
fer lea
r
n
in
g
b
a
s
ed
d
is
ea
s
e
d
etec
tio
n
a
n
d
cla
s
s
ifica
tio
n
o
f to
m
a
to
lea
ve
s
…
(
Mu
n
ir
a
A
kter La
ta
)
1357
2
.
1
.
4
.
Dee
p t
ra
ns
f
er
lea
rning
cla
s
s
if
ica
t
io
n
m
et
ho
ds
T
h
e
d
is
ea
s
es
o
f
to
m
ato
lea
v
es
ar
e
ca
teg
o
r
ized
in
th
i
s
s
t
u
d
y
u
s
i
n
g
s
ev
e
n
d
ee
p
tr
an
s
f
er
lear
n
in
g
class
i
f
icatio
n
tech
n
iq
u
e
s
.
T
h
e
s
e
s
ev
e
n
ca
teg
o
r
izatio
n
tech
n
i
q
u
es
in
cl
u
d
e
I
n
ce
p
tio
n
V3
,
R
esNet
-
1
8
,
R
esNet
-
5
0
,
VGG
-
1
6
,
VGG
-
1
9
,
Go
o
g
L
eNe
t,
a
n
d
A
lex
Net.
T
h
e
f
u
ll
y
co
n
n
ec
ted
la
y
er
a
n
d
th
e
f
i
n
a
l
class
i
f
icat
io
n
la
y
er
ar
e
ad
j
u
s
ted
to
m
atc
h
th
e
tar
g
e
t ta
s
k
’
s
c
lass
s
tr
u
ct
u
r
e
af
ter
th
e
p
r
e
-
tr
ain
ed
m
o
d
els
h
av
e
b
ee
n
lo
ad
ed
.
T
h
e
in
p
u
t
i
m
a
g
es
ar
e
r
e
s
ized
b
ef
o
r
e
class
i
f
icatio
n
to
m
a
k
e
s
u
r
e
it
co
m
p
l
ied
w
it
h
t
h
e
cla
s
s
if
ier
m
o
d
el
’
s
i
n
p
u
t
d
i
m
en
s
io
n
al
it
y
r
eq
u
ir
e
m
en
t
s
.
Deta
ils
o
f
th
e
in
p
u
t
i
m
a
g
e
s
iz
e
in
ac
co
r
d
an
ce
w
it
h
th
e
m
o
d
el
’
s
s
p
ec
i
f
icatio
n
s
ar
e
d
is
p
lay
ed
i
n
T
ab
le
2
.
T
ab
le
2
.
Deta
ils
o
f
th
e
i
n
p
u
t i
m
ag
e
s
ize
i
n
ac
co
r
d
an
ce
w
it
h
th
e
m
o
d
el
’
s
s
p
ec
i
f
icati
o
n
s
M
o
d
e
l
I
n
p
u
t
i
m
a
g
e
si
z
e
I
n
c
e
p
t
i
o
n
V
3
2
9
9
*
2
9
9
R
e
sN
e
t
-
18
2
2
4
*
2
2
4
R
e
sN
e
t
-
50
2
2
4
*
2
2
4
VGG
-
16
2
2
4
*
2
2
4
VGG
-
19
2
2
4
*
2
2
4
G
o
o
g
L
e
N
e
t
2
2
4
*
2
2
4
A
l
e
x
N
e
t
2
2
7
*
2
2
7
a.
I
n
c
e
p
ti
o
n
V
3
:
t
h
e
I
m
a
g
e
N
e
t
d
a
t
a
s
et
c
o
n
t
ain
s
m
o
r
e
th
an
a
m
il
l
i
o
n
im
ag
e
s
th
at
h
av
e
b
e
en
u
s
e
d
t
o
t
r
ai
n
I
n
c
e
p
ti
o
n
V
3
[
1
7
]
,
a
4
8
-
l
ay
e
r
C
N
N
.
T
h
e
1
,
0
0
0
o
b
je
c
t
c
a
t
e
g
o
r
i
e
s
th
a
t
t
h
i
s
n
etw
o
r
k
ca
n
cl
as
s
if
y
p
h
o
t
o
g
r
a
p
h
s
i
n
t
o
c
o
n
t
a
in
a
w
i
d
e
v
a
r
i
ety
o
f
o
b
je
c
t
s
,
in
c
lu
d
in
g
p
en
c
i
ls
,
m
o
u
s
e
,
k
ey
b
o
a
r
d
s
,
an
d
an
im
a
ls
.
R
i
ch
f
ea
tu
r
e
r
e
p
r
e
s
en
t
a
ti
o
n
s
th
a
t
m
ay
b
e
a
p
p
l
i
e
d
t
o
a
b
r
o
a
d
r
a
n
g
e
o
f
im
ag
e
s
h
av
e
th
u
s
b
e
e
n
l
ea
r
n
e
d
.
I
n
th
e
f
i
r
s
t
s
t
ag
e
,
t
h
e
m
o
d
el
a
r
ch
it
e
c
tu
r
e
c
o
n
c
en
tr
a
t
e
s
o
n
e
x
t
r
a
c
ti
n
g
g
e
n
e
r
i
c
f
e
atu
r
es
f
r
o
m
th
e
i
n
p
u
t p
h
o
t
o
s
,
an
d
i
n
th
e
s
e
c
o
n
d
s
t
ag
e
,
th
e
f
ea
tu
r
e
s
a
r
e
u
s
e
d
f
o
r
im
ag
e
c
a
t
eg
o
r
i
z
at
i
o
n
.
I
n
c
e
p
t
i
o
n
V
3
h
as
2
9
.
3
m
il
l
i
o
n
p
a
r
am
ete
r
s
in
t
o
t
al
.
b.
R
esNet
-
1
8
:
s
i
x
teen
co
n
v
o
l
u
ti
o
n
al
la
y
er
s
a
n
d
t
w
o
f
u
ll
y
co
n
n
ec
ted
la
y
er
s
m
ak
e
u
p
th
e
1
8
-
la
y
er
C
NN
k
n
o
w
n
a
s
R
esNet
-
1
8
[
1
8
]
.
T
h
e
n
et
w
o
r
k
w
as
p
r
e
-
tr
ain
ed
u
s
i
n
g
m
o
r
e
th
an
a
m
illi
o
n
p
h
o
to
s
f
r
o
m
t
h
e
I
m
ag
eNe
t
d
ataset.
A
n
i
m
al
s
,
k
e
y
b
o
ar
d
s
,
m
ic
e,
p
en
s
,
an
d
a
v
ar
iet
y
o
f
o
th
er
o
b
j
ec
ts
m
a
y
al
l
b
e
class
if
ied
in
to
1
0
0
0
o
b
j
ec
t
ca
teg
o
r
ies
b
y
t
h
e
p
r
etr
ain
ed
m
o
d
el.
T
h
e
n
et
w
o
r
k
h
as
th
u
s
ac
q
u
ir
ed
e
x
ten
s
i
v
e
f
ea
t
u
r
e
r
ep
r
esen
tatio
n
s
t
h
at
m
a
y
b
e
u
s
ed
to
a
w
id
e
r
an
g
e
o
f
p
ict
u
r
es.
c.
R
esNet
-
5
0
:
R
esNet
-
5
0
[
1
8
]
is
a
50
-
la
y
er
d
ee
p
C
NN.
T
h
e
f
o
u
r
m
ai
n
p
ar
ts
o
f
th
e
R
esNet
-
5
0
ar
ch
itectu
r
e
ar
e
th
e
co
n
v
o
l
u
tio
n
al
la
y
er
s
,
id
en
tit
y
b
lo
ck
,
co
n
v
o
lu
tio
n
al
b
lo
ck
,
an
d
f
u
ll
y
co
n
n
ec
te
d
lay
er
s
.
T
h
e
co
n
v
o
lu
tio
n
al
la
y
er
s
ca
p
tu
r
e
f
ea
tu
r
es
f
r
o
m
t
h
e
i
n
p
u
t
i
m
ag
e,
th
e
id
e
n
tit
y
b
lo
ck
a
n
d
co
n
v
o
lu
tio
n
a
l
b
lo
ck
p
r
o
ce
s
s
an
d
tr
an
s
f
o
r
m
t
h
o
s
e
f
ea
tu
r
es,
an
d
th
e
f
u
ll
y
co
n
n
ec
t
ed
lay
er
s
clas
s
i
f
y
t
h
e
f
ea
t
u
r
es
.
