I
A
E
S
I
n
t
e
r
n
at
io
n
al
Jou
r
n
al
of
A
r
t
if
ic
ia
l
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
V
ol
. 15, No. 1, Febr
ua
r
y 2026
, pp.
568
~
579
I
S
S
N
:
2252
-
8938
,
D
O
I
:
10.11591/
ij
a
i.
v
15
.i
1
.pp
568
-
579
568
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ij
ai
.
ia
e
s
c
or
e
.c
om
A
n
ar
t
i
f
i
c
i
al
i
n
t
e
l
l
i
ge
n
c
e
t
e
c
h
n
ol
ogy f
or
p
r
om
ot
i
n
g h
o
m
-
t
h
on
g
b
an
an
a agr
i
c
u
l
t
u
r
e
sys
t
e
m
R
at
s
am
e
s
T
an
ve
e
n
u
k
ool
1
, S
u
w
it
S
om
s
u
p
h
ap
r
u
n
gyos
1
, B
oon
yar
it
N
ok
k
u
r
t
h
1
, L
ik
i
t
C
h
am
u
t
h
ai
1
,
P
at
u
m
w
ad
e
e
B
on
gu
le
au
m
2
, P
ar
in
ya N
at
h
o
1
1
D
e
pa
r
t
m
e
nt
of
I
nf
or
m
a
t
i
on
S
ys
t
e
m
a
nd B
us
i
ne
s
s
C
om
put
e
r
, F
a
c
ul
t
y of
B
us
i
ne
s
s
A
dm
i
ni
s
t
r
a
t
i
on a
nd I
nf
or
m
a
t
i
on T
e
c
hnol
ogy,
R
a
j
a
m
a
nga
l
a
U
ni
ve
r
s
i
t
y of
T
e
c
hnol
ogy S
uva
r
na
bhum
i
, P
hr
a
N
a
khon S
i
A
yut
t
ha
ya
, T
ha
i
l
a
nd
2
D
e
pa
r
t
m
e
nt
of
A
c
c
ount
i
ng, F
a
c
ul
t
y of
B
us
i
ne
s
s
A
dm
i
ni
s
t
r
a
t
i
on a
nd I
nf
or
m
a
t
i
o
n T
e
c
hnol
ogy, R
a
j
a
m
a
nga
l
a
U
ni
ve
r
s
i
t
y of
T
e
c
hnol
ogy
S
uva
r
na
bhum
i
, P
hr
a
N
a
khon S
i
A
yut
t
ha
ya
, T
ha
i
l
a
nd
A
r
t
ic
le
I
n
f
o
A
B
S
T
R
A
C
T
A
r
ti
c
le
h
is
to
r
y
:
R
e
c
e
iv
e
d
A
ug 18, 2025
R
e
vi
s
e
d
D
e
c
24, 2025
A
c
c
e
pt
e
d
J
a
n 10, 2026
The
hom
-
thong
banana,
being
a
high
-
value
Thai
export
variety,
is
facing
significant
risk
from
disease
outbreaks
affecting
crop
yield
and
q
uality.
Traditional
visual
inspection
methods
in
detection
of
diseases
are
labor
-
consumi
ng,
error
-
prone.
This
research
addresses
these
limitatio
ns
by
developing
a
new
artificial
intell
igence
(AI)
-
based
automatic
disease
detection
system
for
the
hom
-
thong
banana
industry
on
top
of
cuttin
g
-
edge
computer
vision
technolo
gy.
The
study
employed
deep
learning
object
detection
models,
contrast
ing
Roboflow,
you
only
look
once
(
YOL
O
)
v11,
and
YOLOv12
architectures,
which
were
trained
on
a
large
dataset
of
2,576
images
of
Thai
banana
plant
ations.
With
systematic
data
augme
ntation
techniques,
the
dataset
was
augmented
to
6,184
images
of
seven
ty
pes
of
disease
under
varied
environmental
conditions.
The
method
e
ntailed
extensiv
e
preprocessi
ng
and
evaluatio
n
of
performance
through
pre
cision,
recall,
and
mean
average
precision
(mAP)
metrics.
Outcomes
indicat
ed
that
YOLOv12 outperformed wit
h 93.3% accurac
y, 83.3% sensitivity, and
86.3%
mAP@
50
compared
to
standard
inspection
schemes.
This
resea
rch
is
applicabl
e
to
Thailand'
s
smart
agricult
ure
initi
ative
by
providi
ng
f
armers
with
low
-
cost,
accurate,
and
effective
disease
monitoring
equipmen
t.
The
applicati
on
of
this
AI
system
has
the
abilit
y
to
enhance
the
yield
of
crops,
reduce losses,
and enhance the
competiti
veness of Thai
banana exports
in the
global market, in support of sustainable agricultural deve
lopment.
K
e
y
w
o
r
d
s
:
A
gr
ic
ul
tu
r
e
s
ys
te
m
A
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
H
om
-
th
ong ba
na
na
di
s
e
a
s
e
s
P
r
om
ot
in
g hom
-
th
ong
Y
O
L
O
m
ode
l
This is an
open
acce
ss artic
le unde
r the
CC BY
-
SA
license.
C
or
r
e
s
pon
di
n
g A
u
th
or
:
P
a
r
in
ya
N
a
th
o
D
e
pa
r
tm
e
nt
of
I
nf
or
m
a
ti
on S
ys
te
m
a
nd B
us
in
e
s
s
C
om
put
e
r
F
a
c
ul
ty
of
B
us
in
e
s
s
A
dm
in
is
tr
a
ti
on a
nd I
nf
or
m
a
ti
on T
e
c
hnol
ogy
R
a
ja
m
a
nga
la
U
ni
ve
r
s
it
y of
T
e
c
hnol
ogy S
uva
r
na
bhumi
P
hr
a
N
a
khon S
i
A
yut
th
a
ya
, T
ha
il
a
nd
E
m
a
il
:
pa
r
in
ya
.n@
r
m
ut
s
b.a
c
.t
h
1.
I
N
T
R
O
D
U
C
T
I
O
N
T
he
hom
-
th
ong
ba
na
na
,
r
e
now
ne
d
f
or
it
s
gol
de
n
hue
,
a
r
om
a
ti
c
f
r
a
gr
a
nc
e
,
a
nd
s
upe
r
io
r
ta
s
te
,
s
ta
nds
a
s
a
s
ig
ni
f
ic
a
nt
hi
gh
-
va
lu
e
c
r
op
f
or
T
ha
il
a
nd
[
1]
.
C
om
m
a
ndi
ng a
pr
e
m
iu
m
in
bot
h
dom
e
s
ti
c
a
nd
in
te
r
na
ti
ona
l
m
a
r
ke
ts
,
pa
r
ti
c
ul
a
r
ly
in
J
a
pa
n,
th
is
c
ul
ti
va
r
pr
e
s
e
nt
s
a
lu
c
r
a
ti
ve
oppor
tu
ni
ty
f
or
T
ha
i
f
a
r
m
e
r
s
a
nd
is
a
ke
y
c
ont
r
ib
ut
or
to
th
e
na
ti
on'
s
a
g
r
ic
ul
tu
r
a
l
e
c
onomy
[
2]
.
H
ow
e
ve
r
,
th
e
f
ul
l
pot
e
nt
ia
l
o
f
hom
-
th
ong
ba
na
na
c
ul
ti
va
ti
on
is
c
ur
r
e
nt
ly
ha
m
pe
r
e
d
by
a
m
ul
ti
tu
de
o
f
c
ha
ll
e
nge
s
,
r
a
ngi
ng
f
r
om
in
c
ons
is
te
nt
yi
e
ld
s
a
nd
vul
ne
r
a
bi
li
ty
to
pe
s
ts
a
nd
di
s
e
a
s
e
s
to
in
e
f
f
ic
ie
nc
ie
s
w
it
hi
n
th
e
s
uppl
y
c
ha
in
.
T
o
a
ddr
e
s
s
th
e
s
e
c
r
it
ic
a
l
is
s
ue
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
n ar
ti
fi
c
ia
l
in
te
ll
ig
e
nc
e
t
e
c
hnol
ogy
f
or
pr
om
ot
in
g hom
-
th
ong b
anana
…
(
R
at
s
am
e
s
T
anv
e
e
nu
k
ool
)
569
a
nd
s
e
c
ur
e
a
s
u
s
ta
in
a
bl
e
a
nd
pr
os
pe
r
ous
f
ut
ur
e
f
or
th
is
pr
iz
e
d
c
om
m
odi
ty
,
a
pa
r
a
di
gm
s
hi
f
t
to
w
a
r
ds
m
or
e
in
nova
ti
ve
a
nd i
nt
e
ll
ig
e
nt
a
gr
ic
ul
tu
r
a
l
pr
a
c
ti
c
e
s
i
s
not
j
us
t
a
n op
ti
on, but a
ne
c
e
s
s
it
y
[
3]
.
T
he
pa
th
t
o
e
le
va
ti
ng
ho
m
-
t
hong
ba
na
na
p
r
oduc
ti
on
l
ie
s
i
n
t
he
a
dopt
io
n o
f
c
ut
t
in
g
-
e
dge
,
i
nn
ova
ti
ve
a
gr
ic
ul
tu
r
a
l
s
ys
te
m
s
,
w
it
h
a
r
t
if
ic
ia
l
in
te
l
li
ge
nc
e
(
A
I
)
te
c
hnol
o
gy
a
t
t
he
ir
c
or
e
.
T
r
a
di
ti
ona
l
f
a
r
m
i
ng
m
e
t
hods
,
w
hi
le
va
lu
a
bl
e
,
of
te
n
f
a
ll
s
ho
r
t
in
p
r
ovi
di
ng
th
e
p
r
e
c
is
io
n
a
nd
r
e
a
l
-
ti
m
e
da
ta
ne
e
de
d
to
o
pt
i
m
iz
e
c
ul
ti
va
ti
o
n
a
nd
m
it
ig
a
te
r
is
ks
e
f
f
e
c
t
iv
e
ly
[
4]
,
[
5]
.
F
a
r
m
e
r
s
g
r
a
ppl
e
w
i
t
h
is
s
ue
s
s
uc
h
a
s
t
he
de
va
s
ta
ti
ng
i
m
pa
c
ts
o
f
P
a
na
m
a
d
is
e
a
s
e
a
nd
ot
he
r
pa
th
oge
ns
,
th
e
c
om
p
le
xi
t
ie
s
of
nut
r
i
e
nt
a
nd
w
a
te
r
m
a
na
ge
m
e
nt
,
a
n
d
th
e
la
bo
r
io
us
pr
oc
e
s
s
of
m
on
it
o
r
in
g
c
r
op
he
a
l
th
a
nd
p
r
e
di
c
t
in
g
y
ie
ld
s
.
T
h
e
s
e
c
ha
ll
e
nge
s
f
r
e
q
ue
nt
l
y
le
a
d
to
s
ig
n
if
ic
a
nt
pr
e
-
a
nd
pos
t
-
ha
r
ve
s
t
lo
s
s
e
s
,
in
c
ons
is
te
nt
f
r
ui
t
qua
l
it
y,
a
nd
a
n
in
a
b
il
i
ty
to
c
o
ns
is
te
nt
ly
m
e
e
t
th
e
s
t
r
in
ge
nt
de
m
a
nds
o
f
e
xpo
r
t
m
a
r
ke
ts
[
6
]
.
T
r
a
d
it
io
n
a
l
d
is
e
a
s
e
d
ia
g
nos
is
m
e
t
ho
ds
s
uc
h
a
s
vi
s
ua
l
e
xa
m
in
a
ti
o
n
th
r
oug
h
ho
r
t
ic
ul
tu
r
a
l
e
xpe
r
ts
o
r
f
a
r
m
e
r
s
a
r
e
t
ypi
c
a
l
ly
hu
m
a
n
-
e
r
r
o
r
s
us
c
e
p
ti
bl
e
,
t
e
di
ous
,
a
nd
la
bo
r
-
in
te
ns
iv
e
w
he
n
di
s
e
a
s
e
ha
s
s
i
m
i
la
r
v
is
ua
l
f
e
a
tu
r
e
s
[
7
]
.
I
n
r
e
c
e
n
t
ye
a
r
s
,
A
I
a
n
d
c
o
m
p
ut
e
r
vi
s
io
n
ha
ve
s
e
e
n
s
i
gn
if
ic
a
nt
a
dva
nc
e
m
e
n
ts
t
ha
t
ha
v
e
be
e
n
id
e
n
ti
f
ie
d
a
s
p
ot
e
nt
ia
l
t
oo
ls
f
o
r
m
o
ni
to
r
i
ng
pl
a
nt
he
a
lt
h
in
r
e
a
l
-
t
im
e
.
O
bj
e
c
t
de
t
e
c
t
io
n
a
lg
o
r
it
h
m
s
,
i
n
pa
r
ti
c
ul
a
r
,
ha
ve
s
ho
w
n
h
ig
h
p
r
o
m
is
e
in
id
e
n
ti
f
yi
ng
p
la
n
t
d
i
s
e
a
s
e
s
f
r
o
m
i
m
a
ge
s
c
a
pt
u
r
e
d
u
nde
r
v
a
r
yi
ng
f
ie
ld
c
o
nd
it
io
ns
[
8
]
.
O
ne
of
th
e
be
s
t
-
known
A
I
s
ol
ut
io
n
s
f
or
c
om
put
e
r
vi
s
io
n
w
it
h
in
a
gr
ic
ul
tu
r
e
is
th
e
s
e
r
ie
s
of
obj
e
c
t
de
te
c
ti
on
m
ode
ls
you
onl
y
lo
ok
onc
e
(
Y
O
L
O
)
.
