I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
25
,
No
.
2
,
Feb
r
u
ar
y
2
0
2
2
,
p
p
.
8
92
~
8
99
I
SS
N:
2502
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
25
.i
2
.
pp
892
-
8
9
9
892
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Co
mpa
riso
n of f
e
ed f
o
rwa
rd
and c
a
sca
de f
o
rwa
rd n
eura
l
networks
for hum
a
n actio
n recog
nit
io
n
Aditi
J
a
ha
g
irda
r
,
Ra
s
hm
i P
ha
lni
k
a
r
S
c
h
o
o
l
o
f
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
a
n
d
Te
c
h
n
o
l
o
g
y
,
M
I
T
W
o
r
l
d
P
e
a
c
e
U
n
i
v
e
r
si
t
y
,
P
u
n
e
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l
6
,
2
0
2
1
R
ev
is
ed
No
v
30
,
2
0
2
1
Acc
ep
ted
Dec
9
,
2
0
2
1
Hu
m
a
n
s
c
a
n
p
e
rf
o
rm
a
n
e
n
o
rm
o
u
s
n
u
m
b
e
r
o
f
a
c
ti
o
n
s
li
k
e
ru
n
n
i
n
g
,
wa
lk
in
g
,
p
u
sh
i
n
g
,
a
n
d
p
u
n
c
h
in
g
,
a
n
d
c
a
n
p
e
rfo
rm
th
e
m
in
m
u
lt
i
p
le
wa
y
s.
He
n
c
e
re
c
o
g
n
izin
g
a
h
u
m
a
n
a
c
ti
o
n
fr
o
m
a
v
id
e
o
is
a
c
h
a
ll
e
n
g
i
n
g
t
a
sk
.
In
a
su
p
e
rv
ise
d
lea
rn
i
n
g
e
n
v
ir
o
n
m
e
n
t
,
a
c
ti
o
n
s
a
re
first
re
p
re
se
n
ted
u
si
n
g
r
o
b
u
st
fe
a
tu
re
s
a
n
d
th
e
n
a
c
las
sifier
is
t
ra
in
e
d
fo
r
c
las
sifica
ti
o
n
.
T
h
e
se
lec
ti
o
n
o
f
a
c
las
sifier
d
o
e
s
a
ffe
c
t
th
e
p
e
rfo
rm
a
n
c
e
o
f
h
u
m
a
n
a
c
ti
o
n
re
c
o
g
n
i
ti
o
n
.
Th
is
wo
rk
f
o
c
u
se
s
o
n
t
h
e
c
o
m
p
a
riso
n
o
f
tw
o
stru
c
t
u
re
s
o
f
th
e
n
e
u
ra
l
n
e
two
r
k
,
n
a
m
e
ly
,
fe
e
d
fo
rwa
rd
n
e
u
ra
l
n
e
t
wo
rk
a
n
d
c
a
sc
a
d
e
fo
rwa
rd
n
e
u
ra
l
n
e
two
r
k
,
fo
r
h
u
m
a
n
a
c
ti
o
n
re
c
o
g
n
it
io
n
.
H
isto
g
ra
m
o
f
o
rien
ted
g
ra
d
ien
ts
(H
O
G
)
a
n
d
h
isto
g
ra
m
o
f
o
p
ti
c
a
l
fl
o
w
(H
O
F
)
a
re
u
se
d
a
s
fe
a
tu
re
s
f
o
r
re
p
re
se
n
ti
n
g
t
h
e
a
c
ti
o
n
s.
H
O
G
re
p
re
s
e
n
ts
th
e
sp
a
ti
a
l
fe
a
tu
re
s
o
f
th
e
v
id
e
o
w
h
il
e
H
O
F
g
iv
e
s
m
o
ti
o
n
fe
a
tu
re
s
o
f
th
e
v
i
d
e
o
.
Th
e
p
e
rfo
rm
a
n
c
e
o
f
t
wo
n
e
u
ra
l
n
e
two
rk
a
rc
h
it
e
c
tu
re
s
is
c
o
m
p
a
re
d
b
a
se
d
o
n
re
c
o
g
n
it
io
n
a
c
c
u
ra
c
y
.
W
e
ll
-
k
n
o
wn
p
u
b
li
c
a
ll
y
a
v
a
i
lab
le
d
a
tas
e
ts
fo
r
a
c
ti
o
n
a
n
d
i
n
te
ra
c
ti
o
n
d
e
tec
ti
o
n
a
r
e
u
se
d
fo
r
tes
ti
n
g
.
It
is
se
e
n
th
a
t,
f
o
r
h
u
m
a
n
a
c
ti
o
n
re
c
o
g
n
i
ti
o
n
a
p
p
li
c
a
ti
o
n
s,
fe
e
d
fo
rwa
rd
n
e
u
ra
l
n
e
two
rk
g
iv
e
s
b
e
tt
e
r
re
su
lt
s
in
term
s
o
f
h
i
g
h
e
r
re
c
o
g
n
i
ti
o
n
a
c
c
u
ra
c
y
th
a
n
Ca
sc
a
d
e
fo
rwa
rd
n
e
u
ra
l
n
e
two
r
k
.
K
ey
w
o
r
d
s
:
C
FNN
F
FNN
HOF
Hu
m
an
ac
tio
n
r
ec
o
g
n
itio
n
HOG
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
:
Ad
iti J
ah
ag
ir
d
ar
Sch
o
o
l o
f
C
o
m
p
u
ter
E
n
g
in
ee
r
i
n
g
an
d
T
ec
h
n
o
lo
g
y
,
MI
T
W
o
r
ld
Peac
e
Un
iv
er
s
ity
Pu
n
e,
I
n
d
ia
E
m
ail:
ad
iti.jah
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Data
an
aly
tics
is
a
cu
r
r
en
t
b
u
z
zwo
r
d
in
th
e
co
m
p
u
ter
in
d
u
s
tr
y
.
W
ith
im
m
e
n
s
e
d
ev
el
o
p
m
en
t
in
d
i
g
ital
tech
n
o
lo
g
y
,
th
e
am
o
u
n
t
o
f
d
ig
ital
d
ata
g
en
er
ated
is
in
cr
e
asin
g
d
ay
b
y
d
ay
.
Acc
ess
to
ea
s
y
d
ev
ices
lik
e
s
m
ar
tp
h
o
n
es
an
d
clo
s
ed
-
cir
c
u
i
t
telev
is
io
n
(
C
C
T
V
)
ca
m
er
as
h
as
co
n
tr
ib
u
ted
to
v
ast
in
cr
ea
s
e
in
th
e
im
ag
e
an
d
v
id
eo
d
ata.
A
n
aly
zin
g
th
is
d
ata
m
an
u
ally
h
as
b
ec
o
m
e
a
t
ed
io
u
s
an
d
tim
e
-
c
o
n
s
u
m
in
g
t
ask
.
T
o
tack
le
th
is
p
r
o
b
lem
v
ar
io
u
s
alg
o
r
ith
m
s
a
n
d
m
eth
o
d
s
h
a
v
e
b
ee
n
p
r
o
p
o
s
ed
f
o
r
au
to
m
atic
v
id
eo
an
d
i
m
ag
e
an
aly
s
is
.
T
h
is
ar
ea
o
f
r
esear
ch
is
well
k
n
o
wn
b
y
th
e
n
am
e
in
tellig
en
t
v
id
eo
an
aly
s
is
an
d
f
in
d
s
ap
p
l
icatio
n
s
in
in
tellig
en
t
v
id
eo
s
u
r
v
eillan
ce
,
h
u
m
an
-
co
m
p
u
ter
in
te
r
ac
tio
n
,
r
o
b
o
tics
,
s
m
ar
t
h
ea
lth
ca
r
e,
s
m
ar
t
h
o
m
e
[
1
]
.
Hu
m
an
ac
tio
n
r
ec
o
g
n
itio
n
(
HAR)
is
an
in
teg
r
al
p
ar
t o
f
i
n
tellig
en
t v
id
eo
an
a
ly
tics
.
Hu
m
an
ac
tio
n
is
d
ef
in
ed
b
y
Her
ath
et
a
l
.
[
2
]
as
"Ac
tio
n
is
th
e
m
o
s
t
elem
en
tar
y
h
u
m
a
n
-
s
u
r
r
o
u
n
d
in
g
in
ter
ac
tio
n
with
a
m
ea
n
in
g
".
Hu
m
an
ac
tio
n
s
ar
e
b
r
o
a
d
ly
d
i
v
id
ed
in
to
g
estu
r
es,
s
im
p
le
ac
tio
n
s
,
in
ter
ac
tio
n
s
,
an
d
g
r
o
u
p
ac
tiv
ities
.
Mo
v
in
g
o
f
a
p
alm
o
r
n
o
d
d
in
g
o
f
th
e
h
ea
d
is
co
n
s
id
er
ed
as
a
g
estu
r
e.
A
p
er
s
o
n
walk
in
g
,
ju
m
p
in
g
o
r
b
en
d
i
n
g
is
co
n
s
id
er
ed
a
s
a
s
im
p
le
ac
tio
n
.
Han
d
s
h
ak
e
b
y
two
p
eo
p
le
o
r
o
n
e
p
er
s
o
n
p
u
s
h
in
g
o
t
h
er
is
co
n
s
id
er
ed
as
a
h
u
m
an
-
h
u
m
an
in
ter
ac
tio
n
.
A
p
er
s
o
n
wa
lk
in
g
with
a
d
o
g
o
r
p
ick
in
g
a
b
ag
is
co
n
s
id
er
ed
a
h
u
m
an
-
o
b
ject
in
ter
ac
tio
n
.
Mo
r
e
th
an
two
p
e
o
p
le
talk
in
g
o
r
d
an
cin
g
is
co
n
s
id
er
e
d
a
g
r
o
u
p
a
ctio
n
[
3
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
C
o
mp
a
r
is
o
n
o
f fe
ed
fo
r
w
a
r
d
a
n
d
ca
s
ca
d
e
fo
r
w
a
r
d
n
eu
r
a
l n
et
w
o
r
ks fo
r
h
u
ma
n
a
ctio
n
…
(
A
d
iti Ja
h
a
g
ir
d
a
r
)
893
E
v
en
a
f
te
r
b
ei
n
g
r
esear
ch
ed
f
o
r
m
a
n
y
y
ea
r
s
,
h
u
m
an
ac
tio
n
r
ec
o
g
n
itio
n
r
em
ain
s
a
ch
alle
n
g
in
g
task
b
ec
au
s
e
o
f
its
v
ast
s
co
p
e.
T
h
e
m
ain
ch
allen
g
e
in
h
u
m
an
ac
tio
n
r
ec
o
g
n
iti
o
n
is
th
at
th
er
e
is
n
o
lim
it
to
ac
tio
n
s
th
at
ca
n
b
e
p
er
f
o
r
m
e
d
b
y
a
h
u
m
an
b
ein
g
.
Actio
n
s
lik
e
jo
g
g
i
n
g
,
walk
in
g
,
an
d
r
u
n
n
in
g
ca
n
cr
ea
te
co
n
f
u
s
io
n
f
o
r
an
au
to
m
atic
ac
tio
n
r
ec
o
g
n
itio
n
s
y
s
tem
.
An
o
th
er
o
b
s
tacle
in
r
ec
o
g
n
itio
n
is
th
at
th
er
e
is
a
la
r
g
e
d
i
v
er
s
ity
i
n
th
e
way
in
wh
ich
a
p
ar
ticu
lar
ac
ti
o
n
is
p
er
f
o
r
m
ed
.
T
h
is
g
iv
es
r
i
s
e
to
h
ig
h
in
tr
ac
lass
v
ar
iatio
n
.
Oth
er
co
n
d
itio
n
s
lik
e
v
ar
ied
ca
m
er
a
an
g
les,
ca
m
er
a
m
o
tio
n
,
s
ca
le
c
h
an
g
es,
illu
m
in
atio
n
ch
a
n
g
es,
clu
tter
e
d
b
ac
k
g
r
o
u
n
d
,
an
d
o
cc
lu
s
io
n
ad
d
to
th
e
c
h
allen
g
e
s
f
ac
ed
b
y
t
h
e
au
to
m
atic
h
u
m
a
n
ac
tio
n
r
ec
o
g
n
itio
n
s
y
s
tem
.
Mo
s
t
o
f
t
h
e
wo
r
k
i
n
th
is
ar
ea
u
s
es
th
e
s
u
p
er
v
is
ed
lea
r
n
in
g
ap
p
r
o
ac
h
.
T
h
e
m
ain
s
tep
s
in
th
e
HAR
s
y
s
tem
ar
e
f
ea
t
u
r
e
e
x
tr
ac
tio
n
,
f
ea
tu
r
e
s
elec
tio
n
a
n
d
tr
ain
in
g
a
class
if
ier
with
ex
tr
ac
te
d
f
ea
tu
r
es
[
4
]
f
o
r
th
e
class
iifc
atio
n
.
T
h
e
ch
o
ice
o
f
f
ea
tu
r
es to
b
e
ex
tr
ac
ted
f
o
r
r
ep
r
esen
tin
g
th
e
ac
tio
n
d
ep
en
d
s
o
n
th
e
ty
p
e
o
f
ac
tio
n
.
