Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
23
,
No.
1
,
Ju
ly
2021
, p
p.
378
~
386
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v
23
.i
1
.
pp
378
-
386
378
Journ
al h
om
e
page
:
http:
//
ij
eecs.i
aesc
or
e.c
om
Haruma
nis mango le
af
d
isease r
ecogniti
on
sy
stem
u
sing i
mage
processi
ng techn
iqu
e
R.
A.
JM.
Gin
ing
1
, S.
S.
M.
Fau
z
i
2
,
N.
M
.
Yu
s
off
3
, T.
R.
Ra
z
ak
4
,
M.
H.
Ismail
5
, N.
A
.
Z
ak
i
6
,
F.
Ab
dull
ah
7
1,2,3,4,5
Facul
t
y
of C
om
pute
r
and
Mathe
m
at
i
ca
l
Sc
ie
nc
es,
Univ
ersiti
T
eknol
ogi
MA
RA,
Perli
s,
Mal
a
y
sia
6
Facul
t
y
of
Arch
it
e
ct
ure
,
Pl
anni
n
g
and
Surve
y
ing, Unive
rsit
i
T
ekn
ologi
MA
RA,
Pe
rli
s,
Ma
lay
si
a
7
Pus
at
Pen
y
elidi
kan,
Hort
ikul
tu
r M
ard
i
Sintok
,
K
eda
h
,
Mal
a
y
s
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Ma
r 23
, 202
1
Re
vised A
pr
1
9
, 2
021
Accepte
d
Ma
y
1,
2021
Curre
nt
Harum
ani
s
m
ango
far
m
ing
te
chn
ique
in
Ma
lay
sia
stil
l
m
ost
l
y
depe
nds on
the
f
armers'
own e
xp
ert
ise
to
m
onit
or
the
cro
ps from the
at
t
ac
k
of
pests
and
insec
t
s.
Thi
s
appr
oach
is
sus
ce
pti
ble
to
hum
an
err
ors,
and
thos
e
who
do
not
po
sse
ss
thi
s
skill
m
a
y
not
be
able
to
det
e
ct
the
dise
ase
at
the
righ
t
ti
m
e.
As
le
af
dis
ea
ses
seriousl
y
a
ffe
ct
th
e
cro
p
'
s
g
rowth
and
the
qu
al
ity
of
the
y
i
el
d
,
th
is
study
ai
m
s
to
deve
lop
a
rec
ogniti
on
sy
st
em
tha
t
det
ec
ts
th
e
pre
senc
e
of
dis
ea
se
in
the
m
a
ngo
le
af
using
image
proc
essin
g
te
chni
qu
e.
First,
the
image
is
ac
quire
d
thro
ugh
a
sm
art
phone
ca
m
era;
once
it
has
bee
n
pre
-
proc
essed
,
i
t
is
the
n
segm
ented
in
whi
ch
t
h
e
RGB
image
is
c
onver
te
d
to
an
HS
I
image
,
t
hen
th
e
fe
at
ur
es
are
extrac
te
d
.
La
stl
y
,
th
e
cl
ass
ifi
c
at
ion
of
disea
se
is
done
to
det
e
rm
ine
the
t
y
p
e
of
leaf
dis
ea
se.
The
propo
sed
s
y
stem
eff
ectivel
y
de
tec
ts
and
class
if
y
t
he
disea
se
with
an
accurac
y
of
6
8.
89%.
Th
e
findi
ngs
of
thi
s
proje
c
t
wil
l
contribute
to
f
armer
s
and
soci
ety
's
bene
fi
t,
and
rese
arc
h
ers
c
an us
e
the a
ppro
ac
h
to addre
ss
sim
ilar
issues i
n
futur
e
works
.
Ke
yw
or
d
s
:
Harum
anis
Im
age p
r
ocessi
ng
Leaf
disease
Ma
tl
ab
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
B
Y
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
S.
S
. M. Fauzi
Faculty
of Com
pu
te
r
an
d
Ma
them
a
ti
cal
Scie
n
ces
Un
i
ver
sit
i Te
knol
og
i M
ARA
Perlis B
ra
nch
, 026
00,
Ar
a
u, P
erli
s,
Ma
la
ysi
a
Em
a
il
:
sh
ukor
s
anim
@u
itm
.ed
u.
m
y
1.
INTROD
U
CTION
On
e
of
Ma
la
ysi
a'
s
fa
m
ou
s
m
ang
o
va
riet
ie
s
is
Har
um
anis
m
ang
o,
wh
ic
h
ha
s
a
ph
e
nom
enal
com
m
ercial
dem
and
an
d
thri
ve
s
well
,
par
ti
cu
la
rly
in
Perlis,
Ma
la
ysi
a
'
s
no
rt
hern
sta
te
.
Yiel
ding
good
a
nd
ripe
Harum
anis
is
no
t
a
n
ea
sy
ta
sk
;
th
us
,
fa
rm
e
rs
al
ways
m
ust
ta
ke
ext
ra
ca
re
of
t
hem
,
m
a
inly
to
pre
ven
t
them
from
the
diseases
that
m
ay
i
nf
ect
a
nd
r
uin
the
cr
op.
U
su
a
ll
y,
far
m
ers
will
m
ake
regular
m
on
it
or
in
g
t
hro
ug
h
their
nak
e
d
e
ye
s
to
detect
po
te
ntial
disea
ses.
F
or
insta
nce,
t
he
f
ede
r
al
agr
ic
ultur
e
m
ark
et
ing
aut
hority
(
FA
M
A) in M
al
ay
sia
u
ses
hum
an
ex
per
ts
f
or the
fru
it
s
gr
a
ding
process
usi
ng p
e
rcep
ti
on and
hand
m
et
ho
d.
Howe
ver,
this
appr
oach
seem
s
le
ss
ef
fici
ent
as
it
co
nsum
e
s
m
or
e
tim
e
a
nd
fa
rm
ers
m
a
y
ov
e
rlo
ok
so
m
e
le
aves
th
at
are
al
rea
dy
infected
an
d
m
issed
the
ri
gh
t
t
i
m
e
to
pr
e
vent
and
treat
the
m
[1
]
.
Be
sides
,
this
appr
oach
is
al
so
not
app
li
cabl
e
to
ever
y
farm
er
as
they
hu
gely
rely
on
hum
an
exp
erts.
Additi
on
al
ly
,
hum
an
exp
e
rts a
re
pro
ne
to
hum
an
er
ror; h
e
nce t
his
m
ade th
e m
et
ho
d l
ess
reli
able.
Ther
e
f
or
e,
it
is
necessar
y
f
or
red
e
fining
the
op
e
rati
on
of
th
e
far
m
ing
indu
stry,
an
d
the
ke
y
to
this
is
sm
art
far
m
ing
.
This
ap
proac
h'
s
research
ef
f
ort
has
been
dra
m
at
ic
ally
exp
a
nd
e
d
sig
nifica
ntly
with
t
he
rise
of
var
i
ou
s
m
et
ho
ds
,
su
c
h
as
m
a
chine
le
a
rn
i
ng
and
im
age
pro
cessi
ng.
Howe
ver,
m
os
t
of
t
he
re
ported
w
orks
on
sm
art
far
m
ing
wer
e
c
onduct
e
d
on
rice
an
d
tom
at
o
cro
ps,
wh
e
reas
ver
y
l
it
tl
e
on
the
Ha
r
um
anis
m
ang
oe
s.
The
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Ha
r
uma
nis
m
ango le
af d
ise
ase
reco
gnit
ion
s
yst
em usin
g
im
ag
e
proces
sin
g t
echn
i
qu
e
(
R.
A. JM.
G
i
ning
)
379
pr
e
vious
s
olu
t
ion
s
c
onverge
on
the
f
r
uits
rather
tha
n
it
s
le
aves;
wh
i
ch
al
so
play
a
vital
ro
le
in
cr
op
pro
du
ct
ivit
y.
T
hu
s
,
this
stu
dy
inten
ds
t
o
de
velo
p
a
rec
ogni
ti
on
syst
em
th
at
can
qu
ic
kly
help
fa
rm
ers
detect
m
ang
o l
eaf
dis
eases easi
ly
w
i
thout any
hum
an
inte
rf
e
ren
ce
.
Find
i
ng
from
t
his
stu
dy
will
con
t
rib
ute
to
s
ociet
y'
s
ben
efit
as
Ma
la
ysi
a
'
s
econom
ic
back
bones
is
the
a
gr
ic
ultur
e
sec
tor.
T
he
pro
gressi
vely
com
p
et
it
ive
and
vola
ti
le
m
ark
et
ing
and
chall
e
ng
e
s
su
ch
as
insec
ts
a
nd
pests
r
uin
i
ng
t
he
cr
ops,
e
xh
a
us
ti
ve
fa
rm
ing
m
et
ho
ds
a
nd
far
m
ers
at
tempti
ng
t
o
pr
oduc
e
foo
d
in
the
m
os
t
su
sta
ina
ble
m
a
nn
e
r
justi
fy
the
need
for
m
or
e
e
ff
ic
ie
nt
fa
rm
i
ng
m
et
ho
ds.
T
hu
s
,
by
im
ple
m
enting
the
a
ppr
oac
h
der
i
ved
f
r
om
t
his
pro
j
ect
'
s
resu
lt
s,
far
m
ers
can
al
so
ta
ke
pr
eca
utio
ns
or
treat
the
infected
le
aves
at
th
e
right
tim
e.
The
be
gi
nn
e
r
far
m
ers
c
an
al
s
o
detect
the
diseases
e
ve
n
without
pr
i
or
kn
ow
le
d
ge
about
it
.
A
fe
w
m
ajo
r
le
af
diseases
w
ere
detect
ed
i
n
the
Harum
anis
m
ang
o
fam
ily
su
ch
as
An
t
hracn
os
e,
m
ango
scab
,
an
d
powd
e
ry
m
il
dew
.