Fo
r
a
w
id
e
r
an
g
e
o
f
i
m
a
g
e
cla
s
s
i
f
icatio
n
tas
k
s
,
s
u
c
h
as
o
b
j
ec
t
d
etec
tio
n
,
m
ed
ical
i
m
a
g
e
a
n
al
y
s
is
,
an
d
f
ac
ial
r
ec
o
g
n
itio
n
,
R
esNet
-
5
0
is
a
p
o
w
er
f
u
l
m
o
d
el.
I
t
w
as
tr
ai
n
ed
o
n
th
e
ex
ten
s
iv
e
I
m
ag
e
N
et
d
ataset
an
d
ac
h
iev
ed
a
n
er
r
o
r
r
ate
co
m
p
ar
ab
le
to
h
u
m
an
p
er
f
o
r
m
a
n
ce
.
I
t
h
as
al
s
o
b
ee
n
ap
p
lied
as
a
f
ea
tu
r
e
ex
tr
ac
to
r
f
o
r
o
th
er
ap
p
licatio
n
s
,
lik
e
s
e
m
an
tic
s
e
g
m
en
tat
io
n
an
d
o
b
j
e
ct
d
etec
tio
n
.
d.
VGG
-
1
6
:
t
h
ir
tee
n
co
n
v
o
l
u
tio
n
al
lay
er
s
an
d
th
r
ee
f
u
l
l
y
co
n
n
e
cted
lay
er
s
m
a
k
e
u
p
th
e
1
6
-
la
y
er
d
ee
p
C
NN
k
n
o
w
n
as
VG
G
-
1
6
[
1
9
]
.
I
t
w
o
r
k
s
w
ell
b
ec
au
s
e
o
f
its
r
e
m
a
r
k
ab
le
d
ep
th
.
VGG
-
1
6
is
r
en
o
w
n
ed
f
o
r
its
ea
s
e
o
f
u
s
e,
ef
f
ec
tiv
e
n
e
s
s
,
an
d
ex
ce
p
tio
n
al
p
er
f
o
r
m
a
n
ce
o
n
a
v
ar
iety
o
f
co
m
p
u
ter
v
i
s
io
n
ap
p
licatio
n
s
,
s
u
c
h
as
i
m
a
g
e
ca
teg
o
r
izatio
n
an
d
o
b
j
ec
t
r
ec
o
g
n
itio
n
.
T
h
e
m
o
d
el
ar
ch
itectu
r
e
is
m
ad
e
u
p
o
f
m
a
x
-
p
o
o
lin
g
la
y
er
s
a
f
ter
a
s
eq
u
e
n
ce
o
f
co
n
v
o
l
u
tio
n
a
l
la
y
er
s
w
it
h
g
r
ad
u
all
y
d
ee
p
er
la
y
er
s
.
T
h
e
n
et
wo
r
k
ca
n
lear
n
in
tr
icate
h
ier
ar
ch
ica
l
r
ep
r
esen
ta
tio
n
s
o
f
v
is
u
al
f
ea
tu
r
es
b
ec
au
s
e
o
f
th
is
d
esig
n
,
p
r
o
d
u
cin
g
p
r
ed
ictio
n
s
th
at
ar
e
p
r
ec
is
e
an
d
d
ep
en
d
ab
le.
e.
VGG
-
1
9
:
t
h
e
VG
G
-
1
9
[
1
9
]
is
a
d
ee
p
C
NN
o
f
1
9
w
ei
g
h
t
la
y
er
s
,
1
6
co
n
v
o
lu
tio
n
al
la
y
er
s
,
an
d
3
f
u
ll
y
co
n
n
ec
ted
la
y
er
s
.
T
h
e
VGG
-
1
9
m
o
d
el
(
also
k
n
o
w
n
as
VGG
Net
-
1
9
)
is
s
i
m
ilar
to
th
e
VGG
-
1
6
m
o
d
el
in
its
b
asic id
ea
,
ex
ce
p
t th
at
i
t su
p
p
o
r
ts
1
9
lay
er
s
.
f.
Go
o
g
L
eNe
t
:
w
ith
o
v
er
s
e
v
en
m
illi
o
n
p
ar
a
m
eter
s
Go
o
g
L
e
Net
[
2
0
]
,
h
as
n
i
n
e
in
ce
p
tio
n
m
o
d
u
les,
f
o
u
r
co
n
v
o
lu
tio
n
al
la
y
er
s
,
f
o
u
r
m
a
x
-
p
o
o
lin
g
la
y
er
s
,
t
h
r
ee
av
er
a
g
e
p
o
o
lin
g
la
y
er
s
,
f
iv
e
f
u
ll
y
co
n
n
ec
ted
la
y
er
s
,
an
d
th
r
ee
So
f
tMa
x
la
y
er
s
f
o
r
th
e
p
r
in
cip
al
an
d
au
x
iliar
y
class
i
f
ier
s
.
E
ac
h
co
n
v
o
lu
tio
n
al
lay
er
i
n
th
e
n
et
w
o
r
k
u
s
e
s
R
e
L
U
ac
t
iv
ati
o
n
f
u
n
c
tio
n
s
,
a
n
d
th
e
f
u
ll
y
lin
k
ed
la
y
er
s
u
s
e
d
r
o
p
o
u
t
r
eg
u
lar
izatio
n
.
Go
o
g
L
eNe
t
’
s
ef
f
ec
ti
v
e
tr
ad
e
-
o
f
f
b
et
w
ee
n
co
m
p
u
tatio
n
a
l
co
s
t
an
d
p
ar
am
eter
co
u
n
t
m
ak
es
it
id
ea
l
f
o
r
r
ea
l
-
ti
m
e
ap
p
licatio
n
s
a
n
d
d
ep
lo
y
m
e
n
t o
n
d
ev
ices
w
it
h
li
m
it
ed
r
eso
u
r
ce
s.
g.
A
le
x
Net:
A
le
x
Net
[
2
1
]
h
a
s
ei
g
h
t
la
y
er
s
i
n
i
ts
d
es
ig
n
,
t
h
e
la
s
t
th
r
ee
o
f
w
h
ic
h
ar
e
f
u
ll
y
co
n
n
ec
ted
,
an
d
th
e
f
ir
s
t
f
iv
e
o
f
w
h
ic
h
ar
e
co
n
v
o
lu
tio
n
al
la
y
er
s
.
Af
ter
th
e
f
ir
s
t
tw
o
co
n
v
o
l
u
tio
n
a
l
la
y
er
s
,
o
v
er
lap
p
in
g
m
a
x
-
p
o
o
lin
g
la
y
er
s
ar
e
ap
p
lied
to
o
p
tim
ize
f
ea
tu
r
e
ex
tr
ac
tio
n
.
T
h
e
th
ir
d
,
f
o
u
r
th
,
an
d
f
i
f
t
h
co
n
v
o
lu
tio
n
al
la
y
er
s
’
o
u
tp
u
ts
ar
e
d
ir
ec
tl
y
co
n
n
ec
ted
to
th
e
f
u
ll
y
co
n
n
ec
ted
la
y
er
s
.
E
v
er
y
o
u
tp
u
t
tr
an
s
m
it
ted
th
r
o
u
g
h
a
R
eL
U
n
o
n
-
li
n
ea
r
ac
ti
v
atio
n
f
u
n
ct
io
n
co
m
es
f
r
o
m
th
e
co
n
v
o
lu
tio
n
al
an
d
f
u
ll
y
co
n
n
ec
ted
la
y
er
s
.
W
h
en
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
23
,
No
.
5
,
Octo
b
e
r
20
25
:
1
3
5
3
-
1
362
1358
p
air
ed
w
it
h
a
So
f
tMa
x
ac
ti
v
ati
o
n
f
u
n
ctio
n
,
th
e
f
i
n
al
o
u
tp
u
t
la
y
er
cr
ea
tes
a
p
r
o
b
ab
ilit
y
d
i
s
tr
i
b
u
tio
n
ac
r
o
s
s
1
0
0
0
class
lab
els.
3.