Alg
o
r
ith
m
s
p
r
o
p
o
s
ed
f
o
r
g
estu
r
es
r
ec
o
g
n
itio
n
,
s
im
p
le
ac
tio
n
r
ec
o
g
n
itio
n
,
a
n
d
g
r
o
u
p
ac
tio
n
r
ec
o
g
n
itio
n
d
if
f
e
r
m
ain
ly
in
th
e
s
elec
tio
n
o
f
f
ea
tu
r
es.
Var
io
u
s
cla
s
s
if
ier
s
lik
e
th
e
k
-
n
ea
r
est
n
eig
h
b
o
r
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e,
an
d
n
e
u
r
al
n
etwo
r
k
a
r
e
ex
p
lo
r
ed
f
o
r
class
if
icatio
n
p
u
r
p
o
s
e
s
.
I
t
is
o
b
s
er
v
ed
th
at,
alo
n
g
with
th
e
s
elec
tio
n
o
f
ap
p
r
o
p
r
iate
f
ea
tu
r
e,
s
elec
tio
n
o
f
ap
p
r
o
p
r
iate
class
if
ier
p
la
y
s
an
im
p
o
r
tan
t
r
o
le
in
HAR
p
er
f
o
r
m
a
n
ce
.
T
h
is
wo
r
k
f
o
c
u
s
es
o
n
co
m
p
ar
is
o
n
o
f
two
n
e
u
r
al
n
etwo
r
k
ar
c
h
itectu
r
es,
n
am
ely
,
f
ee
d
f
o
r
wa
r
d
n
eu
r
al
n
etwo
r
k
(
FF
NN)
an
d
ca
s
ca
d
e
f
o
r
awa
r
d
n
eu
r
al
n
etwo
r
k
(
C
FNN)
f
o
r
h
u
m
a
n
ac
tio
n
r
ec
o
g
n
itio
n
.
two
h
an
d
-
cr
af
ted
f
ea
tu
r
es,
n
am
ely
,
h
is
to
g
r
am
o
f
o
p
tica
l
f
lo
w
(H
O
F)
an
d
h
is
to
g
r
am
o
f
g
r
a
d
ien
ts
(H
O
G)
ar
e
u
s
ed
f
o
r
r
ep
r
esen
tin
g
th
e
ac
tio
n
.
E
x
p
er
im
en
tatio
n
is
ca
r
r
ied
o
u
t
o
n
w
ell
k
n
o
wn
W
eizm
an
n
,
KT
H,
UT
in
ter
ac
tio
n
,
an
d
Un
iv
er
s
ity
o
f
C
en
tr
al
Flo
r
id
a
(
UC
F
)
s
p
o
r
ts
ac
tio
n
d
atasets
.
R
ec
o
g
n
itio
n
ac
cu
r
ac
y
is
u
s
ed
as
a
p
er
f
o
r
m
an
ce
p
ar
am
eter
f
o
r
co
m
p
a
r
in
g
th
e
a
r
ch
itectu
r
es.
T
h
e
h
ig
h
est
r
ec
o
g
n
itio
n
ac
c
u
r
ac
y
o
f
9
7
.
5
9
%
is
ac
h
iev
ed
f
o
r
UT
1
in
ter
ac
tio
n
d
ataset
with
FF
N
N
ar
ch
itectu
r
e
.
T
h
e
ac
cu
r
ac
y
ca
n
b
e
im
p
r
o
v
ed
f
u
r
th
er
b
y
u
s
in
g
d
if
f
er
en
t
h
a
n
d
-
cr
af
ted
f
ea
tu
r
es.
E
ar
lier
wo
r
k
o
n
h
u
m
a
n
ac
tio
n
r
ec
o
g
n
itio
n
s
h
o
ws
th
e
u
s
e
o
f
v
ar
io
u
s
h
an
d
-
cr
af
ted
f
ea
t
u
r
es.
Han
d
-
cr
af
ted
f
ea
tu
r
es
ar
e
d
iv
id
ed
in
t
o
two
ca
teg
o
r
ies
as
lo
ca
l
f
ea
tu
r
es
an
d
g
lo
b
al
f
ea
tu
r
es.
L
o
ca
l f
ea
tu
r
e
d
ef
i
n
es
th
e
o
b
ject
in
p
ar
ts
an
d
th
en
th
ese
p
ar
ts
ar
e
co
m
b
in
ed
t
o
f
o
r
m
a
l
o
ca
l
f
ea
tu
r
e
d
escr
ip
to
r
.
Glo
b
al
f
ea
tu
r
e
d
ef
in
es
an
o
b
ject
as
a
wh
o
le.
E
ac
h
ty
p
e
o
f
f
ea
tu
r
e
h
as
its
ad
v
an
ta
g
e
a
n
d
d
is
ad
v
an
ta
g
e.
Pre
v
io
u
s
wo
r
k
in
th
is
d
o
m
ain
h
as
em
p
h
asized
th
e
n
ee
d
o
f
u
s
in
g
m
u
ltip
le
f
ea
tu
r
es
t
o
d
escr
ib
e
a
n
ac
tio
n
.
As
o
n
e
ty
p
e
o
f
f
ea
tu
r
e
ca
n
ca
p
tu
r
e
o
n
ly
o
n
e
o
f
th
e
p
r
o
p
er
ties
o
f
a
v
id
e
o
,
m
u
ltip
le
f
ea
tu
r
es
alwa
y
s
h
e
lp
in
d
escr
ib
in
g
an
ac
tio
n
ef
f
icien
tly
as
p
r
o
v
ed
in
[
4
]
.
Ma
n
y
r
esear
ch
er
s
h
av
e
c
o
m
b
in
ed
lo
ca
l
a
n
d
g
lo
b
al
f
ea
t
u
r
es
f
o
r
in
cr
ea
s
in
g
r
ec
o
g
n
itio
n
ac
cu
r
ac
y
.
R
eg
io
n
o
f
in
ter
est is
d
etec
te
d
b
ef
o
r
e
e
x
tr
ac
tin
g
ac
tu
al
f
ea
t
u
r
es in
m
a
n
y
ap
p
r
o
ac
h
es [
5
]
-
[
8
]
.
B
ak
et
a
l.
[
5
]
h
av
e
u
s
ed
a
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
f
o
r
s
alien
t
r
eg
io
n
d
etec
tio
n
.
Var
io
u
s
f
u
s
io
n
m
ec
h
an
is
m
s
ar
e
ex
p
lo
r
e
d
f
o
r
ass
im
ilatin
g
s
p
atial
an
d
te
m
p
o
r
al
f
ea
tu
r
e
s
.
Ab
d
u
lm
u
n
e
m
et
a
l
.
[
6
]
h
a
v
e
p
r
o
p
o
s
ed
th
e
u
s
e
o
f
an
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
es
(
SVM
)
class
if
ier
f
o
r
class
if
y
in
g
o
b
je
cts
d
escr
ib
ed
b
y
a
co
m
b
in
atio
n
o
f
g
lo
b
al
an
d
lo
c
al
f
ea
tu
r
es.
3
D
g
r
ad
i
e
n
t
lo
ca
ti
o
n
an
d
o
r
ien
tatio
n
h
is
to
g
r
am
s
(
GL
OH
)
v
ec
to
r
is
p
r
o
p
o
s
ed
b
y
Ab
d
u
lm
u
n
em
et
a
l.
[
7
]
.
3
D
GL
OH
co
m
b
in
es
g
r
ad
ien
t
lo
ca
tio
n
s
an
d
o
r
ie
n
t
atio
n
h
is
to
g
r
am
.
I
n
Du
ta
et
a
l.
[
8
]
n
ew
f
ea
tu
r
e
en
co
d
in
g
m
eth
o
d
s
n
a
m
ely
v
ec
to
r
o
f
lo
ca
lly
ag
g
r
eg
ated
d
escr
i
p
to
r
s
(
VL
AD
)
an
d
s
p
atio
-
tem
p
o
r
al
(
ST_
VL
AD
)
ar
e
p
r
o
p
o
s
ed
b
y
I
o
n
u
t
C
.
an
d
o
th
er
s
.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
g
iv
es
co
m
p
a
r
ab
le
r
esu
lts
o
n
d
atasets
u
s
ed
f
o
r
test
in
g
.
A
d
etailed
s
tu
d
y
o
f
th
e
b
ag
o
f
v
is
u
al
wo
r
d
m
o
d
el
u
s
in
g
lo
ca
l
f
ea
tu
r
es
ap
p
lied
to
h
u
m
an
ac
tio
n
r
ec
o
g
n
itio
n
is
g
iv
en
in
[
9
]
.
A
n
ew
b
ag
o
f
v
is
u
al
wo
r
d
f
r
am
ewo
r
k
ca
lled
h
y
b
r
id
s
u
p
er
v
ec
to
r
is
also
p
r
o
p
o
s
ed
in
th
is
p
ap
er
wh
ich
g
iv
es
p
r
o
m
is
in
g
r
esu
lts
.
A
n
e
w
f
ea
tu
r
e
d
escr
ip
to
r
u
s
in
g
th
e
f
u
s
io
n
o
f
s
tatio
n
ar
y
wav
elet
tr
an
s
f
o
r
m
(
SW
T
)
an
d
lo
ca
l
b
in
ar
y
p
atter
n
(
L
B
P)
is
p
r
o
p
o
s
ed
in
[
1
0
]
.
A
d
is
cr
ete
wav
elet
tr
an
s
f
o
r
m
(
DW
T
)
b
ased
m
eth
o
d
is
p
r
o
p
o
s
ed
in
[
1
1
]
.
Fo
u
r
-
lev
el
DW
T
is
ap
p
lied
to
f
in
d
th
e
f
ea
tu
r
es
an
d
th
en
th
e
s
tep
wis
e
lin
ea
r
d
is
cr
im
in
an
t
an
aly
s
is
is
ap
p
lied
f
o
r
f
in
d
in
g
k
ey
f
ea
tu
r
es
to
b
e
u
s
ed
in
tr
a
in
in
g
.
3
D
s
tatio
n
ar
y
wav
elet
t
r
an
s
f
o
r
m
is
u
s
ed
to
d
escr
ib
e
th
e
ac
tio
n
in
[
1
2
]
,
[
1
3
]
.
Acc
u
m
u
late
m
o
tio
n
im
a
g
e
(
AM
I
)
an
d
m
o
tio
n
AM
I
h
is
to
r
y
im
ag
e
(
MH
I
)
a
r
e
in
tr
o
d
u
ce
d
in
[
1
4
]
.
DW
T
f
ea
tu
r
es
ar
e
f
u
r
th
er
e
x
tr
ac
ted
f
r
o
m
AM
I
an
d
L
B
P
f
ea
tu
r
es
ar
e
ex
tr
ac
ted
f
r
o
m
MH
I
im
ag
es
to
f
o
r
m
a
f
ea
tu
r
e
d
escr
ip
to
r
.
J
y
o
ts
n
a
et
a
l.
[
1
5
]
,
H
O
G
alo
n
g
with
p
r
in
cip
al
co
m
p
o
n
en
t
an
aly
s
is
(
PC
A
)
is
u
s
ed
to
d
escr
ib
e
th
e
ac
tio
n
af
ter
ap
p
l
y
in
g
th
e
s
eg
m
e
n
t
atio
n
.
K
n
ea
r
est
n
eig
h
b
o
r
cla
s
s
if
ier
is
u
s
ed
as
a
class
if
ier
wh
ich
g
iv
es
r
ec
o
g
n
itio
n
ac
cu
r
ac
y
o
f
9
4
%
o
n
W
ei
zm
an
n
an
d
9
1
.
8
3
%
o
n
th
e
K
T
H
d
ataset.
H
O
G
i
s
s
u
cc
ess
f
u
lly
u
s
ed
f
o
r
f
ac
e
r
ec
o
g
n
itio
n
in
in
tellig
en
t
s
u
r
v
eill
an
ce
s
y
s
tem
in
[
1
6
]
.
A
co
m
b
i
n
atio
n
o
f
H
O
G
an
d
lo
ca
l
f
ea
tu
r
e
s
win
e
co
n
f
in
em
en
t
wo
r
k
er
(
SW
F
)
[
1
7
]
is
s
ee
n
to
g
iv
e
h
ig
h
class
if
icatio
n
ac
cu
r
ac
y
f
o
r
U
T
i
n
ter
ac
tio
n
an
d
UC
F sp
o
r
ts
d
atasets
.
Ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
(
ANN)
h
av
e
r
ef
o
r
m
ed
th
e
d
o
m
ai
n
o
f
m
ac
h
in
e
lear
n
in
g
.
ANNs
ar
e
wid
el
y
u
s
ed
in
h
u
m
a
n
ac
tio
n
r
ec
o
g
n
itio
n
p
r
o
b
lem
s
b
ec
a
u
s
e
o
f
th
eir
ca
p
ab
ilit
y
to
m
ap
c
o
m
p
le
x
in
p
u
ts
to
o
u
t
p
u
ts
.
