I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
8
,
No
.
2
,
A
p
r
il
201
8
,
p
p
.
1092
~
1
1
0
1
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
8
i
2
.
p
p
1
0
9
2
-
1101
1092
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
co
r
e
.
co
m/
jo
u
r
n
a
ls
/in
d
ex
.
p
h
p
/
I
JE
C
E
Fra
m
ew
o
rk
f
o
r C
o
ntextua
l Outlier
Identifica
tion usi
ng
M
ultiva
ria
te
Ana
ly
sis
appro
a
ch and Unsu
pe
rv
ised L
ea
rning
M
a
dh
u Cha
nd
ra
G
1
,
Sree
ra
m
a
Red
dy
G
.
M
.
2
1,
2
De
p
t.
o
f
El
e
c
tro
n
ics
&
Co
m
m
u
n
ica
ti
o
n
En
g
in
e
e
ri
n
g
,
C
By
re
g
o
wd
a
In
stit
u
te o
f
T
e
c
h
n
o
lo
g
y
,
Ko
lar,
In
d
ia
1
V
isv
e
sv
a
ra
y
a
T
e
c
h
n
o
lo
g
ica
l
Un
i
v
e
rsit
y
,
Be
la
g
a
v
i,
In
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
2
9
,
2
0
1
7
R
ev
i
s
ed
Mar
2
0
,
2
0
1
8
A
cc
ep
ted
Ma
r
2
8
,
2
0
1
8
M
a
jo
rit
y
o
f
th
e
e
x
isti
n
g
c
o
m
m
e
rc
ial
a
p
p
li
c
a
ti
o
n
f
o
r
v
id
e
o
su
rv
e
il
lan
c
e
s
y
ste
m
o
n
ly
c
a
p
tu
re
s th
e
e
v
e
n
t
f
r
a
m
e
s
w
h
e
re
th
e
a
c
c
u
ra
c
y
le
v
e
l
o
f
c
a
p
tu
re
s is
to
o
p
o
o
r
.
W
e
re
v
ie
w
e
d
th
e
e
x
isti
n
g
s
y
ste
m
to
f
in
d
th
a
t
a
t
p
re
se
n
t
th
e
re
is
n
o
su
c
h
re
se
a
rc
h
tec
h
n
i
q
u
e
th
a
t
o
f
fe
rs
c
o
n
te
x
tu
a
l
-
b
a
se
d
sc
e
n
e
i
d
e
n
ti
f
ica
ti
o
n
o
f
o
u
tl
iers
.
T
h
e
re
f
o
re
,
w
e
p
re
se
n
t
e
d
a
f
ra
m
e
w
o
rk
th
a
t
u
se
s
u
n
su
p
e
rv
ise
d
lea
rn
in
g
a
p
p
ro
a
c
h
to
p
e
rf
o
rm
p
r
e
c
ise
id
e
n
ti
f
ic
a
ti
o
n
o
f
o
u
tl
iers
fo
r
a
g
iv
e
n
v
id
e
o
f
ra
m
e
s
c
o
n
c
e
rn
in
g
th
e
c
o
n
tex
tu
a
l
i
n
f
o
rm
a
ti
o
n
o
f
th
e
s
c
e
n
e
.
T
h
e
p
ro
p
o
s
e
d
sy
ste
m
u
se
s
m
a
tri
x
d
e
c
o
m
p
o
siti
o
n
m
e
th
o
d
u
sin
g
m
u
lt
iv
a
riat
e
a
n
a
ly
sis
to
m
a
in
tain
a
n
e
q
u
i
li
b
ri
u
m
b
e
tt
e
r
f
a
ste
r
re
sp
o
n
se
ti
m
e
a
n
d
h
ig
h
e
r
a
c
c
u
ra
c
y
o
f
th
e
a
b
n
o
rm
a
l
e
v
e
n
t
/o
b
jec
t
d
e
tec
ti
o
n
a
s
a
n
o
u
t
li
e
r.
Us
in
g
a
n
a
n
a
ly
ti
c
a
l
m
e
th
o
d
o
lo
g
y
,
th
e
p
ro
p
o
se
d
s
y
ste
m
b
lo
c
k
in
g
o
p
e
ra
ti
o
n
f
o
ll
o
w
e
d
b
y
sp
a
rsity
to
p
e
rf
o
rm
d
e
tec
ti
o
n
.
T
h
e
stu
d
y
o
u
tco
m
e
sh
o
w
s
th
a
t
p
ro
p
o
se
d
s
y
ste
m
o
ffe
rs
a
n
in
c
re
a
sin
g
le
v
e
l
o
f
a
c
c
u
ra
c
y
in
c
o
n
tras
t
to
th
e
e
x
isti
n
g
s
y
ste
m
w
it
h
f
a
ste
r
re
sp
o
n
se
ti
m
e
.
K
ey
w
o
r
d
:
A
cc
u
r
ac
y
I
d
en
tif
icat
io
n
Mu
lti
v
ar
iate
an
a
l
y
s
is
Ou
tlier
Scen
e
co
n
te
x
t
Sp
ar
s
it
y
Co
p
y
rig
h
t
©
2
0
1
8
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Ma
d
h
u
C
h
a
n
d
r
a
G
,
Dep
t.
o
f
E
lectr
o
n
i
cs &
C
o
m
m
u
n
ica
tio
n
E
n
g
i
n
ee
r
i
n
g
,
C
B
y
r
eg
o
w
d
a
I
n
s
t
itu
te
o
f
T
ec
h
n
o
lo
g
y
,
Ko
lar
,
I
n
d
ia
.
E
m
ail:
m
ad
h
u
.
g
u
r
u
1
9
8
4
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
u
s
ag
e
an
d
tec
h
n
ica
l
ad
o
p
t
io
n
o
f
th
e
v
id
eo
s
u
r
v
ei
llan
ce
s
y
s
te
m
h
a
s
b
ee
n
in
cr
ea
s
in
g
i
n
f
aster
p
ac
e
o
w
i
n
g
to
t
h
e
i
n
cr
ea
s
i
n
g
s
ec
u
r
it
y
co
n
ce
r
n
s
[
1
]
.
At
p
r
ese
n
t,
t
h
er
e
ar
e
v
ar
io
u
s
p
r
o
ce
s
s
in
g
t
ec
h
n
iq
u
es
o
f
v
id
eo
th
at
h
a
s
s
ig
n
i
f
ica
n
tl
y
b
en
e
f
ite
d
th
e
co
m
p
u
ter
v
i
s
io
n
s
tr
ate
g
ies
to
a
g
r
ea
t
ex
ten
t
[
2
]
,
[
3
]
.
A
lt
h
o
u
g
h
,
e
x
is
tin
g
s
y
s
te
m
ar
e
ca
p
ab
le
o
f
ca
p
tu
r
in
g
th
e
h
ig
h
d
ef
i
n
itio
n
v
id
eo
as
w
ell
a
s
tr
an
s
m
it
t
h
e
h
i
g
h
-
d
e
f
i
n
itio
n
v
id
eo
f
r
a
m
e
s
o
v
er
w
ir
ele
s
s
l
in
k
s
,
b
u
t
t
h
e
y
a
ls
o
s
u
f
f
er
s
f
r
o
m
s
o
m
e
s
ig
n
i
f
ic
an
t
p
it
f
alls
[
4
]
,
[
5
]
.
T
h
e
m
o
s
t
p
r
o
m
i
n
en
t
is
s
u
e
s
i
n
th
e
e
x
i
s
tin
g
s
y
s
te
m
ar
e
to
p
er
f
o
r
m
id
en
ti
f
ica
tio
n
o
f
th
e
o
u
tli
er
s
th
a
t
m
a
y
co
m
e
i
n
d
i
f
f
er
e
n
t
s
h
ap
e
an
d
f
o
r
m
o
f
an
o
b
j
ec
t
p
r
esen
t
in
t
h
e
s
ce
n
e
w
it
h
r
esp
ec
t
to
its
co
n
te
x
t
[
6
]
.
T
h
e
m
ea
n
in
g
o
f
t
h
e
o
u
tlier
p
er
tain
s
to
p
r
esen
ce
o
f
o
b
j
ec
ts
o
r
ev
en
ts
th
a
t
h
as
h
ig
h
l
y
les
s
p
r
o
b
ab
ilit
y
to
o
cc
u
r
w
i
th
r
esp
ec
t
to
th
e
g
i
v
e
n
co
n
tex
t
o
f
t
h
e
s
ce
n
e.
C
o
n
s
tr
u
ctio
n
a
f
r
a
m
e
w
o
r
k
i
n
o
r
d
er
to
p
e
r
f
o
r
m
id
en
tific
atio
n
o
f
s
u
c
h
ev
e
n
t
s
o
r
o
b
j
ec
ts
is
q
u
ite
ch
allen
g
i
n
g
esp
ec
iall
y
co
n
s
id
er
in
g
u
n
s
u
p
er
v
is
ed
m
a
n
n
er
a
n
d
h
e
n
ce
it
h
as
d
r
a
w
n
th
e
a
tten
t
io
n
o
f
t
h
e
r
esear
c
h
co
m
m
u
n
itie
s
.
T
h
e
ex
is
ti
n
g
a
p
p
r
o
ac
h
es
ca
lled
f
o
r
u
s
i
n
g
s
p
atial
an
d
te
m
p
o
r
al
f
ac
to
r
s
[
7
]
,
[
8
]
,
o
p
tical
f
lo
w
[9
-
1
1
]
,
b
ac
k
g
r
o
u
n
d
m
o
d
el
[
1
2
]
,
His
to
g
r
a
m
-
b
ased
[
1
3
]
,
[
1
4
]
,
b
eh
av
io
r
al
-
b
a
s
ed
te
m
p
late
m
atch
i
n
g
[
1
5
]
,
[
1
6
]
,
etc.
Fro
m
p
r
ac
tical
i
m
p
le
m
en
t
atio
n
v
ie
w
p
o
i
n
t,
th
e
e
x
tr
ac
tio
n
o
f
f
ea
tu
r
e
as
w
e
ll
as
co
n
s
tr
u
ctio
n
o
f
f
r
a
m
e
w
o
r
k
is
q
u
ite
s
i
g
n
if
ica
n
t
f
o
r
an
ef
f
ic
ien
t
id
en
ti
f
icat
io
n
o
f
o
u
tlier
s
f
o
r
a
g
iv
en
v
id
eo
f
r
a
m
es.
T
h
er
ef
o
r
e,
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
o
f
f
er
s
a
u
n
iq
u
e
f
r
a
m
e
w
o
r
k
t
h
at
u
ti
lizes
u
n
s
u
p
er
v
is
ed
lear
n
in
g
m
ec
h
an
i
s
m
f
o
r
d
ev
elo
p
in
g
a
n
in
v
o
l
u
n
tar
y
s
y
s
te
m
f
o
r
d
etec
tin
g
ab
n
o
r
m
al
e
v
en
ts
.
W
e
ex
p
lo
r
ed
th
at
th
e
s
et
o
f
f
ea
t
u
r
e
-
b
ased
attr
ib
u
ted
u
tili
ze
d
i
n
ex
i
s
ti
n
g
r
esear
c
h
w
o
r
k
ar
e
n
o
t
ca
p
ab
le
en
o
u
g
h
to
p
er
f
o
r
m
r
ep
r
esen
tatio
n
o
f
co
m
p
le
x
co
n
te
x
t
u
al
b
eh
av
io
u
r
o
f
t
h
e
v
id
eo
f
r
a
m
e
s
.
E
x
is
ti
n
g
s
y
s
te
m
p
er
f
o
r
m
b
etter
ca
p
tu
r
in
g
o
f
i
n
f
o
r
m
atio
n
r
elate
d
to
g
r
ad
ien
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
F
r
a
mewo
r
k
fo
r
C
o
n
textu
a
l O
u
tlier
I
d
en
tifi
ca
tio
n
u
s
in
g
Mu
ltiva
r
ia
te
A
n
a
lysi
s
….
(
Ma
d
h
u
C
h
a
n
d
r
a
G
)
1093
th
at
i
s
b
ett
er
ef
f
ec
ti
v
e
f
o
r
r
esi
s
tin
g
ch
a
n
g
es
a
g
ai
n
s
t
il
lu
m
i
n
at
io
n
an
d
ap
p
ea
r
an
ce
.
Ho
w
e
v
er
,
w
e
h
av
e
a
ls
o
s
ee
n
th
at
v
id
eo
f
r
a
m
es
ar
e
m
o
r
e
li
k
el
y
e
x
h
ib
it
t
h
e
ch
ar
ac
ter
i
s
tics
o
f
g
r
ad
ie
n
ts
a
n
d
h
en
ce
s
u
c
h
e
x
is
t
in
g
s
y
s
te
m
m
a
y
ig
n
o
r
e
a
v
o
l
u
m
i
n
o
u
s
a
m
o
u
n
t
o
f
s
i
g
n
i
f
ica
n
t
co
n
te
x
t
u
al
i
n
f
o
r
m
atio
n
.
T
h
e
ex
is
ti
n
g
s
y
s
te
m
s
ar
e
also
f
o
u
n
d
to
h
av
e
less
er
d
ep
en
d
en
c
ies
o
n
d
ata
in
s
u
ch
a
m
a
n
n
er
th
at
it
is
n
o
t
f
ea
s
ib
le
to
u
s
e
t
h
e
i
n
f
o
r
m
atio
n
s
p
ec
if
ic
to
p
ar
ticu
lar
tas
k
in
th
e
d
ata
s
et.
T
h
e
p
r
o
p
o
s
ed
s
tu
d
y
is
h
i
g
h
l
y
m
o
tiv
a
ted
b
y
t
h
e
ad
v
a
n
ce
m
e
n
t
in
u
s
a
g
e
o
f
f
ea
t
u
r
e
to
w
ar
d
s
id
en
ti
f
icat
io
n
p
r
o
b
lem
s
[
1
7
]
,
[
1
8
]
an
d
th
er
eb
y
p
r
esen
t
s
a
f
r
a
m
e
w
o
r
k
th
a
t
u
s
es
m
atr
i
x
d
ec
o
m
p
o
s
i
tio
n
p
r
in
cip
al
in
o
r
d
er
to
o
p
tim
ize
th
e
lear
n
in
g
p
r
o
ce
s
s
o
f
v
id
e
o
s
.
W
e
f
in
d
t
h
at
o
u
r
p
r
o
p
o
s
ed
s
y
s
te
m
i
s
h
ig
h
l
y
ca
p
ab
le
o
f
ca
p
tu
r
in
g
m
o
r
e
r
el
ev
an
t
i
n
f
o
r
m
atio
n
w
it
h
h
ig
h
er
r
an
g
e
o
f
co
m
p
l
icatio
n
a
n
d
t
h
er
ef
o
r
e
ca
n
h
ar
n
es
s
lo
ts
o
f
tas
k
-
r
elate
d
i
n
f
o
r
m
ati
o
n
p
r
esen
t
i
n
t
h
e
d
atase
t.
T
h
is
m
ec
h
a
n
is
m
is
u
s
ed
f
o
r
ex
tr
ac
ti
n
g
f
ea
t
u
r
es.
T
h
er
ef
o
r
e,
in
th
at
co
n
te
x
t,
it
c
an
b
e
s
aid
th
at
p
r
o
p
o
s
ed
m
ec
h
an
is
m
ca
n
o
f
f
er
b
etter
p
er
f
o
r
m
an
ce
in
co
n
tr
ast
to
ex
is
t
in
g
s
y
s
te
m
.
