Internati
o
nal
Journal of Ele
c
trical
and Computer
Engineering
(IJE
CE)
V
o
l.
6, N
o
. 4
,
A
ugu
st
2016
, pp
. 15
87
~
1
594
I
S
SN
: 208
8-8
7
0
8
,
D
O
I
:
10.115
91
/ij
ece.v6
i
4.9
986
1
587
Jo
urn
a
l
h
o
me
pa
ge
: h
ttp
://iaesjo
u
r
na
l.com/
o
n
lin
e/ind
e
x.ph
p
/
IJECE
Point P
r
ocessin
g
Meth
od for Improvin
g
Dental Rad
i
ology
Image Quality
Retn
o Supri
y
anti
1
, A
r
iep So
ela
i
ma
n
S
e
t
i
a
d
i
1
, Yo
gi
R
a
madh
ani
1
, Haris
B.
W
i
do
do
2
1
Departement of
Electr
i
cal
Engin
eering
,
Jend
eral Soedirman
Univ
ersity
, Indonesia
2
Departement of
Dentistr
y
,
Jend
eral
Soed
irman U
n
iversity
, Indon
esia
Article Info
A
B
STRAC
T
Article histo
r
y:
Received
Ja
n 21, 2016
Rev
i
sed
Mar
29
, 20
16
Accepted Apr 10, 2016
Radiolog
y
field
is ver
y
importan
t
in toda
y's
worl
d, es
peci
al
l
y
in t
h
e field of
medicine in
clu
d
ing dentistr
y
.
Radiolog
y
eq
uipment that is
popular in
dentistr
y is th
e p
a
noram
ic m
achi
n
e. A
panoram
ic
im
age fac
ili
tat
e
the dent
ist
in making a diagnosis of the
abnorma
lity
in
the mouth and
teeth
.
But
unfortunately
, f
o
r develop
i
ng countries
like In
donesia, p
a
noramic machin
e
avai
labl
e ar
e lo
w resolution whi
c
h have
an eff
e
c
t
on the r
e
sultin
g im
age als
o
has low qualit
y.
This resear
ch ai
m
s
to
improve t
h
e quality
of the panoramic
image to have a better qu
ality
. We
use point processing method with
em
phasis on contrast stre
tching
m
e
thod.
W
e
cho
s
e this m
e
thod b
ecause
it
is
quite simple but has a high perf
orma
nce. Bas
e
d on the second o
p
inion from
the hospital, th
e
performance is 8
3
.9%,
ther
efore this method is pr
omising to
be implemented
on the improvement
of d
e
ntal r
a
diolog
y
imag
es.
Keyword:
C
ont
ra
st
st
ret
c
hi
n
g
Den
tistry
Panoram
i
c im
age
Po
in
t pr
o
cessing
R
a
di
ol
o
g
y
fi
el
d
Copyright ©
201
6 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
R
e
t
no Su
pri
y
a
nnt
i
,
Depa
rtem
ent of Elect
ri
cal
E
n
gi
nee
r
i
n
g,
Jenderal Soe
d
irm
a
n Uni
v
ersit
y
,
K
a
m
p
u
s
Blater, Jl. Mayj
end
Su
ngk
ono
K
M
5, Blater
, Pur
b
al
in
gg
a,
In
don
esia
Em
ail: retno_s
upriyanti@uns
o
ed.ac.id
1.
INTRODUCTION
M
o
re t
h
an a
cent
u
ry
de
nt
i
s
t
r
y
pr
o
f
essi
o
n
usi
n
g
radi
og
ra
phi
c e
x
am
i
n
at
ion
as a m
eans t
o
o
b
t
a
i
n
d
i
agn
o
stic in
form
at
io
n
cann
o
t
b
e
o
b
t
ai
n
e
d
fro
m
th
e clin
ical ex
a
m
in
atio
n
an
d
o
t
h
e
r tests b
e
fo
re.
Un
til n
o
w,
dent
al
radi
og
ra
phy
bec
o
m
e
one
of
t
h
e i
m
po
rt
ant
t
o
ol
s
use
d
i
n
t
h
e
t
r
eat
m
e
nt
o
f
m
ode
rn
de
nt
i
s
t
r
y
.
S
h
o
o
t
i
n
g
g
ood
d
e
n
t
al r
a
d
i
og
r
a
ph
ic p
r
oj
ectio
ns bo
th
in
tr
a or
al
and e
x
tra-oral alm
o
st a co
m
m
on proced
u
r
e per
f
o
r
m
ed by
a de
nt
i
s
t
i
n
assi
st
i
ng t
h
e m
a
nagem
e
nt
of a
ca
se.
The
usef
ul
nes
s
of
Dent
al
R
a
di
o
g
ra
p
h
y
of
whi
c
h i
s
t
o
:
(
1
) R
a
di
o
di
ag
n
o
st
i
c
s /
R
ont
g
e
n Di
a
g
n
o
si
s:
use
d
to dia
g
nose suc
h
abnormalities apical or
per ap
ical
n
o
t
d
e
tected cl
in
ically, ab
normali
ties in
th
e j
a
w,
fract
u
r
e
o
f
t
h
e
ja
w
or
t
o
ot
h
ro
ot
a
n
d
hi
dd
e
n
ca
ri
es (t
h
e
pr
ox
im
al o
r
root car
ies)
secondary
ca
ries, ca
ries
i
n
ci
pi
ent
,
t
h
e
d
e
pt
h o
f
cari
e
s a
nd
ot
he
rs. (
2
)
Treat
m
e
nt
Pl
an:
assi
st
i
n
m
a
ki
ng
or
det
e
rm
i
n
i
ng a t
r
eat
m
e
nt pl
an,
su
ch
as th
e d
e
t
e
rm
in
atio
n
o
f
th
e lo
cation
o
f
t
h
e p
i
ns o
r
im
p
l
an
ts, cond
itio
n o
f
th
e roo
t
canal, d
e
term
in
ati
o
n
o
f
th
e typ
e
and
tech
n
i
qu
e
(3)
Sup
p
o
r
ti
n
g
Treat
m
e
n
t: h
e
lp
facilitate th
e co
ndu
ct of a treat
m
e
n
t
, such
as
post
o
perat
i
v
e com
p
l
i
cat
i
ons,
t
r
eatm
e
nt
en
do
don
tic (4) evalu
a
tio
n
o
f
Care: u
s
ed
to
evalu
a
te th
e su
ccess
or
pr
o
g
ress
o
r
c
o
nt
r
o
l
t
r
eat
m
e
nt (5
) R
a
di
o
g
ra
p
h
y
i
s
o
n
e
of t
h
e m
e
di
cal
record t
h
at
i
s
ve
ry
i
m
port
a
nt
an
d a
l
so f
o
r
(6
)
Intere
st f
o
r
e
nsic [
1
]
.
Based
o
n
u
s
abilit
y, it can
b
e
said
th
at t
h
e
Den
t
al Rad
i
ograph
y
p
l
ays an i
m
p
o
r
tan
t
ro
le in
variou
s
m
a
t
t
e
rs i
n
t
h
e f
i
el
d of
de
nt
i
s
t
r
y
.
One
o
f
t
h
e t
ool
s c
o
m
m
onl
y
used i
n
d
e
nt
a
l
radi
o
g
r
ap
hy
i
s
pa
no
ram
i
c
machi
n
e
[2]
.
R
e
fe
rs t
o
t
h
e ad
va
nt
ages
of
usi
n
g
dent
al
radi
og
rap
h
y
,
i
t
i
s
no
w al
m
o
st
al
l
dent
i
s
t
s
use
t
h
em
for t
h
e
b
e
nefi
t
o
f
t
h
eir
d
i
agn
o
sis.