Sev
er
al
ty
p
es
o
f
ANNs
ar
e
e
x
p
lo
r
ed
f
o
r
clas
s
if
y
in
g
d
if
f
e
r
en
t
ty
p
es
o
f
d
ata
.
T
eix
eir
a
a
n
d
Fer
n
an
d
es
[
1
8
]
h
av
e
co
m
p
ar
ed
p
er
f
o
r
m
a
n
ce
s
o
f
f
e
ed
-
f
o
r
war
d
n
e
u
r
al
n
etwo
r
k
a
n
d
ca
s
ca
d
e
f
o
r
war
d
n
eu
r
al
n
etwo
r
k
f
o
r
tim
e
s
er
ies
d
o
m
ain
t
o
p
r
o
v
e
t
h
e
ad
v
an
tag
e
o
f
ca
s
ca
d
e
f
o
r
war
d
n
e
u
r
al
n
etwo
r
k
.
Dh
a
n
aseely
[
1
9
]
h
a
v
e
p
r
esen
ted
r
esu
lts
o
b
tain
ed
b
y
f
ee
d
-
f
o
r
war
d
n
eu
r
al
n
etwo
r
k
an
d
ca
s
ca
d
e
f
o
r
w
ar
d
n
eu
r
al
n
etwo
r
k
f
o
r
f
ac
e
r
ec
o
g
n
itio
n
d
ataset
with
p
r
in
cip
al
co
m
p
o
n
e
n
t
an
a
ly
s
is
u
s
ed
as
a
f
ea
tu
r
e.
B
ad
d
e
et
a
l.
[
2
0
]
an
d
G
o
y
al
[
2
1
]
u
s
e
o
f
f
ee
d
-
f
o
r
war
d
b
ac
k
p
r
o
p
ag
atio
n
n
etwo
r
k
s
a
n
d
ca
s
ca
d
e
f
o
r
war
d
b
ac
k
p
r
o
p
ag
atio
n
n
etwo
r
k
s
ar
e
e
x
p
lo
r
ed
in
th
e
civ
il
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
25
,
No
.
2
,
Feb
r
u
a
r
y
20
22
:
8
9
2
-
8
9
9
894
en
g
in
ee
r
in
g
d
o
m
ain
.
C
ascad
e
f
o
r
war
d
n
etwo
r
k
is
s
h
o
w
n
t
o
g
iv
e
b
etter
ac
c
u
r
ac
y
in
c
o
m
p
ar
is
o
n
to
f
ee
d
-
f
o
r
war
d
n
etwo
r
k
.
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
T
h
e
m
eth
o
d
f
o
r
h
u
m
an
ac
tio
n
r
ec
o
g
n
itio
n
p
r
o
p
o
s
ed
in
th
i
s
wo
r
k
is
g
iv
en
in
th
is
s
ec
tio
n
.
B
lo
ck
s
ch
em
atic
o
f
th
e
p
r
o
p
o
s
ed
m
e
th
o
d
is
g
i
v
en
i
n
Fig
u
r
e
1
.
A
te
s
t
v
id
eo
is
co
n
v
er
ted
to
f
r
am
e
s
an
d
p
r
ep
r
o
ce
s
s
ed
f
o
r
d
e
-
n
o
is
in
g
.
Fo
r
h
u
m
an
a
ctio
n
r
ec
o
g
n
itio
n
,
s
p
atial
as
well
as
tem
p
o
r
al
in
f
o
r
m
ati
o
n
is
im
p
o
r
tan
t.
A
h
is
to
g
r
am
o
f
g
r
ad
ien
ts
is
u
s
ed
h
er
e
to
r
e
p
r
esen
t
s
p
atial
in
f
o
r
m
atio
n
o
f
t
h
e
ac
tio
n
.
His
to
g
r
am
o
f
o
p
tical
f
lo
w
r
ep
r
esen
ts
th
e
tem
p
o
r
al
o
r
m
o
tio
n
f
ea
tu
r
es
o
f
th
e
ac
tio
n
.
Featu
r
e
s
elec
tio
n
is
d
o
n
e
u
s
in
g
Prin
cip
al
co
m
p
o
n
en
t
an
aly
s
is
.
P
C
A
i
s
ap
p
lied
to
b
o
th
th
e
f
ea
tu
r
es
s
ep
ar
ately
to
r
ed
u
ce
th
e
d
im
en
s
io
n
ality
.
H
O
G
an
d
H
O
F
f
ea
tu
r
es
ar
e
th
en
co
n
ca
ten
ated
to
f
o
r
m
a
f
in
a
l
f
ea
tu
r
e
d
escr
ip
to
r
.
A
n
eu
r
al
n
etwo
r
k
is
th
en
tr
ain
e
d
with
th
ese
f
ea
tu
r
e
d
escr
ip
to
r
s
an
d
u
s
ed
to
class
if
y
th
e
test
v
id
eo
.
T
h
e
f
o
llo
win
g
s
u
b
-
s
ec
tio
n
s
d
escr
ib
e
ea
ch
s
te
p
-
in
d
etail.
Fig
u
r
e
1
.
B
lo
ck
s
ch
em
atic
f
o
r
p
r
o
p
o
s
ed
h
u
m
a
n
ac
tio
n
r
ec
o
g
n
itio
n
s
y
s
tem
3
.
1
.
H
is
t
o
g
ra
m
o
f
o
rient
ed
g
ra
dient
(
H
O
G
)
His
to
g
r
am
o
f
o
r
ien
ted
g
r
ad
ie
n
ts
(
HOG)
d
escr
ib
es
an
o
b
ject
in
a
f
r
am
e
u
s
in
g
its
s
p
atial
in
f
o
r
m
atio
n
.
HOG
was
f
ir
s
t
d
ev
elo
p
ed
f
o
r
r
ec
o
g
n
izin
g
h
u
m
an
f
i
g
u
r
es
f
r
o
m
an
im
a
g
e.
Hen
ce
HOG
b
ec
o
m
es
a
p
er
f
ec
t
ch
o
ice
f
o
r
r
e
p
r
esen
tin
g
s
p
atial
f
ea
tu
r
es
in
t
h
e
h
u
m
an
ac
tio
n
r
ec
o
g
n
itio
n
task
.
T
h
e
o
r
ig
i
n
al
alg
o
r
ith
m
[
2
2
]
wh
ich
was
d
ev
elo
p
ed
f
o
r
a
n
im
ag
e
is
ap
p
lied
to
a
v
id
eo
h
e
r
e
b
y
co
n
s
id
er
in
g
ea
ch
f
r
am
e
as
an
im
ag
e.
E
ac
h
f
r
am
e
is
d
iv
id
ed
i
n
to
ce
lls
wh
ich
ar
e
s
m
all
s
p
atial
r
eg
io
n
s
an
d
f
o
r
ea
ch
p
ix
el
m
ag
n
itu
d
e
an
d
o
r
ie
n
tatio
n
o
f
g
r
ad
ien
t a
r
e
ca
lcu
lated
.
1
D
h
is
to
g
r
am
s
ar
e
th
e
n
u
s
ed
to
r
ep
r
e
s
en
t e
ac
h
ce
ll f
o
r
m
in
g
th
e
HO
G
f
ea
tu
r
e.
3
.
2
.
H
is
t
o
g
ra
m
o
f
op
t
ica
l f
lo
w
(HO
F
)
His
to
g
r
am
o
f
o
p
tical
f
lo
w
is
p
r
o
v
ed
to
h
av
e
th
e
ca
p
ab
ilit
y
o
f
r
ep
r
esen
tin
g
h
u
m
a
n
b
o
d
y
m
o
tio
n
[
2
3
]
.
Op
tical
f
lo
w
is
n
o
th
in
g
b
u
t
a
p
atter
n
o
f
ap
p
a
r
en
t
m
o
v
e
m
e
n
t
in
a
s
eq
u
en
ce
th
at
r
ep
r
ese
n
ts
r
elativ
e
m
o
tio
n
b
etwe
en
th
e
o
b
s
er
v
er
an
d
th
e
s
eq
u
en
ce
.
I
t
is
o
b
tain
ed
b
y
f
in
d
in
g
ch
an
g
es
in
th
e
p
o
s
itio
n
o
f
th
e
o
b
ject
in
two
co
n
s
ec
u
tiv
e
f
r
am
es.
Her
e
,
H
OF
is
r
ep
r
esen
ted
b
y
o
p
tical
v
ec
to
r
s
ca
lcu
lated
at
ea
c
h
p
ix
el.
Usi
n
g
th
e
f
ea
tu
r
e
s
elec
tio
n
m
eth
o
d
,
o
p
tical
v
ec
t
o
r
s
h
av
in
g
m
ax
im
u
m
v
alu
e
ar
e
s
elec
ted
to
f
o
r
m
a
f
ea
tu
r
e
d
e
s
cr
ip
to
r
.
3
.
3
.
Neura
l
net
wo
r
k
T
wo
ar
ch
itectu
r
es
o
f
n
eu
r
al
n
etwo
r
k
s
ar
e
u
s
ed
s
ep
ar
ately
f
o
r
ev
alu
atin
g
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
s
y
s
tem
.
T
h
e
ar
ch
itectu
r
es
o
f
f
ee
d
-
f
o
r
war
d
n
eu
r
al
n
etwo
r
k
(
FF
NN)
an
d
ca
s
ca
d
e
f
o
r
war
d
n
e
u
r
al
n
etwo
r
k
(
C
FNN
)
[
2
4
]
ar
e
s
h
o
wn
in
th
e
F
ig
u
r
e
s
2
an
d
3
r
esp
ec
tiv
ely
.
FF
NN
i
s
th
e
m
o
s
t
co
m
m
o
n
ly
u
s
ed
n
e
u
r
al
n
etwo
r
k
m
o
d
el
in
wh
ich
th
e
i
n
p
u
t
lay
er
,
h
id
d
e
n
lay
er
s
,
an
d
o
u
tp
u
t
lay
er
a
r
e
u
s
ed
.
Fig
u
r
e
2
s
h
o
ws
th
e
g
en
er
al
ar
ch
itectu
r
e
o
f
FF
NN.
All t
h
e
in
p
u
t n
o
d
es a
r
e
co
n
n
ec
ted
to
a
ll th
e
n
o
d
es in
1
s
t h
id
d
e
n
lay
e
r
an
d
all
th
e
h
id
d
en
n
o
d
es
o
f
th
e
last
h
id
d
en
la
y
er
ar
e
co
n
n
ec
ted
to
t
h
e
o
u
tp
u
t
lay
er
.
T
h
e
d
ir
ec
tio
n
o
f
d
ata
is
o
n
ly
in
o
n
e
d
i
r
ec
tio
n
i.e
.
in
p
u
t
to
o
u
t
p
u
t.
T
h
e
b
ac
k
p
r
o
p
ag
atio
n
alg
o
r
ith
m
is
u
s
e
d
to
ca
lcu
late
t
h
e
weig
h
ts
b
etwe
en
lay
er
s
.
Mu
ltip
le
lay
er
s
o
f
n
eu
r
o
n
s
an
d
a
b
ac
k
p
r
o
p
ag
atio
n
alg
o
r
ith
m
m
a
k
e
it
p
o
s
s
ib
le
f
o
r
th
e
n
etwo
r
k
to
lea
r
n
lin
ea
r
as
well
as
n
o
n
lin
ea
r
r
elatio
n
s
b
etwe
en
in
p
u
t a
n
d
o
u
tp
u
t.
C
ascad
e
f
o
r
war
d
n
eu
r
al
n
etw
o
r
k
is
s
im
ilar
t
o
FF
N
N
b
u
t
i
s
h
av
in
g
an
ex
tr
a
weig
h
ted
c
o
n
n
ec
tio
n
f
r
o
m
th
e
in
p
u
t
lay
er
to
ea
c
h
h
id
d
en
lay
er
an
d
f
r
o
m
ea
ch
h
id
d
en
lay
er
to
s
u
cc
ess
iv
e
lay
er
.
T
h
is
ex
tr
a
co
n
n
ec
tio
n
f
r
o
m
in
p
u
t
to
ea
ch
lay
er
m
ak
es
th
e
lear
n
in
g
o
f
th
e
n
etwo
r
k
f
aster
.
Fig
u
r
e
3
s
h
o
ws
th
e
ar
ch
itectu
r
e
o
f
C
FNN.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
C
o
mp
a
r
is
o
n
o
f fe
ed
fo
r
w
a
r
d
a
n
d
ca
s
ca
d
e
fo
r
w
a
r
d
n
eu
r
a
l n
et
w
o
r
ks fo
r
h
u
ma
n
a
ctio
n
…
(
A
d
iti Ja
h
a
g
ir
d
a
r
)
895
Fig
u
r
e
2
.
T
h
e
ar
ch
itectu
r
e
o
f
n
eu
r
al
n
etwo
r
k
: FFNN
Fig
u
r
e
3
.
T
h
e
ar
ch
itectu
r
e
o
f
n
eu
r
al
n
etwo
r
k
: CF
NN
Fo
r
f
air
co
m
p
a
r
is
o
n
o
f
th
e
p
er
f
o
r
m
a
n
ce
s
,
s
am
e
p
ar
am
eter
s
ar
e
u
s
ed
f
o
r
b
o
t
h
th
e
n
eu
r
a
l
n
etwo
r
k
ar
ch
itectu
r
es.
