I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
201
9
,
p
p
.
5
2
7
0
~
5
2
7
6
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
9
i
6
.
p
p
5
2
7
0
-
5276
5270
J
o
ur
na
l
ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
co
r
e
.
co
m/
jo
u
r
n
a
ls
/in
d
ex
.
p
h
p
/
I
JE
C
E
H
u
m
a
n activ
i
ty r
ecog
nition
by
usi
ng
co
nv
o
lutiona
l
neura
l net
w
o
rk
H
a
nk
i
l
K
i
m
1
,
Su
ng
o
ck
L
ee
2
,
H
o
ek
y
un
g
J
un
g
3
1
De
p
a
rtme
n
t
o
f
M
u
sic
&
S
o
u
n
d
T
e
c
h
n
o
l
o
g
y
,
Ko
re
a
Un
iv
e
rsit
y
o
f
M
e
d
ia A
rts
,
S
o
u
t
h
Ko
re
a
2
G
lo
b
a
l
HRD
In
c
.
3
4
8
3
8
,
S
o
u
t
h
Ko
e
a
3
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter E
n
g
in
e
e
rin
g
,
P
a
iCh
a
i
Un
iv
e
rsit
y
,
S
o
u
t
h
Ko
re
a
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
A
p
r
2
4
,
2
0
1
9
R
ev
i
s
ed
J
u
l 9
,
2
0
1
9
A
cc
ep
ted
J
u
l 1
7
,
2
0
1
9
In
re
c
e
n
t
y
e
a
rs,
m
a
n
y
re
se
a
rc
h
e
r
s
h
a
v
e
stu
d
ied
t
h
e
HA
R
(Hu
m
a
n
A
c
ti
v
it
y
Re
c
o
g
n
it
io
n
)
sy
ste
m
.
H
A
R
u
sin
g
s
m
a
rt
h
o
m
e
se
n
so
r
is
b
a
se
d
o
n
c
o
m
p
u
ti
n
g
in
sm
a
rt
e
n
v
iro
n
m
e
n
t,
a
n
d
in
tell
i
g
e
n
t
su
rv
e
il
lan
c
e
sy
st
e
m
c
o
n
d
u
c
t
s
in
ten
siv
e
re
se
a
rc
h
o
n
p
e
rip
h
e
ra
l
su
p
p
o
rt
l
if
e
.
T
h
e
p
re
v
io
u
s
sy
ste
m
st
u
d
ied
i
n
so
m
e
o
f
th
e
a
c
ti
v
it
ies
is
a
f
ix
e
d
m
o
ti
o
n
a
n
d
t
h
e
m
e
th
o
d
o
lo
g
y
is
les
s
a
c
c
u
ra
te.
In
th
is
p
a
p
e
r,
v
isio
n
-
b
a
se
d
stu
d
ies
u
s
in
g
th
e
rm
a
l
i
m
a
g
in
g
c
a
m
e
ra
s
im
p
ro
v
e
th
e
a
c
c
u
ra
c
y
o
f
m
o
ti
o
n
re
c
o
g
n
it
io
n
i
n
in
tell
ig
e
n
t
su
rv
e
il
lan
c
e
s
y
ste
m
s.
W
e
u
se
o
n
e
o
f
th
e
d
e
e
p
lea
rn
in
g
a
rc
h
it
e
c
tu
re
s
w
id
e
ly
u
se
d
in
im
a
g
e
re
c
o
g
n
it
io
n
sy
ste
m
s
c
a
ll
e
d
Co
n
v
o
l
u
ti
o
n
a
l
Ne
u
ra
l
Ne
tw
o
rk
s
(CNN
).
T
h
e
re
f
o
re
,
w
e
u
se
CNN
a
n
d
th
e
r
m
a
l
c
a
m
e
r
a
s
to
p
r
o
v
id
e
a
c
c
u
ra
c
y
a
n
d
m
a
n
y
f
e
a
tu
re
s th
ro
u
g
h
th
e
p
ro
p
o
se
d
m
e
th
o
d
.
