I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
p
u
t
er
Science
Vo
l.
21
,
No
.
2
,
Feb
r
u
ar
y
202
1
,
p
p
.
8
3
9
~
8
4
5
I
SS
N:
2
5
02
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/i
j
ee
cs.v
2
1
.i
2
.
p
p
8
3
9
-
8
4
5
839
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Co
ntent
bas
ed vi
deo
ret
ri
ev
a
l usin
g
discrete
cos
ine t
ra
nsfo
r
m
Su
m
a
y
a
H
a
m
a
d
1
,
Ah
m
ee
d S
uli
m
a
n F
a
rha
n
2
,
Do
a
a
Ya
s
ee
n K
hu
dh
ur
3
1,
3
Co
ll
e
g
e
o
f
Co
m
p
u
ter S
c
ien
c
e
a
n
d
I
n
f
o
rm
a
ti
o
n
T
e
c
h
n
o
lo
g
y
,
Un
iv
e
rsit
y
o
f
A
n
b
a
r,
A
n
b
a
r,
Ira
q
2
Co
m
p
u
ter Cen
ter,
Un
iv
e
rsity
o
f
A
n
b
a
r,
A
n
b
a
r,
Ira
q
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
A
p
r
1
8
,
2
0
2
0
R
ev
i
s
ed
J
u
n
15
,
2
0
2
0
A
cc
ep
ted
J
u
n
21
,
2
0
2
0
A
c
o
n
ten
t
b
a
se
d
v
id
e
o
re
tri
e
v
a
l
(CBV
R)
f
ra
m
e
w
o
rk
is
b
u
il
t
i
n
t
h
is
p
a
p
e
r.
On
e
o
f
th
e
e
ss
e
n
ti
a
l
f
e
a
tu
re
s
o
f
v
id
e
o
re
tri
e
v
a
l
p
ro
c
e
ss
a
n
d
CBV
R
is
a
c
o
lo
r
v
a
lu
e
.
T
h
e
d
isc
re
te
c
o
sin
e
tran
sf
o
rm
(DC
T
)
is
u
se
d
to
e
x
trac
t
a
q
u
e
ry
v
id
e
o
f
e
a
tu
re
s
to
c
o
m
p
a
re
w
it
h
th
e
v
id
e
o
f
e
a
tu
re
s
sto
re
d
in
o
u
r
d
a
tab
a
s
e
.
A
v
e
ra
g
e
re
su
lt
o
f
0
.
6
4
7
5
w
a
s
o
b
tain
e
d
b
y
u
sin
g
th
e
DCT
a
f
ter
i
m
p
le
m
e
n
ti
n
g
it
to
t
h
e
d
a
tab
a
se
w
e
c
r
e
a
ted
a
n
d
c
o
ll
e
c
te
d
,
a
n
d
o
n
a
ll
c
a
teg
o
ries
.
T
h
is
tec
h
n
i
q
u
e
w
a
s
a
p
p
li
e
d
o
n
o
u
r
d
a
tab
a
se
o
f
v
id
e
o
,
c
h
e
c
k
1
0
0
d
a
tab
a
se
v
id
e
o
s
,
5
v
id
e
o
s
i
n
e
a
c
h
c
a
teg
o
r
y
.
K
ey
w
o
r
d
s
:
C
o
n
te
n
t b
ased
v
id
eo
r
etr
iev
al
Descr
ete
co
s
in
e
tr
a
n
s
f
o
r
m
Vid
eo
r
etr
iev
al
Vid
eo
r
etr
iev
al
p
r
o
ce
s
s
T
h
is
is
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Su
m
a
y
a
Ha
m
ad
Dep
ar
t
m
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
C
o
lleg
e
o
f
C
o
m
p
u
ter
Scie
n
ce
an
d
I
n
f
o
r
m
atio
n
T
ec
h
n
o
lo
g
y
An
b
ar
Un
i
v
er
s
it
y
,
A
n
b
ar
,
I
r
aq
E
m
ail: s
u
m
a
y
ah
.
h
a
m
ad
@
u
o
an
b
ar
.
ed
u
.
iq
1.
I
NT
RO
D
UCT
I
O
N
Dig
ital
d
ata
s
ee
m
s
to
b
e
an
i
m
p
o
r
tan
t
p
ar
t
o
f
o
u
r
liv
es.
Di
g
it
al
d
ata
in
v
o
l
v
es
v
id
eo
s
,
i
m
ag
e
s
,
au
d
io
s
,
d
o
cu
m
en
ts
,
etc.
Vid
eo
s
co
n
s
ti
tu
te
a
n
a
m
p
le
s
o
u
r
ce
o
f
in
f
o
r
m
atio
n
.
Vid
eo
m
a
y
al
s
o
i
n
clu
d
e
all
o
th
er
d
i
g
ital
co
n
ten
t,
s
u
c
h
as
p
h
o
to
s
,
v
o
ic
es,
an
d
tex
ts
.
I
t
is
f
u
r
t
h
er
d
is
tin
g
u
is
h
ed
b
y
its
te
m
p
o
r
al
co
n
s
is
ten
c
y
.
Dig
ital
d
ev
ices
'
f
ast
d
ev
elo
p
m
e
n
t
is
c
au
s
i
n
g
i
n
f
latio
n
in
th
e
v
id
eo
s
to
r
e.
T
h
e
r
etr
iev
al
o
f
t
h
e
n
ec
ess
ar
y
in
f
o
r
m
at
io
n
f
r
o
m
t
h
e
v
id
eo
d
atab
ase
n
a
m
ed
a
v
id
eo
r
etr
iev
al
p
r
o
ce
s
s
ac
co
r
d
in
g
to
u
s
er
n
ee
d
s
.
Vid
eo
r
et
r
iev
al
is
a
b
r
an
ch
ed
f
ield
ca
lled
in
f
o
r
m
a
tio
n
r
etr
iev
al.
Data
r
etr
iev
al
i
s
co
n
s
id
er
ed
a
s
u
b
-
f
ield
o
f
co
m
p
u
ter
s
cie
n
ce
t
h
at
ar
r
an
g
es
an
d
r
etr
ie
v
es
d
ata
f
r
o
m
lar
g
e
s
ets
o
f
d
atab
ase
s
.
Vid
eo
s
r
etr
iev
al
ap
p
r
o
ac
h
es
ar
e
i
m
p
o
r
tan
t
an
d
n
ec
es
s
ar
y
f
o
r
m
u
lti
m
ed
ia
a
p
p
l
icatio
n
s
lik
e
v
id
eo
s
ea
r
ch
en
g
i
n
es,
d
ig
ital
ar
c
h
iv
e
s
,
v
is
u
al
-
on
-
d
e
m
a
n
d
b
r
o
ad
ca
s
tin
g
,
etc.
[1
-
5]
.
T
o
d
ay
,
th
e
a
m
o
u
n
t
o
f
o
p
en
m
u
lti
m
ed
ia
d
ata
is
g
r
o
w
i
n
g
s
i
g
n
i
f
ican
tl
y
a
s
in
f
o
r
m
at
io
n
te
c
h
n
o
lo
g
y
a
n
d
m
u
lti
m
ed
ia
s
tr
ate
g
ie
s
d
ev
elo
p
r
ap
id
ly
.
I
n
m
aj
o
r
ap
p
licatio
n
s
li
k
e
m
o
n
i
to
r
in
g
,
en
ter
tai
n
m
e
n
t,
h
ea
lt
h
ca
r
e,
lear
n
in
g
,
a
n
d
s
p
o
r
ts
,
v
id
eo
is
h
ea
v
i
l
y
co
n
s
u
m
ed
.
Sear
ch
i
n
g
f
o
r
th
e
ele
m
e
n
t
s
r
eq
u
ir
ed
i
n
t
h
ese
lar
g
e
d
ata
o
n
th
e
I
n
ter
n
et
ca
n
b
e
co
n
s
id
er
e
d
an
i
m
p
o
r
tan
t
ch
al
len
g
e.
T
h
er
ef
o
r
e;
n
u
m
er
o
u
s
v
id
eo
r
etr
iev
al
s
y
s
te
m
s
h
a
v
e
b
ee
n
in
tr
o
d
u
ce
d
f
o
r
th
i
s
i
n
te
n
t.
Vid
eo
r
etr
iev
al
h
as
t
w
o
ap
p
r
o
ac
h
es:
te
x
t
-
b
ased
s
tr
u
ctu
r
e,
an
d
co
n
te
n
t
-
b
ased
f
r
a
m
e
w
o
r
k
s
.
T
ex
t
d
escr
ip
to
r
i
s
u
s
ed
in
te
x
t
-
b
ased
s
tr
u
ct
u
r
e
to
m
a
n
u
all
y
a
n
n
o
tate
v
id
eo
.
T
h
e
s
ea
r
ch
in
th
i
s
m
et
h
o
d
r
elies
o
n
th
e
r
ele
v
an
t
m
etad
ata
f
o
r
.
f
ile,
s
u
ch
a
s
tag
s
,
titl
es,
d
escr
ip
tio
n
a
n
d
k
e
y
wo
r
d
s
.