T
h
e
h
y
p
er
b
o
lic
tan
g
en
t
s
ig
m
o
id
f
u
n
ctio
n
is
u
s
ed
as
an
ac
tiv
atio
n
f
u
n
ctio
n
in
th
e
h
i
d
d
en
lay
e
r
n
o
d
es.
L
i
n
ea
r
f
u
n
ctio
n
is
u
s
ed
as
th
e
ac
tiv
atio
n
f
u
n
cti
o
n
f
o
r
th
e
o
u
tp
u
t
lay
er
.
L
e
v
en
b
er
g
-
Ma
r
q
u
ar
d
t
b
ac
k
p
r
o
p
ag
atio
n
alg
o
r
ith
m
,
w
h
ich
is
p
r
ef
er
r
ed
f
o
r
s
u
p
er
v
is
ed
le
ar
n
i
n
g
an
d
is
f
astes
t,
is
u
s
ed
f
o
r
lea
r
n
in
g
th
e
weig
h
ts
.
Me
an
s
q
u
ar
e
er
r
o
r
is
u
s
ed
as a
lo
s
s
f
u
n
ctio
n
.
3
.
4
.
Da
t
a
s
et
s
Pu
b
licly
av
ailab
le
d
atasets
,
n
am
ely
,
W
eizm
an
n
,
KT
H,
UC
F
s
p
o
r
ts
,
an
d
UT
in
ter
ac
tio
n
ac
tio
n
d
ata
s
ets
ar
e
u
s
ed
f
o
r
ev
alu
atin
g
t
h
e
p
e
r
f
o
r
m
an
ce
o
f
FF
NN
an
d
C
FNN
ar
ch
itectu
r
es.
T
h
ese
d
atasets
ar
e
s
elec
ted
f
o
r
ev
alu
atio
n
b
ec
au
s
e
o
f
th
ei
r
d
is
tin
ct
p
r
o
p
er
ties
.
I
n
th
e
W
eizm
an
n
d
ataset,
v
id
eo
s
o
f
ten
d
ay
-
to
-
d
ay
ac
tio
n
s
lik
e
walk
in
g
,
r
u
n
n
in
g
,
an
d
ju
m
p
in
g
a
r
e
in
co
r
p
o
r
ated
.
T
h
es
e
ac
tio
n
s
ar
e
p
e
r
f
o
r
m
e
d
b
y
n
i
n
e
d
if
f
er
en
t
ac
to
r
s
.
T
h
e
r
ec
o
r
d
i
n
g
is
d
o
n
e
in
a
co
n
tr
o
lled
en
v
ir
o
n
m
e
n
t
wh
er
e
o
n
ly
o
n
e
ac
to
r
is
p
r
esen
t
in
o
n
e
f
r
am
e
an
d
th
e
b
ac
k
g
r
o
u
n
d
is
u
n
clu
tter
ed
.
T
o
tal
n
in
ety
v
id
eo
s
ar
e
av
aila
b
le
in
th
is
d
ataset.
T
h
e
co
m
p
lex
ity
o
f
th
e
KT
H
d
ataset
is
m
o
r
e
th
an
th
e
W
eiz
m
an
n
d
ataset.
I
n
th
e
KT
H
d
ataset,
s
ix
s
im
p
le
ac
tio
n
s
lik
e
h
an
d
clap
p
in
g
,
wav
in
g
,
an
d
b
o
x
in
g
ar
e
p
er
f
o
r
m
ed
b
y
2
5
d
if
f
er
en
t
ac
to
r
s
.
E
ac
h
ac
tio
n
i
s
p
er
f
o
r
m
e
d
b
y
e
v
er
y
ac
to
r
in
f
o
u
r
d
if
f
er
en
t
s
ce
n
ar
i
o
s
.
T
h
e
s
ce
n
ar
io
s
u
s
ed
ar
e
in
d
o
o
r
,
o
u
td
o
o
r
,
ch
a
n
g
e
i
n
s
ca
le,
an
d
c
h
an
g
e
in
th
e
v
iew
an
g
le.
T
h
is
d
ataset
is
h
av
in
g
6
0
0
v
id
eo
s
r
ec
o
r
d
ed
in
a
co
n
tr
o
lled
en
v
ir
o
n
m
e
n
t.
UC
F
s
p
o
r
ts
d
ataset
i
s
also
h
av
i
n
g
o
n
e
ac
to
r
in
ev
e
r
y
v
id
e
o
b
u
t
th
e
r
ec
o
r
d
i
n
g
is
d
o
n
e
at
r
ea
l
-
tim
e
s
p
o
r
ts
ev
en
t
s
.
T
h
er
e
ar
e
v
id
e
o
s
o
f
te
n
d
i
f
f
er
en
t
s
p
o
r
ts
lik
e
h
o
r
s
e
r
i
d
in
g
,
g
o
lf
,
a
n
d
d
iv
in
g
.
As
th
ese
v
id
eo
s
a
r
e
r
ec
o
r
d
ed
in
r
ea
l
-
tim
e,
th
e
y
h
av
e
v
ar
y
i
n
g
b
ac
k
g
r
o
u
n
d
s
,
v
a
r
y
in
g
v
iew
a
n
g
les,
illu
m
in
atio
n
ch
an
g
es,
an
d
d
i
f
f
er
en
t
s
ca
les.
T
h
is
in
cr
ea
s
es
th
e
co
m
p
lex
ity
o
f
th
is
d
ataset.
UT
in
ter
ac
tio
n
d
at
aset
is
d
if
f
er
en
t
f
r
o
m
p
r
ev
io
u
s
ly
d
escr
ib
ed
d
at
asets
b
ec
au
s
e
it
is
h
av
in
g
two
ac
to
r
s
in
ev
er
y
v
id
e
o
.
T
h
er
e
ar
e
s
ix
ac
tio
n
s
lik
e
h
u
g
g
in
g
,
h
an
d
s
h
a
k
in
g
,
p
u
n
ch
i
n
g
,
p
u
s
h
in
g
,
an
d
k
ick
i
n
g
p
er
f
o
r
m
ed
b
y
1
0
d
if
f
er
e
n
t
p
air
s
.
On
ly
th
e
ac
tio
n
o
f
p
o
in
tin
g
a
f
i
n
g
er
is
p
e
r
f
o
r
m
e
d
b
y
a
s
in
g
le
ac
to
r
.
T
h
is
d
ataset
is
d
iv
id
ed
in
to
two
p
ar
ts
as UT
in
ter
ac
tio
n
1
(
UT
1
)
an
d
UT
in
ter
ac
tio
n
2
(
UT
2
)
d
ataset.
I
n
UT
1
d
ataset,
a
ctio
n
s
ar
e
p
er
f
o
r
m
ed
in
a
co
n
t
r
o
lled
en
v
ir
o
n
m
en
t.
I
n
UT
2
d
ataset
s
am
e
ac
tio
n
s
ar
e
p
er
f
o
r
m
ed
with
a
clu
tt
er
e
d
b
ac
k
g
r
o
u
n
d
,
p
ar
tial
o
c
clu
s
io
n
,
illu
m
in
atio
n
ch
an
g
es,
a
n
d
v
iew
an
g
le
ch
a
n
g
es.
I
n
f
ew
v
id
eo
s
o
f
th
e
UT
2
d
ataset,
m
o
r
e
t
h
an
two
ac
to
r
s
ar
e
p
r
esen
t
in
a
f
r
am
e,
m
a
k
in
g
th
e
r
ec
o
g
n
itio
n
task
m
o
r
e
ch
allen
g
in
g
.
Fig
u
r
e
4
s
h
o
ws
s
am
p
le
f
r
am
es
f
r
o
m
all
th
e
d
atasets
u
s
ed
.
Fig
u
r
e
4
(
a
)
s
h
o
ws
a
s
am
p
le
f
r
am
e
f
r
o
m
W
eizm
an
n
d
at
aset
o
f
ac
tio
n
class
‘
walk
’
.
Fig
u
r
e
4
(
b
)
s
h
o
ws
th
e
s
am
p
le
f
r
am
e
o
f
ac
tio
n
class
‘
walk
’
f
r
o
m
KT
H
d
ataset.
Fig
u
r
e
4
(
c)
s
h
o
ws
th
e
s
am
p
le
f
r
am
e
f
r
o
m
v
id
eo
o
f
ac
tio
n
class
‘
Swin
g
b
en
ch
’
f
r
o
m
UC
F
Sp
o
r
ts
d
at
aset.
Fi
g
u
r
e
4
(
d
)
s
h
o
ws
th
e
s
am
p
le
f
r
am
e
o
f
th
e
ac
tio
n
class
‘
s
h
ak
in
g
h
a
n
d
s
’
f
r
o
m
UT
i
n
ter
ac
tio
n
d
ataset.
(
a)
(
b
)
(
c)
(
d
)
Fig
u
r
e
4
.
Sam
p
le
f
r
am
es f
r
o
m
:
(
a)
W
eizm
an
n
,
(
b
)
KT
H
,
(
c
)
UC
F Sp
o
r
ts
,
an
d
(
d
)
UT
in
te
r
a
ctio
n
1
d
atasets
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
25
,
No
.
2
,
Feb
r
u
a
r
y
20
22
:
8
9
2
-
8
9
9
896
Fo
r
ev
alu
atin
g
th
e
p
e
r
f
o
r
m
an
ce
,
8
0
%
o
f
s
am
p
les
ar
e
u
s
ed
f
o
r
th
e
tr
ain
in
g
,
1
0
%
f
o
r
v
alid
atio
n
,
an
d
1
0
%
f
o
r
test
in
g
.
Stra
tifie
d
s
a
m
p
lin
g
is
u
s
ed
to
k
ee
p
th
e
n
u
m
b
er
o
f
s
am
p
les
o
f
ea
ch
class
p
r
o
p
o
r
tio
n
al
to
th
e
n
u
m
b
er
o
f
s
am
p
les
o
f
th
at
class
in
th
e
m
ain
d
ataset.
E
ac
h
s
etu
p
is
r
u
n
6
tim
es
co
n
s
id
er
in
g
d
if
f
er
en
t
s
am
p
les
f
o
r
t
r
ain
in
g
,
v
alid
atio
n
,
a
n
d
te
s
tin
g
,
an
d
an
av
er
ag
e
o
f
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
a
n
d
r
ec
all
a
r
e
ca
lcu
lated
f
o
r
b
o
th
n
eu
r
al
n
etwo
r
k
m
o
d
els
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
E
x
ten
s
iv
e
test
in
g
was
d
o
n
e
t
o
ev
alu
ate
th
e
p
e
r
f
o
r
m
an
ce
o
f
t
h
e
HOG+
HOF
d
escr
ip
to
r
f
o
r
t
h
e
h
u
m
an
ac
tio
n
r
ec
o
g
n
itio
n
s
y
s
tem
.
Fo
r
f
in
d
in
g
a
n
o
p
tim
u
m
n
u
m
b
er
o
f
h
i
d
d
en
lay
er
s
to
b
e
u
s
ed
,
e
x
p
er
im
en
tatio
n
was
p
er
f
o
r
m
ed
o
n
all
d
ata
s
ets
with
a
d
if
f
er
en
t
n
u
m
b
er
o
f
h
id
d
en
lay
er
s
,
an
d
ac
cu
r
ac
y
was
ca
lcu
lated
.
Fig
u
r
e
5
s
h
o
ws
a
g
r
ap
h
o
f
th
e
n
u
m
b
er
o
f
h
id
d
e
n
lay
er
s
p
lo
tted
ag
ain
s
t
ac
cu
r
ac
y
o
b
tain
e
d
.
T
h
e
d
ep
th
o
f
n
eu
r
al
n
etwo
r
k
s
is
in
cr
ea
s
ed
b
y
in
cr
e
asin
g
th
e
n
u
m
b
er
o
f
h
id
d
en
la
y
er
s
f
r
o
m
5
to
1
0
0
.
Fig
u
r
e
5
.
E
f
f
ec
t o
f
n
u
m
b
er
o
f
h
id
d
en
la
y
er
s
o
n
a
v
er
ag
e
ac
c
u
r
ac
y
I
t
is
o
b
s
er
v
ed
th
at
r
ec
o
g
n
iti
o
n
ac
cu
r
ac
y
v
ar
ies
f
r
o
m
8
8
%
to
9
7
%
f
o
r
d
if
f
er
e
n
t
d
ata
s
ets.
Al
s
o
,
r
ec
o
g
n
itio
n
ac
c
u
r
ac
y
c
h
an
g
es
as th
e
n
u
m
b
er
o
f
lay
er
s
ar
e
ch
an
g
ed
.
T
h
e
tim
e
r
eq
u
ir
e
d
f
o
r
t
r
ain
in
g
th
e
n
etwo
r
k
g
o
es
o
n
in
cr
ea
s
in
g
as
th
e
n
u
m
b
er
o
f
la
y
er
s
is
in
cr
ea
s
ed
.
