T
E
L
K
O
M
NIKA
T
elec
o
mm
un
ica
t
io
n,
Co
m
pu
t
ing
,
E
lect
ro
nics
a
nd
Co
ntr
o
l
Vo
l.
1
8
,
No
.
4
,
A
u
g
u
s
t
2020
,
p
p
.
1
9
2
6
~
1
9
3
3
I
SS
N:
1
6
9
3
-
6
9
3
0
,
ac
cr
ed
ited
First Gr
ad
e
b
y
Ke
m
e
n
r
is
te
k
d
i
k
ti,
Dec
r
ee
No
: 2
1
/E/KPT
/2
0
1
8
DOI
: 1
0
.
1
2
9
2
8
/
T
E
L
KOM
NI
KA
.
v
1
8
i
4
.
1
4
7
3
0
1926
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//jo
u
r
n
a
l.u
a
d
.
a
c.
id
/in
d
ex
.
p
h
p
/TELK
OM
N
I
K
A
Deep learni
ng
in
spo
rt
v
ideo
ana
ly
sis
:
a
review
K
ee
rt
ha
na
Ra
ng
a
s
a
m
y
1
,
M
uh
a
mm
a
d A
m
ir
A
s
’
a
ri
2
,
Nur
Az
m
ina
Ra
h
m
a
d
3
,
Nurul F
a
t
hia
h G
ha
za
li
4
,
Sa
ha
rudin
I
s
m
a
il
5
1,
2,
3,
4
,
5
S
c
h
o
o
l
o
f
Bio
m
e
d
ica
l
En
g
in
e
e
rin
g
a
n
d
He
a
lt
h
S
c
ien
c
e
s,
F
a
c
u
lt
y
o
f
En
g
in
e
e
rin
g
,
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
a
lay
sia
,
M
a
la
y
sia
2
S
p
o
rt
In
n
o
v
a
ti
o
n
a
n
d
T
e
c
h
n
o
l
o
g
y
C
e
n
ter,
In
stit
u
te o
f
Hu
m
a
n
Ce
n
tere
d
En
g
in
e
e
rin
g
,
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
a
lay
sia
,
M
a
la
y
sia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
2
,
2019
R
ev
i
s
ed
Ma
r
2
,
2
0
2
0
A
cc
ep
ted
Ma
r
1
3
,
2
0
2
0
S
p
o
rt
is
a
c
o
m
p
e
ti
ti
v
e
f
ield
,
wh
e
re
it
is
a
n
e
lem
e
n
t
o
f
m
e
a
su
re
m
e
n
t
f
o
r
a
c
o
u
n
tri
e
s
d
e
v
e
lo
p
m
e
n
t.
Du
e
to
th
is
re
a
so
n
,
sp
o
r
t
a
n
a
ly
sis
h
a
s
b
e
c
o
m
e
o
n
e
o
f
th
e
m
a
jo
r
c
o
n
tri
b
u
ti
o
n
i
n
a
n
a
ly
sin
g
a
n
d
i
m
p
ro
v
in
g
th
e
p
e
rf
o
r
m
a
n
c
e
lev
e
l
o
f
a
n
a
th
lete
.
V
id
e
o
-
b
a
se
d
m
o
d
a
li
ty
h
a
s
b
e
c
o
m
e
a
c
ru
c
ial
to
o
l
u
se
d
in
sp
o
r
t
a
n
a
ly
sis
b
y
c
o
a
c
h
e
s
a
n
d
p
e
rf
o
rm
a
n
c
e
a
n
a
l
y
sis.
T
h
e
re
w
e
re
w
i
d
e
v
a
riet
y
o
f
tec
h
n
iq
u
e
s
u
se
d
in
sp
o
r
t
v
id
e
o
a
n
a
ly
sis.
T
h
e
m
a
in
p
u
rp
o
se
o
f
th
is
re
v
iew
p
a
p
e
r
is
t
o
c
o
m
p
a
re
a
n
d
u
p
d
a
t
e
re
v
ie
w
b
e
twe
e
n
trad
it
io
n
a
l
h
a
n
d
c
ra
f
ted
a
p
p
ro
a
c
h
a
n
d
d
e
e
p
lea
rn
in
g
a
p
p
r
o
a
c
h
in
sp
o
r
t
v
id
e
o
a
n
a
ly
sis
b
a
se
d
o
n
h
u
m
a
n
a
c
ti
v
it
y
r
e
c
o
g
n
it
io
n
,
o
v
e
rv
iew
o
f
re
c
e
n
t
stu
d
y
in
v
id
e
o
b
a
se
d
h
u
m
a
n
a
c
ti
v
it
y
re
c
o
g
n
it
io
n
i
n
sp
o
r
t
a
n
a
ly
sis
a
n
d
f
in
a
ll
y
c
o
n
c
lu
d
e
d
w
it
h
f
u
tu
re
p
o
ten
ti
a
l
d
irec
ti
o
n
in
sp
o
r
t
v
id
e
o
a
n
a
ly
sis.
K
ey
w
o
r
d
s
:
Dee
p
lear
n
in
g
Hu
m
an
ac
ti
v
it
y
r
ec
o
g
n
itio
n
Sp
o
r
t
v
id
eo
an
al
y
s
i
s
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r
th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Mu
h
a
m
m
ad
Am
ir
A
s
’
ar
i,
Dep
ar
t
m
en
t o
f
B
io
m
ed
ical
E
n
g
in
ee
r
i
n
g
,
Sch
o
o
l o
f
B
io
m
ed
ical
E
n
g
i
n
ee
r
in
g
a
n
d
Hea
lth
Scien
ce
s
,
Facu
lt
y
o
f
E
n
g
i
n
ee
r
i
n
g
,
Un
i
v
er
s
iti T
ek
n
o
lo
g
i M
ala
y
s
ia
,
J
o
h
o
r
B
ah
r
u
,
Ma
lay
s
ia
.
E
m
ail: a
m
ir
.
asar
i
@
b
io
m
ed
ical
.
u
t
m
.
m
y
1.
I
NT
RO
D
UCT
I
O
N
Sp
o
r
t
is
an
ac
ti
v
it
y
t
h
at
i
n
v
o
l
v
es
t
h
e
ac
tiv
e
m
o
v
e
m
en
t
o
f
an
at
h
lete
ei
th
er
i
n
t
h
e
f
o
r
m
o
f
a
tea
m
o
r
in
d
iv
id
u
al
to
co
m
p
lete
w
it
h
th
e
o
p
p
o
s
ite
tea
m
[
1
]
.
I
t
i
s
o
n
e
o
f
t
h
e
ele
m
e
n
ts
o
f
m
ea
s
u
r
e
m
en
t
f
o
r
a
co
u
n
tr
ie
s
d
ev
elo
p
m
en
t
a
s
it
b
r
in
g
s
a
g
o
o
d
r
ep
u
tatio
n
to
a
co
u
n
tr
y
i
f
i
ts
tea
m
w
i
n
i
n
th
e
g
a
m
e
[
2
]
.
I
n
o
r
d
er
to
ac
h
iev
e
h
ig
h
p
er
f
o
r
m
an
ce
in
s
p
o
r
t
b
y
t
h
e
ath
letes,
co
ac
h
e
s
a
n
d
s
p
o
r
ts
p
r
o
f
ess
io
n
al
s
p
la
y
a
v
it
al
r
o
le
to
ev
alu
ate
an
d
tr
ain
th
e
ir
ath
lete
s
[
3
]
.
Am
o
n
g
v
ar
io
u
s
m
eth
o
d
s
,
s
p
o
r
t
v
id
eo
an
al
y
s
is
i
s
o
n
e
o
f
t
h
e
m
et
h
o
d
s
to
ev
alu
ate
th
e
p
er
f
o
r
m
an
ce
le
v
el
o
f
at
h
le
tes
a
n
d
to
e
n
h
an
ce
tr
ain
in
g
tec
h
n
iq
u
es
[
4
]
.
A
p
r
o
ce
s
s
to
as
s
e
s
s
th
e
p
er
f
o
r
m
a
n
ce
lev
el
o
f
a
n
at
h
lete
is
k
n
o
w
n
a
s
p
er
f
o
r
m
a
n
ce
a
n
al
y
s
i
s
[
5
]
.
P
er
f
o
r
m
an
ce
an
a
l
y
s
is
ca
n
b
e
d
iv
id
e
in
to
t
w
o
w
h
ic
h
ar
e
tech
n
ical
a
n
al
y
s
is
a
n
d
tactica
l
o
r
n
o
tatio
n
al
an
al
y
s
i
s
[
6
]
.
T
h
r
o
u
g
h
tech
n
ical
a
n
al
y
s
is
,
w
e
w
o
u
ld
g
e
t
an
a
n
s
w
er
to
t
h
e
q
u
esti
o
n
:
h
o
w
d
o
es
t
h
e
g
a
m
e
is
p
er
f
o
r
m
ed
b
y
t
h
e
a
th
lete
s
.
O
n
t
h
e
o
th
er
h
a
n
d
,
th
r
o
u
g
h
tactica
l o
r
n
o
tatio
n
al
an
a
l
y
s
is
t
h
e
q
u
est
io
n
o
f
w
h
at
ac
ti
v
it
y
is
p
er
f
o
r
m
ed
w
o
u
ld
b
e
an
s
w
er
ed
[
7
]
.
Ov
er
t
h
e
p
ast
f
e
w
y
ea
r
s
,
i
n
t
h
e
f
ield
o
f
co
m
p
u
ter
v
i
s
io
n
,
d
i
f
f
er
e
n
t
ap
p
r
o
ac
h
es
h
av
e
b
ee
n
i
m
p
le
m
en
ted
to
an
al
y
s
e
s
p
o
r
t
v
id
eo
s
.
Du
r
in
g
th
e
ea
r
l
y
s
tag
e
tr
ad
itio
n
al
h
a
n
d
cr
af
te
d
ap
p
r
o
ac
h
es
w
er
e
p
r
o
p
o
s
ed
b
y
m
an
y
r
esear
ch
er
s
f
o
r
s
p
o
r
t
v
id
eo
an
al
y
s
i
s
[
8
]
.
A
f
ter
th
e
d
ev
elo
p
m
e
n
t
o
f
tech
n
o
lo
g
y
,
d
ee
p
lean
in
g
ap
p
r
o
ac
h
h
a
s
b
ec
o
m
e
t
h
e
c
u
r
r
en
t
to
p
ic
o
f
in
te
r
est
in
s
p
o
r
t
v
id
eo
an
al
y
s
i
s
d
u
e
to
its
s
u
cc
es
s
f
u
l
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
Dee
p
lea
r
n
in
g
in
s
p
o
r
t v
id
eo
a
n
a
lysi
s
:
a
r
ev
iew
(
K
ee
r
th
a
n
a
R
a
n
g
a
s
a
my
)
1927
p
er
f
o
r
m
a
n
ce
i
n
o
th
er
co
m
p
u
ter
v
is
io
n
ar
ea
s
s
u
c
h
as
h
u
m
an
-
co
m
p
u
ter
i
n
ter
f
ac
e,
h
a
n
d
w
r
iti
n
g
r
ec
o
g
n
it
io
n
an
d
s
p
ee
ch
r
ec
o
g
n
itio
n
[
9
]
.
Gen
er
all
y
,
s
p
o
r
t
v
id
eo
a
n
al
y
s
i
s
u
s
in
g
s
u
c
h
ap
p
r
o
ac
h
ca
n
b
e
d
iv
id
e
i
n
to
s
e
v
er
al
d
ir
ec
tio
n
s
s
u
c
h
a
s
tr
ac
k
i
n
g
p
l
a
y
er
s
a
s
w
e
ll
a
s
b
all,
tr
aj
ec
to
r
y
d
etec
tio
n
an
d
ac
tio
n
r
ec
o
g
n
itio
n
.
T
h
is
p
ap
er
ai
m
s
to
r
e
v
ie
w
c
u
r
r
en
t
tr
en
d
s
i
n
d
ee
p
lea
n
in
g
ap
p
r
o
ac
h
i
n
h
u
m
a
n
ac
t
iv
i
t
y
r
ec
o
g
n
itio
n
-
b
ased
s
p
o
r
t
v
id
eo
an
al
y
s
is
.
T
h
e
r
e
m
ai
n
d
er
o
f
t
h
i
s
p
ap
er
is
o
r
g
a
n
ized
as
s
ec
tio
n
2
illu
s
tr
ates
tr
ad
itio
n
al
h
an
d
c
r
af
ted
ar
ch
itect
u
r
e.
Nex
t
s
ec
tio
n
3
p
r
esen
ts
d
ee
p
l
ea
r
n
in
g
ar
ch
itectu
r
e.
Sec
tio
n
4
d
is
cu
s
s
es
t
h
e
d
r
a
w
b
ac
k
o
f
s
t
ate
o
f
t
h
e
ar
t
w
o
r
k
s
an
d
Sectio
n
5
p
r
esen
t c
o
n
cl
u
s
i
o
n
an
d
f
u
t
u
r
e
w
o
r
k
s
to
ad
d
r
ess
.
2.
