TELK
O
M
N
I
K
A
Te
l
e
c
ommu
n
i
c
at
i
on
,
C
omp
u
ti
n
g
,
El
e
c
t
r
on
i
c
s
an
d
C
on
tr
o
l
V
ol
.
18
,
N
o
.
2
,
A
pri
l
2020
,
p
p.
1122
~
112
9
IS
S
N
:
1693
-
6930
,
a
c
c
r
e
di
t
e
d
F
i
rs
t
G
r
a
de
by
K
e
m
e
n
ri
s
t
e
kdi
k
t
i
,
D
e
c
r
e
e
N
o:
21
/
E
/
K
P
T
/
2
018
D
O
I:
10
.
129
28/
T
E
L
K
O
M
N
IK
A
.
v1
8i
2
.
1487
8
1122
Jou
r
n
al
h
o
m
e
pa
ge
:
ht
t
p:
/
/
j
our
nal
.
uad
.
a
c
.
i
d/
i
nd
e
x
.
php
/
T
E
L
K
O
MNIKA
A
d
a
p
t
i
v
e
t
h
r
e
sh
o
l
d
f
o
r
mo
v
i
n
g
o
b
j
e
c
t
s dete
c
t
i
o
n
u
si
n
g
g
a
u
ssi
a
n
mi
x
t
u
r
e
m
o
d
e
l
M
oc
h
A
r
i
e
f
S
oe
l
e
man
,
A
r
i
s
N
u
r
h
i
n
d
a
r
to
,
M
u
s
l
i
h
,
K
a
r
i
s
W
.
,
M
u
l
jon
o
,
F
ar
i
k
h
A
l
Za
mi
,
R
.
A
n
gg
i
P
r
amu
n
e
n
d
ar
F
a
c
u
l
t
y
o
f
C
o
m
pu
t
e
r
S
c
i
e
nc
e
D
i
a
n
N
u
s
w
a
n
t
or
o
U
ni
v
e
r
s
i
t
y
,
I
nd
one
s
i
a
A
r
ti
c
l
e
I
n
fo
A
BS
TR
A
C
T
Ar
t
i
c
l
e
h
i
s
t
or
y
:
Re
c
e
i
v
e
d
A
ug
14
,
201
9
Re
vi
s
e
d
D
e
c
30
,
20
19
A
c
c
e
pt
e
d
F
e
b
20
,
202
0
M
ov
i
ng
ob
j
e
c
t
de
t
e
c
t
i
on
be
c
o
m
e
s
t
he
i
m
po
r
t
a
n
t
t
a
s
k
i
n
t
h
e
v
i
de
o
s
ur
ve
i
l
a
n
c
e
s
ys
t
e
m
.
D
e
f
i
n
i
n
g
t
he
t
h
r
e
s
ho
l
d
a
ut
o
m
a
t
i
c
a
l
l
y
i
s
c
ha
l
l
e
ngi
ng
t
o
di
f
f
e
r
e
nt
i
a
t
e
t
he
m
o
vi
n
g
ob
j
e
c
t
f
r
o
m
t
he
ba
c
kgr
ound
w
i
t
hi
n
a
v
i
d
e
o.
T
hi
s
s
t
ud
y
p
r
op
os
e
s
ga
us
s
i
a
n
m
i
x
t
ur
e
m
od
e
l
(
G
M
M
)
a
s
a
t
h
r
e
s
ho
l
d
s
t
r
a
t
e
gy
i
n
m
ov
i
ng
ob
j
e
c
t
de
t
e
c
t
i
o
n.
T
h
e
p
e
r
f
or
m
a
nc
e
o
f
t
h
e
pr
o
pos
e
d
m
e
t
hod
i
s
c
o
m
pa
r
e
d
t
o
t
he
O
t
s
u
a
l
go
r
i
t
h
m
a
n
d
gr
a
y
t
h
r
e
s
ho
l
d
a
s
t
he
ba
s
e
l
i
n
e
m
e
t
h
od
us
i
ng
m
e
a
n
s
q
ua
r
e
e
r
r
or
(
M
S
E
)
a
nd
P
e
a
k
S
i
gn
a
l
N
o
i
s
e
R
a
t
i
o
(
P
S
N
R
)
.
T
he
p
e
r
f
or
m
a
nc
e
c
o
m
pa
r
i
s
o
n
of
t
he
m
e
t
ho
ds
i
s
e
v
a
l
ua
t
e
d
o
n
hu
m
a
n
vi
d
e
o
da
t
a
s
e
t
.
T
he
a
v
e
r
a
ge
r
e
s
u
l
t
o
f
M
S
E
va
l
u
e
G
M
M
i
s
25
7.
1
8,
O
t
s
u
i
s
595
.
36
a
nd
G
r
a
y
i
s
6
45
.
39
,
s
o
t
he
M
S
E
v
a
l
ue
i
s
l
ow
e
r
t
ha
n
O
t
s
u
a
nd
G
r
a
y
t
h
r
e
s
ho
l
d.
T
he
a
v
e
r
a
ge
r
e
s
ul
t
o
f
P
S
N
R
va
l
ue
G
M
M
i
s
2
4.
71,
O
t
s
u
i
s
20
.
66
a
nd
G
r
a
y
i
s
1
9.
3
5
,
s
o
t
he
P
S
N
R
va
l
u
e
is
hi
g
he
r
t
ha
n
O
t
s
u
a
nd
G
r
a
y
t
h
r
e
s
ho
l
d
.
T
h
e
pe
r
f
or
m
a
nc
e
of
t
h
e
p
r
op
os
e
d
m
e
t
hod
out
p
e
r
f
or
m
s
t
h
e
b
a
s
e
l
i
ne
m
e
t
ho
d
i
n
t
e
r
m
of
e
r
r
or
de
t
e
c
t
i
on
.
Ke
y
w
or
d
s
:
D
e
t
e
c
t
i
on
G
a
us
s
i
a
n
m
i
xt
ure
m
ode
l
M
ovi
ng
obj
e
c
t
O
t
s
u
T
hre
s
ho
l
d
T
hi
s
i
s
an
op
e
n
ac
c
e
s
s
ar
t
i
c
l
e
u
nde
r
t
he
C
C
B
Y
-
SA
l
i
c
e
n
s
e
.
Cor
r
e
s
pon
di
n
g
Au
t
h
or
:
M
oc
h
A
r
i
e
f
S
o
e
l
e
m
a
n
,
F
a
c
ul
t
y
of
Com
put
e
r
S
c
i
e
nc
e
,
D
i
a
n
N
us
w
a
nt
oro
U
ni
v
e
rs
i
t
y,
207
I
m
a
m
Bonj
o
l
S
t
.
,
Indon
e
s
i
a
,
E
m
a
i
l
:
a
r
i
e
f2
2208
@
gm
a
i
l
.
c
o
m
1.
I
N
TR
O
D
U
C
TI
O
N
Re
s
e
a
rc
h
i
n
vi
d
e
o
pro
c
e
s
s
i
n
g
ha
s
be
c
o
m
e
on
e
of
t
h
e
e
x
c
i
t
i
ng
fi
e
l
ds
i
n
c
o
m
put
e
r
vi
s
i
on
re
s
e
a
r
c
h
,
a
s
t
he
us
e
of
m
e
t
hods
c
ont
i
nue
s
t
o
b
e
de
ve
l
ope
d
t
o
produ
c
e
q
ua
l
i
t
y
r
e
s
ul
t
s
i
n
t
h
e
fi
e
l
d
of
c
om
put
e
r
vi
s
i
o
n,
e
s
pe
c
i
a
l
l
y
d
e
t
e
c
t
i
o
n
o
f
m
ov
i
ng
ob
j
e
c
t
s
ba
s
e
d
on
s
ur
ve
i
l
l
a
n
c
e
c
a
m
e
r
a
s
.
T
h
e
i
n
c
re
a
s
i
n
g
CT
T
V
t
e
c
hno
l
ogy
,
e
s
pe
c
i
a
l
l
y
i
n
t
h
e
us
e
o
f
s
urve
i
l
l
a
n
c
e
c
a
m
e
r
a
s
i
n
m
on
i
t
or
i
ng
hu
m
a
n
a
c
t
i
v
i
t
i
e
s
or
a
c
t
i
v
i
t
i
e
s
i
n
d
a
i
l
y
l
i
fe
b
ot
h
a
t
ho
m
e
,
offi
c
e
s
,
a
nd
i
n
op
e
n
s
pa
c
e
s
.
I
n
t
he
c
e
n
t
ra
l
pa
rt
o
f
m
on
i
t
or
i
ng
a
c
t
i
vi
t
i
e
s
of
m
ovi
ng
o
bj
e
c
t
s
,
e
s
pe
c
i
a
l
l
y
hum
a
ns
,
t
h
e
pro
c
e
s
s
of
de
t
e
c
t
i
on
of
m
ov
i
ng
ob
j
e
c
t
s
i
s
a
n
e
s
s
e
nt
i
a
l
s
t
e
p
c
a
rri
e
d
ou
t
a
t
a
n
e
a
rl
y
s
t
a
ge
t
o
b
e
a
bl
e
t
o
c
a
rry
out
m
or
e
s
p
e
c
i
fi
c
op
e
r
a
t
i
ons
i
n
ob
j
e
c
t
a
n
a
l
ys
i
s
s
u
c
h
a
s
e
xt
r
a
c
t
i
on
,
c
l
a
s
s
i
fi
c
a
t
i
o
n,
a
nd
i
de
n
t
i
f
i
c
a
t
i
on.
M
ov
i
ng
o
bj
e
c
t
de
t
e
c
t
i
on
ha
s
s
o
m
e
ba
s
i
c
f
or
m
a
t
i
on
,
m
a
t
he
m
a
t
i
c
a
l
l
y
t
h
e
e
xpos
ur
e
of
m
ovi
ng
obj
e
c
t
s
re
s
ul
t
s
fr
om
m
o
de
l
i
ng
t
h
e
b
a
c
kgroun
d
a
nd
f
ore
g
roun
d
[1
,
2]
.
T
hi
s
m
o
de
l
i
ng
a
i
m
s
t
o
ge
t
p
i
xe
l
s
t
ha
t
w
i
l
l
be
c
o
m
e
for
e
gro
und
-
b
a
c
k
ground
.
S
om
e
a
ppro
a
c
h
e
s
i
n
m
o
de
l
i
ng
t
hi
s
s
t
a
t
i
c
ba
c
kgro
und
i
n
c
l
u
de
us
i
ng
s
t
a
t
i
s
t
i
c
a
l
c
o
m
pon
e
nt
m
e
t
hod
s
s
uc
h
a
s
runn
i
ng
a
ve
r
a
g
e
[3,
4],
a
nd
h
i
s
t
ogr
a
m
a
na
l
ys
i
s
[5]
.