W
ith
1
5
n
u
m
b
er
o
f
h
id
d
en
la
y
er
s
,
g
o
o
d
ac
cu
r
ac
y
is
ac
h
iev
ed
f
o
r
all
th
e
d
atasets
with
in
o
p
tim
al
tim
e.
Fo
r
al
l
th
e
f
u
r
th
er
ev
alu
atio
n
s
,
1
5
h
id
d
en
lay
er
s
ar
e
im
p
lem
en
ted
.
Fig
u
r
e
6
s
h
o
ws
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
n
eu
r
al
n
etwo
r
k
m
o
d
el
o
n
th
e
UT
_
2
i
n
ter
ac
tio
n
d
ata
s
et.
Fig
u
r
es
6
(
a)
a
n
d
(
b
)
s
h
o
w
th
e
v
alid
atio
n
p
er
f
o
r
m
an
ce
o
f
FF
NN
an
d
C
FNN
r
esp
ec
tiv
el
y
o
b
tain
e
d
f
o
r
UT
2
i
n
ter
ac
tio
n
d
ataset.
I
t
is
s
ee
n
th
at
m
ea
n
s
q
u
ar
e
er
r
o
r
r
e
d
u
ce
s
with
th
e
n
u
m
b
er
o
f
ep
o
ch
s
a
n
d
af
ter
s
o
m
e
ep
o
ch
s
,
it
is
alm
o
s
t
co
n
s
tan
t
.
Fo
r
FF
NN,
m
ea
n
s
q
u
ar
e
e
r
r
o
r
(
MSE
)
r
e
d
u
ce
s
alm
o
s
t
with
s
am
e
r
ate
f
o
r
tr
ain
in
g
,
test
in
g
an
d
v
alid
atio
n
s
am
p
les.
Af
ter
1
0
e
p
o
ch
s
,
MSE
co
n
v
er
g
es
to
th
e
s
am
e
v
alu
e
an
d
r
em
ain
s
co
n
s
tan
t
th
er
ea
f
ter
.
On
th
e
o
th
er
h
an
d
,
f
o
r
C
FNN,
MSE
r
ed
u
ce
s
f
ast
f
o
r
tr
ain
in
g
s
am
p
les
as
th
er
e
is
a
co
n
n
ec
tio
n
is
p
r
esen
t
f
r
o
m
th
e
in
p
u
t
la
y
er
to
in
ter
m
e
d
iate
h
i
d
d
en
la
y
er
s
.
I
t
is
s
ee
n
th
at
a
l
o
w
v
alu
e
o
f
MSE
is
ac
h
iev
ed
af
ter
o
n
ly
2
ep
o
ch
s
f
o
r
tr
ain
i
n
g
s
am
p
les.
Fo
r
test
a
n
d
v
alid
atio
n
d
ata
s
a
m
p
les,
M
SE
d
o
es
n
o
t
r
ed
u
ce
m
u
ch
a
n
d
r
e
m
ain
s
co
n
s
tan
t
a
f
ter
o
n
ly
o
n
e
ep
o
c
h
.
T
h
is
s
h
o
ws
th
at
b
ec
au
s
e
o
f
co
n
n
ec
tio
n
s
b
etwe
en
th
e
in
p
u
t
l
a
y
e
r
a
n
d
e
v
e
r
y
h
i
d
d
e
n
l
a
y
e
r
,
o
v
e
r
f
i
t
t
i
n
g
t
a
k
es
p
l
a
c
e
w
h
i
c
h
r
esu
l
t
s
i
n
h
i
g
h
MS
E
f
o
r
v
a
l
i
d
at
i
o
n
a
n
d
t
e
s
t
d
at
a
s
et
.
(
a)
(
b
)
Fig
u
r
e
6
.
Sam
p
le
v
alid
atio
n
p
er
f
o
r
m
a
n
ce
o
b
tain
e
d
with
:
(
a)
FF
NN
an
d
(
b
)
C
FNN
o
n
UT
d
ata
s
et
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
C
o
mp
a
r
is
o
n
o
f fe
ed
fo
r
w
a
r
d
a
n
d
ca
s
ca
d
e
fo
r
w
a
r
d
n
eu
r
a
l n
et
w
o
r
ks fo
r
h
u
ma
n
a
ctio
n
…
(
A
d
iti Ja
h
a
g
ir
d
a
r
)
897
E
r
r
o
r
h
is
to
g
r
a
m
s
is
th
e
an
o
th
er
m
ea
s
u
r
e
f
o
r
ev
alu
atin
g
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
class
if
icati
o
n
m
o
d
el.
T
h
e
h
is
to
g
r
am
o
f
class
if
icatio
n
er
r
o
r
p
lo
tted
ag
ain
s
t
t
h
e
in
s
t
an
ce
s
g
iv
es
th
e
d
is
tr
ib
u
tio
n
o
f
class
if
icatio
n
er
r
o
r
.
I
t
is
s
ee
n
th
at
m
o
s
t
o
f
th
e
s
a
m
p
le
p
o
in
ts
f
r
o
m
tr
ain
in
g
,
test
in
g
as
wel
l
as
v
alid
atio
n
d
ata
f
all
in
th
e
b
in
o
f
0
er
r
o
r
f
o
r
FF
NN
as we
ll a
s
C
FNN.
T
h
e
s
p
r
ea
d
o
f
er
r
o
r
h
is
to
g
r
am
is
m
o
r
e
f
o
r
FF
NN
th
an
f
o
r
C
FNN.
T
h
e
g
r
a
p
h
in
Fig
u
r
e
7
s
h
o
ws
r
ec
o
g
n
itio
n
ac
c
u
r
ac
y
o
b
tain
e
d
with
FF
NN
an
d
C
FNN
ar
ch
itectu
r
es.
T
h
e
ac
cu
r
ac
y
o
b
tain
ed
with
b
o
th
ar
ch
itectu
r
es
is
alm
o
s
t
th
e
s
am
e
f
o
r
th
e
W
eizm
an
n
d
ataset.
Fo
r
th
e
r
em
ain
in
g
all
d
atasets
,
th
e
ac
c
u
r
ac
y
o
b
tain
ed
with
FF
NN
is
m
o
r
e
th
an
th
at
o
b
tain
ed
wit
h
C
FNN
ar
ch
itectu
r
e.
Fig
u
r
e.
7
C
o
m
p
a
r
is
o
n
o
f
r
ec
o
g
n
itio
n
ac
cu
r
ac
y
o
b
tain
ed
wit
h
FF
NN
an
d
C
FNN
ar
ch
itectu
r
es
T
ab
le
1
s
h
o
ws
th
e
co
m
p
ar
is
o
n
o
f
r
ec
o
g
n
itio
n
ac
cu
r
ac
y
o
b
tain
ed
f
o
r
UC
F
s
p
o
r
ts
,
UT
i
n
ter
ac
tio
n
,
W
eizm
an
n
,
an
d
KT
H
ac
tio
n
d
atasets
wi
th
s
tate
-
o
f
-
th
e
-
ar
t
m
eth
o
d
s
p
r
esen
ted
in
th
e
liter
atu
r
e.
I
t
is
s
ee
n
th
at
f
o
r
th
e
UC
F
s
p
o
r
ts
d
ataset,
th
e
p
r
o
p
o
s
ed
m
et
h
o
d
with
FF
NN
g
iv
es
th
e
h
ig
h
est
ac
cu
r
ac
y
.
Fo
r
th
e
U
T
in
ter
ac
tio
n
d
ataset,
th
e
p
r
o
p
o
s
ed
m
eth
o
d
with
FF
NN
ar
ch
i
tectu
r
e
o
u
t
p
er
f
o
r
m
s
all
o
th
e
r
m
eth
o
d
s
b
a
s
ed
o
n
r
ec
o
g
n
itio
n
ac
cu
r
ac
y
.
I
t
is
o
b
s
er
v
ed
th
at
f
o
r
W
eizm
an
n
a
n
d
KT
H
d
atasets
,
co
m
p
ar
ab
le
ac
cu
r
ac
y
is
o
b
tain
ed
b
y
th
e
m
eth
o
d
p
r
o
p
o
s
ed
i
n
t
h
is
wo
r
k
.
Fo
r
all
th
e
d
atasets
FF
NN
ar
ch
itectu
r
e
g
iv
es
b
etter
ac
cu
r
ac
y
th
an
C
FNN.
As
in
C
FNN,
th
e
in
p
u
t
lay
er
is
co
n
n
ec
ted
to
m
o
r
e
l
ay
er
s
in
th
e
n
etwo
r
k
,
it
ten
d
s
to
o
v
er
f
it
r
ed
u
cin
g
th
e
o
v
er
all
r
ec
o
g
n
itio
n
ac
cu
r
a
cy
.
T
ab
le
1
.
C
o
m
p
a
r
ativ
e
r
esu
lts
with
s
tate
o
f
ar
t m
eth
o
d
s
S
t
a
t
e
o
f
a
r
t
me
t
h
o
d
s
U
C
F
s
p
o
r
t
s
U
T
i
n
t
e
r
a
c
t
i
o
n
S
t
a
t
e
o
f
a
r
t
me
t
h
o
d
s
W
e
i
z
ma
n
n
K
TH
C
a
r
v
a
j
a
l
e
t
a
l
.
[
2
5
]
8
8
.
6
--
C
h
a
a
r
a
o
u
i
e
t
a
l
.
[
2
6
]
9
2
.
2
8
%
9
6
.
7
0
Yi
a
n
d
Li
n
[
2
7
]
90
9
1
.
8
Ju
n
e
j
o
e
t
a
l
.
[
2
8
]
8
9
%
9
7
.
1
W
a
n
g
a
n
d
Q
i
[
2
9
]
92
8
3
.
3
C
h
i
v
e
r
s
[
3
0
]
9
7
%
9
6
.
9
W
e
n
g
a
n
d
G
u
a
n
[
3
1
]
9
2
.
8
5
8
.
2
S
i
d
d
i
q
i
e
t
a
l
.
[
1
1
]
8
1
%
8
0
.
3
3
C
h
o
e
t
a
l
.
[
3
2
]
8
9
.
7
8
5
%
H
.
N
a
v
e
e
d
e
t
a
l
.
[
3
3
]
9
1
.
6
9
%
9
2
.
2
8
N
a
z
i
r
e
t
a
l
.
[
3
4
]
94
--
M
.
F
.
A
sl
a
n
e
t
a
l
.
[
3
5
]
--
9
5
.
3
3
Ji
e
t
a
l
.
[
3
6
]
--
8
3
.
3
S
.
Ze
n
g
e
t
a
l
.
[
3
7
]
9
8
.
7
--
Th
i
s
W
o
r
k
w
i
t
h
F
F
N
N
9
0
.
3
5
9
7
.
5
9
Th
i
s
W
o
r
k
w
i
t
h
F
F
N
N
9
3
.
1
6
9
4
.
0
8
Th
i
s
w
o
r
k
w
i
t
h
C
F
N
N
8
8
.
8
5
9
5
.
2
Th
i
s
W
o
r
k
w
i
t
h
C
F
N
N
9
3
.
1
6
9
3
.
2
8
4.
CO
NCLU
SI
O
N
E
x
p
er
im
en
tal
r
esu
lts
s
h
o
w
th
at,
f
o
r
h
u
m
a
n
ac
tio
n
r
ec
o
g
n
it
io
n
,
FF
NN
g
iv
es
h
ig
h
e
r
ac
cu
r
ac
y
th
an
C
FNN.
I
t
is
o
b
s
er
v
ed
th
at
m
e
an
s
q
u
ar
e
er
r
o
r
r
e
d
u
ce
s
f
ast
f
o
r
tr
ain
in
g
d
atasets
b
u
t
s
tab
ilizes
at
a
h
ig
h
e
r
lev
el
f
o
r
test
an
d
v
alid
atio
n
d
atasets
in
th
e
ca
s
e
o
f
C
F
NN.
T
h
is
s
h
o
ws
th
at,
in
C
FNN,
o
v
er
f
itti
n
g
o
cc
u
r
s
b
ec
au
s
e
o
f
weig
h
ted
co
n
n
ec
tio
n
s
p
r
esen
t
b
etwe
en
th
e
in
p
u
t
la
y
er
a
n
d
all
h
id
d
e
n
lay
er
s
.
T
h
e
r
e
co
g
n
itio
n
ac
c
u
r
ac
y
ac
h
iev
ed
b
y
C
FNN
r
ed
u
ce
s
as c
o
m
p
ar
e
d
to
FF
NN
b
ec
au
s
e
o
f
o
v
er
f
itti
n
g
.
I
n
th
is
p
ap
er
,
a
f
u
s
io
n
o
f
HO
G
an
d
HOF
f
ea
tu
r
es
is
u
s
ed
to
d
escr
ib
e
h
u
m
an
ac
tio
n
s
.
HOG
an
d
HOF
f
ea
tu
r
es
ar
e
s
elec
ted
f
o
r
th
is
task
as
b
o
th
o
f
th
ese
ar
e
g
lo
b
al
f
ea
tu
r
es.