RE
SU
L
T
S
AND
D
I
SCU
SS
I
O
N
T
h
is
s
t
u
d
y
e
v
al
u
ated
t
h
e
m
o
s
t
ad
v
an
ce
d
p
r
e
-
tr
ain
ed
tr
a
n
s
f
e
r
lear
n
in
g
cla
s
s
i
f
icat
io
n
m
o
d
els
f
o
r
th
e
task
o
f
class
if
y
i
n
g
d
is
ea
s
es
f
r
o
m
t
h
e
i
m
a
g
es
d
ataset
o
f
to
m
ato
cr
o
p
.
Fo
u
r
d
if
f
er
en
t
ass
ess
m
en
t
cr
iter
ia
ar
e
u
s
ed
to
r
eg
u
la
te
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
s
e
m
o
d
els
:
p
r
ec
is
io
n
,
r
ec
all,
F1
-
s
co
r
e,
an
d
ac
cu
r
ac
y
.
T
h
ese
f
o
u
r
m
etr
ics ar
e
d
eter
m
i
n
ed
b
y
(
2
)
-
(
5
)
:
=
+
(
2
)
=
+
(
3
)
1
−
=
2
∗
∗
+
(
4
)
=
+
+
+
+
(
5
)
w
h
er
e,
th
e
s
y
m
b
o
ls
f
o
r
f
alse
p
o
s
itiv
e
s
,
tr
u
e
p
o
s
itiv
es,
f
al
s
e
n
eg
ati
v
es,
an
d
tr
u
e
n
eg
ati
v
e
s
ar
e
v
ie
w
ed
w
it
h
th
e
s
h
o
r
t te
r
m
s
f
al
s
e
p
o
s
iti
v
e
(
FP
)
,
tr
u
e
p
o
s
itiv
e
(
TP
)
,
f
alse n
e
g
a
tiv
e
(
FN
)
,
an
d
tr
u
e
n
eg
a
tiv
e
(
TN
)
,
r
esp
ec
tiv
el
y
.
T
h
e
v
ar
io
u
s
class
if
ica
tio
n
o
u
tco
m
es
o
f
s
e
v
e
n
d
ee
p
tr
an
s
f
er
lear
n
in
g
class
i
f
ier
s
ar
e
d
is
p
la
y
ed
i
n
T
ab
le
3
.
W
e
u
s
e
th
e
“
r
o
u
n
d
h
a
lf
u
p
”
r
u
le
to
ca
lc
u
late
t
h
e
av
e
r
ag
e
p
r
ec
is
io
n
,
r
ec
all,
F1
-
s
co
r
e,
an
d
ac
cu
r
ac
y
.
T
ab
le
3
.
Fin
d
in
g
s
f
o
r
th
e
e
v
al
u
ated
d
ee
p
tr
an
s
f
er
lear
n
in
g
al
g
o
r
ith
m
s
M
o
d
e
l
P
r
e
c
i
si
o
n
R
e
c
a
l
l
F1
-
sco
r
e
A
c
c
u
r
a
c
y
I
n
c
e
p
t
i
o
n
V
3
0
.
7
0
0
.
7
1
0
.
7
0
0
.
7
7
R
e
sN
e
t
-
18
0
.
7
1
0
.
7
2
0
.
7
2
0
.
7
7
R
e
sN
e
t
-
50
0
.
8
2
0
.
8
1
0
.
8
1
0
.
8
5
VGG
-
16
0
.
7
6
0
.
7
7
0
.
7
6
0
.
8
1
VGG
-
19
0
.
7
5
0
.
7
8
0
.
7
6
0
.
8
1
G
o
o
g
L
e
N
e
t
0
.
6
7
0
.
6
8
0
.
6
1
0
.
7
4
A
l
e
x
N
e
t
0
.
7
6
0
.
7
7
0
.
7
6
0
.
8
1
R
esNet
-
5
0
y
ield
s
th
e
h
i
g
h
est
av
er
ag
e
p
r
ec
is
io
n
(
PS
)
,
r
ec
all
(
R
C
A
)
,
F1
-
s
co
r
e
(
F1
-
S)
,
a
n
d
a
cc
u
r
ac
y
(
AC
C
)
,
w
it
h
v
a
lu
e
s
o
f
0
.
8
2
,
0
.
8
1
,
0
.
8
1
,
an
d
0
.
8
5
,
r
esp
ec
tiv
el
y
,
w
h
ile
Go
o
g
L
eNe
t
y
ie
ld
s
th
e
lo
w
est
av
er
ag
e
p
r
ec
is
io
n
,
r
ec
all,
F1
-
s
co
r
e,
an
d
ac
cu
r
ac
y
,
w
it
h
v
alu
e
s
o
f
0
.
6
7
,
0
.
6
8
,
0
.
6
1
,
an
d
0
.
7
4
,
r
esp
ec
tiv
el
y
.
T
ab
le
4
d
is
p
lay
s
th
e
co
n
f
u
s
i
o
n
m
a
tr
ix
o
f
R
esNet
-
5
0
,
th
e
m
o
d
el
t
h
at
p
r
o
d
u
ce
d
th
e
b
est
r
es
u
lt
s
ac
co
r
d
in
g
to
th
e
p
er
f
o
r
m
an
ce
m
etr
ics.
T
ab
le
5
d
is
p
lay
s
t
h
e
r
esu
lt
s
o
f
th
e
R
esNe
t
-
5
0
m
o
d
el
u
s
ed
to
d
eter
m
i
n
e
th
e
p
er
f
o
r
m
a
n
ce
m
e
tr
ics f
o
r
ea
ch
class
.
Fig
u
r
e
5
s
h
o
w
s
t
h
e
p
r
ec
is
io
n
r
ec
all
(
P
R
)
cu
r
v
e
o
f
R
esNet
-
50.
T
ab
le
4
.
C
o
n
f
u
s
io
n
m
atr
i
x
o
f
R
esNet
-
5
0
m
o
d
el
1
2
3
4
5
6
7
8
9
10
1
3
8
9
6
0
2
2
9
0
0
0
24
2
8
1
2
4
0
29
9
10
6
4
0
7
3
0
2
3
0
5
1
2
3
3
10
0
0
4
7
28
0
3
0
9
13
15
1
8
1
10
5
1
10
1
9
1
4
2
8
3
7
4
5
6
9
4
3
15
7
2
9
4
8
19
4
4
7
2
3
4
0
4
5
2
6
1
31
2
8
8
0
4
5
8
2
10
39
1
9
6
0
8
9
0
0
0
1
3
0
3
1
63
2
10
9
19
0
8
6
0
11
5
1
1
0
0
3
N
o
t
e
:
(
1
)
b
a
c
t
e
r
i
a
l
sp
o
t
,
(
2
)
e
a
r
l
y
b
l
i
g
h
t
,
(
3
)
h
e
a
l
t
h
y
,
(
4
)
l
a
t
e
b
l
i
g
h
t
,
(
5
)
l
e
a
f
mo
l
d
,
(
6
)
se
p
t
o
r
i
a
l
e
a
f
s
p
o
t
,
(
7
)
s
p
i
d
e
r
m
i
t
e
s (
8
)
t
a
r
g
e
t
sp
o
t
,
(
9
)
t
o
ma
t
o
mo
sai
c
v
i
r
u
s,
a
n
d
(
1
0
)
y
e
l
l
o
w
l
e
a
f
c
u
r
l
v
i
r
u
s
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
Dee
p
tr
a
n
s
fer lea
r
n
in
g
b
a
s
ed
d
is
ea
s
e
d
etec
tio
n
a
n
d
cla
s
s
ifica
tio
n
o
f to
m
a
to
lea
ve
s
…
(
Mu
n
ir
a
A
kter La
ta
)
1359
T
ab
le
5
.
R
esNet
-
5
0
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
I
mag
e
t
y
p
e
P
r
e
c
i
si
o
n
R
e
c
a
l
l
F1
-
sco
r
e
A
v
e
r
a
g
e
a
c
c
u
r
a
c
y
B
a
c
t
e
r
i
a
l
sp
o
t
0
.
9
2
0
.
9
0
0
.
9
0
0
.
8
5
Ea
r
l
y
b
l
i
g
h
t
0
.
6
2
0
.
6
3
0
.
6
2
H
e
a
l
t
h
y
0
.
9
6
0
.
9
4
0
.
9
5
L
a
t
e
b
l
i
g
h
t
0
.
8
1
0
.
7
9
0
.
8
0
L
e
a
f
mo
l
d
0
.
7
5
0
.
7
5
0
.
7
5
S
e
p
t
o
r
i
a
l
e
a
f
sp
o
t
0
.