Y
O
L
O
is
e
s
pe
c
ia
ll
y
w
e
ll
-
known
f
or
hi
gh
-
s
pe
e
d
a
nd
hi
gh
-
a
c
c
ur
a
c
y
r
e
a
l
-
ti
m
e
obj
e
c
t
de
te
c
ti
on,
a
nd
it
is
th
us
ve
r
y
w
e
ll
-
s
ui
te
d
f
or
a
ppl
ic
a
ti
ons
a
t
th
e
f
ie
ld
le
ve
l,
w
he
r
e
r
a
pi
d
de
c
is
io
n
-
m
a
ki
ng
is
r
e
qui
r
e
d
[
9]
,
[
10]
.
E
xpe
r
im
e
nt
s
ha
ve
e
s
ta
bl
is
he
d
th
e
pe
r
f
or
m
a
nc
e
of
Y
O
L
O
m
ode
ls
in
a
gr
ic
ul
tu
r
a
l
us
e
c
a
s
e
s
,
s
uc
h
a
s
de
te
c
ti
ng
to
m
a
to
le
a
f
bl
ig
ht
[
11]
,
m
oni
to
r
in
g
a
ppl
e
or
c
ha
r
d
pe
s
ts
[
12]
,
a
nd
de
te
c
ti
ng
di
s
e
a
s
e
s
in
gr
a
pe
vi
ne
s
[
13]
.
C
ont
r
a
r
y
to
th
e
s
in
gl
e
-
pa
s
s
d
e
te
c
ti
on
m
ode
pr
ovi
de
d
by
c
onve
nt
io
na
l
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
k
(
C
N
N
)
c
la
s
s
if
ie
r
s
,
Y
O
L
O
'
s
s
in
gl
e
-
s
hot
de
te
c
ti
on
m
ode
of
f
e
r
s
s
im
ul
ta
ne
ous
c
la
s
s
if
ic
a
ti
on
a
nd
lo
c
a
li
z
a
ti
on
of
m
ul
ti
pl
e
di
s
e
a
s
e
s
ym
pt
om
s
in
a
s
in
gl
e
im
a
ge
ir
r
e
s
p
e
c
ti
ve
of
c
om
pl
e
x ba
c
kgr
ounds
[
14]
.
R
e
s
e
a
r
c
h i
nt
o Y
O
L
O
w
it
h
ba
na
na
di
s
e
a
s
e
de
te
c
ti
on
ha
s
be
c
o
m
e
t
r
e
n
dy i
n
r
e
c
e
nt
ye
a
r
s
f
or
e
xa
m
p
le
,
us
e
d
de
e
p
le
a
r
ni
n
g
ob
je
c
t
de
te
c
ti
o
n
to
di
a
gn
os
e
ba
na
na
bunc
hy
to
p
vi
r
us
w
i
th
m
uc
h
a
c
c
u
r
a
c
y
un
de
r
f
ie
l
d
li
ght
i
ll
um
in
a
t
io
n
[
1
5]
.
W
it
h
th
e
m
or
e
r
e
c
e
nt
Y
O
L
O
ve
r
s
io
ns
,
s
uc
h a
s
Y
O
L
O
v5
a
nd
Y
O
L
O
v8
,
th
e
de
te
c
ti
on
a
c
c
ur
a
c
y
i
nc
r
e
a
s
e
s
,
c
o
m
put
a
ti
ona
l
c
a
pa
c
it
y
de
c
r
e
a
s
e
s
,
a
n
d
por
ta
bi
li
ty
a
nd
m
ob
il
i
ty
to
s
m
a
r
t
de
vi
c
e
s
a
ls
o
im
p
r
ove
s
to
t
he
e
xt
e
n
t
w
he
r
e
th
e
y
c
a
n
be
de
p
lo
ye
d
in
por
ta
bl
e
c
on
f
ig
ur
a
t
io
ns
o
r
dr
one
-
ba
s
e
d
m
oni
t
or
i
ng
[
16
]
,
[
17
]
.
T
he
s
e
a
nd
s
ubs
e
que
nt
is
s
ue
s
pr
ovi
de
h
om
-
t
hong
ba
na
na
f
a
r
m
e
r
s
w
it
h
th
e
pos
s
ib
il
i
ty
of
e
a
r
ly
-
w
a
r
ni
ng d
is
e
a
s
e
de
te
c
ti
o
n, t
h
us
ge
ne
r
a
ti
ng
l
e
s
s
c
r
o
p l
o
s
s
a
nd l
e
s
s
ove
r
-
us
e
o
f
pe
s
ti
c
id
e
s
.
I
n
T
ha
i
la
nd,
w
he
r
e
t
he
ho
m
-
th
ong
ba
na
na
is
bot
h
a
s
ta
pl
e
f
ood
a
n
d
a
n
e
xpo
r
t
c
r
o
p,
th
e
us
e
of
AI
-
ba
s
e
d
t
o
ke
e
p
di
s
e
a
s
e
de
te
c
t
io
n
s
ys
te
m
s
on
f
a
r
m
w
oul
d
s
uppl
e
m
e
n
t
th
e
na
t
io
na
l
s
m
a
r
t
a
g
r
ic
ul
tu
r
e
a
ge
nda
.
C
o
m
bi
ne
d
w
it
h
a
f
f
or
da
bl
e
im
a
g
in
g
a
c
qui
s
it
io
n
t
o
ol
s
,
s
m
a
r
tp
h
one
s
,
dr
one
s
,
o
r
I
oT
c
a
m
e
r
a
s
,
th
e
f
a
r
m
e
r
c
a
n
im
a
ge
th
e
c
r
op
r
e
pe
ti
ti
ve
ly
a
nd
r
e
c
e
i
ve
f
e
e
d
ba
c
k
in
r
e
pa
r
a
bl
e
ti
m
e
to
r
e
s
pond
to
is
s
ue
s
c
a
us
in
g
po
te
nt
ia
l
l
os
s
of
yi
e
l
d.
I
t
pr
o
vi
de
s
a
pr
o
m
is
in
g,
s
us
t
a
in
a
bl
e
f
a
r
m
in
g
a
pp
r
oa
c
h
a
n
d
in
c
lu
de
s
o
th
e
r
be
ne
f
it
s
out
li
ne
d
in
th
e
na
ti
ona
l
a
ge
nda
pr
o
duc
ti
vi
ty
,
m
in
i
m
iz
in
g
e
n
vi
r
onm
e
n
ta
l
im
pa
c
t
,
a
nd
in
c
o
m
e
o
f
th
e
f
a
r
m
e
r
[
18]
,
[
19
]
.
T
he
r
e
f
o
r
e
,
th
e
c
ur
r
e
nt
s
tu
dy
pr
opos
e
s
th
e
de
ve
lo
pm
e
n
t
of
AI
te
c
h
nol
o
gy
f
or
p
r
om
ot
io
n
of
hom
-
th
ong
ba
na
na
a
g
r
ic
ul
tu
r
e
s
ys
te
m
us
in
g
t
he
Y
O
L
O
obj
e
c
t
de
te
c
ti
on
a
lg
or
i
th
m
f
o
r
e
a
r
ly
de
te
c
ti
on
a
nd
a
c
c
ur
a
te
r
e
c
ogni
ti
on
of
t
he
ke
y
ba
na
na
di
s
e
a
s
e
s
.
T
he
pr
o
po
s
e
d
s
ys
te
m
is
hope
d
to
be
a
n
e
f
f
ic
ie
nt
a
nd
de
pl
oya
b
le
t
ool
f
or
f
a
r
m
e
r
s
,
c
o
ope
r
a
t
iv
e
s
,
a
nd
a
gr
ic
ul
t
ur
a
l
a
ge
nc
ie
s
to
ul
ti
m
a
te
ly
s
uppo
r
t
th
e
c
om
pe
ti
ti
ve
ne
s
s
a
nd
r
e
s
il
ie
nc
e
o
f
t
he
hom
-
th
o
ng ba
na
na
in
d
us
t
r
y i
n
l
oc
a
l
a
nd g
lo
ba
l
m
a
r
ke
ts
.
T
hi
s
r
e
s
e
a
r
c
h
pr
op
os
e
s
th
e
de
ve
lo
pm
e
n
t
a
nd
im
p
le
m
e
nt
a
ti
on
of
a
n
in
nova
t
iv
e
a
gr
ic
ul
t
ur
e
s
ys
te
m
de
s
ig
ne
d
to
s
pe
c
i
f
ic
a
l
ly
pr
o
m
ot
e
th
e
c
ul
t
iv
a
ti
on
o
f
th
e
hom
-
t
hong
ba
na
na
th
r
o
ugh
th
e
s
t
r
a
te
g
ic
i
nt
e
g
r
a
ti
o
n
of
AI
.
T
h
is
s
ys
te
m
w
il
l
le
ve
r
a
ge
A
I
-
pow
e
r
e
d
to
ol
s
f
o
r
t
he
e
a
r
l
y
de
te
c
t
io
n
a
nd
di
a
gnos
is
of
di
s
e
a
s
e
s
th
r
ou
gh
im
a
ge
r
e
c
o
gni
t
io
n,
e
na
bl
i
ng
ti
m
e
ly
,
a
nd
ta
r
ge
te
d
in
te
r
ve
nt
i
on
s
[
20]
.
B
y
br
in
gi
ng
pr
e
di
c
t
iv
e
a
na
ly
ti
c
s
in
to
pl
a
y,
w
e
’
ll
s
ha
r
pe
n
our
yi
e
ld
f
o
r
e
c
a
s
ti
ng
,
le
t
ti
ng
us
pl
a
n
e
a
r
l
i
e
r
,
pl
a
n
s
m
a
r
te
r
,
a
nd
s
ync
be
tt
e
r
w
it
h
m
a
r
ke
t
ne
e
ds
[
2
1]
.
I
n pa
r
a
ll
e
l,
in
te
l
li
ge
n
t
s
ys
te
m
s
pa
i
r
e
d
w
it
h de
e
p
da
t
a
a
na
ly
s
is
w
il
l
f
in
e
-
tu
ne
ou
r
us
e
o
f
r
e
s
our
c
e
s
,
e
s
pe
c
ia
ll
y
w
he
n
it
c
om
e
s
to
f
ig
ht
in
g
d
is
e
a
s
e
w
hi
le
s
ti
ll
dr
i
vi
n
g
dow
n
th
e
bo
tt
om
l
in
e
o
f
ope
r
a
ti
ona
l
c
os
ts
.
L
e
ve
r
a
g
in
g
A
I
,
th
e
p
r
oj
e
c
t
s
e
e
ks
to
w
e
a
ve
a
p
r
oduc
ti
on
ne
tw
or
k
f
or
th
e
h
om
-
th
on
g
ba
na
na
th
a
t’
s
not
ju
s
t
le
a
ne
r
a
nd
m
o
r
e
r
e
s
il
ie
nt
,
but
a
ls
o
m
or
e
pr
of
i
ta
bl
e
.
T
he
goa
l
f
r
om
th
e
ou
ts
e
t
a
n
d
a
t
e
ve
r
y
s
te
p
is
to
li
f
t
th
e
c
om
pe
ti
ti
ve
ne
s
s
o
f
T
ha
i
a
gr
ic
ul
t
ur
e
in
th
e
g
lo
ba
l
a
r
e
na
a
nd,
j
us
t
a
s
c
r
uc
ia
ll
y
,
t
o
bo
os
t
t
he
in
c
om
e
s
o
f
t
he
f
a
r
m
e
r
s
w
hos
e
w
o
r
k
m
a
ke
s
t
ha
t
a
gr
ic
ul
tu
r
e
pos
s
ib
le
.
2.
M
E
T
H
O
D
I
n
th
is
s
e
c
ti
o
n
pr
o
vi
de
s
a
n
e
nha
nc
e
d
m
a
c
hi
ne
vi
s
i
on
s
ys
te
m
w
it
h
a
de
e
p
le
a
r
ni
ng
a
p
pr
oa
c
h
to
de
te
r
m
in
e
th
e
di
s
e
a
s
e
s
in
h
om
-
th
ong
ba
na
na
s
a
s
s
how
n
in
F
ig
ur
e
1
.
R
G
B
i
m
a
ge
s
o
bt
a
in
e
d
f
r
o
m
hom
-
th
o
ng
ba
na
na
pl
a
nt
a
ti
ons
i
n T
ha
i
la
nd a
r
e
a
n
not
a
te
d t
o
f
or
m
a
tr
a
i
ni
ng
da
ta
s
e
t
a
nd
te
s
ti
ng
da
ta
s
e
t.
T
he
s
e
i
m
a
ge
s
a
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
. 15, No. 1, Febr
ua
r
y 2026
:
568
-
579
570
ta
ke
n
in
bot
h
opt
i
m
a
l
a
nd
s
u
bopt
im
a
l
w
e
a
th
e
r
c
ondi
ti
ons
.
A
s
m
e
nt
io
ne
d
in
th
e
p
r
e
vi
o
us
s
e
c
ti
on,
th
e
de
e
p
le
a
r
ni
n
g
m
ode
ls
a
r
e
de
ve
lo
pe
d
us
in
g
th
e
t
r
a
in
in
g
da
ta
s
e
t,
a
n
d
th
e
pe
r
f
or
m
a
nc
e
a
na
ly
s
is
of
t
he
s
e
gm
e
nt
a
t
io
n
in
s
ta
nc
e
w
a
s
c
o
nduc
te
d
on
th
e
te
s
ti
ng
da
ta
s
e
t.
F
o
ll
ow
in
g
pr
oc
e
s
s
in
g,
th
e
in
f
or
m
a
ti
o
n
w
a
s
ga
th
e
r
e
d
t
o
s
uppor
t
t
he
m
a
na
ge
m
e
n
t
o
f
hom
-
th
ong
ba
na
na
f
a
r
m
in
g b
y ga
th
e
r
in
g
di
s
e
a
s
e
da
ta
.