W
e
h
a
v
e
to
rem
e
m
b
er that th
e i
m
ag
es p
r
od
u
c
ed
b
y
den
t
al rad
i
o
g
raph
y h
a
s
qu
alitie
s th
at
are hi
g
h
l
y
de
p
e
nde
nt
on
t
h
e
resol
u
t
i
o
n
of whi
c
h
i
s
o
w
ne
d by
t
h
e
ra
di
ol
ogy
m
achi
n
e.
Speci
fi
cal
l
y
f
o
r dent
al
radi
ol
o
g
y
,
t
h
e
equi
pm
ent
use
d
i
s
de
nt
al
pa
n
o
ram
i
c
m
achi
n
e as s
h
o
w
n i
n
Fi
gu
re
1.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 4
,
Au
gu
st 2
016
:
15
87
–
1
594
1
588
Figure
1. An e
x
am
ple of
pa
noram
i
c
m
achine
In
1
995
, Mo
lan
d
e
r [3
] d
i
d
a
research
abou
t
i
m
ag
e q
u
a
lities in
p
a
n
o
ram
i
c
rad
i
ograph
y
. In
prin
cip
l
e,
pan
o
r
am
i
c
radi
og
rap
h
y
i
s
di
vi
de
d i
n
t
o
t
w
o cl
assi
fi
cat
i
o
ns t
h
at
c
o
n
v
e
n
t
i
onal
a
n
d di
gi
t
a
l
.
C
o
n
v
e
n
t
i
onal
p
a
noram
ic h
a
s a lower reso
l
u
tio
n th
an
t
h
e d
i
g
ital
p
a
nor
amic.
W
h
ile in reality, p
a
rticu
l
arly in d
e
v
e
l
o
p
i
n
g
cou
n
t
r
i
e
s s
u
ch
as In
do
nesi
a,
pan
o
ram
i
c radi
o
g
ra
p
h
y
avai
l
a
bl
e i
s
t
h
e conve
nt
i
o
nal
t
y
pe and t
h
e
num
ber i
s
li
mited. Low resolution im
age will a
ffect to the accuracy of
diagnosis.
This pa
per
will discuss a
b
out si
m
p
le
m
e
thod
for im
proving im
age quality in l
o
w-resol
u
tion pa
noram
i
c radiography.
Acco
r
d
i
n
g t
o
t
h
e re
searc
h
a
b
out
pa
no
ram
i
c radi
og
ra
phy
,
R
u
sht
o
n
[
4
]
di
d a
researc
h
t
o
i
d
ent
i
f
y
t
h
e
radi
ol
o
g
i
cal
fi
ndi
ng
s fr
om
rout
i
n
e scree
n
i
n
g pa
n
o
ram
i
c radi
o
g
ra
p
h
s t
a
ke
n o
f
ad
ul
t
pat
i
e
nt
s i
n
ge
ne
ral
dent
a
l
p
r
actice.
H
o
wev
e
r
h
i
s study d
i
d
no
t pro
v
i
d
e
ev
i
d
en
ce to
supp
or
t
th
e pr
actice of
ro
u
tin
e
p
a
no
r
a
m
i
c
rad
i
o
g
raph
y
o
f
all n
e
w adu
lt p
a
tien
t
s. C
h
affin
[5
] d
i
d a research
t
o
ex
am
i
n
e th
e
v
a
lid
ity o
f
classi
fyin
g
In
itial
En
try
Train
i
ng
(IET) so
l
d
iers
i
n
to
d
e
n
t
al
fitn
ess
cl
ass
i
fi
cat
i
on
base
d s
o
l
e
l
y
o
n
e
x
am
i
n
i
ng
pan
o
ram
i
c
rad
i
o
g
raph
s. Tak
a
h
a
sh
i
[6
] i
n
v
e
stig
ated
th
e d
e
tectab
ility o
f
t
h
e m
a
n
d
i
bu
lar can
a
l on
d
e
n
t
al CT
reform
at
ted
im
ages of
pat
i
e
nt
s
bef
o
re
i
m
plant
o
p
e
r
at
i
o
n
.
Th
e
d
e
tectab
ility was sign
ifican
tly h
i
gh
er i
n
p
a
noram
ic v
i
e
w
s th
an
in p
a
rax
i
al v
i
ews. Su
o
m
alain
e
n
[7
] con
e
b
eam
co
m
p
u
t
ed
to
mo
grap
h
y
(CBC
T) d
e
v
i
ces in
ord
e
r fo
r im
p
r
ov
ing
qu
ality o
f
p
a
no
ram
i
c radio
l
o
g
y
.
Sezgin [8] did a research
to com
p
are the
effective orga
n doses fr
om
cone
beam
co
m
put
ed t
o
m
o
g
r
ap
hy
(CBCT), m
u
lt
islice co
m
p
u
t
ed
to
m
o
g
r
aph
y
(MSCT), a
nd p
a
no
ram
i
c ra
d
i
og
raph
y. Ert
a
s [9
] presen
ted
a
radi
ol
o
g
i
cal
fi
ndi
ng
s an
d fo
l
l
o
w-
u
p
of t
h
r
ee pat
i
e
nt
s wi
t
h
severe at
he
roscl
e
rosi
s t
h
a
t
was i
n
ci
dent
al
l
y
det
ect
ed o
n
P
a
no
ram
i
c R
a
diol
o
g
y
(PR
s
)
,
a
nd
di
scus
s t
h
e
rol
e
an
d i
m
p
o
rt
a
n
ce o
f
PR
s i
n
t
h
e det
ect
i
on
of
carot
i
d
a
r
t
e
ry
cal
ci
fi
cat
i
ons. R
o
t
o
ndi
[
1
0]
pr
o
pos
ed a t
ech
ni
que t
h
at
can be
used t
o
desi
g
n
a st
udy
m
easuri
n
g
in
terob
s
erv
e
r ag
reem
en
t with
an
y
nu
m
b
er
o
f
ou
tco
m
es an
d
an
y num
b
e
r o
f
raters. Accord
ing
to
th
ese
researc
h
, i
t
l
o
o
k
s t
h
at
t
h
ese
ki
nd
of
resea
r
ch
em
phasi
ze o
n
t
h
e f
unct
i
on
of
t
h
ese ki
n
d
of
m
e
di
cal
devi
ce
s suc
h
as carried
ou
t
b
y
Kao
[11
]
that p
r
o
p
o
s
ed
a
fu
lly au
to
m
a
tic
ab
do
m
i
n
a
l fat seg
m
en
tatio
n
syste
m
fo
r qu
antifyin
g
abd
o
m
i
nal
fat
,
i
n
cl
udi
ng
s
ubc
ut
ane
o
u
s
a
d
i
p
o
s
e t
i
ssue
an
d
visceral adi
pos
e
tissue. Als
o
C
h
ang [12]
pres
ented
a data
hidi
ng
method to c
onceal secret
data into co
l
o
r BTC com
p
ression c
o
de
by rea
r
ranging
high
mean a
nd
low m
ean enc
o
ding se
quences
.