T
h
e
d
o
w
n
s
id
e
o
f
t
h
is
m
et
h
o
d
to
m
an
u
al
an
n
o
tatio
n
i
n
clu
d
e
s
h
u
m
a
n
itar
ia
n
j
o
b
.
C
o
n
ten
t
-
b
ased
s
tr
u
c
tu
r
e
i
s
s
o
l
v
i
n
g
th
e
ab
o
v
e
d
r
a
w
b
ac
k
.
T
h
e
co
n
ten
t
-
b
ased
ar
ch
itect
u
r
e
allo
w
s
t
h
e
ap
p
licatio
n
to
ac
ce
s
s
a
v
id
eo
clip
f
r
o
m
a
lis
t
o
f
v
is
u
al
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
21
,
No
.
2
,
Feb
r
u
ar
y
2021
:
8
3
9
-
8
4
5
840
co
n
ten
t
-
b
ased
v
id
eo
s
th
at
ar
e
en
tire
l
y
a
u
to
m
atica
ll
y
e
x
tr
ac
ted
,
s
u
ch
a
s
co
lo
r
,
tex
tu
r
e
,
s
h
ap
es,
an
d
n
o
t
attr
ib
u
tes t
h
at
ar
e
u
n
r
elate
d
to
th
e
co
n
te
n
t.
Vis
u
al
f
o
r
m
h
o
ld
s
litt
le
to
n
o
s
e
m
an
tic
v
id
eo
co
n
ten
t
[6
-
11]
.
First,
t
h
e
v
id
eo
is
s
p
lit
i
n
to
f
r
a
m
es,
th
e
n
f
r
a
m
es
ar
e
b
r
o
k
en
u
p
i
n
to
i
m
ag
e
s
.
B
y
i
m
ag
e
s
e
g
m
en
tatio
n
th
e
o
b
j
ec
t
is
i
s
o
lated
f
r
o
m
t
h
e
i
m
a
g
e.
A
p
ar
t
o
f
a
n
i
m
a
g
e
is
t
h
e
s
e
g
m
e
n
ted
o
b
j
ec
t.
E
x
tr
ac
t
f
ea
t
u
r
e
f
r
o
m
th
e
s
eg
m
e
n
ted
i
m
a
g
e
(
o
b
j
ec
t)
[
1
2
]
.
E
x
tr
ac
ted
f
ea
tu
r
es
f
r
o
m
v
id
eo
s
co
u
ld
b
e
in
d
ex
ed
an
d
s
ea
r
ch
ed
.
Vid
eo
s
ea
r
ch
b
y
id
e
n
ti
f
y
in
g
it a
s
o
n
e
o
f
th
e
ev
alu
a
tio
n
ta
s
k
s
[
1
3
]
.
C
o
n
te
n
t
B
ased
Vid
eo
R
etr
iev
al
(
C
B
VR
)
r
ec
o
g
n
ized
as
m
o
s
t
o
f
th
e
s
tr
o
n
g
es
t
f
u
n
c
tio
n
al
a
p
p
r
o
ac
h
es
f
o
r
en
h
a
n
ce
d
q
u
ali
t
y
o
f
r
etr
iev
al.
B
ec
au
s
e
o
f
th
e
u
s
e
o
f
w
e
alth
y
m
a
ter
ial
in
th
e
v
id
eo
,
th
er
e
is
m
a
s
s
i
v
e
s
co
p
e
in
t
h
e
f
ield
o
f
v
id
eo
r
etr
iev
al
to
b
o
o
s
t
tr
ad
itio
n
al
s
ea
r
c
h
e
n
g
i
n
e
ef
f
icie
n
c
y
.
T
h
is
lead
s
t
h
e
C
B
VR
f
ie
ld
i
n
a
d
ir
ec
ti
o
n
p
r
o
m
is
i
n
g
f
u
t
u
r
e
d
ev
elo
p
m
e
n
t
o
f
m
o
r
e
s
u
cc
es
s
f
u
l
v
id
eo
s
ea
r
ch
en
g
i
n
es.
Stro
n
g
an
d
ex
p
an
s
i
v
e
r
an
g
e
o
f
C
B
VR
an
d
C
B
VR
s
y
s
te
m
s
is
p
r
esen
ted
in
a
s
i
m
p
le
an
d
d
etailed
m
a
n
n
er
.
T
h
e
p
r
o
ce
s
s
es
ar
e
r
ep
r
esen
ted
in
a
s
y
s
te
m
a
tic
w
a
y
a
t
v
ar
io
u
s
s
tag
es
o
f
C
B
V
R
s
y
s
te
m
s
.
I
t
al
s
o
d
is
p
la
y
s
v
ar
ieties
o
f
f
ea
tu
r
es,
th
eir
v
ar
iatio
n
s
,
an
d
ap
p
r
o
ac
h
es,
tech
n
iq
u
es,
a
n
d
alg
o
r
ith
m
s
f
o
r
th
eir
u
s
e.
Dif
f
er
en
t
q
u
er
y
in
g
m
et
h
o
d
s
,
s
o
m
e
f
ea
t
u
r
es
s
u
ch
a
s
GL
C
M,
Gab
o
r
Ma
g
n
itu
d
e,
s
i
m
ilar
it
y
a
lg
o
r
it
h
m
s
u
c
h
as
K
u
llb
ac
k
-
L
eib
ler
d
is
tan
ce
m
eth
o
d
an
d
R
elev
an
ce
Feed
b
ac
k
m
et
h
o
d
ar
e
ad
d
r
ess
ed
[
1
4
]
.
I
m
p
le
m
e
n
ti
n
g
au
to
m
atic
in
d
ex
i
n
g
o
f
v
id
eo
s
,
v
id
e
o
s
ea
r
ch
in
a
b
r
o
ad
v
id
eo
d
atab
ase
t
h
en
d
is
p
la
y
i
n
g
tailo
r
ed
r
es
u
lts
as
w
ell.
T
h
e
Su
g
g
e
s
ted
s
y
s
te
m
o
p
er
ates i
n
th
r
ee
s
ep
ar
ate
s
tep
s
,
w
it
h
v
id
eo
s
e
g
m
en
tatio
n
in
t
h
e
f
ir
s
t
p
h
ase,
k
e
y
f
r
a
m
e
d
et
ec
tio
n
r
etr
iev
e
r
elev
a
n
t
k
e
y
f
r
a
m
es.
Seco
n
d
,
f
o
r
ex
tr
ac
ti
n
g
te
x
t
k
e
y
w
o
r
d
,
OC
R
,
HOG
an
d
A
S
R
al
g
o
r
ith
m
s
ar
e
u
s
ed
o
v
er
th
e
k
e
y
f
r
a
m
e.
C
o
lo
r
,
tex
tu
r
e
an
d
ed
g
e
f
ea
t
u
r
es
w
ill
al
s
o
b
e
ex
t
r
ac
ted
in
th
e
t
h
ir
d
s
tep
.
E
v
en
t
u
all
y
,
a
n
al
y
s
i
s
o
f
t
h
e
s
ea
r
ch
s
i
m
ilar
it
y
is
ca
r
r
ied
o
u
t
o
n
th
e
e
x
tr
ac
ted
f
ea
t
u
r
es
t
h
at
ar
e
s
to
r
ed
in
th
e
SQ
L
d
at
ab
ase
a
n
d
th
e
r
esu
l
t
is
p
r
o
v
id
ed
to
th
e
u
s
er
s
w
it
h
cu
s
to
m
ized
r
e
-
r
an
k
e
d
r
esu
lt
s
ac
co
r
d
in
g
to
in
ter
est
[
1
5
]
.
Vi
d
eo
r
etr
iev
al
m
ac
h
i
n
e
lear
n
i
n
g
s
y
s
te
m
,
co
m
p
o
s
ed
o
f
s
h
o
t
e
x
tr
ac
tio
n
m
o
d
u
le,
k
e
y
f
r
a
m
e
e
x
tr
ac
tio
n
m
o
d
u
le
an
d
v
id
eo
r
etr
iev
al
m
o
d
u
le
ar
e
als
o
p
r
o
v
id
ed
.
On
e
o
r
m
o
r
e
k
e
y
-
f
r
a
m
es
ca
n
t
h
en
s
u
m
m
ar
ize
ea
c
h
s
h
o
t.
Vid
eo
v
ie
w
i
n
g
a
n
d
r
etr
iev
al
ca
n
also
b
e
co
n
s
id
er
ed
an
is
s
u
e
o
f
clas
s
i
f
icatio
n
.