As
HOG
an
d
HOF
f
ea
tu
r
es
ar
e
ex
tr
ac
ted
f
r
o
m
th
e
f
r
am
e
as
a
wh
o
le,
th
e
r
e
q
u
ir
em
en
t
o
f
t
h
e
cr
u
cial
task
o
f
s
eg
m
e
n
ta
tio
n
an
d
f
o
r
e
g
r
o
u
n
d
ex
tr
ac
tio
n
is
elim
in
ated
.
A
c
o
m
b
in
atio
n
o
f
HOG,
wh
ich
g
iv
es
s
p
atial
in
f
o
r
m
atio
n
,
a
n
d
HOF
wh
ich
g
iv
es
m
o
tio
n
in
f
o
r
m
atio
n
,
f
o
r
m
a
s
t
r
o
n
g
f
ea
tu
r
e
d
escr
ip
to
r
.
T
h
e
r
ec
o
g
n
itio
n
ac
c
u
r
ac
y
will
v
ar
y
as
p
er
th
e
f
etu
r
es
s
elec
ted
f
o
r
r
ep
r
esen
tin
g
t
h
e
ac
tio
n
s
.
I
n
th
is
wo
r
k
as
t
h
e
f
o
cu
s
is
o
n
c
o
m
p
ar
is
o
n
o
f
n
eu
r
al
n
etwo
r
k
ar
ch
itectu
r
es,
v
ar
i
o
u
s
f
ea
tu
r
e
s
ar
e
n
o
t
e
x
p
lo
r
e
d
.
C
o
m
p
a
r
is
o
n
o
f
r
esu
lts
o
b
tain
ed
i
n
t
h
is
wo
r
k
with
th
e
p
r
ev
io
u
s
s
tate
o
f
ar
t
m
et
h
o
d
s
s
h
o
ws
th
at
f
o
r
W
eizm
an
n
an
d
KT
H
d
atasets
,
r
ec
o
g
n
itio
n
ac
cu
r
ac
y
o
b
tain
e
d
is
co
m
p
ar
ab
le
with
o
th
er
m
et
h
o
d
s
.
Fo
r
UC
F
s
p
o
r
ts
an
d
UT
in
ter
ac
tio
n
d
atasets
,
wh
ich
ar
e
m
o
r
e
c
o
m
p
lex
,
r
ec
o
g
n
itio
n
ac
cu
r
ac
y
o
u
t
p
er
f
o
r
m
s
o
th
er
m
eth
o
d
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
25
,
No
.
2
,
Feb
r
u
a
r
y
20
22
:
8
9
2
-
8
9
9
898
RE
F
E
R
E
NC
E
S
[
1
]
V
.
G
h
a
t
e
,
“
H
y
b
r
i
d
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
e
s
f
o
r
smar
t
p
h
o
n
e
s
e
n
s
o
r
-
b
a
se
d
h
u
m
a
n
a
c
t
i
v
i
t
y
r
e
c
o
g
n
i
t
i
o
n
,
”
Mu
l
t
i
m
e
d
.
T
o
o
l
s
A
p
p
l
.
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
1
0
4
2
-
0
2
0
-
1
0
4
7
8
-
4.
[
2
]
S
.
H
e
r
a
t
h
,
M
.
H
a
r
a
n
d
i
,
a
n
d
F
.
P
o
r
i
k
l
i
,
“
G
o
i
n
g
d
e
e
p
e
r
i
n
t
o
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
:
A
su
r
v
e
y
,
”
I
m
a
g
e
Vi
s
.
C
o
m
p
u
t
.
,
v
o
l
.
6
0
,
p
p
.
4
-
2
1
,
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
m
a
v
i
s.
2
0
1
7
.
0
1
.
0
1
0
.
[
3
]
K
.
A
n
u
r
a
d
h
a
a
n
d
N
.
S
a
i
r
a
m,
“
S
p
a
t
i
o
-
t
e
m
p
o
r
a
l
b
a
s
e
d
a
p
p
r
o
a
c
h
e
s
f
o
r
h
u
ma
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
i
n
s
t
a
t
i
c
a
n
d
d
y
n
a
mi
c
b
a
c
k
g
r
o
u
n
d
:
A
su
r
v
e
y
,
”
I
n
d
i
a
n
J
.
S
c
i
.
T
e
c
h
n
o
l
.
,
v
o
l
.
9
,
n
o
.
5
,
2
0
1
6
,
d
o
i
:
1
0
.
1
7
4
8
5
/
i
j
s
t
/
2
0
1
6
/
v
9
i
5
/
7
2
0
6
5
.
[
4
]
F
.
Zh
u
,
L
.
S
h
a
o
,
J.
X
i
e
,
a
n
d
Y
.
F
a
n
g
,
“
F
r
o
m
h
a
n
d
c
r
a
f
t
e
d
t
o
l
e
a
r
n
e
d
r
e
p
r
e
se
n
t
a
t
i
o
n
s
f
o
r
h
u
m
a
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
:
A
su
r
v
e
y
,
”
I
m
a
g
e
Vi
s.
C
o
m
p
u
t
.
,
v
o
l
.
5
5
,
p
p
.
4
2
-
5
2
,
2
0
1
6
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
m
a
v
i
s.
2
0
1
6
.
0
6
.
0
0
7
.
[
5
]
C
.
B
a
k
,
A
.
K
o
c
a
k
,
E
.
Er
d
e
m,
a
n
d
A
.
Er
d
e
m,
“
S
p
a
t
i
o
o
r
a
l
S
a
l
i
e
n
c
y
N
e
t
w
o
r
k
s
f
o
r
D
y
n
a
m
i
c
S
a
l
i
e
n
c
y
P
r
e
d
i
c
t
i
o
n
,
”
I
EEE
T
ra
n
s.
Mu
l
t
i
m
e
d
.
,
v
o
l
.
2
0
,
n
o
.
7
,
p
p
.
1
6
8
8
-
1
6
9
8
,
2
0
1
8
,
d
o
i
:
1
0
.
1
1
0
9
/
T
M
M
.
2
0
1
7
.
2
7
7
7
6
6
5
.
[
6
]
A
.
A
b
d
u
l
mu
n
e
m
,
Y
.
K
.
L
a
i
,
a
n
d
X
.
S
u
n
,
“
S
a
l
i
e
n
c
y
g
u
i
d
e
d
l
o
c
a
l
a
n
d
g
l
o
b
a
l
d
e
scr
i
p
t
o
r
s
f
o
r
e
f
f
e
c
t
i
v
e
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
,
”
C
o
m
p
u
t
.
Vi
s.
Me
d
i
a
,
v
o
l
.
2
,
n
o
.
1
,
p
p
.
9
7
-
1
0
6
,
2
0
1
6
,
d
o
i
:
1
0
.
1
0
0
7
/
s4
1
0
9
5
-
0
1
6
-
0
0
3
3
-
9.
[
7
]
A
.
A
b
d
u
l
m
u
n
e
m
,
Y
.
K
.
L
a
i
,
a
n
d
X
.
S
u
n
,
“
3
D
G
LO
H
f
e
a
t
u
r
e
s
f
o
r
h
u
ma
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
,
”
Pr
o
c
.
-
I
n
t
.
C
o
n
f
.
P
a
t
t
e
r
n
Re
c
o
g
n
i
t
.
,
p
p
.
8
0
5
-
8
1
0
,
2
0
1
6
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
P
R
.
2
0
1
6
.
7
8
9
9
7
3
4
.
[
8
]
I
.
C
.
D
u
t
a
,
B
.
I
o
n
e
sc
u
,
K
.
A
i
z
a
w
a
,
a
n
d
N
.
S
e
b
e
,
“
S
p
a
t
i
o
-
t
e
m
p
o
r
a
l
V
LA
D
e
n
c
o
d
i
n
g
f
o
r
h
u
ma
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
i
n
v
i
d
e
o
s,
”
L
e
c
t
.
N
o
t
e
s
C
o
m
p
u
t
.
S
c
i
.
(
i
n
c
l
u
d
i
n
g
S
u
b
ser.
L
e
c
t
.
N
o
t
e
s
A
rt
i
f
.
I
n
t
e
l
l
.
L
e
c
t
.
N
o
t
e
s B
i
o
i
n
f
o
rm
a
t
i
c
s)
,
v
o
l
.
1
0
1
3
2
LN
C
S
,
p
p
.
3
6
5
-
3
7
8
,
2
0
1
7
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
319
-
5
1
8
1
1
-
4
_
3
0
.
[
9
]
X
.
P
e
n
g
,
L.
W
a
n
g
,
X
.
W
a
n
g
,
a
n
d
Y
.
Q
i
a
o
,
“
B
a
g
o
f
v
i
s
u
a
l
w
o
r
d
s a
n
d
f
u
s
i
o
n
me
t
h
o
d
s
f
o
r
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
:
C
o
mp
r
e
h
e
n
si
v
e
s
t
u
d
y
a
n
d
g
o
o
d
p
r
a
c
t
i
c
e
,
”
C
o
m
p
u
t
.
Vi
s
.
I
m
a
g
e
U
n
d
e
rs
t
.
,
v
o
l
.
1
5
0
,
p
p
.
1
0
9
-
1
2
5
,
2
0
1
6
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
v
i
u
.
2
0
1
6
.
0
3
.
0
1
3
.
[
1
0
]
H
.
M
.
E
.
M
a
r
y
a
m
,
N
.
A
l
-
B
e
r
r
y
,
M
.
A.
-
M
.
S
a
l
e
m
,
A
.
S
.
H
.
a
n
d
M
.
F
.
To
l
b
a
,
“
A
c
t
i
o
n
C
l
a
ssi
f
i
c
a
t
i
o
n
U
si
n
g
W
e
i
g
h
t
e
d
D
i
r
e
c
t
i
o
n
a
l
W
a
v
e
l
e
t
LB
P
H
i
s
t
o
g
r
a
ms,”
i
n
A
d
v
a
n
c
e
s
i
n
I
n
t
e
l
l
i
g
e
n
t
S
y
s
t
e
m
s
a
n
d
C
o
m
p
u
t
i
n
g
,
2
0
1
6
,
v
o
l
.
4
0
7
,
p
p
.
1
5
-
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
3
1
9
-
2
6
6
9
0
-
9
_
2
.
[
1
1
]
M
.
H
.
S
i
d
d
i
q
i
,
R
.
A
l
i
,
M
.
S
.
R
a
n
a
,
E
.
K
.
H
o
n
g
,
E.
S
.
K
i
m,
a
n
d
S
.
Le
e
,
“
V
i
d
e
o
-
b
a
se
d
h
u
ma
n
a
c
t
i
v
i
t
y
r
e
c
o
g
n
i
t
i
o
n
u
si
n
g
m
u
l
t
i
l
e
v
e
l
w
a
v
e
l
e
t
d
e
c
o
m
p
o
si
t
i
o
n
a
n
d
st
e
p
w
i
se
l
i
n
e
a
r
d
i
s
c
r
i
mi
n
a
n
t
a
n
a
l
y
si
s,
”
S
e
n
s
o
rs
(
S
w
i
t
zer
l
a
n
d
)
,
v
o
l
.
1
4
,
n
o
.
4
,
p
p
.
6
3
7
0
-
6
3
9
2
,
2
0
1
4
,
d
o
i
:
1
0
.
3
3
9
0
/
s
1
4
0
4
0
6
3
7
0
.
[
1
2
]
M
.
A
l
-
B
e
r
r
y
,
A
.
-
M
.
M
o
h
a
mm
e
d
,
H
.
Eb
i
e
d
,
A
.
H
u
ss
e
i
n
,
a
n
d
M
.
T
o
l
b
a
,
“
W
e
i
g
h
t
e
d
D
i
r
e
c
t
i
o
n
a
l
3
D
S
t
a
t
i
o
n
a
r
y
W
a
v
e
l
e
t
-
b
a
se
d
A
c
t
i
o
n
C
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
Eg
y
p
t
.
C
o
m
p
u
t
.
S
c
i
.
J
.
,
v
o
l
.
3
9
,
n
o
.
2
,
p
p
.
8
3
-
9
7
,
2
0
1
5
.
[
1
3
]
M
.
N
.
A
l
-
B
e
r
r
y
,
H
.
M
.
E
b
i
e
d
,
A
.
S
.
H
u
sse
i
n
,
a
n
d
M
.
F
.
T
o
l
b
a
,
“
H
u
m
a
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
v
i
a
m
u
l
t
i
-
s
c
a
l
e
3
D
s
t
a
t
i
o
n
a
r
y
w
a
v
e
l
e
t
a
n
a
l
y
si
s
,
”
2
0
1
4
1
4
t
h
I
n
t
.
C
o
n
f
.
H
y
b
ri
d
I
n
t
e
l
l
.
S
y
s
t
.
H
I
S
2
0
1
4
,
p
p
.
2
5
4
-
2
5
9
,
2
0
1
4
,
d
o
i
:
1
0
.
1
1
0
9
/
H
I
S
.
2
0
1
4
.
7
0
8
6
2
0
8
.
[
1
4
]
L.
P
.
S
u
r
e
s
h
,
S
.
S
.
D
a
s
h
,
a
n
d
B
.