A
n
o
t
h
er
s
ig
n
if
ica
n
t
co
n
tr
ib
u
tio
n
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
i
s
its
u
s
ag
e
o
f
p
r
o
b
ab
ilit
y
th
eo
r
y
to
p
er
f
o
r
m
co
m
p
u
tatio
n
o
f
le
v
e
l
o
f
o
u
tlier
s
p
r
ese
n
t
f
r
o
m
th
e
p
ix
el
-
lev
e
ls
u
s
in
g
b
lo
ck
-
b
a
s
ed
tr
an
s
f
o
r
m
a
t
io
n
p
r
o
ce
s
s
.
Fo
r
th
e
p
u
r
p
o
s
e
o
f
r
esis
ti
n
g
d
etec
tio
n
o
f
to
o
m
u
c
h
o
f
lo
ca
l
v
alu
e
s
,
t
h
e
p
r
o
p
o
s
e
d
s
y
s
te
m
ca
r
r
y
o
u
t
ap
p
en
d
in
g
o
f
b
o
t
h
te
m
p
o
r
al
a
s
w
ell
as
s
p
atial
d
ata
t
h
at
b
ea
r
s
m
o
r
e
co
n
tex
tu
al
i
n
f
o
r
m
atio
n
.
I
t
i
s
to
b
e
n
o
ted
th
at
p
r
o
p
o
s
ed
s
y
s
te
m
u
s
es
u
n
s
u
p
er
v
i
s
ed
lear
n
i
n
g
ap
p
r
o
ac
h
th
at
is
co
m
p
letel
y
f
r
ee
f
r
o
m
a
n
y
f
o
r
m
o
f
h
u
m
a
n
in
ter
v
e
n
tio
n
w
it
h
r
esp
ec
t
to
b
o
th
f
ea
t
u
r
e
-
b
a
s
ed
lear
n
i
n
g
a
s
w
ell
a
s
f
r
a
m
e
w
o
r
k
-
b
ased
l
ea
r
n
in
g
.
T
h
e
s
t
u
d
y
o
u
tco
m
e
s
h
o
w
s
b
etter
p
er
f
o
r
m
an
ce
w
it
h
r
esp
ec
t
to
ex
is
tin
g
s
y
s
te
m
an
d
o
f
f
er
s
b
etter
co
m
p
u
ta
tio
n
a
l
p
er
f
o
r
m
a
n
ce
w
h
ile
p
er
f
o
r
m
i
n
g
th
e
p
r
o
ce
s
s
o
f
id
en
ti
f
ica
tio
n
.
Sectio
n
1
.
1
d
is
cu
s
s
e
s
ab
o
u
t
th
e
ex
i
s
ti
n
g
liter
atu
r
es
w
h
er
e
d
if
f
er
e
n
t
tec
h
n
iq
u
es
ar
e
d
is
cu
s
s
ed
f
o
r
d
etec
tio
n
s
ch
e
m
e
s
u
s
ed
in
o
u
tl
ier
lo
ca
lizatio
n
in
v
id
eo
s
u
r
v
eilla
n
ce
s
y
s
te
m
f
o
llo
w
ed
b
y
d
is
c
u
s
s
io
n
o
f
r
esear
c
h
p
r
o
b
le
m
s
i
n
Sectio
n
1
.
2
an
d
p
r
o
p
o
s
ed
s
o
lu
tio
n
i
n
1
.
3
.
Sectio
n
2
d
is
cu
s
s
e
s
ab
o
u
t
al
g
o
r
ith
m
i
m
p
le
m
en
tatio
n
f
o
r
ac
co
m
p
li
s
h
in
g
t
h
e
p
r
o
p
o
s
ed
r
es
ea
r
ch
g
o
als
f
o
llo
w
ed
b
y
d
is
c
u
s
s
io
n
o
f
r
esu
lt
an
al
y
s
i
s
o
b
tain
ed
in
Sect
io
n
3
.
Fin
all
y
,
t
h
e
co
n
cl
u
s
i
v
e
r
e
m
a
r
k
s
ar
e
p
r
o
v
id
ed
in
Sectio
n
4
.
1
.
1
.
B
a
ck
g
ro
un
d
T
h
is
s
ec
tio
n
d
is
c
u
s
s
e
s
th
e
e
x
i
s
tin
g
tec
h
n
iq
u
es
to
w
ar
d
s
th
e
i
d
en
tific
atio
n
o
f
s
ig
n
i
f
ica
n
t
ev
en
ts
i
n
th
e
f
o
r
m
o
f
a
n
o
u
tlier
.
Du
tta
et
al.
[
1
9
]
h
av
e
p
r
esen
ted
a
f
r
a
m
e
w
o
r
k
u
s
in
g
s
p
ar
s
e
co
d
in
g
f
o
r
p
er
f
o
r
m
in
g
s
a
lien
c
y
d
etec
tio
n
as
w
ell
a
s
id
e
n
ti
f
icat
io
n
o
f
o
u
tlier
s
.
Usa
g
e
o
f
s
alie
n
c
y
-
b
ased
ap
p
r
o
ac
h
w
a
s
also
s
ee
n
in
th
e
w
o
r
k
o
f
J
an
g
an
d
P
ar
k
[
2
0
]
to
w
ar
d
s
id
en
ti
f
y
in
g
p
o
th
o
les
f
r
o
m
g
r
a
y
s
ca
le
i
m
ag
e
s
.
W
an
g
et
al.
[
2
1
]
h
av
e
u
s
ed
lo
ca
lized
h
is
to
g
r
a
m
f
o
r
an
a
l
y
zi
n
g
cr
o
w
d
ed
s
ce
n
e
u
s
i
n
g
s
u
p
er
v
i
s
ed
lear
n
in
g
ap
p
r
o
ac
h
.
Z
h
o
u
a
n
d
T
o
r
r
e
[
2
2
]
h
av
e
h
ad
also
ad
o
p
ted
s
p
atial
as
w
ell
as
a
te
m
p
o
r
al
s
ch
e
m
e
f
o
r
an
al
y
z
in
g
h
u
m
an
p
o
s
es
u
s
i
n
g
t
h
r
ee
-
d
i
m
e
n
s
io
n
al
ca
p
tu
r
in
g
m
o
d
el.
Fu
et
al.
[
2
3
]
h
av
e
p
r
esen
ted
a
tech
n
iq
u
e
f
o
r
id
en
tif
icatio
n
o
f
p
o
s
s
ib
le
o
u
tlier
s
f
r
a
m
ed
f
r
o
m
th
e
a
n
n
o
tatio
n
f
r
o
m
t
h
e
v
id
eo
.
L
i
an
d
Ha
u
p
t
[
2
4
]
in
v
es
tig
a
t
e
th
e
p
r
o
b
lem
s
as
s
o
ciate
d
w
it
h
t
h
e
lo
ca
lizi
n
g
t
h
e
o
u
tlier
s
i
n
lar
g
er
s
a
m
p
les
o
f
d
ata
in
f
licted
w
it
h
n
o
i
s
e.
X
u
e
e
t
al.
[
2
5
]
h
av
e
i
n
tr
o
d
u
ce
d
a
te
ch
n
iq
u
e
w
h
er
e
t
h
e
o
u
tlier
’
s
d
etec
tio
n
is
ca
r
r
ied
o
u
t
b
y
e
m
p
h
asizi
n
g
o
n
th
e
e
s
ti
m
atio
n
o
f
f
o
r
eg
r
o
u
n
d
co
n
s
id
er
in
g
t
h
e
s
p
ar
s
it
y
co
n
s
tr
ain
t.
Z
h
o
u
et
al.
[
2
6
]
h
av
e
u
s
ed
a
lo
w
-
r
a
n
k
r
ep
r
esen
tatio
n
f
o
r
id
en
t
if
ica
tio
n
o
f
o
u
tlier
s
o
f
co
n
ti
g
u
o
u
s
t
y
p
e.
Go
p
alan
et
a
l.
[
2
7
]
h
av
e
u
s
ed
lear
n
in
g
-
b
ased
m
eth
o
d
o
lo
g
y
f
o
llo
w
ed
b
y
f
ea
tu
r
e
e
x
t
r
ac
tio
n
f
r
o
m
p
ix
el
h
ier
ar
ch
y
a
n
d
u
s
i
n
g
p
ar
ticle
f
ilter
to
p
er
f
o
r
m
id
en
ti
f
icat
io
n
o
f
t
h
e
o
u
tlier
s
f
r
o
m
t
h
e
tr
af
f
ic
d
ata
co
n
s
id
er
in
g
lan
e
m
ar
k
i
n
g
s
.
A
b
n
o
r
m
al
b
eh
av
io
r
d
etec
tio
n
is
also
in
v
esti
g
ated
o
v
er
a
f
ac
ial
d
ata
b
y
Ya
n
g
a
n
d
B
h
an
u
[
2
8
]
.
Ni
et
a
l.
[
2
9
]
h
a
v
e
u
s
ed
p
r
in
c
i
p
al
co
m
p
o
n
en
t
a
n
al
y
s
i
s
alo
n
g
w
it
h
m
in
i
n
g
-
b
ased
ap
p
r
o
ac
h
i
n
o
r
d
er
to
id
en
ti
f
y
a
p
ec
u
lar
p
atter
n
o
f
ag
e
f
r
o
m
s
o
cial
v
id
eo
s
.
Am
m
ar
an
d
L
a
s
h
k
ar
[
3
0
]
h
a
v
e
p
r
ese
n
ted
a
tech
n
iq
u
e
th
a
t
p
er
f
o
r
m
s
d
iag
n
o
s
i
s
o
f
t
h
e
t
y
p
ical
p
atter
n
o
f
a
s
leep
d
is
ea
s
e
r
i
g
h
t
f
r
o
m
o
p
tical
v
id
eo
f
lo
w
.
I
d
en
ti
f
i
ca
tio
n
o
f
t
h
e
o
u
tlier
w
a
s
al
s
o
ca
r
r
ied
o
u
t
b
y
C
h
o
i
an
d
C
h
o
i
[
3
1
]
f
o
r
as
s
is
t
in
g
i
n
f
ir
e
-
r
esi
s
tiv
e
ap
p
licatio
n
.
Fer
is
et
a
l.
[
3
2
]
h
av
e
p
r
esen
ted
a
co
r
r
ec
tio
n
tech
n
iq
u
e
o
f
l
i
g
h
tn
in
g
co
n
d
itio
n
th
at
s
ig
n
i
f
ica
n
tl
y
ass
is
t
s
i
n
s
h
o
w
-
b
a
s
ed
o
u
tlier
d
etec
tio
n
.
J
a
y
as
u
g
a
n
t
h
i
e
t
al.
[
3
3
]
h
av
e
m
o
d
eled
u
n
if
o
r
m
b
ac
k
g
r
o
u
n
d
u
s
i
n
g
Ga
u
s
s
ia
n
al
g
o
r
ith
m
f
o
llo
w
ed
b
y
s
eg
m
e
n
tatio
n
an
d
u
s
ed
k
-
m
e
an
s
a
lg
o
r
it
h
m
f
o
r
p
er
f
o
r
m
in
g
v
id
eo
s
u
r
v
ei
llan
ce
.
L
iu
et
a
l
.
[
3
4
]
h
av
e
u
s
ed
a
s
p
ar
s
e
co
llab
o
r
ativ
e
m
o
d
el
f
o
r
p
er
f
o
r
m
i
n
g
d
etec
tio
n
o
f
o
u
tl
ier
s
f
r
o
m
a
g
i
v
en
v
id
eo
.
Ma
u
r
y
a
a
n
d
T
o
s
h
n
i
w
a
l
[
3
5
]
h
av
e
u
s
ed
s
u
p
er
v
i
s
ed
lear
n
in
g
alg
o
r
it
h
m
f
o
r
tr
ain
i
n
g
th
e
d
ata
g
at
h
er
ed
f
r
o
m
a
n
u
c
lear
p
o
w
er
p
lan
t
t
o
id
en
ti
f
y
s
et
o
f
o
u
tlier
s
.
T
h
e
w
o
r
k
ca
r
r
ied
o
u
t
b
y
P
an
g
et
al.
[
3
6
]
h
av
e
p
r
esen
ted
a
s
tu
d
y
wh
er
e
th
e
ex
tr
ac
tio
n
o
f
f
ea
t
u
r
es
as
w
e
ll
as
clu
s
ter
in
g
o
f
d
ata
is
ad
o
p
ted
to
p
er
f
o
r
m
d
etec
tio
n
o
f
o
u
tlier
s
f
r
o
m
a
p
u
b
lic
s
ce
n
e
i
m
a
g
es.
Si
m
i
lar
clu
s
ter
in
g
m
e
th
o
d
o
lo
g
y
w
a
s
also
ad
o
p
ted
b
y
P
r
itc
h
et
al.
[
3
7
]
o
n
th
e
v
id
eo
d
ata
to
p
er
f
o
r
m
id
en
ti
f
icatio
n
o
f
ab
n
o
r
m
al
e
v
e
n
ts
.
A
d
o
p
tio
n
o
f
ti
m
e
-
s
er
ie
s
f
o
r
an
al
y
s
is
i
ts
e
f
f
ec
t
o
n
th
e
o
u
tlier
d
etec
tio
n
w
as
s
ee
n
in
th
e
w
o
r
k
o
f
T
en
g
et
al.
[
3
8
]
.
T
h
e
w
o
r
k
o
f
B
a
y
at
et
a
l.
[
39
]
d
is
cu
s
s
ed
t
h
e
d
etec
tio
n
o
f
g
o
al
i
n
s
o
cc
er
b
y
u
s
i
n
g
e
v
en
t
d
etec
tio
n
m
ec
h
a
n
is
m
an
d
ac
h
ie
v
ed
b
etter
a
cc
u
r
ac
y
w
it
h
le
s
s
d
etec
tio
n
f
ail
u
r
e.
A
f
o
r
g
er
y
d
etec
tio
n
m
o
d
el
f
o
r
m
o
b
ile
r
ec
o
r
d
ed
an
d
s
u
r
v
eillan
ce
v
id
eo
s
w
er
e
p
r
esen
ted
in
Sta
f
f
y
et
a
l.
[
40
]
.
T
h
is
m
o
d
el
f
o
u
n
d
ab
le
to
id
en
ti
f
y
t
h
e
ta
m
p
er
in
g
ir
r
esp
ec
ti
v
e
o
f
v
i
d
eo
f
o
r
m
a
t.
T
ed
d
y
et
al.
[
41
]
p
er
f
o
r
m
ed
th
e
p
er
f
o
r
m
a
n
ce
an
al
y
s
is
o
f
au
to
m
atic
n
u
m
b
er
p
late
id
en
ti
f
ica
tio
n
o
v
er
An
d
r
o
id
s
m
ar
tp
h
o
n
e
d
ev
ice
an
d
f
o
u
n
d
ef
f
ec
tiv
e
r
ec
o
g
n
itio
n
at
0
.
9
8
s
p
r
o
ce
s
s
in
g
t
i
m
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
2
,
A
p
r
il
201
8
:
1
0
9
2
–
1101
1094
T
h
er
ef
o
r
e,
it
ca
n
b
e
s
ee
n
t
h
at
th
er
e
h
a
s
b
ee
n
v
ar
io
u
s
f
o
r
m
s
o
f
tech
n
iq
u
e
s
t
h
at
h
as
b
ee
n
ev
o
lv
ed
i
n
m
o
s
t
r
ec
en
t
t
i
m
e
s
f
o
r
s
o
lv
in
g
th
e
p
r
o
b
lem
s
ass
o
ciate
d
w
it
h
o
u
tlier
d
etec
tio
n
.
A
ll
t
h
e
ex
i
s
t
in
g
s
t
u
d
ies
h
a
v
e
a
s
ig
n
i
f
ica
n
t le
v
el
o
f
ad
v
an
ta
g
es
as
w
ell
as c
o
n
tr
ib
u
tio
n
s
.