More
ov
e
r,
s
om
e
far
m
ers
m
ay
al
so
us
e
pestic
ide
to
av
oid
a
disea
se
ou
tb
rea
k
an
d
pest
gro
wth
befor
e
diag
nosin
g
the
diseases
.
Hence
,
the
pro
j
ect
can
reduce
the
us
e
of
hazar
dous
pestic
ides
t
hat
pose
a
si
gnific
a
nt
threat to
hum
a
n healt
h.
2.
RELATE
D
S
TUDIES
2
.
1
.
H
arum
ani
s mang
o
Ma
ny
m
ang
o
(Man
gifer
a
in
dica
Lin
n.)
is
a
tro
pical
f
r
uit
that
bel
ongs
m
or
ph
ologica
ll
y
to
the
delique
scent
dru
pe
s
ub
-
ty
pe
.
It
posse
sses
one
la
r
ge
see
d
wh
ic
h
is
s
urr
ounde
d
by
fles
hy
m
esocarp
an
d
has
m
any
var
ie
ti
es
[
2
]
.
In
Ma
la
ysi
a,
20
9
cl
ones
of
m
ang
o
ha
ve
been
re
gistere
d
by
the
De
par
t
m
ent
of
A
gr
ic
ul
ture.
An
y
cl
ones
th
at
po
ssess
s
uff
ic
ie
ntly
hig
h
qual
it
ie
s
m
ay
be
reco
m
m
end
ed
f
or
gen
e
ral
plantin
g
f
ro
m
thes
e
reg
ist
ere
d
cl
on
es.
Ma
lgo
a
an
d
A
pp
le
Ma
ng
o,
Ha
ru
m
anis,
MA
162,
MA
165,
MA
204,
an
d
MA
205
are
a
m
on
g
the
pro
m
isi
ng
ones
[
3
]
.
Harum
anis
m
ang
o
is
ho
nour
e
d
as
t
he
“
K
ing
of
Ma
ngoe
s”
in
this
c
ount
ry
due
to
it
s
delic
i
ou
s
ness,
unf
orgett
able
sweet
ta
ste
and
a
ro
m
at
ic
s
m
el
l
[4
]
.
The
fruit
is
al
so
ve
ry
rich
in
vitam
in
A,
C, an
d
m
edici
nal q
ualit
ie
s.
Its
le
aves ar
e
m
os
tl
y used
in
r
it
ua
ls as they
are
anti
-
bacte
rial
ag
ai
ns
t
gr
am
-
po
sit
ive
bacteria
[
5].
W
it
hin
t
he
s
pe
ci
es,
two
ty
pe
s
are
recog
nized.
T
he
first
ty
pe
pro
du
ce
s
a
seed
with
a
sing
le
zy
go
ti
c
em
br
yo.
It
is
cal
le
d
m
on
oem
br
yoni
c
m
ang
oes
t
hat
or
igi
nated
in
s
ub
-
tr
op
ic
al
I
nd
ia
,
and
th
e
othe
r
ty
pe
is p
olyem
br
yo
nic m
ang
oes
which c
reate a
se
ed wit
h
a
fe
w
e
m
br
yos [6].
2
.
2
.
H
arum
a
nis ma
ngo le
af dise
as
es
The
Harum
anis
le
aves
a
re
usual
ly
expose
d
to
sp
eci
fic
m
a
ngo
le
af
ty
pe
diseases.
The
anth
racnose
(
colle
totric
hum
glo
eos
porio
dies)
is
on
e
of
the
well
-
known
m
ang
o
le
a
f
diseases.
T
he
anthr
ac
nose
exh
i
bits
sever
e
f
ungal
bu
l
b,
see
ds,
a
nd
le
ave
s
f
ungu
s.
I
n
yo
ung
s
hoots
,
fl
ow
e
rs
a
nd
f
ru
it
s,
t
he
di
sease
causes
s
ever
e
losses
[
7
]
.
Othe
r
tha
n
that,
E
lsi
no
ë
m
ang
ife
rae.
It
is
al
so
known
as
De
nt
ic
ularia
m
anif
erae
cau
ses
m
an
go
scab.
O
nly
li
ve
plant
ti
ss
ue
ca
n
s
urvive
t
his
f
ungus.
The
re
a
re
no
rec
ords
of
this
disease
a
ff
ect
in
g
oth
e
r
plants
oth
e
r
tha
n
m
angoes
.
Sm
al
l
black
spots
de
velo
ped
on
ne
wly
set
f
ru
it
,
and
the
fruit
f
al
ls
off
wh
e
n
m
ul
ti
ple
black
le
sio
ns
infect
it
.
Th
e
rem
ai
nin
g
a
ff
ect
ed
f
ru
it
on
t
he
tree
de
velo
ps
sca
r
ti
ssu
e
that
m
akes
it
un
m
ark
et
ab
le
or down
gr
a
des
it
[
8
].
Ther
e
a
re
ot
he
r
le
af
diseases
li
ke
Xan
t
ho
m
on
as
cam
pestris
pv
.
Ma
ng
i
fer
a
ei
nd
ic
ae,
a
bac
te
rium
that
causes
black
s
po
t
ca
pa
ble
of
at
ta
cking
le
a
ve
s,
bra
nch
es
an
d
f
r
uit
[8
]
.
T
he
re
is
a
s
ubsta
nt
ia
l
season
al
va
riat
ion
in
the
disease
sever
it
y.
It
m
ay
be
s
pr
ea
d
th
rou
gh
wind
-
dri
ven
rain
f
ro
m
tree
to
tree
i
n
the
fiel
d
or
th
r
ough
too
ls
us
e
d
to
c
on
t
ro
l t
as
ks
s
uc
h
as
prunin
g
[
9]
.
The
fun
gu
s
,
O
idium
,
causes
pow
der
y
m
il
de
w
on
t
he
le
af
su
r
face
.
Alth
ough
a
m
od
e
ratel
y
sp
oradi
c
disease,
due
t
o
flo
wer
a
nd
pa
nicle
infecti
on
an
d
s
ubse
qu
e
nt
f
ru
it
set
fail
ur
e
,
it
m
ay
ca
us
e
se
ve
re
c
rop
lo
ss
.
The
c
riti
cal
ke
y
to
diff
e
re
ntiat
ing
this
dis
ease
is
the
presence
of
pan
i
cl
es
an
d
y
oung
fruit
of
a
w
hiti
sh
,
pow
der
y
gro
wth
of
the
fun
gus
.
T
he
in
fected
young
f
ru
it
c
ha
ng
e
s
c
olour
a
nd
fall
s.
T
he
whit
e
grow
t
h
of
youn
g
infected
le
a
ves
can
al
so
be
s
een
on
the
l
ower
la
ye
r
[
10]
.
This
stu
dy
is
fo
c
us
in
g
on
t
wo
ty
pes
of
di
seases
wh
ic
h
a
re
a
nth
rac
nose
a
nd
bacteria
l
blac
k
spot.
T
hese
t
wo
diseases
a
re
sel
ect
ed
a
s
these
a
re
the
m
os
t
pr
e
valent a
nd
prom
inent d
ise
ase, a
nd the
da
ta
f
or it
is e
xtensiv
el
y
ob
ta
ina
ble.
2
.
3
.
Re
la
ted w
orks
T
h
e
r
e
a
r
e
s
e
v
e
r
a
l
t
e
c
h
n
i
q
u
e
s
o
f
d
e
t
e
c
t
i
o
n
a
n
d
i
d
e
n
t
i
f
i
c
a
t
i
o
n
o
f
c
r
o
p
d
i
s
e
a
s
e
s
.
O
n
e
o
f
t
h
o
s
e
t
e
c
h
n
i
q
u
e
s
,
t
h
e
m
o
l
e
c
u
l
a
r
t
e
c
h
n
i
q
u
e
c
a
n
b
e
u
s
e
d
f
o
r
h
i
g
h
-
t
h
r
o
u
g
h
p
u
t
r
e
s
e
a
r
c
h
w
h
e
n
a
n
a
l
y
z
i
n
g
v
a
s
t
n
um
b
e
r
s
o
f
s
a
m
pl
e
s
[1
1
]
.
M
o
l
e
c
u
l
a
r
t
e
c
h
n
i
q
u
e
r
e
f
e
r
s
t
o
t
h
e
m
i
n
i
m
um
a
m
o
u
n
t
o
f
m
i
c
r
o
o
r
g
a
n
i
s
m
t
h
a
t
c
a
n
b
e
f
o
u
n
d
i
n
t
h
e
s
a
m
pl
e
,
a
n
d
E
L
I
S
A
i
s
o
n
e
o
f
t
h
e
w
i
d
e
l
y
u
s
e
d
m
o
l
e
c
u
l
a
r
m
e
t
h
o
d
s
f
o
r
d
i
s
e
a
s
e
d
e
t
e
c
t
i
o
n.
W
i
t
h
E
L
I
S
A
,
t
h
e
m
i
c
r
o
b
i
a
l
p
r
o
t
e
i
n
(
a
n
t
i
g
e
n
)
a
s
s
o
c
i
a
t
e
d
w
i
t
h
a
p
l
a
n
t
d
i
s
e
a
s
e
i
s
i
n
j
e
c
t
e
d
i
nt
o
a
n
a
n
i
m
a
l
w
h
i
c
h
p
r
o
d
u
c
e
s
a
n
t
i
b
o
d
i
e
s
a
g
a
i
n
s
t
t
h
e
a
n
t
i
g
e
n
i
n
t
h
e
d
e
t
e
c
t
i
o
n
o
f
d
i
s
e
a
s
e
.