The
pr
o
p
o
s
ed
m
e
t
hod i
s
at
t
e
m
p
ti
ng t
o
em
bed
m
o
re secr
et
dat
a
i
n
t
o
t
h
e col
o
r B
T
C
c
o
m
p
ressi
o
n
code
. Som
e
research t
h
at
d
i
scusse
d abo
u
t
im
age proces
si
ng are re
sea
r
ch di
d by
G
opi
nat
h
a
n
[
13]
t
h
at
pr
o
pose
d
a st
at
i
ona
ry
and di
s
c
ret
e
wavel
e
t
b
a
sed i
m
age denoi
si
ng sc
hem
e
and an F
F
Tb
ased i
m
age denoi
si
ng
schem
e
to re
move
Ga
ussian
noise
. In
the fi
rst approac
h
, high s
u
bba
n
ds a
r
e added
with
each ot
her a
n
d the
n
so
ft t
h
resho
l
d
i
n
g
is p
e
rfo
r
m
e
d
.
Th
e su
m
o
f
lo
w
subb
an
d
s
is filtered
with
eith
er p
i
ecewise lin
ear (PWL) or
Lag
r
ang
e
or splin
e in
terpo
l
ated
PWL filter. In
th
e seco
nd
ap
pro
ach, FFT i
s
e
m
p
l
o
y
ed
on
th
e no
isy i
m
ag
e an
d
then low fre
quency and
high fre
que
ncy coe
f
ficients are
separated with a s
p
ecified c
u
toff freque
n
cy. Fa
rida
h
[1
4]
di
sc
ussese
d t
h
e a
p
pl
i
cat
i
on
o
f
S
n
ake
t
o
fi
n
d
t
h
e
vis
u
al feature of lip
sha
p
es Als
o
re
search di
d by
Oul
oul
[15] propose
d
a ne
w al
gorithm
that com
b
ines t
h
e RG
B
imag
e with
Dep
t
h
m
a
p
wh
ich
is less sensib
le t
o
illu
m
i
nation changes
.
Sul
o
ng [16] propose
d a
m
e
thod
to
ove
rc
om
e
the low im
age cont
rast, com
b
ining two-
bl
oc
k feat
ure;
m
ean of
g
r
adi
e
nt
m
a
gni
t
ude
and
co
here
nc
e, w
h
ere t
h
e f
i
nge
rp
ri
nt
i
m
age i
s
segm
ent
e
d i
n
t
o
b
ackgr
oun
d, fo
r
e
g
r
ou
nd
or
n
o
i
sy r
e
g
i
on
s,
h
a
s
b
een
d
one. Ex
cep
t
f
o
r
th
e no
isy r
e
g
i
on
s in
th
e fo
r
e
gr
oun
d,
th
ere are still su
ch no
ises ex
i
s
ted
in
t
h
e
b
a
ck
gro
und
who
s
e coh
e
ren
ces are low, and
are mistak
en
ly assig
n
e
d
as fo
r
e
g
r
ou
nd
.
In
t
h
e
d
e
v
e
l
o
p
i
n
g
coun
tries su
ch in
Indo
n
e
sia, av
ailab
ility o
f
p
a
n
o
ram
i
c
d
e
v
i
ce is
no
t ev
en
ly. M
o
st
o
f
th
e eq
u
i
p
m
en
t is o
n
l
y av
ailab
l
e in
m
a
j
o
r h
o
s
p
itals in
th
e p
r
ov
i
n
cial cap
ital, wh
ile for th
e o
t
h
e
r cities are
not
al
way
s
a
v
a
i
l
a
bl
e. I
f
t
h
e
r
e
i
s
, us
ual
l
y
pa
n
o
ram
i
c devi
ces
with
l
o
w reso
l
u
tio
n.
In th
is
pap
e
r,
we
p
r
o
posed
t
o
i
m
p
r
ov
e im
ag
e qu
ality p
r
odu
ced
b
y
low-reso
lu
tion
p
a
n
o
ram
i
c d
e
v
i
ce
b
y
im
p
l
e
m
en
ti
n
g
po
i
n
t pro
c
essin
g
m
e
t
hod.
It
i
s
a
sim
p
l
e
m
e
t
hod an
d s
u
i
t
a
bl
e
fo
r i
m
pl
em
ent
i
ng i
n
devel
opi
ng c
o
unt
ri
es s
u
ch i
n
I
n
do
nes
i
a. In
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Po
in
t Pro
cessin
g
Method
fo
r
Imp
r
o
v
ing
Denta
l
Rad
i
o
l
o
g
y
Imag
e
Qu
a
lity (Retn
o
Sup
riyan
ti)
1
589
ou
r p
r
e
v
i
o
us r
e
search
[1
7]
-
[
22]
we
foc
u
se
d o
u
r
wo
r
k
o
n
t
h
e i
m
pl
em
ent
a
t
i
on
of
di
gi
t
a
l
im
age pr
oc
essi
ng
tech
n
i
qu
es in
th
e d
e
v
e
lop
m
e
n
t of techno
log
i
es su
ppo
r
ting
health se
rvi
ces in
rural a
r
eas, s
u
c
h
as
early
d
e
tectio
n
o
f
cataracts as well as o
p
tim
izat
io
n
of th
e low-
reso
lu
tion
u
ltraso
nog
raph
y, th
e
r
efore in t
h
is research
we still k
eep
em
p
h
a
size o
n
the o
p
tim
iza
tio
n
o
f
h
ealth
serv
ices in
d
e
v
e
l
o
p
i
n
g
co
un
tries. In
th
is case, we
ai
m
to
i
m
p
r
ov
e th
e qu
ality o
f
lo
w-reso
lu
tion
p
a
noram
ic
i
m
ag
e s
o
th
at p
e
o
p
l
e i
n
d
e
v
e
lop
i
ng
co
un
tries can
still b
e
abl
e
t
o
t
a
ke a
d
vant
a
g
e
of
t
h
e
pan
o
r
am
i
c
radi
ol
o
g
y
di
a
g
n
o
si
s o
p
t
i
m
al
l
y
.
2.
R
E
SEARC
H M
ETHOD
Digital im
age processi
ng tec
hni
que
s is an image pr
ocessi
ng to
produce
im
ages in accorda
n
ce
wit
h
o
u
r
wish
es o
r
q
u
a
lity
to
b
e
better.
Im
ag
e
p
r
o
cessing
o
p
e
ratio
n
s
can
b
e
d
i
v
i
d
e
d
can
b
e
div
i
d
e
d
i
n
to
o
p
e
ratio
n
s
t
h
at
ge
nerat
e
a
n
out
put
base
d
o
n
t
h
e
val
u
e
of a
si
n
g
l
e
pi
x
e
l
,
t
h
e o
p
e
r
at
i
o
n
t
h
at
p
r
od
uc
es o
u
t
p
ut
base
d
on a
collection
of
pixels adjace
nt to each
other a
n
d the
ope
rati
on that produces
out
put
base
d on all the pi
xels in the
im
age. Poi
n
t
p
r
oces
si
n
g
m
e
t
h
od i
s
pa
rt
o
f
a
n
o
p
e
r
at
i
o
n
t
h
at
gene
rat
e
s o
u
t
p
ut
base
d
o
n
t
h
e val
u
e
of a
si
ngl
e
pi
xel
t
h
at
co
ve
rs t
h
res
hol
di
n
g
,
co
nt
rast
st
ret
c
hi
n
g
a
n
d
g
r
ay
l
e
vel
re
d
u
ct
i
o
n
[2
3]
.
2.
1.
Pix
e
l Operatio
n
In
th
e im
ag
e pro
cessing
, th
ere is a ter
m
o
p
e
ratio
n
s
p
i
x
e
ls o
r
so
m
e
ti
mes
called
o
p
e
ration
s
p
i
x
e
l-to-
pi
xel
.