Ho
wev
er
,
th
e
e
x
is
tin
g
s
tu
d
ie
s
ar
e
also
ass
o
ciate
d
w
it
h
s
ig
n
i
f
ica
n
t
lo
o
p
h
o
les
w
h
ich
a
r
e
r
eq
u
ir
ed
to
b
e
ad
d
r
ess
ed
.
T
h
e
n
ex
t
s
ec
tio
n
b
r
ief
s
ab
o
u
t
p
r
o
b
lem
s
id
en
t
if
ied
f
r
o
m
ex
is
ti
n
g
liter
at
u
r
e.
1
.
2
.
Resea
rc
h P
ro
ble
m
T
h
e
s
ig
n
i
f
ica
n
t r
esear
c
h
p
r
o
b
l
e
m
s
ar
e
as
f
o
llo
w
s
:
a.
T
h
e
ex
is
tin
g
tech
n
iq
u
e
o
f
o
u
tlier
d
etec
tio
n
h
a
s
b
ee
n
co
n
s
tr
u
cted
d
ep
en
d
in
g
o
n
a
p
ar
ticu
lar
p
atter
n
o
f
an
o
b
j
ec
t
w
it
h
o
u
t c
o
n
s
id
er
in
g
t
h
e
ac
tu
a
l c
o
n
te
x
t o
f
th
e
s
ce
n
e.
b.
Usag
e
o
f
s
u
p
er
v
is
ed
lear
n
i
n
g
ap
p
r
o
ac
h
i
n
cr
ea
s
es
th
e
ac
cu
r
ac
y
o
f
t
h
e
id
en
ti
f
icatio
n
o
f
a
n
a
b
n
o
r
m
al
ev
e
n
t b
u
t a
t t
h
e
co
s
t
o
f
co
m
p
u
tat
io
n
al
co
m
p
le
x
it
y
.
c.
Usag
e
o
f
p
r
io
r
in
f
o
r
m
at
io
n
ab
o
u
t
th
e
o
b
j
ec
t
an
d
its
ty
p
es
m
ak
e
s
th
e
e
x
is
ti
n
g
s
y
s
te
m
m
o
r
e
n
ar
r
o
w
ed
to
th
e
s
p
ec
i
f
ic
r
esear
ch
en
v
ir
o
n
m
e
n
t a
n
d
b
ec
a
m
e
i
n
co
m
p
atib
le
w
h
e
n
it c
h
a
n
g
es.
d.
T
h
e
ex
ten
t
o
f
f
al
s
e
p
o
s
iti
v
es
is
m
o
r
e
i
n
th
e
co
n
v
e
n
tio
n
al
tech
n
iq
u
es
e
v
e
n
w
h
er
e
t
h
e
s
p
ar
s
it
y
co
d
in
g
h
a
s
b
ee
n
ca
r
r
ied
o
u
t in
o
r
d
er
p
er
f
o
r
m
o
u
t
lier
d
etec
tio
n
.
T
h
er
ef
o
r
e,
th
e
p
r
o
b
le
m
s
tate
m
en
t
o
f
t
h
e
p
r
o
p
o
s
ed
s
t
u
d
y
ca
n
b
e
s
tated
a
s
"
T
o
d
esig
n
a
f
r
a
m
e
w
o
r
k
th
at
o
f
f
er
s
m
o
r
e
co
n
te
x
t
u
al
s
ce
n
e
an
al
y
s
i
s
to
en
h
a
n
ce
th
e
p
r
ec
is
io
n
le
v
el
o
f
id
en
ti
f
icatio
n
o
f
o
u
tlier
s
.
"
T
h
e
n
ex
t
s
ec
tio
n
d
is
cu
s
s
es t
h
e
p
r
o
p
o
s
ed
s
o
lu
tio
n
.
1
.
3
.
P
ro
po
s
ed
So
lutio
n
T
h
e
p
r
o
p
o
s
ed
s
tu
d
y
is
a
co
n
ti
n
u
at
io
n
o
f
o
u
r
p
r
io
r
w
o
r
k
[
4
2
]
,
[
4
3
]
w
h
er
e
t
h
e
p
r
esen
t
s
o
l
u
t
io
n
tar
g
et
s
to
d
ev
elo
p
a
f
r
a
m
e
w
o
r
k
f
o
r
v
i
d
eo
s
u
r
v
eilla
n
ce
s
y
s
te
m
t
h
at
i
s
ca
p
ab
le
o
f
id
en
ti
f
y
i
n
g
o
u
tl
ie
r
s
f
o
r
a
g
i
v
en
s
et
o
f
ca
p
tu
r
ed
v
id
eo
f
r
a
m
e
s
.
T
h
e
p
r
im
ar
y
co
n
s
id
er
atio
n
o
f
t
h
is
p
ap
er
is
t
h
at
ex
is
te
n
ce
o
f
o
u
tlier
s
i
s
n
ev
er
in
s
ta
n
ta
n
eo
u
s
a
n
d
n
o
r
m
all
y
e
x
is
t
s
o
n
t
h
e
g
i
v
en
s
ce
n
e
co
n
s
id
er
in
g
b
o
th
ti
m
e
f
ac
to
r
an
d
s
p
atial
f
ac
to
r
w
h
er
e
b
o
th
th
e
s
e
f
ac
to
r
s
ca
n
al
s
o
b
e
s
tated
as
co
n
te
x
t
u
al
f
a
cto
r
s
.
T
h
is
ca
n
b
e
e
m
p
ir
ica
ll
y
r
ep
r
ese
n
ted
b
y
E
q
u
atio
n
(
1
)
,
A
=
A
s
pat
x
A
time
(
1
)
I
n
th
e
ab
o
v
e
ex
p
r
ess
io
n
,
th
e
v
ar
iab
le
A
r
ep
r
esen
ts
an
ag
g
r
eg
atio
n
o
f
lo
ca
l
attr
ib
u
te
w
h
e
r
e
A
spat
an
d
A
time
r
ep
r
ese
n
t
s
p
atial
a
n
d
te
m
p
o
r
al
attr
ib
u
te
as
s
o
ciate
d
w
it
h
t
h
e
lo
ca
l
f
ea
tu
r
e.
T
h
e
i
d
ea
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
is
to
co
m
p
u
te
lo
ca
l
attr
ib
u
te
a
(
A
a
)
f
o
r
all
th
e
p
ix
els
to
d
ev
elo
p
a
lo
ca
l
attr
ib
u
te
Z
i
t
h
a
t
is
lo
ca
lized
at
th
e
ce
n
tr
o
id
al
p
o
s
itio
n
f
o
r
a
g
iv
e
n
p
ix
el.
T
h
is
p
r
o
ce
s
s
r
esu
lts
i
n
th
e
g
e
n
er
atio
n
o
f
h
is
to
g
r
a
m
π(
Z
i
)
,
w
h
er
e
π
r
ep
r
esen
ts
n
e
g
ati
v
e
co
ef
f
icie
n
t
.
T
h
er
ef
o
r
e,
ap
p
ly
i
n
g
p
r
o
b
ab
ilit
y
,
t
h
e
id
en
ti
f
icatio
n
o
f
t
h
e
o
u
tli
er
s
p
r
o
b
(
A
)
ca
n
b
e
em
p
ir
icall
y
ex
p
r
es
s
ed
as,
n
k
k
o
k
Z
Z
z
p
r
o
b
1
)
,
|
(
(
2
)
i
j
j
k
Z
p
r
o
b
/
)
(
(
3
)
I
n
th
e
ab
o
v
e
E
q
u
atio
n
(
2
)
,
z
k
=(
x
k
-
x
o
,y
k
-
y
o
,t
k
-
t
o
)
ca
n
b
e
co
n
s
id
er
ed
to
r
e
p
r
esen
t
th
e
p
o
s
it
io
n
-
b
ased
ass
o
ciatio
n
o
f
Z
k
w
i
th
Z
o
.
T
h
e
ab
o
v
e
ex
p
r
ess
io
n
ca
n
f
u
r
th
er
b
e
s
p
lit
in
t
h
e
f
o
r
m
o
f
p
r
o
b
(z
k
|ϕ
i
,
ϕ
j
)
to
r
ep
r
esen
t
p
r
o
b
a
b
ilis
tic
s
elec
tio
n
f
o
r
a
p
o
s
itio
n
w
it
h
ti
m
e
an
d
s
p
ati
al
-
b
ased
attr
ib
u
te
s
w
h
er
e
v
ar
iab
le
ϕ
r
ep
r
esen
ts
d
ictio
n
ar
y
.
A
ls
o
,
i
n
th
e
ab
o
v
e
e
m
p
ir
ical
ex
p
r
ess
io
n
,
p
r
o
b
(Z
k
=ϕ
j
)
is
eq
u
iv
ale
n
t
to
co
r
r
elatio
n
al
f
ac
to
r
ex
i
s
ti
n
g
b
et
w
ee
n
X
k
a
n
d
d
ictio
n
ar
y
ϕ
j
as
s
h
o
w
n
i
n
E
q
u
at
io
n
(
3
)
.
A
lt
h
o
u
g
h
t
h
e
co
m
p
u
tatio
n
o
f
p
r
o
b
(
A
)
is
s
l
ig
h
tl
y
co
m
p
u
tatio
n
all
y
co
m
p
lex
p
r
o
ce
s
s
o
w
i
n
g
to
its
d
ep
en
d
en
cie
s
o
f
h
is
to
g
r
a
m
f
ac
to
r
s
π
f
o
r
all
th
e
g
i
v
en
v
alu
e
o
f
A
,
it
also
o
f
f
er
s
ca
p
ab
ilit
y
to
co
m
p
u
te
s
u
c
h
f
o
r
m
s
o
f
attr
ib
u
tes
a
n
d
r
eu
t
ilize
h
is
to
g
r
a
m
a
ttrib
u
te
π
f
o
r
all
t
h
e
lo
ca
l
attr
ib
u
te
s
f
o
r
d
if
f
er
en
t
ag
g
r
e
g
ates
A
o
f
p
i
x
els.
T
h
e
co
n
s
tr
u
ct
io
n
o
f
th
e
h
i
s
to
g
r
a
m
f
ac
to
r
o
f
n
eg
a
tiv
e
co
ef
f
icie
n
t π
i
s
e
m
p
ir
icall
y
c
o
m
p
u
ted
as
s
h
o
w
n
in
E
q
u
a
tio
n
(
4
)
as f
o
llo
w
s
:
Z
tim
e
s
p
a
t
i
a
Z
)
,
(
)
(
(
4
)
On
e
o
f
t
h
e
i
n
ter
es
tin
g
f
ac
t
s
to
o
b
s
er
v
e
i
s
t
h
at
u
s
a
g
e
o
f
ab
o
v
e
h
i
s
to
g
r
a
m
attr
ib
u
te
π
is
ap
p
r
o
x
i
m
ate
l
y
eq
u
iv
ale
n
t
to
lo
ca
l
attr
ib
u
te
s
t
h
at
o
f
f
er
s
i
g
n
if
ican
t
a
m
o
u
n
t
o
f
g
r
an
u
lar
it
y
b
y
h
ar
n
es
s
in
g
th
e
m
in
u
te
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
F
r
a
mewo
r
k
fo
r
C
o
n
textu
a
l O
u
tlier
I
d
en
tifi
ca
tio
n
u
s
in
g
Mu
ltiva
r
ia
te
A
n
a
lysi
s
….
(
Ma
d
h
u
C
h
a
n
d
r
a
G
)
1095
in
f
o
r
m
atio
n
f
r
o
m
t
h
e
g
i
v
e
n
m
u
l
ti
m
ed
ia
f
ile
(
i.e
.
,
v
id
e
o
)
to
o
f
f
er
b
etter
ac
cu
r
ac
y
p
er
f
o
r
m
a
n
ce
w
h
ile
p
er
f
o
r
m
in
g
an
o
u
tlier
id
en
ti
f
ic
atio
n
.
T
h
e
n
ex
t
s
tep
o
f
th
e
s
t
u
d
y
is
to
p
er
f
o
r
m
lear
n
in
g
o
p
er
atio
n
w
h
er
e
t
h
e
i
n
itial
lear
n
i
n
g
p
r
o
ce
s
s
is
ap
p
lied
o
n
s
elec
tio
n
o
f
p
r
o
b
a
b
ilit
ies
v
ar
iab
le
p
r
o
b
(π
k
|ϕ
i
,
ϕ
j
)
.
A
p
p
l
y
i
n
g
lear
n
i
n
g
o
p
er
atio
n
is
n
o
t
ch
allen
g
i
n
g
h
er
e
as
it
ca
n
b
e
d
ir
ec
tl
y
e
x
e
cu
ted
o
v
er
t
h
e
tr
ai
n
ed
d
ata.
T
h
e
s
ec
o
n
d
lear
n
i
n
g
p
r
o
ce
s
s
is
ap
p
lied
co
n
ce
r
n
in
g
id
en
ti
f
icatio
n
o
f
c
u
t
-
o
f
f
μ
.
T
h
e
p
r
im
e
id
ea
i
m
p
le
m
en
ted
in
t
h
is
p
ar
t
o
f
th
e
s
t
u
d
y
i
s
if
all
t
h
e
ag
g
r
eg
a
ted
p
o
in
ts
ar
e
o
b
s
er
v
ed
in
t
h
e
f
o
r
m
o
f
t
i
m
e
-
b
ased
ch
r
o
n
o
lo
g
ical
d
ata
th
an
t
h
e
f
r
eq
u
e
n
c
y
o
f
o
cc
u
r
r
en
ce
o
f
o
u
t
lier
s
h
as
s
ig
n
i
f
ica
n
tl
y
lo
w
er
v
alu
e
co
m
p
ar
ed
to
o
th
er
d
ata
p
o
in
ts
.
T
h
er
ef
o
r
e,
w
e
co
n
s
tr
u
ct
a
s
tate
m
e
n
t
t
h
at
w
h
i
le
m
o
n
ito
r
i
n
g
th
e
o
b
j
ec
ts
o
f
th
e
a
g
g
r
e
g
ate
s
ar
e
f
o
u
n
d
to
s
atis
f
y
th
e
f
o
llo
w
in
g
co
n
d
itio
n
s
h
o
w
n
in
E
q
u
atio
n
(
4
)
as
d
ec
is
io
n
m
a
k
i
n
g
:
p
r
o
b
(
A
)
<
μ
(
4
)
I
f
th
e
ab
o
v
e
co
n
d
itio
n
is
f
o
u
n
d
to
b
e
v
alid
th
an
t
h
e
s
y
s
te
m
d
ec
id
es
th
at
t
h
e
m
o
n
ito
r
ed
o
b
j
ec
t
is
co
n
s
id
er
ed
as
o
u
tlier
s
.
T
o
lear
n
th
e
cu
t
-
o
f
f
μ
f
r
o
m
tr
ai
n
ed
d
a
ta,
a
h
ar
d
-
co
d
ed
attr
ib
u
te
o
f
p
r
o
b
ab
ilit
y
o
f
o
u
tl
ier
co
u
ld
b
e
s
elec
ted
b
y
t
h
e
r
eq
u
i
r
e
m
en
t
o
f
th
e
u
s
er
.
W
e
co
n
s
id
er
th
at
s
m
aller
v
al
u
e
o
f
s
u
ch
p
r
o
b
a
b
ilit
y
w
i
ll
n
o
t
lead
to
ef
f
icien
t
d
etec
tio
n
a
s
it
m
a
y
h
av
e
h
ig
h
er
ch
a
n
ce
s
to
g
en
er
ate
f
al
s
e
p
o
s
itiv
e
s
as
o
n
l
y
le
s
s
n
u
m
b
er
o
f
o
u
tlier
s
w
i
ll
b
e
ig
n
o
r
ed
.