F
ig
ur
e
1. T
he
m
e
th
odol
ogy of
t
he
m
ode
l
of
our
w
or
k
2.1. Dat
a c
ol
le
c
t
io
n
I
n
th
is
da
ta
c
ol
le
c
ti
on
in
th
is
s
tu
dy,
w
e
c
ol
le
c
te
d
da
t
a
on
ba
na
na
di
s
e
a
s
e
ph
e
not
ype
s
f
r
om
T
ha
i
hom
-
th
ong
ba
na
na
gr
ow
e
r
s
f
r
om
th
e
ti
m
e
of
pl
a
nt
in
g
unt
il
t
he
y
w
e
r
e
r
e
a
dy
f
or
m
a
r
ke
t
de
m
a
nd,
w
hi
c
h
is
a
ppr
oxi
m
a
te
ly
9
-
10
m
ont
hs
b
e
f
or
e
th
e
c
r
op
c
oul
d
b
e
ha
r
ve
s
te
d.
I
n
a
ddi
ti
on
to
r
e
c
or
di
ng
ba
na
na
de
n
s
it
y,
a
s
il
lu
s
tr
a
te
d
in
F
ig
ur
e
2,
b
a
na
na
phe
not
ype
d
a
ta
c
ol
le
c
ti
on
c
a
n
b
e
c
a
r
r
ie
d
out
in
a
va
r
ie
ty
of
s
e
tt
in
gs
,
in
c
lu
di
ng
th
os
e
w
it
h
va
r
io
us
li
ght
a
nd
w
e
a
th
e
r
c
ondi
ti
ons
.
D
a
ta
i
s
e
s
s
e
n
ti
a
l
to
m
a
c
hi
ne
le
a
r
ni
ng.
G
a
th
e
r
in
g,
la
be
li
ng,
a
nd
a
na
ly
z
in
g
da
ta
a
r
e
one
of
th
e
de
e
p
ne
ur
a
l
ne
twor
ks
(
D
N
N
)
a
lg
or
it
hm
'
s
pr
im
a
r
y
pr
e
-
pr
oc
e
s
s
in
g
r
e
s
pons
ib
il
it
ie
s
.
F
ig
ur
e
1
il
lu
s
tr
a
te
s
it
s
s
e
ve
n
il
ln
e
s
s
f
e
a
tu
r
e
s
.
I
n
th
is
s
te
p,
im
a
ge
s
a
r
e
u
s
e
d
a
s
da
ta
.
E
ve
r
y
pi
c
tu
r
e
is
f
r
om
a
n
or
c
ha
r
d
of
ba
na
na
s
.
T
he
m
a
in
go
a
l
of
th
e
s
ys
te
m
de
s
c
r
ib
e
d
in
th
is
pa
p
e
r
is
to
r
e
c
ogni
z
e
obj
e
c
ts
in
th
is
c
a
s
e
,
th
e
di
s
e
a
s
e
phe
not
ype
s
of
hom
-
th
ong
ba
n
a
na
s
f
r
om
th
e
obt
a
in
e
d
phot
os
.
T
he
a
ut
om
a
ti
c
id
e
nt
if
ic
a
ti
on
of
th
e
s
e
it
e
m
s
in
c
om
put
e
r
vi
s
io
n
pr
e
s
e
nt
s
th
is
di
f
f
ic
ul
ty
.
I
t
f
ol
lo
w
s
th
a
t
th
is
s
ta
ge
m
a
y
in
vol
ve
th
e
us
e
of
s
om
e
A
I
a
lg
or
it
hm
s
.
F
or
th
is
s
tr
a
te
gy
to
w
or
k
e
f
f
e
c
ti
ve
ly
,
a
lo
t
of
da
ta
i
s
ne
e
de
d.
A
s
il
lu
s
tr
a
te
d
in
F
ig
ur
e
2,
w
e
us
e
d
R
obo
f
lo
w
la
be
li
ng
s
of
twa
r
e
f
or
a
na
ly
s
is
of
s
e
ve
n
il
ln
e
s
s
e
s
,
a
pr
ogr
a
m
th
a
t
c
a
n
m
a
nua
ll
y
a
nnot
a
te
th
e
s
e
obj
e
c
t
s
,
to
ga
th
e
r
a
nd
a
nnot
a
t
e
2,576
phot
o
s
of
hom
-
th
ong
ba
na
na
di
s
or
de
r
s
in
or
de
r
to
a
c
c
om
pl
is
h t
hi
s
goa
l.
F
ig
ur
e
2. T
he
da
ta
of
a
nnot
a
ti
on
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
n ar
ti
fi
c
ia
l
in
te
ll
ig
e
nc
e
t
e
c
hnol
ogy
f
or
pr
om
ot
in
g hom
-
th
ong b
anana
…
(
R
at
s
am
e
s
T
anv
e
e
nu
k
ool
)
571
I
n
th
is
in
s
ta
nc
e
,
w
e
ha
v
e
de
c
id
e
d
to
us
e
a
la
be
li
ng
pl
a
tf
or
m
.
T
hi
s
m
a
ke
s
it
pos
s
ib
le
to
s
to
r
e
a
nnot
a
ti
ons
in
a
va
r
ie
ty
of
f
or
m
a
ts
.
A
ll
c
om
pl
e
te
d
a
nnot
a
ti
ons
in
th
e
c
om
m
on
obj
e
c
ts
in
c
ont
e
xt
(
C
O
C
O
)
f
or
m
a
t
m
us
t
be
s
a
ve
d
in
a
s
ui
ta
bl
e
f
il
e
f
or
m
a
t
f
or
la
te
r
vi
e
w
in
g
in
or
de
r
f
or
th
e
tr
a
in
in
g
pr
oc
e
s
s
to
g
e
ne
r
a
te
pr
e
di
c
ti
ve
m
ode
ls
[
22]
.
T
he
s
of
twa
r
e
pr
ogr
a
m
s
s
e
le
c
te
d
f
or
D
N
N
m
ode
li
ng
a
nd
te
s
ti
ng
a
r
e
c
om
pa
ti
bl
e
w
it
h
th
is
f
or
m
a
t.
A
c
or
r
e
s
ponding
J
S
O
N
f
il
e
in
c
lu
di
ng
a
f
r
a
m
e
w
or
k f
or
r
e
c
or
di
ng
th
e
obj
e
c
t
c
a
te
gor
y
a
nd
lo
c
a
ti
on
of
e
a
c
h
a
nnot
a
ti
on
is
in
c
lu
de
d
w
it
h
e
ve
r
y
c
ol
le
c
ti
on
of
a
nnot
a
te
d
phot
os
in
th
e
C
O
C
O
f
or
m
a
t.
A
s
w
il
l
be
c
ove
r
e
d
in
th
e
r
e
s
ul
ts
s
e
c
ti
on
la
te
r
,
th
e
C
O
C
O
s
ta
nda
r
d
a
ls
o
o
f
f
e
r
s
im
por
ta
nt
m
e
tr
ic
s
r
e
qui
r
e
d
to
a
s
s
e
s
s
th
e
a
c
c
ur
a
c
y of
t
he
m
ode
l
[
23]
.
2.2. Dat
a p
r
e
p
ar
at
io
n
I
m
por
ta
nt
s
te
ps
in
th
i
s
s
e
c
ti
on
in
c
lu
de
c
onf
ir
m
in
g
th
a
t
th
e
r
e
is
e
nough
da
ta
f
or
m
ode
l
tr
a
in
in
g,
or
ga
ni
z
in
g
th
e
da
ta
,
a
nd a
ugm
e
nt
in
g
th
e
d
a
ta
w
it
h
th
e
R
obof
lo
w
to
ol
[
24]
.
A
s
n
a
ps
hot
of
R
obof
lo
w
s
how
s
a
n
ove
r
vi
e
w
of
t
he
da
ta
s
e
t.
T
h
e
pr
e
-
pr
oc
e
s
s
in
g
c
a
r
r
ie
d out f
or
t
hi
s
i
nve
s
ti
ga
ti
on i
s
a
ls
o i
ll
us
tr
a
te
d i
n F
ig
ur
e
3.
F
ig
ur
e
3. T
he
da
ta
pr
e
pa
r
a
ti
on f
or
a
ugm
e
nt
a
ti
on
T
he
R
obo
f
lo
w
pl
a
tf
or
m
f
a
c
il
it
a
te
s
e
ve
r
y
a
s
pe
c
t
of
da
ta
,
in
c
l
udi
ng
pr
e
-
pr
oc
e
s
s
in
g,
a
ugm
e
nt
a
ti
on,
a
nnot
a
ti
on,
or
ga
ni
z
a
ti
on
,
m
ode
l
tr
a
in
in
g,
a
nd
d
e
pl
oym
e
nt
.
T
h
is
s
tu
dy'
s
pr
e
-
pr
oc
e
s
s
in
g
s
te
p
w
a
s
c
on
c
e
r
ne
d
onl
y
w
it
h
r
e
s
iz
in
g
a
ny
s
our
c
e
im
a
ge
s
to
f
ol
lo
w
th
e
s
pe
c
if
ic
a
ti
ons
of
th
e
ta
r
ge
t
da
ta
s
e
t.
I
n
te
r
m
s
o
f
pr
e
-
pr
oc
e
s
s
in
g,
th
e
s
our
c
e
im
a
ge
s
unde
r
w
e
nt
a
ut
om
a
ti
c
or
ie
nt
a
ti
on
c
or
r
e
c
ti
on
a
nd
r
e
s
c
a
li
ng
to
a
c
hi
e
v
e
a
nor
m
a
li
z
e
d
s
iz
e
of
640
×
640
pi
xe
ls
to
m
a
tc
h
th
e
ot
he
r
nor
m
a
li
z
e
d
im
a
ge
s
in
th
e
da
ta
s
e
t,
w
hi
c
h
a
id
e
d
in
e
ns
ur
in
g
a
b
e
tt
e
r
pe
r
f
or
m
a
nc
e
in
te
r
m
s
of
c
om
put
a
ti
ona
l
ti
m
e
.
I
n
te
r
m
s
of
a
ugm
e
nt
a
ti
on,
w
e
a
ppl
ie
d
a
va
r
ie
ty
of
te
c
hni
que
s
to
in
c
r
e
a
s
e
th
e
da
ta
s
e
t
s
iz
e
. T
hi
s
in
c
lu
de
d
a
ll
of
th
e
im
a
ge
s
be
in
g
f
li
ppe
d
bot
h
hor
iz
ont
a
ll
y
a
nd
ve
r
ti
c
a
ll
y,
a
s
w
e
ll
a
s
be
in
g
r
ot
a
te
d
90
de
gr
e
e
s
in
e
it
he
r
di
r
e
c
ti
on
to
a
s
s
is
t
th
e
m
ode
l
in
id
e
nt
if
yi
ng
obj
e
c
ts
a
t
a
ngl
e
s
.
W
e
m
odi
f
ie
d
th
e
br
ig
ht
ne
s
s
of
th
e
im
a
ge
s
by
±15
%
in
or
de
r
to
c
r
e
a
te
di
ve
r
s
it
y
in
li
ght
in
g
s
it
ua
ti
ons
th
a
t
c
oul
d
in
c
r
e
a
s
e
th
e
m
ode
l’
s
ove
r
a
ll
pe
r
f
or
m
a
nc
e
r
e
ga
r
di
n
g
di
f
f
e
r
e
nt
li
ght
in
g
e
nvi
r
onm
e
nt
s
.
I
n
te
r
m
s
o
f
c
ol
or
,
w
e
a
dj
us
te
d
s
a
tu
r
a
ti
on
by
±25%
to
e
nha
nc
e
th
e
r
e
s
il
ie
n
c
y
of
th
e
m
ode
l
to
c
ol
or
c
ha
nge
s
.
F
in
a
ll
y,
w
e
r
a
ndoml
y
c
r
oppe
d
20%
a
w
a
y
f
r
om
a
ll
of
th
e
im
a
ge
s
to
tr
y
a
n
d
te
a
c
h
th
e
m
ode
l
to
de
t
e
c
t
obj
e
c
t
s
th
a
t
w
e
r
e
onl
y
pa
r
ti
a
ll
y
vi
s
ib
le
.
B
y
us
in
g
a
ugm
e
nt
a
ti
on,
w
e
in
c
r
e
a
s
e
d
th
e
or
ig
in
a
l
da
ta
s
e
t
of
2,576
im
a
ge
s
to
6,184
im
a
ge
s
.
A
lt
hough
th
e
f
r
e
e
ve
r
s
io
n
o
f
R
obof
lo
w
li
m
it
s
your
da
ta
s
e
t
s
iz
e
a
f
te
r
a
ugm
e
nt
a
ti
on,
6,184
is
s
ti
ll
a
good number
f
or
t
r
a
in
in
g a
m
ode
l.
R
obof
lo
w
or
ga
ni
z
a
ti
on
c
a
pa
bi
li
ti
e
s
m
a
de
it
e
a
s
ie
r
to
a
nnot
a
t
e
th
e
im
a
ge
s
a
nd
gr
oup
th
e
m
in
to
a
c
ont
r
ol
s
ys
te
m
to
c
r
e
a
t
e
a
c
om
pr
e
h
e
ns
iv
e
d
a
ta
s
e
t
of
a
ugm
e
nt
e
d
a
nd
a
nnot
a
te
d
im
a
g
e
s
.
F
ur
th
e
r
m
or
e
,
th
e
pl
a
tf
or
m
a
ls
o
s
im
pl
if
ie
d
th
e
w
or
kf
lo
w
of
m
a
c
hi
ne
le
a
r
ni
ng
be
c
a
us
e
it
s
e
pa
r
a
te
d
th
e
pr
oc
e
s
s
e
d
da
ta
s
e
t
in
to
tr
a
in
in
g,
va
li
da
ti
on,
a
nd
te
s
ti
ng
s
ubs
e
ts
a
ut
om
a
ti
c
a
ll
y.
F
ig
ur
e
3
s
how
s
a
s
na
p
s
hot
of
a
R
obof
lo
w
in
te
r
f
a
c
e
s
how
in
g
th
e
d
a
ta
s
e
t
ove
r
vi
e
w
a
nd
th
e
pr
e
pr
oc
e
s
s
in
g
c
onf
ig
ur
a
ti
ons
th
a
t
w
e
us
e
d
in
th
e
r
e
s
e
a
r
c
h.