K
ey
w
o
r
d
s
:
A
cti
v
it
y
R
ec
o
g
n
itio
n
C
o
n
v
o
lu
tio
n
al
Ne
u
r
al
Net
w
o
r
k
Dee
p
L
ea
r
n
i
n
g
I
n
telli
g
en
t Su
r
v
eilla
n
ce
S
y
s
te
m
T
h
er
m
al
C
a
m
er
a
Co
p
y
rig
h
t
©
2
0
1
9
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Ho
ek
y
u
n
g
J
u
n
g
,
Dep
ar
te
m
en
t
o
f
C
o
m
p
u
ter
E
n
g
in
ee
r
i
n
g
,
P
aich
ai
Un
i
v
er
s
it
y
,
155
-
4
0
B
ae
j
ae
-
r
o
,
Seo
g
u
,
Dae
J
eo
n
,
So
u
th
Ko
r
ea
.
E
m
ail:
h
k
j
u
n
g
@
p
c
u
.
ac
.
k
r
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
HAR
s
y
s
te
m
,
a
w
id
el
y
u
s
ed
p
atter
n
r
ec
o
g
n
i
tio
n
s
y
s
te
m
[1
-
3]
,
ca
n
b
e
d
iv
id
ed
in
to
s
ev
er
a
l
m
o
d
u
les
s
u
c
h
as
s
e
n
s
i
n
g
,
f
e
atu
r
e
ex
tr
ac
tio
n
cla
s
s
i
f
icatio
n
,
s
eg
m
e
n
tatio
n
a
n
d
p
o
s
t
-
p
r
o
ce
s
s
in
g
[
4
]
.
HA
R
s
y
s
te
m
s
ca
n
b
e
ca
teg
o
r
ized
in
to
t
w
o
t
y
p
es:
ti
m
e
-
b
ased
an
d
ac
ce
ler
atio
n
-
b
ased
.
A
cc
eler
ati
o
n
-
b
ased
m
et
h
o
d
s
r
eq
u
ir
e
m
u
ltip
le
ac
ce
ler
o
m
et
er
s
to
b
e
u
s
ed
f
o
r
d
ata
co
lle
ctio
n
,
b
u
t
ti
m
e
-
b
a
s
ed
m
et
h
o
d
s
t
y
p
ica
ll
y
r
eq
u
ir
e
th
e
u
s
e
o
f
o
n
e
o
r
m
o
r
e
ca
m
e
r
as
to
co
llect
d
ata.
T
h
e
d
is
ad
v
an
ta
g
e
o
f
th
e
ac
ce
ler
atio
n
m
eth
o
d
is
th
a
t
it
ca
n
ca
u
s
e
d
is
co
m
f
o
r
t to
th
e
h
u
m
a
n
b
o
d
y
w
h
e
n
p
er
f
o
r
m
i
n
g
ac
tiv
it
ies s
u
ch
as
w
a
lk
i
n
g
,
r
u
n
n
i
n
g
,
an
d
l
y
i
n
g
d
o
w
n
.
Ho
w
e
v
er
,
th
e
v
ar
io
u
s
h
u
m
a
n
ac
tiv
itie
s
to
b
e
m
o
n
ito
r
ed
in
t
h
is
s
tu
d
y
in
cl
u
d
e
h
a
n
d
w
a
v
i
n
g
,
p
u
n
ch
i
n
g
,
k
ick
in
g
,
l
y
in
g
d
o
w
n
,
w
al
k
in
g
,
r
u
n
n
i
n
g
,
an
d
s
ta
n
d
in
g
.
T
h
e
ad
v
an
tag
e
o
f
a
v
is
io
n
-
b
ased
s
y
s
te
m
i
s
t
h
at
th
e
s
e
n
s
o
r
w
o
r
k
s
w
it
h
o
u
t
s
ti
ck
in
g
to
th
e
b
o
d
y
.