T
RAD
I
T
I
O
NAL H
AN
DCR
AF
T
E
D
A
RCH
I
T
E
C
T
UR
E
I
n
th
i
s
s
ec
tio
n
,
o
v
er
v
ie
w
o
f
h
an
d
cr
af
ted
ar
ch
itec
tu
r
e
a
n
d
its
ap
p
licatio
n
in
s
p
o
r
t
v
id
eo
an
al
y
s
i
s
w
a
s
s
tu
d
ied
b
ased
o
n
p
r
io
r
r
esear
c
h
.
2
.
1
.
O
v
er
v
ie
w
o
f
ha
nd
cr
a
f
t
e
d a
rc
hite
ct
ure
B
ef
o
r
e
th
e
ad
v
an
ce
m
e
n
t
o
f
te
ch
n
o
lo
g
y
a
n
d
th
e
e
m
er
g
e
n
ce
o
f
d
ee
p
lear
n
in
g
ar
c
h
itect
u
r
e,
s
p
o
r
t
v
id
eo
an
al
y
s
is
is
m
ain
l
y
p
er
f
o
r
m
ed
u
s
in
g
h
a
n
d
cr
af
ted
f
ea
tu
r
e
s
[
9
]
.
T
r
ad
itio
n
al
h
an
d
cr
a
f
ted
f
ea
tu
r
es
ar
e
h
u
m
an
d
esig
n
f
ea
t
u
r
es
f
o
r
s
p
ec
if
ic
p
r
o
b
lem
s
[
1
0
]
.
I
t
co
n
s
is
ts
o
f
f
ea
t
u
r
e
d
escr
ip
to
r
s
an
d
f
ea
t
u
r
e
ex
t
r
ac
to
r
s
th
at
u
s
ed
to
m
an
u
all
y
ex
tr
ac
t
t
h
e
i
m
p
o
r
tan
t
in
f
o
r
m
atio
n
f
r
o
m
t
h
e
s
p
o
r
t
v
id
eo
f
o
r
f
u
r
th
er
an
al
y
s
is
.
Ho
w
e
v
er
,
th
ese
tr
ad
itio
n
a
l
h
a
n
d
cr
af
ted
f
ea
t
u
r
es
co
u
ld
o
n
l
y
ex
ec
u
te
l
o
w
-
lev
e
l
f
ea
t
u
r
es
e
x
tr
ac
tio
n
[
1
1
]
.
Fu
r
th
er
in
t
h
is
s
ec
tio
n
w
o
u
ld
li
k
e
to
p
r
esen
t
th
e
e
x
is
ti
n
g
tr
ad
itio
n
al
h
an
d
cr
af
ted
ar
ch
itect
u
r
e
th
at
w
a
s
u
s
ed
w
id
el
y
in
co
m
p
u
ter
v
is
io
n
to
r
ec
o
g
n
iz
e
ac
tiv
it
y
in
v
i
d
eo
.
Du
r
in
g
th
e
p
r
eh
is
to
r
ic
a
g
e
s
o
f
s
p
o
r
t
v
id
eo
an
a
l
y
s
is
,
ac
tio
n
r
ec
o
g
n
itio
n
w
a
s
p
er
f
o
r
m
ed
u
s
i
n
g
m
an
u
all
y
h
u
m
a
n
d
esi
g
n
f
ea
t
u
r
es
to
ex
tr
ac
t
an
d
r
ep
r
esen
t
f
ea
t
u
r
es
f
o
r
ac
tio
n
r
ec
o
g
n
it
io
n
b
y
w
h
ic
h
t
h
is
tech
n
iq
u
e
i
s
ca
lled
a
h
a
n
d
cr
af
ted
ap
p
r
o
ac
h
.
A
t
f
ir
s
t,
lo
w
-
l
ev
el
ac
tio
n
f
ea
t
u
r
es
s
u
c
h
a
s
h
is
to
g
r
a
m
o
f
o
p
tical
f
lo
w
s
(
HOF)
,
h
is
to
g
r
a
m
o
f
g
r
ad
ien
ts
(
HOG)
,
s
p
ar
s
e
s
p
ati
al
-
te
m
p
o
r
al
in
ter
est
p
o
in
t
s
(
S
T
I
P
)
f
ea
tu
r
es
w
er
e
d
esig
n
ed
to
ex
tr
ac
t
f
ea
t
u
r
es
i
n
v
id
eo
s
f
o
r
a
n
al
y
s
is
[
9
]
.
T
h
en
,
th
o
s
e
s
p
ar
s
e
f
ea
t
u
r
es
w
er
e
e
m
b
e
llis
h
ed
to
d
en
s
e
s
p
atio
-
te
m
p
o
r
al
f
ea
t
u
r
es
[
1
2
]
.
Af
ter
s
o
m
e
ti
m
e,
i
m
p
r
o
v
ed
d
en
s
e
tr
aj
ec
to
r
ies
(
iDT
)
w
er
e
in
n
o
v
ated
w
h
ic
h
co
m
p
r
is
e
s
o
p
tical
f
lo
w
a
n
d
s
p
ee
d
ed
u
p
r
o
b
u
s
t
f
ea
t
u
r
es
(
S
UR
F)
[
1
3
]
.
T
o
f
u
r
th
er
e
n
h
a
n
ce
iDT
m
u
l
ti
-
la
y
er
s
tak
ed
f
i
s
h
er
v
ec
to
r
(
FV)
w
as
d
ev
elo
p
ed
[
1
4
]
.
Mu
lti
-
s
k
ip
Fe
atu
r
e
s
tak
i
n
g
w
a
s
al
s
o
u
s
ed
f
o
r
f
ea
t
u
r
e
e
x
tr
ac
tio
n
o
f
ac
tio
n
r
ec
o
g
n
i
tio
n
i
n
v
id
eo
s
w
h
ic
h
i
m
p
r
o
v
e
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
ac
tio
n
r
ec
o
g
n
i
ti
o
n
[
1
5
]
.
A
lt
h
o
u
g
h
,
th
ese
h
an
d
cr
af
ted
f
ea
t
u
r
es
wer
e
ev
o
lv
ed
w
it
h
ti
m
e
an
d
a
ls
o
i
m
p
r
o
v
ed
in
p
er
f
o
r
m
a
n
ce
b
u
t
it
co
u
ld
o
n
l
y
p
er
f
o
r
m
i
n
a
s
p
ec
if
ic
p
r
o
b
le
m
.
I
t
i
s
n
o
t
f
le
x
ib
le
to
b
e
u
s
ed
in
o
th
er
d
atase
ts
.
T
ab
le
1
s
h
o
w
s
s
o
m
e
o
f
t
h
e
e
x
is
ti
n
g
p
o
p
u
lar
h
an
d
cr
af
ted
f
ea
tu
r
es
u
s
ed
i
n
ac
ti
o
n
r
ec
o
g
n
itio
n
f
o
r
v
id
eo
an
a
l
y
s
i
s
s
u
c
h
a
s
s
p
o
r
t
v
id
eo
an
al
y
s
is
.
T
ab
le
1
.
L
is
t o
f
p
o
p
u
lar
h
an
d
c
r
af
ted
f
ea
t
u
r
es u
s
ed
f
o
r
ac
tio
n
r
ec
o
g
n
itio
n
in
v
id
eo
s
H
a
n
d
c
r
a
f
t
e
d
f
e
a
t
u
r
e
s
R
e
f
e
r
e
n
c
e
H
i
st
o
g
r
a
m o
f
O
p
t
i
c
a
l
F
l
o
w
(
H
O
F
)
a
n
d
H
i
st
o
g
r
a
m o
f
G
r
a
d
i
e
n
t
s
(
H
O
G
)
[
1
6
,
17]
S
p
a
r
se
S
p
a
t
i
o
-
t
e
mp
o
r
a
l
i
n
t
e
r
e
st
p
o
i
n
t
s
[
1
8
]
I
mp
r
o
v
e
d
D
e
n
se
T
r
a
j
e
c
t
o
r
i
e
s (i
D
T
)
[
1
3
]
F
i
sh
V
e
c
t
o
r
(
F
V
)
[
1
4
]
M
u
l
t
i
-
l
a
y
e
r
F
e
a
t
u
r
e
S
t
a
c
k
i
n
g
(
M
L
F
S
)
[
1
5
]
2
.
2
.
Sp
o
rt
v
ideo
a
na
ly
s
is
us
i
ng
ha
nd
cr
a
f
t
ed
f
ea
t
ures
Du
r
in
g
t
h
e
ea
r
l
y
s
tag
e
s
o
f
v
id
eo
an
al
y
s
is
i
n
co
m
p
u
ter
v
is
io
n
s
u
ch
a
s
in
s
p
o
r
ts
,
h
an
d
cr
af
t
ed
f
ea
tu
r
e
s
w
er
e
u
s
ed
as
m
en
t
io
n
ed
ea
r
lier
.
Fo
r
in
s
tan
ce
,
A
b
d
u
l
m
u
n
e
m
e
t
al.
[
1
9
]
w
er
e
p
r
o
p
o
s
ed
a
m
e
th
o
d
f
o
r
ac
tio
n
r
ec
o
g
n
itio
n
b
ased
o
n
s
al
ien
t
o
b
j
ec
t
s
p
o
ttin
g
u
s
i
n
g
l
o
ca
l
a
n
d
g
lo
b
al
d
escr
ip
to
r
s
in
KT
H
an
d
U
C
F
-
Sp
o
r
ts
d
atasets
.
I
n
t
h
is
m
et
h
o
d
,
th
e
y
w
er
e
u
s
in
g
h
a
n
d
cr
af
ted
3
D
-
SIFT
-
HOO
F
(
SGSH)
f
ea
t
u
r
es
w
h
ic
h
w
er
e
p
ass
th
r
o
u
g
h
SV
M
af
ter
en
co
d
in
g
w
it
h
b
ag
o
f
v
is
u
al
w
o
r
d
s
ap
p
r
o
ac
h
.
B
esid
es,
C
ar
n
o
n
n
ea
u
et
al.
[
2
0
]
p
r
esen
ted
a
n
o
v
el
tech
n
iq
u
e
to
id
en
t
if
y
p
l
a
y
b
r
ea
k
e
v
e
n
ts
in
h
o
c
k
e
y
v
i
d
eo
s
.
T
h
ey
u
s
ed
b
ag
-
of
-
w
o
r
d
s
ev
e
n
t
d
etec
to
r
s
to
id
en
ti
f
y
t
h
e
k
e
y
e
v
e
n
ts
li
k
e
li
n
e
-
c
h
a
n
g
e
a
n
d
p
la
y
-
b
r
ea
k
ST
I
P
w
er
e
u
s
ed
in
f
i
n
al
d
ec
is
io
n
m
ak
in
g
b
y
cr
ea
tin
g
co
n
tex
t d
escr
ip
to
r
.
Fig
u
r
e
1
th
e
s
h
o
w
s
s
c
h
e
m
atic
b
lo
ck
d
ia
g
r
a
m
o
f
t
h
e
ev
e
n
t d
etec
to
r
u
s
ed
i
n
[
2
0
]
.
Z
h
u
et
al.
[
7
]
in
tr
o
d
u
ce
d
a
n
o
v
el
m
et
h
o
d
to
d
etec
t
an
ev
e
n
t
i
n
a
s
o
cc
er
g
a
m
e
s
to
e
x
tr
a
ct
tactica
l
in
f
o
r
m
atio
n
.
I
n
t
h
i
s
r
esear
ch
w
o
r
k
,
t
h
e
y
e
m
p
lo
y
m
u
l
ti
-
o
b
j
ec
t
tr
aj
ec
to
r
y
a
n
d
f
ield
lo
ca
tio
n
s
o
f
th
e
e
v
e
n
t
s
h
o
ts
to
r
ec
o
g
n
ize
th
e
s
e
m
a
n
tic
e
v
e
n
t
in
t
h
e
s
o
cc
er
g
a
m
e
s
.
Mo
r
eo
v
er
,
L
ee
et
al.
[
2
1
]
im
p
le
m
e
n
t
p
atter
n
m
atch
in
g
tech
n
iq
u
es
to
a
u
to
m
at
icall
y
r
e
co
g
n
ize
e
v
en
t
s
i
n
s
o
cc
er
g
a
m
es
u
s
in
g
a
m
u
lti
-
o
b
j
ec
t
tr
ac
k
in
g
u
n
it
an
d
m
o
t
io
n
r
ec
o
g
n
izer
.
T
h
eir
p
r
o
p
o
s
ed
b
l
o
ck
d
iag
r
a
m
i
s
s
h
o
w
n
i
n
Fi
g
u
r
e
2
.
L
ie
n
et
al.
[
2
2
]
s
tu
d
ied
s
ce
n
e
-
b
ased
ev
e
n
t
d
etec
tio
n
f
o
r
b
aseb
all
v
id
eo
s
b
y
i
m
p
le
m
e
n
ti
n
g
v
ar
io
u
s
h
an
d
cr
af
ted
f
ea
t
u
r
es
s
u
c
h
a
s
i
m
a
g
e
-
b
a
s
ed
f
ea
t
u
r
es,
o
b
j
ec
t
-
b
ased
f
ea
tu
r
e
s
an
d
g
lo
b
al
m
o
tio
n
to
ca
p
t
u
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
1
8
,
No
.
4
,
A
u
g
u
s
t
2020:
1
9
2
6
-
1
9
3
3
1928
s
e
m
a
n
tic
i
n
f
o
r
m
atio
n
f
r
o
m
t
h
e
b
aseb
all
s
p
o
r
t
v
id
eo
.