8
3
0
.
8
0
0
.
8
1
S
p
i
d
e
r
mi
t
e
s
0
.
7
8
0
.
8
2
0
.
8
0
T
a
r
g
e
t
sp
o
t
0
.
7
0
0
.
7
2
0
.
7
1
T
o
mat
o
mo
sai
c
v
i
r
u
s
0
.
8
4
0
.
8
6
0
.
8
5
Y
e
l
l
o
w
l
e
a
f
c
u
r
l
v
i
r
u
s
0
.
9
4
0
.
9
4
0
.
9
4
Fig
u
r
e
5
.
P
R
cu
r
v
e
o
f
R
e
s
Net
-
50
Sev
er
al
r
esear
ch
er
s
h
av
e
alr
ea
d
y
ad
v
a
n
ce
d
th
e
f
ield
o
f
p
lan
t
d
is
ea
s
e
id
en
ti
f
icatio
n
th
r
o
u
g
h
th
e
u
s
e
o
f
i
m
a
g
e
p
r
o
ce
s
s
in
g
.
T
o
illu
s
tr
ate
th
e
ef
f
ec
ti
v
e
n
es
s
o
f
o
u
r
s
t
u
d
y
,
w
e
co
n
d
u
cted
a
co
m
p
ar
ativ
e
an
a
l
y
s
is
w
i
th
o
th
er
p
er
tin
en
t
r
esear
ch
p
ap
er
s
,
as
in
d
icate
d
in
T
ab
le
6
.
Ay
u
et
a
l.
[
2
2
]
u
s
ed
th
e
Mo
b
ileNetV2
ap
p
r
o
ac
h
to
d
etec
t
ca
s
s
av
a
leaf
d
is
ea
s
e
s
w
ith
an
ac
cu
r
ac
y
r
ati
n
g
o
f
6
5
.
6
%.
Gad
ad
e
et
a
l.
[
2
3
]
em
p
lo
y
ed
SVM,
KNN,
N
B
,
an
d
DT
to
id
en
ti
f
y
to
m
ato
l
ea
f
d
is
ea
s
es.
W
it
h
7
3
%
ac
c
u
r
ac
y
,
t
h
e
Gab
o
r
f
ea
tu
r
es
co
m
b
in
ed
w
i
th
SV
M
class
i
f
icatio
n
p
r
o
v
id
e
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
.
Usi
n
g
a
C
N
N,
R
a
m
ch
ar
a
n
et
a
l.
[
2
4
]
d
e
tecte
d
ca
s
s
av
a
leaf
d
is
ea
s
es
w
it
h
8
0
.
6
%
class
if
i
ca
tio
n
ac
cu
r
ac
y
o
n
p
ict
u
r
es
an
d
7
0
.
4
%
ac
cu
r
ac
y
o
n
v
id
eo
.
I
n
o
r
d
er
to
d
etec
t p
ests
in
to
m
ato
lea
v
es,
G
u
tier
r
ez
e
t
a
l.
[
2
5
]
em
p
lo
y
ed
K
NN,
m
u
ltil
a
y
er
p
er
ce
p
tr
o
n
(
ML
P
)
,
f
aster
r
eg
io
n
-
b
ased
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
et
w
o
r
k
(
R
-
C
NN)
,
a
n
d
s
i
n
g
le
s
h
o
t
d
ete
cto
r
(
SS
D)
.
A
m
o
n
g
o
u
r
ap
p
lie
d
m
o
d
els,
R
e
s
Net
-
5
0
p
er
f
o
r
m
ed
b
etter
th
a
n
ea
r
lier
s
tu
d
ie
s
,
w
it
h
t
h
e
h
ig
h
es
t
ac
cu
r
ac
y
r
ate
o
f
8
5
%.
I
n
t
h
is
w
o
r
k
,
t
h
e
u
s
e
o
f
R
esNet
-
5
0
f
o
r
to
m
ato
lea
f
d
is
ea
s
e
d
etec
tio
n
a
n
d
class
i
f
icat
i
o
n
r
esu
l
ts
i
n
h
ig
h
p
r
ec
is
io
n
,
r
ec
all,
F1
-
s
co
r
e,
an
d
ac
cu
r
a
c
y
.
T
h
ese
r
esu
l
ts
s
u
g
g
e
s
t
th
at
t
h
e
d
ev
elo
p
ed
m
et
h
o
d
m
a
y
b
e
u
s
ed
in
a
g
r
icu
l
tu
r
al
s
ettin
g
s
to
tr
ac
k
a
n
d
id
en
ti
f
y
d
is
ea
s
es e
ar
l
y
o
n
.
T
ab
le
6
.
C
o
m
p
ar
is
o
n
o
f
v
ar
io
u
s
p
lan
t d
is
ea
s
e
clas
s
i
f
icatio
n
tech
n
iq
u
es
R
e
f
e
r
e
n
c
e
s
A
p
p
l
i
e
d
t
e
c
h
n
i
q
u
e
(
s)
P
l
a
n
t
d
i
se
a
se
C
l
a
ssi
f
i
c
a
t
i
o
n
a
c
c
u
r
a
c
y
(
%)
A
y
u
e
t
a
l
.
[
2
2
]
M
o
b
i
l
e
N
e
t
V
2
C
a
ssav
a
6
5
.
6
G
a
d
a
d
e
e
t
a
l
.
[
2
3
]
S
V
M
,
K
N
N
,
N
a
i
v
e
B
a
y
e
s,
a
n
d
DT
T
o
mat
o
7
3
(
S
V
M
)
R
a
mc
h
a
r
a
n
e
t
a
l
.
[
2
4
]
C
N
N
C
a
ssav
a
8
0
.
6
o
n
i
mag
e
s a
n
d
7
0
.
4
o
n
v
i
d
e
o
G
u
t
i
e
r
r
e
z
e
t
a
l
.
[
2
5
]
K
N
N
,
M
L
P
,
F
a
st
e
r
R
-
C
N
N
,
a
n
d
S
S
D
T
o
mat
o
8
2
.
5
1
(
F
a
st
e
r
R
-
C
N
N
)
T
h
i
s
s
t
u
d
y
I
n
c
e
p
t
i
o
n
V
3
,
R
e
sN
e
t
-
1
8
,
R
e
sN
e
t
-
5
0
,
V
G
G
-
1
6
,
VGG
-
1
9
,
G
o
o
g
L
e
N
e
t
,
a
n
d
A
l
e
x
N
e
t
T
o
mat
o
8
5
(
R
e
sN
e
t
-
50)
4.
CO
NCLU
SI
O
N
Dis
ea
s
e
a
s
s
a
u
lt
s
h
a
v
e
a
n
i
m
p
ac
t
o
n
to
m
ato
p
la
n
t
q
u
al
it
y
a
n
d
a
m
o
u
n
t.
T
h
i
s
s
t
u
d
y
’
s
o
b
j
ec
tiv
e
i
s
to
id
en
ti
f
y
th
e
m
o
s
t
e
f
f
ec
ti
v
e
d
e
ep
tr
an
s
f
er
lear
n
in
g
cla
s
s
i
f
icat
io
n
m
o
d
el
f
o
r
to
m
ato
lea
f
d
is
ea
s
e
id
en
t
if
ica
tio
n
.
I
n
ce
p
tio
n
V3
,
R
esNet
-
1
8
,
R
esNet
-
5
0
,
VGG
-
1
6
,
VGG
-
1
9
,
Go
o
g
L
eNe
t,
an
d
A
le
x
Net
ar
e
th
e
s
ev
e
n
d
ee
p
lear
n
in
g
tr
a
n
s
f
er
lear
n
i
n
g
tech
n
iq
u
e
s
th
a
t
w
e
ass
e
s
s
ed
i
n
t
h
i
s
s
t
u
d
y
.
Gi
v
en
th
e
d
atase
t
an
d
class
i
f
icatio
n
j
o
b
,
th
e
R
esNet
-
5
0
o
u
tp
er
f
o
r
m
ed
t
h
e
o
th
er
te
s
ted
al
g
o
r
ith
m
s
w
it
h
a
p
r
ec
is
io
n
o
f
8
2
%,
r
ec
all
o
f
8
1
%,
F1
-
s
co
r
e
o
f
8
1
%,
an
d
ac
cu
r
ac
y
o
f
8
5
%
m
ee
ti
n
g
t
h
e
ac
cu
r
ac
y
r
eq
u
i
r
e
m
en
t
s
o
f
d
is
ea
s
e
class
i
f
ica
tio
n
.
