T
E
L
K
O
M
N
I
K
A
T
elec
o
m
m
un
ica
t
io
n,
Co
m
pu
t
ing
,
E
lect
ro
nics
a
nd
Co
ntr
o
l
Vo
l.
19
,
No
.
2
,
A
p
r
il
2
0
2
1
,
p
p
.
50
7
~
51
4
I
SS
N:
1693
-
6
9
3
0
,
ac
cr
ed
ited
First Gr
ad
e
b
y
Kem
en
r
is
tek
d
i
k
ti,
Dec
r
ee
No
: 2
1
/E/KPT
/2
0
1
8
DOI
: 1
0
.
1
2
9
2
8
/TE
L
KOM
NI
K
A.
v
1
9
i2
.
1
8
3
2
0
507
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//jo
u
r
n
a
l.u
a
d
.
a
c.
id
/in
d
ex
.
p
h
p
/TELK
OM
N
I
K
A
Autom
a
tic
h
uma
n ear de
tect
io
n
appro
a
ch using
mo
d
ified
a
da
ptive sea
rch w
indo
w t
ec
hnique
Ra
a
d Ahm
ed
H
a
di
1
,
L
o
a
y
E
dwa
r
G
eo
rg
e
2
,
Z
a
ina
b J
a
wa
d Ahm
ed
3
1
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
Co
ll
e
g
e
o
f
E
d
u
c
a
ti
o
n
,
Al
-
Ira
q
ia
Un
iv
e
rsit
y
,
Ba
g
h
d
a
d
,
Ira
q
2
De
p
a
rtme
n
t
o
f
Re
m
o
te S
e
n
si
n
g
a
n
d
G
e
o
g
ra
p
h
ic In
f
o
rm
a
ti
o
n
S
y
ste
m
s,
Un
iv
e
rsity
o
f
Ba
g
h
d
a
d
,
Ba
g
h
d
a
d
,
Ira
q
3
De
p
a
rtme
n
t
o
f
Bi
o
lo
g
y
S
c
ien
c
e
,
Co
ll
e
g
e
o
f
S
c
ien
c
e
,
Un
i
v
e
rsity
o
f
Ba
g
h
d
a
d
,
Ba
g
h
d
a
d
,
Ira
q
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l 3
1
,
2
0
2
0
R
ev
is
ed
Oct
1
,
2
0
2
0
Acc
ep
ted
Oct
2
1
,
2
0
2
0
Th
e
h
u
m
a
n
e
a
r
b
io
m
e
tri
c
re
c
o
g
n
i
ti
o
n
p
lay
s
a
n
imp
o
rtan
t
r
o
le
in
th
e
fo
re
n
sic
s
sp
e
c
ialty
a
n
d
h
a
s
sig
n
ifi
c
a
n
t
imp
a
c
t
fo
r
b
i
o
m
e
tri
c
ian
sc
ien
ti
sts
a
n
d
r
e
se
a
rc
h
e
rs
.
Ac
tu
a
ll
y
,
m
a
n
y
e
a
r
re
c
o
g
n
it
io
n
re
se
a
rc
h
e
s sh
o
we
d
p
ro
m
ise
d
re
su
lt
s
,
b
u
t
so
m
e
issu
e
s
su
c
h
a
s
m
a
n
u
a
l
d
e
tec
ti
o
n
p
ro
c
e
ss
,
e
fficie
n
c
y
a
n
d
ro
b
u
st
n
e
ss
a
re
n
’t
a
tt
a
in
e
d
a
c
e
rtain
lev
e
l
o
f
m
a
tu
ri
ty
.
T
h
e
re
fo
re
,
t
h
e
e
n
h
a
n
c
e
m
e
n
t
d
e
v
e
lo
p
i
n
g
a
p
p
ro
a
c
h
e
s
stil
l
c
o
n
ti
n
u
o
u
s
to
a
c
h
iev
e
li
m
it
e
d
su
c
c
e
ss
e
s.
We
p
ro
p
o
se
a
n
e
fficie
n
t,
re
li
a
b
le
a
n
d
sim
p
le
a
u
t
o
m
a
ti
c
h
u
m
a
n
e
a
r
d
e
te
c
ti
o
n
a
p
p
r
o
a
c
h
.
Th
is
a
p
p
ro
a
c
h
imp
lem
e
n
t
tw
o
sta
g
e
s:
p
re
p
ro
c
e
ss
in
g
a
n
d
e
a
r
lan
d
m
a
rk
s
d
e
tec
ti
o
n
.
We
u
ti
li
z
e
d
th
e
ima
g
e
c
o
n
tras
t,
Lap
lac
e
fil
ter
a
n
d
G
a
u
ss
ian
b
l
u
rrin
g
t
e
c
h
n
iq
u
e
s
to
m
a
d
e
e
n
h
a
n
c
e
m
e
n
t
o
n
a
ll
ima
g
e
s
(in
c
re
a
sin
g
th
e
c
o
n
tras
t,
re
d
u
c
e
th
e
n
o
isy
a
n
d
sm
o
o
t
h
in
g
p
r
o
c
e
ss
e
s).
Afte
r
th
a
t,
we
h
ig
h
li
g
h
ted
t
h
e
e
a
r
e
d
g
e
s
b
y
u
sin
g
th
e
S
o
b
e
l
e
d
g
e
d
e
tec
to
r
a
n
d
d
e
ter
m
in
in
g
th
e
o
n
l
y
wh
i
te p
i
x
e
ls o
f
e
a
r
e
d
g
e
s b
y
a
p
p
ly
in
g
th
e
ima
g
e
su
b
sta
ti
o
n
m
e
t
h
o
d
.
T
h
e
imp
ro
v
e
m
e
n
t
fo
c
u
se
d
o
n
u
sin
g
t
h
e
m
o
d
ifi
e
d
a
d
a
p
ti
v
e
se
a
rc
h
wi
n
d
o
w
(A
S
W)
t
o
d
e
tec
t
t
h
e
e
a
r
re
g
io
n
.
F
u
rth
e
rm
o
re
,
o
u
r
a
p
p
r
o
a
c
h
is
tes
ted
o
n
I
n
d
ian
In
st
it
u
te
o
f
Tec
h
n
o
lo
g
y
(
IIT
)
De
lh
i
sta
n
d
a
rd
e
a
r
b
io
m
e
tri
c
p
u
b
l
ic
d
a
tas
e
t.
Ex
p
e
rime
n
tal
re
su
lt
s
p
re
se
n
ted
a
we
ll
a
v
e
ra
g
e
d
e
tec
ti
o
n
ra
te 9
6
%
f
o
r
4
9
3
ima
g
e
sa
m
p
les
fro
m
1
2
5
p
e
rso
n
s a
n
d
c
o
m
p
u
tati
o
n
a
l
ti
m
e
a
lmo
st
≈
0
.
4
8
5
se
c
o
n
d
s
wh
ich
is
e
v
a
lu
a
ted
with
o
th
e
r
p
re
v
io
u
s wo
r
k
s.
K
ey
w
o
r
d
s
:
Ad
a
p
tiv
e
s
ea
r
ch
win
d
o
w
Au
to
m
atic
ea
r
d
etec
tio
n
B
io
m
e
tr
ic
C
o
m
p
u
tatio
n
al
tim
e
E
ar
r
ec
o
g
n
itio
n
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r
th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
R
aa
d
Ah
m
ed
Had
i
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
C
o
lleg
e
o
f
E
d
u
ca
tio
n
Al
-
I
r
aq
ia
Un
iv
er
s
ity
Ad
am
iy
ah
,
B
ag
h
d
ad
,
I
r
aq
E
m
ail:
r
aa
d
.
h
d
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
E
ar
b
io
m
etr
ics
au
th
en
ticatio
n
is
a
f
o
r
m
o
f
s
ec
u
r
ity
w
h
ich
h
a
s
an
im
p
o
r
tan
t
r
o
le
in
th
e
f
o
r
e
n
s
ics
f
ield
to
d
ef
in
e
th
e
u
n
iq
u
e
p
h
y
s
ical
ch
ar
ac
ter
is
tics
f
o
r
d
is
tin
g
u
is
h
in
g
th
e
p
eo
p
le’
s
id
e
n
tity
,
f
o
r
id
en
tific
atio
n
o
r
v
er
if
icatio
n
task
[
1
-
4
]
.
No
w
ad
ay
s
,
a
m
ajo
r
ity
o
f
t
h
e
we
ll
-
k
n
o
wn
ea
r
b
io
m
et
r
ic
ap
p
r
o
ac
h
es
f
o
cu
s
ed
o
n
r
ec
o
g
n
izin
g
o
f
cr
o
p
p
ed
ea
r
m
an
u
ally
with
o
u
t
u
s
in
g
an
au
t
o
m
atic
ea
r
d
etec
tio
n
an
d
s
eg
m
en
tatio
n
tech
n
iq
u
es.
So
,
th
e
task
au
to
m
ated
ea
r
lo
c
al
izatio
n
r
ep
r
esen
ts
a
b
ig
ch
allen
g
e
f
o
r
m
a
n
y
r
esear
ch
e
r
s
wh
o
s
e
m
ain
co
n
ce
r
n
is
in
h
u
m
an
ea
r
b
io
s
tatis
tics
s
tu
d
ies.
I
n
ad
d
itio
n
,
is
s
u
es
s
u
ch
as
ef
f
icien
c
y
an
d
r
o
b
u
s
tn
ess
ar
e
s
till
n
ee
d
f
o
r
im
p
r
o
v
em
e
n
ts
b
y
b
i
o
m
etr
ici
an
s
cien
tis
ts
an
d
r
esear
ch
er
s
an
d
n
ee
d
to
m
a
k
e
it
in
h
ar
m
o
n
y
with
ea
ch
o
th
er
[
5
-
7
]
.
I
n
f
ac
t,
f
ew
in
tr
o
d
u
ce
d
tech
n
iq
u
es
th
at
ca
n
b
e
u
s
ed
to
d
is
tin
g
u
is
h
th
e
ea
r
lan
d
m
ar
k
au
to
m
atica
lly
.
Mu
r
u
k
esh
et
a
l.
[
2
]
s
u
g
g
ested
a
n
ew
ea
r
r
ec
o
g
n
itio
n
ap
p
r
o
ac
h
th
at
u
tili
ze
d
co
n
to
u
r
let
tr
an
s
f
o
r
m
an
d
a
p
p
ea
r
an
c
e
s
h
ap
e
m
o
d
e
l
(
ASM)
tec
h
n
iq
u
es
f
o
r
f
ea
tu
r
e
ex
tr
ac
tio
n
,
f
o
llo
wed
b
y
class
if
icatio
n
an
d
ea
r
m
atch
in
g
p
r
o
ce
s
s
es
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6
9
3
0
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
,
Vo
l.