A
f
te
r
w
e
c
ol
le
c
te
d,
a
nnot
a
te
d,
pr
e
pr
oc
e
s
s
e
d
a
nd
a
ugm
e
nt
e
d
th
e
da
t
a
s
e
t,
our
da
ta
s
e
t
w
a
s
r
e
a
dy
f
or
m
a
c
hi
ne
le
a
r
ni
ng
e
xpe
r
im
e
nt
a
ti
on a
nd t
he
de
ve
lo
pm
e
nt
of
a
n obje
c
t
de
te
c
ti
on mode
l.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
. 15, No. 1, Febr
ua
r
y 2026
:
568
-
579
572
2.3. Ap
p
li
c
at
io
n
of
ar
t
if
ic
ia
l
in
t
e
ll
ig
e
n
c
e
h
om
-
t
h
on
g b
an
a
n
a
d
is
e
as
e
s
T
he
Y
O
L
O
v12
m
ode
l
w
a
s
tr
a
in
e
d
on
a
n
N
V
I
D
I
A
J
e
ts
on
O
r
in
N
a
no
de
ve
lo
pm
e
nt
ki
t
w
it
h
8
G
B
o
f
R
A
M
. T
he
m
ode
l
w
a
s
t
r
a
in
e
d w
it
h P
yT
or
c
h, a
ba
c
ke
nd f
r
a
m
e
w
or
k t
ha
t
r
uns
on L
in
ux. T
o gi
ve
l
e
s
s
c
ha
nc
e
of
ove
r
f
it
ti
ng,
th
e
m
ode
l
w
a
s
tr
a
in
e
d
w
it
h
0.001
le
a
r
ni
ng
r
a
te
,
ba
t
c
h
s
e
t
to
32
a
nd
dr
opout
s
e
t
to
0.5.
T
he
m
ode
l
w
a
s
tr
a
in
e
d
on
f
or
ove
r
a
th
ou
s
a
nd
e
po
c
hs
.
T
r
a
in
in
g
w
a
s
s
t
oppe
d
w
he
n
162
e
po
c
hs
w
e
r
e
r
e
a
c
he
d
if
th
e
va
li
da
ti
on
s
e
t
di
d
not
s
e
e
im
pr
ove
m
e
nt
f
or
20
m
or
e
e
poc
hs
.
S
uc
h
a
s
te
p
w
a
s
in
tr
oduc
e
d
in
or
de
r
to
e
nha
nc
e
ge
ne
r
a
li
z
a
ti
on a
nd a
voi
d ove
r
f
it
ti
ng
th
e
t
r
a
in
in
g s
e
t.
I
n
bot
h m
o
de
ls
, t
he
l
e
a
r
ni
ng r
a
te
w
a
s
i
ni
ti
a
ll
y s
e
t
to
0.01
.
M
om
e
nt
um
a
nd
d
r
opout
w
e
ig
ht
pa
r
a
m
e
te
r
s
f
or
bo
th
m
ode
ls
s
to
od
a
t
0.937
a
nd
0.0005,
r
e
s
pe
c
ti
ve
ly
.
A
le
a
r
ni
ng
r
a
te
is
e
s
s
e
nt
ia
l
in
s
uc
h
m
ode
ls
in
or
de
r
to
a
voi
d
ov
e
r
f
it
ti
ng
in
th
e
tr
a
in
in
g
da
ta
s
e
t.
I
n
f
a
c
t,
it
c
a
n
in
de
pe
nde
nt
ly
de
c
id
e
upon opti
m
a
l
va
lu
e
s
f
or
t
r
a
in
a
bi
li
ty
i
n or
d
e
r
t
o a
voi
d s
uc
h pr
obl
e
m
s
. A
ddi
ng a
w
a
r
m
-
up
s
te
p
c
a
n
be
e
s
s
e
nt
ia
l
in
or
de
r
to
a
voi
d
s
e
tt
li
ng
in
a
lo
c
a
l
m
in
im
um
poi
nt
.
A
t
th
e
m
om
e
nt
,
th
e
r
e
is
a
m
om
e
nt
um
of
0.8
a
nd
bi
a
s
le
a
r
ni
ng
r
a
te
of
0.1 a
nd
th
is
s
e
c
ti
on w
il
l
c
ont
in
ue
to
s
um
m
a
r
iz
e
th
e
to
ol
s
s
pe
c
if
ie
d
w
it
hi
n t
he
c
ont
e
xt
of
m
a
ki
ng a
m
ode
l
to
m
a
ke
a
pr
e
di
c
ti
on.
P
yt
hon
is
one
of
th
e
m
os
t
us
e
d
pr
og
r
a
m
m
in
g
la
ngua
ge
s
a
nd
w
id
e
ly
us
e
d
f
or
m
a
c
hi
ne
le
a
r
ni
ng.
R
e
a
s
on
s
f
or
th
is
in
c
lu
de
s
im
pl
ic
it
y,
pa
c
ka
g
e
s
a
v
a
il
a
bl
e
f
or
e
f
f
ic
ie
nc
y
in
s
ol
vi
ng
pr
obl
e
m
s
,
a
nd
th
e
num
be
r
of
de
ve
lo
pe
r
s
[
25]
. M
a
ny l
ib
r
a
r
ie
s
a
r
e
a
va
il
a
bl
e
t
ha
t
a
r
e
va
lu
a
bl
e
t
o buil
d
DNN
w
he
n de
ve
lo
pi
ng i
n P
y
th
on. T
he
s
tu
dy'
s
m
a
in
f
oc
us
is
on
obj
e
c
t
de
te
c
ti
on
is
s
ue
s
,
w
it
h
a
pa
r
ti
c
ul
a
r
f
oc
us
on
ba
na
na
di
s
e
a
s
e
a
na
ly
s
i
s
.
B
e
c
a
u
s
e
of
th
is
,
m
a
ny
e
xc
e
ll
e
nt
P
yt
hon
m
odul
e
s
a
r
e
a
va
il
a
bl
e
.
W
e
d
e
c
id
e
d
to
us
e
D
e
te
c
tr
on2,
w
hi
c
h
ut
il
iz
e
s
th
e
P
yT
or
c
h m
odul
e
. A
gr
e
a
te
r
va
r
ie
ty
of
m
a
c
hi
ne
l
e
a
r
ni
ng i
s
s
ue
s
m
a
y be
s
ol
ve
d w
it
h t
he
P
yT
or
c
h l
ib
r
a
r
y, w
hi
c
h
is
e
xt
e
ns
iv
e
ly
ut
il
iz
e
d i
n r
e
s
e
a
r
c
h. A
dva
nt
a
ge
s
of
t
he
l
ib
r
a
r
y i
nc
lu
de
s
im
pl
ic
it
y a
nd f
le
xi
bi
li
ty
.
O
f
te
n P
yT
or
c
h
is
c
om
pa
r
e
d
in
s
om
e
w
a
y
w
it
h
ot
he
r
pr
ogr
a
m
s
th
a
t
pe
r
f
or
m
s
i
m
il
a
r
f
unc
ti
ons
.
I
n
pa
r
ti
c
ul
a
r
,
bot
h
T
e
ns
or
F
lo
w
pr
ogr
a
m
s
a
r
e
us
e
d
to
bui
ld
A
I
s
ol
ut
io
ns
be
c
a
us
e
th
e
r
e
is
no
c
onc
r
e
te
w
a
y
to
de
m
ons
tr
a
te
w
hi
c
h
pr
ogr
a
m
is
s
upe
r
io
r
or
be
s
t
[
26]
. D
e
te
c
tr
on2 a
ls
o ha
s
ve
r
y m
ode
r
n obje
c
t
d
e
te
c
ti
on a
nd s
e
gm
e
nt
a
ti
on a
lg
or
it
hm
s
[
27]
.
T
he
D
e
te
c
tr
on2
pa
c
ka
ge
'
s
in
s
ta
ll
a
ti
on
a
nd
s
e
tu
p
a
r
e
e
nvi
r
onm
e
nt
-
de
pe
nde
nt
a
nd
not
ne
c
e
s
s
a
r
il
y
s
im
pl
e
.
B
e
c
a
us
e
it
is
s
o
s
im
pl
e
to
upgr
a
de
f
r
om
f
r
e
e
to
pr
e
m
iu
m
,
w
e
c
a
m
e
to
th
e
c
onc
lu
s
io
n
th
a
t
th
e
G
oogl
e
C
ol
a
b
e
nvi
r
onm
e
nt
w
a
s
a
good
c
hoi
c
e
f
o
r
r
e
s
e
a
r
c
h
a
nd
e
xpl
or
a
to
r
y
te
s
ti
ng
[
28
]
.
G
oogl
e
C
ol
a
b
ha
s
a
nu
m
be
r
of
to
ol
s
th
a
t
f
a
c
il
it
a
te
a
nd
s
pe
e
d
up
th
e
c
onf
ig
ur
a
ti
on
a
nd
in
s
ta
l
la
ti
on
pr
oc
e
s
s
of
th
e
pa
c
ka
ge
.
T
he
pr
e
m
is
e
of
th
is
te
c
hnol
ogy
is
th
a
t
in
di
vi
dua
ls
c
a
n
r
un
a
vi
r
tu
a
l
m
a
c
hi
ne
f
o
r
a
de
s
ig
na
te
d
pe
r
io
d
of
ti
m
e
,
va
r
ia
bl
e
w
it
hi
n
th
e
f
r
e
e
ve
r
s
io
n,
a
nd
th
e
de
ve
lo
pe
r
is
not
not
if
ie
d
how
lo
ng,
unt
il
th
e
ir
e
ve
nt
ua
l
di
s
c
onne
c
ti
on.
T
he
pr
o
ve
r
s
io
n i
s
e
li
gi
bl
e
f
or
t
he
l
onge
r
t
e
r
m
. T
he
vi
r
tu
a
l
m
a
c
hi
ne
s
e
s
s
io
n c
a
n t
e
r
m
in
a
te
, a
nd t
he
r
e
s
ul
ts
m
ig
ht
not
be
s
a
ve
d,
de
p
e
ndi
ng
on
th
e
m
ode
l
tr
a
in
in
g.
B
y
im
pl
e
m
e
nt
in
g
a
ve
r
y
ti
ny
pi
e
c
e
of
c
od
e
to
s
a
ve
th
e
m
ode
l
c
he
c
kpoi
nt
a
t
di
f
f
e
r
e
nt
s
ta
ge
s
of
th
e
tr
a
in
in
g
pr
oc
e
s
s
,
th
is
i
s
s
ue
c
a
n
b
e
f
ix
e
d.
T
e
c
hni
c
a
ll
y
s
pe
a
ki
ng,
if
w
e
a
r
e
a
w
a
r
e
of
a
nd c
om
pr
e
he
nd t
he
r
e
s
tr
ic
ti
ons
of
G
oogl
e
C
ol
a
b'
s
f
r
e
e
e
di
ti
on, we
c
ons
id
e
r
t
he
f
a
nt
a
s
ti
c
i
ns
tr
um
e
nt
a
t
your
di
s
pos
a
l
to
a
s
s
is
t
you in c
om
pl
e
ti
ng t
he
t
a
s
k a
t
ha
nd.
2.4. P
e
r
f
or
m
an
c
e
e
val
u
at
io
n
F
iv
e
di
f
f
e
r
e
nt
c
r
it
e
r
ia
w
e
r
e
us
e
d
f
or
e
va
lu
a
ti
ng
th
e
pe
r
f
or
m
a
n
c
e
of
th
e
in
s
ta
nc
e
s
e
gm
e
nt
a
ti
on
ta
s
k
f
or
th
e
R
obof
lo
w
,
Y
O
L
O
v11,
a
nd
Y
O
L
O
v12
m
ode
l
s
.
A
c
c
ur
a
c
y
w
a
s
c
ons
id
e
r
e
d
a
s
th
e
r
a
ti
o
of
th
e
c
or
r
e
c
tl
y
pr
e
di
c
te
d
pos
it
iv
e
e
ve
nt
s
to
th
e
to
ta
l
pos
it
iv
e
e
ve
nt
s
pr
e
di
c
t
e
d,
a
s
c
la
r
if
ie
d
in
(
1)
.
T
he
r
e
c
a
ll
m
e
a
s
ur
e
,
s
how
n
in
(
2)
,
w
a
s
us
e
d
f
or
c
om
put
in
g
th
e
pe
r
c
e
nt
a
ge
of
c
or
r
e
c
t
id
e
nt
if
ic
a
ti
ons
f
or
th
e
pos
it
iv
e
e
ve
nt
s
.
T
he
s
e
c
r
it
e
r
ia
w
e
r
e
m
e
a
n
a
ve
r
a
ge
pr
e
c
is
io
n
(
m
A
P
)
a
t
0.5
in
te
r
s
e
c
ti
on
ove
r
uni
on
(
I
oU
)
,
a
c
c
ur
a
c
y,
r
e
c
a
ll
,
a
r
e
a
unde
r
th
e
c
ur
ve
(
A
U
C
)
in
(
3)
f
or
th
e
r
e
c
e
iv
e
r
op
e
r
a
ti
ng
c
ha
r
a
c
te
r
is
ti
c
(
R
O
C
)
,
a
nd
in
f
e
r
e
nc
e
ti
m
e
.
A
s
s
how
n
in
(
4)
,
m
A
P
is
th
e
a
ve
r
a
ge
of
th
e
A
P
v
a
lu
e
s
ove
r
k
obj
e
c
t
c
a
te
gor
ie
s
. A
50%
ove
r
la
p
c
r
it
e
r
io
n
be
tw
e
e
n
th
e
pr
e
di
c
te
d
a
nd
a
c
tu
a
l
it
e
m
bound
a
r
y/
bounding
boxe
s
w
a
s
ut
il
iz
e
d
f
or
A
P
c
om
put
a
ti
on.