A
t
th
e
s
a
m
e
ti
m
e,
a
h
i
g
h
er
v
al
u
e
o
f
s
u
c
h
p
r
o
b
ab
ilit
y
s
h
o
u
ld
also
b
e
r
ej
ec
ted
as
th
e
o
u
tco
m
e
m
a
y
n
o
t
b
e
p
r
ac
tical
in
o
r
i
g
i
n
.
T
h
is
p
r
o
b
le
m
i
s
av
o
id
ed
b
y
p
er
f
o
r
m
i
n
g
s
p
ar
s
it
y
-
b
ased
lear
n
i
n
g
ap
p
r
o
ac
h
i
n
o
r
d
e
r
to
c
o
m
p
u
te
p
r
o
b
(
A
)
f
o
r
all
th
e
ag
g
r
eg
ate
s
ex
is
ti
n
g
in
t
h
e
tr
ain
ed
d
ataset
f
o
r
th
e
p
u
r
p
o
s
e
o
f
ev
alu
a
tin
g
t
h
e
cu
t
-
o
f
f
v
al
u
e
o
f
μ
to
an
y
p
ar
ticu
lar
v
al
u
e
in
s
u
ch
a
w
a
y
th
at
p
r
o
p
o
r
tio
n
o
f
ag
g
r
eg
ate
s
s
atis
f
y
i
n
g
th
e
lo
g
ical
co
n
d
itio
n
o
f
p
r
o
b
(
A
)
<
μ
is
a
s
p
ec
if
ic
p
r
o
b
a
b
ilit
y
,
i.e
.
,
p
r
o
b
(
a)
.
W
h
er
e
A
a
.
An
an
al
y
tica
l
r
esear
ch
m
e
th
o
d
o
lo
g
y
is
ap
p
li
ed
to
ca
r
r
y
o
u
t
t
h
e
i
m
p
le
m
en
t
a
tio
n
o
f
t
h
e
p
r
o
p
o
s
ed
s
tu
d
y
.
Fig
u
r
e
1
h
i
g
h
lig
h
t
s
th
e
s
c
h
e
m
atic
d
iag
r
a
m
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
to
id
en
t
if
y
a
n
y
f
o
r
m
o
f
a
b
n
o
r
m
al
e
v
e
n
ts
a
s
o
u
tlier
s
f
r
o
m
th
e
v
id
eo
f
r
a
m
e
s
.
A
l
g
o
r
i
t
h
m
f
o
r
E
x
t
r
a
c
t
i
o
n
o
f
B
l
o
c
k
E
x
t
r
a
c
t
e
d
b
l
o
c
k
A
l
g
o
r
i
t
h
m
f
o
r
S
p
a
r
s
i
t
y
-
b
a
s
e
d
L
e
a
r
n
i
n
g
D
i
c
t
i
o
n
a
r
y
m
a
t
r
i
x
i
n
d
e
x
e
x
t
r
a
c
t
e
d
f
e
a
t
u
r
e
s
A
l
g
o
r
i
t
h
m
f
o
r
o
u
t
l
i
e
r
d
e
t
e
c
t
i
o
n
D
e
t
e
c
t
i
o
n
o
f
o
u
t
l
i
e
r
s
o
f
a
n
o
b
j
e
c
t
V
i
d
e
o
E
x
t
r
a
c
t
e
d
F
r
a
m
e
s
Fig
u
r
e
1
.
Sch
e
m
atic
A
r
ch
i
tect
u
r
e
o
f
P
r
o
p
o
s
ed
Sy
s
te
m
T
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
is
d
es
ig
n
ed
i
n
t
w
o
p
o
s
s
ib
le
s
t
ep
s
o
f
(
a)
u
n
s
u
p
er
v
is
ed
tr
ain
i
n
g
an
d
(
b
)
v
alid
atin
g
.
T
h
e
s
y
s
te
m
u
s
es
lo
ca
l
-
le
v
el
f
ea
t
u
r
es
in
th
e
p
r
eli
m
in
ar
y
le
v
el
th
a
t
is
c
h
ar
ac
ter
ized
b
y
m
o
r
e
s
o
p
h
is
ticated
p
atter
n
s
.
W
it
h
t
h
e
aid
o
f
p
r
o
b
ab
ilit
y
th
eo
r
y
,
t
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
a
s
s
e
s
s
es
t
h
e
le
v
el
o
f
o
u
tlier
s
ex
is
t
in
g
i
n
g
i
v
e
n
d
ata
u
s
in
g
a
b
lo
ck
in
g
o
p
er
atio
n
.
An
al
g
o
r
ith
m
f
o
r
ex
tr
ac
ti
n
g
b
lo
ck
is
d
es
i
g
n
ed
t
h
at
ta
k
es
t
h
e
in
p
u
t
o
f
v
id
eo
f
r
a
m
e
w
h
ic
h
u
p
o
n
p
r
o
ce
s
s
in
g
y
ie
ld
s
an
o
u
tp
u
t
o
f
ex
tr
ac
ted
b
lo
ck
s
.
T
h
is
b
lo
c
k
ass
is
t
s
in
f
u
r
th
er
ex
tr
ac
tio
n
o
f
lo
ca
l
-
lev
el
in
f
o
r
m
atio
n
th
er
eb
y
co
n
s
tr
u
cti
n
g
a
g
o
o
d
n
u
m
b
er
o
f
f
ea
t
u
r
es
with
t
h
e
aid
o
f
n
e
x
t
alg
o
r
ith
m
o
f
s
p
ar
s
it
y
-
b
a
s
ed
lear
n
in
g
.
B
o
th
th
e
s
p
atial
as
w
e
ll
as
ti
m
e
-
r
elate
d
in
f
o
r
m
atio
n
ca
p
tu
r
ed
f
r
o
m
th
e
tr
ain
ed
b
lo
ck
s
ar
e
u
s
ed
to
p
er
f
o
r
m
d
etec
tio
n
o
f
o
u
tlier
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
2
,
A
p
r
il
201
8
:
1
0
9
2
–
1101
1096
Fin
all
y
,
a
n
al
g
o
r
ith
m
f
o
r
o
u
tli
er
d
etec
tio
n
is
f
o
r
m
u
l
ated
.
T
h
e
s
ig
n
i
f
ica
n
t c
o
n
tr
ib
u
tio
n
o
f
t
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
i
s
as
f
o
llo
w
s
v
iz.
a.
C
o
n
s
tr
u
ctio
n
o
f
an
a
n
al
y
tical
f
r
a
m
e
w
o
r
k
f
o
r
f
ea
tu
r
e
ex
tr
ac
tio
n
f
r
o
m
a
g
i
v
en
v
id
eo
s
et
b.
I
n
co
r
p
o
r
ates b
lo
ck
in
g
o
p
er
atio
n
f
o
r
f
u
r
t
h
er
g
r
an
u
lar
it
y
in
t
h
e
f
ea
t
u
r
e
ex
tr
ac
tio
n
p
r
o
ce
s
s
c.
U
s
in
g
a
d
ictio
n
ar
y
f
o
r
ass
is
ti
n
g
i
n
t
h
e
b
etter
id
en
ti
f
icatio
n
p
r
o
ce
s
s
.
T
h
e
s
ig
n
i
f
ica
n
t
lev
e
l
o
f
th
e
s
tu
d
y
co
n
tr
ib
u
tio
n
is
t
h
at
p
r
o
p
o
s
ed
s
y
s
te
m
i
s
co
m
p
letel
y
ca
p
ab
le
o
f
s
o
l
v
i
n
g
a
n
y
f
o
r
m
o
f
m
u
lti
v
ar
iate
p
r
o
b
lem
s
ex
is
ti
n
g
in
t
h
e
ca
s
e
o
f
v
id
eo
s
u
r
v
eil
la
n
ce
s
y
s
t
e
m
.
T
h
er
ef
o
r
e,
th
e
co
n
tex
t
o
f
th
e
s
ce
n
e
i
s
u
n
d
er
s
to
o
d
w
ell,
an
d
all
t
h
e
d
etailed
i
n
f
o
r
m
at
io
n
is
s
i
g
n
i
f
ica
n
tl
y
ca
p
t
u
r
ed
b
y
th
e
p
r
o
p
o
s
ed
f
r
a
m
e
w
o
r
k
.
T
h
e
u
ti
lizatio
n
o
f
t
h
e
p
r
o
p
o
s
ed
f
r
am
e
w
o
r
k
i
s
m
o
r
e
o
n
ab
n
o
r
m
al
o
b
j
ec
t
b
eh
av
io
r
in
a
n
y
e
n
v
ir
o
n
m
en
t.
T
h
e
n
ex
t
s
ec
tio
n
d
is
c
u
s
s
es
th
e
alg
o
r
ith
m
i
m
p
le
m
en
tatio
n
f
o
llo
w
ed
b
y
a
d
is
c
u
s
s
io
n
o
f
th
e
o
u
tco
m
e
o
b
tain
ed
f
r
o
m
th
e
s
t
u
d
y
.
2.
AL
G
O
RI
T
H
M
I
M
P
L
E
M
E
NT
A
T
I
O
N
T
h
e
p
r
im
e
p
u
r
p
o
s
e
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
i
s
to
p
er
f
o
r
m
a
p
r
ec
is
e
id
en
ti
f
icatio
n
o
f
t
h
e
o
u
tlier
s
f
r
o
m
t
h
e
v
id
eo
f
r
a
m
es.
Ho
wev
er
,
th
i
s
r
esear
c
h
ai
m
is
ca
r
r
ied
o
u
t
co
n
s
id
er
in
g
t
h
e
f
o
r
m
u
latio
n
o
f
t
h
r
ee
d
if
f
er
e
n
t
al
g
o
r
ith
m
s
w
h
er
e
th
e
y
ar
e
r
esp
o
n
s
ib
le
f
o
r
e
x
t
r
ac
t
in
g
b
lo
ck
s
,
ap
p
l
y
in
g
s
p
ar
s
it
y
f
o
r
i
m
p
le
m
e
n
ti
n
g
lear
n
in
g
s
tr
ateg
y
,
an
d
f
o
r
p
er
f
o
r
m
i
n
g
o
u
tl
ier
d
etec
tio
n
.
A
ll t
h
e
al
g
o
r
ith
m
s
ar
e
co
n
s
tr
u
cted
in
a
s
eq
u
en
t
ial
f
o
r
m
an
d
h
en
ce
ar
e
r
esp
ec
ti
v
el
y
i
llu
s
tr
ated
s
eq
u
en
tiall
y
.
T
h
e
s
tep
s
in
v
o
l
v
ed
in
t
h
e
alg
o
r
it
h
m
-
1
a
r
e
as f
o
llo
w
s
:
A
l
g
o
r
i
t
h
m
-
1
f
o
r
E
x
t
r
a
c
t
i
o
n
o
f
B
l
o
c
k
In
p
u
t
:
f
(
n
u
m
b
e
r
o
f
f
r
a
me
s)
Ou
t
p
u
t
:
B
di
c
t
(
Ex
t
r
a
c
t
e
d
b
l
o
c
k
)
S
t
a
r
t
1
.
i
n
i
t
f
2
.
[
n
r
,
n
c
,
k
]
s
i
z
e
(
f
)
3
.
B
b
l
o
c
k
5x5
(f)
4
.
n
rd
p
r
o
d
(
B
S
)
&
n
cd
s
i
z
e
(
B
)
,
o
b
s
BS
1
*
B
S
2
5
.
B
di
c
t
me
a
n
[
(
n
n
-
1
)
*
o
b
s+
1
:
n
n
*
o
b
s
]
6
.
F
o
r
j
=
1
:
si
z
e
(
B
di
c
t
)
7
.
v
=
v
-
m
1
(
j
)
8
.
a
=
v
/
n
o
r
m(
v
)
9
.
B
d
i
c
t
a
1
0
.
E
n
d
E
n
d
T
h
e
alg
o
r
ith
m
-
1
i
n
itiall
y
ta
k
e
s
th
e
v
id
eo
f
r
a
m
e
s
f
as
t
h
e
i
n
p
u
t
(
L
i
n
e
-
1
)
,
w
h
ich
is
f
o
llo
w
ed
b
y
t
h
e
n
u
m
b
er
o
f
o
p
er
atio
n
s
in
th
e
c
o
n
s
ec
u
tiv
e
s
tep
s
f
o
r
p
er
f
o
r
m
i
n
g
b
lo
c
k
co
m
p
u
tatio
n
.
T
h
e
s
i
ze
o
f
t
h
e
f
r
a
m
e
f
i
s
th
en
m
ap
p
ed
in
to
t
h
r
ee
v
ar
iab
les
n
u
m
b
er
o
f
r
o
w
s
nr
,
s
o
m
e
co
lu
m
n
s
nc
a
n
d
in
d
ex
k
(
L
i
n
e
-
2
)
.
T
h
e
n
ex
t
s
tep
i
s
to
co
n
v
er
t
t
h
e
p
ix
el
e
le
m
e
n
t
s
i
n
to
co
lu
m
n
ar
f
o
r
m
to
d
iv
id
e
t
h
e
f
r
a
m
e
i
n
to
d
is
ti
n
ct
5
x
5
b
lo
ck
B
(
L
i
n
e
-
3
)
.
T
w
o
v
ar
iab
le
n
rd
a
n
d
n
cd
co
m
p
u
te
s
th
e
n
u
m
b
er
o
f
r
o
w
s
an
d
c
o
lu
m
n
s
f
o
r
d
ictio
n
ar
y
r
esp
ec
tiv
el
y
alo
n
g
w
it
h
co
m
p
u
tatio
n
o
f
o
n
e
b
lo
ck
s
iz
e
obs
(
L
in
e
-
4
)
.
Fi
n
all
y
,
th
e
d
ictio
n
ar
y
i
s
cr
ea
ted
co
n
s
id
er
i
n
g
n
r
d
,
n
cd
,
a
n
d
th
e
n
u
m
b
er
o
f
f
r
a
m
es
f
o
r
tr
ain
i
n
g
d
iv
id
ed
b
y
b
lo
ck
s
ize.
Fo
r
all
th
e
s
izes
o
f
th
e
d
ictio
n
ar
y
-
b
as
ed
b
lo
ck
s
(
L
in
e
-
6
)
,
all
th
e
f
r
a
m
es
ar
e
co
n
s
id
er
ed
,
w
h
ic
h
ar
e
th
en
f
u
r
t
h
er
d
iv
id
ed
in
to
5
x
5
d
is
t
in
ct
b
lo
ck
s
.
Fi
n
a
ll
y
,
th
e
d
ictio
n
ar
y
-
b
ased
b
lo
ck
s
ar
e
co
m
p
u
ted
a
s
s
h
o
w
n
i
n
L
i
n
e
-
5
to
o
b
tain
ed
B
dict
,
i.e
.
,
d
ictio
n
ar
y
-
b
ase
d
b
lo
ck
.
A
lo
o
p
is
cr
ea
ted
as
s
h
o
w
n
i
n
L
i
n
e
-
6
f
o
r
all
s
ize
s
o
f
B
dict
to
co
m
p
u
te
v
ec
to
r
v
=B
dict
(
j
)
an
d
m
1
r
ep
r
esen
ts
th
e
m
ea
n
v
al
u
e
o
f
B
dict
(
L
in
e
-
7
)
th
at
f
i
n
all
y
le
ad
s
to
th
e
g
e
n
er
atio
n
o
f
e
x
tr
ac
ted
b
lo
ck
s
B
dict
as th
e
o
u
tco
m
e
.