19
,
No
.
2
,
Ap
r
il 2
0
2
1
:
50
7
-
51
4
508
u
s
in
g
f
is
h
er
lin
ea
r
d
is
cr
im
in
an
t
an
aly
s
is
(
FLDA
)
.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
test
ed
o
n
I
n
d
ian
I
n
s
titu
te
o
f
T
ec
h
n
o
lo
g
y
(
IIT
)
Delh
i
ea
r
d
atab
ase
b
y
u
s
in
g
o
f
5
0
ea
r
im
ag
es
f
r
o
m
1
0
p
er
s
o
n
s
an
d
o
wn
ea
r
d
atab
ases
.
T
h
is
ap
p
r
o
ac
h
in
d
icate
d
well
o
v
e
r
a
ll
p
er
f
o
r
m
a
n
ce
with
ac
cu
r
ac
y
as
h
ig
h
as
9
7
%.
I
n
f
ac
t,
t
h
e
tim
e
ev
alu
atio
n
n
o
t
m
en
tio
n
ed
i
n
th
is
ap
p
r
o
ac
h
.
Gh
o
u
alm
i
et
a
l.
p
r
esen
ted
a
n
e
w
ea
r
b
io
m
etr
ics
s
y
s
tem
b
ased
o
n
u
s
in
g
a
n
ar
tific
ial
b
ee
co
l
o
n
y
(
AB
C
)
an
d
s
ca
le
in
v
ar
ian
t
f
ea
tu
r
e
tr
a
n
s
f
o
r
m
(
SIFT
)
tech
n
iq
u
es.
De
f
ac
to
,
h
is
to
g
r
am
eq
u
aliza
tio
n
(
HE
)
an
d
co
n
t
r
ast
lim
ited
ad
ap
tiv
e
h
is
to
g
r
am
eq
u
aliza
tio
n
(
C
L
AHE
)
wer
e
test
ed
in
t
h
is
ap
p
r
o
ac
h
o
n
th
r
ee
s
tan
d
ar
d
ea
r
im
a
g
e
d
atab
ases
(
I
I
T
Delh
i,
USTB
1
an
d
USTB
2
)
.
Mo
r
eo
v
er
,
it
s
h
o
wed
well
r
esu
lts
with
an
a
v
er
ag
e
ac
cu
r
ac
y
o
f
9
7
.
1
5
%
[
4
]
. A
ls
o
,
a
n
ew,
au
to
m
atic
an
d
ef
f
icien
t
ea
r
r
ec
o
g
n
i
tio
n
ap
p
r
o
ac
h
r
ec
o
m
m
en
d
e
d
b
y
Had
i
an
d
Geo
r
g
e.
T
h
is
ap
p
r
o
ac
h
is
b
ased
o
n
u
s
i
n
g
th
e
c
o
lo
r
s
k
in
,
s
o
b
el
e
d
g
e
d
etec
to
r
,
im
ag
e
s
u
b
tr
ac
tio
n
a
n
d
r
e
g
io
n
-
g
r
o
win
g
tech
n
iq
u
es.
T
h
is
ap
p
r
o
ac
h
is
test
ed
o
n
th
e
I
I
T
Delh
i e
ar
im
a
g
e
d
ataset,
an
d
p
r
esen
ted
a
we
ll p
er
f
o
r
m
a
n
ce
with
av
er
ag
e
d
etec
tio
n
r
ate
alm
o
s
t 9
1
.
8
% a
n
d
co
m
p
u
tatio
n
al
tim
e
ar
o
u
n
d
≈
1
.
3
3
s
ec
o
n
d
s
in
ter
m
s
o
f
ef
f
icien
c
y
[
5
]
.
I
n
ter
esti
n
g
ly
,
u
s
in
g
e
n
s
em
b
le
o
f
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
etwo
r
k
(
C
NN)
tech
n
i
q
u
es
f
o
r
a
n
ea
r
lo
ca
lizin
g
s
y
s
tem
is
p
r
o
p
o
s
ed
b
y
Ga
n
ap
ath
i
et
a
l.
[
7]
.
I
n
th
is
s
y
s
tem
,
t
h
r
ee
m
o
d
els
o
f
C
NN
tr
ain
ed
t
h
e
g
iv
e
n
d
ataset.
I
n
f
ac
t,
th
ese
m
o
d
els
p
r
o
v
ed
b
ette
r
p
er
f
o
r
m
an
ce
in
ca
s
e
u
s
ed
to
g
eth
er
.
T
h
e
p
r
o
p
o
s
ed
ea
r
lo
ca
liz
in
g
s
y
s
tem
is
test
ed
o
n
two
d
atab
ases
,
I
I
T
i
n
d
o
r
e
-
co
llectio
n
A
(
I
I
T
-
C
o
l
A)
d
atab
ase
an
d
a
n
n
o
tated
web
ea
r
(
AW
E
)
d
atab
ase
with
th
e
ex
is
ten
ce
o
f
p
o
s
e
v
ar
iati
o
n
s
,
o
cc
lu
s
io
n
an
d
illu
m
i
n
atio
n
co
n
d
itio
n
s
m
atter
s
wh
ich
tak
es
o
n
av
er
ag
e
2
.
1
s
ec
o
n
d
s
.
I
n
ad
d
itio
n
,
a
n
in
n
o
v
ativ
e
ea
r
r
ec
o
g
n
itio
n
alg
o
r
ith
m
b
ased
o
n
u
s
in
g
ex
tr
ac
tio
n
o
f
g
eo
m
etr
ica
l
f
ea
tu
r
es su
ch
as
(
s
h
ap
e,
m
ea
n
,
ce
n
tr
o
id
an
d
E
u
clid
ea
n
d
is
tan
ce
b
etwe
en
p
i
x
els)
s
u
g
g
ested
b
y
An
war
et
a
l
.
[
8
]
.
Alth
o
u
g
h
th
e
ex
p
er
im
en
tal
r
e
s
u
lts
s
h
o
wed
th
at
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
g
iv
es
well
o
u
tco
m
es
an
d
ac
h
iev
ed
av
er
ag
e
ac
c
u
r
ac
y
ar
o
u
n
d
9
8
%,
it
is
co
m
p
u
tatio
n
ally
co
m
p
lex
.
Fu
r
th
er
m
o
r
e,
it
r
e
q
u
ests
f
o
r
m
an
u
al
in
itializatio
n
f
o
r
s
u
cc
ess
f
u
l e
x
ec
u
tio
n
o
f
d
et
ec
tio
n
p
r
o
ce
s
s
.
Ad
d
itio
n
ally
,
an
ef
f
icien
t
ea
r
r
ec
o
g
n
itio
n
tech
n
iq
u
e
b
ased
o
n
n
eu
r
al
n
etwo
r
k
s
(
NN)
is
d
e
m
o
n
s
tr
ated
b
y
Z
h
an
g
an
d
M
u
[
9
]
.
T
h
e
au
th
o
r
s
u
tili
ze
d
m
u
ltip
le
s
ca
le
f
aster
r
eg
io
n
-
b
ased
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
etwo
r
k
s
(
f
aster
R
-
C
NN)
as
a
to
o
l
to
d
etec
t
th
e
2
D
ea
r
r
eg
i
o
n
f
r
o
m
th
e
p
r
o
f
ile
im
ag
e
au
to
m
ati
ca
lly
.
T
h
is
p
r
o
p
o
s
ed
tech
n
iq
u
e
is
test
ed
o
n
a
s
et
o
f
2
0
0
we
b
im
ag
es
u
n
d
e
r
v
ar
ian
t
p
h
o
t
o
g
r
a
p
h
ic
c
o
n
d
itio
n
s
,
an
d
it
is
ac
h
iev
ed
9
8
%
d
etec
tio
n
r
ate.
L
ik
ewise,
an
a
u
to
m
ated
h
u
m
an
ea
r
id
e
n
tific
atio
n
s
y
s
te
m
p
r
o
p
o
s
ed
b
y
T
ar
iq
an
d
Ak
r
am
.
T
h
is
s
y
s
tem
en
co
m
p
ass
ed
f
r
o
m
t
h
r
ee
s
tag
es:
p
r
ep
r
o
ce
s
s
in
g
,
f
ea
tu
r
es
ex
tr
ac
tio
n
an
d
id
en
ti
f
icatio
n
p
r
o
ce
s
s
es
r
esp
ec
tiv
ely
.
T
h
e
e
x
p
er
im
e
n
ta
l
r
esu
lts
illu
s
tr
ate
an
av
er
ag
e
a
cc
u
r
ac
y
o
f
9
7
.
2
%
an
d
9
5
.
2
%
t
h
at
ar
e
ev
alu
ate
d
o
n
th
e
UST
B
an
d
I
I
T
Delh
i e
ar
im
ag
e
d
atab
ases
r
esp
ec
tiv
ely
[
1
0
]
.
B
en
za
o
u
i
et
a
l.
p
r
o
p
o
s
ed
an
e
ar
d
escr
ip
tio
n
an
d
r
ec
o
g
n
itio
n
th
at
u
s
ed
a
r
o
b
u
s
t
ellip
tical
lo
ca
l
b
in
ar
y
p
atter
n
(
E
L
B
P)
an
d
d
is
cr
ete
wav
elet
tr
an
s
f
o
r
m
(
DW
T
)
t
ec
h
n
iq
u
es
t
o
d
e
p
ict
th
e
ad
eq
u
ate
d
etails
o
f
th
e
two
-
d
im
en
s
io
n
al
ea
r
im
ag
es
[
1
1
]
.
H
o
wev
er
,
t
h
e
ev
alu
atio
n
r
esu
lts
s
h
o
wed
a
s
u
cc
ess
r
ec
o
g
n
itio
n
r
ate
ar
o
u
n
d
9
4
%
wh
en
test
ed
o
n
5
0
0
im
a
g
es
f
r
o
m
1
0
0
p
er
s
o
n
s
f
r
o
m
th
e
I
I
T
Delh
i
d
atab
ase
.
B
esid
es,
an
ef
f
icien
t
o
n
lin
e
ear
-
b
ased
p
e
r
s
o
n
al
id
e
n
tific
atio
n
s
y
s
tem
p
r
esen
ted
b
y
Me
r
ao
u
m
ia
et
a
l.
[
1
2
]
.
I
n
th
is
p
ap
er
,
ea
c
h
ea
r
h
a
d
s
p
ec
if
ic
f
ea
tu
r
es
s
et
wh
ich
is
ex
tr
ac
ted
b
y
u
s
in
g
Gab
o
r
f
ilte
r
.