A
s
s
how
n
in
(
5)
,
th
e
A
U
C
f
or
e
a
c
h
m
ode
l
w
a
s
c
a
lc
ul
a
te
d.
T
he
A
U
C
ta
ke
s
in
to
a
c
c
ount
a
ll
pos
s
ib
le
c
ut
-
of
f
va
lu
e
s
.
T
he
s
pe
e
d
of
in
f
e
r
e
nc
e
,
or
th
e
a
m
ount
of
ti
m
e
r
e
qui
r
e
d
to
a
na
ly
z
e
e
a
c
h
in
di
vi
dua
l
im
a
ge
,
a
ls
o
a
ppe
a
r
s
to
be
a
m
e
a
s
ur
e
of
th
e
m
ode
l'
s
c
a
pa
c
it
y t
o pr
oduc
e
pr
e
di
c
ti
on output
s
. T
h
e
s
e
m
e
tr
ic
s
a
r
e
c
a
lc
ul
a
te
d us
in
g:
=
+
(
1)
=
+
(
2)
=
=
+
+
(
3)
=
(
1
)
∑
(
)
=
0
(
4)
=
∫
(
)
−
1
(
)
1
0
(
5
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
n ar
ti
fi
c
ia
l
in
te
ll
ig
e
nc
e
t
e
c
hnol
ogy
f
or
pr
om
ot
in
g hom
-
th
ong b
anana
…
(
R
at
s
am
e
s
T
anv
e
e
nu
k
ool
)
573
w
hi
c
h
in
di
c
a
te
s
in
s
ta
nc
e
s
of
tr
ue
po
s
it
iv
e
,
f
a
ls
e
pos
it
iv
e
, a
nd
f
a
ls
e
ne
ga
ti
ve
us
in
g
th
e
le
tt
e
r
s
,
,
a
nd
.
(
)
,
w
he
r
e
is
th
e
to
ta
l
num
be
r
of
obj
e
c
t
c
la
s
s
e
s
,
is
th
e
a
ve
r
a
ge
a
c
c
ur
a
c
y
f
or
th
e
is
c
la
s
s
a
m
ong
th
e
s
e
c
la
s
s
e
s
.
T
he
a
r
e
a
unde
r
th
e
pr
e
c
is
io
n
-
r
e
c
a
ll
c
ur
ve
f
or
a
c
la
s
s
is
known
a
s
th
e
a
ve
r
a
ge
pr
e
c
is
io
n
(
)
.
s
ta
nds
f
or
f
a
ls
e
pos
it
iv
e
r
a
te
,
f
or
tr
ue
pos
it
iv
e
r
a
te
f
or
a
pa
r
ti
c
ul
a
r
(
s
in
gl
e
)
im
a
ge
,
a
nd
f
or
th
e
m
ode
l'
s
in
f
e
r
e
nc
e
t
im
e
(
in
s
e
c
onds
)
.
3.
R
E
S
U
L
T
S
A
N
D
D
I
S
C
U
S
S
I
O
N
3.1. Com
p
ar
at
iv
e
p
e
r
f
or
m
an
c
e
m
od
e
l
f
or
h
o
m
-
t
h
on
g
d
is
e
as
e
s
e
val
u
at
io
n
T
he
im
a
ge
s
of
2,576
phot
os
in
th
e
c
om
pl
e
te
R
G
B
im
a
ge
da
t
a
s
e
t
w
e
r
e
ta
ke
n
by
T
h
a
i
ba
na
n
a
gr
ow
e
r
s
in
hom
-
th
ong
.
T
e
n
pe
r
c
e
nt
a
r
e
us
e
d
f
or
va
li
da
ti
on,
s
e
ve
nt
y
pe
r
c
e
nt
a
r
e
us
e
d
f
or
tr
a
in
in
g,
a
nd
twe
nt
y
pe
r
c
e
nt
a
r
e
us
e
d
f
or
te
s
ti
ng.
I
n
or
de
r
to
de
te
c
t
hom
-
th
ong
ba
na
n
a
il
ln
e
s
s
,
th
e
m
ode
l
w
a
s
tr
a
in
e
d
a
c
r
os
s
162
e
po
c
hs
on
th
e
c
om
pl
e
te
da
ta
s
e
t
in
a
bout
7.6
hour
s
.
U
s
in
g
a
va
r
ie
ty
of
m
e
tr
ic
s
,
in
c
lu
di
ng
box
lo
s
s
(
box_los
s
)
,
s
e
gm
e
nt
a
ti
on
(
s
e
g_l
o
s
s
)
,
c
l
a
s
s
if
ic
a
ti
on
(
c
l
s
_l
os
s
)
,
a
nd
f
oc
a
l
di
f
f
us
io
n
(
df
l_
lo
s
s
)
,
th
e
tr
a
in
in
g
a
nd
va
li
da
ti
on
s
e
t
gr
a
phs
i
n F
ig
ur
e
4
of
R
obo
f
lo
w
,
F
ig
ur
e
5
of
Y
O
L
O
v11, a
nd
F
ig
ur
e
6
of
Y
O
L
O
v12
s
how
how
t
he
m
ode
l'
s
pe
r
f
or
m
a
nc
e
im
pr
ove
d.
B
y
c
ont
r
ol
li
ng
la
ye
r
im
ba
la
n
c
e
dur
in
g
tr
a
in
in
g
w
it
h
a
ta
r
ge
t
lo
s
s
f
unc
ti
on,
th
e
s
e
m
e
a
s
ur
e
s
e
v
a
lu
a
te
t
he
m
ode
l'
s
a
bi
li
ty
t
o l
oc
a
te
i
ll
ne
s
s
i
n
hom
-
th
ong
ba
na
na
s
by l
a
y
e
r
.
F
ig
ur
e
4. T
r
a
in
in
g a
nd va
li
da
ti
on s
e
ts
a
r
e
pl
ot
te
d u
s
in
g R
obof
lo
w
t
o di
s
pl
a
y e
poc
hs
F
ig
ur
e
5. T
r
a
in
in
g a
nd va
li
da
ti
on s
e
ts
a
r
e
pl
ot
te
d u
s
in
g Y
O
L
O
v11 to di
s
pl
a
y e
poc
hs
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
. 15, No. 1, Febr
ua
r
y 2026
:
568
-
579
574
F
ig
ur
e
6. T
r
a
in
in
g a
nd va
li
da
ti
on s
e
ts
a
r
e
pl
ot
te
d u
s
in
g Y
O
L
O
v12 to di
s
pl
a
y e
poc
hs
T
he
r
e
s
ul
ts
f
r
om
th
e
R
ob
of
l
o
w
a
r
c
hi
te
c
tu
r
e
in
F
i
gur
e
4
s
how
s
te
a
dy
c
on
ve
r
ge
nc
e
th
r
oug
hout
th
e
tr
a
in
in
g
r
un.
L
os
s
c
ur
ve
s
s
te
a
di
ly
de
c
l
in
e
w
it
h
li
tt
le
va
r
ia
nc
e
,
s
ig
na
li
n
g
e
s
ta
bl
is
he
d
,
s
ta
bl
e
le
a
r
ni
ng.
B
ot
h
tr
a
in
in
g
a
nd
va
li
da
ti
on
lo
s
s
e
s
a
dva
nc
e
in
pa
r
a
l
le
l,
ke
e
p
in
g
ju
s
t
th
e
r
i
ght
a
m
ount
of
d
is
ta
nc
e
,
w
h
ic
h
c
on
f
i
r
m
s
th
e
s
ys
te
m
is
le
a
r
n
in
g
c
or
r
e
c
tl
y
w
it
hout
ve
e
r
i
ng
i
nt
o
o
ve
r
f
i
tt
in
g.
T
he
c
on
ve
r
ge
nc
e
pa
c
e
,
t
hough
,
is
not
ic
e
a
bl
y
m
o
r
e
gr
a
d
ua
l
th
a
n
w
it
h
Y
O
L
O
,
a
nd
th
e
a
r
c
hi
te
c
t
ur
e
onl
y
m
a
na
ge
s
e
xpe
c
te
d
ba
s
e
li
ne
s
c
o
r
e
s
.
T
ur
ni
ng
t
o
th
e
Y
O
L
O
v12
r
e
s
u
lt
s
por
tr
a
ye
d
in
F
i
gur
e
5,
th
e
pe
r
f
o
r
m
a
nc
e
e
c
li
ps
e
s
th
e
R
obof
l
ow
s
te
a
dy
ba
s
e
li
ne
in
le
a
r
ni
ng
s
pe
e
d.
T
he
m
ode
l
s
e
ts
r
e
c
o
r
ds
f
o
r
s
w
if
t
in
it
ia
l
c
onve
r
ge
nc
e
,
r
e
f
le
c
te
d
in
s
te
e
pe
r
,
e
a
r
li
e
r
de
c
li
ne
s
o
f
th
e
lo
s
s
.
T
he
r
e
is
m
i
ld
e
le
va
ti
o
n
in
c
ur
ve
os
c
il
la
t
io
n,
ye
t
th
e
dow
nw
a
r
d
tr
e
nd
r
e
m
a
in
s
ti
ght
e
r
,
a
nd
pe
r
f
or
m
a
nc
e
m
e
t
r
ic
s
ou
tp
a
c
e
th
os
e
f
r
om
R
ob
of
l
ow
.
T
he
t
r
a
in
i
ng
a
nd
va
li
da
ti
on
lo
s
s
e
s
,
m
e
a
nw
hi
le
,
do
not
di
ve
r
ge
e
xc
e
s
s
iv
e
ly
,
a
nd
t
he
ne
tw
or
ks
f
in
is
h
s
ugge
s
ti
ng w
i
de
r
ge
ne
r
a
l
iz
a
ti
o
n a
n
d
im
p
r
ove
d
e
f
f
ic
ie
nc
y
in
c
om
put
a
t
io
n
a
l
l
in
a
f
r
a
c
ti
o
n
of
t
he
ti
m
e
t
he
or
i
gi
na
l
a
r
c
hi
t
e
c
tu
r
e
r
e
q
ui
r
e
d.
T
he
Y
O
L
O
v
12
a
r
c
h
it
e
c
tu
r
e
,
il
lu
s
t
r
a
te
d
in
F
i
gur
e
6,
c
le
a
r
ly
e
x
hi
bi
te
d
th
e
hi
ghe
s
t
ove
r
a
l
l
pe
r
f
o
r
m
a
nc
e
a
c
r
os
s
t
he
r
a
nge
of
a
s
s
e
s
s
e
d
m
e
tr
ic
s
.
I
t
r
e
a
li
z
e
d
t
he
f
a
s
te
s
t
c
onve
r
ge
nc
e
s
pe
e
d
,
f
in
is
hi
ng
w
it
h
t
he
be
s
t
te
r
m
i
na
l
lo
s
s
o
f
a
ny
m
ode
l
e
xa
m
in
e
d.
C
ur
ve
s
f
r
om
th
e
t
r
a
in
in
g
pr
oc
e
s
s
s
how
e
d s
te
a
dy,
t
i
g
ht
be
ha
v
io
r
du
r
in
g
t
he
f
in
a
l
e
poc
h, a
s
i
gn
th
a
t
opt
im
iz
a
ti
o
n
r
e
a
c
he
d
a
s
t
r
ong,
f
in
a
l
pl
a
te
a
u.
T
he
s
m
a
ll
ga
p
b
e
twe
e
n
t
r
a
in
i
ng
a
nd
va
li
da
ti
o
n
lo
s
s
e
s
f
ur
t
he
r
unde
r
s
c
o
r
e
s
t
he
a
r
c
hi
te
c
tu
r
e
’
s
a
b
il
i
ty
to
ge
ne
r
a
l
iz
e
, w
h
ic
h
m
a
k
e
s
i
t
pa
r
ti
c
ul
a
r
ly
a
tt
r
a
c
ti
ve
f
o
r
i
nt
e
g
r
a
ti
o
n i
n
to
a
gr
ic
ul
tu
r
a
l
di
s
e
a
s
e
de
te
c
ti
on
a
ppl
ic
a
ti
o
ns
.
T
a
bl
e
1
hi
ghl
ig
ht
s
how
pr
e
c
is
e
a
nd
m
e
m
or
y
-
e
f
f
ic
ie
nt
th
e
Y
O
L
O
v12
-
ba
s
e
d
th
e
r
m
a
l
im
a
gi
ng
m
ode
l
f
or
de
te
c
ti
on i
s
, w
it
h s
c
or
e
s
of
0.933 a
nd 0.893 f
or
Y
O
L
O
v11,
a
nd 0.873 f
or
R
obof
lo
w
. F
ig
u
r
e
s
4 t
o 6 l
a
y out
two
w
a
ys
to
gr
a
s
p
m
A
P
.
T
he
f
ir
s
t,
m
A
P
50,
m
e
a
s
ur
e
s
a
ve
r
a
g
e
a
c
c
ur
a
c
y
a
t
a
n
I
oU
th
r
e
s
hol
d
of
0.5
on
th
e
gi
ve
n
im
a
ge
s
e
t.
H
e
r
e
,
Y
O
L
O
v11
a
ve
r
a
ge
s
0.832,
w
hi
le
R
o
bof
lo
w
'
s
m
ode
l
a
ve
r
a
ge
s
0.817.
T
he
s
e
c
ond
m
e
a
s
ur
e
,
m
A
P
50
-
95,
c
om
put
e
s
a
c
r
os
s
I
oU
va
lu
e
s
f
r
om
0.5
to
0.95
in
0.5
s
te
ps
.
V
a
lu
e
s
a
bove
0.863
s
how
how
qui
c
kl
y
th
e
Y
O
L
O
v12
-
ba
s
e
d
m
ode
l
c
a
n
de
te
c
t
b
a
na
na
le
a
f
di
s
e
a
s
e
a
c
r
os
s
a
r
a
nge
of
li
ght
in
g
c
ondi
ti
ons
.