T
he
s
e
a
nt
i
b
o
d
i
e
s
a
r
e
c
o
l
l
e
c
t
e
d
w
i
t
h
a
f
l
u
o
r
e
s
c
e
n
t
d
y
e
a
n
d
e
n
z
y
m
e
s
f
r
om
t
h
e
a
n
i
m
a
l
'
s
b
o
d
y
a
n
d
u
s
e
d
f
o
r
a
n
t
i
g
e
n
d
e
t
e
c
t
i
o
n
.
T
h
e
s
a
m
p
l
e
w
i
l
l
f
l
u
o
r
e
s
c
e
i
n
t
h
e
p
r
e
s
e
n
c
e
o
f
t
h
e
d
i
s
e
a
s
e
-
c
a
u
s
i
n
g
mi
c
r
o
o
r
g
a
n
i
s
m
(
a
n
t
i
g
e
n
)
,
t
h
u
s
c
o
n
f
i
r
m
i
n
g
t
h
e
e
x
i
s
t
e
n
c
e
o
f
s
p
e
c
i
f
i
c
p
l
a
n
t
d
i
s
e
a
s
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c
En
g
&
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
378
-
386
380
On
t
he
oth
e
r
ha
nd,
the
hy
per
s
pectral
te
c
hn
i
qu
e
can
be
us
e
d
to
c
ol
le
ct
valuab
le
plant
healt
h
inf
or
m
at
ion
ov
er
a
wide
ra
ng
e
of
wav
el
e
ng
t
hs
rangin
g
fro
m
35
0
to
25
00
nm
[1
1
]
.
For
la
rg
e
-
scal
e
c
ulti
vation,
hype
rsp
ect
ral
i
m
aging
is
pr
i
m
aril
y
being
us
e
d
f
or
plant
phe
no
ty
pi
ng
and
f
or
detect
ing
cr
op
diseas
e.
T
he
m
et
ho
dolo
gy
is
ver
y
sta
ble
a
nd
offer
s
a
br
i
ef
analy
sis
of
t
he
im
age
data.
It
is
widely
use
d
f
or
the
dete
ct
ion
of
plant
disease
by
m
easur
ing
t
he
changes
i
n
r
eflect
ion
resu
lt
ing
from
chan
ges
in
biop
hysic
al
and
bi
oc
hem
ic
al
char
act
e
risti
cs at i
nf
ect
io
n.
The
im
age
pro
cessi
ng
te
ch
niq
ue
is
an
oth
e
r
te
chn
iq
ue
us
e
d
for
c
rop
disea
se
detect
io
n.
S
tud
y
in
[
1
2
]
hav
e
decide
d
to
purs
ue
this
te
chn
i
qu
e
to
fi
gure
out
sp
eci
fic
diseases
an
d
to
pro
vid
e
the
treat
m
ent
f
or
the
far
m
ers
for
the
ir
sugarca
ne
c
r
op
s
.
T
he
te
ch
ni
qu
e
sta
rts
wit
h
the
le
a
f
im
a
ge
in
put
ta
ken
in
the
form
of
RGB
us
in
g
the
di
gital
ca
m
era
and
then
tra
ns
f
or
m
ed
into
gray
sca
le
to
el
i
m
inate
th
e
hu
e
a
nd
sat
ur
at
io
n
inf
orm
at
ion
.
Nex
t,
t
hro
ugh
i
m
age
pr
e
-
pro
cessi
ng,
un
nee
ded
pa
rt
of
t
he
i
m
age
is
re
m
ov
e
d
by
filt
ering
t
he
noise
a
nd
t
hen
the
ap
pro
pr
ia
te
portio
n
us
e
d
f
or
extracti
on
is
segm
ented.
A
fter
that
,
te
xt
ure
sta
ti
sti
cs
will
be
c
om
plete
d,
an
d
dise
ase
pr
e
ve
ntion
f
or
t
he
s
ug
arcane
leaves
is g
i
ven acco
rd
i
ng to
t
he
st
ud
y
.
Si
m
il
arly
,
i
m
a
ge
processi
ng
te
chn
i
qu
e
is
al
s
o
us
e
d
to
detect
the
m
os
t
fr
e
qu
e
ntly
occ
urr
ing
disease
s
on
c
otto
n
le
av
es
su
c
h
as
bacteria
l
bligh
t,
al
t
ern
a
ria,
grey
m
il
dew
,
cerc
ospora
,
a
n
d
f
us
ar
ium
wilt
[1
3
]
-
[
1
5
]
.
I
n
these
th
ree
pa
per
s
,
im
age
processin
g
al
s
o
sta
rts
with
im
a
ge
ac
qu
isi
ti
on
wh
e
re
hundre
ds
of
le
af
im
ages
are
captu
red
to
buil
d
a
database
.
The
im
ages
ar
e
la
te
r
pre
-
pro
cessed
t
o
im
pr
ov
e
certai
n
im
age
f
eat
ur
e
s
w
hich
a
re
crit
ic
al
fo
r
further
pr
ocessin
g.
The
pr
e
-
pr
oce
ssing
ste
ps
include
resizi
ng
the
i
m
age
into
250
x
25
0
pixe
l
an
d
app
ly
in
g
the
filt
er.
Ne
xt,
se
gme
ntati
on
of
t
he im
ages is p
erfo
rm
ed
so
that th
e inf
ect
ed regi
on
ca
n be sepa
rated
from
the
healt
hy
reg
io
n.
E
xtracti
on
of
the
f
e
at
ur
es
is
the
nex
t
crit
ic
al
st
ep
to
do
afte
r
segm
entat
ion
so
that
a
set
of
featu
res
represe
nting
e
ach
cha
racter
c
an
be
obta
ine
d.
This
increase
s
the
ackno
wledg
em
ent
rate
with
a
m
ini
m
u
m
nu
m
ber
of
c
om
po
nen
ts
.
Fo
ll
owi
ng
e
xtracti
on
of
the
featu
re,
a
cl
assifi
er
is
us
ed
to
i
den
ti
fy
the
disease
base
d on the
featu
res e
xtracted
.
The
stu
dy
in
[1
6
]
hav
e
op
t
ed
for
an
im
a
ge
processi
ng
te
chn
iq
ue
in
orde
r
to
get
use
fu
l
crit
ic
al
featur
e
s
for
t
he
analy
sis
of
var
i
ou
s
rice
bl
ast
diseases
.
T
he
le
af
im
ages
are
first
ac
qu
ired
t
hr
ough
a
we
b
ca
m
era
and
pr
ocesse
d
by
Ra
sp
be
r
ry
Pi.
Then
,
the
im
age
processi
ng
is
done
by
conver
t
ing
the
RGB
im
ages
into
gray
scal
e,
and
the
disea
sed
c
on
t
ours
a
re
ide
ntifie
d
by
app
ly
in
g
ed
ge
detect
io
n
(
so
be
l).
Nex
t
,
s
ever
al
analy
ti
cs
te
chni
qu
es
a
re
c
onduct
ed
t
o
ide
ntif
y
the
im
ages
accor
ding
t
o
the
sp
eci
fic
pro
ble
m
at
hand.
By
us
in
g
op
ti
m
iz
ation
te
chn
i
qu
e
s,
t
he
i
m
ages w
il
l t
he
n be se
nt to
the
cloud
stora
ge fo
r
c
om
par
ison.
Othe
r
resea
rc
he
rs
hav
e
al
so
us
e
d
im
age
processin
g
te
c
hniqu
e
w
hich
i
nvolv
es
sim
il
ar
ste
ps
s
u
ch
as
i
m
age
acq
uisit
i
on,
im
age
pr
e
-
processi
ng,
im
age
se
gm
entat
i
on,
f
eat
ur
e
ext
racti
on
an
d
cl
a
ssific
at
ion
[
5
]
,
[
17
]
-
[
19
]
.
I
n
these
pap
e
rs,
durin
g
the
i
m
age
segm
entat
ion
,
O
tsu'
s
m
et
ho
d
and
K
-
m
eans
cl
us
te
rin
g
are
use
d
t
o
cl
assify
the
obj
ect
base
d
on
a
set
of
the
fe
at
ur
es
into
K
nu
m
ber
of
cl
asses.
K
-
m
eans
cl
us
te
ring
is
m
or
e
su
pe
r
fici
al
than
ot
her
cl
us
te
r
ing
m
et
ho
ds
a
nd
w
orks
with
a
vast
nu
m
ber
of
var
ia
bles
as
well
.
H
owe
ver,
a
diff
e
re
nt
nu
m
ber
of
cl
ust
er
num
ber
s
and
diff
e
ren
t
init
ia
l
c
entr
oid
val
ues
pro
du
ces
a
d
i
f
fer
e
nt
cl
us
te
r
r
esult
.
Hen
ce
,
init
ia
lizing
the
a
ppropr
ia
te
num
ber
of
cl
us
te
r
k
an
d
the
co
rr
ect
init
ia
l
centro
id
nu
m
ber
is
nec
essary
.
So
m
e
of
the
draw
bac
ks
of
t
he
m
olecular
an
d
hy
per
s
pectra
l
te
chn
iq
ue
a
re
tim
e
-
con
s
um
i
ng,
la
bo
ur
-
inte
ns
ive
and
r
eq
uire
an
el
abo
rate
pro
cess.
The
dra
wb
ac
k
occ
urre
d,
espe
ci
al
ly
du
ri
ng
sam
ple
pr
e
par
at
io
n
(c
ollec
ti
on
and ext
racti
on)
in or
der to
obta
in r
el
ia
ble a
nd acc
ur
at
e
res
ul
ts on the i
den
t
ific
at
ion
of p
la
nt d
isa
bili
ty
[
20
]
.
Im
age
proces
sing
te
ch
niqu
e,
on
t
he
ot
her
ha
nd,
m
i
nim
i
zes
the
su
bject
ivit
y
of
tra
diti
on
al
cl
assifi
cat
ion
m
et
ho
ds
a
nd
m
ist
akes
com
m
itted
by
hu
m
an
bein
gs
[
17
]
.