O
p
erat
i
o
n
of
pi
xel
s
i
s
t
h
e i
m
age pr
oc
essi
ng
o
p
erat
i
o
n t
h
at
m
a
ps ea
ch
pi
xel
rel
a
t
i
ons
hi
p t
h
at
rel
i
e
s on
pi
xel
i
t
s
el
f. I
f
f
(y
, x
)
ex
p
r
ess
e
d t
h
e
val
u
e
of
a pi
xel
i
n
t
h
e i
m
age of
f a
nd
g (y
,
x) e
x
pres
sed
pi
xel
p
r
oce
ssi
n
g
resul
t
s
fr
om
f (
y
, x
)
, t
h
e rel
a
t
i
ons
hi
p [
2
0]
ca
n
be e
x
p
r
esse
d
by
E
quat
i
o
n
1.
(1
)
In t
h
i
s
case, T
st
at
i
ng t
h
e
fu
nct
i
o
n
o
r
o
p
e
r
at
i
on
ki
n
d
i
m
p
o
se
d o
n
pi
xel
f
(y
, x
)
.
O
p
erat
i
ng m
odel
i
s
wh
at will
b
e
d
i
scu
ssed
i
n
th
is p
a
p
e
r.
2.
2.
Brightne
ss Increasing
The
basi
c o
p
e
r
at
i
on i
s
oft
e
n p
e
rf
orm
e
d o
n
t
h
e im
age
is th
e
in
crease i
n
b
r
i
g
h
t
n
e
ss. Th
is
op
eration
is
necessa
ry in
order t
o
m
a
ke the im
age becom
e
s bri
g
hter
. M
a
them
atically, the increa
sed
brightness is
done by
addi
ng
a co
nst
a
nt
t
o
t
h
e
val
u
e of t
h
e e
n
t
i
r
e
pi
xel
[
2
4]
. F
o
r
exam
pl
e, f (y
,
x) e
x
pres
sed t
h
e pi
xel
val
u
es
i
n
t
h
e
gray scale im
age at c
o
ordinat
e
s (y,
x). T
h
us, a
ne
w i
m
age b
ecom
e
s as exp
r
essed
by
E
q
uat
i
on
2.
(2
)
A ne
w i
m
age
has i
n
c
r
eased
t
h
e bri
ght
ness
val
u
e
of al
l
pi
xel
s
o
n
the image of the ori
g
inal
large
β
f (y
, x)
.
If
th
e
β
form
o
f
n
e
g
a
tiv
e
n
u
m
b
e
rs, th
e
brigh
t
n
e
ss will d
e
crease
o
r
b
e
co
m
e
d
a
rk
er.
2.
3.
Contr
a
st Stretching
The c
ont
ra
st
i
n
an i
m
age st
at
i
ng t
h
e
di
st
ri
but
i
o
n
of l
i
g
ht
and
da
rk s
h
a
d
es o
f
c
o
l
o
r.
A g
r
ay
-scal
e
im
age i
s
sai
d
t
o
ha
ve a l
o
w c
ont
rast
w
h
en
t
h
e di
st
ri
b
u
t
i
o
n
of c
o
l
o
r t
e
n
d
t
o
na
rr
ow
t
h
e
ran
g
e
of
gray
l
e
vel
s
.
Conversely, if the im
age ha
s a hi
gh
co
nt
r
a
st
ran
g
e
o
f
g
r
ay
l
e
vel
s
di
st
ri
b
u
t
e
d
o
v
er
wi
de.
C
o
nt
rast
can
b
e
m
easured
base
d o
n
t
h
e di
ffe
r
e
nce
bet
w
ee
n t
h
e val
u
e
of t
h
e
hi
g
h
est
i
n
t
e
ns
i
t
y
and t
h
e l
o
west
i
n
t
e
n
s
i
t
y
val
u
es
that
m
a
ke up the pi
xels in the im
ag
e
[24]. In orde
r
the dist
ribution of pixel
intensity
changes necess
a
ry
t
o
st
ret
c
h t
h
e
c
ont
rast
. T
h
i
s
i
s
d
o
n
e
by
u
s
i
n
g t
h
e
f
o
rm
ul
a as ex
press
e
d
by
E
q
u
a
t
i
on
3.
(3
)
Based
on the a
b
ove
form
ula, the c
ont
rast wil
l
increase i
f
α
> 1
and
th
e con
t
rast
will su
ffer
if
α
<1.
2.
4.
Brightne
ss an
d
Con
t
rast Combination
Ope
r
at
i
o
n f
o
r
i
n
creasi
n
g
bri
ght
ness a
n
d c
ont
rast
st
ret
c
hi
ng
can
be
per
f
o
r
m
e
d sim
u
l
t
aneo
usl
y
i
n
o
r
d
e
r t
o
im
p
r
ov
e im
ag
e q
u
a
lity [2
4
]
.
In
g
e
neral, a co
m
b
inatio
n
of th
e t
w
o
op
eratio
n
s
can
b
e
exp
r
essed
b
y
Equ
a
tio
n 4.
(4
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 4
,
Au
gu
st 2
016
:
15
87
–
1
594
1
590
Howev
e
r, if we in
tend
t
o
m
a
k
e
arrang
em
en
ts so th
at
th
e
gray lev
e
l i
n
th
e im
ag
e o
f
f i
n
t
h
e
rang
e
of
f1
an
d
f2
be
t
h
e im
age o
f
g
w
i
t
h
ceda
r
bet
w
e
e
n
g1
an
d
g
2
, t
h
e
req
u
i
r
e
d
fo
r
m
ul
a i
s
expres
sed
by
E
quat
i
o
n
5.
(5
)
2.
5.
Histog
ra
m Equa
liza
t
io
n
Hi
st
o
g
ram
equal
i
zat
i
on)
pr
o
v
i
d
e a so
phi
st
i
cat
ed
m
e
t
hod
for m
odi
fy
i
n
g t
h
e dy
nam
i
c rang
e and
co
n
t
rast of an i
m
ag
e b
y
altering
th
at im
ag
e su
ch
th
at its in
ten
s
ity h
i
sto
g
ram
h
a
s a desired
sh
ap
e.
Un
lik
e
cont
rast
st
ret
c
hi
n
g
,
hi
st
og
ra
m
m
odel
i
ng
ope
rat
o
rs m
a
y em
pl
oy
no
n-
l
i
n
ear
and
no
n-
mo
n
o
t
o
ni
c
tran
sfer
fun
c
tion
s
to
map
b
e
tween
p
i
x
e
l in
tensity v
a
lu
es in
t
h
e in
pu
t and
ou
tp
u
t
im
ag
es. Histo
g
r
am
eq
u
a
l
i
zatio
n
em
pl
oy
s a m
o
not
oni
c,
n
o
n
-
l
i
n
ear m
a
ppi
n
g
whi
c
h re
-assi
g
n
s t
h
e i
n
t
e
nsi
t
y
val
u
es
of
pi
x
e
l
s
i
n
t
h
e i
n
put
im
age
su
ch
th
at t
h
e
ou
tpu
t
im
ag
e con
t
ain
s
a un
ifo
r
m
d
i
strib
u
tio
n
o
f
in
ten
s
ities
[2
5
]
.
3.
R
E
SU
LTS AN
D ANA
LY
SIS
The
data use
d
i
n
this
researc
h
are pa
noram
i
c
im
ag
es t
a
ken
f
r
om
Dent
al
H
o
spi
t
a
l
Jen
d
eral
Soe
d
i
r
m
a
n
Un
i
v
ersity, Purwok
erto
. Ex
am
p
l
es
o
f
i
n
pu
t i
m
ag
e is shown in
Fi
g
u
re
2
.
Fi
gu
re
2.