T
he
m
ode
ls
b
a
s
e
d
on
Y
O
L
O
v12
w
e
r
e
r
e
la
ti
ve
ly
m
or
e
c
ha
l
le
nge
d
on
s
om
e
m
e
tr
ic
s
,
r
e
f
le
c
t
e
d
in
lo
w
e
r
pe
r
f
or
m
a
nc
e
s
on
tr
a
in
/b
ox_l
os
s
,
tr
a
in
/s
e
g_l
os
s
,
tr
a
in
/c
ls
_l
os
s
,
a
nd
tr
a
in
/d
f
l_
lo
s
s
,
a
s
s
how
n
in
F
ig
ur
e
6.
T
hi
s
im
pl
ie
s
th
a
t
it
ta
ke
s
not
onl
y
s
uf
f
ic
ie
nt
da
ta
,
but
a
ls
o
da
ta
a
ppr
opr
ia
te
f
or
f
a
c
il
it
a
ti
ng
e
f
f
e
c
ti
ve
a
nd
pr
e
c
is
e
de
te
c
ti
on,
to
s
ol
ve
th
e
pr
obl
e
m
pr
e
s
e
nt
e
d
s
uc
c
e
s
s
f
ul
ly
.
T
o
de
s
c
r
ib
e
w
hy
s
om
e
ne
w
m
ode
ls
ba
s
e
d
on
Y
O
L
O
v12
w
e
r
e
de
ve
lo
pe
d,
in
th
is
r
e
s
e
a
r
c
h,
th
e
r
e
w
e
r
e
two
w
a
ys
f
or
im
a
ge
a
ugm
e
nt
a
ti
on
f
or
th
e
m
ode
l
ut
il
iz
in
g
th
e
m
os
a
ic
a
ppr
oa
c
h
a
nd
us
in
g
th
e
c
onve
nt
io
na
l
a
p
pr
oa
c
h,
a
s
li
s
te
d
in
T
a
bl
e
1.
W
it
h
c
ont
in
ui
ng
tr
a
in
in
g,
th
e
r
e
is
a
s
t
e
a
dy
de
c
r
e
a
s
e
in
c
la
s
s
if
ic
a
ti
on
lo
s
s
on
tr
a
in
in
g
(
tr
a
in
/c
ls
_l
os
s
)
on
e
a
c
h
e
poc
h.
S
ta
r
ti
ng
f
r
om
a
r
ound
4.5,
it
e
nd
s
a
r
ound
0.21,
ba
s
e
d
on
F
ig
ur
e
6.
T
hi
s
c
om
pa
r
e
s
w
it
h
va
l/
c
ls
_l
o
s
s
,
w
he
r
e
it
o
s
c
il
la
te
s
f
r
om
a
r
ound
3.1
to
a
r
ound
0.6,
be
in
g
r
oughly
th
r
e
e
ti
m
e
s
hi
g
he
r
th
a
n
tr
a
in
/c
ls
_l
os
s
,
in
di
c
a
ti
ng
a
s
ig
ni
f
ic
a
nt
r
ol
e
f
or
a
ugm
e
nt
a
ti
on i
n i
s
ol
a
ti
ng a
c
ons
id
e
r
a
bl
e
di
f
f
e
r
e
nc
e
i
n l
os
s
va
lu
e
s
f
or
t
r
a
in
in
g a
nd va
li
da
ti
on.
T
a
bl
e
1.
T
he
pe
r
f
or
m
a
nc
e
of
c
om
pa
r
is
on
r
e
s
ul
ts
a
ugm
e
nt
a
ti
on
te
c
hni
que
s
of
m
ode
l
M
ode
l
P
r
e
c
i
s
i
on (
%
)
R
e
c
a
l
l
(
%
)
m
A
P
@
50 (
%
)
R
obof
l
ow
87.3
75.7
81.7
Y
O
L
O
v11
89.3
82.5
86.2
Y
O
L
O
v12
93.3
83.3
86.3
3.2. Re
s
u
lt
s
of
d
e
ve
lo
p
m
e
n
t
T
he
opt
ic
a
l
c
a
m
e
r
a
c
a
n
pr
oduc
e
a
m
or
e
a
c
c
ur
a
te
a
nd
a
c
c
ur
a
te
m
ode
l
to
c
a
lc
ul
a
te
di
s
e
a
s
e
oc
c
ur
r
e
nc
e
in
hom
-
th
ong
ba
na
na
th
a
n
th
e
one
p
r
oduc
e
d
us
in
g
Y
O
L
O
v1
2
in
opt
ic
a
l
c
lo
s
e
d
-
c
ir
c
ui
t
te
le
vi
s
io
n
(
C
C
T
V
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
n ar
ti
fi
c
ia
l
in
te
ll
ig
e
nc
e
t
e
c
hnol
ogy
f
or
pr
om
ot
in
g hom
-
th
ong b
anana
…
(
R
at
s
am
e
s
T
anv
e
e
nu
k
ool
)
575
L
ig
ht
i
nt
e
ns
it
y a
f
f
e
c
ts
i
m
a
ge
qua
li
ty
ob
ta
in
e
d. T
he
di
f
f
ic
ul
ty
i
n
tr
yi
ng t
o de
te
c
t
obj
e
c
ts
unde
r
va
r
yi
ng
l
ig
ht
in
g
w
a
s
a
pr
obl
e
m
w
it
h t
he
pr
e
vi
ous
ly
bui
lt
m
ode
ls
. A
n opti
c
a
l
c
a
m
e
r
a
ha
vi
ng t
he
a
bi
li
ty
t
o c
a
pt
ur
e
i
m
a
ge
s
unde
r
a
lo
t
of
li
ght
in
g
a
nd
w
e
a
th
e
r
c
ondi
ti
ons
c
a
n
be
ut
il
iz
e
d
to
ga
in
a
w
id
e
r
f
ie
ld
of
v
ie
w
to
im
a
ge
m
o
r
e
obj
e
c
ts
.
T
he
r
e
f
or
e
,
th
e
pe
r
f
or
m
a
nc
e
of
th
e
im
a
gi
ng
de
vi
c
e
de
pe
nds
on
how
th
e
a
lg
or
it
hm
is
tr
a
in
e
d
s
how
n
in
F
ig
ur
e
7.
T
he
r
e
f
or
e
,
th
e
m
e
r
ge
d
im
a
ge
s
f
or
m
th
e
i
m
a
ge
da
ta
s
e
t,
w
hi
c
h
in
th
is
r
e
s
e
a
r
c
h
is
us
e
d
to
e
s
ta
bl
is
h
a
r
e
a
l
-
ti
m
e
im
a
ge
Y
O
L
O
v12
m
ode
l
th
a
t
pr
ovi
de
s
di
f
f
e
r
e
nt
vi
e
w
s
of
th
e
di
s
e
a
s
e
s
ta
tu
s
of
hom
-
th
ong
ba
na
na
.
T
he
m
a
jo
r
r
e
a
s
on w
hy opti
c
a
l
c
a
m
e
r
a
s
a
r
e
us
e
d i
s
i
n t
he
pr
ovi
s
i
on of
ha
vi
ng mor
e
a
c
c
ur
a
te
i
m
a
ge
s
i
n or
de
r
t
o
obt
a
in
be
s
t
qu
a
li
ty
of
hom
-
th
ong
ba
na
na
f
a
r
m
in
g.
T
hi
s
pr
oc
e
s
s
is
c
r
uc
ia
l
f
or
th
e
s
ugge
s
te
d
m
ode
l
to
pr
e
di
c
t
di
s
e
a
s
e
in
c
id
e
nc
e
m
or
e
a
c
c
ur
a
te
ly
in
va
r
io
us
lo
c
a
l
or
s
ub
-
c
li
m
a
ti
c
c
ondi
ti
ons
.
R
ough
phot
ogr
a
phs
of
hom
-
th
ong
ba
na
na
s
ta
ke
n
f
r
om
va
r
io
us
pe
r
s
pe
c
ti
ve
s
a
nd
or
ie
nt
a
ti
ons
.
F
ig
ur
e
8
di
s
pl
a
ys
th
e
m
ode
l
out
put
s
,
w
hi
c
h de
m
ons
tr
a
te
i
m
a
ge
a
nd l
ig
ht
de
te
c
ti
on i
n e
ve
r
y
hom
-
th
ong
ba
na
na
s
pe
c
ie
s
.
F
ig
ur
e
7. Y
O
L
O
v12 plot
s
of
c
la
s
s
if
ic
a
ti
on l
os
s
of
m
ode
l
F
ig
ur
e
8. R
e
s
ul
t
of
Y
O
L
O
v12 de
pl
oy mode
l
3.3. Dis
c
u
s
s
io
n
T
he
f
in
di
ngs
f
r
om
th
e
s
tu
dy
s
ugge
s
t
th
a
t
th
e
r
e
is
g
r
e
a
t
pot
e
nt
i
a
l
f
or
di
s
e
a
s
e
de
te
c
ti
on
m
ode
ls
ba
s
e
d
on
AI
to
im
pr
ove
th
e
f
a
r
m
in
g
of
hom
-
th
ong
ba
na
na
s
in
T
ha
il
a
nd.
T
he
th
r
e
e
r
e
pl
a
c
e
m
e
nt
m
ode
ls
th
a
t
w
e
r
e
c
om
pa
r
e
d
-
R
obof
lo
w
,
Y
O
L
O
v11,
a
nd
Y
O
L
O
v12
-
ha
ve
va
r
yi
ng
pe
r
f
or
m
a
nc
e
s
w
it
h
s
ig
ni
f
ic
a
nt
r
e
a
l
-
w
or
l
d
im
pl
ic
a
ti
ons
f
or
us
e
on f
a
r
m
s
. T
he
be
s
t
pe
r
f
or
m
in
g m
ode
l
w
a
s
Y
O
L
O
v12 with t
he
be
s
t
pr
e
c
is
io
n (
93.3%
)
a
nd
r
e
c
a
ll
(
83.3%
)
va
lu
e
s
,
a
s
w
e
ll
a
s
m
A
P
@
50
(
86.3%
)
va
lu
e
s
.
I
m
pr
ove
d
pe
r
f
o
r
m
a
nc
e
m
a
y
be
ju
s
ti
f
ie
d
on
s
e
ve
r
a
l
gr
ounds
.
T
o
s
ta
r
t
w
it
h,
Y
O
L
O
v12
boa
s
ts
a
hi
ghe
r
a
r
c
hi
te
c
tu
r
e
th
a
t
in
te
gr
a
te
s
th
e
qua
li
ty
of
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
a
nd
s
upe
r
io
r
a
nc
hor
box
m
e
c
ha
ni
s
m
s
ta
s
ke
d
w
it
h
de
a
li
ng
w
it
h
th
e
c
om
pl
e
x
vi
s
ua
l
f
e
a
tu
r
e
s
of
ba
na
na
di
s
e
a
s
e
s
. T
he
pr
e
c
is
io
n
c
a
p
a
bi
li
ty
of
th
e
m
ode
l
w
it
hout
a
f
f
e
c
ti
ng
r
e
c
a
ll
r
a
te
s
is
pa
r
ti
c
ul
a
r
ly
be
ne
f
ic
ia
l
in
a
gr
ic
ul
tu
r
e
pr
oduc
ti
on
e
nvi
r
onm
e
nt
s
w
h
e
r
e
f
a
ls
e
pos
it
iv
e
s
w
il
l
le
a
d
to
a
ddi
ti
ona
l
s
pr
a
y
s
of
pe
s
ti
c
id
e
s
a
nd
pr
oduc
ti
on c
os
ts
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
. 15, No. 1, Febr
ua
r
y 2026
:
568
-
579
576
T
he
te
nde
n
c
y
of
c
onve
r
gi
ng
tr
a
in
in
g
in
F
ig
ur
e
s
4
to
6
pr
ovi
de
s
in
s
ig
ht
f
ul
in
f
or
m
a
ti
on
r
e
ga
r
di
ng
th
e
m
ode
l'
s
be
ha
vi
or
.
S
te
e
p
in
it
ia
l
de
c
r
e
a
s
e
in
lo
s
s
a
nd
c
onv
e
r
ge
n
t
tr
e
nds
of
Y
O
L
O
v12
in
di
c
a
te
good
le
a
r
ni
ng
pr
oc
e
s
s
e
s
,
a
nd
th
a
t
th
e
m
ode
l
is
le
a
r
ni
ng
e
f
f
ic
ie
nt
f
e
a
tu
r
e
s
f
r
om
th
e
a
ugm
e
nt
e
d
da
ta
s
e
t.
W
hi
le
th
e
c
om
pa
r
a
ti
ve
ly
lo
w
tr
a
in
in
g
a
nd
va
li
da
ti
on
lo
s
s
e
s
of
a
ll
th
e
m
ode
ls
r
e
la
ti
ve
to
e
a
c
h
ot
he
r
r
e
pr
e
s
e
nt
e
d
good
ge
ne
r
a
li
z
a
ti
on
pe
r
f
or
m
a
nc
e
,
Y
O
L
O
v12
e
xhi
bi
te
d
th
e
m
os
t
i
m
pl
ic
it
pe
r
f
or
m
a
nc
e
.
T
he
da
ta
a
ugm
e
nt
a
ti
on
m
e
th
od
us
e
d
in
th
is
e
xpe
r
im
e
nt
,
w
hi
c
h
a
ll
ow
e
d
th
e
or
ig
in
a
l
d
a
ta
s
e
t
to
e
xpa
nd
f
r
om
2,576
to
6,184
im
a
ge
s
,
e
f
f
e
c
ti
ve
ly
s
tr
e
ngt
he
ne
d
m
ode
l
r
obus
tn
e
s
s
. T
he
di
f
f
e
r
e
nt
a
ugm
e
nt
a
ti
ons
:
hor
iz
ont
a
l
a
nd
ve
r
ti
c
a
l
f
li
p,
r
ot
a
ti
on,
br
ig
ht
ne
s
s
(
±15%
)
,
s
a
tu
r
a
ti
on
(
±25%
)
,
a
nd
r
a
ndom
c
r
oppi
ng
(
20%
)
,
be
tt
e
r
pr
e
pa
r
e
d
th
e
m
od
e
ls
to
de
a
l
w
it
h
di
f
f
e
r
e
nt
e
nvi
r
onm
e
nt
a
l
c
ondi
t
io
ns
in
th
e
f
ie
ld
.