Ho
w
e
v
er
,
r
ec
o
g
n
itio
n
p
er
f
o
r
m
a
n
ce
d
ep
en
d
s
o
n
lig
h
ti
n
g
co
n
d
itio
n
s
,
v
ie
w
in
g
an
g
le,
an
d
o
th
er
f
ac
to
r
s
.
I
n
th
is
p
ap
er
,
w
e
p
r
o
p
o
s
e
a
s
y
s
te
m
th
at
u
s
e
s
a
tim
e
-
b
ased
d
ata
s
et
[
5
-
8
]
ca
p
tu
r
ed
b
y
a
th
er
m
al
ca
m
er
a
[
9
,
1
0
]
an
d
a
C
NN
s
tr
u
ct
u
r
e
[
11
-
14
]
to
s
o
lv
e
th
is
p
r
o
b
lem
.
T
h
is
ca
n
r
ed
u
ce
th
e
p
r
o
ce
d
u
r
es o
f
th
e
h
an
d
icr
af
t p
r
o
ce
s
s
an
d
i
n
cr
ea
s
e
th
e
ac
cu
r
ac
y
.
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
AND
F
E
AT
U
RE
E
XT
RAC
T
I
O
N
E
x
p
lain
i
n
g
T
h
is
s
ec
tio
n
i
n
tr
o
d
u
ce
s
th
e
p
r
o
p
o
s
ed
m
et
h
o
d
an
d
f
ea
t
u
r
e
ex
tr
ac
tio
n
.
Firs
t,
it
ex
p
lain
s
h
o
w
to
p
er
f
o
r
m
th
e
f
u
n
ct
io
n
ex
tr
ac
tio
n
s
tep
-
by
-
s
tep
.
T
h
e
f
ir
s
t
s
tep
is
to
cu
t
t
h
e
o
r
ig
i
n
al
i
m
ag
e
b
y
h
a
n
d
.
T
h
is
is
b
ec
au
s
e
th
e
s
h
ap
e
o
f
th
e
i
m
ag
e
is
cu
t
to
s
o
m
e
s
h
ap
e.
T
h
e
n
ex
t
s
tep
is
to
p
e
r
f
o
r
m
b
ac
k
g
r
o
u
n
d
s
u
b
tr
ac
tio
n
b
ased
o
n
t
h
e
R
OI
co
o
r
d
in
ates
b
et
w
ee
n
th
e
b
ac
k
g
r
o
u
n
d
i
m
a
g
es.
C
NN
[
1
5
]
S
ets
t
h
e
f
in
a
l
b
in
ar
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
Hu
ma
n
a
ctivity
r
ec
o
g
n
itio
n
b
y
u
s
in
g
co
n
v
o
lu
tio
n
a
l
n
eu
r
a
l
n
e
tw
o
r
k
(
Ha
n
kil
K
im
)
5271
i
m
a
g
e
s
ize,
2
0
0
×
2
0
0
i
m
a
g
e,
f
o
r
th
e
i
n
p
u
t
i
m
ag
e,
to
a
f
i
x
ed
s
ize.
T
h
e
t
h
r
esh
o
ld
i
s
t
h
en
u
s
ed
to
o
b
tain
th
e
b
in
ar
y
i
m
ag
e.
T
h
e
t
h
r
esh
o
ld
is
d
ef
i
n
ed
as
5
0
p
o
in
ts
.
A
t
t
h
e
en
d
o
f
th
i
s
p
r
o
ce
s
s
,
a
m
o
r
p
h
o
lo
g
ical
o
p
er
atio
n
is
u
s
ed
an
d
a
GE
I
i
m
a
g
e
o
f
a
b
in
ar
y
i
m
ag
e
is
o
b
tain
ed
.