Exa
m
pl
es of i
n
put
im
age (so
u
r
ce,
W
i
d
o
d
o
[
2
]
)
In
t
h
e
p
r
o
cess
o
f
im
p
r
ov
ing
i
m
ag
e qu
ality b
y
u
s
ing
co
n
t
rast stretch
i
ng
m
e
th
od
, startin
g
b
y
read
ing
i
n
p
u
t
i
m
age as sho
w
n i
n
Fi
gu
r
e
2.
T
h
e
next
s
t
ep i
s
t
o
cha
nge the im
age size to
re
duce
t
h
e com
putational load.
The
n
i
n
or
der
t
o
d
o
c
ont
ra
st
st
ret
c
hi
n
g
pr
oc
ess, t
h
e
first co
lor im
ag
e tran
sfo
r
m
e
d into
a grayscale image.
After
getting
a grayscale image, the
ne
xt is to anal
yze the hist
ogram
of eac
h input im
age as a ba
sed t
o
cal
cul
a
t
e
t
h
e val
u
e of R
1
, S
1
,
R
2
and S
2
.
In
ou
r pre
v
i
ous
research
[2],
we did this anal
ysis and we cl
assified
our im
ages data into three cla
ssifications, there are te
nd t
o
left, ten
d
to
ri
g
h
t
an
d tend
to
left and
righ
t After
knowing the tende
ncies of each histogra
m
,
it can be done furt
her analy
s
is
is to determine the value of the
gray
-R
1, S
1
,
R
2
, S2
, A
1
,
A2 a
nd
A3 as
di
scusse
d i
n
ou
r p
r
evi
o
u
s
r
e
sul
t
s
[2]
an
d
det
a
i
l
e
d resul
t
s are
descri
bed
i
n
Fi
gu
re 3.
Fi
gu
re
3.
C
a
l
c
ul
at
i
on
res
u
l
t
o
f
c
ont
ra
st
st
ret
c
hi
n
g
val
u
es
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Po
in
t Pro
cessin
g
Method
fo
r
Imp
r
o
v
ing
Denta
l
Rad
i
o
l
o
g
y
Imag
e
Qu
a
lity (Retn
o
Sup
riyan
ti)
1
591
Accord
ing
to
th
e d
a
ta ob
tain
ed
it ap
p
e
ars there ar
e som
e
di
ffere
nces
whe
n
t
e
st
i
ng i
n
put
on a
n
y
t
y
pe
of hi
st
og
ram
as fol
l
o
ws (1
)
For t
h
e cat
eg
o
r
y
of b
r
i
g
ht
hi
st
og
ram
,
have a sprea
d
of
gr
ay
t
h
at
dom
i
n
at
es t
h
e
righ
t sid
e
, th
e i
n
pu
t to
S1
is smaller th
an
R1, is cau
sed
b
y
in
pu
t S1
sm
al
le
r can
en
courage ch
ang
e
s in
l
o
catio
n
o
f
p
i
x
e
ls
n
e
w th
at lead
s to th
e left, so
t
h
at the typ
e
of ideal
histogram
can
be ac
hi
eve
d
. (2) For the
category
of
dar
k
hi
st
og
ram
,
ha
ve a sp
rea
d
of
gray
t
h
at
d
o
m
i
nat
e
s t
h
e l
e
ft
si
de, t
h
i
s
wa
s due t
o
t
h
e i
n
put
S
1
l
a
r
g
er s
o
as t
o
enco
u
r
age a
c
h
an
ge
of
pi
xel
s
ne
w l
eads
i
n
t
h
e ri
ght
di
re
ct
i
on s
o
t
h
at
t
h
e t
y
pe
of i
d
e
a
l
hi
st
og
ram
can
be
achi
e
ve
d
by
d
i
st
ri
but
i
o
n
of i
t
s gray
val
u
es
e
v
enl
y
f
r
om
l
e
ft
t
o
ri
g
h
t
.
(3
) F
o
r t
h
e cat
eg
ory
of e
v
e
n
l
y
dar
k
an
d
bri
ght
hi
st
o
g
ra
m
,
have a g
r
a
y
sprea
d
fai
r
l
y
evenl
y
s
h
ape
d
val
l
e
y
pea
k
fr
om
l
e
ft
t
o
ri
ght
. R
e
fer
r
i
n
g
t
o
t
h
e
vi
ew
poi
nt
o
f
t
h
e i
n
p
u
t
alm
o
st
t
h
e sam
e
ty
pe of
da
rk
hi
st
o
g
ram
.
One
o
f
t
h
e t
h
i
ngs
t
h
at
di
st
i
n
g
u
i
s
he
s t
h
at
i
n
d
e
term
in
in
g
the esti
m
a
ted
in
p
u
t
g
r
ay v
a
l
u
e, wh
ich
seek
p
i
x
e
l in
ten
s
ity h
i
sto
g
ram
at
th
e cen
ter so
th
at it can
be see
n
clearl
y
balance t
h
e
sprea
d
of gray
of the im
age histogram
.
Aft
e
r ente
ring
a
value of
gray i
n
eac
h
im
age whi
c
h t
o
t
a
l
e
d 2
18
pan
o
ram
i
c im
age, t
h
e next
st
ep i
s
doi
n
g
f
o
r t
h
e
m
a
i
n
proces
ses
cont
rast
st
ret
c
hi
n
g
.
For com
p
arative a
n
alysis, we com
p
ared
t
h
e res
u
l
t
s
o
f
gra
y
scal
e hi
st
og
ra
m
and co
nt
rast
st
ret
c
hi
n
g
hi
st
og
ram
as desc
ri
be
d i
n
Tabl
e
1.
Tabl
e
1. E
x
am
pl
e o
f
C
o
m
p
ari
s
on
R
e
sul
t
bet
w
een
G
r
ay
scal
e Hi
st
o
g
ram
an
d C
ont
ra
st
St
re
t
c
hi
ng
Hi
st
og
r
a
m
O
r
iginal I
m
age
Grayscale
H
i
st
ogra
m
Contrast S
t
retc
h
i
n
g
I
m
age
Contrast Stretching
H
i
st
ogra
m
C
ont
ra
st
st
ret
c
hi
n
g
m
e
t
hod i
s
a t
ech
ni
q
u
e t
o
get
a
new
i
m
age wi
t
h
bet
t
e
r c
ont
rast
t
h
a
n
t
h
e co
nt
rast
of
the ori
g
inal image. T
h
e idea of
t
h
e co
nt
rast
s
t
ret
c
hi
n
g
p
r
oce
ss i
s
t
o
im
prov
e t
h
e dy
nam
i
c
fi
el
d o
f
g
r
ay
l
e
vel
i
n
t
h
e i
m
age t
o
b
e
p
r
oces
sed
.
We c
oncl
u
de,
cont
rast
st
ret
c
hi
n
g
m
e
t
hod i
s
a m
e
t
hod
of
i
m
age en
ha
nce
m
ent
by
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 4
,
Au
gu
st 2
016
:
15
87
–
1
594
1
592
in
creasing
t
h
e con
t
rast of t
h
e im
ag
e th
rou
g
h
sev
e
ra
l co
nd
itio
ns. Th
e con
d
ition
is
wh
en
t
h
e process of
man
i
p
u
l
ating
th
e
p
i
x
e
ls
d
o
not u
s
e th
e sam
e
o
p
e
ration
o
n
al
l h
i
sto
g
ram
s
, bu
t u
s
i
n
g so
m
e
o
f
t
h
e co
nd
ition
s
,
so
that the curre
n
t im
age settings can be m
o
re
efficient b
eca
use of s
o
m
e
parts of t
h
e im
a
g
e there nee
d
s
to be
sl
i
ght
l
y
en
hanc
ed
bri
ght
ness a
n
d
t
h
e
r
e al
so
n
eeds t
o
be m
u
ch i
m
pro
v
ed
b
r
i
ght
ness.