I
n
r
ea
li
t
y
,
th
eir
r
etr
iev
al
m
o
d
u
le
is
b
a
s
ed
o
n
k
er
n
el
-
b
ased
SVM
(
S
u
p
p
o
r
t
Vec
to
r
Ma
ch
i
n
e)
,
w
h
ic
h
is
n
o
w
r
ep
r
ese
n
ted
in
t
h
e
ac
tiv
e
lear
n
i
n
g
m
et
h
o
d
,
u
s
e
s
th
e
ex
tr
ac
ted
k
e
y
f
r
a
m
e
s
[
1
6
]
.
Af
ter
t
w
o
y
ea
r
s
,
a
g
r
o
u
n
d
b
r
ea
k
in
g
ap
p
r
o
ac
h
f
o
r
ac
h
iev
i
n
g
s
ig
n
if
ican
t
l
y
h
ig
h
-
q
u
ali
t
y
co
n
tex
t
-
b
ased
v
id
eo
r
etr
iev
al
is
p
r
o
p
o
s
ed
b
y
id
en
tify
i
n
g
te
m
p
o
r
al
tr
en
d
s
i
n
v
id
eo
co
n
te
n
t.
B
ased
o
n
th
e
te
m
p
o
r
al
p
atter
n
s
d
is
co
v
er
ed
,
an
ef
f
ec
t
iv
e
i
n
d
e
x
in
g
s
tr
ateg
y
a
n
d
an
e
f
f
icie
n
t
tech
n
iq
u
e
f
o
r
m
atc
h
i
n
g
s
eq
u
en
ce
s
ar
e
co
m
b
i
n
ed
r
ed
u
cin
g
co
m
p
u
ti
n
g
ex
p
en
s
e
s
an
d
i
m
p
r
o
v
e
th
e
p
r
ec
is
io
n
o
f
th
e
r
etr
iev
al
r
esp
ec
tiv
el
y
.
A
n
ex
p
er
i
m
en
tal
f
i
n
d
in
g
s
h
o
w
t
h
at
th
eir
m
et
h
o
d
is
p
r
ett
y
g
o
o
d
in
ter
m
s
o
f
ef
f
icie
n
c
y
an
d
e
f
f
ec
ti
v
en
e
s
s
in
t
h
e
i
m
p
r
o
v
e
m
e
n
t
o
f
co
n
ten
t
-
b
ased
v
id
eo
r
etr
iev
al
[
1
7
]
.
A
m
et
h
o
d
tr
an
s
f
er
s
t
h
r
o
u
g
h
d
atab
ase
v
id
eo
to
s
ce
n
es
u
s
i
n
g
s
ce
n
e
ch
a
n
g
e
d
etec
tio
n
alg
o
r
ith
m
b
ased
o
n
co
lo
r
h
is
to
g
r
a
m
a
n
d
ex
tr
ac
t
k
e
y
f
r
a
m
e
s
is
s
u
g
g
es
ted
.
Usi
n
g
s
tr
aig
h
t
f
o
r
w
ar
d
r
u
les
m
u
l
tip
le
f
ea
t
u
r
es
ar
e
o
b
tain
ed
f
o
r
k
e
y
f
r
a
m
es.
T
h
en
u
s
e
th
e
m
u
ltip
le
f
ea
t
u
r
e
v
ec
to
r
s
to
co
m
p
ar
e
q
u
er
y
an
d
d
atab
ase
v
id
eo
s
b
y
m
ea
s
u
r
i
n
g
E
u
clid
ea
n
d
i
s
tan
ce
[
1
8
]
.
A
ls
o
,
a
s
y
s
te
m
m
o
d
el
f
o
r
lar
g
e
v
id
eo
d
ata
p
r
o
ce
s
s
in
g
i
n
C
B
VR
s
y
s
te
m
s
is
p
r
esen
ted
,
Usa
g
e
o
f
a
d
is
tr
ib
u
ted
C
o
m
p
u
ter
in
f
r
astr
u
ctu
r
e,
w
it
h
a
Ma
p
R
ed
u
ce
-
b
a
s
ed
Had
o
o
p
s
y
s
te
m
.
T
h
e
y
h
a
v
e
also
o
p
ted
f
o
r
E
l
Ou
ad
r
h
ir
i
et
al.
C
B
V
R
f
r
a
m
e
w
o
r
k
as
a
ca
s
e
in
p
o
in
t
f
o
r
th
eir
ap
p
r
o
ac
h
.
T
h
e
ch
al
len
g
e
i
n
th
e
c
u
r
r
en
t
s
c
en
ar
io
is
t
h
er
ef
o
r
e
th
e
o
p
ti
m
iz
atio
n
o
f
t
h
e
B
C
MH
co
m
p
u
tatio
n
ti
m
e
[
1
9
]
.
T
h
en
an
a
u
to
m
a
ted
v
id
eo
co
n
te
n
t
an
al
y
s
is
a
n
d
r
etr
iev
al
s
y
s
te
m
to
en
ab
le
t
h
e
d
is
co
v
er
y
o
f
GDR
T
V
v
id
eo
s
in
h
is
to
r
ical
co
llec
tio
n
s
i
s
p
r
o
v
id
ed
.
I
t
is
b
ased
o
n
s
er
v
ice
-
o
r
ien
ted
d
is
tr
ib
u
te
d
ar
ch
itect
u
r
e
an
d
i
n
cl
u
d
es
al
g
o
r
ith
m
s
f
o
r
s
h
o
t
b
o
u
n
d
ar
y
d
ete
ctio
n
v
id
eo
an
al
y
s
i
s
,
id
ea
clas
s
if
ica
tio
n
,
p
er
s
o
n
a
l
r
ec
o
g
n
itio
n
,
tex
t r
ec
o
g
n
it
io
n
an
d
s
i
m
ilar
it
y
s
ea
r
ch
[
2
0
]
.
Me
an
w
h
ile,
a
v
id
eo
r
etr
iev
al
s
y
s
te
m
f
o
r
v
is
u
al
co
n
te
n
t
S
in
c
e
a
u
s
er
s
en
d
s
a
v
id
eo
s
ea
r
ch
q
u
er
y
in
a
n
atu
r
al
la
n
g
u
a
g
e
is
d
ev
elo
p
ed
.
A
g
e
n
er
al
p
ip
eli
n
e
o
f
a
co
n
t
en
t
-
b
ased
v
id
eo
r
ec
o
v
er
y
s
y
s
t
e
m
co
m
p
r
i
s
es
t
w
o
m
ai
n
s
ec
t
i
o
n
s
:
a)
T
h
e
o
f
f
li
n
e
tr
ain
i
n
g
p
r
o
ce
s
s
in
w
h
ich
v
ec
t
o
r
s
ar
e
ex
tr
ac
ted
f
r
o
m
a
b
r
o
ad
v
id
eo
d
atab
ase
to
tr
ain
th
e
s
y
s
te
m
o
f
v
id
eo
f
ea
t
u
r
es
an
d
th
e
co
r
r
esp
o
n
d
in
g
v
id
eo
ca
p
tio
n
s
th
at
d
escr
ib
e
th
e
n
atu
r
al
lan
g
u
ag
e
v
id
eo
co
n
tex
t,
an
d
b
)
On
li
n
e
c
o
n
ten
t
p
r
o
ce
s
s
i
n
g
s
tep
s
in
w
h
i
ch
q
u
er
y
f
ea
t
u
r
es
ar
e
p
ic
k
ed
u
p
an
d
th
e
n
u
s
ed
to
r
etr
iev
e
v
id
eo
s
r
elate
d
to
th
e
d
atab
ase
[
2
1
]
.
A
s
p
ec
i
f
ic
v
id
eo
s
eg
m
e
n
tatio
n
u
s
in
g
d
etec
tio
n
o
f
ch
a
n
g
e
o
f
s
ce
n
e
ac
co
m
p
a
n
ied
b
y
r
an
k
i
n
g
is
p
r
o
p
o
s
ed
.
T
h
e
ap
p
r
o
ac
h
s
u
g
g
e
s
t
ed
w
i
ll
co
n
s
is
t
o
f
th
e
f
o
llo
w
in
g
p
h
a
s
es:
p
r
i
m
ar
il
y
Sep
ar
ate
s
ce
n
es
ar
e
id
e
n
ti
f
ied
f
r
o
m
v
id
eo
s
to
d
ef
i
n
e
s
ce
n
e
b
o
u
n
d
ar
ies.
Dete
cted
b
o
u
n
d
a
r
ies
h
elp
to
d
iv
id
e
v
id
eo
s
i
n
to
c
h
u
n
k
s
o
f
v
ar
io
u
s
s
ce
n
es
an
d
th
e
n
cla
s
s
i
f
y
t
h
e
m
.
T
h
e
v
id
eo
clas
s
if
ier
clas
s
if
ie
s
t
h
e
t
y
p
e
o
f
a
p
ar
ticu
lar
ch
u
n
k
o
f
ea
ch
f
r
a
m
e.
T
h
e
v
id
eo
clas
s
if
ier
o
u
t
p
u
t
f
o
r
a
g
iv
e
n
c
h
u
n
k
p
r
ed
i
cts
cla
s
s
o
f
h
i
g
h
e
s
t
p
r
o
b
a
b
ilit
y
.