I
n
o
r
d
er
to
r
ea
lize
an
id
ea
l
m
u
lti
-
r
ep
r
esen
tatio
n
s
y
s
tem
,
th
e
f
u
s
io
n
p
h
ase
is
ap
p
lied
b
y
tr
y
in
g
o
f
s
ev
er
al
co
m
b
in
atio
n
s
o
f
u
s
in
g
th
ese
f
ea
tu
r
es
(
p
h
ase,
m
o
d
u
le
an
d
r
ea
l
(
im
ag
in
ar
y
)
p
ar
ts
m
ix
tu
r
es
)
.
T
h
is
s
y
s
tem
tes
ted
o
n
I
I
T
Delh
i
d
atab
ase
o
f
2
2
1
u
s
er
s
an
d
y
ield
s
a
well
p
er
f
o
r
m
an
ce
o
f
ea
r
id
e
n
tific
atio
n
p
r
o
ce
s
s
.
An
o
th
er
ea
r
d
etec
tio
n
m
o
d
el
s
u
g
g
ested
b
y
[
1
3
]
.
T
h
e
au
th
o
r
s
u
s
ed
two
tech
n
iq
u
es
as
f
o
llo
ws:
s
n
ak
e
-
b
ased
b
ac
k
g
r
o
u
n
d
r
em
o
v
al
(
SB
R
)
an
d
s
n
a
ke
-
b
ased
ea
r
lo
ca
lizatio
n
(
SEL
)
.
C
o
n
v
er
s
ely
,
th
is
m
o
d
el
s
h
o
ws
well
r
esu
lt
b
u
t
it
s
u
f
f
er
s
f
r
o
m
th
e
h
ig
h
co
m
p
u
tatio
n
al
tim
e;
i
t
is
ar
o
u
n
d
3
.
8
6
s
p
er
im
ag
e.
In
[
1
4
]
,
th
e
au
th
o
r
s
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
f
o
r
ea
r
lo
ca
lizatio
n
b
ased
o
n
co
lo
r
(
YC
b
C
r
co
lo
r
s
p
ac
e
)
d
etec
tio
n
an
d
ed
g
e
m
a
p
p
in
g
tech
n
iq
u
es.
I
n
d
ee
d
,
t
h
e
ea
r
d
e
tectio
n
av
er
a
g
e
tim
e
is
ar
o
u
n
d
7
.
9
5
s
wh
ic
h
is
c
o
n
s
u
m
ed
m
o
r
e
r
eso
u
r
ce
s
f
r
o
m
co
m
p
u
tatio
n
al
co
m
p
lex
ity
asp
ec
t.
Alter
n
ativ
e
wo
r
k
was
s
u
g
g
ested
b
y
Ho
u
r
ali
an
d
Gh
ar
r
a
v
i
[
1
5
]
.
T
h
ey
u
s
ed
a
mod
if
ied
f
o
r
m
o
f
d
is
cr
ete
co
s
i
n
e
tr
an
s
f
o
r
m
(
tr
a
n
s
f
o
r
m
e
d
DC
T
)
.
I
t
is
test
ed
o
n
two
d
atasets
USTB
s
u
b
s
et
I
I
an
d
I
I
T
Delh
i
s
u
b
s
et
I
I
a
n
d
ev
al
u
a
ted
with
g
o
o
d
ef
f
icien
c
y
.
Nev
er
th
eless
,
th
is
wo
r
k
ass
u
m
ed
th
at
th
e
ea
r
r
e
g
io
n
is
cr
o
p
p
e
d
m
an
u
ally
b
y
s
p
ec
if
ied
ea
r
d
ete
cto
r
ad
v
a
n
ce
ly
.
So
,
it is
n
’
t a
n
au
to
m
ated
ea
r
lo
ca
li
za
tio
n
s
y
s
tem
.
T
h
e
m
ain
id
ea
o
f
o
u
r
wo
r
k
is
to
p
r
esen
t
a
n
ef
f
icien
t,
r
eliab
le
a
n
d
s
im
p
le
a
u
to
m
atic
h
u
m
an
ea
r
d
etec
tio
n
ap
p
r
o
ac
h
wh
ich
is
b
ased
o
n
m
o
d
if
ied
ASW
.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
wh
ich
is
r
ec
a
p
itu
lated
b
y
two
p
h
ases
p
r
ep
r
o
ce
s
s
in
g
an
d
ea
r
lan
d
m
ar
k
s
d
etec
tio
n
.
A
d
d
itio
n
ally
,
ex
p
er
m
in
tal
test
s
s
h
o
w
p
r
o
m
is
ed
r
esu
lts
o
f
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
.
T
h
is
p
ap
e
r
is
s
tr
u
ctu
r
ed
as
f
o
llo
ws:
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
o
f
ea
r
b
io
m
etr
ic
d
etec
tio
n
is
p
r
esen
ted
i
n
s
ec
tio
n
2
.
Du
r
in
g
th
e
s
ec
tio
n
3
,
th
e
r
esu
lts
an
d
d
is
s
ec
tio
n
is
illu
s
tr
ated
.
Fin
ally
,
co
n
clu
s
io
n
a
n
d
f
u
tu
r
e
wo
r
k
s
u
g
g
esti
o
n
s
ar
e
s
u
m
m
ar
ized
in
s
ec
tio
n
4
.
2.
T
H
E
P
RO
P
O
SE
D
AP
P
RO
A
CH
O
F
E
AR
B
I
O
M
E
T
R
I
C
DE
T
E
C
T
I
O
N
Her
e,
a
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
o
f
ea
r
b
io
m
etr
ic
d
etec
tio
n
is
illu
s
tr
ated
th
at
is
e
x
tr
ac
ted
ef
f
ic
ien
tly
an
d
s
im
p
lify
in
i
m
p
lem
en
tatio
n
t
h
e
ea
r
’
s
lan
d
m
a
r
k
s
in
f
o
r
m
ati
o
n
as
s
h
o
wn
in
Fig
u
r
e
1
.
M
o
r
e
s
p
ec
if
ically
,
th
is
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
A
u
to
ma
tic
h
u
ma
n
ea
r
d
etec
tio
n
a
p
p
r
o
a
ch
u
s
in
g
… (
R
a
a
d
A
h
med
Ha
d
i
)
509
ap
p
r
o
ac
h
im
p
lem
e
n
ts
two
s
tag
es
as
f
o
llo
w:
f
ir
s
tly
,
we
m
a
d
e
a
p
r
ep
r
o
ce
s
s
in
g
f
o
r
im
ag
e
en
h
an
ce
m
e
n
t;
th
r
ee
o
p
er
atio
n
s
ar
e
u
s
ed
f
o
r
in
cr
e
asin
g
th
e
co
n
tr
ast
(
s
tr
etch
in
g
)
,
r
ed
u
ce
o
r
b
lu
r
t
h
e
n
o
is
y
(
Gau
s
s
ian
b
lu
r
)
an
d
s
m
o
o
th
in
g
(
lap
lace
f
ilter
)
o
f
al
l e
ar
im
ag
es.
Seco
n
d
ly
,
we
ap
p
lied
a
So
b
el
ed
g
e
d
etec
to
r
tec
h
n
iq
u
e
to
h
i
g
h
lig
h
t
th
e
ea
r
lan
d
m
ar
k
ed
g
es
an
d
y
i
eld
a
b
in
ar
y
m
ask
im
ag
e.
Af
t
er
th
at,
we
u
s
ed
th
e
im
ag
e
s
u
b
tr
ac
tio
n
p
r
o
ce
s
s
to
d
eter
m
in
e
th
e
o
n
l
y
wh
ite
p
ix
el
s
o
f
ea
r
ed
g
es.
I
n
f
ac
t,
we
s
u
b
t
r
ac
ted
th
e
r
esu
ltan
t
im
ag
e
o
f
S
o
b
el
ed
g
e
d
etec
tio
n
p
r
o
ce
s
s
f
r
o
m
th
e
r
esu
ltan
t
im
ag
e
o
f
s
m
o
o
t
h
in
g
p
r
o
ce
s
s
b
y
L
ap
lace
f
ilter
.
F
in
ally
,
th
e
n
ew
d
etec
tio
n
tech
n
iq
u
e
wh
ich
is
in
s
p
ir
e
d
f
r
o
m
[
1
6
,
1
7
]
a
p
p
lied
to
d
etec
t
th
e
ea
r
l
an
d
m
ar
k
s
r
eg
io
n
.
T
h
is
n
ew
t
ec
h
n
iq
u
e
u
s
ed
f
o
u
r
d
etec
to
r
s
to
s
ca
n
,
co
llect
wh
it
e
p
ix
els
o
f
ea
r
ed
g
es
a
n
d
r
ec
o
g
n
ized
it.
Du
r
in
g
th
e
test
in
g
p
r
o
ce
s
s
,
we
ap
p
lied
o
u
r
ap
p
r
o
ac
h
o
n
4
9
3
s
am
p
l
e
b
elo
n
g
to
I
I
T
Delh
i
s
tan
d
ar
d
ea
r
b
io
m
etr
ic
p
u
b
lic
d
ataset.
B
e
s
id
es,
two
p
er
f
o
r
m
an
ce
m
ea
s
u
r
es
ar
e
ev
a
lu
ated
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
as
in
d
icato
r
o
f
its
ef
f
icien
cy
as
f
o
llo
w
th
e
ac
c
u
r
ac
y
an
aly
s
is
an
d
co
m
p
u
tatio
n
al
ti
m
e
r
esp
ec
tiv
ely
.
T
h
e
av
er
ag
e
a
cc
u
r
ac
y
o
f
t
h
e
p
r
esen
t
wo
r
k
s
h
o
ws
p
r
o
m
is
ed
r
esu
lt
b
u
t
litt
le
b
it
less
f
r
o
m
s
o
m
e
p
r
ev
io
u
s
wo
r
k
s
,
b
u
t
in
c
o
n
tr
ar
y
th
e
ef
f
icien
cy
o
f
r
ea
l
-
tim
e
d
ep
icts
f
aster
r
esu
lt
in
ea
r
d
etec
tio
n
th
r
o
u
g
h
co
m
p
u
ti
n
g
ea
ch
p
r
o
ce
s
s
p
er
im
a
g
e.
Fig
u
r
e
1
.
L
a
y
o
u
t
o
f
p
r
o
p
o
s
ed
ea
r
d
etec
tio
n
ap
p
r
o
ac
h
2
.
1
.