T
h
en
th
o
s
e,
e
x
tr
ac
ted
f
ea
t
u
r
es
wer
e
f
ed
in
to
h
id
d
en
Ma
r
v
o
n
m
o
d
el
(
HM
M)
to
clas
s
if
y
t
h
e
d
etec
ted
e
v
en
ts
.
C
h
e
n
et
al.
[
2
3
]
d
ev
elo
p
ed
a
n
o
v
el
h
an
d
cr
af
ted
m
o
d
el
b
ased
o
n
p
la
y
er
tr
aj
ec
to
r
ies
r
ec
o
n
s
tr
u
ct
io
n
m
et
h
o
d
w
h
ic
h
i
n
clu
d
e
s
co
u
r
t
d
et
ec
tio
n
,
p
la
y
e
r
d
etec
tio
n
,
p
la
y
er
tr
ac
k
in
g
a
n
d
h
o
m
o
g
r
ap
h
y
tr
a
n
s
f
o
r
m
a
tio
n
i
n
b
r
o
ad
ca
s
ted
b
as
k
etb
all
v
id
eo
s
.
Fig
u
r
e
1
.
Sch
e
m
a
tic
b
lo
ck
d
iag
r
a
m
o
f
e
v
e
n
t d
etec
to
r
s
in
[
2
0
]
Fig
u
r
e
2
.
B
lo
ck
d
iag
r
a
m
o
f
s
o
cc
er
ev
en
t d
etec
tio
n
s
y
s
te
m
[
2
1
]
L
u
et
al.
[
2
4
]
co
n
s
tr
u
cted
a
s
y
s
te
m
to
a
u
to
m
atica
ll
y
tr
a
ck
v
ar
io
u
s
h
o
c
k
e
y
p
la
y
er
s
a
s
w
e
ll
as
r
ec
o
g
n
ize
t
h
eir
ac
tiv
i
t
y
co
n
cu
r
r
en
tl
y
i
n
b
r
o
ad
ca
s
ted
h
o
ck
e
y
v
id
eo
.
I
n
t
h
is
s
tu
d
y
,
H
O
G
d
escr
ip
to
r
s
w
er
e
u
tili
s
ed
to
d
etec
t
t
h
e
p
la
y
er
s
an
d
to
m
o
d
el
th
e
ap
p
ea
r
an
c
e
o
f
t
h
e
p
la
y
er
s
a
p
r
o
b
ab
ilis
tic
f
r
a
m
e
w
o
r
k
w
a
s
d
esig
n
ed
w
ith
a
m
i
x
t
u
r
e
o
f
lo
ca
l
s
u
b
s
p
ac
es.
A
d
d
itio
n
all
y
,
b
o
o
s
ted
p
ar
ticle
f
ilter
(
B
P
F)
w
as
e
m
p
lo
y
ed
i
n
t
h
i
s
tr
ac
k
in
g
s
y
s
te
m
.
M
u
k
h
er
j
ee
et
al.
[
2
5
]
p
r
o
p
o
s
ed
a
n
o
v
el
d
e
s
cr
ip
to
r
b
ased
o
n
im
p
r
o
v
ed
d
en
s
e
tr
aj
ec
to
r
ies
in
h
u
m
a
n
ac
ti
o
n
an
d
e
v
en
t
r
ec
o
g
n
i
tio
n
.
T
h
eir
d
esig
n
al
s
o
u
s
ed
Fis
h
er
Vec
to
r
(
FV)
an
d
th
e
p
r
o
p
o
s
ed
n
o
v
el
ap
p
r
o
ac
h
w
as tr
ai
n
ed
w
it
h
b
in
ar
y
s
u
p
p
o
r
t v
ec
to
r
m
ac
h
i
n
e
(
S
VM
)
.
T
h
e
d
ataset
u
s
ed
f
o
r
t
h
is
r
esear
ch
w
as U
C
F
s
p
o
r
ts
,
C
MU
Mo
ca
p
an
d
Ho
lly
w
o
o
d
2
d
atasets
f
o
r
ev
en
t a
n
d
ac
t
io
n
r
ec
o
g
n
it
io
n
.
3.
DE
E
P
L
E
A
RNIN
G
I
n
th
is
s
ec
tio
n
,
o
v
er
v
ie
w
o
f
d
ee
p
lear
n
in
g
ar
ch
ictu
r
e
a
n
d
it
s
ap
p
licatio
n
i
n
s
p
o
r
t
v
id
eo
an
al
y
s
i
s
w
a
s
r
ev
ie
w
ed
b
ased
o
n
p
r
ev
io
u
s
s
t
u
d
ies.
3
.
1
.
O
v
er
v
ie
w
o
f
deep
lea
rning
a
rc
hite
ct
ure
Dee
p
lear
n
i
n
g
ar
ch
itec
tu
r
e
wo
r
k
s
a
u
to
m
at
icall
y
b
y
d
ir
ec
tly
cla
s
s
i
f
y
in
g
r
a
w
i
n
p
u
t
i
m
ag
es
o
r
v
id
eo
f
r
a
m
e
s
t
h
r
o
u
g
h
m
u
ltip
le
p
r
o
ce
s
s
i
n
g
la
y
er
s
s
o
as
to
lear
n
an
d
r
ep
r
ese
n
t
d
ata
[
2
6
]
.
Un
li
k
e
tr
ad
itio
n
a
l
h
an
d
cr
af
ted
ar
ch
itect
u
r
e,
it
d
o
es
n
o
t
r
eq
u
ir
es
an
y
f
ea
t
u
r
e
d
escr
ip
to
r
s
o
r
f
ea
tu
r
e
ex
tr
a
cto
r
s
.
Fo
r
in
s
tan
ce
,
d
ee
p
lear
n
in
g
ar
ch
i
tectu
r
e
u
s
e
s
lo
ca
l
p
er
ce
p
tio
n
,
d
o
w
n
p
o
o
lin
g
,
w
ei
g
h
t
s
h
ar
in
g
,
a
m
u
lti
-
c
o
n
v
o
l
u
tio
n
k
er
n
e
l,
etc.
to
au
to
m
at
icall
y
lear
n
lo
ca
l
f
ea
t
u
r
es
f
r
o
m
j
u
s
t
a
s
e
g
m
en
t
o
f
a
n
i
m
a
g
e
r
ath
er
t
h
a
n
w
h
o
le
i
m
ag
e
[
2
7
]
.
Dee
p
lear
n
in
g
tec
h
n
iq
u
es
ab
le
to
class
if
y
h
i
g
h
-
le
v
el
o
r
co
m
p
le
x
ac
tio
n
r
ec
o
g
n
itio
n
wh
ich
attr
ac
t
s
h
u
g
e
r
esear
ch
o
f
in
ter
est
[
2
8
]
.
E
x
a
m
p
les
o
f
w
id
el
y
u
s
ed
d
ee
p
l
ea
r
n
in
g
m
o
d
els
i
s
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
et
w
o
r
k
(
C
NN)
,
r
ec
u
r
r
en
t n
e
u
r
al
n
e
t
wo
r
k
(
R
NN)
a
n
d
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
(
L
ST
M)
.
E
x
ce
lle
n
t
p
er
f
o
r
m
an
ce
w
it
h
o
v
er
w
h
e
l
m
in
g
ac
cu
r
ac
y
o
f
d
ee
p
lear
n
in
g
i
n
v
i
s
u
a
l
task
in
s
p
ir
ed
th
e
ex
p
lo
itatio
n
o
f
d
ee
p
lear
n
i
n
g
i
n
v
id
eo
an
al
y
s
is
.
I
n
i
tiall
y
,
C
NN
w
o
r
k
s
i
n
d
ep
en
d
en
tl
y
to
ex
tr
ac
t
in
f
o
r
m
atio
n
f
r
o
m
th
e
s
till
i
m
ag
e
s
[
2
9
]
.
Ho
w
e
v
er
,
2
D
-
C
NN
f
ails
to
ex
tr
ac
t
te
m
p
o
r
al
in
f
o
r
m
atio
n
in
v
id
eo
s
eq
u
en
ce
s
.
I
n
o
r
d
er
to
o
v
er
co
m
e
th
i
s
is
s
u
e,
3
D
-
C
NN
i
s
th
e
n
co
n
s
tr
u
cte
d
to
ex
tr
ac
t
b
o
th
s
p
atial
a
n
d
te
m
p
o
r
al
in
f
o
r
m
at
io
n
o
f
v
id
eo
f
r
a
m
es
[
3
0
]
.
Fo
llo
w
i
n
g
t
h
is
R
NN
w
as
u
s
ed
in
ac
t
io
n
r
ec
o
g
n
itio
n
.
R
NN
b
ased
m
et
h
o
d
ef
f
ec
ti
v
el
y
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
Dee
p
lea
r
n
in
g
in
s
p
o
r
t v
id
eo
a
n
a
lysi
s
:
a
r
ev
iew
(
K
ee
r
th
a
n
a
R
a
n
g
a
s
a
my
)
1929
ca
p
tu
r
e
te
m
p
o
r
al
in
f
o
r
m
atio
n
w
h
er
eb
y
its
c
u
r
r
en
t
p
r
ed
ictio
n
is
b
ased
o
n
b
o
th
cu
r
r
en
t
o
b
s
er
v
atio
n
s
a
s
w
ell
as
p
ast
in
f
o
r
m
atio
n
[
3
1
]
.
B
u
t
in
s
p
ite
o
f
t
h
at,
R
NN
ar
c
h
itect
u
r
e
o
n
l
y
h
as
s
h
o
r
t
ter
m
m
e
m
o
r
y
,
w
h
ic
h
co
u
ld
n
o
t
ap
p
ly
in
t
h
e
r
ea
l
-
w
o
r
ld
s
ce
n
ar
io
.
T
o
allev
iate
th
is
p
r
o
b
le
m
,
L
ST
M
m
o
d
el
w
as
p
r
o
p
o
s
ed
.
T
h
is
m
o
d
el
ab
le
to
ex
tr
ac
t
te
m
p
o
r
al
i
n
f
o
r
m
atio
n
f
r
o
m
s
eq
u
e
n
tial
v
id
eo
d
ata.
L
S
T
M
m
o
d
el
h
as
a
m
e
m
o
r
y
u
n
i
t
th
at
d
ec
id
es
w
h
en
to
r
em
e
m
b
er
an
d
f
o
r
g
et
h
id
d
e
n
s
tates
[
3
2
]
.
Du
e
to
its
s
u
p
er
io
r
ity
,
t
h
e
L
ST
M
m
o
d
el
b
r
o
ad
l
y
u
s
ed
i
n
co
m
p
u
ter
v
is
io
n
ap
p
licatio
n
s
s
u
c
h
as a
ct
io
n
r
ec
o
g
n
it
io
n
.
T
ab
le
2
s
h
o
w
a
co
m
p
ar
is
o
n
b
et
w
ee
n
d
ee
p
lear
n
in
g
m
o
d
els.
T
ab
le
2
.
C
o
m
p
ar
is
o
n
b
et
w
ee
n
d
ee
p
lear
n
in
g
m
o
d
el
M
o
d
e
l
A
d
v
a
n
t
a
g
e
D
r
a
w
b
a
c
k
2D
-
C
N
N
A
u
t
o
mat
i
c
a
l
l
y
c
a
p
t
u
r
e
t
h
e
s
p
a
t
i
a
l
i
n
f
o
r
ma
t
i
o
n
i
n
t
h
e
i
m
a
g
e
p
a
t
c
h
e
s
C
o
u
l
d
n
o
t
a
b
l
e
t
o
c
a
p
t
u
r
e
t
h
e
t
e
mp
o
r
a
l
i
n
f
o
r
mat
i
o
n
i
n
v
i
d
e
o
d
a
t
a
3D
-
C
N
N
A
u
t
o
mat
i
c
a
l
l
y
c
a
p
t
u
r
e
b
o
t
h
sp
a
t
i
a
l
a
n
d
t
e
mp
o
r
a
l
i
n
f
o
r
mat
i
o
n
Ex
p
e
n
si
v
e
mo
d
e
l
d
u
e
t
o
3
D
R
N
N
A
u
t
o
mat
i
c
a
l
l
y
c
a
p
t
u
r
e
t
h
e
t
e
m
p
o
r
a
l
i
n
f
o
r
ma
t
i
o
n
i
n
se
q
u
e
n
t
i
a
l
d
a
t
a
H
a
s sh
o
r
t
me
mo
r
y
a
b
i
l
i
t
y
,
c
o
u
l
d
n
o
t
a
p
p
l
y
i
n
r
e
a
l
si
t
u
a
t
i
o
n
G
r
a
d
i
e
n
t
e
x
p
l
o
si
o
n
L
S
T
M
A
u
t
o
mat
i
c
a
l
l
y
c
a
p
t
u
r
e
t
h
e
t
e
m
p
o
r
a
l
i
n
f
o
r
ma
t
i
o
n
i
n
se
q
u
e
n
t
i
a
l
d
a
t
a
N
I
L
3.
2
.