T
h
en
,
ev
e
n
t
u
all
y
,
u
s
er
q
u
er
y
is
m
atc
h
ed
w
ith
ea
ch
ch
u
n
k
lab
el
to
o
b
tain
t
h
e
a
p
p
r
o
p
r
iate
p
iece
o
f
n
e
w
s
[
2
2
]
.
B
y
i
n
teg
r
at
in
g
m
u
lti
-
m
o
d
al
f
ea
tu
r
es
f
o
r
b
i
n
ar
y
h
as
h
lear
n
in
g
i
n
b
o
th
o
f
f
li
n
e
tr
ain
i
n
g
a
n
d
o
n
l
in
e
q
u
er
y
p
h
ases
,
m
u
l
ti
-
m
o
d
al
h
a
s
h
i
n
g
m
et
h
o
d
s
m
a
y
allo
w
s
u
c
ce
s
s
f
u
l
m
u
lti
m
ed
ia
r
etr
iev
a
l.
E
v
en
s
o
,
w
h
e
n
o
n
l
y
o
n
e
o
r
s
o
m
e
o
f
t
h
e
m
o
d
alit
ies
ar
e
g
iv
e
n
,
cu
r
r
en
t
m
u
lti
-
m
o
d
al
ap
p
r
o
ac
h
es
co
u
l
d
n
’
t
b
in
ar
ize
q
u
er
ies.
T
h
er
ef
o
r
e,
to
r
eso
lv
e
th
i
s
i
s
s
u
e
,
a
n
e
w
f
le
x
ib
le
m
u
lti
-
m
o
d
al
h
as
h
i
n
g
(
FMH)
m
et
h
o
d
is
s
u
g
g
e
s
ted
.
W
ith
i
n
a
s
in
g
le
m
o
d
el,
FMH
lear
n
s
v
ar
io
u
s
m
o
d
alit
y
-
s
p
ec
i
f
ic
h
a
s
h
co
d
es
an
d
m
u
lti
-
m
o
d
al
co
o
p
er
ativ
e
h
as
h
co
d
es
co
n
cu
r
r
en
tl
y
.
B
ased
o
n
th
e
n
e
w
l
y
ar
r
iv
i
n
g
q
u
er
ies,
th
a
t
o
f
f
er
i
n
g
a
n
y
o
n
e
o
r
co
m
b
i
n
atio
n
o
f
m
o
d
alit
y
ch
ar
ac
ter
is
tic
s
,
t
h
e
h
as
h
co
d
es
ar
e
f
lex
ib
l
y
cr
ea
ted
.
I
n
ad
d
itio
n
,
t
h
e
h
as
h
i
n
g
lear
n
i
n
g
m
eth
o
d
is
e
f
f
ec
ti
v
el
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
C
o
n
ten
t b
a
s
ed
vid
e
o
r
etri
ev
a
l u
s
in
g
d
is
crete
co
s
in
e
tr
a
n
s
fo
r
m
(
S
u
ma
ya
Ha
ma
d
)
841
m
o
n
ito
r
ed
b
y
t
h
e
p
air
-
w
is
e
s
e
m
an
tic
m
atr
i
x
to
b
o
o
s
t
t
h
e
ab
i
lit
y
to
d
i
f
f
e
r
e
n
tiate
[
2
3
]
.
A
s
er
ies
o
f
e
x
p
er
i
m
e
n
ts
m
ea
s
u
r
in
g
h
o
w
w
ell
t
h
e
b
en
ch
m
ar
k
o
u
tco
m
e
s
r
ep
r
esen
t
th
e
r
ea
l
p
r
o
g
r
ess
in
s
o
lv
i
n
g
t
h
e
task
o
f
m
o
m
e
n
t
r
etr
iev
al
is
p
r
ese
n
ted
.
T
h
e
f
in
d
in
g
s
s
h
o
w
m
aj
o
r
b
iases
in
t
h
e
co
m
m
o
n
d
atasets
a
n
d
u
n
ex
p
ec
ted
ac
tio
n
s
o
f
t
h
e
s
ta
te
-
of
-
th
e
-
ar
t
m
o
d
el
s
.
I
n
ad
d
itio
n
,
th
e
au
th
o
r
s
p
r
ese
n
t
n
e
w
s
tu
d
ie
s
an
d
s
tr
ate
g
ies
f
o
r
v
i
s
u
aliz
in
g
t
h
e
ef
f
ec
t
s
o
f
h
ea
lth
ch
ec
k
s
.
A
t
la
s
t,
t
h
e
y
r
ec
o
m
m
e
n
d
p
o
ten
t
ial
w
a
y
s
o
f
i
m
p
r
o
v
i
n
g
t
h
e
m
o
m
e
n
t
o
f
r
et
r
iev
al
i
n
t
h
e
f
u
t
u
r
e.
[
2
4
]
.
M
o
r
e
d
etailed
r
esear
ch
r
ev
ie
w
ca
n
b
e
f
o
u
n
d
in
[
2
5
]
.
T
h
e
m
ai
n
tar
g
et
o
f
t
h
is
w
o
r
k
is
to
b
u
ild
a
C
B
VR
s
y
s
te
m
th
at
s
u
p
p
o
r
ts
q
u
er
y
i
n
g
b
y
e
x
a
m
p
le
to
r
etr
iev
e
s
i
m
ilar
v
id
eo
f
r
o
m
a
d
atab
ase
ac
co
r
d
in
g
to
th
eir
f
e
atu
r
es.
C
o
lo
r
f
ea
t
u
r
es
h
av
e
b
e
en
u
s
ed
to
id
en
tify
an
d
d
escr
ib
e
th
e
co
n
te
n
ts
o
f
t
h
e
v
id
eo
.
T
h
e
r
est
o
f
th
e
p
ap
er
is
s
tr
u
ct
u
r
ed
ac
co
r
d
in
g
to
th
e
f
o
llo
w
i
n
g
:
Sect
io
n
2
p
r
esen
ts
r
elate
d
r
esear
ch
a
n
d
d
atasets
.
Sectio
n
3
i
n
tr
o
d
u
c
es
th
e
C
B
VR
s
tr
u
ct
u
r
e.
Sectio
n
4
d
escr
ib
es
DC
T
f
ea
t
u
r
e
ex
tr
ac
tio
n
,
in
c
lu
d
i
n
g
f
r
a
m
e
ex
tr
ac
tio
n
a
n
d
r
esizin
g
,
d
is
cr
ete
co
s
in
e
tr
an
s
f
o
r
m
a
tio
n
(
DC
T
)
.
Sectio
n
5
r
ep
o
r
ts
o
n
th
e
test
m
ater
ial.
S
ec
tio
n
6
p
r
esen
ts
r
es
u
lt
s
an
d
a
n
al
y
s
i
s
o
f
t
h
e
ex
p
er
i
m
en
tal
s
t
u
d
y
o
n
t
h
e
d
ataset.
Sectio
n
7
co
n
clu
d
e
s
th
e
p
ap
er
.
2.
RE
S
E
ARCH
M
E
T
H
O
D
2
.
1
.
C
B
VR
s
t
ruct
ure
T
h
e
p
r
o
p
o
s
ed
f
r
a
m
e
w
o
r
k
co
m
p
r
is
es
f
o
u
r
m
ai
n
m
o
d
u
les,
as
s
h
o
w
n
in
t
h
e
Fi
g
u
r
e
1
.
E
v
er
y
m
o
d
u
le
h
a
s
its
o
w
n
s
p
ec
i
f
ic
f
u
n
ctio
n
s
.
T
h
e
f
o
u
r
m
o
d
u
le
s
ar
e:
a)
Featu
r
e
E
x
tr
ac
t
io
n
m
o
d
u
le:
T
h
e
co
lo
r
f
ea
tu
r
es a
r
e
ex
tr
ac
ted
f
r
o
m
ea
ch
v
id
eo
an
d
lis
ted
i
n
th
e
d
atab
ase.
b)
Qu
er
y
i
n
g
m
o
d
u
le:
T
h
r
o
u
g
h
th
e
GUI
o
f
th
is
m
o
d
u
le
r
etr
iev
e
th
e
m
o
s
t
s
i
m
ilar
v
id
eo
lis
ted
in
th
e
d
atab
ase,
th
e
u
s
er
m
a
y
i
s
s
u
e
h
is
/
h
er
q
u
er
y
.
c)
Si
m
i
lar
it
y
m
atc
h
i
n
g
m
o
d
u
le
:
W
ith
in
th
i
s
m
o
d
u
le,
u
s
i
n
g
E
u
clid
ian
d
i
s
tan
ce
a
n
d
m
ea
n
s
q
u
ar
e
er
r
o
r
s
i
m
ilar
it
y
m
e
tr
ics,
s
i
m
ilar
i
t
y
b
et
w
ee
n
q
u
er
y
v
id
eo
an
d
d
atab
ase
v
id
eo
is
a
s
s
e
s
s
ed
.