P
re
pro
ce
s
s
ing
Du
r
in
g
th
e
p
r
ep
r
o
c
ess
in
g
,
th
e
wh
o
le
im
ag
es
r
ea
d
o
n
e
b
y
o
n
e
an
d
co
n
v
er
t
to
g
r
ay
s
ca
le
im
ag
es;
th
e
im
ag
e
co
n
tr
ast
s
tr
etch
in
g
is
a
p
p
lied
to
ad
j
u
s
t
th
e
in
te
n
s
ities
co
n
tr
ast
o
f
i
m
ag
es
f
o
r
attain
i
n
g
m
o
r
e
s
h
ar
p
en
i
n
g
o
f
th
e
ea
r
e
d
g
es.
Af
ter
th
at,
g
au
s
s
ian
b
lu
r
an
d
l
a
p
lace
f
ilter
ar
e
ap
p
lied
to
en
h
a
n
ce
th
e
q
u
ality
o
f
ea
r
im
a
g
es
(
in
cr
ea
s
e
th
e
d
is
p
ar
ity
o
f
ea
r
lan
d
m
ar
k
s
an
d
r
em
o
v
e
th
e
n
o
is
e
o
r
b
lu
r
o
f
im
ag
e
)
.
Her
e
,
o
n
e
o
f
th
e
co
m
m
o
n
im
ag
e
illu
m
in
atio
n
en
h
a
n
ce
m
en
t
tech
n
iq
u
es
is
co
n
tr
ast
s
tr
e
tch
in
g
wh
ich
is
u
s
ed
.
I
t
wo
r
k
s
b
y
s
p
r
ea
d
in
g
th
e
g
r
ay
-
lev
els
(
b
r
ig
h
tn
ess
v
alu
es
)
o
f
th
e
h
an
d
led
im
a
g
e
i
n
to
d
y
n
am
ic
r
an
g
e.
T
h
e
n
,
t
h
e
r
esu
l
tan
t
im
ag
e
will
b
e
g
iv
in
g
m
o
r
e
in
f
o
r
m
ati
o
n
f
o
r
an
aly
s
is
p
r
o
ce
s
s
[
1
8
]
.
Actu
ally
,
ca
lcu
latin
g
th
e
p
ix
el
illu
m
in
atio
n
v
alu
es
ar
e
illu
s
tr
ated
b
y
co
m
p
u
tin
g
o
f
th
e
p
o
wer
in
ter
v
al
th
at
is
s
elec
ted
th
e
m
in
im
u
m
an
d
m
ax
im
u
m
v
alu
es
an
d
s
tr
etch
ed
m
ain
p
o
wer
i
n
ter
v
al
o
f
th
e
h
is
to
g
r
am
to
t
h
e
f
u
ll r
a
n
g
e
(
0
-
2
5
5
)
as sh
o
wn
in
(
2
)
[
1
9
]
:
(
,
)
=
{
255
(
,
)
≥
0
(
,
)
≤
255
∗
(
(
,
)
−
−
)
ℎ
}
(
1
)
wh
er
e
G
(
x
,
y
)
is
th
e
im
ag
e
co
o
r
d
in
atio
n
,
m
ax
a
n
d
m
i
n
ca
lcu
lated
as f
o
llo
ws:
=
−
=
+
(
2
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6
9
3
0
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
,
Vo
l.
19
,
No
.
2
,
Ap
r
il 2
0
2
1
:
50
7
-
51
4
510
wh
er
e
μ
a
n
d
σ
s
y
m
b
o
ls
in
d
ic
ated
th
e
m
ea
n
an
d
s
tan
d
ar
d
d
ev
iatio
n
v
al
u
es
o
f
th
e
im
ag
e,
r
esp
ec
tiv
ely
.
T
h
e
p
ar
am
eter
α
is
ap
p
lied
to
c
o
n
tr
o
l th
e
s
tr
en
g
th
o
f
im
p
lem
e
n
ted
lin
ea
r
ex
ten
t.
Af
ter
th
at,
t
h
e
Gau
s
s
ian
b
lu
r
r
in
g
o
p
er
atio
n
m
ea
n
s
s
im
p
ly
t
o
tr
an
s
f
o
r
m
o
n
e
co
lo
r
v
al
u
e
t
o
th
e
o
th
er
v
er
y
s
m
o
o
th
.
Als
o
,
it
c
o
n
s
id
er
s
o
n
e
o
f
c
o
m
m
o
n
ess
en
tial
o
p
er
atio
n
s
b
ef
o
r
e
m
an
y
task
s
s
u
ch
as
ed
g
e
d
etec
tio
n
in
im
ag
e
p
r
o
ce
s
s
in
g
s
p
ec
ialt
y
.
T
h
e
Gau
s
s
ian
s
m
o
o
th
in
g
av
er
ag
i
n
g
o
p
er
ato
r
o
r
f
ilte
r
r
ep
r
esen
ts
a
2
D
co
n
v
en
tio
n
al
o
p
er
ato
r
th
at
tak
es
th
e
s
h
ap
e
o
f
a
Gau
s
s
ian
(
b
ell
-
s
h
ap
ed
)
h
u
n
ch
,
it
will
r
em
o
v
e
th
e
im
ag
e’
s
n
o
is
e
with
th
e
h
ig
h
s
p
atial
f
r
e
q
u
e
n
cies
an
d
y
iel
d
s
a
s
m
o
o
th
i
n
g
o
u
tco
m
e.
I
n
two
d
im
e
n
s
io
n
s
,
an
is
o
tr
o
p
ic
(
i.e
.
,
cir
cu
lar
ly
s
y
m
m
etr
ic)
Ga
u
s
s
ian
b
lu
r
f
ilter
f
u
n
ctio
n
h
as th
e
f
o
r
m
[
2
0
,
2
1
]
:
(
,
)
=
1
√
2
2
−
(
2
+
2
)
/
2
2
(
3
)
wh
er
e
σ
s
y
m
b
o
l
is
th
e
s
tan
d
ar
d
d
ev
iatio
n
o
f
t
h
e
2
D
d
is
tr
ib
u
tio
n
an
d
it
co
n
tr
o
ls
th
e
d
eg
r
ee
o
f
im
ag
e
s
m
o
o
th
in
g
.
Als
o
,
th
e
L
ap
lace
f
ilter
will
b
e
u
s
ed
in
th
is
p
r
o
p
o
s
ed
ap
p
r
o
a
ch
.
I
t
is
u
s
ed
as
a
m
ea
s
u
r
e
o
f
t
h
e
s
ec
o
n
d
s
p
atial
d
er
iv
ativ
e
o
f
an
im
ag
e
wh
ich
is
a
2
D
id
en
tical
(
is
o
tr
o
p
ic)
d
eg
r
ee
.
Ho
wev
e
r
,
it
is
a
c
o
n
v
en
tio
n
al
o
p
er
ato
r
th
at
h
ig
h
lig
h
ts
th
e
c
h
an
g
i
n
g
o
f
r
ap
id
in
ten
s
ity
ar
ea
s
.
So
,
it
is
o
f
ten
u
s
ed
a
f
ter
s
m
o
o
th
i
n
g
ap
p
r
o
x
im
atin
g
p
r
o
ce
s
s
o
f
an
im
ag
e
b
y
Gau
s
s
i
an
b
lu
r
f
ilter
to
r
ed
u
ce
s
en
s
itiv
ity
to
n
o
is
e
ef
f
ec
t
[
2
2
,
2
3
]
.
T
h
e
f
o
llo
win
g
eq
u
atio
n
d
ep
icts
o
f
an
im
a
g
e
wh
ich
is
d
e
n
o
ted
b
y
L
ap
lacia
n
L
(
x
,
y
)
with
p
i
x
el
in
ten
s
ity
v
alu
es I
(
x
,
y
)
is
ass
u
m
ed
b
y
:
(
,
)
=
2
2
+
2
2
(
4)
Ad
d
itio
n
ally
,
we
co
m
p
ar
e
e
v
e
r
y
p
ix
el
o
f
th
e
r
esu
ltan
t
im
ag
e
(
L
ap
_
im
ag
e(
x
)
)
f
r
o
m
L
ap
la
ce
f
iltra
tio
n
p
r
o
ce
s
s
with
a
r
an
g
e
f
r
o
m
(
7
0
>=
L
a
p
_
i
m
ag
e(
x
)
<=
2
0
0
)
to
g
et
m
o
r
e
w
h
ite
p
ix
els,
an
d
h
ig
h
lig
h
tin
g
th
e
ea
r
ed
g
es
(
b
in
ar
y
m
ask
im
ag
e)
f
o
r
n
e
x
t step
(
So
b
el
ed
g
e
d
etec
tio
n
p
r
o
ce
s
s
).
2
.
2
.
E
a
r
la
nd
ma
rk
s
det
ec
t
io
n
T
h
e
m
ain
id
ea
an
d
co
n
tr
i
b
u
tio
n
o
f
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
is
to
u
tili
ze
n
ew
tech
n
iq
u
e
f
o
r
r
e
co
g
n
itio
n
th
e
ea
r
lan
d
m
ar
k
s
wh
ic
h
is
ad
a
p
ted
b
y
[
1
6
,
1
7
]
.
Af
ter
th
e
p
r
e
p
r
o
ce
s
s
in
g
p
h
ase,
we
h
av
e
g
o
t
an
en
h
a
n
ce
d
im
a
g
e
with
s
h
ar
p
ed
ed
g
e
an
d
n
o
is
eless
o
f
ea
r
lan
d
m
ar
k
s
.
Firstl
y
,
a
S
o
b
el
ed
g
e
d
etec
to
r
a
p
p
lied
o
n
t
h
e
en
h
an
ce
d
im
ag
e
to
d
is
cr
im
in
ate
t
h
e
h
u
m
an
ea
r
ed
g
es.
An
e
d
g
e
is
t
h
e
b
o
u
n
d
a
r
y
b
etwe
en
r
e
g
io
n
s
o
f
two
im
ag
es,
wh
ich
h
as
th
e
d
is
tin
ct
p
r
o
p
er
ties
ac
co
r
d
in
g
to
s
o
m
e
f
ea
tu
r
es
s
u
ch
as
(
g
r
ad
i
en
t,
co
lo
r
,
tex
t
u
r
e
o
r
g
r
ay
lev
e
l)
.
I
t
im
p
lies
a
p
air
o
f
o
r
th
o
g
o
n
al
g
r
a
d
ien
t
o
p
er
ato
r
,
it
wo
r
k
s
to
f
in
d
th
e
ed
g
e
s
tr
en
g
th
a
n
d
d
ir
ec
tio
n
at
l
o
ca
tio
n
(
x
,
y
)
o
f
an
im
ag
e
f
,
it is
ca
lled
th
e
g
r
ad
ie
n
t,
d
e
n
o
ted
b
y
∇
,
an
d
is
d
ef
in
ed
as a
v
ec
to
r
[
5
].