Thu
s
,
the
reli
abili
ty
of
the
est
i
m
ation
is
en
han
ce
d,
a
nd
reli
able
dat
a
f
or
disease
s
tud
ie
s
a
re
giv
e
n.
The
ap
proa
ch
is
al
s
o
sim
ple,
i
n
wh
ic
h
i
t
on
ly
need
s
m
erely
c
om
pu
te
rs,
dig
it
al
ca
m
eras,
as
well
as
the
sof
tware
program
s.
The
im
age
proces
sin
g
ap
proac
h
al
so
m
ini
m
iz
es
the
subj
ect
ivit
y
of
tradit
io
nal
cl
assifi
cat
ion
m
et
ho
ds
a
nd
is
m
or
e
strai
ghtf
orward.
Othe
r
i
m
age
processi
ng
te
c
hn
i
qu
e
s
that
use
s
hybr
i
d
ap
pro
ach [
21]
,
im
a
ge
segm
entat
ion
[
22
]
an
d
cl
ust
er
-
ba
sed
featu
re
[23]
al
so
pro
ven
to
be
us
ef
ul
in
the
process
of
i
m
age
pr
oces
sing
te
ch
nique
.
Hen
ce
,
acco
rd
i
ng
to
the
s
ta
te
d
ben
e
fits,
the
pro
po
se
d
rec
ogniti
on
syst
em
i
m
ple
m
ents
the
i
m
age
proces
sin
g
te
ch
nique
to
de
te
ct
the
Harum
anis
m
a
ngo
le
a
f disea
s
es.
3.
METHO
DOL
OGY
As
sta
te
d
in
th
e
pr
e
vious
sect
ion
,
the
te
ch
ni
qu
e
util
iz
ed
for
the
devel
opm
ent
of
the
pr
oto
ty
pe
is
the
i
m
age
pr
oc
essing
te
c
hn
i
qu
e
.
Im
age
pr
oces
s
ing
te
ch
nolo
gy
us
ed
for
pla
nt
disease
detec
ti
on
el
i
m
inate
s
the
su
bject
ivit
y
of
tradit
ion
al
cl
as
sific
at
ion
m
et
h
od
s
a
nd
hum
a
n
-
i
nduce
d
erro
r.
Th
us,
the
est
i
m
ation
reli
abi
li
ty
is
i
m
pr
oved
,
an
d
accurate
data
acqu
ire
d
f
or
disease
stu
dies
.
The
te
ch
niqu
e
is
al
so
conv
enient,
wh
ic
h
needs
com
pu
te
rs,
dig
it
al
ca
m
eras
with
the
com
bin
at
io
n
of
ne
cessary
softwa
re
pro
gr
am
s
t
o
reali
ze
for
disease
detect
ion sy
ste
m
[24]
.
T
h
e
n
e
c
e
s
s
a
r
y
s
t
e
p
s
f
o
r
d
i
s
e
a
s
e
d
e
t
e
c
t
i
o
n
u
s
i
n
g
i
m
a
g
e
p
r
o
c
e
s
s
i
n
g
a
r
e
-
i
m
a
g
e
a
c
q
ui
s
i
t
i
o
n
,
p
r
e
-
p
r
o
c
e
s
s
i
n
g
,
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
,
s
e
gm
e
n
t
a
t
i
o
n
a
n
d
f
i
n
a
l
l
y
,
c
l
a
s
s
i
f
i
c
a
t
i
o
n
[
5
]
,
[
17
]
-
[
19
].
I
n
t
h
i
s
s
t
ud
y
,
t
h
e
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Ha
r
uma
nis
m
ango le
af d
ise
ase
reco
gnit
ion
s
yst
em usin
g
im
ag
e
proces
sin
g t
echn
i
qu
e
(
R.
A. JM.
G
i
ning
)
381
M
A
T
L
A
B
’
s
I
m
a
g
e
P
r
o
c
e
s
s
i
n
g
T
o
o
l
b
o
x
w
a
s
u
t
i
l
i
z
e
d
.
I
t
p
r
o
v
i
d
e
s
a
c
o
m
p
r
e
h
e
n
s
i
v
e
c
o
l
l
e
c
t
i
o
n
o
f
s
t
a
n
d
a
r
d
r
e
f
e
r
e
n
c
e
a
l
g
o
r
i
t
hm
s
a
n
d
w
o
r
k
f
l
o
w
f
r
a
m
e
w
o
r
k
s
f
o
r
i
m
a
g
e
p
r
o
c
e
s
s
i
n
g
,
a
n
a
l
y
s
i
s
,
v
i
s
u
a
l
i
z
a
t
i
o
n
,
a
n
d
a
l
g
o
r
i
t
hm
s
d
e
v
e
l
o
pm
e
n
t
.
T
h
i
s
t
o
o
l
b
o
x
w
a
s
u
s
e
d
t
o
p
e
r
f
o
r
m
t
h
e
i
m
a
g
e
p
r
e
-
p
r
o
c
e
s
s
i
n
g
,
i
m
a
g
e
s
e
gm
e
n
t
a
t
i
o
n
,
e
x
t
r
a
c
t
i
o
n
o
f
f
e
a
t
u
r
e
a
n
d
c
l
a
s
s
i
f
i
c
a
t
i
o
n
p
r
o
c
e
s
s
.
T
h
e
M
A
T
L
A
B
’
s
G
u
i
d
e
a
l
s
o
e
m
p
l
oy
e
d
t
o
d
e
v
e
l
o
p
t
h
e
G
U
I
o
f
t
h
e
p
r
o
t
o
t
y
p
e
.
3
.
1
.
Im
age
ac
quisi
tion
T
h
e
p
r
o
c
e
s
s
o
f
a
c
q
u
i
r
i
n
g
l
e
a
v
e
s
i
m
a
g
e
s
b
y
c
a
p
t
u
r
i
n
g
i
m
a
g
e
s
o
f
d
i
f
f
e
r
e
n
t
t
y
p
e
s
o
f
l
e
a
v
e
s
us
i
n
g
a
hi
g
h
-
r
e
s
o
l
u
t
i
o
n
c
a
m
e
r
a
t
o
o
b
t
a
i
n
g
o
o
d
r
e
s
u
l
t
s
a
n
d
e
f
f
i
c
i
e
n
c
y
[
2
5
]
.
T
h
e
l
e
a
v
e
s
'
i
m
a
g
e
s
w
e
r
e
t
a
k
e
n
m
a
n
u
a
l
l
y
u
s
i
n
g
a
R
e
dm
i
N
o
t
e
7
s
m
a
r
t
p
h
o
n
e
c
a
m
e
r
a
,
t
a
k
e
n
a
t
t
w
o
h
a
r
u
m
a
n
i
s
m
a
n
g
o
f
a
r
m
s
l
o
c
a
t
e
d
a
t
R
e
p
o
h
a
n
d
B
a
t
u
B
e
r
t
a
ng
k
u
p
,
P
e
r
l
i
s
.
T
h
e
r
e
a
r
e
1
0
3
h
e
a
l
t
hy
l
e
a
f
i
m
a
ge
s
,
a
n
d
4
3
i
m
a
g
e
s
o
f
t
h
e
d
i
s
e
a
s
e
d
l
e
a
v
e
s
(
a
n
t
h
r
a
c
n
o
s
e
a
n
d
b
a
c
t
e
r
i
a
l
b
l
a
c
k
s
p
o
t
)
w
e
r
e
o
b
t
a
i
n
e
d
.
E
a
c
h
t
y
pe
o
f
i
m
a
g
e
s
s
a
m
p
l
e
w
a
s
d
i
v
i
de
d
i
n
t
o
t
w
o
c
a
t
e
g
o
r
i
e
s
;
t
r
a
i
n
i
ng
(
±
7
0
%
)
a
n
d
t
e
s
t
i
n
g
(
±
3
0
%
)
i
m
a
g
e
s
,
s
a
v
e
d
i
nt
o
a
f
i
l
e
n
a
m
e
d
"
d
a
t
a
s
e
t
"
.
B
o
t
h
c
a
t
e
g
o
r
i
e
s
u
n
d
e
r
g
o
t
h
e
s
a
m
e
p
r
o
c
e
s
s
e
s
t
h
r
o
u
g
h
o
u
t
t
h
e
m
e
t
h
o
d
o
l
o
g
y
p
h
a
s
e
s
.
H
o
w
e
v
e
r
,
t
h
e
t
r
a
i
n
i
n
g
i
s
d
o
n
e
f
i
r
s
t
b
e
f
o
r
e
t
h
e
t
e
s
t
i
n
g.
H
e
n
c
e
,
t
r
a
i
ni
n
g
t
h
e
s
y
s
t
e
m
u
s
i
n
g
t
r
a
i
ni
n
g
i
m
a
g
e
s
w
i
l
l
b
e
d
i
s
c
u
s
s
e
d
f
i
r
s
t
,
f
o
l
l
o
w
e
d
b
y
t
h
e
t
e
s
t
i
ng
,
w
h
i
c
h
u
s
e
s
t
e
s
t
i
n
g
i
m
a
g
e
s
.
Fo
r
the
syst
e
m
trai
ning,
each
trai
ning
im
age
is
up
l
oa
ded
int
o
the
syst
em
,
ru
nnin
g
t
hro
ugh
a
se
ries
of
functi
ons
im
pl
e
m
enting
the
pre
-
processi
ng,
segm
entat
ion
,
and
featu
re
e
xtracti
on
phases
consecuti
vely
-
this
will
be
disc
us
s
ed
m
or
e
in
t
he
desc
riptio
n
of
each
phase.
T
he
syst
em
will
then
prom
pt
to
la
bel
t
he
upl
oad
e
d
i
m
ages
us
in
g
the
f
ollo
wing
ta
gs
:
0
f
or
he
al
thy,
1
f
or
a
nthracn
os
e
a
nd
2
f
or
bacter
ia
l
black
s
pot.