T
h
e
b
est
s
i
m
ilar
v
id
eo
w
il
l b
e
s
h
o
w
n
ac
co
r
d
in
g
to
th
eir
s
i
m
ilar
i
t
y
d
eg
r
ee
.
d)
R
etr
iev
al
m
o
d
u
le:
T
h
e
d
atab
a
s
e
v
id
eo
w
ith
h
ig
h
er
s
i
m
ilar
it
y
r
an
k
w
ill
b
e
r
etr
ie
v
ed
an
d
d
is
p
la
y
ed
to
t
h
e
u
s
er
.
T
h
e
ab
o
v
e
m
e
n
tio
n
m
o
d
u
le
s
w
il
l b
e
d
is
cu
s
s
ed
in
d
etails i
n
t
h
e
n
e
x
t sect
io
n
s
.
Fig
u
r
e
1
.
C
B
VR
s
y
s
te
m
m
o
d
u
les
2
.
2
.
DCT f
ea
t
ure
ex
t
ra
ct
i
o
n
(
DC
T
)
m
ea
n
s
d
is
cr
ete
co
s
i
n
e
tr
an
s
f
o
r
m
.
I
t
h
as
f
o
u
n
d
ap
p
licatio
n
s
f
o
r
d
ig
ital
s
i
g
n
al
an
d
i
m
a
g
e
p
r
o
ce
s
s
in
g
,
an
d
p
a
r
ticu
lar
l
y
f
o
r
d
ata
c
o
m
p
r
ess
io
n
/
d
ec
o
m
p
r
es
s
io
n
tr
an
s
f
o
r
m
a
tio
n
co
d
in
g
s
y
s
te
m
s
.
I
n
th
i
s
w
o
r
k
DC
T
u
s
ed
as
s
h
o
w
n
i
n
th
is
f
lo
w
c
h
ar
t
Fi
g
u
r
e
2
.
T
h
e
DC
T
p
r
o
ce
s
s
in
v
o
l
v
es
r
ea
d
in
g
t
h
e
v
id
eo
an
d
th
e
n
ex
tr
ac
t
1
0
f
r
a
m
es
f
r
o
m
v
id
eo
an
d
r
e
s
ize
t
h
e
m
to
2
5
6
×2
5
6
p
ix
els.
A
f
ter
th
at
t
h
e
D
C
T
eq
u
atio
n
i
s
ap
p
lied
to
t
h
e
r
esized
f
r
a
m
e.
(
1
9
2
)
DC
co
ef
f
icien
t
s
f
o
r
ea
ch
i
m
a
g
e
ar
e
o
b
ta
in
ed
an
d
th
e
n
s
to
r
ed
in
t
h
e
s
y
s
te
m
d
atab
ase.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
21
,
No
.
2
,
Feb
r
u
ar
y
2021
:
8
3
9
-
8
4
5
842
Fig
u
r
e
2
.
DC
T
f
ea
tu
r
e
ex
tr
ac
ti
o
n
p
r
o
ce
s
s
2
.
2
.
1
.
F
ra
m
e
e
x
t
ra
ct
io
n a
nd
re
s
izing
E
x
tr
ac
tin
g
1
0
f
r
a
m
e
s
f
r
o
m
v
i
d
eo
,
ac
co
r
d
in
g
to
th
e
n
u
m
b
er
o
f
f
r
a
m
e
s
th
e
r
eq
u
ir
ed
f
r
a
m
es
w
ill
b
e
s
elec
ted
,
r
esizi
n
g
t
h
e
i
n
p
u
t
f
r
a
m
e
to
f
ix
ed
s
ize,
s
o
all
f
r
a
m
e
w
i
ll
b
e
in
n
e
w
f
i
x
ed
w
id
th
a
n
d
h
i
g
h
.
O
f
all
fr
a
m
e
s
,
th
e
n
e
w
s
ize
i
s
2
5
6
×2
5
6
.
I
t is si
g
n
i
f
ica
n
t h
o
w
t
h
i
s
p
r
o
ce
s
s
g
et
s
th
e
b
est r
es
u
lt.
2
.
2
.
2
.
Dis
cr
et
e
co
s
ine t
ra
ns
f
o
r
m
a
t
io
n
(
DC
T
)
A
p
p
l
y
in
g
D
C
T
to
th
e
th
r
ee
f
r
a
m
e
m
atr
ices
R
,
G
a
n
d
B
to
co
n
v
er
t
f
r
a
m
e
d
ata
f
r
o
m
s
p
atial
d
o
m
ai
n
to
f
r
eq
u
en
c
y
d
o
m
a
in
b
y
s
p
litt
in
g
ea
ch
f
r
a
m
e
i
n
to
3
2
×3
2
p
ix
el
b
lo
ck
s
a
n
d
ap
p
ly
DC
T
tr
an
s
f
o
r
m
to
ea
c
h
b
lo
ck
v
alu
e
to
p
r
o
d
u
ce
1
0
2
4
co
ef
f
icien
t
s
.
T
h
e
o
u
tp
u
t
m
atr
i
x
co
n
s
is
ts
o
f
D
C
T
co
ef
f
icien
t
s
,
th
e
v
al
u
ab
le
an
d
i
m
p
o
r
tan
t
d
ata
r
ep
r
esen
ti
n
g
th
e
f
r
a
m
e
p
lace
d
i
n
th
e
to
p
le
f
t
h
an
d
co
r
n
er
o
f
t
h
e
m
atr
i
x
w
h
i
le
th
e
les
s
v
alu
ab
l
e
d
etails
ab
o
u
t
t
h
e
co
ef
f
ic
ien
t
s
lies
o
n
th
e
lo
w
er
r
i
g
h
t
-
h
a
n
d
.
I
n
th
e
to
p
lef
t
h
a
n
d
co
r
n
er
o
f
th
e
m
atr
i
x
is
DC
T
co
ef
f
icie
n
t
at
p
o
s
itio
n
(
0
,
0
)
w
h
ich
i
n
d
icate
s
t
h
e
a
v
er
ag
e
o
f
t
h
e
o
th
er
1
0
2
4
v
al
u
e
i
n
th
e
m
at
r
ix
.
T
h
e
p
o
in
t
(
0
,
0
)
is
ca
lled
DC
co
ef
f
icien
t
an
d
th
e
o
th
er
p
o
in
t
ca
lled
AC
co
ef
f
icie
n
ts
.
So
w
e
u
s
e
t
h
e
DC
co
ef
f
icien
t
f
r
o
m
ea
c
h
b
lo
ck
f
o
r
ev
er
y
f
r
a
m
e
i
n
d
atab
ase
an
d
u
s
e
it a
s
D
C
T
co
lo
r
f
ea
tu
r
e.
2
.
3
.
T
est
m
a
t
er
ia
l
T
h
e
d
atab
ase
o
f
v
id
eo
s
u
s
ed
in
th
i
s
p
ap
er
b
elo
n
g
s
to
o
u
r
v
id
eo
s
th
at
co
llected
f
r
o
m
r
ea
l
wo
r
ld
.
v
id
eo
d
atab
ase
co
n
s
i
s
ts
o
f
8
cla
s
s
e
s
ea
ch
co
n
ta
in
s
ab
o
u
t
1
0
v
id
eo
s
.
So
,
th
e
v
id
eo
te
s
ted
to
tal
n
u
m
b
er
is
1
0
0
.
T
h
ese
Vid
eo
ar
e
o
f
d
if
f
er
en
t
s
izes
w
it
h
co
lo
r
r
eso
lu
tio
n
2
4
b
it/p
ix
el.
E
ig
h
t
v
id
eo
ca
teg
o
r
ies
o
f
d
atab
a
s
e
ar
e
u
s
ed
(
G
a
m
e
,
P
er
s
o
n
,
B
uil
din
g
,
Child
re
n
,
Ca
t
s
,
F
lo
w
er
s
,
B
ea
ches,
L
a
nd
s
ca
pe
)
.
Fi
g
u
r
e
3
s
h
o
w
s
Sa
m
p
le
o
f
v
id
eo
f
o
r
co
llected
Data
b
ase.
Fig
u
r
e
3
.
Sa
m
p
le
o
f
v
id
eo
f
o
r
co
llected
d
atab
ase
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
C
o
n
ten
t b
a
s
ed
vid
e
o
r
etri
ev
a
l u
s
in
g
d
is
crete
co
s
in
e
tr
a
n
s
fo
r
m
(
S
u
ma
ya
Ha
ma
d
)
843
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
p
er
f
o
r
m
an
ce
o
f
a
r
etr
iev
al
s
y
s
te
m
ca
n
b
e
ev
al
u
ated
w
it
h
r
esp
ec
t
to
its
p
r
ec
is
io
n
an
d
r
ec
all.