(
,
)
=
(
)
=
[
]
=
(
)
(
5
)
T
h
is
v
ec
to
r
o
f
co
n
tin
u
o
u
s
f
u
n
c
tio
n
f
(
x
,
y
)
h
as
im
p
o
r
tan
t
g
eo
m
etr
ical
p
r
o
p
er
ties
o
f
lo
ca
tio
n
(
x
,
y
)
.
T
h
e
v
alu
e
o
f
ch
an
g
in
g
r
ate
in
th
e
d
ir
ec
tio
n
o
f
g
r
ad
ien
t
v
ec
to
r
is
ca
lled
th
e
m
ag
n
itu
d
e
(
len
g
th
)
o
f
th
e
v
ec
to
r
∇
,
an
d
is
d
en
o
te
d
as M
(
x
,
y
)
,
wh
er
e:
(
,
)
=
(
)
=
√
2
+
2
(
6
)
I
n
ad
d
itio
n
,
th
e
d
ir
ec
tio
n
a
n
g
l
e
o
f
g
r
ad
ien
t v
ec
t
o
r
is
d
ec
lar
e
d
as:
(
,
)
=
−
1
[
]
(
7
)
So
b
el
ed
g
e
d
etec
to
r
is
wo
r
k
e
d
b
y
co
n
v
o
l
u
tio
n
co
n
ce
p
t;
it
u
s
es
a
s
et
o
f
3
x
3
co
n
v
o
lu
tio
n
k
e
r
n
els.
On
e
o
f
th
e
(
Gx
)
k
er
n
els
s
et
is
u
s
ed
to
d
is
tin
g
u
is
h
th
e
b
r
ig
h
tn
ess
s
tr
en
g
th
o
f
e
d
g
es
in
th
e
h
o
r
iz
o
n
tal
d
ir
ec
tio
n
,
an
d
th
e
o
th
er
o
n
e
(
Gy
)
is
u
s
ed
to
d
is
tin
g
u
is
h
th
e
b
r
ig
h
tn
ess
s
tr
en
g
th
o
f
ed
g
es
in
th
e
v
er
tical
d
ir
ec
tio
n
[
5
,
2
3
-
2
6
]
.
Seco
n
d
ly
,
an
a
r
ith
m
etic
o
p
e
r
at
io
n
is
a
p
r
o
ce
s
s
p
er
f
o
r
m
ed
b
et
wee
n
p
ix
el
-
to
-
p
ix
el
is
ca
lled
im
ag
e
s
u
b
tr
ac
tio
n
.
I
t
is
o
f
ten
u
s
ed
to
d
etec
t th
e
d
if
f
er
en
ce
s
b
etwe
en
two
im
ag
es (
e.
g
.
r
em
o
v
in
g
o
f
r
elev
a
n
t c
o
n
t
en
ts
f
r
o
m
th
e
im
ag
e
o
r
d
etec
t
th
e
r
elev
a
n
t
o
b
ject
m
o
tio
n
b
etwe
en
two
f
r
am
es
o
f
a
v
id
eo
s
eq
u
e
n
ce
)
.
T
h
e
r
ef
o
r
e,
t
h
e
r
esu
ltan
t
im
ag
e
f
r
o
m
p
r
ev
io
u
s
p
r
o
ce
s
s
will
b
e
s
u
b
tr
ac
ted
f
r
o
m
th
e
e
n
h
an
ce
d
i
m
ag
e
to
g
et
th
e
f
in
al
b
in
ar
y
m
a
s
k
im
ag
e
with
o
n
l
y
th
e
s
am
e
ed
g
es
(
wh
ite
p
ix
els)
b
etwe
en
t
h
e
two
-
im
ag
e
m
en
ti
o
n
ed
b
ef
o
r
e
.
Giv
e
n
a
2
D
ar
r
a
y
o
f
(
X)
im
ag
e
an
d
an
o
th
er
2
D
ar
r
ay
o
f
(
Y)
im
ag
e
,
th
e
r
esu
ltin
g
is
a
n
ew
ar
r
ay
,
s
ca
lar
(
Z
)
,
is
o
b
tain
e
d
b
y
ca
lcu
l
atin
g
[
5
,
2
4
]
:
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
A
u
to
ma
tic
h
u
ma
n
ea
r
d
etec
tio
n
a
p
p
r
o
a
ch
u
s
in
g
… (
R
a
a
d
A
h
med
Ha
d
i
)
511
−
=
(
8)
Fin
ally
,
a
m
o
d
if
ie
d
tech
n
i
q
u
e
th
at
is
k
in
d
o
f
s
im
ilar
to
[
1
6
,
1
7
]
,
it
will
b
e
u
s
ed
f
o
r
e
f
f
ic
ien
tly
an
d
s
im
p
licity
d
is
tin
g
u
is
h
in
g
th
e
ea
r
lan
d
m
ar
k
f
r
o
m
th
e
f
ac
e
s
id
e
r
eg
io
n
.
Ho
wev
er
,
it
is
a
v
ar
iab
le
-
s
ize
s
ea
r
ch
win
d
o
w
th
at
is
u
s
ed
to
d
etec
t
t
h
e
e
ar
lan
d
m
ar
k
s
r
eg
io
n
b
y
ad
j
u
s
tin
g
th
e
s
ize
s
ca
les
(
h
eig
h
t
a
n
d
wid
th
)
.
Actu
ally
,
it
is
a
f
lex
ib
le
s
ea
r
ch
win
d
o
w
b
y
u
s
in
g
f
o
u
r
d
ir
ec
tio
n
s
(
to
p
,
b
o
tto
m
,
lef
t
an
d
r
i
g
h
t
d
etec
to
r
s
)
f
o
r
lo
ca
tin
g
an
d
d
etec
tio
n
th
e
ea
r
lan
d
m
a
r
k
s
h
e
ig
h
t
an
d
wig
h
t
d
im
en
s
io
n
s
.
Pr
ac
tic
ally
,
th
is
tech
n
iq
u
e
is
ap
p
lied
o
n
th
e
r
esu
ltan
t
lo
w
-
lev
el
b
in
ar
y
im
ag
e
(
p
ix
el
co
lo
u
r
v
alu
es
with
wh
ite
an
d
b
lack
)
f
r
o
m
im
ag
e
s
u
b
tr
ac
tio
n
p
r
o
ce
s
s
,
it’s
s
ca
n
n
ed
th
e
im
ag
e
f
r
o
m
f
o
u
r
d
ir
ec
tio
n
s
to
f
in
d
th
e
wh
ite
p
ix
els
b
y
u
s
in
g
f
lex
i
b
le
d
etec
to
r
s
as sh
o
wn
in
Fig
u
r
e
2
.
T
h
ese
d
etec
to
r
s
in
itialized
b
y
u
s
in
g
th
r
esh
o
ld
s
as
tr
ai
n
in
g
v
alu
es
to
s
ca
n
t
h
e
im
ag
e
p
ix
els
an
d
r
ed
u
ce
th
e
s
ea
r
ch
win
d
o
w
p
r
o
ce
s
s
with
ac
cu
m
u
lato
r
s
to
co
u
n
t
wh
ite
p
ix
els
f
o
r
ea
r
la
n
d
m
ar
k
s
lo
ca
lizatio
n
f
r
o
m
th
e
f
ac
e
s
id
e
r
eg
io
n
.
Fin
ally
,
th
ese
accu
m
u
l
ato
r
s
will
b
e
s
elec
ted
b
ased
o
n
th
e
h
ig
h
e
r
o
n
e
o
f
w
h
ite
p
ix
e
ls
co
u
n
tin
g
v
alu
es
,
an
d
it will d
eter
m
in
e
th
e
co
r
r
d
in
ates o
f
th
e
d
etec
ted
ea
r
r
eg
i
o
n
.
Fig
u
r
e
2
.
Ad
a
p
tiv
e
s
ea
r
ch
win
d
o
w
tec
h
n
iq
u
e
f
o
r
ea
r
d
etec
tio
n
3.
RE
S
U
L
T
S AN
D
D
I
SCU
SS
I
O
N:
E
VALU
AT
I
O
N
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
h
ad
b
ee
n
test
ed
o
n
th
e
I
n
d
ia
n
I
n
s
titu
te
o
f
T
ec
h
n
o
lo
g
y
(
I
I
T
)
Delh
i
ea
r
im
ag
e
d
ataset
an
d
ev
alu
ated
in
ter
m
s
o
f
p
r
o
ce
s
s
in
g
tim
e
an
d
d
etec
tio
n
ac
cu
r
a
cy
.
T
h
e
I
I
T
d
ataset
co
n
s
is
ts
o
f
4
9
3
R
GB
ea
r
im
ag
e
,
wh
ich
was
t
ak
en
f
r
o
m
1
2
5
d
if
f
er
en
t
(
s
u
b
jects)
p
er
s
o
n
s
th
at
a
r
e
o
f
ag
es
b
etwe
en
1
8
an
d
5
8
y
ea
r
s
.
I
n
a
d
d
itio
n
,
th
ese
im
ag
es
h
av
e
a
r
eso
lu
tio
n
o
f
2
7
2
x
2
0
4
p
ix
els
f
o
r
ea
ch
.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
is
ca
r
r
ied
o
u
t
u
s
in
g
Mic
r
o
s
o
f
t
v
is
u
al
C
#
2
0
1
7
s
o
f
twar
e
o
n
a
p
en
tiu
m
I
V
C
o
r
e
i
5
(
1
.
6
0
GH
z)
lap
to
p
.
Fig
u
r
e
3
p
r
esen
t so
m
e
s
am
p
les f
r
o
m
th
e
d
ataset
im
ag
es.
Fig
u
r
e
3
.
I
I
T
Delh
i e
ar
im
a
g
e
s
am
p
les
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6
9
3
0
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
,
Vo
l.
19
,
No
.
2
,
Ap
r
il 2
0
2
1
:
50
7
-
51
4
512
3
.
1
.
Det
ec
t
io
n
a
cc
ura
cy
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
R
OI
(
R
eg
io
n
o
f
I
n
ter
est)
o
r
ea
r
d
etec
tio
n
ca
n
b
e
m
ea
s
u
r
ed
[
3
]
as f
o
llo
ws:
=
×
100
%
(
9
)
Ho
wev
er
,
th
e
ea
r
d
etec
tio
n
ac
cu
r
ac
y
o
f
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
a
ch
s
h
o
wed
a
well
r
esu
lt
alm
o
s
t
(
9
6
.
5
%)
co
m
p
ar
ed
to
o
t
h
er
p
r
ev
i
o
u
s
wo
r
k
s
as
s
h
o
wn
in
T
ab
le
1
.