Thes
e
la
belle
d data ar
e save
d
i
n
a
da
ta
base
nam
ed
db.m
at
.
3
.
2
.
Im
age p
re
-
pr
ocessin
g
Be
fore
ext
racti
on
a
nd
cl
assifi
cat
ion
,
it
is
sta
nd
a
rd
pract
ic
e
for
le
af
i
m
ages
to
go
t
hro
ugh
pr
e
-
processi
ng.
Va
rio
us
pre
-
proc
essing
te
ch
niques
ca
n
be
ta
ken
int
o
acc
ount
t
o
el
im
inate
the
no
ise
f
r
om
the
acqu
i
red
im
ag
es;
the
us
a
ge
of
im
age
cl
ipp
in
g
to
get
the
a
re
a
in
quest
ion
by
crop
ping,
t
he
sm
oo
thin
g
filt
er
f
or
i
m
age
s
m
oo
thing,
an
d
i
m
age
enh
a
ncem
ent
t
o
co
ntr
ol
the
con
t
ra
st
le
vel
[1
7].
I
n
this
ph
ase,
the
noise
on
th
e
trai
ning
im
age
s
was
re
du
c
ed
us
in
g
a
filt
erin
g
te
ch
nique
-
im
age
enh
a
nce
m
ent,
wh
ic
h
increase
d
the
i
m
age'
s
con
t
rast.
Tw
o
MATLAB'
s
filt
ering
te
c
hn
i
que
functi
ons
we
re
ap
plied
in
t
his
phase,
im
adjust
to
inc
rea
se
th
e
con
t
rast
an
d
stret
chlim
to
ad
just
the
im
age
intensit
y.
T
he
GUI
for
te
sti
ng
was
al
s
o
devel
op
e
d
to
il
lustrate
th
e
ov
e
rall
te
sti
ng
pr
oce
sses.
T
he
imrea
d
f
un
ct
ion
is
util
ized
to
ena
ble
the
te
sti
ng
im
a
ge
uploa
d
fun
ct
ion
.
Figure
1
il
lustr
at
es the tw
o functi
ons'
ex
ecut
ion
with a
cli
ck on
the e
nhan
ce i
m
age contr
ast
butt
on.
Figure
1
.
Im
age enhan
cem
ent
3
.
3
.
Im
age se
gmen
tatio
n
Im
age
seg
m
entat
ion
is
a
m
e
tho
d
by
wh
ic
h
an
i
m
age
is
m
o
re
m
eaningfu
l
and
c
onve
nient
to
analy
ze
[25].
Se
gm
entat
ion
c
om
pr
ise
s
pa
rtit
ion
in
g
t
he
im
age,
or
an
y
relat
ion
,
int
o
sepa
rate
par
ts
of
the
sam
e
featur
es.
Upo
n
e
nh
a
nci
ng
the
im
age,
the
RGB
pictu
re
was
t
hen
tr
ansfo
rm
ed
to
hu
e
,
sat
ur
at
io
n,
an
d
inten
sit
y
(
HS
I
).
The
tra
nsfo
rm
at
ion
w
as
ena
bled
us
in
g
im2
double
f
un
ct
io
n
that
c
onve
rts
true
R
GB
col
our
im
age
to
doubl
e
pr
eci
sio
n
a
nd,
rescale
the
da
ta
and
extra
c
t
the
in
div
id
ua
l
com
po
ne
nt
of
the
im
age
w
her
e
a
ppr
opriat
e.
Af
te
r
wa
rd
s
,
th
e
co
nv
e
rsion
e
qu
at
io
n
is
a
ppli
ed
to
tra
nsfo
r
m
the
RGB
im
age
int
o
H
SI
for
m
at
.
Fi
gure 2
de
picts
the se
gm
ented
i
m
ages af
te
r
cl
ic
kin
g o
n
t
he
S
egm
ent I
m
age b
utto
n.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c
En
g
&
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
378
-
386
382
Figure
2
.
Im
age seg
m
entat
ion
3
.
4
.
Fe
ature
ext
r
act
i
on
The
e
xtracti
on
of
t
he
featu
re
play
s
a
vital
ro
le
in
im
age
processi
ng.
Fe
at
ur
e
e
xtracti
on
is
us
e
d
in
m
any
ap
plica
tio
ns
of
im
age
pr
oces
sin
g.
Col
our,
te
xture,
a
nd
m
or
phol
ogy
can
be
con
si
der
e
d
as
a
featur
e
f
or
le
af
dis
ease
de
te
ct
ion
.
T
he
gr
ey
le
vel
c
o
-
oc
currence
m
at
rix
(
GLCM)
for
each
pi
xel
m
a
p
f
or
H
a
nd
S
i
m
ages
of
i
nf
ect
e
d
cl
ust
ers
was
c
reat
ed
f
or
this
syst
e
m
.
GLCM
sta
ti
sti
cs
su
ch
as
m
ean,
sta
ndar
d
dev
ia
ti
on,
e
nt
ropy,
RM
S,
va
riance
, corr
el
at
io
n, e
nergy, a
nd
hom
og
eneit
y wer
e extracte
d.
3
.
5
.
Disease
c
lassifica
tion
Af
te
r
t
he
pr
e
vi
ou
s
phase'
s
ex
ecuti
on,
the
re
su
lt
co
uld
be
vi
ewed
by
cl
ic
ki
ng
t
he
vie
w
r
esult
butt
on.
Figure
3
sho
w
s
the
syst
em
s
uccess
fu
ll
y
pr
oc
essed
the
im
a
ge
a
nd
ide
ntifi
ed
t
he
le
af
dis
ease,
in
wh
ic
h
in
this
dem
on
strat
io
n, the lea
f
w
as
in
fected
by a
nthr
acnose.
Figure
3
.
Re
su
l
t display
4.
RESU
LT
S
A
ND
D
IS
C
USS
ION
The
syst
em
'
s
perform
ance
was
e
v
al
uate
d
to
te
st
it
s
ac
cur
acy
,
recall
,
preci
sion,
a
nd
F
1
-
sco
re
i
n
detect
ing
t
he
l
eaf
disea
ses.
T
he
detai
ls
of
th
is
evaluati
on
a
re
disc
us
se
d
i
n
this
sect
io
n.
T
o
deliver
the
s
yst
e
m
evaluati
on,
a
n
exp
e
rim
ent
was
carried
out
on
14
6
i
m
ages
of
the
t
hr
ee
c
hose
n
cl
asses
w
hich
wer
e
healt
hy
le
af
as w
el
l as le
af th
at
w
as inf
ect
ed
by an
t
hr
ac
nose an
d
bacte
rial
b
la
ck
spot.
Table 1
s
hows t
he
i
m
age d
at
aset
that
was
di
vid
e
d
in
to
two
s
ubset
s
wh
e
re
a
set
of
113
im
ages
are
us
ed
for
trai
ni
ng
a
nd
a
set
of
33
im
ages
are
us
e
d
for
t
he
test
.
Af
te
r
div
i
ding
the
i
m
ages
into
the
tw
o
subs
et
s,
the
33
im
a
ges
f
ro
m
the
test
data
wer
e
te
ste
d
an
d
th
e
resu
lt
s
wer
e
re
corde
d
in a
co
nfusion
m
at
rix
ta
ble.
A
m
at
rix
of
co
nfusi
on
is
a
ta
ble d
epict
ing
t
he
values
of
t
rue
neg
at
ive
,
true
po
sit
ive
,
false
po
sit
ive
as
well
as
false
neg
at
ive.
This
al
lo
w
s
tho
r
ough
res
ults
rather
tha
n
on
ly
the correct
proporti
on
of
gues
ses (
acc
ur
acy
) [3
]
. Fi
gure
4 s
hows
t
he ge
neral
co
nf
us
io
n
m
at
rix
ta
ble.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Ha
r
uma
nis
m
ango le
af d
ise
ase
reco
gnit
ion
s
yst
em usin
g
im
ag
e
proces
sin
g t
echn
i
qu
e
(
R.
A. JM.
G
i
ning
)
383
Table
1.
Divi
din
g i
m
age d
at
aset
s into
trai
ning
data an
d
te
st
data
Diseas
e class
Total i
m
ag
e
datas
e
ts
Tr
ain
in
g
data
Test data
Health
y
103
83
20
An
th
racno
se
30
20
10
Bacterial Bl
ack Sp
o
t
13
10
3
Total
146
113
33
Figure
4
.
Ge
ne
ral co
nfusi
on
m
at
rix
ta
ble
The
ne
xt
ste
p
was
to
cal
culat
e
the
accuracy,
pr
eci
sio
n,
rec
al
l
and
F1
sco
rin
g.
Acc
ur
acy
rep
rese
nt
s
the
am
ou
nt
of
t
est
data
that
w
ere
accu
ratel
y
identifie
d
ove
r
the
s
um
of
te
st
data.
I
n
(
1
)
disp
la
ys
the
f
or
m
ula
of
ov
e
rall
accu
rac
y.
Pr
eci
sio
n
is
the
rati
o
of
c
orrectl
y
exp
ect
ed
po
sit
ive obser
v
at
ions
to
the
a
ct
ual
total
pr
e
di
ct
ed
ob
s
er
vation.
I
n
(
2
)
il
lustrate
s
the
eq
uation
of
pr
eci
sio
n.
Re
c
al
l
is
the
rati
o
of
c
orrectl
y
predict
ed
posit
iv
es
to
al
l
act
ual
cl
ass
obser
vatio
ns
.
In
(
3
)
in
dicat
es
the
recall
f
orm
ula.
Last
ly
,
F1
is
t
he
har
m
on
ic
m
ean
of
p
r
eci
sion
and recall
a
nd is a
bette
r
m
eas
ur
em
ent than a
ccur
acy
.
I
n (
4
)
sh
ows
the
F1
-
s
cor
e
equati
on.