P
r
ec
is
io
n
m
ea
s
u
r
es
ca
n
b
e
d
escr
ib
ed
as
th
e
s
y
s
te
m
's
ab
ilit
y
to
r
etr
iev
e
o
n
l
y
r
ele
v
an
t
m
o
d
els;
W
h
er
ea
s
r
ec
all
m
ea
s
u
r
es
ar
e
r
ep
r
esen
ted
th
e
s
y
s
te
m
's
ab
ilit
y
to
r
etr
iev
e
all
r
elev
an
t
m
o
d
el
s
d
ef
i
n
ed
th
e
n
u
m
b
er
o
f
r
elev
a
n
t
v
id
eo
s
r
etr
iev
ed
.
I
n
th
is
p
ap
er
,
w
e
e
v
alu
a
te
5
v
id
eo
s
in
ea
ch
d
atab
ase
class
an
d
f
o
r
ea
ch
ca
teg
o
r
y
3
v
id
eo
s
r
etr
iev
ed
w
il
l b
e
s
i
m
ilar
as s
h
o
w
n
in
t
h
e
T
ab
le
1
.
P
r
ec
is
io
n
=
=
R
ec
all=
=
W
h
e
r
e
A
b
e
i
n
g
t
h
e
n
u
m
b
e
r
o
f
r
e
l
e
v
a
n
t
v
i
d
e
o
s
r
e
t
r
i
e
v
e
d
,
B
b
e
i
n
g
t
h
e
n
u
m
b
e
r
o
f
i
r
r
e
l
e
v
a
n
t
i
t
e
m
s
a
n
d
C
b
e
i
n
g
t
h
e
n
u
m
b
e
r
o
f
r
e
l
e
v
a
n
t
i
t
e
m
s
n
o
t
r
e
t
r
i
e
v
e
d
i
n
t
h
e
d
a
t
a
b
a
s
e
.
F
i
g
u
r
e
4
s
h
o
w
s
S
a
m
p
l
e
o
f
v
i
d
e
o
r
e
t
r
i
e
v
a
l
r
e
s
u
l
t
s
.
T
ab
le
1
.
T
h
e
p
r
ec
is
io
n
r
esu
lt
C
l
a
ss
G
a
me
P
e
r
so
n
B
u
i
l
d
i
n
g
C
h
i
l
d
r
e
n
Cats
F
l
o
w
e
r
s
B
e
a
c
h
e
s
L
a
n
d
sca
p
e
A
v
e
r
a
g
e
A
v
e
r
a
g
e
P
r
e
c
i
si
o
n
0
.
4
0
.
8
6
0
.
6
0
.
5
3
0
.
6
6
0
.
8
0
.
6
0
.
7
3
0
.
6
4
7
5
Fig
u
r
e
4
.
Sa
m
p
le
o
f
v
id
eo
r
etr
iev
al
r
es
u
lts
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
21
,
No
.
2
,
Feb
r
u
ar
y
2021
:
8
3
9
-
8
4
5
844
4.
CO
NCLU
SI
O
N
T
h
e
co
n
ten
t
-
b
ased
m
o
d
el
allo
w
s
th
e
s
y
s
te
m
to
ac
q
u
ir
e
a
v
i
d
eo
f
r
o
m
a
lis
t
o
f
v
i
s
u
al
co
n
t
en
t
-
b
ased
v
id
eo
s
t
h
at
ar
e
en
tire
l
y
a
u
to
m
atica
ll
y
e
x
tr
ac
ted
,
s
u
ch
as
co
lo
u
r
,
tex
t
u
r
e,
s
h
ap
es,
b
u
t
n
o
t
attr
ib
u
tes
t
h
at
ar
e
u
n
r
elate
d
to
th
e
co
n
ten
t.
T
h
e
v
is
u
al
co
n
te
n
t
o
f
th
i
s
v
id
eo
ca
r
r
ies
litt
le
to
n
o
s
em
a
n
tic
co
n
ten
t.
I
n
th
is
p
ap
er
,
th
e
co
lo
r
co
n
ten
t
o
f
v
id
eo
is
u
s
ed
to
r
etr
iev
e
v
id
eo
s
.
T
h
e
ex
p
er
im
e
n
tal
r
e
s
u
l
ts
s
h
o
w
t
h
at
t
h
e
av
er
a
g
e
r
ati
n
g
o
f
0
.
6
4
7
5
w
a
s
ca
lcu
la
ted
u
s
i
n
g
t
h
e
DC
T
af
ter
ap
p
l
y
in
g
it
to
o
u
r
d
atab
ase
f
o
r
all
c
ateg
o
r
ies,
w
h
en
i
t’
s
u
s
ed
in
co
n
ten
t
-
b
ased
v
id
eo
r
etr
iev
al
(
C
B
VR
)
.
As
t
h
e
i
m
p
le
m
e
n
ted
alg
o
r
ith
m
s
r
u
n
o
n
th
e
v
id
eo
'
s
co
lo
r
f
ea
tu
r
es
an
d
v
alu
e
s
,
th
e
d
i
v
is
io
n
o
f
v
id
eo
in
to
b
lo
ck
s
o
r
p
iece
s
th
er
ef
o
r
e
g
av
e
m
o
r
e
ab
ilit
y
to
r
etr
iev
e
t
h
e
r
elev
a
n
t v
id
eo
s
.
RE
F
E
R
E
NC
E
S
[1
]
S
.
T
.
M
.
T
o
riah
,
A
.
Z.
G
h
a
lw
a
s
h
,
a
n
d
A
.
A
.
A
.
Yo
u
ss
if
,
“
S
e
m
a
n
ti
c
-
Ba
se
d
V
id
e
o
Re
tri
e
v
a
l
S
u
rv
e
y
,
”
J
.
Co
mp
u
t
.
Co
mm
u
n
.
,
v
o
l.
6
,
n
o
.
8
,
p
p
.
2
8
-
4
4
,
2
0
1
8
.
[2
]
S
.
S
.
G
o
rn
a
le,
A
.
K.
Ba
b
a
les
h
w
a
r,
a
n
d
P
.
L
.
Ya
n
n
a
w
a
r,
“
A
n
a
l
y
sis
a
n
d
De
tec
ti
o
n
o
f
Co
n
ten
t
b
a
se
d
V
i
d
e
o
Re
tri
e
v
a
l,
”
In
t.
J
.
Ima
g
e
,
Gr
a
p
h
.
S
ig
n
a
l
Pro
c
e
ss
.
,
v
o
l.
1
1
,
n
o
.
3
,
p
.
4
3
,
2
0
1
9
.
[3
]
R.
A
sh
ra
f
,
M
.
A
h
m
e
d
,
U.
A
h
m
a
d
,
M
.
A
.
Ha
b
ib
,
S
.
Ja
b
b
a
r,
a
n
d
K.
Na
se
e
r,
“
M
DCBIR
-
M
F
:
m
u
lt
im
e
d
ia
d
a
ta
f
o
r
c
o
n
ten
t
-
b
a
se
d
im
a
g
e
re
t
riev
a
l
b
y
u
sin
g
m
u
lt
ip
le
f
e
a
tu
re
s,”
M
u
lt
ime
d
.
T
o
o
ls
A
p
p
l
.
,
v
o
l.
7
9
,
n
o
.
1
3
,
p
p
.
8
5
5
3
-
8
5
7
9
,
2
0
2
0
.
[4
]
L
.
T
o
n
g
,
R.
T
o
n
g
,
a
n
d
L
.
Ch
e
n
,
“
Eff
icie
n
t
re
tri
e
v
a
l
a
lg
o
rit
h
m
f
o
r
m
u
lt
i
m
e
d
ia
i
m
a
g
e
in
f
o
r
m
a
ti
o
n
,
”
M
u
lt
ime
d
.
T
o
o
ls
Ap
p
l
.
,
v
o
l.
7
9
,
n
o
.
1
3
,
p
p
.
9
4
6
9
-
9
4
8
7
,
2
0
2
0
.
[5
]
R.
G
a
ss
e
r,
L
.
Ro
ss
e
tt
o
,
S
.
He
ll
e
r,
a
n
d
H.
S
c
h
u
ld
t,
“
Co
t
to
n
tail
DB:
A
n
Op
e
n
S
o
u
rc
e
Da
tab
a
se
S
y
ste
m
f
o
r
M
u
lt
im
e
d
ia
Re
tri
e
v
a
l
a
n
d
A
n
a
ly
s
is,
”
in
Pro
c
e
e
d
in
g
s
o
f
t
h
e
2
8
t
h
A
CM
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
M
u
lt
ime
d
ia
,
p
p
.
4
4
6
5
-
4
4
6
8
,
2
0
2
0
.
[
6
]
M
.
E
.
A
b
d
u
l
m
u
n
e
m
a
n
d
E
.
H
a
t
o
,
“
S
e
m
a
n
t
i
c
B
a
s
e
d
V
i
d
e
o
R
e
t
r
i
e
v
a
l
S
y
s
t
e
m
:
S
u
r
v
e
y
,
”
I
r
a
q
i
J
.