Af
ter
t
h
e
b
lo
c
k
s
ar
e
ex
tr
ac
ted
f
r
o
m
th
e
g
i
v
e
n
f
r
a
m
es,
t
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
i
m
p
le
m
e
n
ts
a
n
o
v
el
f
o
r
m
o
f
m
atr
i
x
d
ec
o
m
p
o
s
it
io
n
to
p
er
f
o
r
m
m
u
l
tiv
ar
ia
te
an
al
y
s
is
u
s
in
g
s
p
ar
s
it
y
-
b
a
s
ed
lear
n
i
n
g
p
r
o
ce
s
s
.
T
h
e
s
tep
s
o
f
an
alg
o
r
it
h
m
-
2
f
o
r
s
p
ar
s
it
y
-
b
ased
lear
n
in
g
an
d
s
tep
s
o
f
an
al
g
o
r
ith
m
-
3
f
o
r
o
u
tlier
d
etec
tio
n
ar
e
g
iv
e
n
.
T
h
e
ab
o
v
e
-
m
e
n
tio
n
ed
al
g
o
r
it
h
m
-
2
f
ir
s
t
in
i
tialize
s
t
h
e
i
n
d
ex
o
f
t
h
e
m
atr
ix
k
to
s
et
d
i
m
e
n
s
i
o
n
o
f
t
h
e
d
ictio
n
ar
y
f
o
r
ap
p
ly
in
g
it
to
m
u
lti
-
v
ar
iate
an
al
y
s
i
s
to
it (
L
i
n
e
-
1
)
.
A
s
tr
u
ctu
r
e
is
m
ai
n
tai
n
ed
f
o
r
th
e
d
ict
io
n
ar
y
f
o
llo
w
ed
b
y
f
o
r
m
atio
n
o
f
a
lo
o
p
as
s
h
o
w
n
in
L
i
n
e
-
3
.
A
m
atr
ix
Dt
i
s
cr
ea
ted
f
o
r
s
to
r
in
g
all
th
e
d
ictio
n
ar
y
-
r
elate
d
v
a
l
u
e
s
w
it
h
i
n
it
s
el
f
(
L
in
e
-
4
)
f
o
llo
w
ed
b
y
th
e
cr
ea
tio
n
o
f
a
s
u
p
er
-
in
d
ex
X
(
L
in
e
-
6
)
th
a
t
m
a
in
ta
in
s
a
ll
th
e
m
u
lti
-
v
ar
iate
m
atr
ices
o
f
Dt.
T
h
e
n
e
x
t
s
tep
is
to
ap
p
l
y
a
f
u
n
c
tio
n
ϕ
t
h
at
p
er
f
o
r
m
s
s
p
ar
s
e
m
atr
i
x
d
ec
o
m
p
o
s
itio
n
u
s
in
g
lin
ea
r
alg
eb
r
a
(
L
in
e
-
8
)
o
v
er
th
e
m
at
r
ix
in
d
e
x
k
an
d
s
u
p
er
-
i
n
d
ex
X.
As
th
e
o
u
tco
m
e
o
f
th
i
s
m
atr
i
x
is
al
w
a
y
s
p
o
s
itiv
e,
th
er
ef
o
r
e,
it
is
ea
s
ier
f
o
r
co
m
p
u
ti
n
g
t
h
e
r
es
u
lti
n
g
m
atr
i
x
.
T
h
e
o
u
tco
m
e
o
f
t
h
is
a
lg
o
r
i
th
m
r
es
u
lts
i
n
t
h
e
g
en
er
atio
n
o
f
m
u
ltip
le
f
ea
tu
r
es,
e.
g
.
,
X
(
m
atr
i
x
w
it
h
m
i
x
ed
s
ig
n
s
)
,
A
(
b
asis
m
atr
i
x
)
,
an
d
Y
(
co
ef
f
icien
t
m
atr
i
x
)
.
T
h
e
d
ictio
n
ar
y
ev
o
l
v
ed
f
r
o
m
th
is
al
g
o
r
ith
m
i
n
L
in
e
-
8
w
ill
b
e
r
eu
s
ed
f
o
r
p
er
f
o
r
m
i
n
g
id
e
n
ti
f
icatio
n
o
f
th
e
o
u
tlier
s
f
o
r
th
e
test
f
r
a
m
e
s
.
A
f
ter
t
h
e
f
ea
t
u
r
es
h
a
v
e
b
ee
n
ex
tr
ac
ted
f
r
o
m
t
h
e
m
u
lti
v
ar
ia
te
an
al
y
s
is
co
n
ce
p
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
F
r
a
mewo
r
k
fo
r
C
o
n
textu
a
l O
u
tlier
I
d
en
tifi
ca
tio
n
u
s
in
g
Mu
ltiva
r
ia
te
A
n
a
lysi
s
….
(
Ma
d
h
u
C
h
a
n
d
r
a
G
)
1097
u
s
ed
i
n
t
h
e
lear
n
in
g
p
h
a
s
e,
t
h
e
n
ex
t
p
ar
t
o
f
al
g
o
r
ith
m
i
m
p
le
m
en
tatio
n
w
ill
b
e
to
p
er
f
o
r
m
i
d
en
tific
atio
n
o
f
t
h
e
o
u
tlier
s
.
T
h
e
in
p
u
t
to
alg
o
r
ith
m
-
3
f
o
r
o
u
tlier
d
etec
tio
n
is
t
h
e
d
ictio
n
ar
y
t
h
at
h
a
s
b
ee
n
f
o
r
m
ed
in
its
p
r
io
r
alg
o
r
ith
m
(
L
in
e
-
1
)
.
A
ll
t
h
e
i
m
a
g
es,
a
s
well
as
g
r
o
u
n
d
tr
u
t
h
i
m
a
g
es,
ar
e
co
n
s
id
er
ed
in
t
h
i
s
s
tu
d
y
p
h
as
e.
I
t
is
f
o
llo
w
ed
b
y
i
m
p
le
m
en
ta
tio
n
o
f
f
ir
s
t
al
g
o
r
ith
m
f
o
r
b
lo
ck
ex
tr
ac
tio
n
(
L
i
n
e
-
2
)
,
w
h
er
e
s
i
m
ilar
s
tep
s
,
e.
g
.
,
cr
ea
tio
n
o
f
d
ictio
n
ar
y
o
f
5
x
5
b
lo
ck
s
ize,
r
ea
d
in
g
th
e
te
s
t
f
r
a
m
e,
d
iv
id
i
n
g
th
e
f
r
a
m
e
s
i
n
to
5
x
5
b
lo
ck
s
,
co
m
p
u
tatio
n
o
f
d
ictio
n
ar
y
-
co
ef
f
icie
n
ts
(
n
rd
,
n
cd
,
an
d
o
b
s
)
,
co
m
p
u
tatio
n
o
f
m
ea
n
o
f
B
dict
,
an
d
esti
m
atio
n
o
f
d
ictio
n
ar
y
-
b
ased
v
ec
to
r
v
(
L
in
e
-
3
)
.
A
n
e
w
m
atr
ix
M
Dict
i
s
f
o
r
m
u
lated
b
y
-
p
r
o
d
u
ct
o
f
a
d
ictio
n
ar
y
o
f
all
s
izes
o
f
B
dict
an
d
s
u
p
er
-
in
d
ex
(
li
n
e
-
4
)
.
A
d
is
tan
ce
co
m
p
u
tat
io
n
o
f
t
h
e
t
w
o
li
n
ea
r
p
o
in
ts
is
co
m
p
u
ted
u
s
in
g
p
r
o
b
ab
ilit
y
p
r
o
b
b
et
w
ee
n
M
dict
an
d
n
o
r
m
alize
d
v
ec
to
r
a
f
o
llo
w
ed
b
y
co
m
p
u
tatio
n
o
f
f
i
n
al
p
r
o
b
ab
ilit
y
,
i.e
.
,
E
prob
(
L
i
n
e
-
6
)
.
T
h
e
alg
o
r
ith
m
th
en
e
s
ti
m
ate
s
th
e
m
i
n
i
m
u
m
v
alu
e
o
f
E
prob
f
o
llo
w
ed
b
y
f
i
n
al
co
m
p
u
tatio
n
o
f
th
e
f
i
n
a
l
f
ea
t
u
r
e
Dim
t
h
at
tr
an
s
f
o
r
m
s
co
l
u
m
n
ar
in
f
o
r
m
at
io
n
in
th
e
f
o
r
m
o
f
an
i
m
a
g
e
(
f
o
r
h
ig
h
li
g
h
ti
n
g
t
h
e
o
u
tlier
s
)
.
Fin
all
y
,
u
s
in
g
th
e
in
f
o
r
m
atio
n
r
esto
r
ed
f
r
o
m
g
r
o
u
n
d
tr
u
th
i
m
a
g
e,
t
h
e
s
y
s
te
m
ch
ec
k
s
if
t
h
e
f
i
n
al
f
ea
t
u
r
e
Dim
i
s
m
o
r
e
th
a
n
1
(
L
in
e
-
8
)
.
I
t
is
i
m
p
o
r
tan
t
to
u
n
d
er
s
tan
d
t
h
at
g
r
o
u
n
d
tr
u
t
h
i
m
ag
e
s
p
la
y
a
s
i
g
n
i
f
ican
t
r
o
le
in
o
u
tlier
d
etec
tio
n
w
h
er
e
f
ir
s
t
th
e
g
r
o
u
n
d
tr
u
th
G
T
im
a
g
es
ar
e
r
ea
d
an
d
co
n
v
er
t
ed
to
b
in
ar
y
i
m
a
g
es.
T
h
is
p
r
o
ce
s
s
i
s
f
o
llo
w
ed
b
y
ap
p
ly
i
n
g
an
o
t
h
er
f
u
n
ctio
n
Ω
w
h
ic
h
is
m
ea
n
t
f
o
r
p
er
f
o
r
m
i
n
g
o
u
tlier
d
etec
tio
n
b
y
g
r
o
u
n
d
tr
u
th
i
m
a
g
e
GT
,
b
in
ar
y
i
m
a
g
e
bw
f
o
r
t
h
e
co
n
d
itio
n
m
ap
p
in
g
w
i
th
Dim
>1
,
a
n
d
f
i
n
al
f
ea
t
u
r
e
D
im
.
T
h
e
f
u
n
cti
o
n
Ω
i
s
d
es
ig
n
ed
i
n
f
o
llo
w
in
g
s
tep
s
:
f
ir
s
t
t
h
e
r
eg
i
o
n
-
b
ased
p
r
o
p
er
ties
o
f
g
r
o
u
n
d
tr
u
th
i
m
a
g
e
GT
is
esti
m
ated
b
ased
o
n
ce
n
tr
o
id
al
f
ac
to
r
o
f
r
eg
io
n
f
o
llo
w
ed
b
y
co
n
ca
ten
at
io
n
o
f
i
t
w
i
th
ce
n
tr
o
id
an
d
f
o
llo
w
ed
b
y
d
i
latio
n
o
p
er
atio
n
o
f
m
o
r
p
h
o
lo
g
y
in
o
r
d
er
to
o
b
tai
n
Dim
a
n
d
bw
v
al
u
es.
I
m
p
le
m
en
tatio
n
o
f
th
e
f
u
n
c
tio
n
Ω
l
ea
d
s
to
th
e
f
o
r
m
a
tio
n
o
f
p
r
o
b
ab
ilit
y
m
ap
P
r
o
b
m
ap
,
w
h
ic
h
is
c
h
ec
k
ed
f
o
r
its
cu
t
-
o
f
f
v
al
u
e
in
o
r
d
er
to
id
en
tify
th
e
o
u
tl
ier
s
o
b
j
ec
t
(
L
in
e
-
1
1
to
L
i
n
e
-
1
3
)
.
T
h
is
c
alcu
latio
n
is
f
in
al
l
y
f
o
llo
w
ed
b
y
es
ti
m
a
tio
n
o
f
ac
c
u
r
ac
y
p
er
f
o
r
m
an
ce
i
n
later
s
tag
e
s
.
A
clo
s
er
lo
o
k
i
n
to
t
h
e
alg
o
r
ith
m
f
o
r
m
a
tio
n
w
ill
s
h
o
w
th
at
p
r
o
p
o
s
ed
s
y
s
te
m
u
s
es
u
n
s
u
p
er
v
i
s
ed
lear
n
in
g
alg
o
r
ith
m
i
n
o
r
d
er
to
a
d
d
r
ess
th
e
u
n
ce
r
tain
t
y
p
r
o
b
le
m
s
as
s
o
ciate
d
w
it
h
d
etec
tio
n
o
f
p
a
tter
n
s
e
x
h
ib
ited
b
y
an
o
m
al
y
d
etec
tio
n
.
T
h
er
ef
o
r
e,
ap
p
l
y
in
g
th
e
ap
p
r
o
ac
h
o
f
m
u
l
tiv
ar
iate
a
n
al
y
s
i
s
o
v
er
t
h
e
m
at
r
ix
d
ec
o
m
p
o
s
itio
n
,
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
e
n
s
u
r
es
th
e
p
r
ese
n
ce
o
f
p
o
s
iti
v
e
e
le
m
e
n
ts
al
w
a
y
s
d
ep
icti
n
g
ca
s
e
o
f
n
o
n
-
o
u
t
lier
s
.
P
r
esen
ce
o
f
an
y
f
o
r
m
o
n
n
o
n
-
p
o
s
itiv
e
ele
m
en
ts
i
s
th
e
o
n
l
y
p
o
s
s
ib
le
ca
s
e
o
f
o
u
tl
ier
s
.
T
h
er
ef
o
r
e,
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
i
s
ea
s
ier
to
im
p
le
m
e
n
t
an
d
r
eq
u
ir
es
less
er
n
u
m
b
er
o
f
iter
atio
n
in
o
r
d
er
to
r
ea
ch
its
c
o
n
v
er
g
e
n
ce
p
h
ase
.
An
o
th
er
s
i
g
n
i
f
ica
n
t
ad
v
an
ta
g
e
o
f
t
h
is
al
g
o
r
it
h
m
is
i
ts
cl
u
s
ter
in
g
c
h
ar
ac
ter
is
tic
s
t
h
at
i
n
v
o
lu
n
tar
il
y
p
er
f
o
r
m
clu
s
ter
i
n
g
o
f
t
h
e
co
lu
m
n
ar
d
ata
o
f
th
e
v
ec
to
r
v
.
I
n
s
h
o
r
t,
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
h
ar
n
es
s
es
t
h
e
p
o
ten
tial
o
f
s
p
ar
s
it
y
-
b
ased
m
atr
i
x
d
ec
o
m
p
o
s
itio
n
co
n
ce
p
t
u
s
i
n
g
m
u
lti
v
ar
iate
an
a
l
y
s
is
to
o
b
tain
u
n
iq
u
e
n
u
m
er
ical
f
ea
tu
r
e
s
th
at
h
a
v
e
en
o
u
g
h
in
f
o
r
m
at
io
n
to
as
s
is
t
i
n
id
en
tific
atio
n
o
f
o
u
t
lier
s
.
T
h
e
m
atr
ix
h
o
ld
s
en
o
u
g
h
in
f
o
r
m
a
tio
n
w
it
h
o
u
t
ca
u
s
i
n
g
an
y
f
o
r
m
s
o
f
d
eg
r
ad
atio
n
to
s
y
s
te
m
p
er
f
o
r
m
an
ce
.
A
n
o
t
h
er
u
n
iq
u
en
e
s
s
o
f
t
h
is
al
g
o
r
ith
m
i
s
less
er
v
ar
iab
le
d
ep
en
d
en
cie
s
w
il
l
also
r
ed
u
ce
less
er
r
eso
u
r
ce
d
ep
en
d
en
cies
w
it
h
f
as
ter
co
m
p
u
tatio
n
al
ti
m
e.