In
or
de
r t
o
kn
ow c
h
a
r
act
eri
s
t
i
c
of co
nt
rast
st
ret
c
hi
n
g
hi
s
t
og
ram
,
we t
ook t
w
o sam
p
l
e
s of
ran
d
o
m
im
ages. C
h
ara
c
t
e
ri
st
i
c
s obt
ai
ned
by
i
n
vest
i
g
at
i
ng
di
ffe
renc
es location
of the pi
xels in t
h
e
grayscale
histogram
and
co
nt
rast
st
r
e
t
c
hi
ng
hi
st
og
r
a
m
on eac
h i
m
age t
h
at
i
s
p
r
es
ent
e
d i
n
Fi
gu
re
4.
Fi
gu
re
4.
A
n
e
x
am
pl
e of
di
f
f
e
rence
pi
xel
va
l
u
e i
n
a
n
i
m
age (Im
age 1
8
,
R
1
= 3
3
,
S
1
=
5,
R
2
=
1
9
8
,
S
2
=
2
2
6
)
Acco
r
d
i
n
g t
o
Fi
gu
re 4
usi
n
g
cont
ra
st
st
ret
c
hi
n
g
m
e
t
hod t
h
ere a
r
e t
h
ree
areas st
ret
c
hi
n
g
nam
e
l
y
A1
(first
a
r
ea), A2
(second
area
) and A3
(thi
rd area). The
a
r
ea
is an area stret
c
hing scaling
whic
h is base
d on t
h
e
basi
c eq
uat
i
o
n
cont
rast
st
ret
c
h
i
ng.
I
n
m
a
t
h
em
at
i
c
s, t
h
e
are
a
shows
the
va
lue of a strai
g
ht line gra
d
ient. If t
h
e
g
r
ad
ien
t
of a straigh
t
lin
e g
e
neratin
g
v
a
lu
e <1
th
en
th
e
n
e
w
p
i
x
e
l area will b
e
n
a
rro
w
ed
area o
f
p
i
xels, and
th
e
merg
er in
ten
s
ity v
a
lu
es of th
e o
r
ig
i
n
al p
i
x
e
ls. In
stead
grad
ien
t
a straig
h
t
l
i
n
e
> 1
th
ere will b
e
a wid
e
n
i
n
g
o
r
stretch
i
ng
t
h
e
p
i
x
e
l area. Imp
r
ov
ing
im
ag
e qu
ality u
s
ing Con
t
rast St
retch
i
ng
m
e
th
o
d
d
o
e
s
n
o
t
ch
ang
e
t
h
e
v
a
lu
e
of th
e in
t
e
n
s
ity o
f
an imag
e.
B
a
sed o
n
t
h
e
fo
rm
ul
at
i
on of t
h
e
pr
o
b
l
e
m
t
h
at
we
mad
e
prev
i
o
usly ab
ou
t op
timi
zatio
n
of the
g
r
ayscale im
ag
e cau
sing
so
m
e
li
mitat
i
o
n
s
in th
e research
prob
lem
.
It sh
o
w
ed
in
th
e
research
is still fo
cused
o
n
a
n
a
rrow sco
p
e
and
th
ere
are still
m
a
n
y
sh
ortco
m
in
g
s
.
Esp
ecially o
n
t
h
e grayscale valu
es R1
, S1
,
R2
and
S2 w
h
i
c
h
has
no
re
fere
nce
v
a
l
u
es. I
n
ot
he
r wo
rd
s,
t
h
er
e
are no
param
e
ter
s
or
refe
rence t
o
whet
her the
value
of a
n
est
i
m
a
t
e
d i
n
put
by
t
h
e
user
ge
nerat
i
n
g t
h
e pa
n
o
ram
i
c im
age out
p
u
t
m
o
re cl
earl
y
or n
o
t
.
B
a
sed
o
n
t
h
es
e
p
r
ob
lem
s
, we are testin
g th
e
h
ypo
th
esis i
n
t
h
e
form
o
f
'seco
nd
op
in
ion
'
to th
e
d
e
n
t
al
h
o
sp
ital Purwok
erto
b
y
attaching the data of the original im
age and the new im
age of the result
s of our
r
e
s
e
arc
h
.
In
p
r
a
c
tic
e,
th
e
radi
ological
merely com
p
are and c
h
oose the
one im
age of
th
e orig
in
al im
a
g
e with
a
ne
w im
age. Therefore
,
it
is ex
p
ected
throug
h
h
ypo
th
esis testin
g
in
th
e fo
rm
o
f
'seco
nd
o
p
i
n
i
on
' th
at can
ju
stify th
e resu
lts of ou
r
t
e
st
i
ng. Hy
pot
hesi
s t
e
st
i
ng ba
sed o
n
S
ugi
y
o
no
[2
6]
i
s
com
p
ri
se
d o
f
t
w
o t
y
pes:
(1
) The
n
u
l
l
hy
pot
hesi
s
(H
0) i
s
th
e h
ypo
th
esis th
at a
lack
o
f
co
rrelation
b
e
tween
th
e ind
e
p
e
nd
en
t v
a
riable (X) and
th
e d
e
p
e
n
d
e
n
t
v
a
riab
le
(Y
). (
2
) T
h
e w
o
r
k
i
n
g hy
pot
he
si
s (H1
)
i
s
t
h
e hy
p
o
t
h
esi
s
t
h
at
t
h
e rel
a
t
i
onshi
p bet
w
ee
n t
h
e i
nde
pe
nde
nt
va
r
i
abl
e
(X) an
d
t
h
e
d
e
p
e
nd
en
t v
a
riable (Y) being
inv
e
stig
ated. Th
e calcu
latio
n
resu
lts H1
,
will be u
s
ed
as t
h
e
basis o
f
researc
h
data search. Based
on the
hypothes
is test re
ports, it
turns out H1
(worki
ng hypothesis
)
"acceptable"
i.e. "There is a
signi
ficant e
f
fect of t
h
e c
ont
rast stretc
hing
m
e
thods
on the quality of the
pa
noram
i
c image".
It
has
bee
n
pr
o
v
e
n
by
t
h
e
fact
f
r
o
m
t
h
e n
u
m
b
er o
f
new
i
m
ag
es that
have
been m
o
re th
an
th
e
o
r
i
g
in
al im
ag
e th
at
are: (i) the original im
age of a
sel
ected num
ber of 35 pieces
. (ii) The
new
i
m
age of a sele
cted num
b
er of 183
pieces.
It m
eans is 83.9% s
u
c
cessful im
age.
While
16.1
%
declared
unfit image accord
i
n
g to t
h
e ra
diol
ogy. It
is cau
sed
b
y
some fo
llo
wi
n
g
reason
s: (1
) there is a lin
gu
istic in
fo
rm
atio
n
fro
m
sev
e
ral p
a
n
o
ram
i
c i
m
ag
es are
reduce
d.
Whereas in the
m
e
dical worl
d, th
e in
form
at
io
n
is really
i
m
p
o
r
tan
t
, as a referen
ce to
d
i
agno
se th
e
p
a
tien
t
's
co
nd
itio
n
.