O
t
s
u
t
hre
s
h
ol
d
i
ng
[6]
i
s
a
t
hr
e
s
ho
l
d
m
e
t
h
od
i
n
s
e
gm
e
nt
a
t
i
on
t
e
c
hni
q
ue
s
,
t
h
e
a
p
pl
i
c
a
t
i
o
n
of
t
he
O
t
s
u
m
e
t
h
o
d
m
a
ke
s
i
t
e
a
s
i
e
r
t
o
do
ho
m
og
e
ne
o
us
di
v
i
s
i
on
of
pa
rt
s
b
a
s
e
d
on
s
i
m
i
l
a
r
i
t
y
c
ri
t
e
r
i
a
t
o
re
c
ogn
i
z
e
obj
e
c
t
s
.
B
e
for
e
obs
e
rvi
ng
t
h
e
s
e
g
m
e
n
t
e
d
i
m
a
ge
,
t
h
e
pro
c
e
s
s
m
us
t
fi
rs
t
go
t
hrou
gh
t
he
i
m
a
ge
i
n
put
pro
c
e
s
s
s
o
t
ha
t
i
t
i
s
e
a
s
y
t
o
p
roc
e
e
d
t
o
t
h
e
ne
x
t
p
roc
e
s
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
A
dapt
i
v
e
t
hr
e
s
h
ol
d
f
or
m
ov
i
ng
obj
e
c
t
s
d
e
t
e
c
t
i
o
n
us
i
ng
gaus
s
i
an
m
i
x
t
ur
e
m
o
de
l
(
M
oc
h
A
r
i
e
f
S
oe
l
e
m
an
)
1123
T
he
ne
xt
proc
e
s
s
i
s
t
o
a
d
d
br
i
gh
t
ne
s
s
t
o
t
he
i
m
a
ge
t
o
i
m
pro
ve
i
m
a
ge
qu
a
l
i
t
y
.
A
f
t
e
r
t
h
a
t
,
t
he
i
m
a
ge
s
e
gm
e
nt
a
t
i
on
pr
oc
e
s
s
i
s
don
e
w
i
t
h
t
he
O
t
s
u
t
hre
s
ho
l
di
ng
m
e
t
hod
a
nd
t
hroug
h
di
s
c
ri
m
i
n
a
n
t
a
n
a
l
ys
i
s
a
p
proa
c
h
s
o
t
ha
t
i
t
c
a
n
m
a
xi
m
i
z
e
t
he
s
e
v
a
ri
a
bl
e
s
s
o
t
ha
t
obj
e
c
t
s
w
i
t
h
a
b
a
c
kg
round
c
a
n
s
e
p
a
ra
t
e
a
u
t
om
a
t
i
c
a
l
l
y.
In
a
p
re
v
i
ous
s
t
udy
,
S
o
e
l
e
m
a
n
,
M
a
ur
i
dhi
H
e
ry,
H
a
ri
a
di
i
n
[7]
us
e
d
t
h
e
a
da
pt
i
ve
t
h
re
s
ho
l
d
a
pp
roa
c
h
t
o
g
e
t
t
he
ba
c
kgr
ound
of
m
ovi
ng
ob
j
e
c
t
s
us
i
n
g
f
uz
z
y
c
-
m
e
a
ns
c
l
us
t
e
ri
ng,
a
nd
t
h
e
f
uz
z
y
m
e
t
ho
d
p
rodu
c
e
d
b
e
t
t
e
r
pe
rfor
m
a
nc
e
t
ha
n
t
h
e
O
t
s
u
t
hr
e
s
ho
l
d.
In
a
no
t
he
r
s
t
ud
y,
Z
e
ng,
J
i
a
a
nd
C
he
n
i
n
[5]
us
e
d
fuz
z
y
t
o
o
bt
a
i
n
fore
gro
und
o
bj
e
c
t
s
w
i
t
h
a
hi
s
t
o
gra
m
a
ppro
a
c
h
t
o
pa
r
t
i
t
i
on
t
h
e
fore
gro
und
fr
om
t
h
e
b
a
c
kg
round
vi
d
e
o
s
e
que
nc
e
.
In
a
not
h
e
r
s
t
udy
,
Y
e
s
ong
,
X
i
a
op
i
ng
L
i
,
N
a
F
u
a
nd
Q
i
o
ngxi
n
L
i
u
i
n
[8]
us
e
d
t
h
e
g
a
us
s
i
a
n
m
i
xt
ur
e
m
ode
l
a
pproa
c
h
t
o
de
t
e
c
t
m
ov
i
ng
ob
j
e
c
t
s
ba
s
e
d
on
b
a
c
kgr
ound
s
ubt
r
a
c
t
i
o
n.
V
i
ve
k
M
a
i
k
,
H
yung
t
a
e
K
i
m
,
D
a
e
dh
e
e
K
i
m
,
E
un
j
ung
Ch
a
e
a
nd
J
oonk
i
P
a
i
k
i
n
[9
]
m
odi
f
i
e
d
t
he
g
a
us
s
i
a
n
m
i
xt
u
re
m
od
e
l
t
o
ove
rc
o
m
e
s
udd
e
n
c
ha
n
ge
s
i
n
t
he
o
bj
e
c
t
'
s
ba
c
kgro
und.
T
h
e
s
a
m
e
m
e
t
h
od
i
n
t
he
g
a
us
s
i
a
n
m
i
x
t
ur
e
m
o
de
l
i
s
a
l
s
o
us
e
d
i
n
[
10
-
12]
.
F
rom
s
om
e
of
t
h
e
s
e
s
t
ud
i
e
s
i
t
c
a
n
be
c
onc
l
ud
e
d
t
h
a
t
d
e
t
e
c
t
i
ng
m
ov
i
ng
ob
j
e
c
t
s
i
s
a
d
i
ffi
c
u
l
t
c
ha
l
l
e
nge
i
n
t
he
d
e
t
e
c
t
i
on
m
e
t
hod
.
K
i
t
t
i
p
op
P
e
uw
nu
a
n
,
K
unt
p
ong
W
ora
r
a
t
pa
ny
a
a
nd
K
i
t
s
uc
i
n
[13]
us
i
ng
i
n
t
e
g
ra
l
i
m
a
g
e
for
t
hre
s
ho
l
di
ng
a
d
a
pt
i
ve
i
n
i
m
a
g
e
,
a
ut
ho
rs
pr
opos
e
d
t
w
o
s
t
a
ge
for
pr
oc
e
s
s
i
n
l
ow
a
nd
h
i
gh
i
n
t
e
ns
i
t
y
fo
r
t
hre
s
ho
l
di
ng
a
nd
s
e
g
m
e
nt
a
t
i
on.
S
e
nt
hi
l
kum
a
ra
n
a
n
d
V
a
i
t
he
gi
i
n
[1
4]
us
i
n
g
t
hr
e
s
ho
l
di
ng
t
e
c
hn
i
qu
e
for
s
e
gm
e
nt
a
t
i
on
m
e
d
i
c
a
l
i
m
a
ge
,
t
he
p
ropos
e
d
m
e
t
h
od
ha
ve
fa
s
t
p
e
rfor
m
a
nc
e
t
o
e
xt
ra
c
t
fo
re
gro
und
fr
om
ba
c
k
ground
.
In
d
i
ffe
re
n
t
m
e
t
ho
d
m
a
ny
r
e
s
e
a
r
c
h
e
r
us
i
ng
b
a
c
kgr
ound
m
od
e
l
l
i
ng
[1
5
-
17]
t
o
de
t
e
c
t
m
ovi
ng
obj
e
c
t
s
i
n
vi
d
e
o,
bu
t
b
a
c
kg
round
m
o
de
l
l
i
ng
h
a
ve
a
prob
l
e
m
w
h
e
n
i
t
s
ud
de
n
l
y
o
c
c
urs
a
n
i
nt
e
ns
i
t
y
c
ha
ng
e
a
nd
dyn
a
m
i
c
ba
c
kgr
ound
,
s
o
t
he
m
e
t
hod
ha
v
e
di
s
a
dva
nt
a
g
e
t
o
de
t
e
c
t
m
o
vi
ng
ob
j
e
c
t
.
In
ot
h
e
r
m
e
t
h
od
N
u
r
A
yu
ni
M
o
ha
m
e
d
a
nd
M
o
hd
A
s
yr
a
f
Z
u
l
ki
f
l
e
y
i
n
[18
]
us
i
ng
opt
i
c
a
l
f
l
ow
t
o
de
t
e
c
t
t
h
e
m
ot
i
on
of
h
um
a
n
fa
l
l
-
da
w
n,
t
ha
t
e
va
l
ua
t
i
o
n
us
i
ng
i
nt
e
rs
e
c
t
i
on
ov
e
r
u
ni
on
for
m
e
a
s
uri
ng
.
In
t
h
i
s
p
a
p
e
r,
t
h
e
g
a
us
s
i
a
n
m
i
xt
ur
e
m
od
e
l
m
e
t
hod
i
s
pr
opos
e
d
t
o
d
e
t
e
c
t
m
ovi
n
g
ob
j
e
c
t
s
w
i
t
h
a
da
p
t
i
v
e
t
hre
s
ho
l
d
a
ppro
a
c
h
t
o
ob
t
a
i
n
t
he
f
ore
gr
ound
f
rom
t
he
dyn
a
m
i
c
b
a
c
k
ground
of
a
fra
m
e
s
e
que
nc
e
.
A
s
t
hre
s
ho
l
d
s
t
ra
t
e
gi
c
g
a
us
s
i
a
n
m
i
x
t
ur
e
m
od
e
l
w
e
re
c
o
m
pa
r
e
d
a
ga
i
ns
t
gra
y
t
h
re
s
ho
l
d
a
n
d
O
t
s
u
a
l
gor
i
t
h
m
t
o
e
va
l
ua
t
e
t
he
a
c
hi
e
ve
m
e
nt
.
T
he
re
s
t
of
pa
p
e
r
i
s
a
rra
n
ge
d
a
s
f
ol
l
ow
s
.
P
a
s
s
a
ge
2
d
i
s
c
us
s
e
d
our
re
s
e
a
r
c
h
m
e
t
hod
i
n
de
t
e
c
t
i
on
m
ovi
ng
ob
j
e
c
t
s
.
P
a
s
s
a
g
e
3
s
how
s
t
h
e
r
e
s
ul
t
s
a
nd
a
na
l
ys
i
s
o
f
t
he
e
xp
e
ri
m
e
nt
a
l
ou
r
propos
e
d
a
nd
d
i
s
c
us
s
i
on
.