Fi
g
u
r
e
1
also
s
h
o
w
s
th
e
s
y
s
te
m
a
r
ch
itect
u
r
e
f
o
r
u
s
to
u
n
d
er
s
ta
n
d
th
e
s
y
s
te
m
ea
s
ier
.
Fig
u
r
e
1
.
E
f
f
ec
t
s
o
f
s
elec
tin
g
d
if
f
er
e
n
t s
w
i
tch
i
n
g
u
n
d
er
d
y
n
a
m
ic
co
n
d
itio
n
I
n
Fi
g
u
r
e
1
,
o
u
r
m
ain
s
y
s
te
m
ar
ch
itect
u
r
e
is
th
at
w
e
p
r
ep
r
o
ce
s
s
t
h
e
u
s
er
s
ec
tio
n
an
d
c
u
t
s
o
m
e
w
o
r
k
in
to
h
an
d
icr
a
f
ts
.
T
h
e
ap
p
lica
tio
n
s
ec
t
io
n
i
s
o
u
r
m
ai
n
p
r
o
p
o
s
ed
m
e
th
o
d
u
s
ed
b
y
th
e
C
NN
(
VGG
16
-
Net)
ar
ch
itect
u
r
e
[
1
6
-
18]
,
an
d
p
a
r
t
o
f
th
e
tr
ain
i
n
g
is
to
u
s
e
o
u
r
m
o
d
el.
W
e
u
s
e
th
e
Ker
as
m
o
d
el
an
d
test
it
as
a
p
ar
t
o
f
th
e
test
.
Ou
r
d
ataset
is
b
y
a
tr
ain
ed
m
o
d
el.
A
n
d
o
u
r
r
ec
o
g
n
itio
n
s
y
s
te
m
i
s
illu
s
tr
ated
in
th
e
f
lo
w
c
h
ar
t
o
f
Fig
u
r
e
2
.
Ou
r
f
ea
t
u
r
e
i
m
a
g
e
is
s
h
o
w
n
i
n
Fi
g
u
r
e
3
.
Fig
u
r
e
1
.
Flo
w
c
h
ar
t o
f
o
u
r
p
r
o
p
o
s
ed
m
et
h
o
d
I
n
Fig
u
r
e
2
,
th
e
T
h
(
i)
im
ag
e
is
a
th
er
m
al
i
m
a
g
e
b
y
B
G
an
d
O
(
i)
is
a
b
ac
k
g
r
o
u
n
d
s
u
b
tr
ac
tio
n
.
B
G
i
m
a
g
e
is
a
b
ac
k
g
r
o
u
n
d
i
m
ag
e
an
d
O
(
i)
i
m
a
g
e
i
s
an
o
r
i
g
in
a
l
i
m
a
g
e.
A
s
s
h
o
w
n
i
n
Fi
g
u
r
e
2
,
it
cu
ts
th
e
v
ar
io
u
s
s
izes
a
n
d
ap
p
ea
r
s
as
a
T
h
(
i)
i
m
ag
e.
Fo
r
C
(
i)
i
m
a
g
es,
cr
ea
t
e
a
2
2
4
×
2
2
4
b
lan
k
i
m
ag
e
f
i
x
ed
s
ize
an
d
ce
n
ter
th
e
cr
o
p
p
ed
im
ag
e
o
f
t
h
e
b
lan
k
i
m
a
g
e.
I
t
th
en
cr
ea
tes
a
n
Av
(
i)
av
er
ag
e
i
m
ag
e
o
f
t
h
e
GE
I
i
m
ag
e.
F,
(
i)
is
th
e
E
GE
I
ed
it
ed
s
tate
i
m
a
g
e
an
d
F
(
i)
is
a
s
a
n
E
GE
I
i
m
ag
e.
T
h
e
last
t
h
r
ee
ch
a
n
n
els
s
h
o
w
n
i
n
Fi
g
u
r
e
3
co
m
b
i
n
e
A
v
(
i)
,
F,
(
i)
,
an
d
F
(
i
)
im
a
g
es.