No
n
et
h
eles
s
,
Go
o
g
L
eNe
t
y
ield
ed
th
e
least
ef
f
icie
n
t
p
er
f
o
r
m
an
ce
in
co
m
p
ar
i
s
o
n
to
th
e
o
th
er
la
y
o
u
ts
.
R
esNet
-
5
0
ca
n
b
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
23
,
No
.
5
,
Octo
b
e
r
20
25
:
1
3
5
3
-
1
362
1360
in
te
g
r
ated
in
to
m
o
b
ile
ap
p
lica
tio
n
s
o
r
I
o
T
-
co
m
p
atib
le
d
ev
ic
es
to
f
ac
ilit
ate
r
ea
l
-
t
i
m
e
ill
n
es
s
id
en
t
if
ica
tio
n
f
o
r
f
ar
m
er
s
.
A
d
d
itio
n
all
y
,
th
e
ap
p
r
o
ac
h
ca
n
b
e
s
ca
led
to
s
u
p
p
o
r
t
o
th
er
cr
o
p
s
o
f
s
a
m
e
f
a
m
i
l
y
a
n
d
b
ac
k
g
r
o
u
n
d
v
ar
iatio
n
s
b
y
r
etr
ai
n
in
g
t
h
e
m
o
d
el
w
it
h
d
iv
er
s
e
a
g
r
ic
u
lt
u
r
al
d
atasets
.
S
u
ch
i
m
p
le
m
e
n
tatio
n
s
co
u
ld
p
la
y
a
v
ital
r
o
le
in
p
r
ec
is
io
n
a
g
r
ic
u
lt
u
r
e
an
d
h
e
lp
r
ed
u
ce
cr
o
p
l
o
s
s
es.
Ho
w
e
v
er
,
t
h
e
d
ataset
u
s
ed
in
t
h
is
w
o
r
k
is
i
m
b
alan
ce
d
,
w
h
ic
h
m
a
y
lead
th
e
class
i
f
ier
to
p
er
f
o
r
m
p
o
o
r
ly
o
n
r
ar
e
d
is
ea
s
es
an
d
f
av
o
r
th
e
m
aj
o
r
it
y
class
i
f
icatio
n
s
.
T
o
v
er
if
y
ca
teg
o
r
izatio
n
r
es
u
lts
,
a
cr
o
s
s
-
v
alid
atio
n
tec
h
n
iq
u
e
w
il
l
b
e
ad
d
ed
t
o
im
a
g
e
p
r
o
c
ess
in
g
i
n
t
h
e
f
u
t
u
r
e.
F
UNDIN
G
I
NF
O
RM
AT
I
O
N
T
h
is
r
esear
ch
w
a
s
f
u
n
d
ed
b
y
Un
i
v
er
s
it
y
Gr
an
t
s
C
o
m
m
i
s
s
io
n
at
Un
i
v
er
s
it
y
o
f
Fro
n
t
ier
T
e
ch
n
o
lo
g
y
,
B
an
g
lad
es
h
.
C
o
n
tr
ac
t n
u
m
b
er
:
3
7
.
0
1
.
4
1
0
4
.
0
0
0
.
7
1
.
0
0
1
.
1
9
.
3
1
0
9
.
AUTHO
R
CO
NT
RIB
UT
I
O
NS ST
A
T
E
M
E
NT
T
h
is
j
o
u
r
n
al
u
s
e
s
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT
)
to
r
ec
o
g
n
ize
in
d
i
v
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
t
h
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
lla
b
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Mu
n
ir
a
Ak
ter
L
ata
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Ma
r
j
ia
Su
ltan
a
✓
✓
✓
✓
✓
✓
✓
I
f
f
at
A
r
a
B
ad
h
a
n
✓
✓
✓
✓
✓
✓
Ma
s
tu
r
a
J
ah
a
n
Ma
r
ia
✓
✓
✓
✓
Far
ih
a
T
asn
i
m
Nu
h
a
✓
✓
✓
✓
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
si
s
I
:
I
n
v
e
st
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
si
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
t
h
o
r
s
s
tate
n
o
co
n
f
lic
t o
f
i
n
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
in
g
s
o
f
t
h
i
s
s
t
u
d
y
ar
e
av
a
ilab
l
e
f
r
o
m
t
h
e
co
r
r
esp
o
n
d
in
g
a
u
t
h
o
r
,
[
MS]
,
u
p
o
n
r
ea
s
o
n
ab
le
r
eq
u
est.
RE
F
E
R
E
NC
E
S
[
1
]
J.
D
a
w
i
d
,
“
T
h
e
r
o
l
e
o
f
t
o
mat
o
p
r
o
d
u
c
t
s
f
o
r
h
u
ma
n
h
e
a
l
t
h
(
so
l
a
n
u
m
l
y
c
o
p
e
r
si
c
u
m
)
-
a
r
e
v
i
e
w
,
”
J
o
u
r
n
a
l
o
f
H
e
a
l
t
h
,
Me
d
i
c
i
n
e
a
n
d
N
u
rs
i
n
g
,
v
o
l
.
3
3
,
p
p
.
6
6
–
7
4
,
J
a
n
.
2
0
1
6
.
[
2
]
M
.
M
.
H
o
ssa
i
n
a
n
d
F
.
A
b
d
u
l
l
a
,
“
O
n
t
h
e
p
r
o
d
u
c
t
i
o
n
b
e
h
a
v
i
o
r
s
a
n
d
f
o
r
e
c
a
st
i
n
g
t
h
e
t
o
m
a
t
o
e
s
p
r
o
d
u
c
t
i
o
n
i
n
B
a
n
g
l
a
d
e
sh
,
”
J
o
u
rn
a
l
o
f
Ag
ri
c
u
l
t
u
r
a
l
Ec
o
n
o
m
i
c
s
a
n
d
D
e
v
e
l
o
p
m
e
n
t
,
v
o
l
.
4
,
n
o
.
5
,
p
p
.
6
6
-
7
4
,
A
u
g
.
2
0
1
5
.
[
3
]
M
.
K
h
a
t
u
n
a
n
d
M
.
R
a
h
m
a
n
,
“
P
o
st
h
a
r
v
e
st
l
o
ss
a
sse
ssm
e
n
t
o
f
t
o
m
a
t
o
i
n
se
l
e
c
t
e
d
l
o
c
a
t
i
o
n
s
o
f
B
a
n
g
l
a
d
e
sh
,
”
B
a
n
g
l
a
d
e
s
h
J
o
u
r
n
a
l
o
f
Ag
ri
c
u
l
t
u
r
a
l
Re
s
e
a
rc
h
,
v
o
l
.
4
5
,
n
o
.
1
,
p
p
.
4
3
-
5
2
,
J
u
n
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
2
9
/
b
j
a
r
.
v
4
5
i
1
.
5
9
8
3
7
.
[
4
]
R
.
T
h
a
n
g
a
r
a
j
,
S
.
A
n
a
n
d
a
m
u
r
u
g
a
n
,
a
n
d
V
.
K
.
K
a
l
i
a
p
p
a
n
,
“
A
u
t
o
mat
e
d
t
o
mat
o
l
e
a
f
d
i
se
a
se
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
-
b
a
se
d
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
,
”
J
o
u
r
n
a
l
o
f
P
l
a
n
t
D
i
s
e
a
ses
a
n
d
Pro
t
e
c
t
i
o
n
,
v
o
l
.
1
2
8
,
p
p
.
7
3
-
8
6
,
N
o
v
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
s
4
1
3
4
8
-
0
2
0
-
0
0
4
0
3
-
0.
[
5
]
O
.
A
t
t
a
l
l
a
h
,
“
T
o
mat
o
l
e
a
f
d
i
se
a
se
c
l
a
ssi
f
i
c
a
t
i
o
n
v
i
a
c
o
m
p
a
c
t
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
w
i
t
h
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
a
n
d
f
e
a
t
u
r
e
se
l
e
c
t
i
o
n
,
”
H
o
rt
i
c
u
l
t
u
r
a
e
,
v
o
l
.
9
,
n
o
.
2
,
p
.
1
4
9
,
Ja
n
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
h
o
r
t
i
c
u
l
t
u
r
a
e
9
0
2
0
1
4
9
.