F
i
na
l
l
y
,
c
onc
l
us
i
o
n
a
re
g
i
ve
n
i
n
p
a
s
s
a
g
e
4
.
2.
R
ES
EA
R
C
H
M
ET
H
O
D
T
he
pr
opos
e
d
m
e
t
hod
b
e
c
om
p
os
e
d
of
s
e
ve
r
a
l
s
t
e
ps
t
o
t
h
re
s
hol
d
s
t
ra
t
e
g
i
c
i
n
t
h
e
m
o
vi
ng
o
bj
e
c
t
de
t
e
c
t
i
on
,
a
s
i
l
l
us
t
r
a
t
e
d
i
n
F
i
gur
e
1.
In
t
h
e
i
ni
t
i
a
l
s
t
a
ge
,
t
h
e
fra
m
e
bre
a
kdow
n
be
g
i
ns
i
n
t
he
vi
d
e
o
d
a
t
a
s
e
t
,
t
he
n
a
d
a
pt
i
ve
t
hr
e
s
hol
d
us
i
n
g
g
a
us
s
i
a
n
m
i
xt
ur
e
m
od
e
l
,
O
t
s
u
a
nd
gra
y
t
hr
e
s
hol
d
i
n
pa
ra
l
l
e
l
pro
c
e
s
s
t
o
g
e
t
t
he
fore
ground
fro
m
v
i
de
o
d
a
t
a
s
e
t
,
t
he
ne
x
t
s
t
e
p
t
h
e
b
i
n
a
ry
m
ov
i
ng
o
bj
e
c
t
s
i
n
t
h
e
fr
a
m
e
i
s
c
o
nt
i
nue
d
by
t
he
m
o
rphol
ogy
ope
ra
t
i
on
un
t
i
l
t
he
m
ovi
ng
ob
j
e
c
t
i
s
d
e
t
e
c
t
e
d
,
f
i
na
l
l
y
w
e
e
va
l
ua
t
e
a
l
l
p
e
rfor
m
a
n
c
e
of
a
da
p
t
i
v
e
t
hr
e
s
hol
d
m
e
t
hod
.
F
i
gure
1
.
F
l
ow
m
e
t
hod
i
n
t
h
re
s
ho
l
d
for
obj
e
c
t
d
e
t
e
c
t
i
o
n
2.
1
.
G
au
s
s
i
an
m
i
xtu
r
e
m
od
e
l
S
t
a
u
f
f
e
r
et
.
a
l
i
n
[19]
prop
os
e
d
t
h
e
g
a
us
s
i
a
n
m
i
xt
u
re
m
od
e
l
(G
M
M
)
m
e
t
hod
a
s
a
t
ype
de
ns
i
t
y
m
ode
l
c
ons
i
s
t
i
ng
of
g
a
us
s
i
a
n
fun
c
t
i
on
c
o
m
pon
e
nt
s
.
T
he
c
o
m
pon
e
n
t
of
t
hi
s
func
t
i
on
c
ons
i
s
t
s
of
d
i
ffe
re
n
t
w
e
i
gh
t
s
t
o
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
1693
-
6
930
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
,
V
ol
.
18
,
N
o.
2
,
A
pri
l
2
020:
1122
-
1129
1124
produc
e
m
ul
t
i
-
m
od
e
l
d
e
ns
i
t
y.
T
h
e
g
a
us
s
i
a
n
m
i
xt
u
re
m
od
e
l
m
e
t
hod
pr
opos
e
d
by
G
r
i
m
s
on
i
s
v
e
ry
e
ffi
c
i
e
nt
i
n
s
e
pa
r
a
t
i
ng
t
he
fo
re
gro
und
a
nd
b
a
c
k
ground
fro
m
t
h
e
i
npu
t
fr
a
m
e
s
e
qu
e
n
c
e
T
h
e
n
u
m
b
e
r
o
f
G
M
M
i
n
[
8
,
1
7
]
us
e
d
a
ffe
c
t
s
t
h
e
n
um
b
e
r
of
ba
c
kgrou
nd
m
od
e
l
s
.
T
he
gr
e
a
t
e
r
t
he
num
be
r
of
G
M
M
m
od
e
l
s
us
e
d
,
t
he
m
or
e
b
a
c
kgroun
d
m
ode
l
s
a
pi
xe
l
h
a
s
.
T
h
e
re
a
r
e
s
e
ve
r
a
l
s
t
a
g
e
s
of
t
he
pro
c
e
s
s
for
t
h
i
s
m
e
t
hod
,
na
m
e
l
y
t
h
e
s
t
a
g
e
of
m
a
t
c
hi
ng
i
nput
t
o
t
he
di
s
t
ri
bu
t
i
o
n
a
nd
t
h
e
s
t
a
ge
of
s
e
l
e
c
t
i
ng
t
he
di
s
t
r
i
bu
t
i
on
t
h
a
t
re
fl
e
c
t
s
t
h
e
b
a
c
kgroun
d.
In
G
M
M
e
a
c
h
pi
xe
l
In
t
h
e
m
a
t
c
hi
ng
s
t
a
ge
,
t
he
r
e
i
s
a
p
a
ra
m
e
t
e
r
upda
t
e
s
t
a
g
e
.
T
he
G
M
M
e
qu
a
t
i
on
m
od
e
l
[8]
i
s
a
s
(1)
.
(
)
=
∑
(
,
,
,
∑
,
)
=
1
(1)
T
he
p
a
ra
m
e
t
e
r
K
i
n
(1)
i
s
t
he
nu
m
e
ra
l
of
di
s
t
ri
bu
t
i
o
ns
,
w
hi
l
e
µ
i
s
t
h
e
a
ve
r
a
ge
v
a
l
ue
of
g
a
us
s
i
a
n
a
t
t
e
r
m
un
i
t
t
,
a
nd
∑
i
s
t
h
e
c
ova
ri
a
nc
e
m
a
t
ri
x
a
t
I
t
hr
e
s
hol
d
on
g
a
us
s
i
a
n
a
nd
ꞷ
i
s
t
he
w
e
i
g
ht
.
T
h
e
fo
l
l
ow
i
ng
e
qu
a
t
i
on
i
s
a
g
a
us
s
i
a
n
p
roba
bi
l
i
t
y
d
e
ns
i
t
y
f
unc
t
i
on
.
(
,
,
,
,
)
=
1
(
2
)
2
1
/
2
1
2
(
−
)
−
1
(
−
)
(2)
,
=
(
1
−
)
,
−
1
+
(
,
)
(3)
T
he
p
a
ra
m
e
t
e
r
n
i
s
t
he
g
a
us
s
i
a
n
di
s
t
r
i
bu
t
i
on
d
i
m
e
ns
i
o
n,
t
he
v
a
l
ue
of
n
=
1
w
h
e
n
t
h
e
ba
c
kgr
ound
m
ode
l
i
s
a
gr
a
y
i
m
a
g
e
,
∑I
,
t
i
s
a
c
o
l
or
i
m
a
ge
,
a
nd
t
h
e
a
m
ou
nt
of
n
=
3
i
s
a
n
RG
B
i
m
a
g
e
.
W
h
e
n
t
h
e
p
a
ra
m
e
t
e
r
M
i
s
1,
t
he
n
i
t
m
a
t
c
he
s
,
a
nd
v
i
c
e
v
e
rs
a
,
t
h
e
pa
r
a
m
e
t
e
r
M
i
s
0.
T
he
pa
r
a
m
e
t
e
r
va
l
ue
s
µ
a
nd
α
a
re
up
da
t
e
d
w
i
t
h
t
he
fol
l
ow
i
ng
e
qu
a
t
i
on
.
,
=
(
1
−
)
,
−
1
+
,
(4)
,
2
=
(
1
−
)
,
−
1
2
+
(
−
,
)
2
(5)
=
(
,
,
,
,
)
(6)
=
a
r
g
(
Σ
=
1
>
)
(7)
U
s
i
ng
(3)
i
n
t
h
e
pa
r
a
m
e
t
e
r
upda
t
e
s
t
a
ge
,
t
he
v
a
l
u
e
s
of
G
M
M
pa
r
a
m
e
t
e
rs
a
re
us
e
d
t
o
pro
c
e
s
s
t
h
e
n
e
xt
i
n
put
.
U
pda
t
e
d
va
l
ue
s
c
ons
i
s
t
of
w
e
i
gh
t
s
,
a
v
e
ra
g
e
s
,
a
n
d
v
a
ri
a
n
t
s
.
T
h
e
w
e
i
gh
t
v
a
l
u
e
i
s
up
da
t
e
d
e
v
e
ry
t
i
m
e
.
A
ft
e
r
t
he
w
e
i
g
ht
v
a
l
u
e
s
ha
ve
b
e
e
n
nor
m
a
l
i
z
e
d
,
t
he
t
ot
a
l
w
e
i
g
ht
of
a
l
l
di
s
t
ri
bu
t
i
ons
i
s
not
m
or
e
t
ha
n
1.
T
h
e
n
t
h
e
m
e
a
n
v
a
l
u
e
of
a
di
s
t
r
i
bu
t
i
on
i
s
upd
a
t
e
d
e
ve
ry
t
i
m
e
t
h
e
re
i
s
a
pi
x
e
l
v
a
l
ue
t
ha
t
m
a
t
c
he
s
t
he
di
s
t
r
i
bu
t
i
on
va
l
ue
.
U
s
i
ng
(5)
a
nd
(6)
t
he
s
t
a
nd
a
rd
d
e
v
i
a
t
i
on
v
a
l
ue
of
a
di
s
t
ri
b
ut
i
o
n
i
s
upda
t
e
d
e
v
e
ry
t
i
m
e
t
he
r
e
i
s
a
pi
x
e
l
v
a
l
ue
t
ha
t
m
a
t
c
h
e
s
t
ha
t
di
s
t
ri
bu
t
i
o
n.
T
he
n
e
xt
s
t
e
p
by
(7)
t
he
p
i
x
e
l
s
w
hi
c
h
a
r
e
prob
a
bi
l
i
s
t
i
c
m
a
t
c
h
e
s
t
h
e
g
a
us
s
i
a
n
m
o
d
e
l
,
i
f
t
h
e
y
a
r
e
a
ga
i
ns
t
t
he
ba
c
kgr
ound
,
t
he
y
w
i
l
l
b
e
c
l
a
s
s
i
fi
e
d
a
s
ba
c
k
ground
pi
xe
l
s
a
nd
t
h
e
i
r
oppo
ne
n
t
s
a
s
t
h
e
fore
gro
und.