T
h
e
n
ex
t sectio
n
d
is
c
u
s
s
e
s
a
b
o
u
t th
e
r
es
u
lt
s
b
ein
g
o
b
tain
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
2
,
A
p
r
il
201
8
:
1
0
9
2
–
1101
1098
3.
RE
SU
L
T
ANAL
YSI
S
T
h
e
im
p
le
m
e
n
tatio
n
o
f
t
h
e
al
g
o
r
ith
m
s
f
o
r
ac
h
ie
v
i
n
g
t
h
e
r
e
s
ea
r
ch
g
o
al
o
f
o
u
tlier
d
etec
tio
n
h
a
s
b
ee
n
ca
r
r
ied
o
u
t
u
s
in
g
p
ed
estria
n
d
ata
o
f
U
C
SD
o
n
Ma
tlab
.
T
h
e
d
ataset
co
n
s
is
t
s
o
f
ap
p
r
o
x
i
m
atel
y
6
8
0
0
i
m
a
g
e
s
eq
u
en
ce
s
w
h
er
e
v
ar
io
u
s
p
ed
estrian
s
ar
e
f
o
u
n
d
to
b
e
w
al
k
i
n
g
i
n
d
i
f
f
er
e
n
t
d
ir
ec
tio
n
s
o
n
t
h
e
g
i
v
e
n
p
ath
.
T
h
e
an
o
m
al
y
o
b
j
ec
t
w
ill
b
e
an
i
n
d
iv
id
u
al
c
y
cl
in
g
alo
n
g
t
h
e
p
a
th
o
f
a
p
ed
estrian
.
Si
m
ilar
l
y
,
th
e
te
s
t
-
i
m
ag
e
s
ar
e
f
o
u
r
ti
m
e
s
t
h
e
n
u
m
b
er
o
f
tr
ain
i
m
a
g
es
a
l
o
n
g
w
it
h
g
r
o
u
n
d
tr
u
t
h
i
m
a
g
es.
T
h
e
v
i
s
u
a
l
o
u
tco
m
e
s
o
f
t
h
e
u
n
s
u
p
er
v
is
ed
tr
ain
i
n
g
o
p
er
atio
n
ar
e
as sh
o
w
n
b
elo
w
i
n
T
ab
le
1
.
T
ab
le
1
.
Sam
p
le
Vis
u
al
o
u
tco
m
es o
f
T
r
ain
in
g
se
t
F
r
a
me
-
1
F
r
a
me
-
50
F
r
a
me
-
100
1
10
20
30
T
ab
le
1
s
h
o
w
s
t
h
e
s
a
m
p
le
v
i
s
u
al
o
u
tco
m
es
o
f
1
st
,
5
0
th
,
an
d
1
0
0
th
f
r
a
m
e
f
o
r
1
st
,
1
0
th
,
2
0
th
,
an
d
3
0
th
tr
ain
i
n
g
d
ataset.
T
h
e
co
m
p
let
e
tr
ain
in
g
o
f
6
8
0
0
im
a
g
e
s
to
o
k
ap
p
r
o
x
i
m
atel
y
0
.
7
6
2
1
s
ec
o
n
d
s
in
t
h
e
co
r
e
-
i5
p
r
o
ce
s
s
o
r
alo
n
g
w
it
h
t
h
e
ex
tr
ac
tio
n
o
f
th
e
f
ea
t
u
r
es.
W
h
ile
p
er
f
o
r
m
in
g
th
e
tr
ai
n
i
n
g
,
all
t
h
e
f
ea
t
u
r
es
f
o
r
an
y
m
o
b
ile
o
b
j
ec
ts
ar
e
ex
tr
ac
ted
an
d
s
u
b
j
ec
ted
to
th
e
n
e
x
t
p
h
a
s
e
o
f
al
g
o
r
ith
m
i
m
p
le
m
e
n
tat
i
o
n
,
i.e
.
,
test
in
g
.
T
h
e
co
m
p
lete
tr
ain
i
n
g
is
ca
r
r
ied
o
u
t
co
n
s
id
er
in
g
p
ar
a
m
eter
o
f
s
p
a
r
s
e
r
eg
u
lar
izat
io
n
o
f
1
0
w
it
h
a
b
lo
ck
s
ize
o
f
5
x
5
.
T
h
e
p
o
s
s
ib
le
ch
al
len
g
i
n
g
s
i
tu
atio
n
i
s
to
e
n
s
u
r
e
t
h
e
d
etec
ti
o
n
o
f
an
o
m
al
y
o
b
j
ec
t,
i.e
.
,
a
c
y
clis
t
o
r
a
p
er
s
o
n
m
o
v
i
n
g
w
ith
ca
r
t,
etc.
T
h
is
is
ch
alle
n
g
i
n
g
as
th
e
d
ataset
h
a
s
d
if
f
er
en
t
n
u
m
b
er
s
o
f
o
b
j
ec
ts
(
p
er
s
o
n
)
m
o
v
in
g
at
d
if
f
er
e
n
t
s
p
ee
d
,
in
a
d
if
f
er
en
t
d
ir
ec
tio
n
,
w
it
h
d
i
f
f
er
en
t
m
o
b
ilit
y
p
atter
n
s
,
w
h
er
e
n
o
s
p
e
cif
ic
p
atter
n
ca
n
b
e
d
r
a
w
n
f
r
o
m
t
h
i
s
.
T
h
er
ef
o
r
e,
th
is
ca
s
e
s
t
u
d
y
ex
ac
t
l
y
m
a
tc
h
es
w
it
h
t
h
e
r
ea
l
-
ti
m
e
s
ce
n
a
r
io
.
Ho
w
e
v
er
,
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
s
o
l
v
es
t
h
is
p
r
o
b
lem
b
y
u
n
d
er
ta
k
in
g
t
h
e
b
lo
ck
in
g
o
p
er
atio
n
th
a
t
allo
w
s
ex
p
licit
ex
tr
ac
tio
n
o
f
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
F
r
a
mewo
r
k
fo
r
C
o
n
textu
a
l O
u
tlier
I
d
en
tifi
ca
tio
n
u
s
in
g
Mu
ltiva
r
ia
te
A
n
a
lysi
s
….
(
Ma
d
h
u
C
h
a
n
d
r
a
G
)
1099
a
f
ea
tu
r
e
f
r
o
m
s
p
ec
i
f
ic
b
lo
ck
s
o
f
t
h
e
i
m
ag
e.
T
h
is
o
p
er
atio
n
ass
i
s
ts
i
n
t
h
e
s
i
g
n
if
ica
n
t
f
o
r
m
u
la
tio
n
o
f
m
u
lti
v
ar
iate
an
a
l
y
s
is
o
f
d
if
f
e
r
en
t
co
ef
f
icie
n
ts
e
x
tr
ac
ted
f
r
o
m
th
e
o
b
j
ec
t
in
s
u
c
h
a
li
n
e
ar
p
atter
n
th
at
t
h
e
d
ec
o
m
p
o
s
ed
m
atr
i
x
,
as
w
ell
as
o
r
ig
in
al
m
atr
ice
s
,
d
o
h
av
e
o
n
l
y
p
o
s
iti
v
e
ele
m
en
t
s
.
T
h
is
co
n
ce
p
t
u
s
ed
in
tr
ain
i
n
g
h
as
o
n
e
d
o
m
i
n
an
t
a
d
v
an
ta
g
e,
i.e
.
,
id
en
ti
f
icatio
n
o
f
an
y
an
o
m
al
y
o
b
j
ec
t
d
u
r
in
g
th
e
tr
ai
n
i
n
g
is
m
ai
n
tai
n
ed
in
a
d
i
f
f
er
en
t
m
atr
i
x
w
h
ich
b
ea
r
s
s
ep
ar
ate
in
d
e
x
o
f
n
e
g
ati
v
e
ele
m
en
ts
.
He
n
ce
,
t
h
e
s
y
s
te
m
p
r
o
p
o
s
es
a
s
u
p
er
m
a
tr
ix
,
w
h
er
e
o
n
e
m
at
r
ix
h
o
ld
s
o
n
l
y
n
o
n
-
an
o
m
alo
u
s
in
f
o
r
m
atio
n
,
w
h
er
e
o
t
h
er
h
o
l
d
s
o
n
l
y
in
d
e
x
es
o
f
th
e
ce
ll
p
o
s
itio
n
m
ap
p
i
n
g
w
it
h
a
n
o
m
alo
u
s
in
f
o
r
m
at
io
n
.
T
h
i
s
co
n
ce
p
t
o
f
u
n
s
u
p
er
v
is
ed
tr
ai
n
in
g
n
o
t
o
n
l
y
tr
ai
n
s
f
aster
b
u
t
also
o
f
f
er
en
h
a
n
ce
d
ac
cu
r
ac
y
p
er
f
o
r
m
a
n
ce
w
it
h
9
0
%
o
f
m
e
m
o
r
y
ef
f
ic
ien
c
y
as
th
e
m
atr
i
x
o
n
l
y
s
to
r
es
th
e
in
d
ices
o
f
a
n
o
m
alo
u
s
o
b
j
ec
ts
.
Fo
r
b
etter
in
f
er
en
ce
o
f
th
e
s
tu
d
y
o
u
tco
m
e,
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
i
s
also
co
m
p
ar
ed
w
it
h
o
n
e
o
f
t
h
e
m
o
s
t r
elev
a
n
t s
tu
d
ies b
y
C
o
n
g
et
al.
[
44
]
as sh
o
w
n
i
n
Fi
g
u
r
e
2
.
T
h
e
w
o
r
k
ca
r
r
ied
o
u
t
b
y
C
o
n
g
et
al.
[
44
]
h
av
e
w
o
r
k
ed
to
w
ar
d
s
ad
d
r
ess
i
n
g
a
s
i
m
ilar
p
r
o
b
lem
,
i.e
.
,
d
etec
tio
n
o
f
t
h
e
an
o
m
al
y
f
r
o
m
t
h
e
v
id
eo
f
o
r
as
s
is
t
in
g
i
n
e
v
en
t
d
etec
t
io
n
s
y
s
te
m
.
Ho
w
e
v
er
,
th
e
au
t
h
o
r
s
h
a
v
e
u
s
ed
th
e
s
e
g
m
en
ta
tio
n
-
b
ased
ap
p
r
o
ac
h
o
n
th
e
s
i
m
ila
r
d
atab
ase.
Ho
w
ev
er
,
w
e
h
y
p
o
t
h
eti
ca
ll
y
co
m
p
ar
e
th
e
th
eo
r
etica
l
o
u
tco
m
es
o
f
t
h
e
ex
is
t
in
g
s
y
s
te
m
w
it
h
a
p
r
o
p
o
s
ed
s
y
s
te
m
co
n
ce
r
n
i
n
g
co
n
v
en
t
io
n
al
ac
c
u
r
ac
y
p
ar
am
eter
s
,
e.
g
.
,
r
ec
all,
p
r
ec
is
io
n
,
s
p
ec
if
ic
it
y
,
an
d
F1
-
Sc
o
r
e
to
f
in
d
th
at
p
r
o
p
o
s
ed
s
y
s
te
m
o
f
f
er
s
b
etter
ac
cu
r
ac
y
i
n
id
en
ti
f
icat
io
n
o
f
o
u
tlier
s
co
n
ce
r
n
i
n
g
e
x
is
tin
g
s
y
s
te
m
.
Fig
u
r
e
2
.
C
o
m
p
ar
ativ
e
O
u
tco
m
es o
f
A
cc
u
r
ac
y
4.
CO
NCLU
SI
O
N
T
h
is
p
a
p
er
p
r
esen
ts
a
n
o
v
el
f
r
am
e
w
o
r
k
th
at
em
p
h
asizes
o
n
t
h
e
co
n
t
ex
tu
al
in
f
o
r
m
atio
n
o
f
a
s
ce
n
e.
A
s
a
s
c
en
e
c
an
h
av
e
m
u
ltip
le
n
u
m
b
e
r
s
o
f
h
et
er
o
g
en
e
o
u
s
co
n
t
ex
ts
,
h
en
c
e
w
e
ap
p
ly
m
u
ltiv
ar
i
ate
an
aly
s
is
to
p
er
f
o
r
m
m
atr
ix
d
ec
o
m
p
o
s
it
io
n
.
T
h
e
p
r
es
en
te
d
t
ec
h
n
i
q
u
e
s
i
g
n
if
ican
t
ly
ass
is
ts
in
i
d
en
tif
y
in
g
an
o
u
tlie
r
th
a
t
is
a
ls
o
ca
p
a
b
le
o
f
ex
t
r
a
ctin
g
all
th
e
c
o
n
tex
tu
al
f
e
atu
r
es
.
A
n
alg
o
r
ith
m
is
d
esig
n
e
d
f
o
r
b
l
o
ck
in
g
o
p
e
r
at
io
n
,
u
n
s
u
p
er
v
is
ed
lea
r
n
in
g
u
s
in
g
s
p
a
r
s
ity
f
ac
t
o
r
,
an
d
f
in
ally
,
p
e
r
f
o
r
m
id
en
tif
ic
at
io
n
o
f
th
e
o
b
jects
.
RE
F
E
RE
NCES
[1
]
O.
Ja
v
e
d
,
M
.
S
h
a
h
,
“
A
u
to
m
a
te
d
M
u
lt
i
-
Ca
m
e
ra
S
u
rv
e
il
lan
c
e
:
A
l
g
o
rit
h
m
s
a
n
d
P
ra
c
ti
c
e
"
,
S
p
ri
n
g
e
r
S
c
ien
c
e
&
Bu
sin
e
ss
M
e
d
ia
,
p
p
.
1
1
0
,
2
0
0
8
[2
]
T
.
Wad
a
,
F
.
Hu
a
n
g
,
S
.
L
in
,
"
A
d
v
a
n
c
e
s
in
Im
a
g
e
a
n
d
V
id
e
o
T
e
c
h
n
o
lo
g
y
:
T
h
ird
P
a
c
if
ic
Rim
S
y
m
p
o
siu
m
,
P
S
I
V
T
2
0
0
9
,
T
o
k
y
o
,
Ja
p
a
n
,
Ja
n
u
a
ry
1
3
-
1
6
,
2
0
0
9
,
P
ro
c
e
e
d
i
n
g
s"
,
S
p
rin
g
e
r S
c
ien
c
e
&
Bu
sin
e
ss
M
e
d
ia
,
p
p
.
1
1
1
9
,
2
0
0
9
[3
]
T
.
W
.
Na
tt
k
e
m
p
e
r,
“
M
u
lt
iv
a
riate
ima
g
e
a
n
a
l
y
sis
in
b
i
o
m
e
d
icin
e
”
,
J
o
u
rn
a
l
o
f
Bi
o
me
d
ic
a
l
I
n
fo
rm
a
ti
c
s
,
Vo
l.
3
7
,
n
o
.
5
,
p
p
.
3
8
0
-
3
9
1
,
2
0
0
4
)
[4
]
A
.
Kh
a
n
,
L
.
S
u
n
,
a
n
d
E.
If
e
a
c
h
o
r,
"
Co
n
te
n
t
-
b
a
se
d
v
id
e
o
q
u
a
l
it
y
p
re
d
ictio
n
f
o
r
M
P
EG
4
v
id
e
o
stre
a
m
in
g
o
v
e
r
w
irele
ss
n
e
tw
o
rk
s"
,
J
o
u
rn
a
l
o
f
M
u
lt
ime
d
i
a
,
v
o
l
.
4
,
No
.
4
,
2
0
0
9
[5
]
A
.
Kh
a
n
,
L
.
S
u
n
,
a
n
d
E.