Dee
p
lea
rning
a
rc
hite
ct
ure
in s
po
rt
v
ideo
a
na
ly
s
is
Desp
ite
th
e
a
s
to
n
i
s
h
in
g
p
er
f
o
r
m
a
n
ce
o
f
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
in
v
ar
io
u
s
co
m
p
u
t
er
v
is
io
n
ap
p
licatio
n
s
u
ch
as
v
o
ice
r
ec
o
g
n
itio
n
,
te
x
t
r
ec
o
g
n
itio
n
it
also
ac
h
ie
v
es
o
u
t
s
tan
d
i
n
g
r
es
u
lts
in
s
p
o
r
t
v
id
eo
an
al
y
s
is
i
n
r
ec
en
t
y
ea
r
s
.
A
lt
h
o
u
g
h
it
is
s
t
ill
in
t
h
e
ea
r
l
y
s
ta
g
e
o
f
ap
p
licatio
n
in
s
p
o
r
t
v
id
eo
an
al
y
s
i
s
an
d
o
n
l
y
v
er
y
f
e
w
r
e
s
ea
r
ch
h
as
b
ee
n
d
o
n
e,
y
et
s
o
f
ar
i
ts
p
er
f
o
r
m
a
n
ce
is
m
o
r
e
ac
c
u
r
ate
as
co
m
p
ar
ed
to
tr
ad
itio
n
al
ap
p
r
o
ac
h
es
an
d
it’
s
g
ett
in
g
m
o
r
e
atte
n
tio
n
p
r
ese
n
tl
y
[
3
3
]
.
T
o
r
a
et
al
.
[
3
4
]
p
r
o
p
o
s
ed
a
n
o
v
el
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
in
clas
s
i
f
y
in
g
p
u
ck
p
o
s
s
ess
io
n
e
v
en
t
s
i
n
ice
h
o
c
k
e
y
.
T
h
e
y
u
s
ed
p
r
e
-
tr
ain
ed
C
NN
to
f
ir
s
t
e
x
tr
ac
t
th
e
f
ea
t
u
r
es
t
h
e
n
u
s
e
L
ST
M
f
o
r
class
i
f
icatio
n
o
f
th
e
f
i
v
e
t
y
p
es
o
f
e
v
en
t
s
w
h
ich
ar
e
d
u
m
p
i
n
,
d
u
m
p
o
u
t,
lo
o
s
e
p
u
ck
r
ec
o
v
er
y
,
p
as
s
a
n
d
s
h
o
t.
So
z
y
k
i
n
et
al.
[
3
5
]
p
r
esen
ted
a
3
D
C
NN
b
ased
ac
tio
n
r
ec
o
g
n
itio
n
s
y
s
te
m
f
o
r
m
u
lti
-
clas
s
i
m
b
ala
n
ce
d
in
ice
h
o
ck
e
y
.
T
h
e
y
f
ir
s
t
e
x
tr
a
ct
f
ea
tu
r
e
s
f
r
o
m
b
o
th
s
in
g
le
i
m
a
g
e
an
d
a
s
lice
o
f
f
r
a
m
e
s
u
s
in
g
C
NN.
T
h
en
,
to
s
o
lv
e
th
e
m
u
lt
i
-
cla
s
s
i
m
b
al
an
ce
,
th
e
y
i
n
tr
o
d
u
ce
t
w
o
d
if
f
er
en
t
s
tr
ateg
ies
th
a
t
ca
n
co
m
p
ar
e
i
ts
p
er
f
o
r
m
an
ce
s
ep
ar
atel
y
w
h
ic
h
ar
e
th
e
e
n
s
e
m
b
le
o
f
k
i
n
d
ep
en
d
en
t
s
i
n
g
le
-
lab
el
lear
n
i
n
g
n
et
w
o
r
k
s
an
d
a
s
i
n
g
le
m
u
lti
-
l
ab
el
k
-
o
u
tp
u
t
n
et
w
o
r
k
.
I
t
is
f
o
u
n
d
t
h
at
as
co
m
p
ar
ed
to
th
e
en
s
e
m
b
le
m
o
d
el,
s
in
g
le
m
u
lti
-
lab
el
k
o
u
tp
u
t
m
o
d
el
ac
h
ie
v
e
s
h
ig
h
p
er
f
o
r
m
an
ce
.
L
o
n
g
ten
g
et
al.
[
3
6
]
d
esig
n
ed
a
j
o
in
t
f
r
a
m
e
w
o
r
k
co
m
p
r
is
i
n
g
o
f
b
o
th
ath
lete
tr
ac
k
in
g
a
n
d
ac
t
io
n
r
ec
o
g
n
itio
n
in
s
p
o
r
t
v
id
eo
s
u
s
in
g
s
ca
lin
g
an
d
o
cc
lu
s
io
n
r
o
b
u
s
t
tr
ac
k
er
w
it
h
s
p
atial
p
y
r
a
m
id
p
o
o
lin
g
n
et
w
o
r
k
(
SP
P
-
n
e
t)
[
3
7
]
an
d
L
ST
M
to
ex
tr
ac
t
m
o
tio
n
,
s
p
atial
a
n
d
te
m
p
o
r
al
f
ea
tu
r
es.
J
ian
g
et
al.
[
3
8
]
in
tr
o
d
u
ce
d
ee
p
lear
n
in
g
b
ased
a
u
to
m
atic
s
o
cc
er
v
id
eo
ev
e
n
t d
etec
tio
n
b
y
u
s
i
n
g
C
NN
an
d
R
NN.
T
h
is
p
ap
er
f
o
cu
s
ed
o
n
4
t
y
p
es
o
f
e
v
en
t
w
h
ich
ar
e
g
o
al,
g
o
al
atte
m
p
t,
co
r
n
er
an
d
ca
r
d
.
C
NN
m
o
d
el
w
a
s
u
s
ed
to
e
x
tr
ac
t
i
m
ag
e
f
e
atu
r
es.
An
d
R
NN
s
tr
u
ct
u
r
e
w
a
s
u
s
ed
to
ca
p
tu
r
e
te
m
p
o
r
al
r
elatio
n
.
Ho
w
e
v
er
,
in
t
h
is
p
ap
er
th
r
ee
d
i
f
f
er
en
t
t
y
p
es
o
f
R
NN
s
tr
u
ctu
r
e
s
(
tr
ad
itio
n
al
R
N
N
m
o
d
el,
L
ST
M
m
o
d
el
an
d
g
ated
r
ec
u
r
r
en
t
u
n
i
t (
G
R
U)
m
o
d
el)
w
er
e
u
s
ed
to
d
eter
m
in
e
th
e
o
p
ti
m
u
m
m
o
d
el.
L
ST
M
is
th
e
b
e
s
t
-
p
er
f
o
r
m
ed
m
o
d
el
a
m
o
n
g
t
h
o
s
e
t
h
r
ee
R
NN
s
tr
u
ct
u
r
es.
Ho
n
g
et
al.
[
3
9
]
d
eliv
er
ed
d
ee
p
tr
an
s
f
er
lear
n
i
n
g
b
ased
en
d
-
to
-
e
n
d
s
o
cc
er
v
id
eo
s
ce
n
e
an
d
ev
e
n
t
clas
s
i
f
i
ca
tio
n
.
Scen
e
clas
s
i
f
icatio
n
in
clu
d
es
lo
n
g
v
ie
w
a
n
d
clo
s
e
-
u
p
v
ie
w
w
h
ile
e
v
en
t
class
i
f
icatio
n
i
n
clu
d
e
s
co
r
n
er
,
f
r
ee
-
k
ic
k
,
g
o
al
an
d
p
en
al
t
y
.
T
h
eir
class
i
f
icatio
n
m
o
d
el
u
s
ed
C
NN
m
o
d
el
s
a
s
w
ell
a
s
tr
an
s
f
er
lear
n
in
g
m
et
h
o
d
.
W
h
er
ea
s
,
Yu
et
al.
[
4
0
]
d
esig
n
ed
a
d
ee
p
lear
n
in
g
-
b
ased
s
o
c
ce
r
ev
en
t
d
etec
tio
n
th
at
i
n
clu
d
es
e
v
en
t
d
etec
tio
n
as
w
el
l
as
s
to
r
y
g
e
n
er
atio
n
r
ef
er
to
Fig
u
r
e
3
w
h
ic
h
b
eg
in
s
w
it
h
ev
e
n
t
clip
s
an
d
en
d
w
it
h
r
ep
la
y
ev
e
n
t.
I
n
th
is
d
es
ig
n
C
NN
an
d
L
ST
M
w
er
e
u
s
ed
.
B
esid
es,
in
ten
n
is
s
p
o
r
t,
Mo
r
a
et
al.
[
4
1
]
h
av
e
p
r
o
p
o
s
ed
a
d
o
m
ai
n
-
s
p
ec
if
ic
d
ee
p
lear
n
in
g
ac
tio
n
r
ec
o
g
n
itio
n
m
et
h
o
d
b
y
u
tili
s
i
n
g
p
r
e
-
tr
ai
n
ed
C
NN
w
it
h
th
r
ee
-
la
y
er
ed
L
ST
M
m
o
d
el.
T
h
is
p
ap
er
w
as
ai
m
to
r
ec
o
g
n
is
ed
f
i
n
e
-
g
r
ai
n
ed
ac
tio
n
in
ten
n
i
s
s
p
o
r
ts
.
T
h
e
d
ee
p
L
ST
M
n
et
w
o
r
k
u
s
ed
in
t
h
is
r
esea
r
ch
is
ab
le
to
lear
n
h
ig
h
-
le
v
el
s
tr
u
ct
u
r
es
a
n
d
p
r
o
v
id
es
h
i
g
h
ac
cu
r
ac
y
.
A
p
ar
t
f
r
o
m
t
h
at,
I
b
r
ah
i
m
et
a
l.
[
4
2
]
in
n
o
v
ated
a
d
ee
p
m
o
d
el
to
ex
tr
ac
t
d
y
n
a
m
ic
te
m
p
o
r
al
i
n
f
o
r
m
atio
n
in
v
o
lle
y
b
a
ll
u
s
i
n
g
L
ST
M
m
o
d
els.
T
h
e
g
a
m
e
s
t
ate
w
as
d
ed
u
ce
d
b
y
ca
p
tu
r
in
g
p
la
y
er
s
’
s
tate
t
h
r
o
u
g
h
t
h
e
t
w
o
-
s
tag
e
L
ST
M
m
o
d
el.
R
am
a
n
at
h
a
n
et
al.
[
4
3
]
estab
lis
h
ed
a
d
ee
p
lean
i
n
g
m
eth
o
d
to
d
etec
t
a
n
d
ca
teg
o
r
is
ed
b
as
k
etb
all
e
v
e
n
ts
u
s
i
n
g
R
NN.
T
h
is
w
o
r
k
u
s
ed
a
n
atte
n
tio
n
m
o
d
el
to
d
etec
t
k
e
y
p
la
y
er
s
f
r
o
m
m
u
lti
-
p
er
s
o
n
v
id
eo
s
f
ir
s
t.
T
h
en
u
s
ed
C
NN
a
n
d
L
ST
M
to
ex
tr
a
ct
f
e
atu
r
e
an
d
d
etec
t
an
ev
en
t i
n
b
as
k
etb
all.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
1
8
,
No
.
4
,
A
u
g
u
s
t
2020:
1
9
2
6
-
1
9
3
3
1930
Fig
u
r
e
3
.
P
r
o
p
o
s
ed
s
o
cc
e
r
s
to
r
y
d
etec
tio
n
f
r
a
m
e
w
o
r
k
[
4
0
]
4.
DIS
CU
S
SI
O
N
Fo
r
th
e
p
as
t
f
e
w
d
ec
ad
es,
t
h
er
e
w
as
q
u
ite
a
lo
t
o
f
r
esear
ch
f
o
cu
s
i
n
g
o
n
s
p
o
r
t
v
id
eo
an
a
l
y
s
i
s
b
ased
o
n
h
an
d
cr
af
ted
ar
ch
i
tectu
r
es.
H
o
w
e
v
er
,
th
is
h
a
n
d
cr
af
ted
ar
c
h
itect
u
r
e
w
h
ic
h
co
n
s
i
s
ts
o
f
f
ea
t
u
r
e
d
escr
ip
to
r
s
an
d
ex
tr
ac
to
r
s
ar
e
p
r
o
b
lem
-
d
ep
en
d
en
t.
I
n
o
th
er
w
o
r
d
s
,
th
o
s
e
h
an
d
cr
af
ted
f
ea
t
u
r
es
ca
n
o
n
l
y
ap
p
l
y
to
s
p
ec
if
i
c
p
r
o
b
lem
s
.
I
t
co
u
ld
n
o
t
b
e
u
s
e
d
f
o
r
o
th
er
d
atasets
o
r
p
r
o
b
lem
s
w
h
ic
h
is
ir
r
ele
v
an
t
to
ap
p
ly
in
a
r
ea
l
s
ce
n
ar
io
.
A
p
ar
t
f
r
o
m
th
a
t,
th
e
tr
ad
itio
n
al
h
a
n
d
cr
af
ted
ar
ch
itect
u
r
e
u
s
ed
f
o
r
s
p
o
r
t
v
id
eo
an
al
y
s
is
co
u
ld
o
n
l
y
e
x
tr
ac
t
lo
w
-
le
v
el
f
ea
t
u
r
es,
it
is
v
er
y
ch
a
llen
g
i
n
g
i
n
ca
p
tu
r
i
n
g
f
o
r
h
i
g
h
-
lev
el
s
e
m
a
n
tic
in
f
o
r
m
atio
n
[
4
4
]
.