2.
2
.
O
ts
u
a
l
g
or
i
th
m
T
he
O
t
s
u
m
e
t
hod
[
6]
a
i
m
s
t
o
s
p
l
i
t
t
h
e
gr
e
y
i
m
a
g
e
hi
s
t
ogra
m
upon
t
w
o
d
i
ffe
re
n
t
re
g
i
ons
a
u
t
o
m
a
t
i
c
a
l
l
y
w
i
t
hou
t
a
n
y
he
l
p
fro
m
t
h
e
us
e
r
t
o
e
nt
e
r
t
h
e
t
hr
e
s
hol
d
va
l
ue
.
O
t
s
u
a
s
t
h
e
t
hr
e
s
hol
d
m
e
t
hod
i
s
s
e
t
t
o
t
h
e
m
a
rk
us
i
ng
z
.
T
he
s
c
ope
va
l
ue
of
z
i
s
i
n
t
h
e
m
i
ds
t
of
1
a
n
d
R
,
w
he
re
ve
r
=
255
.
T
he
l
i
k
e
l
i
hood
o
f
e
v
e
ry
p
i
xe
l
i
n
th
e
th
i
de
gre
e
c
ou
l
d
b
e
re
s
o
l
ve
d
by
e
m
p
l
oye
d
(
8
).
=
/
(8)
W
he
r
e
i
n
i
n
i
s
t
h
e
a
m
ount
o
f
pi
x
e
l
i
n
t
h
e
th
i
de
gre
e
a
nd
N
i
s
t
he
u
nq
ua
l
i
fi
e
d
of
a
m
oun
t
of
pi
xe
l
s
.
T
he
m
e
a
n
gre
y
de
g
re
e
of
a
n
i
m
a
ge
fr
a
m
e
us
e
(9)
.
=
∑
×
−
1
=
1
(9)
T
ow
a
rds
s
i
ng
l
e
t
h
re
s
ho
l
d,
O
t
s
u
s
pl
i
t
t
h
e
p
i
xe
l
s
i
ns
i
de
t
w
o
gro
up
1
=
{
0
,
1
,
.
.
.
,
}
a
nd
2
=
{
+
1
,
+
2
,
.
.
.
,
−
1
}
.
T
h
e
l
i
k
e
l
i
hood
of
gro
up
c
a
n
b
e
c
a
l
c
u
l
a
t
e
d
by
a
ppl
y
i
ng
(
10
).
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
A
dapt
i
v
e
t
hr
e
s
h
ol
d
f
or
m
ov
i
ng
obj
e
c
t
s
d
e
t
e
c
t
i
o
n
us
i
ng
gaus
s
i
an
m
i
x
t
ur
e
m
o
de
l
(
M
oc
h
A
r
i
e
f
S
oe
l
e
m
an
)
1125
1
(
)
=
∑
=
1
2
(
)
=
∑
−
1
=
1
(10)
1
(
)
=
∑
×
1
(
)
=
1
2
(
)
=
∑
×
2
(
)
−
1
=
1
(11)
T
he
a
m
oun
t
of
z
c
a
n
be
c
a
l
c
ul
a
t
e
d
a
pp
l
yi
ng
(
12
)
a
nd
(13)
.
′
=
1
≤
<
2
(
)
(12)
2
(
)
=
1
(
)
(
1
(
)
−
)
2
+
2
(
)
(
2
(
)
−
)
2
(13)
2.
3
.
G
r
ay
t
h
r
e
s
h
ol
d
A
bi
n
a
ry
i
m
a
ge
[20]
c
ons
t
ru
c
t
by
t
hre
s
ho
l
di
n
g
t
hrou
gh
a
gr
e
y
s
c
a
l
e
or
c
o
l
or
i
m
a
g
e
by
s
urroun
di
ng
pi
xe
l
gra
de
t
o
1
o
r
0
re
l
y
on
w
he
t
he
r
t
he
y
a
r
e
a
bov
e
or
b
e
l
ow
t
h
e
t
h
re
s
ho
l
d
va
l
ue
.
T
h
i
s
fre
q
ue
n
t
l
y
us
u
a
l
t
o
di
s
t
i
n
c
t
o
r
por
t
i
on
a
s
e
c
t
i
on
obj
e
c
t
w
i
t
hi
n
t
h
e
i
m
a
g
e
b
a
s
e
d
upon
t
h
e
pi
x
e
l
va
l
ue
s
.
T
he
b
a
s
i
c
op
e
ra
t
i
o
n
t
hre
s
ho
l
di
ng
o
pe
r
a
t
e
s
on
a
n
i
m
a
ge
a
s
fo
l
l
ow
s
:
F
or
pi
x
e
l
(
,
)
s
c
op
e
t
he
i
m
a
ge
Z
(
,
)
>
ℎ
ℎ
(
,
)
=
1
(
,
)
=
0
T
ow
a
rds
s
c
a
l
e
of
gr
a
y
i
m
a
g
e
fr
a
m
e
,
t
h
a
t
pi
x
e
l
s
ha
v
e
s
ubs
t
a
n
c
e
a
s
i
ng
l
e
i
nt
e
ns
i
t
y
v
a
l
u
e
,
a
s
i
ng
l
e
t
hr
e
s
hol
d
m
us
t
be
p
i
c
ke
d
,
a
nd
for
c
ol
our
i
m
a
ge
,
a
de
t
a
c
he
d
t
hre
s
ho
l
d
c
a
n
b
e
s
pe
c
i
fi
e
d
f
or
e
a
c
h
na
rr
ow
.
2.
4
.
Bi
n
a
r
y
fr
a
me
mov
i
n
g
ob
je
c
t
T
he
re
s
u
l
t
of
t
h
e
t
hr
e
s
hol
di
ng
pro
c
e
s
s
w
i
t
h
t
he
a
da
p
t
i
v
e
t
hr
e
s
hol
d
w
i
t
h
g
a
us
s
i
a
n
m
i
x
t
ur
e
m
od
e
l
,
O
t
s
u
a
nd
gr
a
y
t
hre
s
ho
l
d
t
ha
t
prod
uc
e
s
bi
n
a
ry
i
m
a
ge
s
[
21]
z
e
ro
a
nd
one
w
h
e
re
t
h
e
va
l
ue
1
re
pre
s
e
nt
s
a
s
t
h
e
obj
e
c
t
or
fore
gro
und
a
nd
t
he
a
m
oun
t
0
a
s
t
h
e
ba
c
kgro
und.
2.
5
.
M
or
p
h
ol
o
gy
M
orphol
o
gi
c
a
l
f
i
l
t
e
r
r
e
s
ul
t
s
t
o
n
oi
s
e
fro
m
a
for
e
gro
und
di
s
t
r
i
but
i
on
by
us
i
ng
a
n
e
ros
i
on
fi
l
t
e
r
w
hi
c
h
a
i
m
s
t
o
r
e
du
c
e
t
h
e
di
m
e
ns
i
ons
of
t
he
v
e
hi
c
l
e
'
s
di
m
e
ns
i
ons
t
o
be
c
l
os
e
t
o
t
he
a
c
t
ua
l
s
i
z
e
,
w
hi
l
e
re
d
uc
i
ng
s
m
a
l
l
m
ove
m
e
nt
s
i
f
i
t
do
e
s
n
ot
n
e
e
d
t
o
be
s
e
gm
e
nt
e
d
.
A
l
t
ho
ugh
t
h
e
re
s
u
l
t
s
do
no
t
l
oo
k
c
o
m
pl
e
t
e
l
y
c
l
e
a
n
,
t
h
i
s
c
ondi
t
i
o
n
i
s
go
od
e
noug
h
f
or
t
he
d
e
t
e
c
t
i
on
of
m
o
vi
ng
ob
j
e
c
t
s
.
M
orphol
o
gy
ope
ra
t
i
ons
i
n
t
he
f
orm
of
e
ros
i
o
n
a
nd
d
i
l
a
t
i
on
w
e
re
done
t
o
ge
t
be
t
t
e
r
ob
j
e
c
t
re
s
u
l
t
s
b
a
s
e
d
on
t
h
e
s
ha
pe
of
t
he
o
bj
e
c
t
[22]
.
D
i
l
a
t
i
o
n
i
s
t
he
pro
c
e
s
s
of
a
dd
i
ng
pi
x
e
l
s
t
o
t
he
bound
a
ry
of
a
n
ob
j
e
c
t
i
n
a
d
i
gi
t
a
l
i
nput
i
m
a
ge
,
w
h
i
l
e
e
ros
i
on
i
s
t
he
pr
oc
e
s
s
of
m
o
vi
ng
/
re
duc
i
n
g
pi
xe
l
s
a
t
t
h
e
b
orde
r
of
a
n
obj
e
c
t
M
or
phol
ogi
c
a
l
ope
ra
t
i
ons
i
nc
l
ud
e
e
ros
i
on
′
=
⊝
w
he
r
e
′
i
s
n
e
w
bi
n
a
ry
i
m
a
ge
fro
m
i
ma
g
e
ha
ve
s
t
ru
c
t
uri
ng
e
l
e
m
e
nt
s
,
t
h
e
di
l
a
t
i
on
not
a
t
i
on
of
i
m
a
g
e
f
w
i
t
h
s
t
ru
c
t
ur
i
ng
e
l
e
m
e
nt
s
i
s
′
=
⊕
,
ope
ni
ng
,
a
nd
c
l
os
i
ng
[23]
.
A
t
t
h
i
s
s
t
a
ge
,
t
h
e
d
i
s
pos
a
l
p
roc
e
s
s
i
s
c
a
rri
e
d
out
[2
4]
i
m
p
e
rf
e
c
t
p
i
xe
l
obj
e
c
t
s
fro
m
t
he
d
e
t
e
c
t
i
on
re
s
u
l
t
.
3.
R
ES
U
LTS
A
N
D
A
N
A
LY
S
I
S
In
t
h
i
s
s
e
c
t
i
o
n,
t
he
re
s
u
l
t
s
of
t
h
e
e
xpe
r
i
m
e
n
t
a
l
a
l
gori
t
h
m
t
ha
t
h
a
v
e
b
e
e
n
a
pp
l
i
e
d
t
o
t
h
e
t
hre
s
h
ol
d
of
m
ovi
ng
ob
j
e
c
t
s
f
or
o
bj
e
c
t
d
e
t
e
c
t
i
on
a
r
e
pr
e
s
e
n
t
e
d
i
n
F
i
gure
2
.
T
h
e
d
a
t
a
s
e
t
us
e
d
c
ons
i
s
t
s
of
200
f
ra
m
e
s
e
que
n
c
e
s
.