If
e
a
c
h
o
r,
"
Co
n
te
n
t
c
lu
ste
rin
g
b
a
se
d
v
id
e
o
q
u
a
l
it
y
p
re
d
ictio
n
mo
d
e
l
fo
r
M
PE
G4
v
id
e
o
stre
a
min
g
o
v
e
r wire
les
s n
e
two
rk
s"
,
I
n
Co
m
m
u
n
ica
ti
o
n
s,
ICC'
0
9
.
IE
EE
In
ter
n
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
,
p
p
.
1
-
5
,
2
0
0
9
.
[6
]
A
.
Kh
a
n
,
L
.
S
u
n
,
a
n
d
E.
If
e
a
c
h
o
r,
"
Imp
a
c
t
o
f
v
id
e
o
c
o
n
ten
t
o
n
v
id
e
o
q
u
a
li
ty
fo
r
v
id
e
o
o
v
e
r
wire
le
ss
n
e
two
rk
s"
,
In
A
u
to
n
o
m
ic an
d
A
u
to
n
o
m
o
u
s S
y
st
e
m
s,
2
0
0
9
.
ICA
S
'
0
9
.
F
if
th
In
ter
n
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
,
p
p
.
2
7
7
-
2
8
2
,
2
0
0
9
.
[7
]
A
.
Zah
a
re
sc
u
a
n
d
R.
W
il
d
e
s,
“
A
n
o
m
a
lo
u
s
b
e
h
a
v
io
u
r
d
e
tec
ti
o
n
u
sin
g
sp
a
ti
o
tem
p
o
ra
l
o
rien
ted
e
n
e
rg
ies
,
su
b
se
t
in
c
lu
sio
n
h
isto
g
ra
m
c
o
m
p
a
riso
n
a
n
d
e
v
e
n
t
-
d
r
iv
e
n
p
ro
c
e
ss
in
g
,
”
i
n
E
CCV
,
2
0
1
0
.
[8
]
K.
G
.
D
e
rp
a
n
is,
M
.
S
izin
tse
v
,
K.
Ca
n
n
o
n
s,
a
n
d
R.
P
.
W
il
d
e
s,
“
Eff
icie
n
t
a
c
ti
o
n
sp
o
tt
in
g
b
a
se
d
o
n
a
sp
a
c
e
ti
m
e
o
rien
ted
stru
c
tu
re
re
p
re
se
n
tatio
n
,
”
in
P
r
o
c
.
C
o
m
p
u
ter V
isi
o
n
a
n
d
P
a
tt
e
rn
Re
c
o
g
n
it
i
o
n
,
2
0
1
0
.
[9
]
J.
Ki
m
a
n
d
K.
G
ra
u
m
a
n
,
“
Ob
se
r
v
e
lo
c
a
ll
y
,
in
fer
g
lo
b
a
ll
y
:
A
sp
a
c
e
c
ti
me
m
rf
fo
r
d
e
tec
ti
n
g
a
b
n
o
rm
a
l
a
c
ti
v
it
ies
wit
h
in
c
re
me
n
ta
l
u
p
d
a
tes
,
”
in
P
ro
c
.
C
o
m
p
u
ter V
isi
o
n
a
n
d
P
a
tt
e
rn
Re
c
o
g
n
it
io
n
,
2
0
0
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
2
,
A
p
r
il
201
8
:
1
0
9
2
–
1101
1100
[1
0
]
A
.
A
d
a
m
,
E.
Riv
li
n
,
I.
S
h
im
sh
o
n
i,
a
n
d
D.
Re
in
i
tz,
“
R
o
b
u
st
re
a
l
-
t
i
me
u
n
u
s
u
a
l
e
v
e
n
t
d
e
tec
ti
o
n
u
si
n
g
mu
lt
i
p
le
fi
x
e
d
-
lo
c
a
ti
o
n
m
o
n
i
to
rs
,
”
IEE
E
T
ra
n
s.
P
a
tt
A
n
a
l.
M
a
c
h
In
tell
,
2
0
0
8
.
[1
1
]
R.
M
e
h
ra
n
,
A
.
O
y
a
m
a
,
a
n
d
M
.
S
h
a
h
,
“
A
b
n
o
rm
a
l
c
ro
wd
b
e
h
a
v
i
o
r
d
e
tec
ti
o
n
u
sin
g
so
c
ia
l
f
o
rc
e
mo
d
e
l,
”
i
n
P
ro
c
.
Co
m
p
u
ter V
isi
o
n
a
n
d
P
a
tt
e
rn
Re
c
o
g
n
it
io
n
,
2
0
0
9
.
[1
2
]
K.
Ki
m
,
T
.
H.
Ch
a
li
d
a
b
h
o
n
g
se
,
D.
Ha
r
w
o
o
d
,
a
n
d
L
.
Da
v
is,
“
Re
a
l
-
ti
m
e
f
o
re
g
ro
u
n
d
b
a
c
k
g
ro
u
n
d
se
g
m
e
n
tatio
n
u
si
n
g
c
o
d
e
b
o
o
k
m
o
d
e
l,
”
Re
a
l
-
T
i
m
e
I
m
a
g
.
,
2
0
0
5
.
[1
3
]
P
.
F
.
F
e
lze
n
sz
w
a
lb
,
R.
B.
G
irsh
ick
,
D.
A
.
M
c
A
ll
e
ste
r,
a
n
d
D.
Ra
m
a
n
a
n
,
“
Ob
jec
t
d
e
tec
ti
o
n
wi
th
d
isc
rimin
a
ti
v
e
ly
tra
in
e
d
p
a
rt
-
b
a
se
d
mo
d
e
ls,
”
IE
E
E
T
ra
n
s.
P
a
tt
A
n
a
l.
M
a
c
h
In
tell
,
2
0
10
[1
4
]
N.
Da
lal
a
n
d
B.
T
ri
g
g
s,
“
Histo
g
ra
ms
o
f
o
rie
n
ted
g
ra
d
ien
ts
fo
r
h
u
ma
n
d
e
tec
ti
o
n
,
”
i
n
P
r
o
c
.
Co
m
p
u
ter
V
isio
n
a
n
d
P
a
tt
e
r
n
Re
c
o
g
n
it
i
o
n
,
2
0
0
5
[1
5
]
E.
B.
Erm
is,
V
.
S
a
li
g
ra
m
a
,
P
.
M
.
Jo
d
o
i
n
,
a
n
d
J.
Ko
n
ra
d
,
“
M
o
ti
o
n
se
g
me
n
ta
ti
o
n
a
n
d
a
b
n
o
rm
a
l
b
e
h
a
v
io
r
d
e
tec
ti
o
n
v
ia
b
e
h
a
v
io
r cl
u
ste
rin
g
,
”
i
n
P
r
o
c
.
IEE
E
I
n
t.
C
o
n
f
.
Im
a
g
e
P
ro
c
e
ss
in
g
,
2
0
0
8
[1
6
]
Y.
Be
n
e
z
e
th
,
P
.
M
.
Jo
d
o
i
n
,
V.
S
a
li
g
ra
m
a
,
a
n
d
C.
Ro
se
n
b
e
rg
e
r,
“
Ab
n
o
rm
a
l
e
v
e
n
ts
d
e
tec
ti
o
n
b
a
se
d
o
n
sp
a
ti
o
-
tem
p
o
ra
l
c
o
-
o
c
c
u
re
n
c
e
s,”
i
n
P
r
o
c
.
Co
m
p
u
ter
V
isio
n
a
n
d
P
a
tt
e
rn
Re
c
o
g
n
it
io
n
,
2
0
0
9
.
[1
7
]
D
.
D.
L
e
e
a
n
d
H.
S
.
S
e
u
n
g
,
“
L
e
a
r
n
in
g
t
h
e
p
a
rts
o
f
o
b
jec
ts
b
y
n
o
n
n
e
g
a
ti
v
e
m
a
tri
x
f
a
c
to
riza
ti
o
n
,
”
Na
tu
re
,
1
9
9
9
.
[1
8
]
Y.
G
u
o
,
G
.
Din
g
,
X.
Jin
,
a
n
d
J.
W
a
n
g
,
“
L
e
a
rn
in
g
p
re
d
icta
b
le
a
n
d
d
isc
rimin
a
ti
v
e
a
tt
ri
b
u
tes
f
o
r
v
is
u
a
l
re
c
o
g
n
it
io
n
,
”
in
P
r
o
c
.
A
AA
I
Co
n
f
.
A
rti
f
icia
l
I
n
telli
g
e
n
c
e
,
2
0
1
5
.
[1
9
]
J.
K.
Du
tt
a
,
B.
Ba
n
e
rjee
a
n
d
C.
K.
Re
d
d
y
,
"
ROD
S
:
Ra
rit
y
b
a
se
d
Ou
tl
ier
De
tec
ti
o
n
in
a
S
p
a
rse
Co
d
i
n
g
F
ra
m
e
w
o
r
k
,
"
in
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
K
n
o
w
led
g
e
a
n
d
Da
t
a
E
n
g
i
n
e
e
rin
g
,
v
o
l.
2
8
,
n
o
.
2
,
p
p
.
4
8
3
-
4
9
5
,
F
e
b
.
1
2
0
1
6
.
[2
0
]
D.
W
.
Ja
n
g
a
n
d
R.
H.
P
a
rk
,
"
P
o
th
o
le
d
e
tec
ti
o
n
u
sin
g
sp
a
t
io
-
t
e
m
p
o
ra
l
sa
li
e
n
c
y
,
"
in
IET
I
n
tell
ig
e
n
t
T
ra
n
s
p
o
r
t
S
y
ste
ms
,
v
o
l.
1
0
,
n
o
.
9
,
p
p
.
6
0
5
-
6
1
2
,
1
1
2
0
1
6
.
[2
1
]
S
iq
i
W
a
n
g
,
En
Zh
u
,
Jia
n
p
i
n
g
Yin
a
n
d
F
.
P
o
r
ik
li
,
"
An
o
ma
ly
d
e
tec
ti
o
n
in
c
ro
wd
e
d
sc
e
n
e
s
b
y
S
L
-
HO
F
d
e
sc
rip
to
r
a
n
d
fo
re
g
ro
u
n
d
c
l
a
ss
if
i
c
a
ti
o
n
,
"
2
0
1
6
2
3
r
d
In
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
P
a
tt
e
r
n
Re
c
o
g
n
it
io
n
(IC
P
R)
,
Ca
n
c
u
n
,
2
0
1
6
,
p
p
.
3
3
9
8
-
3
4
0
3
.
[2
2
]
F
.
Zh
o
u
a
n
d
F
.
D.
l
.
T
o
rre
,
"
S
p
a
ti
o
-
T
e
m
p
o
ra
l
M
a
tch
in
g
f
o
r
Hu
m
a
n
P
o
se
Esti
m
a
ti
o
n
i
n
Vi
d
e
o
,
"
in
IE
EE
T
ra
n
sa
c
ti
o
n
s
o
n
Pa
tt
e
rn
A
n
a
lys
is
a
n
d
M
a
c
h
i
n
e
In
telli
g
e
n
c
e
,
v
o
l.
3
8
,
n
o
.
8
,
p
p
.
1
4
9
2
-
1
5
0
4
,
A
u
g
.
1
2
0
1
6
.
[2
3
]
Y.
F
u
e
t
a
l
.
,
"
Ro
b
u
st
S
u
b
jec
ti
v
e
V
isu
a
l
P
r
o
p
e
rty
P
re
d
icti
o
n
f
ro
m
Cro
w
d
so
u
rc
e
d
P
a
irw
ise
L
a
b
e
ls,
"
in
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Pa
tt
e
rn
A
n
a
lys
is
a
n
d
M
a
c
h
i
n
e
In
telli
g
e
n
c
e
,
v
o
l.
3
8
,
n
o
.
3
,
p
p
.
5
6
3
-
5
7
7
,
M
a
rc
h
1
2
0
1
6
.
[2
4
]
X
.
L
i
a
n
d
J.
Ha
u
p
t,
"
Id
e
n
ti
fy
in
g
Ou
tl
iers
in
L
a
rg
e
M
a
tri
c
e
s
v
ia
R
a
n
d
o
m
ize
d
A
d
a
p
ti
v
e
Co
m
p
re
ss
iv
e
S
a
m
p
li
n
g
,
"
in
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
S
i
g
n
a
l
Pr
o
c
e
ss
in
g
,
v
o
l.
6
3
,
n
o
.
7
,
p
p
.
1
7
9
2
-
1
8
0
7
,
A
p
ril
1
,
2
0
1
5
.
[2
5
]
G
.
X
u
e
,
L
.
S
o
n
g
a
n
d
J.
S
u
n
,
"
F
o
re
g
ro
u
n
d
Esti
m
a
ti
o
n
Ba
se
d
o
n
L
i
n
e
a
r
Re
g
re
ss
io
n
M
o
d
e
l
W
it
h
F
u
s
e
d
S
p
a
rsity
o
n
Ou
tl
iers
,
"
in
IEE
E
T
ra
n
s
a
c
ti
o
n
s
o
n
Circ
u
it
s
a
n
d
S
y
ste
ms
fo
r
Vi
d
e
o
T
e
c
h
n
o
l
o
g
y
,
v
o
l.
2
3
,
n
o
.
8
,
p
p
.
1
3
4
6
-
1
3
5
7
,
A
u
g
.
2
0
1
3
.
[2
6
]
X
.
Z
h
o
u
,
C.
Ya
n
g
a
n
d
W
.
Yu
,
"
M
o
v
in
g
Ob
jec
t
De
tec
ti
o
n
b
y
De
tec
ti
n
g
Co
n
ti
g
u
o
u
s
Ou
t
li
e
rs
in
th
e
L
o
w
-
Ra
n
k
Re
p
re
se
n
tatio
n
,
"
in
IEE
E
T
r
a
n
sa
c
ti
o
n
s
o
n
Pa
tt
e
rn
A
n
a
lys
is
a
n
d
M
a
c
h
in
e
In
telli
g
e
n
c
e
,
v
o
l.
3
5
,
n
o
.
3
,
p
p
.
5
9
7
-
6
1
0
,
M
a
rc
h
2
0
1
3
.
[2
7
]
R.
G
o
p
a
lan
,
T
.
Ho
n
g
,
M
.
S
h
n
e
ie
r
a
n
d
R.
Ch
e
l
lap
p
a
,
"
A
Lea
rn
in
g
A
p
p
ro
a
c
h
T
o
w
a
rd
s
D
e
tec
ti
o
n
a
n
d
T
ra
c
k
in
g
o
f
L
a
n
e
M
a
rk
in
g
s
,
"
in
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
I
n
telli
g
e
n
t
T
ra
n
sp
o
rta
t
i
o
n
S
y
ste
ms
,
v
o
l.
1
3
,
n
o
.
3
,
p
p
.
1
0
8
8
-
1
0
9
8
,
S
e
p
t
.
2
0
1
2
.
[2
8
]
S
.
Ya
n
g
a
n
d
B.
Bh
a
n
u
,
"
Un
d
e
rst
a
n
d
in
g
Disc
re
te
F
a
c
ial
Ex
p
re
ss
io
n
s
in
V
id
e
o
Us
in
g
a
n
Em
o
ti
o
n
A
v
a
tar
I
m
a
g
e
,
"
in
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
S
y
ste
ms
,
M
a
n
,
a
n
d
Cy
b
e
rn
e
ti
c
s,
Pa
rt
B
(
Cy
b
e
rn
e
ti
c
s)
,
v
o
l.
4
2
,
n
o
.
4
,
p
p
.
9
8
0
-
9
9
2
,
A
u
g
.
2
0
1
2
.
[2
9
]
B.
Ni,
Z.
S
o
n
g
a
n
d
S
.