K
.
P
a
n
i
g
r
a
h
i
,
“
A
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
a
n
d
e
v
o
l
u
t
i
o
n
a
r
y
a
l
g
o
r
i
t
h
ms
i
n
e
n
g
i
n
e
e
r
i
n
g
sy
s
t
e
ms
:
P
r
o
c
e
e
d
i
n
g
s
o
f
I
C
A
EES
2
0
1
4
,
v
ol
.
2
,
”
A
d
v
.
I
n
t
e
l
l
.
S
y
st
.
C
o
m
p
u
t
.
,
v
o
l
.
3
2
5
,
p
p
.
3
0
9
-
3
1
6
,
2
0
1
5
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
81
-
3
2
2
-
2
1
3
5
-
7.
[
1
5
]
E.
Jy
o
t
sn
a
,
P
.
V
A
k
h
i
l
,
a
n
d
A
.
K
u
mar,
“
S
i
l
h
o
u
e
t
t
e
b
a
se
d
h
u
ma
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
u
si
n
g
P
C
A
a
n
d
I
S
O
M
A
P
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
Ad
v
a
n
c
e
d
R
e
se
a
rc
h
i
n
C
o
m
p
u
t
e
r
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
2
,
n
o
.
1
1
,
p
p
.
4
1
9
2
-
4
1
9
8
,
2
0
1
3
.
[
1
6
]
R
.
C
.
N
g
,
K
.
M
.
Li
m
,
C
.
P
.
Le
e
,
a
n
d
S
.
F
.
A
.
R
a
z
a
k
,
“
S
u
r
v
e
i
l
l
a
n
c
e
s
y
st
e
m
w
i
t
h
m
o
t
i
o
n
a
n
d
f
a
c
e
d
e
t
e
c
t
i
o
n
u
si
n
g
h
i
s
t
o
g
r
a
ms
o
f
o
r
i
e
n
t
e
d
g
r
a
d
i
e
n
t
s,”
I
n
d
o
n
e
si
a
n
J
o
u
r
n
a
l
o
f
E
l
e
c
t
ri
c
a
l
En
g
i
n
e
e
r
i
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
I
J
EE
C
S
)
,
v
o
l
.
1
4
,
n
o
.
2
,
p
p
.
8
6
9
-
8
7
6
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s.
v
1
4
.
i
2
.
p
p
8
6
9
-
8
7
6
.
[
1
7
]
A
.
S
.
Jah
a
g
i
r
d
a
r
a
n
d
M
.
S
.
N
a
g
m
o
d
e
,
“
A
N
o
v
e
l
H
u
ma
n
A
c
t
i
o
n
R
e
c
o
g
n
i
t
i
o
n
a
n
d
B
e
h
a
v
i
o
u
r
A
n
a
l
y
s
i
s
Te
c
h
n
i
q
u
e
u
s
i
n
g
S
W
F
H
O
G
,
”
I
n
t
.
J
.
A
d
v
.
C
o
m
p
u
t
.
S
c
i
.
Ap
p
l
.
,
v
o
l
.
1
1
,
n
o
.
4
,
p
p
.
5
7
1
-
5
8
0
,
2
0
2
0
,
d
o
i
:
1
0
.
1
4
5
6
9
/
I
JA
C
S
A
.
2
0
2
0
.
0
1
1
0
4
7
5
.
[
1
8
]
J.
P
.
Te
i
x
e
i
r
a
a
n
d
P
.
O
.
F
e
r
n
a
n
d
e
s,
“
C
o
m
p
a
r
i
so
n
o
f
A
r
t
i
f
i
c
i
a
l
N
e
u
r
a
l
N
e
t
w
o
r
k
A
r
c
h
i
t
e
c
t
u
r
e
s
i
n
t
h
e
Ta
sk
o
f
To
u
r
i
sm
T
i
me
S
e
r
i
e
s
F
o
r
e
c
a
st
,
”
I
n
t
.
J
.
C
o
m
p
u
t
.
I
n
f
.
E
n
g
.
,
v
o
l
.
6
,
n
o
.
6
,
p
p
.
8
3
0
-
8
3
5
,
2
0
1
2
,
d
o
i
:
1
0
.
5
2
8
1
/
z
e
n
o
d
o
.
1
0
7
2
8
9
4
.
[
1
9
]
A
.
Jo
h
n
D
h
a
n
a
se
e
l
y
,
S
.
H
i
ma
v
a
t
h
i
,
a
n
d
E.
S
r
i
n
i
v
a
s
a
n
,
“
P
e
r
f
o
r
ma
n
c
e
c
o
m
p
a
r
i
so
n
o
f
c
a
s
c
a
d
e
a
n
d
f
e
e
d
f
o
r
w
a
r
d
n
e
u
r
a
l
n
e
t
w
o
r
k
f
o
r
f
a
c
e
r
e
c
o
g
n
i
t
i
o
n
s
y
st
e
m,
”
I
ET
S
e
m
i
n
.
D
i
g
.
,
v
o
l
.
2
0
1
2
,
n
o
.
4
,
2
0
1
2
,
d
o
i
:
1
0
.
1
0
4
9
/
i
c
.
2
0
1
2
.
0
1
5
4
.
[
2
0
]
D
.
S
.
B
a
d
d
e
,
A
.
G
u
p
t
a
,
a
n
d
V
.
K
.
P
a
t
k
i
,
“
C
a
sca
d
e
a
n
d
F
e
e
d
F
o
r
w
a
r
d
B
a
c
k
p
r
o
p
a
g
a
t
i
o
n
A
r
t
i
f
i
c
i
a
l
N
e
u
r
a
l
N
e
t
w
o
r
k
M
o
d
e
l
s
f
o
r
P
r
e
d
i
c
t
i
o
n
o
f
C
o
m
p
r
e
ss
i
v
e
S
t
r
e
n
g
t
h
o
f
R
e
a
d
y
M
i
x
C
o
n
c
r
e
t
e
,
”
I
O
S
R
J
.
M
e
c
h
.
C
i
v
.
En
g
.
,
n
o
.
2
2
7
8
-
1
6
8
4
,
p
p
.
1
-
6
,
2
0
0
9
.
[
2
1
]
S
.
G
o
y
a
l
a
n
d
G
.
K
.
G
o
y
a
l
,
“
C
a
s
c
a
d
e
a
n
d
F
e
e
d
f
o
r
w
a
r
d
B
a
c
k
p
r
o
p
a
g
a
t
i
o
n
A
r
t
i
f
i
c
i
a
l
N
e
u
r
a
l
N
e
t
w
o
r
k
M
o
d
e
l
s
F
o
r
P
r
e
d
i
c
t
i
o
n
o
f
S
e
n
s
o
r
y
Q
u
a
l
i
t
y
o
f
I
n
st
a
n
t
C
o
f
f
e
e
F
l
a
v
o
u
r
e
d
S
t
e
r
i
l
i
z
e
d
D
r
i
n
k
,
”
C
a
n
.
J
.
Art
i
f
.
I
n
t
e
l
l
.
Ma
c
h
.
L
e
a
r
n
.
Pa
t
t
e
r
n
R
e
c
o
g
n
i
t
.
,
v
o
l
.
2
,
n
o
.
6
,
p
p
.
7
8
-
8
2
,
2
0
1
1
.
[
2
2
]
N
.
D
a
l
a
l
a
n
d
B
.
Tr
i
g
g
s
,
“
H
i
st
o
g
r
a
ms
o
f
O
r
i
e
n
t
e
d
G
r
a
d
i
e
n
t
s
f
o
r
H
u
ma
n
D
e
t
e
c
t
i
o
n
,
”
I
EE
E
C
o
m
p
u
t
.
S
o
c
.
C
o
n
f
.
C
o
m
p
u
t
.
V
i
s.
P
a
t
t
e
r
n
Re
c
o
g
n
i
t
.
,
2
0
0
5
,
p
p
.
8
8
6
-
8
9
3
,
d
o
i
:
1
0
.
1
1
0
9
/
C
V
P
R
.
2
0
0
5
.
1
7
7
.
[
2
3
]
J.
P
e
r
š,
V
.
S
u
l
i
ć
,
M
.
K
r
i
st
a
n
,
M
.
P
e
r
še,
K
.
P
o
l
a
n
e
c
,
a
n
d
S
.
K
o
v
a
č
i
č
,
“
H
i
s
t
o
g
r
a
ms
o
f
o
p
t
i
c
a
l
f
l
o
w
f
o
r
e
f
f
i
c
i
e
n
t
r
e
p
r
e
s
e
n
t
a
t
i
o
n
o
f
b
o
d
y
m
o
t
i
o
n
,
”
P
a
t
t
e
r
n
Re
c
o
g
n
i
t
.
L
e
t
t
.
,
v
o
l
.
3
1
,
n
o
.
1
1
,
p
p
.
1
3
6
9
-
1
3
7
6
,
2
0
1
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
a
t
r
e
c
.
2
0
1
0
.
0
3
.
0
2
4
.
[
2
4
]
A
.
S
.
Ja
h
a
g
i
r
d
a
r
a
n
d
M
.
S
.
N
a
g
m
o
d
e
,
"
S
i
l
h
o
u
e
t
t
e
-
b
a
s
e
d
h
u
ma
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
b
y
e
m
b
e
d
d
i
n
g
H
O
G
a
n
d
P
C
A
f
e
a
t
u
r
e
s,
"
2
n
d
S
p
r
i
n
g
e
r
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
I
n
t
e
l
l
i
g
e
n
t
C
o
m
p
u
t
i
n
g
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
–
I
C
I
C
C
2
0
1
7
A
t
:
P
u
n
e
,
I
n
d
i
a
,
v
o
l
.
6
7
3
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
9
8
1
-
10
-
7
2
4
5
-
1
.
[
2
5
]
J.
C
a
r
v
a
j
a
l
,
A
.
W
i
l
i
e
m,
C
.
M
c
C
o
o
l
,
B
.
L
o
v
e
l
l
,
a
n
d
C
.
S
a
n
d
e
r
so
n
,
“
C
o
m
p
a
r
a
t
i
v
e
e
v
a
l
u
a
t
i
o
n
o
f
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
me
t
h
o
d
s
v
i
a
r
i
e
ma
n
n
i
a
n
ma
n
i
f
o
l
d
s
,
f
i
sh
e
r
v
e
c
t
o
r
s
a
n
d
G
M
M
s:
I
d
e
a
l
a
n
d
c
h
a
l
l
e
n
g
i
n
g
c
o
n
d
i
t
i
o
n
s,
”
L
e
c
t
.
N
o
t
e
s
C
o
m
p
u
t
.
S
c
i
.
(
i
n
c
l
u
d
i
n
g
S
u
b
s
e
r.
L
e
c
t
.
N
o
t
e
s
Art
i
f
.
I
n
t
e
l
l
.
L
e
c
t
.
N
o
t
e
s B
i
o
i
n
f
o
rm
a
t
i
c
s)
,
v
o
l
.
9
7
9
4
,
p
p
.
8
8
-
1
0
0
,
2
0
1
6
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
3
1
9
-
4
2
9
9
6
-
0
_
8
.
[
2
6
]
A
.
A
.
C
h
a
a
r
a
o
u
i
,
P
.
C
l
i
me
n
t
-
P
é
r
e
z
,
a
n
d
F
.
F
l
ó
r
e
z
-
R
e
v
u
e
l
t
a
,
“
S
i
l
h
o
u
e
t
t
e
-
b
a
sed
h
u
ma
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
u
s
i
n
g
s
e
q
u
e
n
c
e
s
o
f
k
e
y
p
o
s
e
s,
”
P
a
t
t
e
r
n
Re
c
o
g
n
i
t
.
L
e
t
t
.
,
v
o
l
.
3
4
,
n
o
.
1
5
,
p
p
.
1
7
9
9
-
1
8
0
7
,
2
0
1
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
a
t
r
e
c
.
2
0
1
3
.
0
1
.
0
2
1
.
[
2
7
]
Y
.
Y
i
a
n
d
Y
.
L
i
n
,
“
H
u
ma
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
w
i
t
h
s
a
l
i
e
n
t
t
r
a
j
e
c
t
o
r
i
e
s,
”
S
i
g
n
a
l
Pr
o
c
e
ssi
n
g
,
v
o
l
.
9
3
,
n
o
.
1
1
,
p
p
.
2
9
3
2
-
2
9
4
1
,
2
0
1
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
s
i
g
p
r
o
.
2
0
1
3
.
0
5
.
0
0
2
.
[
2
8
]
I
.
N
.
J
u
n
e
j
o
,
K
.
N
.
J
u
n
e
j
o
,
a
n
d
Z.
A
l
A
g
h
b
a
r
i
,
“
S
i
l
h
o
u
e
t
t
e
-
b
a
se
d
h
u
m
a
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
u
si
n
g
S
A
X
-
S
h
a
p
e
s,”
Vi
s
.
C
o
m
p
u
t
.
,
v
o
l
.
3
0
,
n
o
.
3
,
p
p
.
2
5
9
-
2
6
9
,
2
0
1
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
3
7
1
-
0
1
3
-
0
8
4
2
-
0.