I
n
or
de
r
t
o
j
udge
m
e
nt
our
a
d
a
pt
i
v
e
t
hre
s
ho
l
d
pra
c
t
i
c
e
,
w
e
us
e
a
d
a
t
a
s
e
t
i
n
r
e
a
l
i
s
t
i
c
w
i
t
h
pe
r
m
i
s
s
i
o
n
t
o
L
e
2i
“
L
a
bor
a
t
or
i
e
E
l
e
t
ro
ni
qu
e
,
I
nfor
m
a
t
i
qu
e
e
t
Im
a
ge
”
vi
d
e
o
s
urve
i
l
l
a
n
c
e
s
t
a
t
e
by
us
i
ng
a
n
u
na
c
c
o
m
p
a
ni
e
d
c
a
m
e
r
a
[2
5
]
.
T
h
e
v
i
de
o
h
a
s
fra
m
e
ra
t
e
i
s
2
5
p
e
r
s
e
c
ond
,
a
nd
w
i
t
h
re
s
o
l
ut
i
on
o
n
320
x
2
40
p
i
xe
l
s
.
T
h
e
v
i
de
o
da
t
a
e
xp
l
a
i
ne
d
w
i
t
h
pr
i
nc
i
p
a
l
a
d
ve
rs
i
t
y
of
r
e
a
s
on
a
bl
e
fra
m
e
s
e
que
n
c
e
s
w
hi
c
h
c
a
n
c
om
e
a
c
ros
s
a
t
a
h
um
a
n
a
s
w
e
l
l
a
s
i
n
a
s
t
r
a
i
gh
t
forw
a
rd
o
ffi
c
e
roo
m
.
T
h
e
fr
a
m
e
s
e
que
n
c
e
s
c
om
pri
s
e
fl
u
c
t
ua
t
i
ng
of
i
l
l
um
i
na
t
i
on
,
a
nd
di
s
t
i
nc
t
i
v
e
ha
r
ds
hi
p
a
s
t
h
ough
oc
c
l
us
i
ons
or
c
on
found
a
nd
t
e
x
t
ure
d
b
a
c
kgroun
d.
T
he
a
c
t
ors
br
i
ng
a
bo
ut
c
e
r
t
a
i
n
nor
m
a
l
ha
bi
t
ua
l
a
c
t
i
vi
t
i
e
s
.
T
h
e
d
a
t
a
s
e
t
c
o
m
pri
s
e
s
nu
m
be
r
of
fra
m
e
s
.
F
or
j
udg
e
m
e
nt
g
oa
l
,
w
i
t
h
e
xt
r
a
i
nfor
m
a
t
i
on
r
e
pre
s
e
nt
i
ng
t
h
e
gr
ound
-
t
rut
h
of
t
he
de
t
e
c
t
i
on
po
s
i
t
i
on
i
n
t
he
i
m
a
g
e
s
e
qu
e
nc
e
.
T
he
n
,
e
a
c
h
fr
a
m
e
o
f
e
a
c
h
vi
d
e
o
i
s
l
a
b
e
l
l
e
d
:
t
h
e
l
oc
a
l
i
z
a
t
i
on
of
t
h
e
b
ody
i
s
a
u
t
om
a
t
i
c
a
s
s
i
gn
e
d
w
i
t
h
bou
ndi
n
g
b
oxe
s
.
T
he
e
xpe
ri
m
e
n
t
us
e
s
M
A
T
L
A
B
2016b
a
nd
runs
o
n
a
CP
U
w
i
t
h
a
n
i
5
-
8
265U
proc
e
s
s
or
w
i
t
h
RA
M
L
P
D
D
R3
8
G
B
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
1693
-
6
930
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
,
V
ol
.
18
,
N
o.
2
,
A
pri
l
2
020:
1122
-
1129
1126
F
or
t
h
e
d
e
t
a
i
l
e
d
e
x
p
l
a
n
a
t
i
on
,
w
e
a
ppl
i
e
d
our
a
pp
roa
c
h
i
n
F
i
gure
2,
w
he
r
e
a
s
w
e
c
om
pa
r
e
d
ga
us
s
i
a
n
m
i
x
t
ure
m
ode
l
w
i
t
h
O
t
s
u
a
nd
g
ra
y
t
hr
e
s
hol
d
.
F
or
s
a
ke
t
he
s
i
m
pl
i
c
i
t
y
,
w
e
t
a
k
e
e
xa
m
pl
e
a
s
fo
l
l
ow
s
:
i
n
t
h
e
f
ra
m
e
num
be
r
90,
t
h
e
s
ub
j
e
c
t
a
rri
ve
d
a
t
roo
m
,
t
h
us
t
he
bou
ndi
n
g
box
s
t
a
r
t
t
o
d
e
t
e
c
t
t
he
s
ub
j
e
c
t
,
t
he
n
fol
l
ow
e
d
by
fr
a
m
e
nu
m
be
r
1
02
a
nd
f
ra
m
e
nu
m
b
e
r
168
.
In
t
he
fra
m
e
n
um
b
e
r
168
a
nd
179
,
w
e
c
a
n
s
e
e
t
h
a
t
ga
us
s
i
a
n
m
i
x
t
ur
e
m
od
e
l
a
b
l
e
t
o
m
a
i
nt
a
i
n
t
h
e
bound
i
ng
bo
x
t
o
t
he
s
ubj
e
c
t
,
w
he
re
t
h
e
o
t
h
e
r
m
e
t
hod
i
s
not
s
uc
c
e
s
s
fu
l
.
T
h
us
,
f
rom
vi
s
u
a
l
e
v
a
l
u
a
t
i
on
,
w
e
c
a
n
c
on
c
l
ud
e
t
h
a
t
t
he
ga
us
s
i
a
n
m
i
xt
u
re
m
od
e
l
i
s
r
obus
t
t
h
a
n
o
t
he
r
m
e
t
hods
.
T
he
fo
l
l
ow
i
ng
w
e
re
s
how
n
t
h
e
re
s
u
l
t
i
ng
f
ra
m
e
of
t
h
e
e
xpe
ri
m
e
n
t
:
F
ra
m
e
90
102
168
179
M
e
t
hod
O
ri
gi
n
a
l
F
ra
m
e
G
a
us
s
i
a
n
M
i
xt
ur
e
M
ode
l
O
t
s
u
G
ra
y
T
hre
s
ho
l
d
F
i
gure
2
.
R
e
s
ul
t
o
f
t
he
t
h
re
s
ho
l
d
a
nd
d
e
t
e
c
t
i
on
by
ga
us
s
i
a
n
m
i
xt
ur
e
m
od
e
l
,
O
t
s
u
a
nd
gr
a
y
t
hr
e
s
hol
d
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
A
dapt
i
v
e
t
hr
e
s
h
ol
d
f
or
m
ov
i
ng
obj
e
c
t
s
d
e
t
e
c
t
i
o
n
us
i
ng
gaus
s
i
an
m
i
x
t
ur
e
m
o
de
l
(
M
oc
h
A
r
i
e
f
S
oe
l
e
m
an
)
1127
3.
1
.
Eva
l
u
ati
on
In
t
he
e
x
pe
r
i
m
e
nt
s
c
a
rr
i
e
d
ou
t
t
o
m
e
a
s
ur
e
t
h
e
p
e
rfo
rm
a
nc
e
of
e
a
c
h
a
d
a
pt
i
ve
t
hre
s
ho
l
d
m
e
t
ho
d
t
ha
t
ha
s
be
e
n
a
ppl
i
e
d
fo
r
t
he
d
e
t
e
c
t
i
on
of
m
ov
i
ng
obj
e
c
t
s
,
i
n
c
l
udi
n
g
t
he
ga
us
s
i
a
n
m
i
x
t
ure
m
od
e
l
,
t
he
O
t
s
u
a
l
go
ri
t
hm
a
nd
t
he
g
ra
y
t
hr
e
s
hol
d
,
M
S
E
a
nd
P
S
N
R
a
r
e
us
e
d
.
E
va
l
ua
t
i
o
n
of
t
h
e
t
hre
s
ho
l
d
r
e
s
ul
t
s
i
n
t
he
de
t
e
c
t
i
on
m
e
t
hod
i
s
us
e
d
pe
a
k
s
i
g
na
l
t
o
noi
s
e
r
a
t
i
o
(P
S
N
R)
[2
6]
w
h
i
c
h
i
s
a
m
e
t
hod
of
c
o
m
pa
r
i
ng
t
h
e
m
a
x
i
m
a
l
v
a
l
ue
of
t
he
s
t
a
t
e
d
s
i
gna
l
w
i
t
h
t
h
e
qu
a
n
t
i
t
y
of
n
oi
s
e
i
nt
e
ns
i
t
y
o
n
t
h
e
s
i
g
na
l
.
B
e
f
ore
d
e
t
e
rm
i
ni
ng
t
h
e
P
S
N
R
v
a
l
u
e
,
fi
rs
t
,
c
a
l
c
u
l
a
t
e
t
he
va
l
ue
[2
7]
.
W
he
r
e
a
s
m
e
a
n
s
qu
a
re
e
rror
(
M
S
E
)
[28]
i
s
t
h
e
a
v
e
ra
ge
s
qua
r
e
e
rro
r
v
a
l
u
e
be
t
w
e
e
n
t
h
e
or
i
gi
na
l
i
m
a
ge
a
n
d
t
he
i
m
pro
ve
d
i
m
a
ge
,
w
h
i
c
h
i
s
for
m
ul
a
t
e
d
i
n
t
h
e
(1
4)
d
a
n
(
15)
.
(
,
)
=
1
∑
∑
[
(
,
)
−
(
,
)
]
=
1
=
1
(14)
(
,
)
=
10
.
10
(
2
(
,
)
)
(15)
T
he
v
a
ri
a
b
l
e
F
i
n
di
c
a
t
e
s
t
h
a
t
t
he
f
ra
m
e
gr
ound
t
rut
h
,
w
hi
l
e
t
h
e
pa
r
a
m
e
t
e
r
X
i
s
a
fra
m
e
s
e
que
nc
e
t
h
a
t
h
a
s
a
s
i
z
e
of
K
x
Q
,
t
h
e
m
a
x
va
l
ue
i
nd
i
c
a
t
e
s
t
h
e
m
a
x
i
m
um
a
m
ount
of
t
h
e
p
i
xe
l
fr
a
m
e
.
T
h
e
r
e
s
ul
t
s
of
t
h
e
m
e
a
n
s
qua
re
e
rro
r
a
nd
P
S
N
R
v
a
l
u
e
s
of
t
h
e
g
a
us
s
i
a
n
m
i
xt
ure
m
ode
l
a
nd
t
he
O
t
s
u
t
e
c
hni
que
c
a
n
b
e
s
e
e
n
i
n
T
a
b
l
e
1
.