Ya
n
,
"
W
e
b
Im
a
g
e
a
n
d
V
id
e
o
M
i
n
in
g
T
o
w
a
rd
s Un
iv
e
rsa
l
a
n
d
Ro
b
u
st A
g
e
Esti
m
a
to
r,
"
in
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
M
u
lt
ime
d
i
a
,
v
o
l.
1
3
,
n
o
.
6
,
p
p
.
1
2
1
7
-
1
2
2
9
,
De
c
.
2
0
1
1
.
[3
0
]
H.
Am
m
a
r
a
n
d
S
.
L
a
sh
k
a
r,
"
Ob
stru
c
ti
v
e
sle
e
p
a
p
n
e
a
d
iag
n
o
sis
b
a
s
e
d
o
n
a
sta
ti
stica
l
a
n
a
l
y
sis
o
f
th
e
o
p
ti
c
a
l
f
lo
w
in
v
id
e
o
re
c
o
rd
i
n
g
s,"
2
0
1
6
In
tern
a
ti
o
n
a
l
S
y
m
p
o
siu
m
o
n
S
ig
n
a
l,
Im
a
g
e
,
V
id
e
o
a
n
d
C
o
m
m
u
n
ica
ti
o
n
s
(
IS
IV
C)
,
T
u
n
is,
2
0
1
6
,
p
p
.
1
8
-
2
3
.
[3
1
]
J.
Ch
o
i
a
n
d
J.
Y.
Ch
o
i,
"
P
a
tch
-
b
a
se
d
f
ire
d
e
tec
ti
o
n
w
it
h
o
n
li
n
e
o
u
tl
ier
le
a
rn
i
n
g
,
"
2
0
1
5
1
2
t
h
IEE
E
In
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
A
d
v
a
n
c
e
d
V
i
d
e
o
a
n
d
S
ig
n
a
l
Ba
se
d
S
u
rv
e
il
lan
c
e
(AV
S
S
)
,
Ka
rlsru
h
e
,
2
0
1
5
,
p
p
.
1
-
6.
[3
2
]
R.
F
e
ris,
L
.
M
.
Bro
w
n
,
S
.
P
a
n
k
a
n
ti
a
n
d
M
.
T
.
S
u
n
,
"
A
p
p
e
a
ra
n
c
e
-
Ba
se
d
Ob
jec
t
De
tec
ti
o
n
Un
d
e
r
Va
ry
in
g
En
v
iro
n
me
n
ta
l
Co
n
d
i
ti
o
n
s,"
2
0
1
4
2
2
n
d
I
n
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
P
a
tt
e
rn
Re
c
o
g
n
it
io
n
,
S
t
o
c
k
h
o
lm
,
2
0
1
4
,
p
p
.
1
6
6
-
1
7
1
.
[3
3
]
P
.
Ja
y
a
su
g
a
n
th
i,
V
.
Je
y
a
p
ra
b
h
a
,
P
.
M
.
A
.
Ku
m
a
r
a
n
d
V.
V
a
id
e
h
i,
"
De
tec
ti
o
n
o
f
d
ro
p
p
e
d
n
o
n
p
ro
tru
d
i
n
g
o
b
jec
ts
i
n
v
id
e
o
su
rv
e
il
la
n
c
e
u
sin
g
c
lu
ste
re
d
d
a
t
a
stre
a
m,"
2
0
1
3
In
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
Re
c
e
n
t
T
r
e
n
d
s
in
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
lo
g
y
(ICR
T
I
T
)
,
Ch
e
n
n
a
i,
2
0
1
3
,
p
p
.
3
7
1
-
3
7
5
.
[3
4
]
Y.
L
iu
,
H.
L
iu
,
Y.
L
iu
a
n
d
F
.
S
u
n
,
"
O
u
tl
ier
-
a
tt
e
n
u
a
ti
n
g
su
mm
a
riza
ti
o
n
f
o
r
u
se
r
-
g
e
n
e
ra
ted
-
v
id
e
o
,
"
2
0
1
4
I
EE
E
In
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
M
u
lt
im
e
d
ia an
d
Ex
p
o
(ICM
E)
,
C
h
e
n
g
d
u
,
2
0
1
4
,
p
p
.
1
-
6.
[3
5
]
C.
K.
M
a
u
ry
a
a
n
d
D.
T
o
sh
n
iw
a
l,
"
A
n
o
ma
ly
d
e
tec
ti
o
n
i
n
n
u
c
lea
r
p
o
we
r
p
la
n
t
d
a
t
a
u
si
n
g
su
p
p
o
rt
v
e
c
to
r
d
a
ta
d
e
sc
rip
ti
o
n
,
"
S
tu
d
e
n
ts'
T
e
c
h
n
o
l
o
g
y
S
y
m
p
o
siu
m
(Tec
h
S
y
m
),
2
0
1
4
I
EE
E
,
Kh
a
ra
g
p
u
r
,
2
0
1
4
,
p
p
.
8
2
-
86.
[3
6
]
K.
L
.
P
a
n
g
,
G
.
H.
Ch
e
n
a
n
d
W
.
G
.
T
e
n
g
,
"
Disc
o
v
e
rin
g
u
n
u
su
a
l
b
e
h
a
v
io
r
p
a
tt
e
rn
s
fro
m
mo
ti
o
n
d
a
t
a
,
"
2
0
1
3
IEE
E
In
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
Co
n
s
u
m
e
r
El
e
c
tro
n
ics
(ICCE)
,
L
a
s V
e
g
a
s,
NV
,
2
0
1
3
,
p
p
.
2
4
2
-
2
4
3
.
[3
7
]
Y.
P
rit
c
h
,
S
.
Ra
to
v
it
c
h
,
A
.
He
n
d
e
l
a
n
d
S
.
P
e
leg
,
"
Clu
ste
re
d
S
y
n
o
p
sis
o
f
S
u
rv
e
il
l
a
n
c
e
Vi
d
e
o
,
"
2
0
0
9
S
ix
th
IEE
E
In
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
A
d
v
a
n
c
e
d
Vid
e
o
a
n
d
S
ig
n
a
l
Ba
se
d
S
u
r
v
e
il
lan
c
e
,
Ge
n
o
v
a
,
2
0
0
9
,
p
p
.
1
9
5
-
2
0
0
.
[3
8
]
M
in
g
y
a
n
Ten
g
,
"
An
o
ma
ly
d
e
te
c
ti
o
n
o
n
ti
me
se
rie
s,"
2
0
1
0
I
EE
E
I
n
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
P
ro
g
re
ss
in
In
f
o
rm
a
ti
c
s an
d
Co
m
p
u
ti
n
g
,
S
h
a
n
g
h
a
i,
2
0
1
0
,
p
p
.
6
0
3
-
6
0
8
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
F
r
a
mewo
r
k
fo
r
C
o
n
textu
a
l O
u
tlier
I
d
en
tifi
ca
tio
n
u
s
in
g
Mu
ltiva
r
ia
te
A
n
a
lysi
s
….
(
Ma
d
h
u
C
h
a
n
d
r
a
G
)
1101
[3
9
]
F
a
rsh
a
d
Ba
y
a
t,
M
o
h
a
m
m
a
d
S
h
a
h
ra
m
M
o
in
,
F
a
rh
a
d
Ba
y
a
t,
"
G
o
a
l
De
tec
ti
o
n
in
S
o
c
c
e
r
V
id
e
o
:
R
o
l
e
-
Ba
se
d
Ev
e
n
ts
De
tec
ti
o
n
A
p
p
ro
a
c
h
"
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
mp
u
ter
E
n
g
in
e
e
rin
g
(
IJ
ECE
)
Vo
l.
4
,
N
o
.
6
,
p
p
.
9
7
9
~
9
8
8
,
De
c
e
m
b
e
r
2
0
1
4
.
[4
0
]
S
taffy
Kin
g
ra
,
Na
v
e
e
n
A
g
g
a
r
wa
l,
Ra
a
h
a
t
De
v
e
n
d
e
r
S
in
g
h
,
"
V
i
d
e
o
In
ter
-
f
ra
m
e
F
o
rg
e
r
y
D
e
tec
ti
o
n
A
p
p
ro
a
c
h
f
o
r
S
u
rv
e
il
lan
c
e
a
n
d
M
o
b
il
e
Re
c
o
r
d
e
d
V
i
d
e
o
s"
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
t
e
r
En
g
i
n
e
e
rin
g
(
IJ
ECE
)
V
o
l.
7
,
N
o
.
2
,
p
p
.
8
3
1
~
8
4
1
,
A
p
ril
2
0
1
7
.
[4
1
]
T
e
d
d
y
S
u
ry
a
G
u
n
a
w
a
n
,
A
b
d
u
l
M
u
th
o
li
b
,
M
ira
Ka
rti
w
i,
"
P
e
rf
o
rm
a
n
c
e
Ev
a
lu
a
ti
o
n
o
f
A
u
to
m
a
ti
c
Nu
m
b
e
r
P
late
Re
c
o
g
n
it
io
n
o
n
A
n
d
r
o
id
S
m
a
rtp
h
o
n
e
P
latf
o
rm
"
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
t
e
r
En
g
in
e
e
rin
g
(
IJ
ECE
)
V
o
l.
7
,
N
o
.
4
,
p
p
.
1
9
7
3
~
1
9
8
2
,
A
u
g
u
st
2
0
1
7
.
[4
2
]
M
a
d
h
u
Ch
a
n
d
ra
G
.
a
n
d
S
re
e
ra
m
a
Re
d
d
y
G
M
.
“
In
sig
h
ts
to
V
id
e
o
A
n
a
l
y
ti
c
M
o
d
e
ll
in
g
A
p
p
ro
a
c
h
with
F
u
t
u
re
L
in
e
o
f
Re
se
a
r
c
h
”
.
In
ter
n
a
t
io
n
a
l
J
o
u
r
n
a
l
o
f
Co
m
p
u
ter
A
p
p
l
ica
ti
o
n
s
1
4
7
(
7
):1
5
-
2
4
,
A
u
g
u
st 2
0
1
6
[4
3
]
M
a
d
h
u
Ch
a
n
d
ra
G
.
a
n
d
S
re
e
ra
m
a
Re
d
d
y
G
M
.
“
A
n
a
l
y
ti
c
a
l
F
ra
m
e
w
o
rk
f
o
r
Id
e
n
ti
f
ica
ti
o
n
o
f
Ou
tl
ier
s
f
o
r
Un
sc
rip
ted
V
id
e
o
”
,
J
o
u
rn
a
l
o
f
Co
m
p
u
ter
S
c
i
e
n
c
e
, v
o
l.
1
4
,
Oc
to
b
e
r
1
6
,
2
0
1
6
[4
4
]
Y.
Co
n
g
,
J.
Yu
a
n
,
Y.
T
a
n
g
,
“
V
i
d
e
o
A
n
o
m
a
l
y
S
e
a
rc
h
in
Cro
w
d
e
d
S
c
e
n
e
s
v
ia
S
p
a
ti
o
-
tem
p
o
ra
l
M
o
ti
o
n
C
o
n
tex
t”
,
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
In
f
o
rm
a
ti
o
n
Fo
re
n
sic
s
a
n
d
S
e
c
u
rity
,
v
o
l.
8
,
I
ss
.
1
0
,
2
0
1
3
B
IO
G
RA
PHI
ES
O
F
AU
TH
O
R
S
M
r
.
M
a
d
h
u
C
h
a
n
d
r
a
G
,
P
u
rs
u
in
g
Do
c
to
ra
l
De
g
re
e
in
th
e
d
o
m
a
in
o
f
I
m
a
g
e
a
n
d
Vid
e
o
p
ro
c
e
ss
in
g
f
ro
m
V
isv
e
s
v
a
ra
y
a
Tec
h
n
o
l
o
g
ica
l
Un
iv
e
rsity
(V
T
U),
Be
lag
a
v
i.
Re
c
e
iv
e
d
th
e
B.
E
a
n
d
M
.
T
e
c
h
De
g
re
e
f
ro
m
V
T
U,
Be
lag
a
v
i,
In
d
ia
i
n
2
0
0
6
a
n
d
2
0
0
8
,
re
sp
e
c
ti
v
e
l
y
.
P
re
se
n
tl
y
h
e
is
w
o
rk
in
g
a
s
A
ss
ist
a
n
t
p
ro
f
e
ss
o
r
in
th
e
De
p
t
.
o
f
El
e
c
tro
n
ics
a
n
d
C
o
m
m
u
n
ica
ti
o
n
En
g
in
e
e
ri
n
g
a
t
RL
JI
T
,
Do
d
d
a
b
a
ll
a
p
u
r
.
He
is
h
a
v
i
n
g
to
tal
tea
c
h
in
g
e
x
p
e
rien
c
e
o
f
1
0
y
e
a
r
s an
d
a
lso
su
c
c
e
ss
f
u
ll
y
g
u
id
e
d
m
a
n
y
U
G
P
ro
jec
ts
a
n
d
a
lso
p
a
rti
c
ip
a
ted
v
a
rio
u
s
w
o
rk
sh
o
p
s
a
n
d
f
a
c
u
lt
y
d
e
v
e
lo
p
m
e
n
t
p
ro
g
ra
m
m
e
.
He
is
a
m
e
m
b
e
r
o
f
I
EE
E,
M
I
E,
a
n
d
IET
E.
Dr
.
S
r
e
e
r
a
m
a
Re
d
d
y
G
M
,
P
r
o
f
e
ss
o
r
&
HO
D
o
f
El
e
c
tr
o
n
ics
a
n
d
C
o
m
m
u
n
ica
ti
o
n
En
g
in
e
e
rin
g
,
C
By
re
g
o
w
d
a
In
stit
u
te
o
f
T
e
c
h
n
o
lo
g
y
-
Ko
lar,
In
d
i
a
.
Re
c
e
iv
e
d
P
h
.
D
i
n
M
icr
o
El
e
c
tro
n
ics
/
V
L
S
I
De
sig
n
f
ro
m
D
e
p
t.
o
f
El
e
c
tro
n
ics
a
n
d
C
o
m
m
u
n
ica
ti
o
n
E
n
g
in
e
e
rin
g
,
JN
T
U,
A
n
d
h
ra
P
ra
d
e
sh
,
i
n
2
0
0
9
a
n
d
M
E
f
ro
m
UV
CE
Ba
n
g
a
lo
re
in
th
e
y
e
a
r
1
9
9
8
.
He
h
a
s
2
6
y
e
a
rs
o
f
tea
c
h
in
g
a
n
d
9
y
e
a
rs
o
f
re
se
a
rc
h
e
x
p
e
rien
c
e
a
n
d
p
u
b
l
ish
e
d
a
ro
u
n
d
3
0
p
lu
s
i
n
n
a
ti
o
n
a
l
a
n
d
in
tern
a
ti
o
n
a
l
j
o
u
r
n
a
ls
a
n
d
c
o
n
f
e
r
e
n
c
e
s.
Co
n
d
u
c
ted
2
5
v
a
rio
u
s
w
o
rk
sh
o
p
s
a
n
d
h
a
n
d
s
o
n
sk
il
l
s
a
c
ti
v
it
ies
f
o
r
b
o
th
sta
f
f
a
n
d
stu
d
e
n
ts.
He
h
a
s
i
n
it
iate
d
v
a
rio
u
s
a
c
ti
v
it
ies
a
s
P
r
in
c
ip
a
l
in
p
re
v
io
u
s
c
o
ll
e
g
e
s an
d
re
c
e
iv
e
d
v
a
rio
u
s f
u
n
d
s f
ro
m
lea
d
in
g
o
rg
a
n
iza
ti
o
n
s.
Evaluation Warning : The document was created with Spire.PDF for Python.