[
2
9
]
X
.
W
a
n
g
a
n
d
C
.
Q
i
,
“
A
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
u
si
n
g
e
d
g
e
t
r
a
j
e
c
t
o
r
i
e
s
a
n
d
m
o
t
i
o
n
a
c
c
e
l
e
r
a
t
i
o
n
d
e
scri
p
t
o
r
,
”
Ma
c
h
.
V
i
s.
A
p
p
l
.
,
v
o
l
.
2
7
,
n
o
.
6
,
p
p
.
8
6
1
-
8
7
5
,
2
0
1
6
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
1
3
8
-
0
1
6
-
0
7
4
6
-
x.
[
3
0
]
D
.
S
.
C
h
i
v
e
r
s,
“
H
u
ma
n
A
c
t
i
o
n
R
e
c
o
g
n
i
t
i
o
n
b
y
P
r
i
n
c
i
p
a
l
C
o
m
p
o
n
e
n
t
A
n
a
l
y
s
i
s
o
f
M
o
t
i
o
n
C
u
r
v
e
s
H
u
m
a
n
A
c
t
i
o
n
R
e
c
o
g
n
i
t
i
o
n
b
y
P
r
i
n
c
i
p
a
l
C
o
m
p
o
n
e
n
t
A
n
a
l
y
si
s
o
f
M
o
t
i
o
n
C
u
r
v
e
s
A
d
i
ss
e
r
t
a
t
i
o
n
su
b
mi
t
t
e
d
i
n
p
a
r
t
i
a
l
f
u
l
f
i
l
l
m
e
n
t
b
y
,
”
Br
o
w
s
e
a
l
l
T
h
e
ses
a
n
d
D
i
ssert
a
t
i
o
n
s
,
p
.
6
4
2
,
2
0
1
2
.
[
3
1
]
Z.
W
e
n
g
a
n
d
Y
.
G
u
a
n
,
“
A
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
u
si
n
g
l
e
n
g
t
h
-
v
a
r
i
a
b
l
e
e
d
g
e
t
r
a
j
e
c
t
o
r
y
a
n
d
sp
a
t
i
o
-
t
e
m
p
o
r
a
l
mo
t
i
o
n
s
k
e
l
e
t
o
n
d
e
scri
p
t
o
r
,
”
Eu
r
a
si
p
J
.
I
m
a
g
e
V
i
d
e
o
Pr
o
c
e
ss.
,
v
o
l
.
2
0
1
8
,
n
o
.
1
,
2
0
1
8
,
d
o
i
:
1
0
.
1
1
8
6
/
s1
3
6
4
0
-
0
1
8
-
0
2
5
0
-
5.
[
3
2
]
J.
C
h
o
,
M
.
L
e
e
,
H
.
J
.
C
h
a
n
g
,
a
n
d
S
.
O
h
,
“
R
o
b
u
s
t
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
u
s
i
n
g
l
o
c
a
l
m
o
t
i
o
n
a
n
d
g
r
o
u
p
sp
a
r
si
t
y
,
”
P
a
t
t
e
r
n
R
e
c
o
g
n
i
t
.
,
v
o
l
.
4
7
,
n
o
.
5
,
p
p
.
1
8
1
3
-
1
8
2
5
,
2
0
1
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
a
t
c
o
g
.
2
0
1
3
.
1
2
.
0
0
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
C
o
mp
a
r
is
o
n
o
f fe
ed
fo
r
w
a
r
d
a
n
d
ca
s
ca
d
e
fo
r
w
a
r
d
n
eu
r
a
l n
et
w
o
r
ks fo
r
h
u
ma
n
a
ctio
n
…
(
A
d
iti Ja
h
a
g
ir
d
a
r
)
899
[
3
3
]
H
.
N
a
v
e
e
d
,
G
.
K
h
a
n
,
A
.
U
.
K
h
a
n
,
A
.
S
i
d
d
i
q
i
,
a
n
d
M
.
U
.
G
.
K
h
a
n
,
“
H
u
m
a
n
a
c
t
i
v
i
t
y
r
e
c
o
g
n
i
t
i
o
n
u
s
i
n
g
mi
x
t
u
r
e
o
f
h
e
t
e
r
o
g
e
n
e
o
u
s
f
e
a
t
u
r
e
s
a
n
d
se
q
u
e
n
t
i
a
l
m
i
n
i
m
a
l
o
p
t
i
m
i
z
a
t
i
o
n
,
”
I
n
t
.
J
.
Ma
c
h
.
L
e
a
r
n
.
C
y
b
e
r
n
.
,
v
o
l
.
1
0
,
n
o
.
9
,
p
p
.
2
3
2
9
-
2
3
4
0
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
0
4
2
-
0
1
8
-
0
8
7
0
-
1.
[
3
4
]
S
.
N
a
z
i
r
,
M
.
H
.
Y
o
u
saf,
a
n
d
S
.
A
.
V
e
l
a
s
t
i
n
,
“
E
v
a
l
u
a
t
i
n
g
a
b
a
g
-
of
-
v
i
s
u
a
l
f
e
a
t
u
r
e
s
a
p
p
r
o
a
c
h
u
si
n
g
s
p
a
t
i
o
-
t
e
m
p
o
r
a
l
f
e
a
t
u
r
e
s
f
o
r
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
,
”
C
o
m
p
u
t
.
El
e
c
t
r
.
En
g
.
,
p
p
.
6
6
0
-
6
6
9
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
e
l
e
c
e
n
g
.
2
0
1
8
.
0
1
.
0
3
7
.
[
3
5
]
M
.
F
.
A
sl
a
n
,
A
.
D
u
r
d
u
,
a
n
d
K
.
S
a
b
a
n
c
i
,
“
H
u
ma
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
w
i
t
h
b
a
g
o
f
v
i
s
u
a
l
w
o
r
d
s
u
s
i
n
g
d
i
f
f
e
r
e
n
t
mac
h
i
n
e
l
e
a
r
n
i
n
g
met
h
o
d
s
a
n
d
h
y
p
e
r
p
a
r
a
me
t
e
r
o
p
t
i
mi
z
a
t
i
o
n
,
”
N
e
u
r
a
l
C
o
m
p
u
t
.
Ap
p
l
.
,
v
o
l
.
9
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
0
7
/
s0
0
5
2
1
-
0
1
9
-
0
4
3
6
5
-
9.
[
3
6
]
X
.
Ji
,
C
.
W
a
n
g
,
a
n
d
Z
.
Ju
,
“
A
n
e
w
f
r
a
mew
o
r
k
o
f
h
u
ma
n
i
n
t
e
r
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
b
a
s
e
d
o
n
m
u
l
t
i
p
l
e
st
a
g
e
p
r
o
b
a
b
i
l
i
t
y
f
u
si
o
n
,
”
Ap
p
l
.
S
c
i
.
,
v
o
l
.
7
,
n
o
.
6
,
2
0
1
7
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
7
0
6
0
5
6
7
.
[
3
7
]
S
.
Ze
n
g
,
G
.
L
u
,
a
n
d
P
.
Y
a
n
,
“
En
h
a
n
c
i
n
g
h
u
ma
n
a
c
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
v
i
a
st
r
u
c
t
u
r
a
l
a
v
e
r
a
g
e
c
u
r
v
e
s
a
n
a
l
y
s
i
s,”
S
i
g
n
a
l
,
I
m
a
g
e
Vi
d
e
o
Pro
c
e
ss.
,
v
o
l
.
1
2
,
n
o
.
8
,
p
p
.
1
5
5
1
-
1
5
5
8
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
7
6
0
-
0
1
8
-
1
3
1
1
-
z.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Dr
.
Adi
ti
J
a
h
a
g
ird
a
r
re
c
e
iv
e
d
h
e
r
Ba
c
h
e
lo
r‟
s
d
e
g
re
e
i
n
1
9
9
3
in
El
e
c
tro
n
ics
&
Tele
c
o
m
m
u
n
ica
ti
o
n
En
g
in
e
e
ri
n
g
a
n
d
a
M
a
ste
r‟
s
d
e
g
re
e
in
El
e
c
tro
n
ics
En
g
in
e
e
ri
n
g
in
1
9
9
9
fr
o
m
th
e
Un
iv
e
rsit
y
o
f
P
u
n
e
.
S
h
e
c
o
m
p
lete
d
h
e
r
P
h
.
D.
in
t
h
e
d
o
m
a
in
o
f
c
o
m
p
u
ter
v
isi
o
n
i
n
2
0
2
1
fro
m
S
a
v
it
rib
a
i
P
h
u
le
P
u
n
e
U
n
iv
e
rsi
ty
,
P
u
n
e
,
M
a
h
a
ra
sh
tra,
I
n
d
ia.
S
h
e
is
c
u
rre
n
tl
y
wo
rk
in
g
a
s
As
sista
n
t
P
ro
fe
ss
o
r
in
th
e
S
c
h
o
o
l
o
f
C
o
m
p
u
ter
En
g
in
e
e
rin
g
a
n
d
Tec
h
n
o
lo
g
y
,
a
t
M
IT
W
o
rl
d
P
e
a
c
e
Un
iv
e
rsity
,
P
u
n
e
.
S
h
e
is
h
a
v
in
g
a
t
o
tal
o
f
2
6
y
e
a
rs
o
f
tea
c
h
in
g
e
x
p
e
rie
n
c
e
.
S
h
e
is
a
li
fe
m
e
m
b
e
r
o
f
th
e
Co
m
p
u
ter
S
o
c
iety
o
f
In
d
ia
a
n
d
I
n
stit
u
ti
o
n
o
f
El
e
c
tr
o
n
ics
a
n
d
Tele
c
o
m
m
u
n
ica
ti
o
n
En
g
i
n
e
e
rs.
He
r
re
se
a
r
c
h
in
tere
sts
in
c
lu
d
e
c
o
m
p
u
ter
v
isi
o
n
,
Im
a
g
e
p
ro
c
e
ss
in
g
,
v
i
d
e
o
p
r
o
c
e
ss
in
g
,
a
n
d
p
a
tt
e
r
n
re
c
o
g
n
it
i
o
n
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
a
d
it
i.
ja
h
a
g
i
rd
a
r@m
it
wp
u
.
e
d
u
.
i
n
.
Dr
.
Ra
shm
i
P
h
a
l
n
ik
a
r
is
a
n
As
so
c
iate
P
r
o
fe
ss
o
r
i
n
th
e
S
c
h
o
o
l
o
f
C
o
m
p
u
ter
En
g
i
n
e
e
rin
g
a
n
d
Tec
h
n
o
lo
g
y
a
t
M
IT
Wo
r
ld
P
e
a
c
e
Un
iv
e
rsit
y
,
P
u
n
e
wh
e
re
sh
e
h
a
s
b
e
e
n
a
fa
c
u
lt
y
m
e
m
b
e
r
sin
c
e
2
0
0
8
.
S
h
e
h
a
s
c
o
m
p
lete
d
h
e
r
P
h
.
D.
(S
o
ftwa
re
En
g
in
e
e
rin
g
)
fro
m
S
a
rd
a
r
Va
ll
a
b
h
a
i
Na
ti
o
n
a
l
In
st
it
u
te
o
f
Tec
h
n
o
l
o
g
y
,
S
u
ra
t,
I
n
d
ia.
He
r
re
se
a
rc
h
t
o
p
ic
wa
s
“
A
F
ra
m
e
wo
rk
fo
r
An
a
ly
sis
o
f
Co
m
p
le
x
Us
e
r
Re
q
u
irem
e
n
ts
u
sin
g
As
p
e
c
t
Orie
n
ted
Us
e
Ca
s
e
M
e
th
o
d
a
n
d
G
ra
p
h
Th
e
o
ry
”
.
He
r
p
u
b
li
sh
e
d
wo
r
k
t
o
d
a
te,
(m
o
re
t
h
a
n
5
0
)
i
n
re
p
u
ted
C
o
n
fe
re
n
c
e
s
a
n
d
Jo
u
rn
a
ls
c
o
v
e
r
th
e
th
e
o
re
ti
c
a
l
a
n
d
e
x
p
e
rime
n
tal
wo
rk
i
n
t
h
e
d
o
m
a
in
re
late
d
t
o
S
o
ftwa
re
En
g
i
n
e
e
rin
g
,
R
o
le
o
f
No
n
F
u
n
c
ti
o
n
a
l
Re
q
u
irem
e
n
ts,
A
p
p
li
c
a
ti
o
n
Are
a
s
o
f
Da
ta
M
i
n
i
n
g
a
n
d
Da
ta
A
n
a
ly
sis.
S
h
e
is
a
Li
fe
M
e
m
b
e
r
o
f
I
n
d
ia
n
S
o
c
iety
f
o
r
Tec
h
n
ica
l
E
d
u
c
a
ti
o
n
(
IS
TE
)
a
n
d
Co
m
p
u
ter
S
o
c
iety
o
f
In
d
ia
(CS
I).
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
ra
sh
m
i.
p
h
a
l
n
ik
a
r@m
it
wp
u
.
e
d
u
.
in
.
Evaluation Warning : The document was created with Spire.PDF for Python.