F
rom
T
a
b
l
e
1
s
how
s
t
h
a
t
t
he
M
S
E
G
M
M
va
l
ue
i
s
l
ow
e
r
t
ha
n
O
t
s
u
a
nd
gr
a
y,
a
nd
t
h
e
P
S
N
R
G
M
M
s
how
s
hi
gh
e
r
t
h
a
n
t
h
e
O
t
s
u
a
nd
G
r
a
y
A
l
go
ri
t
hm
.
T
a
b
l
e
1
.
A
v
e
ra
g
e
pe
r
form
a
n
c
e
e
v
a
l
u
a
t
i
on
a
d
a
p
t
i
v
e
t
hre
s
h
ol
d
w
i
t
h
M
S
E
a
nd
P
S
N
R
(
dB)
E
v
a
l
u
a
t
i
o
n
G
M
M
T
h
re
s
h
o
l
d
O
t
s
u
T
h
re
s
h
o
l
d
G
ra
y
T
h
re
s
h
o
l
d
M
S
E
2
5
7
.
1
8
5
9
5
.
3
6
6
4
5
.
3
9
P
S
N
R
2
4
.
7
1
2
0
.
6
6
1
9
.
3
5
3.
2
.
Exp
e
r
i
me
n
ta
l
r
e
s
u
l
ts
T
he
re
s
u
l
t
of
t
hr
e
s
hol
d
i
ng
i
n
hu
m
a
n
m
ovi
ng
i
n
v
i
d
e
o
d
a
t
a
s
e
t
,
F
i
g
ur
e
3
a
nd
F
i
g
ur
e
4
s
how
t
he
M
S
E
a
nd
P
S
N
R
of
a
da
pt
i
ve
t
hre
s
ho
l
d
i
n
hu
m
a
n
de
t
e
c
t
i
on,
s
u
c
c
e
s
s
i
v
e
l
y
.
T
h
e
di
a
gr
a
m
of
r
e
s
ul
t
s
how
t
he
b
e
s
t
a
c
c
om
p
l
i
s
h
m
e
nt
o
f
GMM
(M
S
E
=
89
.
11,
PSNR
=
28
.
63
)
a
t
fra
m
e
33
.
A
ve
r
a
g
e
l
y
,
M
S
E
of
GMM
i
s
25
7.
1
8,
O
t
s
u
i
s
59
5.
36
a
nd
gr
a
y
t
hr
e
s
hol
d
i
s
64
5,
3
8.
A
l
s
o
,
P
S
N
R
of
GMM
i
s
24.
71,
O
t
s
u
i
s
20
.
66
a
nd
gr
a
y
t
hr
e
s
hol
d
i
s
19
.
36
.
F
i
gure
3
.
P
e
rfor
m
a
n
c
e
M
S
E
(dB)
of
a
da
p
t
i
v
e
t
hr
e
s
hol
d
F
i
gure
4
.
P
e
rfor
m
a
n
c
e
P
S
N
R
(
dB)
of
a
d
a
pt
i
ve
t
h
re
s
ho
l
d
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
1693
-
6
930
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
,
V
ol
.
18
,
N
o.
2
,
A
pri
l
2
020:
1122
-
1129
1128
Ba
s
e
d
on
t
he
t
e
s
t
r
e
s
ul
t
s
t
h
a
t
t
h
e
l
ow
e
r
t
he
va
l
ue
of
m
e
a
n
s
q
ua
re
e
rror
(M
S
E
),
t
h
e
t
hre
s
h
ol
d
i
ng
p
roc
e
s
s
of
f
i
nd
i
ng
ob
j
e
c
t
s
i
n
m
ov
i
ng
o
n
t
h
e
vi
d
e
o
fr
a
m
e
t
he
b
e
t
t
e
r
t
he
r
e
s
ul
t
s
o
r
ha
v
e
a
l
ow
e
rror
r
a
t
e
,
w
h
e
re
a
s
i
f
t
he
m
a
rk
of
t
h
e
p
e
a
k
s
i
g
n
a
l
t
o
noi
s
e
ra
t
i
o
(P
S
N
R)
i
s
hi
g
he
r
t
he
b
e
t
t
e
r
t
h
e
qu
a
l
i
t
y
t
hr
e
s
hol
d
m
o
vi
ng
obj
e
c
t
s
i
n
t
he
vi
d
e
o
fra
m
e
.
T
h
e
n
i
t
c
a
n
b
e
s
how
n
i
n
F
i
gure
3
gr
a
ph
t
ha
t
s
how
s
t
he
s
uc
c
e
s
s
ra
t
e
of
t
h
e
m
ovi
ng
obj
e
c
t
t
hre
s
ho
l
d
pro
c
e
s
s
.
3.
3
.
D
i
s
c
u
s
s
i
on
T
he
pos
s
i
b
l
e
r
e
a
s
on
O
t
s
u
t
hre
s
ho
l
d
m
e
t
h
od
i
s
u
nde
r
pe
rf
orm
e
d
c
o
m
p
a
re
d
w
i
t
h
G
M
M
a
re
t
he
O
t
s
u
do
e
s
not
us
e
s
pa
t
i
a
l
c
ohe
re
n
c
e
a
nd
a
ny
obj
e
c
t
s
t
ru
c
t
ur
e
.
A
l
s
o,
t
he
O
t
s
u
t
hre
s
h
ol
d
s
t
i
l
l
us
e
a
s
s
u
m
pt
i
on
i
n
us
i
ng
bi
na
ry
c
l
a
s
s
,
w
h
i
c
h
pa
r
t
i
t
i
on
t
h
e
gra
ys
c
a
l
e
hi
s
t
o
gra
m
v
a
l
ue
i
n
t
o
b
i
na
ry
c
l
a
s
s
.
D
u
e
t
o
da
t
a
s
e
t
w
h
i
c
h
c
a
n
be
s
e
e
n
a
t
F
i
gure
3
,
i
t
s
how
s
t
ha
t
t
h
e
v
i
d
e
o
c
on
t
a
i
ns
m
or
e
t
ha
n
t
w
o
s
e
g
m
e
n
t
c
l
a
s
s
.
T
h
us
,
i
n
fr
a
m
e
num
b
e
r
179
,
t
h
e
O
t
s
u
t
hre
s
ho
l
d
s
e
g
m
e
n
t
t
he
b
e
d
a
nd
p
e
rs
on
i
nt
o
on
e
due
t
o
c
ol
o
r
o
f
be
d
a
nd
c
o
l
or
of
c
l
ot
h
e
s
h
a
ve
s
a
m
e
gr
a
y
c
o
l
or
.
T
he
pos
s
i
b
l
e
re
a
s
on
g
ra
y
t
h
re
s
ho
l
d
i
s
und
e
rp
e
rfor
m
e
d
c
om
p
a
re
d
w
i
t
h
G
M
M
a
r
e
t
he
G
ra
y
t
hre
s
ho
l
d
fa
i
l
e
d
t
o
de
t
e
c
t
non
-
e
dg
e
p
i
xe
l
s
.
A
s
w
e
c
a
n
s
e
e
a
t
F
i
gur
e
3
,
t
he
G
r
a
y
o
nl
y
s
e
g
m
e
n
t
t
he
pa
n
t
s
du
e
t
o
i
t
c
on
t
a
i
ns
m
ore
v
a
l
u
e
t
h
a
n
t
h
e
c
l
ot
h
e
s
w
hi
c
h
m
or
e
s
a
m
e
l
i
ke
t
he
fo
re
gr
o
und.
A
l
t
houg
h
G
M
M
pos
s
e
s
s
s
a
m
e
c
ha
r
a
c
t
e
r
i
s
t
i
c
l
i
k
e
O
t
s
u
t
hre
s
ho
l
d
,
w
h
i
c
h
us
e
s
h
i
s
t
ogr
a
m
-
ba
s
e
d
t
e
c
hn
i
qu
e
s
,
but
t
he
G
M
M
i
t
s
e
l
f
a
s
s
um
e
s
t
he
h
i
s
t
ogr
a
m
di
s
t
ri
but
i
on
i
s
re
pre
s
e
nt
e
d
by
t
w
o
g
a
us
s
i
a
n
c
u
rve
s
,
w
hi
c
h
a
bl
e
t
o
fi
nd
i
ng
a
t
h
r
e
s
hol
d
a
nd
d
e
s
c
ri
be
t
he
t
w
o
re
gi
o
ns
ob
t
a
i
ne
d
.
4.
C
O
N
C
LU
S
I
O
N
T
hi
s
p
a
pe
r
pr
ov
i
de
d
a
n
a
da
pt
i
ve
t
hre
s
h
ol
d
t
e
c
hno
l
ogy
on
a
s
s
i
gn
t
h
e
t
hr
e
s
ho
l
d
of
ob
j
e
c
t
m
ov
i
ng
fo
r
e
s
pe
c
i
a
l
l
y
i
n
hum
a
n
d
e
t
e
c
t
i
on
.
T
he
out
c
o
m
e
s
how
e
d
t
ha
t
t
he
a
c
h
i
e
ve
m
e
n
t
of
a
n
o
bj
e
c
t
m
o
vi
ng
i
n
v
i
d
e
o
de
t
e
c
t
i
on
us
i
ng
a
g
a
us
s
i
a
n
m
i
x
t
ure
m
o
de
l
i
s
b
e
t
t
e
r
t
h
a
n
w
i
t
h
s
t
a
nd
a
rd
o
t
s
u
a
l
go
ri
t
hm
a
nd
gra
y
t
hr
e
s
ho
l
d
.
T
he
a
v
e
ra
ge
of
M
S
E
va
l
ue
s
i
s
257
.
18
dB
s
m
a
l
l
e
r
t
ha
n
t
h
e
o
t
he
r
a
nd
t
he
P
S
N
R
va
l
u
e
i
s
24.
7
1
dB
hi
g
he
r
t
h
a
n
O
t
s
u
a
nd
gr
a
y
t
hre
s
ho
l
d.
F
u
t
ur
e
w
ork
i
s
a
n
a
rr
a
ng
e
m
e
n
t
t
o
e
va
l
ua
t
e
o
t
he
r
t
hr
e
s
hol
d
t
e
c
hn
i
que
s
,
t
o
i
m
p
rove
pe
rfor
m
a
nc
e
i
n
t
hr
e
s
hol
d
s
t
r
a
t
e
gy
.