(2
) Visu
al
ly,
th
e
boundary area of an a
r
ea of
seve
ral
panoram
i
c images are less clea
r and
une
quivocal. T
h
is is
because t
h
e c
ont
rast is t
o
o hi
gh
thus a
f
fecting bounda
r
y or
a
ffirm
ation of a
n
a
r
ea.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Po
in
t Pro
cessin
g
Method
fo
r
Imp
r
o
v
ing
Denta
l
Rad
i
o
l
o
g
y
Imag
e
Qu
a
lity (Retn
o
Sup
riyan
ti)
1
593
4.
CO
NCL
USI
O
N
B
a
sed
on t
h
e
d
e
pl
oy
m
e
nt
of
p
i
xel
s
, hi
st
og
ra
m
panoram
i
c
im
age can
be
gr
ou
pe
d i
n
t
o
t
h
re
e gr
o
ups;
(i
)
fo
r t
h
e t
y
pe
o
f
bri
ght
hi
st
o
g
ra
m
,
t
h
e spread
of
gray
val
u
es
dom
i
n
at
ed i
n
t
h
e ri
ght
si
de. T
h
ere
f
o
r
e t
h
e
va
l
u
e o
f
S1
is sm
al
ler th
an
R
1
. (ii) For th
e typ
e
of d
a
rk
h
i
stog
ram
,
t
h
e spread
of gray v
a
lu
es do
m
i
n
a
ted
in
t
h
e left sid
e
.
Th
erefo
r
e t
h
e
v
a
lu
e
of S1
is
g
r
eater th
an
R
1
.
(iii) Fo
r th
e
typ
e
of ev
en
ly
d
a
rk
and
b
r
igh
t
h
i
stog
ram
,
th
e spread
o
f
g
r
ay valu
es ev
en
ly fro
m
l
e
ft to
righ
t, wh
ere the in
pu
t
v
a
lu
e is alm
o
st eq
u
a
l to
th
at d
a
rk
h
i
stog
ram S1
great
er
t
h
a
n
R
1
.
H
o
we
ve
r, t
h
e ra
nge
bet
w
ee
n R
1
t
o
S
1
i
s
g
r
eat
er t
h
an
t
h
e
dar
k
hi
st
o
g
ram
.
R
2
val
u
e
m
u
st
be
greater tha
n
S2 because a refe
rence
poi
nt on
the left, ther
efore the stretc
h shoul
d
lead
to right, it applies bot
h
to
d
a
rk,
b
r
i
g
h
t
o
r
ev
en
ly d
a
rk
an
d brigh
t
.
After im
p
r
o
v
i
ng
im
ag
e q
u
a
lity, abo
u
t
8
sam
p
les rep
r
esen
ti
n
g
218
i
m
ag
es, in
d
i
-catin
g
a sig
n
i
fican
t
i
m
p
r
ov
em
e
n
t to
ward
s
a better i
m
ag
e q
u
a
lity. It
is b
a
sed
on
a b
e
tter visu
al
and
ne
w
i
m
ag
e hi
st
o
g
r
a
m
has bee
n
e
v
e
n
l
y
sprea
d
.
I
f
a
n
a
r
ea o
r
a
g
r
adi
e
n
t
of
a st
rai
g
ht
l
i
n
e g
e
ne
rat
i
n
g
val
u
e
sm
a
ller th
an
1
th
en
th
e
n
e
w
pix
e
l area will b
e
n
a
rrowed
an
d
th
e m
e
rg
er inten
s
ity v
a
lu
es o
f
th
e
o
r
i
g
in
al
p
i
x
e
ls.
Co
nv
ersely, if an
area or g
r
adien
t
a straig
h
t
lin
e g
r
eater
th
an
1
th
ere will b
e
a wid
e
n
i
ng
or stretch
i
ng
th
e p
i
x
e
l
area.
Im
provi
ng im
age quality using Co
ntrast
Stretching m
e
thod does not
c
h
a
nge
int
e
nsity values
of a
n
im
age. Acc
o
r
d
i
ng t
o
t
h
e
sec
o
nd
o
p
i
n
i
o
n
f
r
o
m
Dent
al
Ho
sp
ital Jend
eral
So
ed
irm
a
n
Univ
ersity, p
e
rforman
ce
of
o
u
r
m
e
t
hod
i
s
8
3
.
9
%.
Whi
l
e ab
out
1
6
.
1
%
fai
l
e
d
due
t
o
l
i
ng
ui
st
i
c
i
n
f
o
rm
at
i
on a
n
d
vi
s
u
a
l
bo
u
nda
ry
are
a
.
ACKNOWLE
DGE
M
ENTS
We
wou
l
d lik
e to
t
h
ank
s
v
e
ry
m
u
ch
t
o
Den
t
al Ho
sp
ital Jen
d
e
ral So
ed
irman
Un
iv
ersity Puwok
e
rto
fo
r pa
n
o
ram
i
c
im
age data. Th
is researc
h
is s
u
p
p
o
rte
d
b
y
Dan
a
BLU
Jenderal So
ed
irm
a
n
Un
iv
ersity th
roug
h
Hi
ba
h Un
g
gul
a
n
Uns
o
e
d
.
REFERE
NC
ES
[1]
D. J. I
a
nnucci
an
d R. M
.
L. J
.
Ho
werton, “Dental
Radiograph
y
,” 4
t
h
ed
. United
States,
Elsevier, 20
12.
[2]
H. B. Widodo,
et al.
, “Calculating Contrast Str
e
tch
i
ng Variab
les in
Order to Improve Dental
Radiolog
y
Imag
e
Qualit
y,
” in
In
ternational Conference on
E
ngineering and Techn
o
logy
for Sus
tainable Developm
ent (
I
CET4SD)
,
2015.
[3]
B. Molander,
et
al
, “Image Quality
in Panoramic
Radiograph
y
,”
Dentomaxillofac Radiol.
, vol/issue: 24(1), pp. 17–
22, 1995
.
[4]
V. Rushton,
et al.
, “Screen
ing panoramic radiolo
g
y
of adults in
general dental practice:
Rad
i
ological findings
,”
Br
.
Dent.
J.
, vo
l/issu
e: 190(9)
, pp
. 49
5–501, 1995
.
[5]
J. G. Chaffin,
et al.
, “Valid
ity
o
f
using a panoramic radiograph
for in
iti
al den
t
al
clas
s
i
fic
a
tion of
arm
y
recru
its
,
”
Mil.
Med
.
, vol/issue: 169(5), pp
.
368–372, 2004
.
[6]
A. Tak
a
hashi,
et al.
, “Localizing
the mandibu
lar
canal on den
t
al
CT
reformatted
images: Usefuln
e
ss of panoramic
views
,
”
Surg
.
Ra
diol.
Anat.
, vol/issue: 35(9), pp.
803–809, 2013
.
[7]
A. S. Suomalainen,
et al.
, “
D
entom
a
xillof
aci
al
im
aging with panoram
ic vi
ews and cone bea
m
CT.,”
Insights
Imaging
, vol/issue: 6(1)
, pp
. 1–1
6, 2015
.
[8]
Ö.
S.
Sez
g
in,
et al.
, “Comparativ
e dosimetr
y
of
d
e
ntal con
e
b
eam
computed
tomograph
y
, p
a
noramic radiogr
aph
y
,
and m
u
ltisl
i
ce
c
o
m
puted tom
ograph
y
,”
Oral Rad
i
ol.
, vol/issue: 2
8
(1), pp
. 32–37
,
2012.
[9]
E. T
.
Ert
a
s,
et al.
, “Incid
ental findings of carotid ar
ter
y
stenos
is detec
t
ed
b
y
ca
lc
ific
atio
ns on panorami
c
radiographs: Report of
three
cas
es.,”
Oral Radio
l
.