(1)
(2)
(3)
(4)
Wh
e
re
T
P
-
T
ru
e
P
os
it
ive,
TN
-
T
ru
e
Ne
gative,
F
P
-
F
al
se
p
os
it
ive,
FN
-
False
Ne
gative
.
Th
e
conf
us
io
n
m
at
rix
ta
ble
of
Ha
r
um
anis
Ma
ngo
Leaf
Disease
Detect
ion
Syst
e
m
is
sh
ow
n
in
Table
2.
The
c
la
sses
wer
e
a
healt
hy
le
af,
the
leaf i
nfect
ed by a
nthr
acnose a
nd the
le
af in
fe
ct
ed b
y t
he
bacte
rial
b
la
ck sp
ot.
Table
2
.
C
onf
usi
on m
at
rix
ta
ble o
f
the syste
m
Diseas
e class
TP
TN
FP
FN
Health
y
10
4
6
An
th
racno
se
7
3
Bacterial Bl
ack Sp
o
t
2
1
Fr
om
Table
2,
TP
in
dicat
es
th
at
the
le
af
is
ei
ther
healt
hy
or
infected
an
d
was
c
orrectl
y
identifie
d
by
the
syst
em
.
T
N
in
dicat
es
th
at
the
le
af
is
no
t
healt
hy
or
not
infecte
d
and
was
co
rr
e
ct
ly
identifie
d
by
the
syst
e
m
.
FN
indi
cat
es
that
the
l
eaf
is
infected
,
bu
t
the
syst
em
identifie
d
it
as
healt
hy.
FP
in
dicat
es
that
the
le
af
is healt
hy, b
ut
the syst
em
iden
ti
fied
it
as
infec
te
d.
Few
facto
rs
m
igh
t
c
on
t
rib
ute
to
wh
y
the
sy
stem
inaccur
at
el
y
detect
t
he
diseases.
Th
e
f
irst
facto
r
i
s
the
an
gle
from
wh
ic
h
the
le
af
i
m
age
was
ta
ken.
I
f
the
ca
m
era
was
not
po
sit
io
ne
d
wel
l,
the
captu
red
i
m
age
m
igh
t
be
intri
cat
e
for
t
he
sy
stem
to
proces
s
pri
m
arily
durin
g
the
se
gme
ntati
on
proce
ss.
T
he
le
a
f
w
as
sti
ll
consi
der
e
d
he
al
thy
even
th
ough
the
re
we
r
e
a
few
blac
k
or
brown
s
pots
on
the
le
a
f.
T
he
syst
e
m
m
igh
t
interp
ret
these
sp
ots
as
a
n
i
nfect
ed
le
af
w
he
n
it
wa
s
no
t.
T
he
la
ck
of
le
a
f
i
m
ages
infecte
d
by
ant
hr
ac
no
se
an
d
bacteria
l blac
k spo
t
m
igh
t al
so
contrib
ute to
wh
y t
he
le
a
f'
s inaccu
rate cl
ass
ific
at
ion
.
Table
3
show
s
the
accu
racy,
pr
eci
sio
n,
reca
ll
and
F
1
-
sc
or
e
for
eac
h
cl
ass
base
d
on
t
he
conf
us
io
n
m
at
rix
from
Ta
ble 2
. T
hese
re
su
lt
s w
ere
obta
ined
a
fter p
e
rfor
m
ing
the cal
culat
ion
base
d on
t
he
f
our
e
quat
ion
s
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c
En
g
&
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
378
-
386
384
pr
ese
nted
.
A
s
uccess
fu
l
cl
ass
ifie
r'
s
pr
eci
sio
n
s
hould
prefe
rab
ly
be
1,
an
d
high
re
cal
l
i
m
plies
that
the
c
la
ss
(a
sm
a
ll
nu
m
ber
of
F
N)
was
ide
ntifie
d
co
rr
ect
l
y.
F1
-
sc
ore
is
on
ly
one
if
both
recall
and
preci
sion
are
1.
The
F
1
scor
e
only
g
et
s
h
ig
h
i
f bo
t
h
re
cal
l and
preci
s
ion are
h
i
gh.
Af
te
r
the
e
xpe
rim
ent,
the
res
ults
obta
ine
d
wer
e
t
he
n
co
m
par
ed
an
d
a
naly
sed.
Table
4
de
picts
the
com
par
ison
be
tween
the
pro
po
s
ed
a
ppro
a
c
h
an
d
the
ot
he
r
two
e
xisti
ng
appro
ac
hes
ba
sed
on
the
cl
assifi
er
us
e
d
a
nd the t
hree sy
ste
m
s
'
ac
cur
acy
rate.
T
he
overall
acc
uracy
o
f
the
pro
pose
d
syst
em
is
68.89%
.
Table
3
.
T
he
a
ccur
acy
,
preci
s
ion
,
r
ecal
l a
nd
F1
-
sco
re
of the
syst
e
m
Diseas
e class
Accurac
y
(
%)
Precisio
n
Recall
F1
-
sco
re
Health
y
70
0
.71
1
0
.77
An
th
racno
se
70
1
0
.70
0
.82
Bacterial Bl
ack Sp
o
t
6
6
.67
1
0
.67
0
.80
Table
4
.
C
om
par
iso
n betwee
n t
he pr
opos
e
d s
yst
e
m
w
it
h
the
existi
ng syst
em
Pap
er
Accurac
y
(
%)
Health
y
8
7
.80
An
th
racno
se
8
3
.26
Bacterial Bl
ack Sp
o
t
6
8
.89
Fr
om
the
ta
ble
above
,
it
is
sh
own
t
hat
the
pro
po
s
ed
syst
em
had
the
l
ow
est
a
ccur
acy
rate.
T
he
la
ck
of
trai
ning
im
age
s
causes
this
lo
west
res
ult.
For
instance
,
both
existi
ng
syst
e
m
s
had
trai
ne
d
62
9
an
d
20
2
i
m
age
s
com
par
ed
t
o
the
propose
d
s
yst
e
m
,
wh
ic
h
on
ly
trai
ne
d
113
im
ages.
Both
existi
ng
sys
tem
s
al
so
us
e
d
SVM
cl
assifi
ers,
giv
i
ng
m
or
e accu
r
at
e resu
lt
s,
espe
ci
al
ly
w
hen m
or
e
im
ages w
er
e trai
ne
d.
Nex
t,
t
he
feat
ures
that
the
tw
o
existi
ng
syst
e
m
s
and
the
propose
d
syst
em
had
m
ay
al
so
i
m
pact
the
accuracy
of
t
he
syst
e
m
s.
Re
search
i
n
[
26
]
e
xt
racted
the
le
av
es'
te
xtu
re,
col
our,
a
nd
sh
a
pe
fe
at
ur
e
s,
wh
e
re
as
in
[27]
extracte
d
t
he
te
xture an
d
colo
ur
f
eat
ur
es
d
ur
i
ng
the
featur
e ext
racti
on
process
. Th
e pro
posed
syst
e
m
o
nly
extracte
d t
he
te
xture feat
ur
e;
t
hu
s
, t
his m
igh
t co
ntribute
to
a
n ov
e
rall
lack
of accu
racy.
5.
CONCL
US
I
O
N
This
pa
per
de
s
cribe
d
a
stud
y
on
dev
el
op
i
ng
a
syst
e
m
that
d
et
ect
s
disease
i
n
the
m
ango
le
af
us
in
g
an
i
m
age
process
ing
te
c
hn
i
qu
e
.
The
pro
po
se
d
syst
em
eff
ect
ively
detect
s
and
cl
assify
the
disease
wi
th
an
accuracy
of
68
.89%,
wh
ic
h
is
low,
re
su
lt
ed
f
ro
m
the
inadequacy
of
trai
ning
im
ages.
Ad
di
ti
on
al
diseases
will
be
inte
gr
at
e
d wit
h
the
r
ec
oginit
ion
syst
e
m
as p
a
rt of
fu
t
ur
e
works.
6.
ACKN
OWLE
DGE
MENT
We
grat
ef
ully
ackno
wled
ge
su
pp
or
t
by
Un
i
ver
sit
iTe
kn
ologi
MARA
Ca
wanga
n
Per
li
s,
Ma
la
ysi
a
unde
r
dan
a
pem
bu
dayaa
n
pe
nyel
idika
n
da
lam
an
(DPP
D)
un
der
G
ra
nts
No
60
0
-
T
NCPI
5/
3
/D
D
N
(09)
(
019
/20
20
).
REFERE
NCE
S
[1]
Chuanl
ei,
Z
.
,
Sh
anwe
n
,
Z
.
,
Juch
eng,
Y.,
Yan
cui,
S.,
and
Jia,
C
.
,
“
Apple
le
af
disea
se
ide
n
ti
fi
cati
on
using
gene
tic
al
gorit
hm
and
c
orre
lation
base
d
fea
tur
e
select
io
n
m
et
hod,
”
Int
e
rnational
Jour
n
al
of
Agric
u
lt
ur
al
and
Bi
olog
ical
Engi
ne
ering
,
vol
.
10
,
no
.
2
,
pp
.
7
4
-
83
,
2017
, d
oi
:
10.
3965/j.ijabe
.
2
0171002.
2166
.
[2]
Huang,
C.
Y.,
Kuo,
C.
H.,
W
u
,
C.
H.,
Kuan,
A.
W
.
,
Guo,
H.
R.
,
Li
n,
Y.
H.,
and
W
ang,
P.
K.,
“
Free
Radi
c
al
-
Scave
nging
,
Anti
-
Infl
amm
at
or
y
,
and
Antiba
c
terial
Act
ivi
t
ie
s
of
W
at
er
and
Et
hanol
Ext
r
ac
t
s
Prepa
red
fro
m
Com
pre
ss
iona
l
-
Puffing
Pretre
a
t
ed
Ma
ngo
(Mangif
era
indi
c
a
L.