A
C
K
N
O
WL
ED
G
E
M
EN
TS
T
hi
s
r
e
s
e
a
rc
h
w
a
s
f
unde
d
by
a
P
D
U
P
T
Re
s
e
a
r
c
h
G
r
a
nt
num
b
e
r
026
/
L
6/
A
K
/
S
P
2H
/
RE
S
E
A
RCH
/
2
019
,
from
t
h
e
M
i
ni
s
t
ry
of
R
e
s
e
a
rc
h,
T
e
c
h
nol
o
gy,
a
nd
H
i
gh
e
r
E
du
c
a
t
i
on
,
Re
pu
bl
i
c
of
Indon
e
s
i
a
.
R
EF
ER
EN
C
ES
[
1]
M
.
A
.
K
a
n
da
v
a
l
l
i
,
"
A
no
ve
l
a
ppr
oa
c
h
o
n
o
bj
e
c
t
d
e
t
e
c
t
i
o
n
a
nd
t
r
a
c
k
i
n
g
u
s
i
ng
a
da
pt
i
ve
b
a
c
kg
r
ou
nd
s
ub
t
r
a
c
t
i
on
m
e
t
ho
d,
"
i
n
S
e
c
on
d
I
n
t
e
r
na
t
i
o
nal
C
o
nf
e
r
e
nc
e
on
C
om
p
ut
i
ng
M
e
t
ho
dol
o
gi
e
s
and
C
om
m
u
ni
c
at
i
on
(
I
C
C
M
C
)
,
201
8.
[
2]
X
i
a
o
f
e
ng
L
u
a
nd
C
a
i
di
X
u
,
"
N
ove
l
G
a
us
s
i
a
n
m
i
xt
ur
e
m
ode
l
ba
c
kg
r
oun
d
s
ubt
r
a
c
t
i
on
m
e
t
hod
f
o
r
d
e
t
e
c
t
i
n
g
m
ov
i
ng
obj
e
c
t
,
"
I
n
t
e
r
na
t
i
o
nal
C
o
nf
e
r
e
nc
e
o
f
Saf
e
t
y
pr
o
duc
e
I
nf
or
m
a
t
i
on
(
I
I
C
SP
I
)
,
201
8.
[
3]
A
.
R
a
h
m
a
n
,
M
.
A
hm
e
d
a
n
d
H
o
s
i
a
n
,
"
A
n
a
d
a
p
t
i
v
e
ba
c
kgr
ound
m
od
e
l
i
ng
b
a
s
e
d
on
m
o
di
f
i
e
d
r
u
nn
i
ng
G
a
u
s
s
i
a
n
a
ve
r
a
ge
m
e
t
h
od,
"
I
nt
e
r
n
at
i
on
al
C
o
nf
e
r
e
n
c
e
o
n
E
l
e
c
t
r
i
c
a
l
,
C
om
p
ut
e
r
and
C
om
m
u
ni
c
at
i
o
n
E
ng
i
ne
e
r
i
ng
(
E
C
C
E
)
,
B
angl
ade
s
h
,
2
017
.
[
4]
S
ye
d
T
a
r
i
q
A
l
i
,
K
a
l
pa
n
a
G
oya
l
a
n
d
J
y
ot
i
S
i
ngh
a
i
,
"
M
o
vi
n
g
O
bj
e
c
t
D
e
t
e
c
t
i
on
U
s
i
n
g
S
e
l
f
A
d
a
p
t
i
v
e
G
a
u
s
s
i
a
n
M
i
xt
ur
e
M
od
e
l
f
o
r
R
e
a
l
T
i
m
e
A
pp
l
i
c
a
t
i
o
ns
,
"
P
r
oc
e
e
d
i
ng
I
n
t
e
r
na
t
i
ona
l
c
on
f
e
r
e
nc
e
on
R
e
c
e
nt
I
nno
v
at
i
on
s
i
s
S
i
gn
al
P
r
o
c
e
s
s
i
ng
and
E
m
be
d
de
d
Sy
s
t
e
m
s
(
R
I
SE
-
201
7)
,
2017
.
[
5]
Z
.
Z
e
ng
,
J
.
J
i
a
a
n
d
Y
.
C
he
n
,
“
P
i
xe
l
m
od
e
l
i
ng
us
i
ng
hi
s
t
ogr
a
m
s
b
a
s
e
d
on
f
uz
z
y
p
a
r
t
i
t
i
ons
f
o
r
d
yna
m
i
c
ba
c
kgr
ound
s
ubt
r
a
c
t
i
on
,
"
I
E
E
E
T
r
ans
ac
t
i
o
n
I
n
F
uz
z
y
Sy
s
t
e
m
,
vo
l
.
2
5,
no
.
3
,
2
01
6.
[
6]
O
t
s
u
N
,
"
A
t
hr
e
s
h
ol
d
s
e
l
e
c
t
i
on
m
e
t
ho
d
f
r
o
m
gr
a
y
-
l
e
ve
l
hi
s
t
ogr
a
m
,
"
I
E
E
E
T
r
ans
ac
t
.
M
an
a
nd
C
y
b
e
r
n
e
t
,
vo
l
.
9,
no.
1
,
pp.
6
2
-
6
6,
19
79.
[
7]
M
.
S
o
e
l
e
m
a
n
,
M
a
ur
i
dh
y
H
e
r
y
a
nd
H
a
r
i
a
di
M
,
"
A
d
a
pt
i
v
e
T
hr
e
s
ho
l
d
f
or
B
a
c
kg
r
ou
nd
S
ub
t
r
a
c
t
i
o
n
i
n
M
ov
i
ng
O
b
j
e
c
t
D
e
t
e
c
t
i
on
u
s
i
n
g
F
uz
z
y
C
-
M
e
a
n
s
C
l
u
s
t
e
r
i
ng
,
"
I
E
E
E
T
E
N
C
O
N
2012
,
C
e
bu,
P
hi
l
l
i
ph
i
n
e
,
20
12.
[
8]
Y
e
s
ong
,
X
i
a
op
i
ngL
i
,
N
a
F
u
a
nd
Q
i
o
ngx
i
n
L
i
u
,
"
F
a
s
t
M
o
vi
n
g
O
bj
e
c
t
D
e
t
e
c
t
i
on
U
s
i
ng
I
m
p
r
ove
d
G
a
us
s
i
a
n
M
i
x
t
ur
e
M
od
e
l
s
,
"
I
C
A
L
I
P
2014
,
201
4.
[
9]
V
i
ve
k
M
a
i
k
,
H
y
ung
t
a
e
K
i
m
,
D
a
e
dhe
e
K
i
m
,
E
u
nj
u
ng
C
h
a
e
a
nd
J
oo
nki
P
a
i
k,
"
R
o
bus
t
B
a
c
kg
r
ou
nd
G
e
n
e
r
a
t
i
on
U
s
i
ng
a
M
od
i
f
i
e
d
M
i
xt
u
r
e
o
f
G
a
us
s
i
a
n
M
o
de
l
f
or
O
b
j
e
c
t
D
e
t
e
c
t
i
on
,
"
I
E
E
E
I
SC
E
,
201
4.
[
10]
X
ue
hua
S
on
g,
J
i
ng
z
hu
C
he
n
,
C
hon
g
H
e
a
n
d
X
i
a
ng
Z
ho
u,
"
A
R
obu
s
t
M
ov
i
ng
O
bj
e
c
t
s
D
e
t
e
c
t
i
on
B
a
s
e
d
o
n
I
m
pr
o
ve
d
G
a
us
s
i
a
n
M
i
x
t
ur
e
M
ode
l
,
"
I
nt
e
r
n
at
i
on
al
C
o
nf
e
r
e
nc
e
o
n
A
r
t
i
f
i
c
i
al
I
n
t
e
l
l
i
g
e
nc
e
and
C
om
pu
t
at
i
o
nal
I
n
t
e
l
l
i
ge
nc
e
,
20
10
.
[
11]
Y
i
ng
S
hi
,
S
h
u
C
he
ng
,
J
i
e
C
he
n
a
nd
D
i
C
h
e
n
,
"
M
o
vi
n
g
O
b
j
e
c
t
s
D
e
t
e
c
t
i
o
n
by
G
a
u
s
s
i
a
n
M
i
xt
ur
e
M
ode
l
:
a
c
o
m
pa
r
a
t
i
ve
a
na
l
ys
i
s
,
"
2
011
.
[
12]
Y
i
Z
he
ng
a
nd
P
hi
ng
Z
h
e
ng
,
"
H
a
n
d
S
e
gm
e
n
t
a
t
i
o
n
ba
s
e
d
on
I
m
p
r
ove
d
G
a
us
s
i
a
n
M
i
xt
ur
e
M
ode
l
,
"
2015
I
n
t
e
r
na
t
i
ona
l
C
onf
e
r
e
n
c
e
on
C
om
p
ut
e
r
Sc
i
e
n
c
e
and
A
p
pl
i
c
a
t
i
on
s
,
20
15
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
A
dapt
i
v
e
t
hr
e
s
h
ol
d
f
or
m
ov
i
ng
obj
e
c
t
s
d
e
t
e
c
t
i
o
n
us
i
ng
gaus
s
i
an
m
i
x
t
ur
e
m
o
de
l
(
M
oc
h
A
r
i
e
f
S
oe
l
e
m
an
)
1129
[
13]
K
i
t
t
i
pop
P
e
uw
nua
n
,
K
un
t
po
ng
W
or
a
r
a
t
p
a
nya
a
nd
K
i
t
s
uc
,
"
M
od
i
f
i
e
d
a
da
pt
i
ve
t
h
r
e
s
hol
di
ng
us
i
ng
i
n
t
e
g
r
a
l
i
m
a
ge
,
"
13
th
I
n
t
e
r
na
t
i
ona
l
J
o
i
nt
C
o
nf
e
r
e
nc
e
on
C
om
p
ut
e
r
Sc
i
e
n
c
e
an
d
S
of
t
w
ar
e
E
ng
i
ne
e
r
i
ng
(
J
C
SSE
)
,
2
016
.
[
14]
S
e
nt
hi
l
ku
m
a
r
a
n
a
nd
V
a
i
t
he
g
i
S
,
"
I
m
a
ge
S
e
g
m
e
n
t
a
t
i
on
B
y
U
s
i
ng
T
hr
e
s
ho
l
d
i
ng
T
e
c
hn
i
qu
e
s
f
or
M
e
d
i
c
a
l
I
m
a
ge
s
,
"
C
om
put
e
r
Sc
i
e
n
c
e
&
E
ngi
ne
e
r
i
ng:
A
n
I
nt
e
r
n
at
i
on
al
J
o
ur
n
al
(
C
SE
I
J
)
,
vo
l
.