[
6
]
N
.
K
h
a
saw
n
e
h
,
E.
F
a
o
u
r
i
,
a
n
d
M
.
F
r
a
i
w
a
n
,
“
A
u
t
o
mat
i
c
d
e
t
e
c
t
i
o
n
o
f
t
o
mat
o
d
i
se
a
se
s
u
si
n
g
d
e
e
p
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
,
”
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s
,
v
o
l
.
1
2
,
n
o
.
1
7
,
p
.
8
4
6
7
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
1
2
1
7
8
4
6
7
.
[
7
]
T
.
S
a
n
i
d
a
,
A
.
S
i
d
e
r
i
s,
M
.
V
.
S
a
n
i
d
a
,
a
n
d
M
.
D
a
sy
g
e
n
i
s,
“
T
o
mat
o
l
e
a
f
d
i
se
a
se
i
d
e
n
t
i
f
i
c
a
t
i
o
n
v
i
a
t
w
o
–
st
a
g
e
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
,
”
S
m
a
rt
A
g
ri
c
u
l
t
u
r
a
l
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
5
,
p
.
1
0
0
2
7
5
,
O
c
t
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
t
e
c
h
.
2
0
2
3
.
1
0
0
2
7
5
.
[
8
]
A
.
A
b
b
a
s,
S
.
Ja
i
n
,
M
.
G
o
u
r
,
a
n
d
S
.
V
a
n
k
u
d
o
t
h
u
,
“
T
o
mat
o
p
l
a
n
t
d
i
se
a
se
d
e
t
e
c
t
i
o
n
u
s
i
n
g
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
w
i
t
h
C
-
G
A
N
s
y
n
t
h
e
t
i
c
i
mag
e
s,
”
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
ri
c
u
l
t
u
r
e
,
v
o
l
.
1
8
7
,
p
.
1
0
6
2
7
9
,
A
u
g
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
a
g
.
2
0
2
1
.
1
0
6
2
7
9
.
[
9
]
M
.
S
.
A
l
z
a
h
r
a
n
i
a
n
d
F
.
W
.
A
l
saa
d
e
,
“
T
r
a
n
sf
o
r
m
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms
f
o
r
t
h
e
e
a
r
l
y
d
e
t
e
c
t
i
o
n
a
n
d
r
e
c
o
g
n
i
t
i
o
n
o
f
t
o
mat
o
l
e
a
f
d
i
se
a
se
,
”
Ag
ro
n
o
m
y
,
v
o
l
.
1
3
,
n
o
.
5
,
p
.
1
1
8
4
,
M
a
y
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
o
n
o
my
1
3
0
5
1
1
8
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
Dee
p
tr
a
n
s
fer lea
r
n
in
g
b
a
s
ed
d
is
ea
s
e
d
etec
tio
n
a
n
d
cla
s
s
ifica
tio
n
o
f to
m
a
to
lea
ve
s
…
(
Mu
n
ir
a
A
kter La
ta
)
1361
[
1
0
]
G
.
P
a
t
t
n
a
i
k
,
V
.
K
.
S
h
r
i
v
a
s
t
a
v
a
,
a
n
d
K
.
P
a
r
v
a
t
h
i
,
“
T
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
-
b
a
se
d
f
r
a
me
w
o
r
k
f
o
r
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
p
e
st
i
n
t
o
m
a
t
o
p
l
a
n
t
s
,
”
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
4
,
n
o
.
1
3
,
p
p
.
1
-
1
3
,
J
u
l
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
8
0
/
0
8
8
3
9
5
1
4
.
2
0
2
0
.
1
7
9
2
0
3
4
.
[
1
1
]
A
.
S
.
P
a
y
mo
d
e
a
n
d
V
.
B
.
M
a
l
o
d
e
,
“
T
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
f
o
r
m
u
l
t
i
-
c
r
o
p
l
e
a
f
d
i
se
a
se
i
m
a
g
e
c
l
a
ssi
f
i
c
a
t
i
o
n
u
si
n
g
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s VG
G
,
”
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
i
n
A
g
r
i
c
u
l
t
u
r
e
,
v
o
l
.
6
,
p
p
.
2
3
-
3
3
,
J
a
n
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
i
i
a
.
2
0
2
1
.
1
2
.
0
0
2
.
[
1
2
]
A
.
S
a
e
e
d
,
A
.
A
.
A
b
d
e
l
-
A
z
i
z
,
A
.
M
o
ssad
,
M
.
A
.
A
b
d
e
l
h
a
mi
d
,
A
.
Y
.
A
l
k
h
a
l
e
d
,
a
n
d
M
.
M
a
y
h
o
u
b
,
“
S
mart
d
e
t
e
c
t
i
o
n
o
f
t
o
m
a
t
o
l
e
a
f
d
i
se
a
se
s
u
s
i
n
g
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
-
b
a
s
e
d
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
,
”
Ag
ri
c
u
l
t
u
r
e
,
v
o
l
.
1
3
,
n
o
.
1
,
p
.
1
3
9
,
Ja
n
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
a
g
r
i
c
u
l
t
u
r
e
1
3
0
1
0
1
3
9
.
[
1
3
]
S
.
M
.
H
a
ss
a
n
,
A
.
K
.
M
a
j
i
,
M
.
Jas
i
ń
sk
i
,
Z
.
L
e
o
n
o
w
i
c
z
,
a
n
d
E.
Jas
i
ń
sk
a
,
“
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
se
a
se
s
u
si
n
g
C
N
N
a
n
d
t
r
a
n
sf
e
r
-
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
,
”
E
l
e
c
t
r
o
n
i
c
s,
v
o
l
.
1
0
,
n
o
.
1
2
,
p
.
1
3
8
8
,
J
u
n
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
e
l
e
c
t
r
o
n
i
c
s
1
0
1
2
1
3
8
8
.
[
1
4
]
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
,
“
T
o
L
e
D
:
To
mat
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
.
[
1
5
]
A
.
A
li
,
“
P
l
a
n
t
V
i
l
l
a
g
e
D
a
t
a
se
t
,
”
K
a
g
g
l
e
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s:
/
/
w
w
w
.
k
a
g
g
l
e
.
c
o
m/
d
a
t
a
se
t
s/
a
b
d
a
l
l
a
h
a
l
i
d
e
v
/
p
l
a
n
t
v
i
l
l
a
g
e
-
d
a
t
a
se
t
[
1
6
]
M
.
A
n
c
o
n
a
,
E
.
C
e
o
l
i
n
i
,
C
.
Ö
z
t
i
r
e
l
i
,
a
n
d
M
.
G
r
o
ss,
“
T
o
w
a
r
d
s
b
e
t
t
e
r
u
n
d
e
r
st
a
n
d
i
n
g
o
f
g
r
a
d
i
e
n
t
-
b
a
se
d
a
t
t
r
i
b
u
t
i
o
n
me
t
h
o
d
s
f
o
r
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
ren
c
e
o
n
L
e
a
r
n
i
n
g
Re
p
r
e
se
n
t
a
t
i
o
n
s
,
F
e
b
.
2
0
1
8
,
d
o
i
:
1
0
.
3
9
2
9
/
e
t
h
z
-
b
-
0
0
0
2
4
9
9
2
9
.
[
1
7
]
C
.
S
z
e
g
e
d
y
,
V
.
V
a
n
h
o
u
c
k
e
,
S
.
I
o
f
f
e
,
J.
S
h
l
e
n
s,
d
a
n
Z
.
W
o
j
n
a
,
“
R
e
t
h
i
n
k
i
n
g
t
h
e
i
n
c
e
p
t
i
o
n
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
c
o
mp
u
t
e
r
v
i
si
o
n
,
”
i
n
Pro
c
e
e
d
i
n
g
s o
f
t
h
e
I
EE
E
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
s
i
o
n
a
n
d
P
a
t
t
e
r
n
Re
c
o
g
n
i
t
i
o
n
(
C
VPR)
,
L
a
s Ve
g
a
s,
N
V
,
U
S
A
,
2
0
1
6
,
p
p
.
2
8
1
8
-
2
8
2
6
,
d
o
i
:
1
0
.
1
1
0
9
/
C
V
P
R
.
2
0
1
6
.
3
0
8
.
[
1
8
]
K
.
H
e
,
X
.
Z
h
a
n
g
,
S
.
R
e
n
,
d
a
n
J
.