Fi
g
u
r
e
3
s
elec
ts
a
s
am
p
le
f
r
o
m
a
f
in
a
l
f
u
n
ctio
n
,
s
u
ch
as
a)
s
h
ak
i
n
g
b
y
h
an
d
,
b
)
p
u
n
c
h
c)
w
al
k
i
n
g
,
d
)
k
ick
,
e)
lie
d
o
w
n
,
an
d
f
)
r
u
n
n
i
n
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
201
9
:
5
2
7
0
-
5
2
7
6
5272
Fig
u
r
e
2
.
E
x
a
m
p
le
o
f
o
u
r
f
i
n
al
f
ea
tu
r
e
s
3.
E
XP
E
R
I
M
E
NT
A
L
RE
SUL
T
I
n
t
h
is
s
ec
tio
n
,
w
e
b
r
ief
l
y
d
escr
ib
e
o
u
r
d
atab
ases
,
co
m
p
ar
is
o
n
an
d
ac
c
u
r
ac
y
.
First,
it
'
s
ab
o
u
t
d
atab
ase
1
,
c
o
llected
in
m
u
lti
p
le
en
v
ir
o
n
m
en
ts
w
it
h
o
b
j
ec
ts
th
at
ar
e
d
if
f
er
en
t
f
r
o
m
p
eo
p
le
i
m
ag
e
s
tak
e
n
w
it
h
a
th
er
m
al
ca
m
er
a
in
a
d
ar
k
e
n
v
ir
o
n
m
e
n
t.
So
,
as
m
e
n
tio
n
e
d
ea
r
lier
,
w
e
h
a
v
e
co
m
b
i
n
ed
all
d
atab
ases
w
it
h
d
atab
ase
1
.
T
h
er
e
ar
e
s
ix
d
if
f
er
en
t
p
eo
p
le
an
d
to
p
ics
ca
u
g
h
t
in
d
if
f
er
e
n
t
en
v
ir
o
n
m
en
ts
.
T
h
er
e
is
a
s
a
m
p
le
o
f
th
e
d
atab
ase
i
m
a
g
e
s
h
o
w
n
i
n
Fig
u
r
e
4
.
W
e
ca
n
v
is
u
alize
t
h
e
d
atab
ase
d
escr
ip
tio
n
in
m
o
r
e
d
etail
in
T
ab
le
1
.
Fig
u
r
e
3
.
E
x
a
m
p
le
o
f
t
h
e
d
ata
b
ase
o
f
d
if
f
er
en
t o
b
j
ec
ts
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
Hu
ma
n
a
ctivity
r
ec
o
g
n
itio
n
b
y
u
s
in
g
co
n
v
o
lu
tio
n
a
l
n
eu
r
a
l
n
e
tw
o
r
k
(
Ha
n
kil
K
im
)
5273
T
ab
le
1
.
Descr
ip
tio
n
o
f
d
atab
a
s
e
D
a
t
a
se
t
s
D
e
t
a
i
l
e
d
d
e
scri
p
t
i
o
n
F
i
g
u
r
e
4a
-
o
u
t
si
d
e
n
e
x
t
t
o
r
o
a
d
-
man
i
s
w
a
l
k
i
n
g
r
i
g
h
t
a
n
d
l
e
f
t
F
i
g
u
r
e
4b
-
o
u
t
si
d
e
a
n
d
f
r
o
n
t
o
f
b
u
i
l
d
i
n
g
-
t
h
e
m
a
n
i
s
k
i
c
k
i
n
g
F
i
g
u
r
e
4c
-
o
u
t
si
d
e
n
e
x
t
t
o
r
o
a
d
-
t
h
e
m
a
n
i
s r
u
n
n
i
n
g
T
ab
le
2
s
h
o
w
s
t
h
e
n
u
m
b
er
o
f
i
m
ag
e
s
in
o
u
r
d
atab
ase.