T
hi
s
is
c
r
it
ic
a
l
w
he
n
im
pl
e
m
e
nt
in
g
th
e
s
ys
te
m
s
in
to
r
e
a
l
-
w
or
ld
s
e
tt
in
gs
, w
he
r
e
i
ll
um
in
a
ti
on, vie
w
in
g c
ondi
ti
ons
, a
n
d a
m
bi
e
nt
c
ondi
ti
ons
va
r
y s
ig
ni
f
ic
a
nt
ly
.
I
t
is
im
por
ta
nt
to
not
e
di
s
pa
r
it
y
c
la
s
s
if
ic
a
ti
on
lo
s
s
be
twe
e
n
th
e
tr
a
in
in
g
a
nd
va
li
da
ti
on
s
e
t,
is
m
os
t
a
c
ut
e
ly
s
e
e
n
in
Y
O
L
O
v12
(
tr
a
in
/c
ls
_l
os
s
:
0.21
vs
.
va
l/
c
ls
_l
os
s
:
0.6)
.
A
n
a
c
tu
a
l
th
r
e
e
f
ol
d
di
f
f
e
r
e
nc
e
in
di
s
pa
r
it
y
in
di
c
a
te
s
,
w
hi
le
a
ugm
e
nt
a
ti
on
te
c
hni
que
s
ty
pi
c
a
ll
y
im
pr
ove
m
ode
l
ge
ne
r
a
li
z
a
bi
li
ty
,
th
e
y
a
ls
o
m
e
a
n
th
a
t
di
m
e
ns
io
na
li
ty
a
nd
c
om
pl
e
xi
ty
ha
s
in
c
r
e
a
s
e
d,
m
e
a
ni
ng
it
m
us
t
be
ta
m
e
d
th
r
ough
hype
r
pa
r
a
m
e
te
r
tu
ni
ng
.
F
ut
ur
e
r
e
s
e
a
r
c
h
w
il
l
tr
y
to
m
ove
be
yond
a
ugm
e
nt
a
ti
on
te
c
h
ni
que
s
to
m
in
im
iz
e
s
uc
h
a
di
s
pa
r
it
y
w
it
hout
c
om
pr
om
is
in
g
de
te
c
ti
on
pe
r
f
or
m
a
nc
e
.
T
he
a
bi
li
ty
of
th
e
m
ode
l
to
ope
r
a
te
unde
r
va
r
yi
ng
li
ght
in
g
c
ondi
ti
ons
a
ddr
e
s
s
e
s
one
of
th
e
bi
gge
s
t
hur
dl
e
s
to
A
I
u
s
e
a
t
th
e
f
a
r
m
l
e
ve
l.
T
h
e
pr
a
c
ti
c
a
l
r
e
a
l
-
w
or
ld
e
f
f
e
c
t
of
s
uc
h
f
in
di
ngs
is
e
nor
m
ous
.
A
c
c
or
di
ng
to
th
e
a
c
c
ur
a
c
y
of
Y
O
L
O
v12,
a
gr
ic
ul
tu
r
a
l
f
a
r
m
e
r
s
c
a
n
r
e
ly
on
th
e
s
y
s
te
m
to
id
e
nt
if
y
di
s
e
a
s
e
a
c
c
ur
a
te
ly
w
it
hout
undue
w
or
r
y
of
f
a
ls
e
a
la
r
m
s
.
A
lt
hough
th
e
83.3%
r
e
c
a
ll
r
a
te
is
w
or
s
e
c
om
pa
r
e
d
to
pr
e
c
is
io
n,
it
is
s
ti
ll
s
ig
ni
f
ic
a
nt
ly
be
tt
e
r
th
a
n
th
a
t
p
os
s
ib
le
th
r
ough
hum
a
n
in
s
pe
c
ti
on
a
nd
c
a
r
r
ie
s
e
a
r
ly
w
a
r
ni
ng c
a
pa
c
it
y w
it
h t
he
a
bi
li
ty
t
o a
ve
r
t
huge
l
os
s
e
s
i
n c
r
ops
.
T
he
r
e
a
r
e
c
e
r
ta
in
c
ons
tr
a
in
ts
th
a
t
ne
e
d
to
be
m
e
nt
io
ne
d,
how
e
ve
r
.
T
he
s
tu
dy
s
pe
c
if
ic
a
ll
y
ta
c
kl
e
d
th
e
pe
r
f
or
m
a
nc
e
of
s
e
ve
n
s
pe
c
if
ic
di
s
e
a
s
e
ty
pe
s
of
th
e
hom
-
th
ong
ba
na
na
s
,
a
nd
it
is
ne
c
e
s
s
a
r
y
to
go
ba
c
k
a
nd
c
he
c
k
w
it
h
th
e
m
ode
l
on
th
e
pe
r
f
or
m
a
nc
e
r
e
ga
r
di
ng
ot
he
r
ba
na
na
va
r
ie
ti
e
s
or
ot
he
r
ty
pe
s
of
di
s
e
a
s
e
s
.
A
ddi
ti
ona
ll
y, Y
O
L
O
v12'
s
c
om
put
a
ti
ona
l
lo
a
d, a
lt
hough r
e
a
s
ona
bl
e
on t
oda
y'
s
ha
r
dw
a
r
e
, m
a
y be
c
um
be
r
s
om
e
to
de
pl
oy
on
h
a
r
dw
a
r
e
-
c
ons
tr
a
in
e
d
r
ur
a
l
e
nvi
r
onm
e
nt
s
.
I
m
pl
e
m
e
nt
a
ti
on
of
th
is
t
e
c
hnol
ogy
in
e
xi
s
ti
ng
a
gr
ic
ul
tu
r
a
l
pr
a
c
ti
c
e
m
us
t
c
ons
id
e
r
f
a
r
m
e
r
tr
a
in
in
g,
in
f
r
a
s
tr
uc
tu
r
e
r
e
qui
r
e
m
e
nt
s
,
a
nd
c
os
t
a
s
s
e
s
s
m
e
nt
.
I
t
m
us
t
be
r
e
s
e
a
r
c
he
d
in
th
e
f
ut
ur
e
w
it
h
a
f
oc
us
on
e
xpl
or
in
g
th
e
pos
s
ib
il
it
y
of
c
r
e
a
ti
ng
li
ght
w
e
ig
ht
m
ode
l
im
pl
e
m
e
nt
a
ti
ons
f
or
m
obi
le
a
nd e
dge
c
om
put
in
g pl
a
tf
or
m
s
t
o i
m
pr
ove
a
c
c
e
s
s
.
4.
C
O
N
C
L
U
S
I
O
N
T
hi
s
r
e
s
e
a
r
c
h e
f
f
e
c
ti
ve
ly
de
ve
lo
pe
d
a
nd
v
a
li
da
te
d
a
uni
que
A
I
-
ba
s
e
d s
ys
te
m
f
or
e
na
bl
in
g
hom
-
th
ong
ba
na
na
c
ul
ti
va
ti
on
on
th
e
ba
s
is
of
a
ut
om
a
ti
c
di
s
e
a
s
e
id
e
nt
if
ic
a
ti
on.
T
he
e
xt
e
ns
iv
e
c
om
pa
r
a
ti
ve
a
na
ly
s
is
of
th
r
e
e
di
f
f
e
r
e
nt
de
e
p
le
a
r
ni
ng
m
od
e
ls
R
obof
lo
w
,
Y
O
L
O
v11,
a
nd
Y
O
L
O
v12
pr
ovi
de
d
va
lu
a
bl
e
in
f
or
m
a
ti
on
a
bout
th
e
a
ppl
ic
a
ti
on
of
c
om
put
e
r
vi
s
io
n
te
c
hnol
ogy
in
th
e
a
gr
ic
ul
tu
r
e
s
e
c
to
r
in
r
e
a
l
li
f
e
.
T
he
ke
y
r
e
s
ul
ts
in
di
c
a
te
th
a
t
Y
O
L
O
v12
pe
r
f
or
m
s
s
ig
ni
f
ic
a
nt
ly
be
tt
e
r
th
a
n
o
th
e
r
a
ppr
oa
c
he
s
,
a
c
hi
e
vi
ng
93.3%
a
c
c
ur
a
c
y,
83.3%
r
e
c
a
ll
,
a
nd
86.3%
m
A
P
@
50
in
de
te
c
ti
ng
s
e
ve
n
ba
na
na
di
s
e
a
s
e
s
.
T
he
s
e
pe
r
f
or
m
a
nc
e
r
e
s
ul
ts
a
r
e
m
uc
h
be
tt
e
r
th
a
n
m
a
nua
l
in
s
pe
c
ti
on
m
e
th
ods
,
pr
ovi
di
ng
f
a
r
m
e
r
s
w
it
h
a
f
a
s
t,
r
e
li
a
bl
e
,
a
nd
obj
e
c
ti
ve
m
e
th
od
of
di
s
e
a
s
e
d
e
te
c
ti
on.
M
or
e
ove
r
,
th
e
m
ode
l'
s
c
a
p
a
bi
li
ty
to
ge
ne
r
a
li
z
e
unde
r
c
ha
ngi
ng
e
nvi
r
onm
e
nt
a
l
f
a
c
to
r
s
a
nd
th
e
us
e
of
r
obus
t
da
ta
a
ugm
e
nt
a
ti
on
m
e
th
ods
s
uppor
t
it
s
pot
e
nt
ia
l
a
ppl
ic
a
bi
li
ty
in
pr
a
c
ti
c
a
l
f
ie
ld
a
ppl
ic
a
ti
ons
in
T
ha
i
ba
n
a
na
f
a
r
m
s
.
T
he
r
e
s
e
a
r
c
h
c
ont
r
ib
ut
e
s
to
s
om
e
of
th
e
ke
y
c
ha
ll
e
ng
e
s
th
a
t
pr
e
s
e
nt
ly
hi
nde
r
th
e
hom
-
th
ong ba
na
na
i
ndus
tr
y, i
nc
lu
di
ng e
r
r
a
ti
c
yi
e
ld
, s
us
c
e
pt
ib
il
it
y t
o di
s
e
a
s
e
s
a
nd pe
s
ts
, a
nd i
ne
f
f
ic
ie
nt
s
uppl
y
c
ha
in
s
.
B
y
f
a
c
il
it
a
ti
ng
e
a
r
ly
de
te
c
ti
on
of
di
s
e
a
s
e
s
,
th
e
s
y
s
te
m
c
a
n
he
lp
a
f
a
r
m
e
r
to
ti
m
e
ly
in
te
r
ve
nt
io
n,
r
e
duc
e
c
r
op l
os
s
, a
nd l
e
s
s
e
n pe
s
ti
c
id
e
us
e
, w
hi
c
h c
a
n a
ls
o c
ont
r
ib
ut
e
t
o a
m
or
e
s
us
ta
in
a
bl
e
a
nd pr
of
it
a
bl
e
a
gr
ic
ul
tu
r
a
l
pr
a
c
ti
c
e
.
T
hi
s
A
I
s
y
s
te
m
c
a
n
be
im
pl
e
m
e
nt
e
d
s
u
c
c
e
s
s
f
ul
ly
a
nd
in
di
c
a
te
s
a
s
hi
f
t
to
w
a
r
ds
pr
e
c
is
io
n
a
gr
ic
ul
tu
r
e
in
T
ha
il
a
nd'
s
ba
na
n
a
in
dus
tr
y.
T
he
t
e
c
hnol
ogy'
s
pot
e
nt
ia
l
a
ls
o
li
e
s
b
e
yond
di
s
e
a
s
e
de
te
c
ti
on,
f
or
e
xa
m
pl
e
,
w
it
h
w
id
e
r
a
ppl
ic
a
ti
ons
in
c
r
op
m
oni
to
r
in
g,
yi
e
ld
pr
e
di
c
ti
on
a
nd
qua
li
ty
in
s
pe
c
ti
on,
a
s
a
c
om
pl
e
m
e
nt
to
T
ha
il
a
nd'
s
n
a
ti
ona
l
pol
ic
ie
s
f
or
s
m
a
r
t
a
gr
ic
ul
tu
r
e
.
F
ut
ur
e
r
e
s
e
a
r
c
h
s
houl
d
in
ve
s
ti
ga
te
how
to
w
id
e
n
th
e
s
ys
te
m
'
s
s
c
ope
to
in
c
lu
de
ot
he
r
c
a
t
e
gor
ie
s
of
di
s
e
a
s
e
s
a
nd
ty
pe
s
of
ba
na
n
a
s
,
d
e
ve
lo
p
m
obi
le
-
f
r
ie
ndl
y
a
pps
s
o
f
a
r
m
e
r
s
c
a
n
a
c
c
e
s
s
th
e
m
e
a
s
il
y,
a
nd
c
om
m
e
nc
e
lo
ngi
tu
di
na
l
f
ie
ld
s
tu
di
e
s
to
a
s
s
e
s
s
lo
ng
-
te
r
m
e
f
f
e
c
ti
ve
ne
s
s
a
nd e
c
onomi
c
pa
yba
c
k.