T
h
u
s
,
t
h
e
ad
v
an
ce
m
e
n
t
in
tec
h
n
o
lo
g
y
lead
s
to
th
e
e
m
er
g
e
n
ce
o
f
d
ee
p
lear
n
in
g
-
b
ased
s
p
o
r
t
v
id
eo
an
al
y
s
i
s
.
Sin
ce
d
ee
p
lear
n
i
n
g
-
b
ased
s
p
o
r
t
v
id
eo
an
al
y
s
is
is
s
till
a
n
e
w
a
n
d
g
r
o
w
in
g
r
esear
c
h
f
iel
d
,
th
er
e
w
er
e
o
n
l
y
a
f
e
w
s
t
u
d
ies
f
o
u
n
d
.
Mo
s
t
o
f
th
e
s
t
u
d
ies
ar
e
f
o
cu
s
i
n
g
o
n
s
o
cc
er
g
a
m
es,
ten
n
is
,
b
aseb
all
a
n
d
b
as
k
etb
all
[
4
5
]
.
On
l
y
l
i
m
ited
r
eso
u
r
ce
s
ar
e
f
o
u
n
d
o
n
th
e
o
th
er
s
p
o
r
ts
s
u
ch
as
h
o
ck
e
y
a
n
d
b
ad
m
i
n
to
n
.
Ho
w
e
v
er
,
th
e
m
ai
n
d
r
a
w
b
ac
k
o
f
u
s
i
n
g
d
ee
p
lear
n
in
g
m
o
d
els
ar
e
its
d
ata
-
h
u
n
g
r
y
m
ea
n
i
n
g
to
lear
n
f
ea
tu
r
es
au
to
m
at
icall
y
f
r
o
m
r
a
w
in
p
u
ts
d
ata
it
is
d
ep
en
d
s
o
n
th
o
u
d
s
a
n
d
s
o
f
in
p
u
t
d
ata
[
4
6
]
.
B
esid
es,
to
elev
ate
th
e
p
er
f
o
r
m
a
n
ce
lev
el
o
f
th
e
d
ee
p
lear
n
in
g
m
o
d
el
it
n
ee
d
s
h
i
g
h
-
p
er
f
o
r
m
an
ce
GP
U
s
[
4
7
]
.
B
u
t,
th
e
r
ec
en
t
ad
v
an
c
e
m
en
t
in
tech
n
o
lo
g
y
an
d
th
e
g
r
o
w
t
h
i
n
b
ig
d
ata
h
a
v
e
o
v
er
co
m
e
t
h
ese
i
s
s
u
es.
P
r
ev
io
u
s
l
y
,
C
NN
m
o
d
el
w
h
ic
h
is
o
n
e
o
f
t
h
e
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
h
as
s
h
o
w
n
tr
e
m
e
n
d
o
u
s
s
u
cc
es
s
in
i
m
ag
e
r
ec
o
g
n
itio
n
,
s
p
ee
ch
r
ec
o
g
n
itio
n
au
d
io
r
ec
o
g
n
i
tio
n
,
etc
[
4
8
]
.
Ho
w
e
v
er
,
in
t
h
e
an
a
l
y
s
i
s
o
f
v
id
eo
in
p
u
t
d
ata
r
esear
ch
er
s
f
ac
e
m
a
n
y
c
h
allen
g
es
b
ec
au
s
e
v
id
eo
s
eq
u
e
n
ce
s
d
y
n
a
m
icall
y
e
v
o
l
v
e
w
ith
ti
m
e.
I
t
is
d
if
f
icu
lt
to
e
x
tr
ac
t
te
m
p
o
r
al
i
n
f
o
r
m
ati
o
n
.
W
ith
th
e
co
n
t
in
u
o
u
s
s
tu
d
y
i
n
v
id
eo
a
n
al
y
s
is
u
s
i
n
g
d
ee
p
lear
n
in
g
,
ca
u
s
e
th
e
estab
li
s
h
m
e
n
t
o
f
a
s
eq
u
e
n
t
ial
m
o
d
els
s
u
ch
a
s
R
NN
a
n
d
L
ST
M.
T
h
ese
m
o
d
els
ar
e
ab
le
to
ex
tr
ac
t
te
m
p
o
r
al
in
f
o
r
m
atio
n
i
n
v
id
eo
in
p
u
t
d
ata.
T
h
er
e
w
er
e
s
o
m
e
r
ese
ar
ch
es
w
o
r
k
o
n
t
h
e
co
m
b
in
atio
n
o
f
b
o
th
C
N
N
an
d
L
ST
M
m
o
d
el
to
e
x
tr
ac
t
s
p
atio
-
te
m
p
o
r
al
i
n
f
o
r
m
atio
n
.
B
u
t
o
n
l
y
a
f
e
w
r
e
s
ea
r
ch
es
w
er
e
f
o
u
n
d
i
n
e
x
tr
ac
tio
n
h
ig
h
-
le
v
el
s
e
m
a
n
tic
i
n
f
o
r
m
ati
o
n
in
s
p
o
r
t
v
id
eo
an
al
y
s
is
.
D
esp
ite
asto
n
is
h
i
n
g
p
er
f
o
r
m
an
c
e
o
f
d
ee
p
lear
n
i
n
g
b
ased
ar
ch
ictu
r
e,
t
h
e
ad
v
a
n
ce
m
en
t
ac
h
iev
e
s
i
n
i
m
ag
e
clas
s
i
f
icatio
n
h
a
v
e
n
o
t
b
ee
n
r
ea
c
h
ed
in
ce
r
tai
n
f
ield
li
k
e
v
id
eo
class
i
f
icat
io
n
o
r
s
p
o
r
t
v
id
eo
an
al
y
s
is
[
4
9
]
.
I
t
is
s
till
a
n
o
p
en
is
s
u
e
in
d
ee
p
lear
n
i
n
g
-
b
ased
r
esear
ch
i
n
w
h
ic
h
m
a
n
y
r
e
s
ea
r
ch
er
s
tr
y
to
s
o
lv
e
an
d
it is
a
n
o
n
g
o
in
g
r
es
ea
r
ch
w
o
r
k
[
5
0
]
.
5.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
co
n
tr
ib
u
tes
a
co
m
p
r
e
h
e
n
s
i
v
e
s
u
r
v
e
y
o
n
s
p
o
r
t
v
id
eo
an
al
y
s
is
b
y
co
m
p
a
r
in
g
b
o
th
h
an
d
cr
af
ted
an
d
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
.
I
n
s
u
m
m
ar
y
,
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
h
as
o
v
er
co
m
e
t
h
e
li
m
itatio
n
s
en
c
o
u
n
ter
ed
b
y
tr
ad
itio
n
al
m
eth
o
d
s
in
ac
ti
v
it
y
r
ec
o
g
n
itio
n
o
f
s
p
o
r
t
v
id
eo
an
al
y
s
is
.
Ho
w
e
v
er
,
o
n
l
y
a
f
e
w
r
esear
ch
h
as
f
o
cu
s
ed
o
n
s
p
o
r
t
v
id
eo
an
al
y
s
is
.
So
,
in
f
u
t
u
r
e
s
tu
d
ies,
t
h
e
r
esear
ch
er
s
ca
n
f
o
cu
s
o
n
ex
tr
ac
tio
n
h
ig
h
-
le
v
el
s
e
m
an
tic
in
f
o
r
m
ati
o
n
in
s
p
o
r
t
an
al
y
s
is
w
h
ich
w
i
ll
b
e
u
s
ed
b
y
co
ac
h
es
an
d
s
p
o
r
ts
p
r
o
f
ess
io
n
al
s
i
n
ev
alu
a
tin
g
p
la
y
er
s
’
tactica
l
p
er
f
o
r
m
an
ce
i
n
t
h
e
g
a
m
e.
Mo
r
eo
v
er
,
f
u
t
u
r
e
r
esear
c
h
s
h
o
u
ld
al
s
o
co
n
ce
n
tr
ate
m
o
r
e
o
n
s
p
o
r
t
v
id
eo
s
th
at
ar
e
ap
ar
t
f
r
o
m
s
o
cc
er
g
a
m
e
s
,
te
n
n
i
s
,
b
aseb
all
an
d
b
as
k
etb
all
as
al
m
o
s
t
8
0
%
o
f
p
r
io
r
r
esear
ch
h
ad
b
ee
n
f
o
c
u
s
ed
o
n
th
o
s
e
s
p
o
r
ts
.
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
e
au
th
o
r
s
w
o
u
ld
lik
e
to
ex
p
r
ess
th
e
ir
ap
p
r
ec
iatio
n
to
Un
i
v
er
s
iti
T
ek
n
o
lo
g
i
Ma
la
y
s
ia
(
UT
M
)
f
o
r
en
d
o
w
th
i
s
r
esear
c
h
a
n
d
t
h
e
Min
is
ter
o
f
Hig
h
er
E
d
u
ca
tio
n
(
MO
HE
)
,
Ma
la
y
s
ia
f
o
r
s
u
p
p
o
r
tin
g
th
is
r
e
s
ea
r
c
h
w
o
r
k
u
n
d
er
Z
a
m
a
lah
Sc
h
o
lar
s
h
ip
an
d
R
e
s
ea
r
ch
Gr
an
t
No
.
Q.
J
1
3
0
0
0
0
.
2
5
4
5
.
1
9
H8
8
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
Dee
p
lea
r
n
in
g
in
s
p
o
r
t v
id
eo
a
n
a
lysi
s
:
a
r
ev
iew
(
K
ee
r
th
a
n
a
R
a
n
g
a
s
a
my
)
1931
RE
F
E
R
E
NC
E
S
[1
]
L
e
x
ico
,
“
sp
o
rt
|
De
f
in
it
io
n
o
f
sp
o
rt
in
En
g
li
s
h
b
y
Ox
f
o
rd
Dic
ti
o
n
a
ries
,
”
[
On
l
in
e
].
A
v
a
il
a
b
le:
h
tt
p
s:/
/en
.
o
x
f
o
rd
d
ictio
n
a
ries
.
c
o
m
/
d
e
f
in
it
io
n
/sp
o
r
t.
A
c
c
e
ss
e
d
:
3
1
Ja
n
u
a
ry
2019.
[2
]
D.
G
u
,
“
A
n
a
l
y
sis
o
f
tac
ti
c
a
l
in
f
o
r
m
a
ti
o
n
c
o
ll
e
c
ti
o
n
in
s
p
o
rts
c
o
m
p
e
ti
ti
o
n
b
a
se
d
o
n
t
h
e
in
telli
g
e
n
t
p
r
o
m
p
t
a
u
to
m
a
ti
c
c
o
m
p
letio
n
a
lg
o
rit
h
m
,
”
J
o
u
rn
a
l
o
f
In
telli
g
e
n
t
a
n
d
F
u
zz
y
S
y
ste
ms
,
v
o
l.
3
5
,
n
o
.
3
,
p
p
.
2
9
2
7
-
2
9
3
6
,
2
0
1
8
.
[3
]
N.
Ho
m
a
y
o
u
n
f
a
r,
S
.
F
i
d
ler,
a
n
d
R.
Urta
su
n
,
“
S
p
o
r
ts
f
ield
l
o
c
a
li
z
a
ti
o
n
v
ia
d
e
e
p
stru
c
t
u
re
d
m
o
d
e
ls,”
2
0
1
7
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
m
p
u
ter
Vi
si
o
n
a
n
d
Pa
tt
e
rn
Rec
o
g
n
it
io
n
(
CVP
R)
,
v
o
l.
2
0
1
7
-
Ja
n
u
a
ry
,
p
p
.
4
0
1
2
–
4
0
2
0
,
2
0
1
7
.
[4
]
M
.
S
tei
n
e
t
a
l.
,
“
Bri
n
g
It
t
o
t
h
e
P
i
tch
:
C
o
m
b
in
in
g
V
id
e
o
a
n
d
M
o
v
e
m
e
n
t
Da
ta
to
En
h
a
n
c
e
T
e
a
m
S
p
o
rt
A
n
a
ly
sis,”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Vi
su
a
li
za
t
i
o
n
a
n
d
Co
m
p
u
ter
Gr
a
p
h
ics
,
v
o
l.
2
4
,
n
o
.
1
,
p
p
.
1
3
–
2
2
,
2
0
1
8
.
[5
]
E.
E.
Cu
st,
A
.
J.
S
w
e
e
ti
n
g
,
K.
B
a
ll
,
a
n
d
S
.
Ro
b
e
rtso
n
,
“
M
a
c
h
in
e
a
n
d
d
e
e
p
lea
rn
in
g
f
o
r
sp
o
rt
-
sp
e
c
i
f
ic
m
o
v
e
m
e
n
t
re
c
o
g
n
it
io
n
:
a
s
y
ste
m
a
ti
c
re
v
ie
w
o
f
m
o
d
e
l
d
e
v
e
lo
p
m
e
n
t
a
n
d
p
e
rf
o
rm
a
n
c
e
,
”
J
o
u
rn
a
l
o
f
S
p
o
rts
S
c
i
e
n
c
e
s
,
v
o
l.
3
7
,
n
o
.
5
,
p
p
.
5
6
8
–
6
0
0
,
2
0
1
9
.
[6
]
N.
A
.
R
a
h
m
a
d
,
M
.
A
.
As
’a
ri,
N.
F
.