T
ab
le
3
s
h
o
w
s
th
e
n
u
m
b
er
s
o
f
i
m
ag
e
s
an
d
th
e
t
y
p
es
o
f
m
o
tio
n
i
n
ea
ch
d
ataset.
T
h
e
s
o
u
r
ce
d
atab
ase
r
ep
r
esen
ts
a
th
r
ee
-
ch
a
n
n
el
f
ea
tu
r
e
i
m
a
g
e.
T
h
e
au
g
m
e
n
ted
d
atab
ase
d
is
p
la
y
s
ar
ti
f
iciall
y
au
g
m
e
n
ted
f
e
atu
r
e
i
m
a
g
es.
I
t
th
e
n
d
iv
id
e
s
in
to
o
n
e
o
r
t
w
o
s
et
s
o
f
d
ata
an
d
th
e
o
r
ig
i
n
al
a
u
g
m
en
ted
d
atab
ase.
Fo
r
ex
a
m
p
le,
an
a
u
g
m
e
n
te
d
d
ata
s
et
w
a
s
u
s
ed
to
s
tu
d
y
th
e
VGG
-
1
6
Net
[
1
9,
20
]
a
n
d
th
e
o
r
ig
i
n
al
d
ata
s
e
t
w
as
u
s
ed
to
t
es
t
th
e
VGG
-
1
6
Net.
I
n
ad
d
itio
n
,
w
e
u
s
e
t
h
e
i
n
v
er
s
e
to
esti
m
ate
d
o
u
b
le
cr
o
s
s
v
a
lid
atio
n
.
A
d
e
s
cr
ip
tio
n
o
f
th
e
H
AR
s
y
s
te
m
d
atab
ase
s
a
m
p
le
i
s
g
iv
e
n
ab
o
v
e.
T
ab
le
1
.
Nu
m
b
er
o
f
i
m
a
g
es i
n
d
atab
ase
D
a
t
a
se
t
s
T
o
t
a
l
O
r
i
g
i
n
a
l
d
a
t
a
b
a
se
2
101
A
u
g
me
n
t
e
d
d
a
t
a
b
a
se
4
2
0
2
0
N
u
mb
e
r
o
f
p
e
o
p
l
e
6
D
i
f
f
e
r
e
n
t
e
n
v
i
r
o
n
me
n
t
5
T
ab
le
2
.
Nu
m
b
er
s
o
f
i
m
a
g
e
s
an
d
th
e
t
y
p
e
s
o
f
m
o
t
io
n
i
n
ea
ch
d
ataset
A
c
t
i
v
i
t
y
F
r
a
me
B
e
h
a
v
i
o
r
H
a
n
d
w
a
v
i
n
g
1
1
6
3
1
2
0
P
u
n
c
h
i
n
g
9
5
4
3
1
7
K
i
c
k
i
n
g
8
0
7
99
W
a
l
k
i
n
g
3
6
2
1
6
5
R
u
n
n
i
n
g
1
4
2
63
S
i
t
t
i
n
g
44
21
S
t
a
n
d
i
n
g
56
11
L
a
y
i
n
g
d
o
w
n
5
8
0
2
0
7
T
o
t
a
l
1
2
1
0
8
1
0
0
3
Seco
n
d
,
w
e
in
tr
o
d
u
ce
th
e
co
m
p
ar
i
s
o
n
a
n
d
ac
cu
r
ac
y
.
I
n
o
u
r
s
tu
d
y
w
e
u
s
e
t
h
e
C
NN
m
et
h
o
d
in
H
A
R
s
y
s
te
m
s
to
i
m
p
r
o
v
e
t
h
e
s
e
p
r
o
b
le
m
s
.
C
NN
allo
w
s
s
o
m
e
f
ea
tu
r
es
th
at
y
o
u
d
o
n
o
t
n
ee
d
to
tr
a
ck
to
d
etec
t
h
u
m
a
n
leg
o
r
h
a
n
d
lo
ca
tio
n
i
n
f
o
r
m
at
io
n
.