S
u
n
,
“
D
e
e
p
r
e
si
d
u
a
l
l
e
a
r
n
i
n
g
f
o
r
i
m
a
g
e
r
e
c
o
g
n
i
t
i
o
n
,
”
i
n
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
I
E
EE
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
Re
c
o
g
n
i
t
i
o
n
(
C
VP
R)
,
2
0
1
6
,
p
p
.
7
7
0
-
7
7
8
,
d
o
i
:
1
0
.
4
8
5
5
0
/
a
r
X
i
v
.
1
5
1
2
.
0
3
3
8
5
.
[
1
9
]
K
.
S
i
mo
n
y
a
n
a
n
d
A
.
Z
i
sse
r
man
,
“
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
-
sca
l
e
i
mag
e
r
e
c
o
g
n
i
t
i
o
n
,
”
C
o
m
p
u
t
e
r
V
i
si
o
n
a
n
d
Pa
t
t
e
r
n
Re
c
o
g
n
i
t
i
o
n
,
A
p
r
.
1
0
,
2
0
1
5
,
d
o
i
:
1
0
.
4
8
5
5
0
/
a
r
X
i
v
.
1
4
0
9
.
1
5
5
6
.
[
2
0
]
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
,
”
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
I
EEE
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
si
o
n
a
n
d
P
a
t
t
e
r
n
Re
c
o
g
n
i
t
i
o
n
(
C
VPR),
2
0
1
5
,
p
p
.
1
-
9
,
d
o
i
:
1
0
.
4
8
5
5
0
/
a
r
X
i
v
.
1
4
0
9
.
4
8
4
2
.
[
2
1
]
A
.
K
r
i
z
h
e
v
sk
y
,
I
.
S
u
t
sk
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
,
”
A
d
v
a
n
c
e
s
i
n
N
e
u
ra
l
I
n
f
o
rm
a
t
i
o
n
Pr
o
c
e
ssi
n
g
S
y
s
t
e
m
s,
v
o
l
.
2
5
,
N
e
u
rI
PS
2
0
1
2
,
p
p
.
1
0
9
7
-
1
1
0
5
,
M
a
y
2
0
1
2
.
[
2
2
]
H
.
R
.
A
y
u
,
A
.
S
u
r
t
o
n
o
,
a
n
d
D
.
K
.
A
p
r
i
y
a
n
t
o
,
“
D
e
e
p
l
e
a
r
n
i
n
g
f
o
r
d
e
t
e
c
t
i
o
n
c
a
ssav
a
l
e
a
f
d
i
se
a
se
,
”
J
o
u
r
n
a
l
o
f
P
h
y
s
i
c
s:
C
o
n
f
e
re
n
c
e
S
e
r
i
e
s,
v
o
l
.
1
7
5
1
,
p
.
0
1
2
0
7
2
,
Ja
n
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
8
8
/
1
7
4
2
-
6
5
9
6
/
1
7
5
1
/
1
/
0
1
2
0
7
2
.
[
2
3
]
H.
D.
G
a
d
a
d
e
a
n
d
K
.
R
.
K
i
r
a
n
g
e
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
t
o
w
a
r
d
s
t
o
mat
o
l
e
a
f
d
i
se
a
se
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Ad
v
a
n
c
e
d
T
r
e
n
d
s
i
n
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
ri
n
g
,
v
o
l
.
9
,
n
o
.
1
,
p
p
.
4
9
0
–
4
9
5
,
F
e
b
.
2
0
2
0
,
d
o
i
:
1
0
.
3
0
5
3
4
/
i
j
a
t
c
se
/
2
0
2
0
/
6
7
9
1
2
0
2
0
.
[
2
4
]
A
.
R
a
mc
h
a
r
a
n
e
t
a
l
.
,
“
A
mo
b
i
l
e
-
b
a
s
e
d
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
f
o
r
c
a
ssav
a
d
i
se
a
se
d
i
a
g
n
o
si
s
,
”
Fro
n
t
i
e
r
s
i
n
Pl
a
n
t
S
c
i
e
n
c
e
,
v
o
l
.
1
0
,
M
a
r
.
2
0
1
9
,
d
o
i
:
1
0
.
3
3
8
9
/
f
p
l
s.2
0
1
9
.
0
0
2
7
2
.
[
2
5
]
A
.
G
u
t
i
e
r
r
e
z
,
A
.
A
n
su
a
t
e
g
i
,
L
.
S
u
sp
e
r
r
e
g
i
,
C
.
T
u
b
í
o
,
I
.
R
a
n
k
i
ć
,
a
n
d
L
.
L
e
n
ž
a
,
“
A
B
e
n
c
h
mark
i
n
g
o
f
l
e
a
r
n
i
n
g
st
r
a
t
e
g
i
e
s
f
o
r
p
e
st
d
e
t
e
c
t
i
o
n
a
n
d
i
d
e
n
t
i
f
i
c
a
t
i
o
n
o
n
t
o
m
a
t
o
p
l
a
n
t
s
f
o
r
a
u
t
o
n
o
mo
u
s
s
c
o
u
t
i
n
g
r
o
b
o
t
s
u
s
i
n
g
i
n
t
e
r
n
a
l
d
a
t
a
b
a
se
s
,
”
J
o
u
r
n
a
l
o
f
S
e
n
s
o
rs,
v
o
l
.
2
0
1
9
,
n
o
.
1
,
p
p
.
1
–
1
5
,
M
a
y
2
0
1
9
,
d
o
i
:
1
0
.
1
1
5
5
/
2
0
1
9
/
5
2
1
9
4
7
1
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
M
u
n
ira
A
k
te
r
L
a
ta
is
c
u
rre
n
tl
y
w
o
rk
in
g
a
s
a
n
A
s
sista
n
t
P
r
o
f
e
ss
o
r
in
th
e
De
p
a
rtme
n
t
o
f
Ed
u
c
a
ti
o
n
a
l
T
e
c
h
n
o
lo
g
y
a
n
d
E
n
g
in
e
e
rin
g
w
it
h
i
n
t
h
e
F
a
c
u
lt
y
o
f
Dig
it
a
l
T
ra
n
s
f
o
r
m
a
ti
o
n
E
n
g
in
e
e
rin
g
a
t
Un
iv
e
rsity
o
f
F
ro
n
ti
e
r
T
e
c
h
n
o
l
o
g
y
,
Ba
n
g
lad
e
sh
.
S
h
e
re
c
e
iv
e
d
h
e
r
B.
S
c
.
(H
o
n
s)
a
n
d
M
.
S
c
.
d
e
g
re
e
in
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
f
ro
m
Ja
h
a
n
g
irn
a
g
a
r
Un
iv
e
rsit
y
,
S
a
v
a
r,
Dh
a
k
a
,
Ba
n
g
lad
e
sh
.
He
r
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
im
a
g
e
p
ro
c
e
ss
in
g
,
m
a
c
h
in
e
lea
rn
in
g
,
d
e
e
p
lea
rn
i
n
g
,
h
u
m
a
n
-
c
o
m
p
u
ter
i
n
tera
c
ti
o
n
,
n
a
tu
ra
l
la
n
g
u
a
g
e
p
ro
c
e
ss
in
g
,
d
a
ta
m
in
in
g
,
d
a
ta
a
n
a
l
y
sis,
c
o
m
p
u
ter
v
isio
n
,
h
e
a
lt
h
in
f
o
rm
a
ti
c
s,
a
n
d
i
n
tern
e
t
o
f
th
i
n
g
s.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
m
u
n
ira0
0
0
1
@b
d
u
.
a
c
.
b
d
.
M
a
r
jia
S
u
lta
n
a
is cu
rre
n
tl
y
wo
rk
in
g
a
s a
n
A
ss
istan
t
P
ro
f
e
ss
o
r
in
th
e
De
p
a
rtm
e
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
E
n
g
in
e
e
rin
g
w
it
h
in
t
h
e
F
a
c
u
lt
y
o
f
En
g
in
e
e
rin
g
a
n
d
T
e
c
h
n
o
l
o
g
y
a
t
Be
g
u
m
Ro
k
e
y
a
Un
iv
e
rsit
y
,
Ra
n
g
p
u
r,
Ba
n
g
lad
e
sh
.
S
h
e
re
c
e
iv
e
d
h
e
r
B.
S
c
.