)
Pee
ls,”
Jou
rnal
of
Food
Qualit
y
,
2018,
d
oi
:
10.
1155/2018/
10
25387
.
[3]
Yin,
M.
O.
,
a
nd
Na
y
,
C.
H
.
,
“
Plant
L
ea
f
Disea
se
Det
ec
t
i
on
and
Cla
ss
i
fi
ca
t
ion
using
I
m
age
Proce
ss
ing,
”
Inte
rnational
Jo
urnal
of Re
searc
h
and
Eng
ineeri
ng
,
vol
.
5
,
no
.
9
,
pp.
516
-
523
,
20
18,
d
oi
:
10
.
2127
6/i
jre.2018.
5
.
9.
4
.
[4]
Tha
ra
n
at
han
,
R.
N.,
Yashoda,
H.
M.,
and
Prabh
a,
T
.
N.,
“
Mango
(Mangif
era
in
dic
a
L
.
),
“
the
ki
ng
of
fruit
s”
-
An
over
vie
w,
”
In
F
ood
Revie
ws
Int
ernati
onal
, v
ol
.
22,
no
.
2
,
2006
,
doi
:
10
.
1080/87
559120600574493
.
[5]
Sethupa
th
y
,
J.
,
&
Veni,
S.,
“
OpenCV
base
d
disea
se
ide
nt
ifi
c
ation
of
m
ango
l
ea
ves,
”
Inte
rnat
ional
Journal
of
Engi
ne
ering
and
Technol
og
y
,
vol
.
8
,
no
.
5
,
pp
.
19
90
-
1998
,
2016
,
doi
:
10
.
21817/ij
et
/2016/
v8
i5/
16
0805417
.
[6]
Ploet
z
,
R.
C.
,
“
Di
sea
ses
of
tropi
c
al
frui
t
cro
ps,”
Dise
ases
of
Tr
opic
al
Fruit
Cro
ps
,
2003,
doi
:
10.
1079/978085
1993904.
0000
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Ha
r
uma
nis
m
ango le
af d
ise
ase
reco
gnit
ion
s
yst
em usin
g
im
ag
e
proces
sin
g t
echn
i
qu
e
(
R.
A. JM.
G
i
ning
)
385
[7]
El
qassas,
R.
,
an
d
Abu
-
nase
r,
S.
S.,
“
Expe
rt
S
ystem
for
the
Di
agnosis
of
Mango
Disea
ses,”
2
018
Inte
rnation
al
Confe
renc
e
on
Digit
al Arts, Me
dia
and
Te
chnology
(
ICDAMT)
,
vol.
2
,
no
.
8
,
pp
.
10
-
18
,
2018
.
[8]
Conde,
B
.
D
.
,
Pitke
thle
y
,
R.
N.,
Sm
it
h,
E. S.
C.
,
Kulkar
ni, V.
J.,
and
Darwin
.
,
“
Agnote. August,
2
007,
pp
.
1
-
5.
[9]
Pitke
thle
y
,
R.
,
P
at
hologi
st
,
P.
P.
,
and
Servi
ce
s,
D.,
“
Agnote
Ba
ct
er
i
al
B
la
ck
Spot of
Mangoe
s,”
Jul
y
,
2006
.
[10]
Nelson,
S.,
“
Mango
Pow
der
y
M
il
dew,”
Depart
ment
of
Pl
ant
a
nd
Env
ironmen
t
al
Protecti
on
Sc
ie
nc
es
,
6
,
2008,
htt
p://citese
erx
.
i
st.psu.
edu/
v
ie
wd
oc/
downlo
ad
?
do
i=
10.
1
.
1.
731
.
42
52&re
p=r
ep1&
t
y
pe
=pdf
[11]
Fang,
Y.,
and
R
amasam
y
,
R.
P.
,
“
Curre
nt
and
prospec
ti
v
e
m
et
ho
ds
for
pla
nt
dise
ase
det
e
ction,
”
B
iosensors
,
vol.
5,
no.
3
,
pp
.
537
-
5
61,
2015
,
doi
:
10
.
3390/bi
os50305
37
.
[12]
Mitka
l
,
P.,
Pawar,
P.,
Naga
n
e,
M.,
Bhosale
,
P.,
P
adwa
l,
M.,
and
Naga
ne,
P.,
“
Le
af
Disea
se
Dete
ction
an
d
Preve
nti
on
Us
ing
Im
age
Proc
essing
using Ma
tl
ab
,” p
p
.
26
-
30
,
201
6
.
[13]
Adhao,
A.
S.,
a
nd
Pawar,
V.
R.
,
“
Autom
at
ic
Cott
on
Leaf
Disea
s
e
Diagnosis
and
Control
li
ng
Us
ing
Raspber
r
y
Pi
and
IoT
,
”
Lec
ture
Not
es
in
Net
wo
rks and
Syst
ems
,
pp.
157
-
167
,
20
18,
doi
:
10
.
1007
/978
-
981
-
10
-
552
3
-
2_15
.
[14]
B
hong,
V.
S.
,
and
Pawar,
P
.
B.
V.,
“
Stud
y
and
Anal
y
sis
o
f
Cott
on
L
ea
f
Disea
se
D
etec
t
i
on
Us
ing
Im
age
Proce
ss
ing,
”
Int
ernati
onal
Jour
nal
of
Adv
anc
ed
Re
search
in
Scienc
e
,
Engi
n
ee
r
i
ng
and
Technol
ogy
,
vol.
3
,
no.
2
,
pp.
1447
-
1454
,
2016
.
[15]
Sara
ngdhar
,
A.
A.,
and
Pawar,
V.,
“
Mac
hin
e
l
e
arn
ing
r
egr
essio
n
te
chn
ique
for
cot
ton
leaf
dise
a
se
dete
ction
and
cont
rolling
usin
g
IoT,
”
Procee
dings
of
the
I
nte
rnational
Co
nfe
renc
e
on
Elec
troni
cs,
Comm
unic
ati
on
and
Ae
ros
pace
Te
ch
nology
,
ICECA
2017
,
pp
.
449
-
4
54,
2017
,
d
oi
:
10
.
1109/ICE
CA.20
17.
8212855
.
[16]
Ramesh,
S.,
&
Raj
ar
am,
B.
,
“
Io
t
base
d
cro
p
dis
ea
se
id
ent
if
ic
a
tion
sy
st
e
m
using
opti
m
iz
ation
technique
s,
”
ARP
N
Journal
of
Engi
n
ee
ring a
nd
App
lied
Sc
ie
n
ce
s
,
vol. 13, no.
4,
pp.
13
92
-
1395
,
2018
.
[17]
Pati
l,
B.
,
Hem
a
nt
,
P.
,
Shubha
m
,
Y.,
Arvind
,
S.,
and
Patil,
D.,
“
Plant
Mo
n
it
oring
Us
ing
I
m
age
Proce
ss
in
g
,
Raspber
r
y
P
I
an
d
IoT,
”
Int
ernational
Re
search
J
ournal
of
Engi
n
ee
ring
and
Tech
nology
(
IRJ
ET)
,
vol.
4,
no
.
10
,
p
p.
1337
-
1342
,
201
7
.
[18]
Patki
,
S.
S.
,
and
S
abl
e,
G.
S.
,
“
Cott
on
Leaf
Disea
se
Det
ec
t
ion
&
Cla
ss
ifi
c
at
ion
using
Multi
SV
M,”
Inte
rnat
ion
al
Journal
of
Adv
a
nce
d
Re
search
i
n
Computer
and
Comm
unic
ati
on
Engi
nee
ring
,
vol.
5,
no.
10
,
pp
.
165
-
168
,
2016,
d
oi
:
10
.
17148/IJ
ARCCE.2016.
5
1034
.
[19]
Rastog
i,
A.
,
Aro
ra,
R.
,
and
Shar
m
a,
S.,
“
Le
af
d
i
sea
se
detec
t
ion
and
gra
ding
usi
ng
computer
v
ision
te
chno
log
y
&
fuz
z
y
logic,
”
2nd
Inte
rnational
Confe
renc
e
on
Signal
Proce
ss
in
g
and
Inte
grate
d
Net
works
,
SP
IN
2015,
pp.
500
-
505
,
2015
,
d
oi
:
1
0.
1109/SP
IN.2015.
7095350
.
[20]
Sankar
a
n,
S.,
Mishra,
A.,
Eh
sani,
R
.
,
and
Davis,
C
.
,
“
A
re
vie
w
of
adv
anced
techniqu
es
for
det
e
ct
ing
plan
t
disea
ses,”
Co
mputers
and
El
ectronics
in
Agric
ul
ture
,
vol.
72,
n
o.
1
,
pp.
1
-
13
,
2010
,
d
oi
:
10.
1016/j.c
om
pa
g.
2010
.
02
.
007
.
[21]
Muham
ad
Farid
Mavi,
Zu
lki
fl
i
H
usin,
Badlishah
Ahm
ad,
Yasm
in
Mohd
Yac
ob,
R
ohani
S.
Moham
ed
Farook,
W
e
i
Keong
Ta
n
,
“
Mango
ripe
n
ess
c
la
ss
ifi
c
at
ion
s
y
s
te
m
using
h
y
br
i
d
te
chn
ique,”
I
ndonesian
Jour
nal
of
E
lectric
a
l
Engi
ne
ering
and
Computer
Sci
ence
(
IJE
ECS)
,
vol.
14,
no.
2,
pp.
859
-
868,
Ma
y
2019
,
doi
:
10.
11591/ijeecs.
v14.
i2.
pp859
-
86
8
.
[22]
Al
-
Furaij
i
,
O.
J.
M.,
Tua
n
,
N.
A.,
and
Yurev
ich,
T.,
“
A
new
fast
eff
i
ci
en
t
non
-
m
axi
m
um
supp
ression
al
gori
th
m
base
d
on
image
segm
ent
at
ion
,
”
I
ndonesian
Journal
of
El
ectric
al
Engi
ne
ering
and
Computer
Sci
enc
e
(
IJE
ECS)
,
vol.