S
c
i
.
,
p
p
.
7
3
9
-
7
5
3
,
2
0
1
8
.
[7
]
S
.
P
u
th
e
ti
,
M
.
N.
S
.
Ha
rsh
a
,
a
n
d
A
.
V
ish
n
u
v
a
rd
h
a
n
,
“
M
o
ti
o
n
De
t
e
c
ti
o
n
i
n
V
i
d
e
o
Re
tri
e
v
a
l
Us
in
g
Co
n
ten
t
-
Ba
se
d
V
id
e
o
Re
tri
e
v
a
l,
”
in
I
n
n
o
v
a
ti
o
n
s i
n
Co
m
p
u
ter
S
c
ie
n
c
e
a
n
d
E
n
g
i
n
e
e
rin
g
,
S
p
ri
n
g
e
r
,
p
p
.
2
3
5
-
2
4
2
,
2
0
1
9
.
[8
]
A
.
F
.
S
.
De
v
a
ra
j
e
t
a
l.
,
“
A
n
Eff
ic
ien
t
F
ra
m
e
w
o
rk
f
o
r
S
e
c
u
re
Im
a
g
e
A
r
c
h
iv
a
l
a
n
d
Re
tri
e
v
a
l
S
y
ste
m
Us
in
g
M
u
lt
i
p
le
S
e
c
re
t
S
h
a
re
Cre
a
ti
o
n
S
c
h
e
m
e
,
”
I
EE
E
Acc
e
ss
,
v
o
l.
8
,
p
p
.
1
4
4
3
1
0
-
1
4
4
3
2
0
,
2
0
2
0
.
[9
]
P
.
S
a
x
e
n
a
,
S
.
K
.
S
i
n
g
h
,
a
n
d
M
.
S
riv
a
sta
v
a
,
“
Co
n
ten
t
-
Ba
se
d
Re
tri
e
v
a
l
o
f
M
u
lt
im
e
d
ia
In
f
o
r
m
a
ti
o
n
Us
in
g
M
u
lt
i
p
le
S
im
il
a
rit
y
In
d
e
x
e
s,” in
S
o
ft
Co
mp
u
ti
n
g
:
T
h
e
o
rie
s a
n
d
Ap
p
li
c
a
t
io
n
s
,
S
p
ri
n
g
e
r
,
p
p
.
1
2
3
5
-
1
2
4
2
,
2
0
2
0
.
[1
0
]
S
.
He
ll
e
r,
L
.
S
a
u
ter,
H.
S
c
h
u
ld
t,
a
n
d
L
.
Ro
ss
e
tt
o
,
“
M
u
lt
i
-
S
tag
e
Qu
e
ries
a
n
d
T
e
m
p
o
ra
l
S
c
o
rin
g
in
V
it
riv
r,
”
in
2
0
2
0
IEE
E
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
M
u
lt
ime
d
i
a
&
Exp
o
W
o
rk
sh
o
p
s
(
ICM
E
W
)
,
p
p
.
1
-
5
,
2
0
2
0
.
[1
1
]
H.
W
a
n
g
,
Z.
L
i,
Y.
L
i,
B.
B.
G
u
p
ta,
a
n
d
C.
Ch
o
i,
“
V
isu
a
l
sa
l
ien
c
y
g
u
id
e
d
c
o
m
p
lex
i
m
a
g
e
r
e
tri
e
v
a
l,
”
Pa
tt
e
rn
Rec
o
g
n
it
.
L
e
tt
.
,
v
o
l.
1
3
0
,
p
p
.
6
4
-
7
2
,
2
0
2
0
.
[1
2
]
J.
K.
B.
B.
Ra
m
u
,
“
A
S
u
rv
e
y
o
f
Ob
jec
t
V
i
d
e
o
Re
tri
e
v
a
l
S
y
ste
m
s
Us
in
g
S
c
a
le
-
In
v
a
rian
t
F
e
a
tu
re
T
ra
n
sf
o
r
m
(S
if
t)
,
”
In
t.
J
.
Ad
v
.
E
n
g
.
Res
.
De
v
.
,
v
o
l
.
5
,
n
o
.
3
,
p
p
.
1
2
1
1
-
1
2
1
6
,
2
0
1
8
.
[1
3
]
N.
G
a
y
a
th
ri
a
n
d
K.
M
a
h
e
sh
,
“
A
S
y
ste
m
a
ti
c
S
tu
d
y
o
n
V
id
e
o
In
d
e
x
in
g
,
”
In
t.
J
.
P
u
re
Ap
p
l.
M
a
t
h
.
,
v
o
l.
1
1
8
,
n
o
.
8
,
p
p
.
425
-
4
2
8
,
2
0
1
8
.
[1
4
]
A
.
A
n
s
a
ri
a
n
d
M
.
H.
M
o
h
a
m
m
e
d
,
“
Co
n
ten
t
b
a
se
d
v
id
e
o
re
tri
e
v
a
l
s
y
ste
m
s
-
m
e
th
o
d
s,
tec
h
n
iq
u
e
s,
tren
d
s
a
n
d
c
h
a
ll
e
n
g
e
s,”
In
t.
J
.
C
o
mp
u
t.
A
p
p
l
.
,
v
o
l.
1
1
2
,
n
o
.
7
,
2
0
1
5
.
[1
5
]
P
.
Ch
iv
a
d
sh
e
tt
i,
K.
S
a
d
a
f
a
le,
a
n
d
K.
T
h
a
k
a
re
,
“
Co
n
ten
t
b
a
se
d
v
id
e
o
re
tri
e
v
a
l
u
sin
g
i
n
teg
ra
ted
f
e
a
tu
re
e
x
trac
ti
o
n
a
n
d
p
e
rso
n
a
li
z
a
ti
o
n
o
f
re
su
lt
s,” in
2
0
1
5
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
In
f
o
rm
a
ti
o
n
Pr
o
c
e
ss
in
g
(
ICIP)
,
p
p
.
1
7
0
-
1
7
5
,
2
0
1
5
.
[
1
6
]
G
.
C
á
m
a
r
a
-
C
h
á
v
e
z
,
F
.
P
r
e
c
i
o
s
o
,
M
.
C
o
r
d
,
S
.
P
h
i
l
l
i
p
-
F
o
l
i
g
u
e
t
,
a
n
d
A
.
d
e
A
.
A
r
a
ú
j
o
,
“
A
n
i
n
t
e
r
a
c
t
i
v
e
v
i
d
e
o
c
o
n
t
e
n
t
-
b
a
s
e
d
r
e
t
r
i
e
v
a
l
s
y
s
t
e
m
,
”
i
n
2
0
0
8
1
5
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
S
y
s
t
e
m
s
,
S
i
g
n
a
l
s
a
n
d
I
m
a
g
e
P
r
o
c
e
s
s
i
n
g
,
p
p
.
1
3
3
-
136
,
2
0
0
8
.
[1
7
]
J.
-
H.
S
u
,
Y.
-
T
.
Hu
a
n
g
,
H.
-
H.
Ye
h
,
a
n
d
V.
S
.
T
se
n
g
,
“
Eff
e
c
ti
v
e
c
o
n
ten
t
-
b
a
se
d
v
id
e
o
re
tri
e
v
a
l
u
sin
g
p
a
tt
e
rn
-
in
d
e
x
in
g
a
n
d
m
a
tch
in
g
tec
h
n
iq
u
e
s,”
Exp
e
rt
S
y
st.
A
p
p
l
.
,
v
o
l.
3
7
,
n
o
.
7
,
p
p
.
5
0
6
8
-
5
0
8
5
,
2
0
1
0
.
[1
8
]
D.
As
h
a
,
M
.
Lata
,
a
n
d
V
.
S
.
K.
Re
d
d
y
,
“
Co
n
ten
t
b
a
se
d
v
id
e
o
re
tri
e
v
a
l
s
y
ste
m
u
sin
g
M
u
lt
ip
le
F
e
a
tu
r
e
s,”
In
t.
J
.
p
u
re
Ap
p
l
.
M
a
t
h
.
,
v
o
l
.
1
1
8
,
p
p
.
2
8
7
-
2
9
4
,
2
0
1
8
.
[1
9
]
E.
M
.
S
a
o
u
d
i
,
A
.
A
.
El
Ou
a
d
rh
i
ri,
O.
El
W
a
rra
k
,
S
.
J.
A
n
d
a
lo
u
ss
i,
a
n
d
A
.
S
e
k
k
a
k
i,
“
I
m
p
ro
v
in
g
Co
n
ten
t
Ba
se
d
V
id
e
o
Re
tri
e
v
a
l
P
e
rf
o
rm
a
n
c
e
b
y
Us
in
g
Ha
d
o
o
p
-
M
a
p
Re
d
u
c
e
M
o
d
e
l,
”
in
2
0
1
8
2
3
rd
C
o
n
fer
e
n
c
e
o
f
Op
e
n
In
n
o
v
a
ti
o
n
s
Asso
c
ia
ti
o
n
(
FR
UCT)
,
p
p
.