Ad
d
itio
n
ally
,
we
ca
n
n
o
tice
in
th
e
tab
le
m
en
tio
n
ed
b
ef
o
r
e
th
at
t
h
e
m
o
s
t
o
f
r
esear
c
h
er
s
test
ed
th
eir
wo
r
k
s
o
n
o
w
n
d
atasets
,
wh
ich
ar
e
m
o
s
tly
n
o
t
s
tan
d
ar
d
d
ataset.
Mo
r
eo
v
er
,
s
m
all
n
u
m
b
e
r
s
o
f
s
am
p
les
o
f
th
eir
o
wn
d
atasets
an
d
I
I
T
Delh
i
d
atasets
ar
e
u
s
ed
f
o
r
test
in
g
an
d
ev
alu
atio
n
th
eir
ea
r
r
ec
o
g
n
itio
n
s
y
s
tem
s
.
T
ab
le
1
.
E
ar
an
d
o
th
er
p
h
y
s
io
l
o
g
ical
h
u
m
a
n
tr
aits
co
m
p
a
r
is
o
n
P
u
b
l
i
c
a
t
i
o
n
A
p
p
r
o
a
c
h
N
a
me
o
f
D
a
t
a
s
e
t
Ea
r
i
m
a
g
e
s
a
m
p
l
e
s
A
c
c
u
r
a
c
y
M
u
r
u
k
e
s
h
.
C
e
t
a
l
.
[
2
]
C
o
n
t
o
u
l
e
t
a
n
d
P
C
A
I
I
T
D
e
l
h
i
50
9
6
%
H
a
d
i
a
n
d
G
e
o
r
g
e
[
5
]
C
o
l
o
r
sk
i
n
,
e
d
g
e
d
e
t
e
c
t
i
o
n
a
n
d
i
m
a
g
e
su
b
t
r
a
c
t
i
o
n
I
I
T
D
e
l
h
i
4
9
3
9
1
%
A
n
w
a
r
,
A
.
S
.
e
t
a
l
.
[
8
]
G
e
o
me
t
r
i
c
a
l
f
e
a
t
u
r
e
s
I
I
T
D
e
l
h
i
4
5
0
9
8
%
Ta
r
i
q
a
n
d
A
k
r
a
m
[
1
0
]
H
a
a
r
w
a
v
e
l
e
t
s a
n
d
n
o
r
m
a
l
i
z
e
d
c
r
o
ss
c
o
r
r
e
l
a
t
i
o
n
(
N
C
C
)
I
I
T
D
e
l
h
i
1
2
5
s
u
b
j
e
c
t
9
5
%
Ji
t
e
n
d
r
a
,
B
.
[
2
7
]
G
e
o
me
t
r
i
c
a
l
f
e
a
t
u
r
e
Th
e
i
r
O
w
n
D
a
t
a
s
e
t
30
9
0
%
A
l
a
r
a
j
e
t
a
l
.
[
2
8
]
P
r
i
n
c
i
p
a
l
c
o
mp
o
n
e
n
t
s
a
n
a
l
y
s
i
s
(
P
C
A
)
a
n
d
M
LFF
N
N
s
Th
e
i
r
O
w
n
D
a
t
a
s
e
t
85
9
6
%
Th
e
p
r
o
p
o
s
e
d
a
p
p
r
o
a
c
h
G
a
u
ss
i
a
n
,
La
p
l
a
c
e
,
e
d
g
e
d
e
t
e
c
t
i
o
n
,
i
m
a
g
e
su
b
t
r
a
c
t
i
o
n
a
n
d
m
o
d
i
f
i
e
d
a
d
a
p
t
i
v
e
s
e
a
r
c
h
w
i
n
d
o
w
(ASW)
I
I
T
D
e
l
h
i
4
9
3
9
6
%
3
.
2
.
P
r
o
ce
s
s
ing
t
im
e
T
h
e
p
r
o
ce
s
s
in
g
tim
e
o
f
ea
c
h
p
r
o
ce
d
u
r
e
f
o
r
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
is
m
ea
s
u
r
ed
in
s
ec
o
n
d
wh
ich
is
u
tili
ze
d
to
ev
alu
ate
th
e
c
o
m
p
u
tatio
n
al
tim
e
co
s
t
o
f
th
e
p
r
o
p
o
s
e
d
ap
p
r
o
ac
h
.
T
h
e
av
e
r
ag
e
tim
e
f
o
r
ev
er
y
p
r
o
ce
d
u
r
e
is
co
m
p
u
ted
v
ia
s
ix
p
r
o
c
ed
u
r
e
s
/o
p
er
atio
n
s
s
u
ch
as
s
tr
etch
in
g
,
Gau
s
s
ian
,
L
ap
lace
,
So
b
el
e
d
g
e
d
etec
tio
n
,
i
m
ag
e
s
u
b
tr
ac
tio
n
an
d
ea
r
d
etec
tio
n
.
I
n
th
is
r
eg
ar
d
,
o
u
r
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
ac
h
iev
e
d
an
ef
f
icien
t
co
m
p
u
tatio
n
al
tim
e
(
≈
0
.
4
8
5
s
ec
o
n
d
s
)
.
Fig
u
r
e
4
s
h
o
ws
th
e
ca
lcu
latio
n
o
f
th
e
p
r
o
ce
d
u
r
es
o
f
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
th
at
a
r
e
u
s
ed
to
g
eth
er
to
d
etec
t
ea
r
lan
d
m
ar
k
s
.
T
ab
le
2
s
u
m
m
a
r
izes
th
e
p
r
ev
io
u
s
wo
k
s
r
esu
lts
co
m
p
a
r
e
d
with
th
e
ef
f
icien
t
ac
q
u
ir
ed
c
o
m
p
u
tatio
n
al
tim
e
r
esu
lt f
r
o
m
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
.
I
n
s
p
ite
o
f
o
u
r
r
esu
lts
g
iv
e
litt
le
b
it
less
ac
cu
r
ac
y
o
f
d
etec
tio
n
r
elatio
n
to
s
o
m
e
o
th
er
r
esear
ch
er
s
,
b
u
t
it
is
m
o
r
e
ef
f
icien
t
in
ter
m
s
o
f
co
m
p
u
tatio
n
al
tim
e
as
s
h
o
wn
in
T
ab
le
2
.
On
o
th
er
h
an
d
,
th
e
ap
p
r
o
ac
h
h
ad
s
o
m
e
m
is
d
etec
tio
n
r
esu
lts
wh
ich
is
s
h
o
wn
i
n
Fig
u
r
e
5
f
o
r
m
o
r
e
clar
if
icatio
n
.
Actu
ally
,
we
ca
n
n
o
tice
th
at
th
e
p
r
o
b
lem
o
f
m
is
d
etec
tio
n
is
th
e
d
is
jo
in
t
p
ix
els
o
f
ea
r
ed
g
es.
I
n
f
ac
t,
th
e
d
en
s
ity
o
f
wh
ite
p
ix
els
is
af
f
ec
ted
o
n
th
e
w
o
r
k
o
f
ea
r
d
etec
tio
n
p
r
o
ce
s
s
u
s
in
g
ASW
tech
n
iq
u
e
to
s
eg
m
e
n
t
th
e
R
OI
o
b
ject
c
o
r
r
e
ctly
,
b
e
ca
u
s
e
th
e
d
etec
to
r
s
d
ep
en
d
e
d
o
n
co
u
n
tin
g
th
ese
wh
ite
p
ix
els
o
f
ea
r
ed
g
e
an
d
if
th
er
e
ar
e
m
an
y
g
a
p
s
(
b
lack
p
i
x
els)
in
s
h
ap
e
o
f
ea
r
’
s
ed
g
e;
th
e
r
esu
lt
o
f
d
etec
tio
n
ac
cu
r
ac
y
r
ate
will
b
e
d
ec
r
ea
s
ed
.
I
n
th
is
r
eg
ar
d
,
we
will
wo
r
k
o
n
m
a
k
e
s
o
m
e
en
h
an
ce
m
e
n
t
s
o
n
th
e
ap
p
r
o
a
ch
to
y
ield
m
o
r
e
ac
cu
r
ate
d
e
tectio
n
ac
cu
r
ac
y
r
ate
in
h
a
r
m
o
n
y
with
ef
f
icien
t
co
m
p
u
tatio
n
al
tim
e
i
n
th
e
f
u
tu
r
e.
Fig
u
r
e
4
.
E
a
r
d
etec
tio
n
a
p
p
r
o
a
ch
’
s
co
m
p
u
tatio
n
al
tim
e
r
esu
lt
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
A
u
to
ma
tic
h
u
ma
n
ea
r
d
etec
tio
n
a
p
p
r
o
a
ch
u
s
in
g
… (
R
a
a
d
A
h
med
Ha
d
i
)
513
Fig
u
r
e
5
.
Dis
jo
in
t p
ix
els o
f
ea
r
r
eg
io
n
(
s
o
m
e
s
am
p
les)
T
ab
le
2
.
C
o
m
p
a
r
is
o
n
b
etwe
en
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
an
d
o
th
er
p
r
ev
i
o
u
s
wo
r
k
s
in
ter
m
s
o
f
co
m
p
u
tatio
n
al
tim
e
P
u
b
l
i
c
a
t
i
o
n
A
p
p
r
o
a
c
h
C
o
m
p
u
t
a
t
i
o
n
a
l
t
i
m
e
(
se
c
o
n
d
s)
H
a
d
i
a
n
d
G
e
o
r
g
e
[
5
]
C
o
l
o
r
sk
i
n
,
e
d
g
e
d
e
t
e
c
t
i
o
n
a
n
d
i
m
a
g
e
su
b
t
r
a
c
t
i
o
n
1
.
3
3
Ta
r
i
q
a
n
d
A
k
r
a
m
[
10
]
H
a
a
r
w
a
v
e
l
e
t
s a
n
d
N
o
r
m
a
l
i
z
e
d
C
r
o
ss
C
o
r
r
e
l
a
t
i
o
n
(
N
C
C
)
0
.
6
0
Th
e
p
r
o
p
o
s
e
d
a
p
p
r
o
a
c
h
G
a
u
ss
i
a
n
,
La
p
l
a
c
e
,
e
d
g
e
d
e
t
e
c
t
i
o
n
,
i
m
a
g
e
s
u
b
t
r
a
c
t
i
o
n
a
n
d
mo
d
i
f
i
e
d
A
d
a
p
t
i
v
e
sea
r
c
h
w
i
n
d
o
w
(
A
S
W
)
0
.
4
8
4.