A
C
K
N
O
WL
E
D
G
M
E
N
T
S
W
e
w
oul
d
li
ke
to
th
a
nk
th
e
R
e
s
e
a
r
c
h
a
nd
D
e
v
e
lo
pm
e
nt
I
ns
ti
t
ut
e
in
th
e
R
a
ja
m
a
ng
a
la
U
ni
ve
r
s
it
y
of
T
e
c
hnol
ogy
S
uva
r
na
bhumi
f
or
th
e
he
lp
a
nd
a
s
s
i
s
ta
nc
e
th
a
t
th
e
te
a
m
r
e
c
e
iv
e
d
in
th
e
f
or
m
a
ti
on
of
th
is
r
e
s
e
a
r
c
h. T
h
e
ir
c
ont
r
ib
ut
io
n t
o t
hi
s
r
e
s
e
a
r
c
h ha
s
s
ig
ni
f
ic
a
nt
va
lu
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
n ar
ti
fi
c
ia
l
in
te
ll
ig
e
nc
e
t
e
c
hnol
ogy
f
or
pr
om
ot
in
g hom
-
th
ong b
anana
…
(
R
at
s
am
e
s
T
anv
e
e
nu
k
ool
)
577
F
U
N
D
I
N
G
I
N
F
O
R
M
A
T
I
O
N
T
he
pr
oj
e
c
t
obt
a
in
e
d
f
in
a
nc
ia
l
a
s
s
is
ta
nc
e
f
r
om
th
e
R
a
ja
m
a
nga
la
U
ni
ve
r
s
it
y
of
T
e
c
hnol
ogy
S
uva
r
na
bhumi
th
r
ough
gr
a
nt
F
R
B
680051/0173
-
5,
w
hi
c
h
he
lp
e
d
in
f
undi
ng
th
e
va
r
io
us
a
c
ti
vi
ti
e
s
in
th
e
s
tu
dy,
s
uc
h a
s
a
na
ly
s
is
a
nd da
ta
ga
th
e
r
in
g.
A
U
T
H
O
R
C
O
N
T
R
I
B
U
T
I
O
N
S
S
T
A
T
E
M
E
N
T
T
hi
s
jo
ur
na
l
us
e
s
th
e
C
ont
r
ib
ut
or
R
ol
e
s
T
a
xonomy
(
C
R
e
di
T
)
to
r
e
c
ogni
z
e
in
di
vi
dua
l
a
ut
hor
c
ont
r
ib
ut
io
ns
, r
e
duc
e
a
ut
hor
s
hi
p di
s
put
e
s
,
a
nd f
a
c
il
it
a
te
c
ol
la
bo
r
a
ti
on.
N
am
e
o
f
A
u
t
h
or
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
R
a
ts
a
m
e
s
T
a
n
ve
e
n
uk
ool
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
S
u
w
i
t
S
o
m
s
u
p
h
a
p
r
u
n
g
y
o
s
✓
✓
✓
✓
✓
✓
✓
✓
B
oonya
r
it
N
okkur
th
✓
✓
✓
✓
✓
✓
✓
L
ik
it
C
ha
m
ut
ha
i
✓
✓
✓
✓
✓
✓
✓
P
a
t
u
m
w
a
d
e
e
B
o
n
g
u
l
e
a
u
m
✓
✓
✓
✓
✓
✓
P
a
r
in
ya
N
a
th
o
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
C
:
C
onc
e
pt
ua
l
i
z
a
t
i
on
M
:
M
e
t
hodol
ogy
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
da
t
i
on
Fo
:
Fo
r
m
a
l
a
na
l
ys
i
s
I
:
I
nve
s
t
i
ga
t
i
on
R
:
R
e
s
our
c
e
s
D
:
D
a
t
a
C
ur
a
t
i
on
O
:
W
r
i
t
i
ng
-
O
r
i
gi
na
l
D
r
a
f
t
E
:
W
r
i
t
i
ng
-
R
e
vi
e
w
&
E
di
t
i
ng
Vi
:
Vi
s
ua
l
i
z
a
t
i
on
Su
:
Su
pe
r
vi
s
i
on
P
:
P
r
oj
e
c
t
a
dm
i
ni
s
t
r
a
t
i
on
Fu
:
Fu
ndi
ng a
c
qui
s
i
t
i
on
C
O
N
F
L
I
C
T
O
F
I
N
T
E
R
E
S
T
S
T
A
T
E
M
E
N
T
T
he
a
ut
hor
s
de
c
la
r
e
th
a
t
th
e
y
h
a
ve
no
kno
w
n
c
om
pe
ti
ng
f
in
a
nc
ia
l
in
te
r
e
s
ts
or
p
e
r
s
ona
l
r
e
la
ti
ons
hi
p
s
th
a
t
c
oul
d ha
ve
a
ppe
a
r
e
d t
o i
nf
lu
e
nc
e
t
h
e
w
or
k r
e
por
te
d i
n t
hi
s
pa
pe
r
.
I
N
F
O
R
M
E
D
C
O
N
S
E
N
T
W
e
ha
ve
obt
a
in
e
d i
nf
or
m
e
d c
ons
e
nt
f
r
om
a
ll
i
ndi
vi
dua
ls
i
nc
lu
de
d i
n t
hi
s
s
tu
dy.
E
T
H
I
C
A
L
A
P
P
R
O
V
A
L
T
he
r
e
s
e
a
r
c
h
r
e
la
te
d
to
hum
a
n
u
s
e
h
a
s
be
e
n
c
om
pl
ie
d
w
it
h
a
ll
th
e
r
e
le
va
nt
na
ti
ona
l
r
e
gul
a
ti
ons
a
nd
in
s
ti
tu
ti
ona
l
pol
ic
ie
s
in
a
c
c
or
da
nc
e
w
it
h
th
e
te
ne
ts
of
th
e
H
e
ls
i
nki
D
e
c
la
r
a
ti
on
a
nd
ha
s
be
e
n
a
ppr
ove
d
by
th
e
a
ut
hor
s
'
i
ns
ti
tu
ti
ona
l
r
e
vi
e
w
boa
r
d or
e
qui
va
le
nt
c
om
m
it
te
e
.
D
A
T
A
A
V
A
I
L
A
B
I
L
I
T
Y
D
a
ta
a
v
a
il
a
bi
li
ty
i
s
no
t
a
p
pl
ic
a
bl
e
t
o t
hi
s
pa
p
e
r
a
s
no
ne
w
da
t
a
w
e
r
e
c
r
e
a
t
e
d or
a
na
ly
z
e
d i
n
t
hi
s
s
t
udy.
R
E
F
E
R
E
N
C
E
S
[
1]
S
.
V
i
s
e
t
noi
a
nd
S
.
S
i
r
i
s
opons
i
l
p,
“
U
pl
i
f
t
i
ng
T
ha
i
l
a
nd’
s
a
gr
i
c
ul
t
ur
e
t
hr
ough
a
gr
i
c
ul
t
ur
a
l
e
duc
a
t
i
on:
a
pa
r
a
di
gm
s
hi
f
t
f
or
f
ut
u
r
e
f
a
r
m
e
r
s
,”
I
nt
e
r
nat
i
onal
J
our
nal
of
A
gr
i
c
ul
t
ur
e
I
nnov
at
i
on,
T
e
c
hnol
ogy
and
G
l
obal
i
s
at
i
on
,
vol
.
1,
no.
1,
pp.
44
–
56,
2019,
doi
:
10.1504/
i
j
a
i
t
g.2019.099600.
[
2]
Y
.
L
uo
a
nd
Z
.
Y
i
n,
“
M
a
r
ke
r
-
a
s
s
i
s
t
e
d
b
r
e
e
di
ng
o
f
T
h
a
i
f
r
a
g
r
a
nc
e
r
i
c
e
f
or
s
e
m
i
-
d
w
a
r
f
p
he
n
ot
ype
,
s
ub
m
e
r
ge
nc
e
t
o
l
e
r
a
nc
e
a
n
d
di
s
e
a
s
e
r
e
s
i
s
t
a
n
c
e
t
o
r
i
c
e
b
l
a
s
t
a
n
d b
a
c
t
e
r
i
a
l
b
l
i
gh
t
,”
M
ol
e
c
u
l
ar
B
r
e
e
di
ng
, v
ol
. 3
2,
no
.
3,
p
p.
709
–
7
21,
2
013
,
do
i
:
10
.10
07
/
s
11
03
2
-
01
3
-
99
04
-
2.
[
3]
M
.
S
a
k
t
h
i
g
a
ne
s
h
a
n
d S
.
D
i
n
e
s
h
ku
m
a
r
, “
A
s
t
u
d
y o
n c
o
ns
t
r
a
i
n
t
s
f
a
c
e
d
b
y
t
h
e
b
a
n
a
na
g
r
o
w
e
r
s
i
n t
h
e
p
r
o
du
c
t
i
on
a
n
d m
a
r
ke
t
i
n
g o
f
b
a
na
n
a
,”
E
P
R
A
I
n
t
e
r
n
a
t
i
o
na
l
J
ou
r
n
a
l
o
f
A
g
r
i
c
u
l
t
ur
e
a
n
d
R
ur
a
l
E
c
o
n
o
m
i
c
R
e
s
e
ar
c
h
,
vo
l
.
1
0,
no
.
1
1
,
p
p
.
15
–
1
7
,
2
0
2
2,
d
o
i
:
1
0
.
36
7
1
3
/
e
p
r
a
11
6
5
1
.
[
4]
N
.
F
.
K
a
c
ho,
M
.
H
us
s
a
i
n,
a
nd
S
.
B
a
nda
y,
“
A
r
t
i
f
i
c
i
a
l
i
nt
e
l
l
i
ge
nc
e
i
n
a
gr
i
c
ul
t
u
r
e
,”
i
n
F
ut
ur
i
s
t
i
c
T
r
e
nds
i
n
A
gr
i
c
ul
t
ur
e
E
ngi
ne
e
r
i
ng
&
F
ood Sc
i
e
nc
e
s
, C
hi
kka
m
a
ga
l
ur
u, I
ndi
a
:
I
t
e
r
a
t
i
ve
I
nt
e
r
na
t
i
ona
l
P
ubl
i
s
he
r
, 202
4, pp. 652
–
658. doi
:
10.58532/
v3bc
a
gp1c
h51.
[
5]
P
.
N
a
t
ho,
S
.
B
oonyi
ng,
P
.
B
ongul
e
a
um
,
N
.
T
a
nt
i
dont
a
ne
t
,
a
nd
L
.
C
h
a
m
ut
ha
i
,
“
A
n
e
nha
nc
e
d
m
a
c
hi
ne
vi
s
i
on
s
ys
t
e
m
f
or
s
m
a
r
t
poul
t
r
y f
a
r
m
s
us
i
ng de
e
p l
e
a
r
ni
ng,”
Sm
ar
t
A
gr
i
c
ul
t
ur
al
T
e
c
hnol
ogy
, vol
. 12, 20
25, doi
:
10.1016/
j
.a
t
e
c
h.2025.101083.
[
6]
R
.
S
i
ngh
a
nd
K
.
K
.
S
i
ngh,
“
E
nha
nc
i
ng
a
gr
i
c
ul
t
u
r
a
l
e
f
f
i
c
i
e
nc
y
t
hr
ough
s
m
a
r
t
f
a
r
m
i
ng
a
nd
i
nt
e
r
ne
t
of
t
h
i
ngs
e
na
bl
e
d
pr
e
c
i
s
i
on
a
gr
i
c
ul
t
ur
e
,”
A
gr
i
c
ul
t
ur
al
Sc
i
e
nc
e
D
i
ge
s
t
-
A
R
e
s
e
ar
c
h J
our
nal
, 2024, doi
:
10.1
8805/
a
g.d
-
6039.
[
7]
M
.
U
.
M
a
noj
,
V
.
P
r
a
de
e
p,
B
.
V
.
C
hi
nda
n,
N
.
G
ow
r
i
s
h,
a
nd
H
.
G
.
P
.
G
o
w
da
,
“
A
I
t
e
c
hni
que
s
f
or
pl
a
nt
di
s
e
a
s
e
de
t
e
c
t
i
on,
”
I
nt
e
r
nat
i
onal
J
our
nal
of
A
dv
anc
e
d
R
e
s
e
ar
c
h
i
n
Sc
i
e
nc
e
,
C
om
m
uni
c
at
i
on
an
d
T
e
c
hnol
ogy
,
vol
.
4,
no.
3,
pp.
200
–
207,
2024,
doi
:
10.48175/
i
j
a
r
s
c
t
-
22832.
[
8]
P
. A
. B
a
i
na
l
w
a
r
, S
. M
. B
or
ka
r
, S
. S
. S
ha
m
bha
r
ka
r
, a
nd P
. S
. M
oon, “
I
nt
e
l
l
i
ge
nt
s
ys
t
e
m
t
o a
na
l
y
s
i
s
of
pl
a
nt
di
s
e
a
s
e
s
us
i
ng m
a
c
hi
n
e
l
e
a
r
ni
ng
t
e
c
hni
que
s
,”
i
n
P
r
oc
e
e
di
ngs
of
t
he
5t
h
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on
I
nf
or
m
at
i
on
M
anage
m
e
nt
&
M
ac
hi
ne
I
nt
e
l
l
i
ge
nc
e
,
J
a
i
pur
, I
ndi
a
:
A
s
s
oc
i
a
t
i
on f
or
C
om
put
i
ng M
a
c
hi
ne
r
y
,
2023, pp. 1
–
8. doi
:
10.1145/
3647444.3652478.
[
9]
S
.
K
.
K
a
nna
,
K
.
R
a
m
a
l
i
nga
m
,
P
.
P
a
z
ha
ni
v
e
l
a
n,
R
.
J
a
ga
de
e
s
w
a
r
a
n,
a
nd
P
.
P
.C
.,
“
Y
O
L
O
de
e
p
l
e
a
r
ni
ng
a
l
gor
i
t
hm
f
or
obj
e
c
t
de
t
e
c
t
i
on i
n a
gr
i
c
ul
t
ur
e
:
a
r
e
vi
e
w
,”
J
our
nal
of
A
gr
i
c
ul
t
ur
al
E
ngi
ne
e
r
i
ng
, vol
. 55
, no. 4, 2024, doi
:
10.4081/
j
a
e
.2024.1641.
Evaluation Warning : The document was created with Spire.PDF for Python.