(Ho
n
s)
d
e
g
re
e
in
Co
m
p
u
ter S
c
ien
c
e
a
n
d
En
g
in
e
e
ri
n
g
f
ro
m
Ja
h
a
n
g
irn
a
g
a
r
Un
iv
e
rsit
y
,
S
a
v
a
r,
Dh
a
k
a
in
2
0
1
6
a
n
d
M
.
S
c
.
d
e
g
re
e
f
ro
m
th
e
sa
m
e
u
n
iv
e
rsity
in
2
0
1
8
.
He
r
re
se
a
rc
h
in
tere
sts
i
n
c
lu
d
e
m
a
c
h
in
e
lea
rn
in
g
,
d
e
e
p
lea
rn
i
n
g
,
d
a
ta
m
in
in
g
,
c
o
m
p
u
ter
v
isio
n
,
im
a
g
e
p
ro
c
e
ss
in
g
,
a
n
d
c
o
m
p
u
ter
n
e
tw
o
rk
in
g
.
S
h
e
h
a
s
se
v
e
ra
l
re
se
a
rc
h
p
a
p
e
rs
p
u
b
li
s
h
e
d
in
in
te
rn
a
ti
o
n
a
l
c
o
n
f
e
re
n
c
e
s
a
n
d
jo
u
r
n
a
ls.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
m
a
rji
a
.
c
se
@b
ru
r.
a
c
.
b
d
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
23
,
No
.
5
,
Octo
b
e
r
20
25
:
1
3
5
3
-
1
362
1362
Iffa
t
Ar
a
B
a
d
h
a
n
is
c
u
rr
e
n
tl
y
w
o
rk
in
g
a
s
a
n
A
ss
istan
t
P
r
o
f
e
ss
o
r
in
th
e
De
p
a
rtme
n
t
o
f
El
e
c
tri
c
a
l
a
n
d
El
e
c
tro
n
ic
En
g
in
e
e
rin
g
w
it
h
in
th
e
F
a
c
u
lt
y
o
f
En
g
in
e
e
rin
g
a
n
d
T
e
c
h
n
o
lo
g
y
a
t
Be
g
u
m
Ro
k
e
y
a
Un
iv
e
rsit
y
,
Ra
n
g
p
u
r,
Ba
n
g
lad
e
s
h
.
S
h
e
re
c
e
iv
e
d
h
e
r
B.
S
c
.
(En
g
g
.
)
d
e
g
re
e
in
A
p
p
li
e
d
P
h
y
sic
s
a
n
d
El
e
c
tro
n
ic
E
n
g
in
e
e
rin
g
f
ro
m
Un
iv
e
r
sit
y
o
f
Ra
jsh
a
h
i
in
2
0
1
5
a
n
d
M
.
S
c
.
(E
n
g
g
.
)
d
e
g
re
e
f
ro
m
th
e
sa
m
e
u
n
iv
e
rsit
y
in
2
0
1
7
.
He
r
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
m
a
c
h
in
e
lea
rn
in
g
,
d
e
e
p
lea
rn
in
g
,
d
a
ta
m
in
in
g
,
c
o
m
p
u
ter
v
isio
n
,
im
a
g
e
p
ro
c
e
ss
in
g
,
c
o
m
m
u
n
ica
ti
o
n
,
a
n
d
c
o
m
p
u
ter
n
e
tw
o
rk
in
g
.
S
h
e
h
a
s
se
v
e
ra
l
re
se
a
rc
h
p
a
p
e
rs
p
u
b
li
sh
e
d
i
n
in
tern
a
ti
o
n
a
l
j
o
u
r
n
a
ls.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
iab
a
d
h
a
n
@
b
r
u
r.
a
c
.
b
d
.
M
a
stu
r
a
J
a
h
a
n
M
a
r
ia
i
s
a
se
c
o
n
d
-
y
e
a
r
stu
d
e
n
t
i
n
th
e
De
p
a
rtm
e
n
t
o
f
Ed
u
c
a
ti
o
n
a
l
T
e
c
h
n
o
l
o
g
y
a
n
d
En
g
in
e
e
rin
g
w
it
h
in
t
h
e
F
a
c
u
lt
y
o
f
Dig
it
a
l
T
ra
n
s
f
o
r
m
a
ti
o
n
En
g
in
e
e
rin
g
a
t
Un
iv
e
rsit
y
o
f
F
ro
n
ti
e
r
T
e
c
h
n
o
l
o
g
y
,
B
a
n
g
lad
e
sh
.
He
r
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
m
a
c
h
in
e
lea
rn
in
g
,
d
e
e
p
lea
rn
i
n
g
,
d
a
ta
a
n
a
ly
sis,
a
d
a
p
ti
v
e
lea
rn
in
g
sy
ste
m
s,
a
n
d
i
n
tern
e
t
o
f
th
in
g
s.
S
h
e
is p
la
n
n
i
n
g
to
o
f
f
e
r
a
c
o
n
tri
b
u
ti
o
n
to
t
h
e
in
terd
isc
ip
l
in
a
ry
r
e
se
a
r
c
h
th
a
t
w
il
l
h
e
lp
i
n
li
n
k
in
g
a
d
v
a
n
c
e
d
tec
h
n
o
lo
g
ies
w
it
h
t
h
e
p
re
se
n
t
e
d
u
c
a
ti
o
n
.
S
h
e
se
e
k
s
to
m
a
k
e
a
c
a
re
e
r
th
a
t
a
lso
in
v
o
lv
e
s
re
se
a
rc
h
,
in
n
o
v
a
ti
o
n
a
n
d
p
ra
c
ti
c
a
l
so
lu
ti
o
n
s
i
n
e
d
u
c
a
ti
o
n
a
n
d
tec
h
n
o
lo
g
y
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
m
a
stu
ra
0
0
0
1
@s
td
.
b
d
u
.
a
c
.
b
d
.
Fa
r
ih
a
Ta
s
n
i
m
Nuh
a
is
a
se
c
o
n
d
-
y
e
a
r
stu
d
e
n
t
in
th
e
De
p
a
rtm
e
n
t
o
f
Ed
u
c
a
ti
o
n
a
l
T
e
c
h
n
o
lo
g
y
a
n
d
En
g
in
e
e
rin
g
w
i
th
in
t
h
e
F
a
c
u
lt
y
o
f
Dig
it
a
l
T
ra
n
sf
o
r
m
a
ti
o
n
En
g
in
e
e
rin
g
a
t
Un
iv
e
rsit
y
o
f
F
ro
n
ti
e
r
T
e
c
h
n
o
lo
g
y
,
Ba
n
g
lad
e
sh
.
He
r
e
d
u
c
a
ti
o
n
al
b
a
c
k
g
ro
u
n
d
is
a
c
o
m
b
in
a
ti
o
n
o
f
e
d
u
c
a
ti
o
n
a
n
d
tec
h
n
o
l
o
g
y
w
it
h
e
m
p
h
a
sis
o
n
i
n
n
o
v
a
ti
v
e
m
e
c
h
a
n
is
m
s
o
f
tea
c
h
in
g
a
n
d
lea
rn
i
n
g
.
P
a
rti
c
u
lar
ly
,
sh
e
is
i
n
tere
ste
d
i
n
t
h
e
st
u
d
i
e
s
th
a
t
c
o
m
b
in
e
e
d
u
c
a
ti
o
n
w
it
h
n
e
w
te
c
h
n
o
l
o
g
ies
to
a
c
h
iev
e
b
e
tt
e
r
in
stru
c
ti
o
n
a
l
p
ra
c
ti
c
e
s
,
stu
d
e
n
t
e
n
g
a
g
e
m
e
n
t
a
n
d
c
o
n
se
q
u
e
n
tl
y
b
e
tt
e
r
lea
rn
i
n
g
re
su
lt
s
.
He
r
f
o
c
u
s
a
re
a
s
o
f
re
se
a
rc
h
a
re
a
d
a
p
ti
v
e
lea
rn
in
g
,
b
len
d
e
d
lea
rn
i
n
g
a
n
d
i
n
c
o
r
p
o
ra
ti
o
n
o
f
d
ig
it
a
l
t
o
o
ls
i
n
t
h
e
p
re
se
n
t
d
a
y
p
e
d
a
g
o
g
i
c
a
l
m
o
d
e
l.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
f
a
rih
a
n
u
h
a
3
5
6
@g
m
a
il
.
c
o
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.