CO
NCLU
SI
O
N
An
ef
f
icien
t,
r
eliab
le
a
n
d
s
im
p
le
ap
p
r
o
ac
h
h
as
s
u
cc
ess
f
u
lly
p
r
o
p
o
s
ed
f
o
r
a
u
to
m
atic
h
u
m
an
ea
r
d
etec
tio
n
; th
is
ap
p
r
o
ac
h
is
b
as
ed
o
n
u
s
in
g
m
o
d
i
f
ied
ASW
.
I
t
is
co
n
s
is
ted
o
f
two
p
h
ases
: p
r
ep
r
o
ce
s
s
in
g
an
d
ea
r
lan
d
m
ar
k
s
d
etec
tio
n
.
Firstl
y
,
th
r
ee
o
p
er
atio
n
s
(
im
ag
e
co
n
tr
a
s
t
s
tr
etch
in
g
,
Gau
s
s
ian
b
lu
r
,
an
d
L
ap
lace
f
ilter
)
o
f
im
ag
e
en
h
an
c
em
en
t a
r
e
u
tili
ze
d
to
m
ak
e
im
p
r
o
v
em
en
t o
n
all
im
ag
es (
in
cr
ea
s
in
g
th
e
c
o
n
tr
as
t,
r
ed
u
ce
th
e
n
o
is
y
an
d
s
m
o
o
t
h
in
g
p
r
o
ce
s
s
es).
Seco
n
d
ly
,
two
o
p
er
atio
n
s
(
So
b
e
l
ed
g
e
d
etec
to
r
an
d
im
ag
e
s
u
b
tr
ac
tio
n
)
a
r
e
u
s
ed
h
ig
h
lig
h
ted
an
d
d
eter
m
in
e
d
th
e
o
n
ly
w
h
ite
p
ix
els
o
f
th
e
ea
r
e
d
g
es
r
esp
ec
tiv
ely
.
I
n
a
d
d
itio
n
,
a
m
o
d
if
ied
m
eth
o
d
wh
ich
is
ad
o
p
ted
f
r
o
m
ASW
is
u
s
ed
to
d
etec
t
th
e
ea
r
lan
d
m
a
r
k
s
r
eg
io
n
.
L
ik
ewise,
o
u
r
ap
p
r
o
ac
h
is
test
ed
o
n
I
I
T
Delh
i
s
tan
d
ar
d
ea
r
b
io
m
etr
ic
p
u
b
lic
d
ataset.
E
x
p
e
r
im
en
tal
r
e
s
u
lts
p
r
esen
ted
a
well
av
er
ag
e
d
etec
tio
n
r
ate
9
6
%
f
o
r
4
9
3
im
ag
e
s
am
p
les
f
r
o
m
1
2
5
p
er
s
o
n
s
an
d
co
m
p
u
tatio
n
al
tim
e
alm
o
s
t
≈
0
.
4
8
5
s
ec
o
n
d
s
wh
ich
is
ev
alu
ated
with
o
th
er
p
r
ev
io
u
s
wo
r
k
s
.
I
n
t
h
e
f
u
tu
r
e,
we
will
ex
p
an
d
o
u
r
t
est
in
m
an
y
ea
r
d
atab
ases
wi
th
d
if
f
er
en
t
s
ce
n
ar
io
s
s
u
ch
as p
o
s
e
v
ar
iatio
n
,
o
cc
lu
s
i
o
n
,
s
ca
le
an
d
illu
m
i
n
atio
n
c
h
a
n
g
es m
atter
s
.
ACK
NO
WL
E
DG
E
M
E
NT
S
T
h
e
au
th
o
r
s
ar
e
g
r
atef
u
l
an
d
a
p
p
r
ec
iativ
e
to
th
e
Min
is
tr
y
o
f
Hig
h
er
E
d
u
ca
tio
n
a
n
d
Scien
tif
ic
R
esear
ch
(
MO
HE
SE)
,
I
r
aq
,
f
o
r
s
u
p
p
o
r
ti
n
g
th
is
r
esear
ch
g
r
an
t.
RE
F
E
R
E
NC
E
S
[1
]
Vé
lez
J.
F
.
,
S
.
Á.,
M
o
re
n
o
B
.
a
n
d
S
u
ra
l
S
.
,
"
Ro
b
u
st
Ear
De
tec
ti
o
n
fo
r
Bio
m
e
tri
c
Ve
rifi
c
a
ti
o
n
,
"
IADI
S
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
n
C
o
mp
u
ter
S
c
ien
c
e
a
n
d
In
f
o
rm
a
ti
o
n
S
y
ste
ms
,
v
o
l.
8
,
n
o
.
1
,
p
p
.
31
-
4
6
,
2
0
1
3
.
[2
]
M
u
ru
k
e
sh
C.
,
A.
P
a
riv
a
z
h
a
g
a
n
,
a
n
d
K.
Th
a
n
u
sh
k
o
d
i,
"
A
N
o
v
e
l
Ear
Re
c
o
g
n
i
ti
o
n
P
r
o
c
e
ss
Us
in
g
A
p
p
e
a
ra
n
c
e
S
h
a
p
e
M
o
d
e
l,
F
is
h
e
r
Li
n
e
a
r
Disc
rimin
a
n
t
An
a
ly
sis
a
n
d
Co
n
to
u
rlet
Tran
sfo
rm
,
"
Pro
c
e
d
i
a
E
n
g
in
e
e
rin
g
,
v
o
l
.
38
,
p
p
.
7
7
1
-
7
7
8
,
2
0
1
2
.
[3
]
P
ra
k
a
sh
S
.
,
a
n
d
G
u
p
ta
P
.
,
"
Ear
Bio
m
e
tri
c
s
in
2
D
a
n
d
3
D
,
"
A
u
g
me
n
ted
Vi
si
o
n
a
n
d
Rea
li
ty
,
e
d
.
L.
B.
W.
Riad
I
.
Ha
m
m
o
u
d
,
v
o
l
.
1
0
,
2
0
1
5
.
[4
]
G
h
o
u
a
lmi
L.
,
A.
Dra
a
,
a
n
d
S
.
C
h
i
k
h
i,
"
An
e
a
r
b
i
o
m
e
tri
c
sy
ste
m
b
a
se
d
o
n
a
rti
ficia
l
b
e
e
s
a
n
d
th
e
sc
a
le
in
v
a
rian
t
fe
a
tu
re
tran
sfo
rm
,
"
Ex
p
e
rt
S
y
ste
ms
wit
h
Ap
p
li
c
a
ti
o
n
s,
v
o
l
.
57
,
p
p
.
4
9
-
61
,
2
0
1
6
.
[5
]
Ha
d
i,
R.
A.
a
n
d
L.
E.
G
e
o
rg
e
,
"
A
n
Au
t
o
m
a
ted
Ear
De
tec
ti
o
n
M
e
t
h
o
d
b
a
se
d
o
n
S
o
b
e
l
E
d
g
e
De
tec
t
io
n
a
n
d
Im
a
g
e
S
u
b
trac
ti
o
n
Tec
h
n
iq
u
e
s
,
"
2
0
1
8
T
h
irtee
n
th
I
n
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
Dig
it
a
l
In
f
o
rm
a
ti
o
n
M
a
n
a
g
e
me
n
t
(ICDI
M
2
0
1
8
)
,
2
0
1
8
.
[6
]
Ci
n
tas
C.
,
e
t
a
l
.,
"
Au
t
o
m
a
ti
c
Ear
De
tec
ti
o
n
a
n
d
S
e
g
m
e
n
tati
o
n
o
v
e
r
P
a
rti
a
ll
y
Oc
c
lu
d
e
d
P
ro
fil
e
F
a
c
e
Im
a
g
e
s
,
"
J
o
u
rn
a
l
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
19
,
n
o
.
01
,
p
p
.
8
1
-
9
0
,
2
0
1
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6
9
3
0
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
,
Vo
l.
19
,
No
.
2
,
Ap
r
il 2
0
2
1
:
50
7
-
51
4
514
[7
]
G
a
n
a
p
a
th
i
I.
I.
,
e
t
a
l.
,
"
Un
c
o
n
str
a
in
e
d
e
a
r
d
e
tec
ti
o
n
u
si
n
g
e
n
se
m
b
le
-
b
a
se
d
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
two
rk
m
o
d
e
l
,
"
Co
n
c
u
rr
e
n
c
y
a
n
d
C
o
mp
u
ta
ti
o
n
:
P
ra
c
ti
c
e
a
n
d
Exp
e
rie
n
c
e
,
v
o
l.
32
,
n
o
.
1
,
2
0
2
0
.
[8
]
An
wa
r
A.
S
.
,
K
.
K.
A.
G
h
a
n
y
,
a
n
d
H.
El
m
a
h
d
y
,
"
Hu
m
a
n
Ear
Re
c
o
g
n
it
i
o
n
Us
in
g
G
e
o
m
e
tri
c
a
l
F
e
a
tu
re
s
Ex
trac
ti
o
n
,
"
Pro
c
e
d
ia
C
o
mp
u
ter
S
c
ien
c
e
,
v
o
l.
65
,
p
p
.
5
2
9
-
5
3
7
,
2
0
1
5
.
[9
]
Zh
a
n
g
Y.
,
a
n
d
Z
.
M
u
,
"
Ear
De
t
e
c
ti
o
n
u
n
d
e
r
Un
c
o
n
tr
o
ll
e
d
Co
n
d
i
ti
o
n
s
wit
h
M
u
lt
ip
le
S
c
a
le
F
a
ste
r
Re
g
io
n
-
Ba
se
d
Co
n
v
o
l
u
ti
o
n
a
l
Ne
u
ra
l
Ne
two
r
k
s
,
"
S
y
mm
e
try
,
v
o
l.
9
,
n
o
.
4
,
2
0
1
7
.
[1
0
]
Tariq
,
A.
a
n
d
Ak
ra
m
M
.
U.
,
"
P
e
rso
n
a
l
i
d
e
n
ti
f
ica
ti
o
n
u
si
n
g
c
o
m
p
u
teriz
e
d
h
u
m
a
n
e
a
r
re
c
o
g
n
it
io
n
sy
ste
m
,
"
T
EL
KOM
NIKA
T
e
lec
o
mm
u
n
ic
a
ti
o
n
C
o
mp
u
ti
n
g
El
e
c
tro
n
ics
a
n
d
Co
n
tro
l
,
v
o
l
.
10
,
n
o
.
2
,
p
p
.
3
2
1
-
3
2
6
,
2
0
1
2
.
[1
1
]
Be
n
z
a
o
u
i
A.,
A.
Kh
e
i
d
e
r,
a
n
d
A.