G
h
a
z
a
li
,
N.
S
h
a
h
a
r,
a
n
d
N.
A
.
J.
S
u
f
ri,
“
A
su
rv
e
y
o
f
v
i
d
e
o
b
a
se
d
a
c
ti
o
n
re
c
o
g
n
it
io
n
in
s
p
o
rts
,
”
In
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
in
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
,
v
o
l.
1
1
,
n
o
.
3
,
p
p
.
9
8
7
–
9
9
3
,
2
0
1
8
.
[7
]
G
.
Zh
u
e
t
a
l.
,
“
Ev
e
n
t
tac
ti
c
a
n
a
ly
sis
b
a
se
d
o
n
b
r
o
a
d
c
a
st
sp
o
r
ts
v
id
e
o
,
”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
M
u
lt
i
me
d
ia
,
v
o
l.
1
1
,
n
o
.
1
,
p
p
.
4
9
–
6
7
,
2
0
0
9
.
[8
]
A
.
Ka
r,
N.
Ra
i,
K.
S
ik
k
a
,
a
n
d
G
.
S
h
a
rm
a
,
“
A
d
a
S
c
a
n
:
A
d
a
p
ti
v
e
sc
a
n
p
o
o
li
n
g
in
d
e
e
p
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s
f
o
r
h
u
m
a
n
a
c
ti
o
n
re
c
o
g
n
it
io
n
in
v
id
e
o
s,”
2
0
1
7
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
mp
u
ter
Vi
sio
n
a
n
d
Pa
tt
e
rn
Rec
o
g
n
it
i
o
n
(
CVP
R)
,
p
p
.
3
3
7
6
–
3
3
8
5
,
2
0
1
7
.
[9
]
B
.
M
e
n
g
,
X.
L
i
u
,
a
n
d
X.
W
a
n
g
,
“
H
u
m
a
n
b
o
d
y
a
c
t
i
o
n
re
c
o
g
n
i
t
i
o
n
b
a
s
e
d
o
n
q
u
a
t
e
r
n
i
o
n
s
p
a
t
i
a
l
-
tem
p
o
ra
l
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
tw
o
r
k
,
”
Y
i
Q
i
Y
i
B
i
a
o
X
u
e
B
a
o
/
C
h
i
n
e
s
e
J
.
S
c
i
.
I
n
s
t
r
u
m
.
,
v
o
l
.
3
8
,
n
o
.
1
1
,
p
p
.
2
6
4
3
–
2
6
5
0
,
2
0
1
7
.
[1
0
]
H.
Ya
n
g
e
t
a
l.
,
“
A
s
y
m
m
e
tri
c
3
D
Co
n
v
o
l
u
ti
o
n
a
l
Ne
u
ra
l
Ne
tw
o
rk
s
f
o
r
a
c
ti
o
n
re
c
o
g
n
it
i
o
n
,
”
Pa
t
ter
n
Rec
o
g
n
it
.
,
v
o
l.
8
5
,
p
p
.
1
–
1
2
,
2
0
1
9
.
[1
1
]
Z.
Hu
i
q
u
n
,
W
.
Hu
i,
a
n
d
W
.
X
ia
o
li
n
g
,
“
A
p
p
li
c
a
ti
o
n
re
se
a
rc
h
o
f
v
id
e
o
a
n
n
o
tati
o
n
in
s
p
o
rts
v
i
d
e
o
a
n
a
ly
sis,”
2011
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
F
u
t
u
re
Co
mp
u
ter
S
c
ie
n
c
e
a
n
d
E
d
u
c
a
ti
o
n
,
p
p
.
6
2
–
6
6
,
2
0
1
1
.
[1
2
]
S
.
Do
ll
á
r,
P
.
,
Ra
b
a
u
d
,
V.,
Co
tt
re
l
l,
G
.
a
n
d
Be
lo
n
g
ie,
“
G
o
in
g
d
e
e
p
e
r
in
to
a
c
ti
o
n
re
c
o
g
n
it
io
n
:
A
su
rv
e
y
,
”
Ima
g
e
a
n
d
Vi
sio
n
Co
mp
u
ti
n
g
,
v
o
l.
6
0
,
p
p
.
4
–
2
1
,
2
0
1
7
.
[1
3
]
H.
Ch
e
n
,
J.
C
h
e
n
,
R
.
Hu
,
C.
Ch
e
n
,
a
n
d
Z.
W
a
n
g
,
“
A
c
ti
o
n
re
c
o
g
n
it
io
n
w
it
h
tem
p
o
ra
l
sc
a
le
-
in
v
a
rian
t
d
e
e
p
lea
rn
i
n
g
f
ra
m
e
w
o
rk
,
”
Ch
in
a
C
o
mm
u
n
ic
a
ti
o
n
s
,
v
o
l
.
1
4
,
n
o
.
2
,
p
p
.
1
6
3
–
1
7
2
,
2
0
1
7
.
[1
4
]
X
.
P
e
n
g
,
C.
Zo
u
,
Y.
Qia
o
,
a
n
d
Q
.
P
e
n
g
,
“
A
c
ti
o
n
Re
c
o
g
n
it
io
n
w
it
h
S
tac
k
e
d
F
ish
e
r
V
e
c
to
rs,”
Eu
ro
p
e
a
n
Co
n
fer
e
n
c
e
o
n
C
o
mp
u
ter
Vi
sio
n
-
ECC
V
2
0
1
4
:
Co
m
p
u
ter
Vi
si
o
n
–
ECC
V
2
0
1
4
,
p
p
.
5
8
1
–
5
9
5
,
2
0
1
4
.
[1
5
]
Zh
e
n
z
h
o
n
g
L
a
n
,
M
in
g
L
in
,
Xu
a
n
c
h
o
n
g
L
i,
A
.
G
.
H
a
u
p
tm
a
n
n
,
a
n
d
B.
Ra
j,
“
Be
y
o
n
d
G
a
u
ss
ian
P
y
r
a
m
id
:
M
u
lt
i
-
sk
ip
F
e
a
tu
re
S
tac
k
in
g
f
o
r
a
c
ti
o
n
re
c
o
g
n
it
io
n
,
”
2
0
1
5
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
mp
u
ter
Vi
sio
n
a
n
d
P
a
tt
e
rn
Rec
o
g
n
it
io
n
(
CVP
R)
,
p
p
.
2
0
4
–
2
1
2
,
2
0
1
5
.
[1
6
]
N.
Da
lal,
B.
T
ri
g
g
s,
a
n
d
C.
S
c
h
m
id
,
“
Hu
m
a
n
De
t
e
c
ti
o
n
Us
in
g
Orie
n
ted
Histo
g
ra
m
s
o
f
F
lo
w
a
n
d
A
p
p
e
a
ra
n
c
e
,
”
Eu
ro
p
e
a
n
C
o
n
fer
e
n
c
e
o
n
C
o
mp
u
t
e
r V
isio
n
(
ECCV
’0
6
)
,
p
p
.
4
2
8
–
4
4
1
,
2
0
0
6
.
[1
7
]
N
.
D
a
la
l
,
B
.
T
r
ig
g
s
,
“
H
i
s
t
o
g
r
a
m
s
o
f
o
r
ie
n
t
e
d
g
r
a
d
i
e
n
t
s
f
o
r
h
u
m
a
n
d
e
te
c
t
i
o
n
,
”
2
0
0
5
I
E
E
E
C
o
m
p
u
t
e
r
S
o
c
i
e
t
y
C
o
n
f
e
r
e
n
c
e
on
C
o
m
p
u
t
e
r
V
i
s
i
o
n
a
n
d
P
a
t
t
e
r
n
R
e
c
o
g
n
i
t
i
o
n
(
C
V
P
R
'0
5
)
,
2
0
0
5
.
[1
8
]
P
.
D
o
l
l
á
r
,
V.
R
a
b
a
u
d
,
G
.
C
o
t
t
r
e
l
l
,
S
.
B
e
l
o
n
g
i
e
,
“
Be
h
a
v
i
o
r
r
e
c
o
g
n
i
t
i
o
n
v
i
a
s
p
a
rs
e
s
p
a
t
i
o
-
t
e
m
p
o
r
a
l
f
e
a
t
u
r
e
s
,
”
2
0
0
5
I
E
E
E
I
n
t
e
r
n
a
t
i
o
n
a
l
W
o
r
k
s
h
o
p
o
n
V
i
s
u
a
l
S
u
r
v
e
i
l
l
a
n
c
e
a
n
d
P
e
r
f
o
r
m
a
n
c
e
E
v
a
l
u
a
t
i
o
n
o
f
T
r
a
c
k
i
n
g
a
n
d
S
u
r
v
e
i
l
l
a
n
c
e
,
2
0
0
5
.
[1
9
]
A
.
A
b
d
u
lm
u
n
e
m
,
Y.
K.
L
a
i,
a
n
d
X.
S
u
n
,
“
S
a
li
e
n
c
y
g
u
id
e
d
lo
c
a
l
a
n
d
g
lo
b
a
l
d
e
sc
rip
to
rs
f
o
r
e
ff
e
c
ti
v
e
a
c
ti
o
n
re
c
o
g
n
it
io
n
,
”
Co
mp
u
t
a
ti
o
n
a
l
Vi
su
a
l
M
e
d
ia
,
v
o
l.
2
,
p
p
.
9
7
–
1
0
6
,
2
0
1
6
.
[2
0
]
M
.
A
.
Ca
rb
o
n
n
e
a
u
,
A
.
J.
Ra
y
m
o
n
d
,
E.
G
ra
n
g
e
r,
a
n
d
G
.
G
a
g
n
o
n
,
“
Re
a
l
-
ti
m
e
v
isu
a
l
p
la
y
-
b
re
a
k
d
e
t
e
c
ti
o
n
in
sp
o
rt
e
v
e
n
ts
u
sin
g
a
c
o
n
tex
t
d
e
sc
rip
t
o
r,
”
2
0
1
5
IEE
E
In
ter
n
a
ti
o
n
a
l
S
y
mp
o
si
u
m
o
n
Circ
u
it
s
a
n
d
S
y
ste
ms
(
IS
CAS
)
,
p
p
.
2
8
0
8
–
2
8
1
1
,
2
0
1
5
.
[2
1
]
J.
L
e
e
,
D.
Na
m
,
S
.
M
o
o
n
,
J.
L
e
e
,
a
n
d
W
.
Yo
o
,
“
S
o
c
c
e
r
Ev
e
n
t
Re
c
o
g
n
it
io
n
T
e
c
h
n
iq
u
e
b
a
se
d
o
n
P
a
tt
e
rn
M
a
tch
i
n
g
,
”
2
0
1
7
Fe
d
e
ra
ted
C
o
n
fer
e
n
c
e
o
n
C
o
mp
u
ter
S
c
ien
c
e
a
n
d
I
n
fo
rm
a
ti
o
n
S
y
ste
ms
(
Fed
CS
IS
),
p
p
.
6
4
3
–
6
4
6
,
2
0
1
7
.
[2
2
]
C.
C.
L
ien
,
C.
L
.
Ch
ian
g
,
a
n
d
C.
H.
Lee
,
“
S
c
e
n
e
-
b
a
se
d
e
v
e
n
t
d
e
tec
ti
o
n
f
o
r
b
a
se
b
a
ll
v
id
e
o
s,”
J
o
u
rn
a
l
o
f
Vi
su
a
l
Co
mm
u
n
ica
ti
o
n
a
n
d
Ima
g
e
Rep
re
se
n
ta
ti
o
n
,
v
o
l.
1
8
,
n
o
.
1
,
p
p
.
1
–
1
4
,
2
0
0
7
.
[2
3
]
L
.
H.
Ch
e
n
,
C.
W
.
S
u
,
a
n
d
H.
A
.
Hs
iao
,
“
P
lay
e
r
traje
c
to
r
y
r
e
c
o
n
stru
c
ti
o
n
f
o
r
tac
ti
c
a
l
a
n
a
l
y
sis,”
M
u
lt
ime
d
ia
T
o
o
l
s
a
n
d
Ap
p
li
c
a
ti
o
n
s
,
v
o
l.
7
7
,
p
p
.
3
0
4
7
5
–
3
0
4
8
6
,
2
0
1
8
.
[2
4
]
W
.
L
.
L
u
,
K.
Ok
u
m
a
,
a
n
d
J.
J.
L
it
tl
e
,
“
T
ra
c
k
in
g
a
n
d
re
c
o
g
n
izin
g
a
c
ti
o
n
s
o
f
m
u
lt
ip
le
h
o
c
k
e
y
p
la
y
e
rs
u
sin
g
th
e
b
o
o
ste
d
p
a
rti
c
le f
il
ter,”
Ima
g
e
a
n
d
Vi
si
o
n
C
o
mp
u
ti
n
g
,
v
o
l.
2
7
,
n
o
.
1
,
p
p
.
1
8
9
–
2
0
5
,
2
0
0
9
.
[2
5
]
S
.
M
u
k
h
e
rjee
a
n
d
K.
Ka
ra
n
,
“
H
u
m
a
n
a
c
ti
o
n
a
n
d
e
v
e
n
t
re
c
o
g
n
it
io
n
u
si
n
g
a
n
o
v
e
l
d
e
sc
rip
to
r
b
a
se
d
o
n
im
p
ro
v
e
d
d
e
n
se
traje
c
to
ries
,
”
M
u
lt
ime
d
i
a
T
o
o
ls
a
n
d
A
p
p
li
c
a
ti
o
n
s v
o
l
u
me
,
v
o
l
.