, vol/issue: 26(
2), pp
. 116–121
, 2010.
[10]
M. A. Rotondi
and A. Donner, “A c
onfidence interval appro
a
ch to samp
le size estimation f
o
r interobserv
e
r
agreem
ent
studi
e
s
with m
u
ltip
le
r
a
ters
and ou
tco
m
es,”
J
.
C
lin.
E
p
idemiol
.
, vol/issue: 65(7), pp
. 7
78–784, 2012
.
[11]
K.
Pa
n-Fu,
et al.
, “Fully
Automatic Abdominal
Fat Segmentati
o
n
S
y
stem from a Low Resolu
tio
n CT Image,”
J.
Comput.
, vo
l/iss
u
e: 26(2)
, pp
. 64
–77, 2015
.
[12]
H.
H.
Chang,
et al.
, “A High
Pay
l
o
a
d Stegan
ograph
y
Schem
e
fo
r Color Images Based on
BTC and Hy
br
id
Stra
te
gy
,”
J. Co
mput.
, vol/issue: 26(2), pp. 46–5
5, 2015
.
[13]
S. Gopinathan,
et al.
, “Wavelet and FFT Based I
m
age Denoi
sing Using Non-Lin
ear Filters,”
Int
.
J. El
ec
, vol/issue
:
5(5), 2015
.
[14]
B. A. Far
i
dah
a
nd B. L
i
st
yana
S., “
L
ip
im
age
featur
e ex
tra
c
tio
n util
izing
snak
e’s control
poin
t
s for lip
read
in
g
applications,”
In
t.
J. Electr. Com
put.
Eng.
, vo
l/issue: 5(4)
, pp
. 720
–728, 2015
.
[15]
M.
I.
Ouloul,
et al.
, “An Efficient Face R
ecog
n
ition Using
SI
FT Descriptor
in RGB-D Images,”
In
t. J.
El
ec
tr
.
Comput. Eng
.
, v
o
l/issue: 5
(
6), pp
. 1227–1233
, 20
15.
[16]
G. Sulong, “Se
g
mentation of Fi
ngerprin
t
Image Based on Gradie
nt Magnitud
e
and Coherence,”
Int. J.
E
l
ec
tr
.
Comput. Eng
.
, v
o
l/issue: 5
(
5), pp
. 1202–1215
, 20
15.
[17]
R.
Supriy
a
n
ti
,
et al.
, “A Simple Screening for
High-Risk Pregnanc
ie
s in Rura
l Area
s Ba
se
d Expe
rt S
y
ste
m
,
”
Telkomnika
, vo
l/issue: 13(2), pp.
661–669, 2015
.
[18]
R. Supriy
an
ti,
et
al
., “Separab
ili
t
y
f
ilt
er for
locali
zing
abnorm
a
l p
upil: Id
entif
icati
on of input
im
ag
e,”
Telkomnika
,
vol/issue: 11(4), pp.
783–790
,
20
13.
[19]
R. Supriy
anti,
et al.
, “Comparing edge detection
methods to lo
calize uterus area on ultrasound image,” in
Proc. of
2013 3rd Int. C
onf. on Instrumentation
,
Communications
, Info
rmation Technol., and Biomed
ical Engin
eering
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 4
,
Au
gu
st 2
016
:
15
87
–
1
594
1
594
Scien
c
e and Technol.
for Improvement o
f
He
a
lth, Safety, and
Environ., ICICI-
BM
E 2013
, pp. 152
–155, 2013
.
[20]
R.
Supriy
a
n
ti
,
et al.
, “Extracting
appear
ance infor
m
ation insi
de th
e pupil for catar
act screen
ing,” in
Proceedings o
f
the 11
th I
APR C
onference on
Ma
chine Vision
Ap
plications
, pp
. 3
42–345, 2009
.
[21]
R. Supriy
anti,
et al.
, “A simple and robust method to screen catar
acts using specular reflectio
n appearan
ce,” in
Proc. S
P
IE 6915
Medical Imagin
g
, 2008
.
[22]
R.
Supriy
a
n
ti
,
et al.
, “
C
om
pa
ct Ca
tar
act S
c
r
eening S
y
s
t
em
:
Design and
Practi
cal
dat
a
Acquisition
,
” in
International C
onference on Instrumentation, Co
mmunication, Information Technology and Biomedica
l
Engineering (
I
CICI-BME)
, 2009.
[23]
B. W
illkinson
a
nd M. Allen
,
“
P
aral
lel Progr
am
m
i
ng: Tech
n
i
qu
es and Appli
cat
i
on Usi
ng Networked W
o
rkstatio
ns
and Parallel Co
mputers,” Secon
d
Edi. Ne
w J
e
rsey
: Pearson Pr
entice Hall, 2005
.
[24]
A. Kadir
and
A.
Soesanto, “Teori da
n Aplikasi Pengolahan
Citr
a,” Yog
y
akar
ta Ind
onesia: Penerbit
Andi, 2013
.
[25]
R. Fisher,
et al.
, “
H
istogram
Equali
z
a
tion
,
”
200
3. [Online]
. Available:
http://homepages.inf.ed
.
ac.uk
/
rbf
/
HIPR2/histeq.htm.
BIOGRAP
HI
ES OF
AUTH
ORS
Re
tno Supr
iy
anti
is
an a
cad
em
ic
s
t
aff at
El
ectr
i
c
a
l Eng
i
neer
ing
Departm
e
nt,J
end
e
ral S
o
ed
irm
a
n
Universit
y
, Indo
nesia. She recei
ved her PhD in
March 2010 fro
m
Nara Institut
e
of Scien
ce
and
Techno
log
y
J
a
p
a
n. Also, she receiv
e
d her M.
S degree
and Bachelor degr
ee in
2001 and 1998,
respectively
,
fro
m Electrical En
gineer
ing Depar
t
ment, Gadjah
Mada Univ
ersity
Indonesia. Her
research
in
terests includ
e imag
e processing, co
mputer vision,
pattern
recognition, biomedical
appli
cat
ion,
e-h
e
alth
, t
e
l
e
-hea
lth
and t
e
lem
e
d
i
cin
e
.
Ar
ie
p Soe
l
aiman Se
tiadi
receiv
ed his Bachelor
degree from
Electrical Engin
eering Depratment,
Jenderal Soedir
man University
I
ndone
sia. His research
interest
is
Decision
Support S
y
stem field
.
Yogi Ramadhani
is
an
ac
a
d
em
ic s
t
aff
a
t
El
ectr
i
c
a
l
En
gineer
ing Dep
a
rtm
e
nt, J
e
nd
era
l
Soedirman University
, Indon
esia. He receiv
ed hi
s MS
Gadjah Mada Univ
ersir
t
I
ndonesia,
and
his Bachelor d
e
gree from Jen
d
eral Soedirman Un
iversity
In
donesia. His research in
terest
including
Computer Network, D
ecision
Support
S
y
etem,
Telemedicin
e
and
Medical imaging
Har
i
s Budi Widodo
is
an aca
dem
i
c s
t
aff a
t
P
ublic Heal
th
Departm
e
nt, J
e
n
d
eral S
o
edi
r
m
a
n
University
, Indo
nesia. He received his Ph.D fr
om Airlangga University
Indo
nesia. Also He
received his M.S degree and bachelor degree
from Gadjah Mada Univ
ersity
Indonesia.
His
research
in
terest
including
public
health, e-health
and telemed
i
cin
e
Evaluation Warning : The document was created with Spire.PDF for Python.