19,
no
.
2
,
pp
.
11
55
-
1163
,
2020
,
doi
:
10
.
11591/ij
ee
cs.
v19
.
i2
.
pp10
62
-
1070
.
[23]
A
lha
ran
,
A.
F.
,
Fatl
awi
,
H.
K
.
,
and
Al
i
,
N
.
S.,
“
A
cl
ust
er
-
ba
sed
feature
se
lecti
on
m
et
hod
fo
r
image
t
ext
ur
e
cl
assifi
ca
t
ion,”
Indone
si
an
Journal
of
El
ectric
al
Engi
ne
ering
and
Computer
Sci
enc
e
(
IJE
ECS)
,
vol.
14,
no.
3
,
pp
.
1433
-
1442
,
201
9
,
doi
:
10
.
11591
/i
jeec
s.v14
.
i3
.
pp
1433
-
1442
.
[24]
Gupta,
M.
,
“
Pla
nt
Disea
se
De
tecti
on
using
Dig
it
al
Im
age
Proc
essing,
”
In
te
rna
ti
onal
Journal
of
Innov
at
io
n
&
Adv
anc
eme
nt
in compute
r sci
ence
,
vo
l. 7, no. 5, 2
018
.
[25]
Mainka
r,
P.
M.
,
Ghor
pade
,
S.,
an
d
Adawadka
r,
M.,
“
Plant
Le
af
Disea
se
Detect
io
n
and
Cla
ss
ifi
cat
ion
Us
ing
Im
age
Proce
ss
ing
Te
ch
nique
s,”
In
te
rnat
ional
Journal
of
Innov
ative
and
Eme
rging
Re
sea
rch
in
Engi
ne
eri
ng
,
vol.
2
,
no.
4
,
pp.
139
-
144
,
20
15
.
[26]
Es
-
Saad
y
,
Y.
,
El
Mass
i,
I.
,
E
l
Ya
ss
a,
M.,
Mam
m
ass,
D.,
and
Benazoun,
A.,
“
Auto
m
at
ic
re
cogni
t
io
n
of
pla
nt
l
ea
ves
disea
ses
base
d
on
seria
l
com
bination
of
two
SVM
cl
assifie
rs,”
Proce
ed
ings
of
2016
Inte
rnatio
nal
Confe
r
ence
on
El
e
ct
rica
l
and
In
formation
Techn
ologi
es,
ICEIT
2
016
,
2016
,
pp
.
5
61
-
566
, d
oi
:
10
.
1109/E
IT
ec
h.
20
16.
7519661
.
[27]
Adhao,
A.
S.,
a
nd
Pawar,
V.
R.
,
“
Autom
at
ic
Cott
on
Leaf
Disea
s
e
Diagnosis
and
Control
li
ng
Us
ing
Raspber
r
y
Pi
and
IoT
,
”
Lec
ture
Not
es
in
Net
wo
rks and
Syst
ems
,
pp.
157
-
167,
20
18,
d
oi
:
10
.
1007
/978
-
981
-
10
-
552
3
-
2_15
.
BIOGR
AP
HI
ES OF
A
UTH
ORS
R.
A.
J.
M.
Gi
nin
g
is
now
serving
as
a
le
c
ture
r
a
t
th
e
Facul
t
y
of
Co
m
pute
r
and
Mathe
m
at
i
ca
l
S
ci
en
ce
s,
Univer
s
it
i
Te
kno
logi
MA
R
Henc
e,
acco
rding
to
the
stated
bene
fi
ts,
the
proposed
r
e
cogni
ti
on
s
y
st
e
m
implements
the
image
pro
ce
s
sing
te
chn
ique
t
o
det
e
ct
th
e
Harum
ani
s
m
an
go
le
af
disea
ses
.
A,
Perli
s
Branc
h,
Malay
si
a.
H
e
rec
e
ive
d
his
Diploma
in
Com
pute
r
Science
and
Ba
chelor
Scie
nc
e
(Hons
)
spec
ia
lizin
g
in
Inform
at
io
n
Sy
st
ems
Engi
ne
eri
ng
fro
m
the
Univer
siti
Te
knolog
i
MA
RA,
and
Mas
te
r
of
Scie
nc
e
in
Cloud
Com
pute
r
from
Newca
stle Univ
ersity
,
Unit
ed
Ki
ngdom
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c
En
g
&
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
378
-
386
386
As
sociate
Pr
ofe
ss
or
Ts.
Dr
Sh
u
kor
San
im
Mohd
Fau
z
i
is
a
Deput
y
R
ec
to
r
(Re
sea
rch
and
Industria
l
Li
nk
a
ges)
at
Ui
TM
Perli
s
Branc
h
,
and
al
so
a
f
ac
u
lty
m
ember
of
Facul
t
y
of
Com
pute
r
and
Mathe
m
atical
Scie
nc
es,
Uni
ver
siti
Te
kno
lo
gi
MA
RA,
Perli
s
Bran
ch,
Malay
s
ia.
He
r
ec
e
ive
d
Mast
er
of
Sci
enc
e
(
Com
pute
r
Science
-
Real
-
T
i
m
e
Software
Engi
ne
eri
ng)
f
r
om
the
Cen
tre
for
Advanc
ed
S
oftwa
re
Engi
n
eering,
Univ
ersit
i
Te
knologi
Malay
s
ia.
He
th
en
obtained
his
PhD
in
Software
Engi
n
ee
ring
fr
om
the
Univer
si
t
y
of
New
South Wal
es
(U
NS
W
)
N.
M
.
Yu
soff
is
a
rese
arc
h
stude
nt
from
th
e
Facu
lty
of
Com
pute
r and
Math
ematic
al
Sc
ie
nc
es
and
is und
er
gu
i
danc
e
and
dir
ec
t
ion
from
th
e
r
ese
arc
her
s
of
Appli
ed
Com
puti
ng
a
nd
Te
chno
log
y
Res
ea
rch
Group.
T.
R
.
Raz
ak
cu
rre
ntly
works
a
t
Univer
siti
Te
kn
ologi
MA
RA,
Perli
s
Bran
ch,
M
al
a
y
si
a
and
al
so
a
fa
cul
t
y
m
ember
of
Fac
ulty
of
Com
pute
r
and
Math
ematica
l
Scie
n
ce
s.
He
recei
v
ed
Bac
he
lor
of
Inform
at
ion
Te
chn
olog
y
(Hons
)
spec
i
al
i
zi
ng
in
Artifi
c
ial
Inte
l
li
ge
nt
from
the
Univer
siti
Utar
a
Malay
si
a,
and
Master
of
Scie
nce
(Intelli
g
ent
Sy
stems
)
from
th
e
Univer
s
it
i
Utar
a
Mal
a
y
s
ia
.
He
the
n
obtain
ed
his
PhD
in
Com
pute
r
Scie
n
ce
from
the
Univer
sit
y
of
Notti
ngham,
Uni
te
d
Kingdom
.
M.
H
.
Is
mail
i
s
a
le
ct
ure
r
and
rese
arc
he
r
from
Facul
t
y
of
Com
pute
r
and
M
at
hemat
ica
l
Scie
nc
es,
Univer
siti
Te
kno
logi
M
ARA
,
Perli
s
Bra
nch,
Mal
a
y
s
ia.
He
obta
in
ed
his
first
degr
e
e
in
Data
Com
m
unic
a
ti
on
and
Ne
t
work
and
his
m
aste
r’s
degr
ee
in
i
nform
at
ion
tech
nolog
y
.
His
primar
y
r
ese
ar
c
h
int
er
est
is
Mobile
and
Pe
rva
si
ve
Com
puti
ng
a
nd
is
ac
ti
v
ely
i
nvolve
d
in
m
obil
e
t
ec
hnolo
g
y
soluti
on
in
h
i
s c
om
m
unity
.
N.
A.
Z
a
ki,
holds
a
Doctor
of
Philosoph
y
(Ph
D)
from
Univer
siti
T
eknol
ogi
MA
RA,
Shah
Alam
in
2018
in
the
fi
el
d
of
G
eomati
cs
(Remo
te
Sensing)
and
le
ads
to
th
e
fi
eld
of
ca
rbon
stocks
for
fore
sts.
Dr
Nurul
Ain
is
cur
ren
tly
se
rvi
ng
at
Univer
sit
i
Te
knologi
MA
R
A
(UiTM)
Perli
s Bra
n
ch at
the
Fa
cul
t
y
of
A
rch
itect
ur
e, Pla
n
ning
and
Surve
ying.
F.
Ab
du
ll
ah
is
cur
ren
tly
work
ing
as
Senio
r
R
ese
arc
h
Offic
er
at
M
al
a
y
si
an
Agric
ult
ur
al
Resea
rch
and
Deve
lopment
Insti
tut
e
(MA
RDI)
and
now
holds
a
positi
on
of
H
ea
d
of
Station
at
Hortic
u
lt
ure
Resea
rch
Cen
tre
MA
RDI
Sintok,
Keda
h,
Mal
a
ysia.
He
recei
v
e
d
Bac
hel
or
Scie
nc
e
Bio
-
ind
ustr
y
,
m
aj
oring
in
Plant
Phy
sio
lo
g
y
from
the
Univer
sit
y
Putr
a
Malay
s
ia.
He
the
n
obt
ai
ned
P
ost
Gradua
te
Di
ploma
in
Scie
n
c
e
(Hortic
u
lt
ura
l
Scie
nc
e)
and
Ph
.
D
in
Plant
Scie
nc
e
(Fruit
Crop
Ph
y
siolog
y
)
from
the
M
asse
y
Unive
rsit
y
,
Pa
lmerston
North,
New
Ze
a
la
nd.
Evaluation Warning : The document was created with Spire.PDF for Python.