Bo
u
k
r
o
u
c
h
e
,
"
Ear
d
e
sc
rip
t
io
n
a
n
d
re
c
o
g
n
it
i
o
n
u
sin
g
EL
BP
a
n
d
wa
v
e
lets
,
"
2
0
1
5
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
A
p
p
l
ied
Res
e
a
rc
h
in
Co
mp
u
ter
S
c
ien
c
e
a
n
d
En
g
i
n
e
e
rin
g
(ICA
R)
,
2
0
1
5
.
[1
2
]
M
e
ra
o
u
m
ia
A.,
S
.
Ch
it
r
o
u
b
,
a
n
d
A
.
Bo
u
rid
a
n
e
,
"
A
n
a
u
to
m
a
ted
e
a
r
i
d
e
n
ti
fica
ti
o
n
sy
ste
m
u
sin
g
G
a
b
o
r
fi
lt
e
r
re
sp
o
n
se
s
,
"
2
0
1
5
IE
EE
1
3
t
h
In
ter
n
a
ti
o
n
a
l
Ne
w
Circ
u
it
s a
n
d
S
y
ste
ms
Co
n
fer
e
n
c
e
(NEW
CAS
)
,
2
0
1
5
.
[1
3
]
De
e
p
a
k
R.
,
A.V.
Na
y
a
k
,
a
n
d
K
.
M
a
n
i
k
a
n
tan
,
"
Ear
d
e
tec
ti
o
n
u
s
in
g
a
c
ti
v
e
c
o
n
t
o
u
r
m
o
d
e
l
,
"
2
0
1
6
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Eme
rg
in
g
T
re
n
d
s i
n
En
g
i
n
e
e
rin
g
,
T
e
c
h
n
o
lo
g
y
a
n
d
S
c
ien
c
e
(ICET
ET
S
)
,
2
0
1
6
.
[1
4
]
P
ra
k
a
sh
S
.
,
a
n
d
P
.
G
u
p
ta,
"
A
n
e
f
ficie
n
t
e
a
r
lo
c
a
li
z
a
ti
o
n
tec
h
n
i
q
u
e
,
"
Ima
g
e
a
n
d
Vi
si
o
n
C
o
mp
u
ti
n
g
,
v
o
l.
30
,
n
o
.
1
,
p
p
.
3
8
-
5
0
,
2
0
1
2
.
[1
5
]
Ho
u
ra
li
F
.
a
n
d
G
h
a
rra
v
i
S
.
,
"
A
n
Ear
Re
c
o
g
n
it
io
n
M
e
t
h
o
d
Ba
s
e
d
o
n
Ro
tatio
n
I
n
v
a
rian
t
Tran
s
fo
rm
e
d
DCT
,
"
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
(IJ
E
CE)
,
v
o
l.
7
,
n
o
.
5
,
p
p
.
2
8
9
5
-
2
9
0
1
,
2
0
1
7
.
[1
6
]
Ha
d
i
R.
A.,
"
M
o
v
i
n
g
Ve
h
icle
s
De
tec
ti
o
n
a
n
d
Trac
k
in
g
A
p
p
r
o
a
c
h
Via
M
o
d
ifi
e
d
Ba
c
k
g
ro
u
n
d
S
u
b
trac
ti
o
n
a
n
d
In
n
o
v
a
ti
v
e
Ad
a
p
ti
v
e
S
e
ra
c
h
Wi
n
d
o
w
,
"
P
h
D
th
e
sis,
F
a
c
u
lt
y
o
f
Co
m
p
u
ti
n
g
Un
i
v
e
rsiti
Tek
n
o
lo
g
i
M
a
lay
sia
:
M
a
lay
sia
,
p
p
.
2
1
5
,
2
0
1
5
.
[1
7
]
Ha
d
i
R.
A.,
L.
E.
G
e
o
rg
e
,
a
n
d
M
.
J.
M
o
h
a
m
m
e
d
,
"
A
Co
m
p
u
tati
o
n
a
ll
y
Ec
o
n
o
m
ic
No
v
e
l
Ap
p
ro
a
c
h
f
o
r
Re
a
l
-
Ti
m
e
M
o
v
i
n
g
M
u
l
ti
-
Ve
h
icle
De
tec
ti
o
n
a
n
d
Trac
k
in
g
t
o
wa
rd
Eff
icie
n
t
Tr
a
ffic
S
u
rv
e
il
lan
c
e
,
"
Ar
a
b
i
a
n
J
o
u
r
n
a
l
f
o
r
S
c
ien
c
e
a
n
d
En
g
i
n
e
e
rin
g
,
v
o
l.
42
,
n
o
.
2
,
p
p
.
8
1
7
-
8
3
1
,
2
0
1
7
.
[1
8
]
Ac
h
a
ry
a
T.
a
n
d
A.
K.
Ra
y
,
"
Im
a
g
e
P
ro
c
e
ss
in
g
:
P
ri
n
c
ip
les
a
n
d
Ap
p
li
c
a
ti
o
n
s
,
"
H
o
b
o
k
e
n
,
NJ
:
J
o
h
n
W
il
e
y
a
n
d
S
o
n
s
,
2
0
0
5
.
[1
9
]
Ab
d
u
l
-
Na
sir
A.
S
.
,
M
.
Y.
M
a
sh
o
r
,
a
n
d
Z
.
M
o
h
a
m
e
d
,
"
M
o
d
ifi
e
d
g
lo
b
a
l
a
n
d
m
o
d
ifi
e
d
li
n
e
a
r
c
o
n
tras
t
stre
tch
in
g
a
lg
o
rit
h
m
s:
n
e
w
c
o
lo
u
r
c
o
n
tras
t
e
n
h
a
n
c
e
m
e
n
t
tec
h
n
iq
u
e
s
fo
r
m
icro
sc
o
p
ic
a
n
a
ly
sis
o
f
m
a
laria
slid
e
i
m
a
g
e
s
,
"
Hin
d
a
w
i
Pu
b
li
s
h
in
g
C
o
rp
o
r
a
ti
o
n
C
o
mp
u
ta
ti
o
n
a
l
a
n
d
M
a
t
h
e
ma
ti
c
a
l
M
e
th
o
d
s
in
M
e
d
icin
e
,
v
o
l.
2
0
1
2
,
p
p
.
1
-
1
6
,
2
0
1
2
.
[2
0
]
Nix
o
n
M
.
a
n
d
A
.
S
.
Ag
u
a
d
o
,
"
F
e
a
t
u
re
Ex
trac
ti
o
n
&
Im
a
g
e
P
r
o
c
e
ss
in
g
fo
r
Co
m
p
u
ter
Visio
n
,
"
Th
ir
d
E
d
i
ti
o
n
,
Aca
d
e
mic
Pre
ss
,
In
c
,
2
0
1
2
.
[2
1
]
S
h
ih
F
.
Y
.
,
"
Im
a
g
e
P
ro
c
e
ss
in
g
a
n
d
P
a
tt
e
rn
Re
c
o
g
n
it
io
n
:
F
u
n
d
a
m
e
n
tals a
n
d
Tec
h
n
iq
u
e
s
,
"
W
il
e
y
,
2
0
1
0
.
[2
2
]
Ha
ra
li
c
k
R.
M
.
a
n
d
L
.
G
.
S
h
a
p
ir
o
,
"
C
o
m
p
u
ter
a
n
d
Ro
b
o
t
Visi
o
n
,
"
Ad
d
iso
n
-
W
e
sle
y
L
o
n
g
m
a
n
Pu
b
l
ish
in
g
Co
.
,
I
n
c
,
2
0
0
2
.
[2
3
]
G
o
n
z
a
lez
R.
C.
,
a
n
d
R.
E.
Wo
o
d
s
,
"
Dig
it
a
l
Im
a
g
e
P
ro
c
e
ss
in
g
,
"
Up
p
e
r S
a
d
d
le
Ri
v
e
r,
NJ
:
Pre
n
ti
c
e
Ha
ll
,
2
0
1
8
.
[2
4
]
M
a
rq
u
e
s
O.,
"
P
ra
c
ti
c
a
l
Im
a
g
e
a
n
d
Vid
e
o
P
ro
c
e
ss
in
g
u
sin
g
M
ATL
AB
,
"
Ho
b
o
k
e
n
,
Ne
w
J
e
rs
e
y
:
J
o
h
n
W
il
e
y
&
S
o
n
s,
In
c
,
2
0
1
1
.
[2
5
]
G
a
o
W.
,
e
t
a
l.
,
"
An
Im
p
ro
v
e
d
S
o
b
e
l
Ed
g
e
De
tec
ti
o
n
,
"
2
0
1
0
3
rd
IEE
E
In
ter
n
a
t
io
n
a
l
Co
n
fer
e
n
c
e
o
n
C
o
mp
u
ter
S
c
ien
c
e
a
n
d
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
lo
g
y
(IC
CS
IT
)
,
C
h
e
n
g
d
u
,
Ch
in
a
p
p
.
6
7
-
71
,
2
0
1
0
.
[2
6
]
P
ra
tt
W.
K.
,
"
Di
g
it
a
l
Im
a
g
e
P
r
o
c
e
ss
in
g
:
P
IK
S
In
si
d
e
,
"
F
o
u
rth
e
d
.
J
o
h
n
W
il
e
y
&
S
o
n
s,
In
c
,
2
0
0
7
.
[2
7
]
Jiten
d
ra
B.
Ja
wa
le
a
n
d
S
.
M
.
T.
An
jali
S
.
Bh
a
lch
a
n
d
ra
,
"
Th
e
Hu
m
a
n
Id
e
n
ti
fica
ti
o
n
S
y
ste
m
Us
in
g
M
u
lt
i
p
le
G
e
o
m
e
tri
c
a
l
F
e
a
tu
re
Ex
trac
ti
o
n
o
f
Ear
–
An
I
n
n
o
v
a
ti
v
e
Ap
p
ro
a
c
h
,
"
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Eme
rg
i
n
g
T
e
c
h
n
o
lo
g
y
a
n
d
Ad
v
a
n
c
e
d
E
n
g
in
e
e
rin
g
,
v
o
l.
2
,
n
o
.
3
,
p
p
.
6
6
2
-
6
6
6
,
2
0
1
2
.
[2
8
]
Ala
ra
j
M
.
,
Ho
u
J.
,
a
n
d
F
u
k
a
m
i
T.
,
"
A
Ne
u
ra
l
Ne
two
r
k
Ba
se
d
Hu
m
a
n
Id
e
n
ti
f
ica
ti
o
n
F
ra
m
e
wo
rk
Us
in
g
Ear,
"
Pro
c
e
e
d
in
g
s
o
f
T
ENCON 2
0
1
0
IE
EE
Reg
i
o
n
1
0
Co
n
fer
e
n
c
e
,
F
u
k
u
o
k
a
,
Ja
p
a
n
,
2
0
1
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.