7
7
,
p
p
.
1
3
6
6
1
–
1
3
6
7
8
,
2
0
1
8
.
[2
6
]
A
.
V
o
u
lo
d
im
o
s,
N.
Do
u
la
m
is,
A
.
Do
u
lam
is,
a
n
d
E.
P
r
o
to
p
a
p
a
d
a
k
i
s,
“
De
e
p
Lea
rn
in
g
f
o
r
Co
m
p
u
ter
V
isio
n
:
A
Brie
f
Re
v
ie
w
,
”
Co
mp
u
ta
ti
o
n
a
l
I
n
telli
g
e
n
c
e
a
n
d
Ne
u
ro
sc
ien
c
e
,
v
o
l
.
2
0
1
8
,
p
p
.
1
-
13
,
2
0
1
8
.
[2
7
]
A
.
S
a
r
g
a
n
o
,
P
.
A
n
g
e
lo
v
,
a
n
d
Z.
Ha
b
ib
,
“
A
Co
m
p
re
h
e
n
siv
e
R
e
v
i
e
w
o
n
Ha
n
d
c
ra
f
ted
a
n
d
Lea
rn
in
g
-
Ba
se
d
Ac
ti
o
n
Re
p
re
se
n
tatio
n
A
p
p
r
o
a
c
h
e
s f
o
r
Hu
m
a
n
A
c
ti
v
it
y
Re
c
o
g
n
it
io
n
,
”
Ap
p
l
li
e
d
S
c
i
e
n
c
e
,
v
o
l.
7
,
n
o
.
1
,
p
p
.
1
1
0
,
2
0
1
7
.
[2
8
]
A
.
El
b
o
u
sh
a
k
i,
R.
Ha
n
n
a
n
e
,
K.
Af
d
e
l,
a
n
d
L
.
Ko
u
tt
i,
“
M
u
lt
i
D
-
CNN
:
A
m
u
lt
i
-
d
im
e
n
sio
n
a
l
fe
a
tu
re
lea
rn
in
g
a
p
p
ro
a
c
h
b
a
se
d
o
n
d
e
e
p
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
tw
o
rk
s
f
o
r
g
e
stu
re
r
e
c
o
g
n
it
io
n
in
RG
B
-
D
i
m
a
g
e
se
q
u
e
n
c
e
s,”
Exp
e
rt
S
y
ste
ms
wit
h
Ap
p
li
c
a
ti
o
n
s
,
v
o
l
.
1
3
9
,
p
p
.
1
1
2
8
2
9
,
2
0
2
0
.
[2
9
]
B
.
M
e
n
g
,
X.
L
i
u
,
a
n
d
X.
W
a
n
g
,
“
H
u
m
a
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
b
a
s
e
d
o
n
q
u
a
t
e
r
n
i
o
n
s
p
a
t
i
a
l
-
t
e
m
p
o
r
a
l
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
rk
a
n
d
L
S
T
M
i
n
RG
B
v
i
d
e
o
s
,
”
M
u
l
t
i
me
d
i
a
T
o
o
l
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
38
,
n
o
.
11
,
p
p
.
2
6
9
0
1
-
2
6
9
1
8
,
2
0
1
8
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
1
8
,
No
.
4
,
A
u
g
u
s
t
2020:
1
9
2
6
-
1
9
3
3
1932
[3
0
]
M
.
A
s
a
d
i
-
a
g
h
b
o
l
a
g
h
i
e
t
a
l
.
,
“
A
S
u
r
v
e
y
o
n
D
e
e
p
L
e
a
r
n
i
n
g
B
a
s
e
d
A
p
p
r
o
a
c
h
e
s
f
o
r
A
c
t
i
o
n
a
n
d
G
e
s
t
u
r
e
R
e
c
o
g
n
i
t
i
o
n
i
n
I
m
a
g
e
S
e
q
u
e
n
c
e
s
,
”
F
G
2
0
1
7
-
1
2
t
h
I
E
E
E
C
o
n
f
e
r
e
n
c
e
o
n
A
u
t
o
m
a
t
i
c
F
a
c
e
a
n
d
G
e
s
t
u
r
e
R
e
c
o
g
n
i
t
i
o
n
,
p
p
.
4
7
6
–
483
,
2
0
1
7
.
[3
1
]
X.
Y
a
n
g
,
P
.
M
o
l
c
h
a
n
o
v
,
a
n
d
J
.
K
a
u
t
z
,
“
M
u
l
t
i
l
a
y
e
r
a
n
d
M
u
l
t
i
m
o
d
a
l
F
u
s
i
o
n
o
f
D
e
e
p
N
e
u
r
a
l
N
e
tw
o
rk
s
f
o
r
Vi
d
e
o
C
l
a
s
s
if
i
c
a
t
i
o
n
,
”
M
M
'1
6
:
P
r
o
c
e
e
d
in
g
s
o
f
t
h
e
2
4
t
h
A
C
M
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
M
u
l
t
i
m
e
d
i
a
,
p
p
.
9
7
8
–
987
,
2
0
1
6
.
[3
2
]
J.
Y.
H.
Ng
,
M
.
Ha
u
sk
n
e
c
h
t
,
S
.
V
ij
a
y
a
n
a
ra
si
m
h
a
n
,
O.
V
in
y
a
ls,
R.
M
o
n
g
a
,
a
n
d
G
.
T
o
d
e
rici,
“
Be
y
o
n
d
S
h
o
rt
S
n
ip
p
e
ts:
De
e
p
Ne
tw
o
rk
s
f
o
r
V
id
e
o
Clas
sif
ica
ti
o
n
,
”
2
0
1
5
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
mp
u
ter
Vi
sio
n
a
n
d
Pa
t
ter
n
Rec
o
g
n
it
io
n
(
CVP
R)
,
2
0
1
5
.
[3
3
]
N.
A
.
Ra
h
m
a
d
,
N.
A
.
J.
S
u
f
ri,
N.
H.
M
u
z
a
m
il
,
a
n
d
M
.
A
.
A
s’
a
ri,
“
Ba
d
m
in
to
n
p
lay
e
r
d
e
tec
ti
o
n
u
si
n
g
f
a
ste
r
re
g
io
n
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
,
”
In
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
E
lec
trica
l
En
g
in
e
e
rin
g
a
n
d
Co
mp
u
ter
S
c
ien
c
e
,
v
o
l.
1
4
,
n
o
.
3
,
p
p
.
1
3
3
0
–
1
3
3
5
,
2
0
1
9
.
[3
4
]
M
.
R.
T
o
ra
a
n
d
J.
J.
L
it
tl
e
,
“
C
las
sif
ic
a
ti
o
n
o
f
P
u
c
k
P
o
ss
e
s
sio
n
Ev
e
n
ts
in
Ic
e
Ho
c
k
e
y
,
”
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
mp
u
ter
V
isio
n
a
n
d
Pa
t
ter
n
Rec
o
g
n
it
io
n
W
o
rk
sh
o
p
s (
CVP
RW
)
, p
p
.
1
4
7
-
1
5
4
,
2
0
1
7
.
[3
5
]
K.
S
o
z
y
k
in
,
S
.
P
ro
tas
o
v
,
A
.
Kh
a
n
,
R.
Hu
ss
a
in
,
a
n
d
J.
L
e
e
,
“
M
u
lt
i
-
lab
e
l
c
las
s
-
im
b
a
lan
c
e
d
a
c
ti
o
n
re
c
o
g
n
it
io
n
i
n
h
o
c
k
e
y
v
id
e
o
s v
i
a
3
D co
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s,”
a
rX
iv
:1
7
0
9
.
0
1
4
2
1
,
2
0
1
8
.
[3
6
]
L
.
K
o
n
g
,
D
.
H
u
a
n
g
,
J
.
Q
i
n
,
a
n
d
Y
.
W
a
n
g
,
“
A
J
o
i
n
t
F
r
a
m
e
w
o
r
k
f
o
r
A
t
h
l
e
t
e
T
ra
c
k
i
n
g
a
n
d
A
c
t
i
o
n
R
e
c
o
g
n
i
t
i
o
n
i
n
S
p
o
r
t
s
Vi
d
e
o
s
,
”
I
E
E
E
T
r
a
n
s
a
c
t
i
o
n
s
o
n
C
i
r
c
u
i
t
s
a
n
d
S
y
s
t
e
m
s
f
o
r
V
i
d
e
o
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
3
0
,
n
o
.
2
,
p
p
.
5
3
2
–
5
4
8
,
2
0
2
0
.
[3
7
]
J.
He
,
K.,
Zh
a
n
g
,
X
.
,
Re
n
,
S
.
a
n
d
S
u
n
,
“
S
p
a
ti
a
l
P
y
ra
m
id
P
o
o
l
i
n
g
in
De
e
p
Co
n
v
o
lu
t
io
n
a
l
Ne
two
rk
s
f
o
r
V
isu
a
l
Re
c
o
g
n
it
io
n
,
”
IEE
E
T
ra
n
s
.
Pa
tt
e
rn
An
a
l.
M
a
c
h
.
I
n
tell.
,
v
o
l.
3
7
,
n
o
.
9
,
p
p
.
1
9
0
4
–
1
9
1
6
,
2
0
1
5
.
[3
8
]
H.
Jia
n
g
,
Y.
L
u
,
a
n
d
J.
X
u
e
,
“
A
u
to
m
a
ti
c
so
c
c
e
r
v
id
e
o
e
v
e
n
t
d
e
tec
ti
o
n
b
a
se
d
o
n
a
d
e
e
p
n
e
u
ra
l
n
e
t
w
o
rk
c
o
m
b
in
e
d
CNN
a
n
d
RNN
,
”
2
0
1
6
IE
EE
2
8
t
h
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
T
o
o
ls
wit
h
Arti
f
icia
l
I
n
telli
g
e
n
c
e
(
ICT
AI)
,
p
p
.
4
9
0
–
4
9
4
,
2
0
1
7
.
[3
9
]
Y.
Ho
n
g
,
C.
L
in
g
,
a
n
d
Z.
Ye
,
“
En
d
-
to
-
e
n
d
so
c
c
e
r
v
id
e
o
sc
e
n
e
a
n
d
e
v
e
n
t
c
las
si
f
ica
ti
o
n
w
it
h
d
e
e
p
tra
n
sf
e
r
lea
rn
in
g
,
”
2
0
1
8
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
In
telli
g
e
n
t
S
y
ste
ms
a
n
d
Co
m
p
u
te
r V
isio
n
(
IS
CV)
,
p
p
.
1
–
4
,
2
0
1
8
.
[4
0
]
J.
Yu
,
A
.
L
e
i,
a
n
d
Y.
Hu
,
“
S
o
c
c
e
r
Vid
e
o
Ev
e
n
t
De
tec
ti
o
n
Ba
se
d
o
n
De
e
p
L
e
a
rn
in
g
,
”
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
M
u
lt
ime
d
ia
M
o
d
e
li
n
g
-
M
M
M
2
0
1
9
:
M
u
l
ti
M
e
d
ia
M
o
d
e
li
n
g
,
S
p
ri
n
g
e
r,
v
o
l.
8
9
3
6
,
p
p
.
3
7
7
-
3
8
9
,
2
0
1
9
.
[4
1
]
S
.
V
.
M
o
ra
a
n
d
W
.
J.
Kn
o
tt
e
n
b
e
lt
,
“
De
e
p
L
e
a
rn
in
g
f
o
r
Do
m
a
in
-
S
p
e
f
icis
Ac
ti
o
n
Re
c
o
g
n
it
io
n
i
n
T
e
n
n
is,”
2
0
1
7
IE
E
E
Co
n
fer
e
n
c
e
o
n
Co
m
p
u
ter
Vi
si
o
n
a
n
d
Pa
tt
e
rn
Rec
o
g
n
it
io
n
W
o
rk
sh
o
p
s (
CVP
RW
)
,
p
p
.
1
7
0
–
1
7
8
,
,
2
0
1
7
.
[4
2
]
M
.
G
.
I
b
r
a
h
i
m
M
S
,
M
u
r
a
l
i
d
h
a
r
a
n
S
,
D
e
n
g
Z
,
V
a
h
d
a
t
A
,
“
A
h
i
e
r
a
r
c
h
i
c
a
l
d
e
e
p
t
e
m
p
o
r
a
l
m
o
d
e
l
f
o
r
g
r
o
u
p
a
c
t
i
v
i
t
y
r
e
c
o
g
n
i
t
i
o
n
,
”
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
I
E
E
E
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
t
e
r
V
i
s
i
o
n
a
n
d
P
a
t
t
e
r
n
R
e
c
o
g
n
i
t
i
o
n
,
p
p
.
1
9
7
1
–
1980
,
2
0
1
6
.
[4
3
]
V
.
Ra
m
a
n
a
th
a
n
,
J.
Hu
a
n
g
,
S
.
A
b
u
-
el
-
h
a
ij
a
,
A
.
G
o
rb
a
n
,
K.
M
u
rp
h
y
,
a
n
d
L
.