1
-
6
,
2
0
1
8
.
[2
0
]
M
.
M
ü
h
l
in
g
e
t
a
l.
,
“
Co
n
ten
t
-
b
a
se
d
v
id
e
o
re
tri
e
v
a
l
in
h
ist
o
rica
l
c
o
ll
e
c
ti
o
n
s
o
f
t
h
e
g
e
rm
a
n
b
ro
a
d
c
a
sti
n
g
a
rc
h
iv
e
,
”
In
t.
J
.
Dig
it
.
L
ib
r
.
,
v
o
l.
2
0
,
n
o
.
2
,
p
p
.
167
-
1
8
3
,
2
0
1
9
.
[2
1
]
R.
Ba
n
sa
l
a
n
d
S
.
C
h
a
k
ra
b
o
rty
,
“
V
isu
a
l
c
o
n
ten
t
b
a
se
d
v
id
e
o
re
tri
e
v
a
l
o
n
n
a
t
u
ra
l
lan
g
u
a
g
e
q
u
e
ries
,
”
in
Pr
o
c
e
e
d
in
g
s
o
f
t
h
e
3
4
th
ACM
/S
IGAPP
S
y
mp
o
s
iu
m o
n
A
p
p
li
e
d
Co
m
p
u
t
in
g
,
p
p
.
2
1
2
-
2
1
9
,
2
0
1
9
.
[2
2
]
M
.
A
.
Ha
s
sa
n
,
S
.
S
a
lee
m
,
M
.
Z.
Kh
a
n
,
a
n
d
M
.
U.
G
.
Kh
a
n
,
“
S
t
o
ry
b
a
se
d
v
id
e
o
re
tri
e
v
a
l
u
sin
g
d
e
e
p
v
i
su
a
l
a
n
d
tex
tu
a
l
in
f
o
r
m
a
ti
o
n
,
”
in
2
0
1
9
2
n
d
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Co
mm
u
n
ica
ti
o
n
,
Co
mp
u
ti
n
g
a
n
d
Di
g
it
a
l
sy
ste
ms
(
C
-
COD
E)
,
p
p
.
1
6
6
-
1
7
1
,
2
0
1
9
.
[2
3
]
L
.
Zh
u
,
X
.
L
u
,
Z.
C
h
e
n
g
,
J.
L
i,
a
n
d
H.
Zh
a
n
g
,
“
F
lex
ib
le
m
u
lt
i
-
m
o
d
a
l
h
a
sh
i
n
g
f
o
r
sc
a
l
a
b
le
m
u
lt
ime
d
ia
re
tri
e
v
a
l,
”
ACM
T
ra
n
s.
I
n
tell.
S
y
st.
T
e
c
h
n
o
l.
,
v
o
l.
1
1
,
n
o
.
2
,
p
p
.
1
-
2
0
,
2
0
2
0
.
[2
4
]
M
.
Ota
n
i,
Y.
Na
k
a
sh
ima
,
E.
R
a
h
tu
,
a
n
d
J.
He
ik
k
il
ä
,
“
Un
c
o
v
e
rin
g
Hid
d
e
n
Ch
a
ll
e
n
g
e
s
in
Qu
e
ry
-
Ba
se
d
V
id
e
o
M
o
m
e
n
t
Re
tri
e
v
a
l,
”
a
rXiv P
re
p
r.
a
rXiv2
0
0
9
.
0
0
3
2
5
,
2
0
2
0
.
[2
5
]
N.
S
p
o
laô
r,
H.
D
.
L
e
e
,
W
.
S
.
R.
T
a
k
a
k
i,
L
.
A
.
En
sin
a
,
C.
S
.
R
.
Co
y
,
a
n
d
F
.
C.
W
u
,
“
A
s
y
ste
m
a
ti
c
re
v
ie
w
o
n
c
o
n
ten
t
-
b
a
se
d
v
id
e
o
re
tri
e
v
a
l,
”
E
n
g
.
Ap
p
l.
Arti
f.
I
n
tell.
,
v
o
l.
9
0
,
p
.
1
0
3
5
5
7
,
2
0
2
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
C
o
n
ten
t b
a
s
ed
vid
e
o
r
etri
ev
a
l u
s
in
g
d
is
crete
co
s
in
e
tr
a
n
s
fo
r
m
(
S
u
ma
ya
Ha
ma
d
)
8
45
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
S
u
m
a
y
a
H
a
m
a
d
re
c
e
iv
e
d
th
e
B.
S
c
.
(g
o
o
d
)
(f
irst
c
las
s)
d
e
g
re
e
in
c
o
m
p
u
ter
sc
ien
c
e
f
ro
m
Un
iv
e
rsit
y
o
f
A
n
b
a
r,
A
n
b
a
r,
Ira
q
,
in
2
0
0
2
,
a
n
d
th
e
M
.
S
c
.
d
e
g
re
e
in
c
o
m
p
u
ter
sc
ien
c
e
f
ro
m
Un
iv
e
rsit
y
o
f
A
n
b
a
r,
A
n
b
a
r,
Ira
q
,
in
2
0
1
2
.
S
h
e
is
c
u
rre
n
tl
y
P
h
D
stu
d
e
n
t
in
Co
m
p
u
ter
S
c
ien
c
e
De
p
a
rtme
n
t
a
t
Un
iv
e
rsity
o
f
Te
c
h
n
o
l
o
g
y
,
Ba
g
h
a
d
a
d
,
Ira
q
,
a
n
d
sh
e
a
n
A
ss
ist.
tea
c
h
e
r.
a
t
t
h
e
Co
ll
e
g
e
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
lo
g
y
,
U
n
iv
e
rsity
o
f
A
n
b
a
r.
S
h
e
h
a
s
p
u
b
li
sh
e
d
5
re
f
e
re
e
d
jo
u
rn
a
l
a
n
d
c
o
n
f
e
re
n
c
e
p
a
p
e
rs.
He
r
c
u
rre
n
t
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
m
o
b
il
e
c
o
m
p
u
ti
n
g
,
a
rti
f
icia
l
in
telli
g
e
n
c
e
,
A
d
Ho
c
n
e
tw
o
rk
s,
se
a
r
c
h
e
n
g
in
e
s,
a
n
d
in
f
o
rm
a
ti
o
n
tec
h
n
o
lo
g
y
.
Ahm
e
e
d
S
u
li
m
a
n
Fa
r
h
a
n
h
a
s
r
e
c
e
iv
e
d
B.
S
c
.
i
n
Co
m
p
u
ter
S
c
ien
c
e
,
A
l
-
A
n
b
a
r
Un
iv
e
rsit
y
,
Ira
q
,
(2
0
0
7
-
2
0
1
1
).
M
.
S
c
d
e
g
re
e
(2
0
1
2
-
2
0
1
4
)
in
Co
m
p
u
ter
S
c
ien
c
e
De
p
a
rt
m
e
n
t,
A
l
-
A
n
a
b
a
r
Un
iv
e
rsit
y
.
F
ield
s
o
f
in
tere
st:
M
u
lt
im
e
d
ia.
He
tau
g
h
t
m
a
n
y
s
u
b
jec
ts
su
c
h
a
s
c
ry
p
to
g
ra
p
h
y
,
o
p
e
ra
ti
o
n
sy
ste
m
,
Ro
b
o
t
S
y
ste
m
a
n
d
n
e
tw
o
rk
s.
Do
a
a
Y
a
se
e
n
K
h
u
d
h
u
r
re
c
e
i
v
e
d
th
e
B.
S
c
.
(g
o
o
d
)
(f
i
rst
c
las
s)
d
e
g
re
e
in
c
o
m
p
u
ter
sc
ien
c
e
f
ro
m
Un
iv
e
rsit
y
o
f
A
n
b
a
r,
A
n
b
a
r,
IRA
Q,
in
2
0
0
4
,
a
n
d
th
e
M
.
S
c
.
d
e
g
re
e
in
c
o
m
p
u
ter
sc
ien
c
e
f
ro
m
Un
iv
e
rsit
y
o
f
A
n
b
a
r,
A
n
b
a
r,
IRA
Q,
in
2
0
1
2
.
S
h
e
is
c
u
rre
n
tl
y
a
tea
c
h
e
r.
a
t
th
e
Co
ll
e
g
e
o
f
Co
m
p
u
ter,
Un
iv
e
rsity
o
f
A
n
b
a
r.
S
h
e
h
a
s
p
u
b
l
ish
e
d
3
re
f
e
re
e
d
jo
u
r
n
a
ls.
He
r
c
u
rre
n
t
re
se
a
rc
h
in
tere
sts in
c
lu
d
e
E
-
c
o
m
m
e
rc
e
,
c
r
y
p
to
g
ra
p
h
y
,
Da
ta se
c
u
rit
y
,
a
n
d
in
f
o
r
m
a
ti
o
n
tec
h
n
o
lo
g
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.