6
,
no
.
1,
pp
.
1
-
12
,
2
016
.
[
15]
K
.
S
r
i
n
i
va
s
a
n
,
K
.
P
o
r
ku
m
a
r
a
n
a
n
d
G
.
S
a
i
na
r
a
y
a
na
n
,
"
I
m
pr
o
ve
d
B
a
c
kgr
ound
S
ub
t
r
a
c
t
i
o
n
T
e
c
hn
i
qu
e
s
f
or
S
e
c
ur
i
t
y
i
n
V
i
de
o
A
p
pl
i
c
a
t
i
on
s
,
"
P
r
o
c
.
o
f
3
rd
I
n
t
e
r
na
t
i
o
na
l
C
onf
e
r
e
nc
e
o
n
A
n
t
i
-
c
o
unt
e
r
f
e
i
t
i
ng
,
Se
c
ur
i
t
y
,
and
I
de
n
t
i
f
i
c
a
t
i
on
i
n
C
om
m
uni
c
at
i
on
,
200
9.
[
16]
M
a
r
w
a
a
bd
e
l
A
z
e
e
m
M
a
r
z
ouk
,
"
M
od
i
f
i
e
d
ba
c
kgr
ound
s
ubt
r
a
c
t
i
o
n
a
l
go
r
i
t
h
m
f
o
r
m
ot
i
on
de
t
e
c
t
i
on
i
n
s
u
r
ve
i
l
l
a
n
c
e
s
ys
t
e
m
s
,
"
J
our
na
l
o
f
A
m
e
r
i
c
an
A
r
a
bi
c
A
c
ade
m
y
f
o
r
S
c
i
e
nc
e
and
T
e
c
hno
l
og
y
,
vo
l
.
1
,
no
.
2,
pp
.
112
-
12
3,
20
10
.
[
17]
Y
.
Z
a
n
d
F
.
L
i
a
ngz
hong
,
"
M
ovi
ng
ob
j
e
c
t
de
t
e
c
t
i
on
b
a
s
e
d
on
r
unn
i
n
g
a
ve
r
a
g
e
ba
c
kg
r
ou
nd
a
n
d
t
e
m
po
r
a
l
di
f
f
e
r
e
nc
e
,
"
H
a
ngz
h
ou
,
2
010
.
[
18]
M
.
A
.
N
ur
a
nd
A
.
Z
.
M
oh
d
,
"
M
ov
i
ng
o
bj
e
c
t
de
t
e
c
t
i
on
v
i
a
T
V
-
L
1
o
pt
i
c
a
l
f
l
ow
i
n
f
a
l
l
-
dow
n
v
i
d
e
os
,
"
B
ul
l
e
t
i
n
o
f
E
l
e
c
t
r
i
c
al
E
ng
i
n
e
e
r
i
ng
a
nd
I
n
f
or
m
at
i
c
s
,
vo
l
.
8
,
no
.
3,
p
.
839
-
846
,
2
019.
[
19]
C
.
S
t
a
u
f
f
e
r
a
nd
W
.
E
.
G
r
i
m
s
on
,
"
A
d
a
pt
i
v
e
b
a
c
k
gr
o
und
m
i
x
t
u
r
e
m
od
e
l
s
f
o
r
r
e
a
l
-
t
i
m
e
t
r
a
c
k
i
ng
,
"
I
E
E
E
C
om
pu
t
e
r
Soc
i
e
t
y
C
onf
e
r
e
nc
e
o
n
C
om
pu
t
e
r
V
i
s
i
o
n
an
d
P
a
t
e
r
n
R
e
c
ogn
i
t
i
o
n
,
2
007.
[
20]
Q
I
A
O
B
a
om
i
ng
,
J
i
n
g
L
i
u
a
nd
Y
A
N
G
Y
uf
a
n
,
"
A
n
A
d
a
p
t
i
v
e
A
l
go
r
i
t
h
m
f
o
r
G
r
e
y
I
m
a
ge
E
dge
D
e
t
e
c
t
i
on
B
a
s
e
d
o
n
G
r
e
y
C
or
r
e
l
a
t
i
on
A
na
l
ys
i
s
,
"
1
2
th
I
nt
e
r
n
at
i
on
al
C
on
f
e
r
e
n
c
e
on
C
om
p
ut
a
t
i
ona
l
I
n
t
e
l
l
i
ge
nc
e
an
d
Se
c
u
r
i
t
y
I
E
E
E
,
2
016
.
[
21]
S
.
O
.
I
m
o
c
ha
a
n
d
S
.
T
e
j
m
a
n
i
,
"
L
oc
a
l
C
o
nt
r
a
s
t
a
nd
M
e
a
n
ba
s
e
d
T
hr
e
s
ho
l
di
ng
T
e
c
hn
i
que
i
n
I
m
a
g
e
B
i
n
a
r
i
z
a
t
i
on
,
"
I
nt
e
r
n
at
i
on
al
J
o
ur
n
al
o
f
C
om
put
e
r
A
pp
l
i
c
a
t
i
ons
,
vo
l
.
51
,
no
.
6,
20
1
2.
[
22]
S
.
C
.
N
a
g
a
r
a
j
u,
G
.
N
a
ga
m
a
ni
a
nd
P
.
G
.
R
a
ke
s
h
,
"
M
o
r
pho
l
ogi
c
a
l
E
dge
D
e
t
e
c
t
i
on
A
l
gor
i
t
h
m
B
a
s
e
d
on
M
ul
t
i
-
S
t
r
uc
t
ur
e
E
l
e
m
e
n
t
s
o
f
D
i
f
f
e
r
e
nt
D
i
r
e
c
t
i
o
ns
,
"
I
n
t
e
r
nat
i
o
nal
J
our
n
al
o
f
I
nf
or
m
a
t
i
on
and
C
om
m
un
i
c
a
t
i
on
T
e
c
hn
ol
o
gy
R
e
s
e
a
r
c
h
,
vo
l
.
1
,
no.
1
,
201
2.
[
23]
A
.
A
m
e
r
,
"
N
e
w
b
i
n
a
r
y
m
or
p
hol
og
i
c
a
l
o
pe
r
a
t
i
on
s
f
or
e
f
e
c
t
i
v
e
l
ow
-
c
os
t
boun
da
r
y
d
e
t
e
c
t
i
o
n,
"
I
nt
e
r
n
at
i
on
al
J
ou
r
na
l
o
f
Pa
t
t
e
r
n
R
e
c
og
ni
t
i
on
and
A
r
t
i
t
c
i
a
l
I
n
t
e
l
l
i
g
e
nc
e
,
vo
l
.
17,
n
o.
2
,
2002
.
[
24]
K
.
M
a
he
s
h
a
nd
K
.
K
uppu
s
a
m
y
,
"
A
N
e
w
H
yb
r
i
d
V
i
d
e
o
S
e
g
m
e
nt
a
t
i
on
A
l
go
r
i
t
h
m
u
s
i
n
g
F
uz
z
y
C
M
e
a
ns
C
l
us
t
e
r
i
ng
,
"
I
nt
e
r
n
at
i
on
al
J
o
ur
n
al
o
f
C
om
put
e
r
Sc
i
e
nc
e
,
vo
l
.
2
,
p.
9
,
201
2.
[
25]
C
ha
r
f
i
,
J
.
M
i
t
È
r
a
n,
J
.
D
ub
oi
s
,
M
.
A
t
r
i
,
R
.
T
ou
r
k
i
,
"
O
p
t
i
m
i
s
e
d
s
pa
t
i
o
-
t
e
m
por
a
l
de
s
c
r
i
p
t
o
r
s
f
o
r
r
e
a
l
-
t
i
m
e
f
a
l
l
de
t
e
c
t
i
on:
c
om
pa
r
i
s
on
of
S
V
M
a
nd
A
da
boo
s
t
ba
s
e
d
c
l
a
s
s
i
f
i
c
a
t
i
on
î
,
”
J
ou
r
na
l
o
f
E
l
e
c
t
r
on
i
c
I
m
a
gi
n
g
(
J
E
I
)
,
v
o
l
.
22
,
no
.
4,
20
13
.
[
26]
A
.
Z
.
S
e
y
e
d
M
oh
a
m
m
a
d
,
J
u
n
Z
ho
u,
X
i
n
L
i
u
a
nd
K
.
T
o
ny
T
a
e
-
H
you
ng,
"
A
n
A
r
e
a
-
a
nd
E
n
e
r
gy
-
E
f
f
i
c
i
e
nt
F
I
F
O
D
e
s
i
gn
U
s
i
ng
E
r
r
or
-
R
e
du
c
e
d
D
a
t
a
C
o
m
p
r
e
s
s
i
on
a
nd
N
e
a
r
-
T
hr
e
s
ho
l
d
O
pe
r
a
t
i
on
f
or
I
m
a
ge
/
V
i
de
o
A
pp
l
i
c
a
t
i
ons
,
"
I
E
E
E
T
r
ans
ac
t
i
ons
on
V
e
r
y
L
a
r
ge
Sc
al
e
I
n
t
e
g
r
at
i
o
n
(
V
L
S
I
)
Sy
s
t
e
m
s
,
vol
.
23,
no
.
11
,
pp.
2
408
-
24
16,
2
015
.
[
27]
K
.
B
ho
ya
r
a
n
d
O
.
K
a
kd
e
,
"
C
ol
or
i
m
a
g
e
s
e
g
m
e
n
t
a
t
i
on
ba
s
e
d
on
j
nd
c
ol
o
r
h
i
s
t
o
gr
a
m
,
"
I
nt
e
r
n
at
i
ona
J
o
ur
na
l
of
I
m
ag
e
P
r
oc
e
s
s
i
ng
(
I
J
I
P
)
,
vo
l
.
3
,
no
.
6,
pp
.
282
-
29
3,
20
10
.
[
28]
W
a
ng
,
Z
,
B
o
vi
k
,
A
.
C
.
,
S
h
e
i
k
h,
H
.
R
a
n
d
S
i
m
oc
e
l
l
i
,
"
I
m
a
g
e
Q
ua
l
i
t
y
A
s
s
e
s
s
m
e
n
t
:
F
r
o
m
E
r
r
or
V
i
s
i
b
i
l
i
t
y
t
o
S
t
r
uc
t
ur
a
l
V
i
s
i
bi
l
i
t
y
,
”
I
E
E
E
T
r
an
s
ac
t
i
on
s
o
n
I
m
ag
e
P
r
oc
e
s
s
i
n
g
,
vo
l
.
4
,
pp
.
60
0
-
61
2,
20
04.
Evaluation Warning : The document was created with Spire.PDF for Python.