F
e
i
-
f
e
i,
“
De
te
c
ti
n
g
e
v
e
n
ts
a
n
d
k
e
y
a
c
to
rs
in
m
u
lt
i
-
p
e
rso
n
v
id
e
o
s
,
”
2
0
1
6
I
E
EE
Co
n
fer
e
n
c
e
o
n
C
o
mp
u
ter
Vi
si
o
n
a
n
d
P
a
tt
e
rn
Rec
o
g
n
it
io
n
(
C
VP
R),
2
0
1
6
.
[4
4
]
J.
Lee
,
D.
W
.
Na
m
,
J.
S
.
L
e
e
,
S
.
M
o
o
n
,
K.
Ki
m
,
a
n
d
H.
Ki
m
,
“
A
stu
d
y
o
n
c
o
m
p
o
siti
o
n
o
f
c
o
n
tex
t
-
b
a
se
d
so
c
c
e
r
a
n
a
ly
sis
s
y
ste
m
,
”
2
0
1
7
1
9
t
h
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Ad
v
a
n
c
e
d
Co
mm
u
n
ica
ti
o
n
T
e
c
h
n
o
lo
g
y
(
ICACT
)
,
p
p
.
8
8
6
–
8
8
9
,
2
0
1
7
.
[4
5
]
H.
C.
S
h
ih
,
“
A
S
u
rv
e
y
o
f
Co
n
ten
t
-
Aw
a
r
e
V
id
e
o
A
n
a
l
y
sis
f
o
r
S
p
o
r
ts,”
IEE
E
T
r
a
n
sa
c
ti
o
n
s
o
n
Circ
u
i
ts
a
n
d
S
y
ste
ms
fo
r V
i
d
e
o
T
e
c
h
n
o
l
o
g
y
,
v
o
l.
2
8
,
n
o
.
5
,
p
p
.
1
2
1
2
–
1
2
3
1
,
2
0
1
8
.
[4
6
]
P
.
F
e
lse
n
,
P
.
A
g
ra
w
a
l,
a
n
d
J.
M
a
li
k
,
“
W
h
a
t
w
il
l
Ha
p
p
e
n
Ne
x
t?
F
o
re
c
a
stin
g
P
lay
e
r
M
o
v
e
s
in
S
p
o
r
ts
V
i
d
e
o
s,”
2
0
1
7
IEE
E
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
Co
m
p
u
ter
Vi
si
o
n
(
ICCV)
,
p
p
.
3
3
6
2
–
3
3
7
1
,
2
0
1
7
.
[4
7
]
Z.
X
u
,
Y.
Ya
n
g
,
a
n
d
A
.
G
.
Ha
u
p
tm
a
n
n
,
“
A
d
isc
ri
m
in
a
ti
v
e
CNN
v
id
e
o
re
p
re
se
n
tatio
n
f
o
r
e
v
e
n
t
d
e
tec
ti
o
n
,
”
2
0
1
5
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
mp
u
ter
V
isio
n
a
n
d
P
a
tt
e
rn
Rec
o
g
n
it
io
n
(
CVP
R)
,
p
p
.
1
7
9
8
–
1
8
0
7
,
2
0
1
5
.
[4
8
]
Y.
L
e
c
u
n
,
Y.
Be
n
g
io
,
a
n
d
G
.
Hin
t
o
n
,
“
De
e
p
lea
rn
i
n
g
,
”
Na
t
u
re
,
v
o
l.
5
2
1
,
n
o
.
7
5
5
3
,
p
p
.
4
3
6
–
4
4
4
,
2
0
1
5
.
[4
9
]
I.
Ro
d
ríg
u
e
z
-
M
o
re
n
o
,
J.
M
.
M
a
rt
ín
e
z
-
Otz
e
ta,
B.
S
ierra
,
I.
Ro
d
rig
u
e
z
,
a
n
d
E.
Ja
u
re
g
i,
“
V
i
d
e
o
a
c
ti
v
it
y
r
e
c
o
g
n
it
io
n
:
S
tate
-
of
-
th
e
-
a
rt,
”
S
e
n
so
rs
(
S
wit
ze
rla
n
d
)
,
v
o
l.
1
9
,
n
o
.
1
4
,
p
p
.
1
–
2
5
,
2
0
1
9
.
[5
0
]
A
.
Kh
a
n
,
A
.
S
o
h
a
il
,
U.
Zah
o
o
ra
,
a
n
d
A
.
S
.
Qu
re
sh
i,
“
A
S
u
rv
e
y
o
f
th
e
Re
c
e
n
t
A
rc
h
it
e
c
tu
re
s
o
f
De
e
p
Co
n
v
o
l
u
ti
o
n
a
l
Ne
u
ra
l
Ne
tw
o
rk
s,”
a
rX
iv
:1
9
0
1
.
0
6
0
3
2
,
p
p
.
1
–
6
2
,
2
0
1
9
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
K
e
e
r
th
a
n
a
Ra
n
g
a
sa
m
y
re
c
e
iv
e
d
B.
E
(Ho
n
s)
i
n
Bi
o
m
e
d
ica
l
En
g
in
e
e
rin
g
f
ro
m
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
a
lay
sia
(U
T
M
)
in
2
0
1
8
.
Cu
rre
n
t
ly
,
sh
e
p
u
rsu
i
n
g
h
e
r
P
h
D
i
n
Bio
m
e
d
ica
l
En
g
in
e
e
rin
g
u
n
d
e
r
S
c
h
o
o
l
o
f
Bio
m
e
d
ica
l
En
g
in
e
e
rin
g
a
n
d
He
a
lt
h
S
c
ien
c
e
,
F
a
c
u
lt
y
o
f
En
g
in
e
e
rin
g
in
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
a
lay
si
a
.
He
r
re
se
a
r
c
h
in
tere
sts
a
re
ar
ti
f
icial
in
teli
g
e
n
ce
,
co
m
p
u
ter
v
is
io
n
an
d
d
ee
p
lear
n
in
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
Dee
p
lea
r
n
in
g
in
s
p
o
r
t v
id
eo
a
n
a
lysi
s
:
a
r
ev
iew
(
K
ee
r
th
a
n
a
R
a
n
g
a
s
a
my
)
1933
Dr
.
M
u
h
a
m
m
a
d
A
m
ir
B
in
As
'
a
r
i
h
o
ld
s
a
P
h
D
i
n
Bio
m
e
d
ica
l
En
g
in
e
e
rin
g
f
ro
m
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
a
lay
sia
.
His
P
h
’s
w
o
rk
w
a
s
in
th
e
f
ield
o
f
a
ss
isti
v
e
tec
h
n
o
l
o
g
y
,
c
o
m
p
u
ter
v
isio
n
a
n
d
im
a
g
e
p
ro
c
e
ss
in
g
a
n
d
h
is
wo
rk
f
o
c
u
se
d
o
n
d
e
v
e
lo
p
in
g
a
n
o
v
e
l
3
D
sh
a
p
e
d
e
sc
rip
to
rs
f
o
r
re
c
o
g
n
izin
g
th
e
a
c
ti
v
it
ies
o
f
d
a
il
y
li
v
in
g
b
a
se
d
o
n
Kin
e
c
t
-
li
k
e
d
e
p
th
im
a
g
e
.
He
p
u
rsu
e
d
h
is
mas
ter
d
e
g
r
e
e
a
n
d
b
a
c
h
e
lo
r
d
e
g
r
e
e
a
t
F
a
c
u
lt
y
o
f
El
e
c
tri
c
a
l
En
g
in
e
e
rin
g
,
Un
iv
e
rsiti
Tek
n
o
lo
g
i
M
a
la
y
sia
,
m
a
jo
re
d
in
El
e
c
tri
c
a
l
En
g
in
e
e
rin
g
.
C
u
rre
c
tl
y
,
h
e
is
w
o
rk
in
g
o
n
th
e
d
e
v
e
lo
p
m
e
n
t
o
f
sig
n
a
l
a
n
d
im
a
g
e
p
ro
c
e
ss
in
g
a
p
p
ro
a
c
h
e
s
w
it
h
in
telli
g
e
n
t
in
terv
e
n
ti
o
n
i
n
a
s
sistiv
e
tec
h
n
o
lo
g
y
a
n
d
re
h
a
b
il
it
a
ti
o
n
a
s
w
e
ll
a
s
sp
o
rt
p
e
rf
o
rm
a
n
c
e
tec
h
n
o
lo
g
y
e
sp
e
c
iall
y
in
a
u
to
m
a
ted
h
u
m
a
n
m
o
v
e
m
e
n
t
re
c
o
g
n
it
io
n
Nur
Az
m
i
na
Ra
h
m
a
d
r
ec
eiv
ed
B
.
E
(
Ho
n
s
)
i
n
B
io
m
ed
ical
E
n
g
in
ee
r
i
n
g
f
r
o
m
Un
i
v
er
s
iti
T
ek
n
o
lo
g
i
Ma
la
y
s
i
a
(
UT
M)
in
2
0
1
7
.
C
u
r
r
en
tl
y
,
s
h
e
p
u
r
s
u
i
n
g
h
er
P
h
D
i
n
B
io
m
ed
ical
E
n
g
in
ee
r
i
n
g
u
n
d
er
Sch
o
o
l
o
f
B
io
m
ed
ical
E
n
g
i
n
ee
r
i
n
g
an
d
Hea
l
t
h
Scien
ce
,
Facu
l
t
y
o
f
E
n
g
in
ee
r
in
g
in
U
n
iv
er
s
iti
T
ek
n
o
lo
g
i
Ma
la
y
s
ia.
Her
r
esear
c
h
in
ter
est
s
ar
e
ar
tif
icial
i
n
teli
g
e
n
ce
,
co
m
p
u
t
er
v
is
io
n
an
d
d
ee
p
lear
n
in
g
.
Nurul
F
a
t
hia
h
G
ha
za
li
r
ec
eiv
ed
Dip
lo
m
a
in
E
lectr
ical
E
n
g
i
n
ee
r
i
n
g
(
P
o
w
er
)
f
r
o
m
Un
i
v
er
s
iti T
ek
n
o
lo
g
i M
ala
y
s
i
a,
Ku
ala
L
u
m
p
u
r
(
2
0
1
3
)
an
d
B
ac
h
elo
r
o
f
E
n
g
i
n
ee
r
in
g
i
n
B
io
m
ed
ical
f
r
o
m
Un
i
v
er
s
i
ti
T
ek
n
o
lo
g
i
Ma
la
y
s
ia,
J
o
h
o
r
B
ah
r
u
(
2
0
1
7
)
.
Sh
e
is
cu
r
r
en
tl
y
w
o
r
k
i
n
g
o
n
h
er
P
h
D
s
t
u
d
ies
i
n
B
io
m
ed
ical
E
n
g
i
n
ee
r
i
n
g
at
U
n
i
v
er
s
it
i
T
ek
n
o
lo
g
i
Ma
la
y
s
ia,
J
o
h
o
r
B
ah
r
u
an
d
h
er
cu
r
r
en
t
r
e
s
ea
r
ch
is
ab
o
u
t
ac
ti
v
it
y
r
ec
o
g
n
itio
n
s
tu
d
y
i
n
b
ad
m
in
t
o
n
s
p
o
r
t
u
s
i
n
g
i
n
er
tial
s
e
n
s
o
r
w
h
ic
h
c
o
llab
o
r
ated
w
it
h
Sp
o
r
t
I
n
n
o
v
atio
n
an
d
T
ec
h
n
o
lo
g
y
C
en
tr
e
(
SIT
C
)
an
d
I
n
s
ti
tu
te
o
f
H
u
m
an
C
e
n
ter
ed
E
n
g
i
n
ee
r
i
n
g
(
iHu
m
E
n
)
f
r
o
m
UT
M
J
B
.
Her
r
esear
ch
i
n
ter
est
in
cl
u
d
es
ac
ti
v
it
y
r
ec
o
g
n
itio
n
s
t
u
d
ies,
s
i
g
n
a
l p
r
o
ce
s
s
i
n
g
,
a
n
d
m
ac
h
in
e
lear
n
i
n
g
.
Sa
ha
rudin
I
s
m
a
il
g
r
ad
u
ated
f
r
o
m
B
ac
h
elo
r
o
f
Occ
u
p
ati
o
n
al
T
h
er
ap
y
(
Ho
n
o
u
r
s
)
,
UKM
in
2
0
0
8
.
He
p
u
r
s
u
ed
h
is
Ma
s
ter
o
f
P
h
ilo
s
o
p
h
y
i
n
R
eh
ab
ilit
atio
n
T
ec
h
n
o
lo
g
y
o
n
2
0
1
3
an
d
attain
ed
h
is
P
HD
in
Hea
lt
h
Scien
ce
o
n
2
0
1
9
f
r
o
m
U
n
iv
er
s
iti
T
ek
n
o
lo
g
i
Ma
la
y
s
ia.
Mo
s
t
o
f
h
is
w
o
r
k
i
s
r
e
l
a
te
d
t
o
A
s
s
is
ti
v
e
D
e
v
i
c
es
,
R
eh
a
b
i
li
t
at
i
o
n
,
B
i
o
s
t
a
ti
s
t
i
cs
,
M
o
t
o
r
C
o
n
t
r
o
l
,
M
o
t
o
r
L
e
a
r
n
in
g
an
d
K
in
es
i
o
l
o
g
y
th
r
o
u
g
h
th
e
a
p
p
l
i
ca
t
i
o
n
o
f
th
e
i
n
er
tial se
n
s
o
r
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.