Yo
u
ca
n
also
r
ec
o
g
n
ize
m
an
y
ac
ti
v
it
ies.
B
ec
au
s
e
if
we
tr
ain
t
h
e
d
ata
s
et
r
ea
d
y
f
o
r
th
e
i
n
p
u
t
s
o
f
th
e
m
o
d
u
le,
m
a
n
y
f
u
n
ct
io
n
s
w
il
l
b
e
r
ec
o
g
n
ize
d
as
w
e
e
x
p
ec
ted
.
T
ab
le
4
s
h
o
w
s
th
e
p
r
o
s
an
d
co
n
s
b
et
w
ee
n
th
e
p
r
ev
io
u
s
m
et
h
o
d
an
d
o
u
r
m
et
h
o
d
.
I
n
ad
d
itio
n
,
w
e
tes
te
d
o
u
r
m
et
h
o
d
an
d
co
n
s
tr
u
cted
it to
est
i
m
ate
d
o
u
b
le
-
cr
o
s
s
v
alid
atio
n
i
n
Fi
g
u
r
e
s
5
-
6.
I
n
Fi
g
u
r
es
5
a
n
d
6
,
w
e
tr
ai
n
e
d
ten
ep
o
ch
s
a
n
d
it
s
h
o
w
s
t
h
e
g
o
o
d
ac
cu
r
ac
y
en
d
o
f
th
e
l
ast
ep
o
ch
.
I
n
NN
tr
ain
in
g
,
a
s
tr
o
k
e
r
ep
r
esen
t
s
o
n
e
co
m
p
lete
s
tep
th
r
o
u
g
h
a
g
iv
e
n
s
et
o
f
d
ata.
T
h
e
u
p
p
er
d
ir
ec
tio
n
o
f
th
e
f
ig
u
r
e
is
t
h
e
ac
cu
r
ac
y
,
th
e
lo
w
er
le
f
t a
cc
u
r
ac
y
is
0
~
1
an
d
th
e
n
u
m
b
er
o
f
ep
o
ch
is
0
~
1
0
.
A
cc
u
r
ac
y
v
al
u
es
ar
e
b
et
w
ee
n
0
an
d
1
.
Me
as
u
r
e
0
m
ea
n
s
t
h
at
th
e
f
ea
t
u
r
e
v
alu
es
m
a
tch
a
n
d
d
o
es
n
o
t
r
ec
o
g
n
i
ze
th
e
i
n
p
u
t
i
m
ag
e.
Me
asu
r
ed
v
al
u
e
1
in
d
icate
s
th
e
f
u
n
c
tio
n
m
atc
h
i
n
g
v
al
u
e,
an
d
th
e
co
r
r
esp
o
n
d
in
g
co
m
p
en
s
atio
n
v
alu
e
i
s
d
is
p
la
y
ed
as
tr
u
e.
Fro
m
0
to
1
0
f
o
r
ep
o
ch
w
e
h
a
v
e
tr
ai
n
ed
1
0
s
ets
o
f
d
ata.
Af
ter
d
o
u
b
le
-
cr
o
s
s
i
n
g
tr
ai
n
in
g
,
o
u
r
d
ata
s
et
is
r
ea
d
y
f
o
r
test
i
n
g
.
T
ab
le
5
s
h
o
w
s
th
e
s
u
m
m
ar
y
o
f
co
m
p
ar
is
o
n
s
a
n
d
ac
cu
r
ac
ie
s
.
T
ab
le
3
.
Sy
s
te
m
o
f
m
et
h
o
d
co
m
p
ar
at
iv
e
a
n
al
y
s
i
s
M
e
t
h
o
d
D
e
scri
p
t
i
o
n
P
r
e
v
i
o
u
s me
t
h
o
d
s
-
F
i
x
e
d
t
a
sk
s
e
v
e
n
t
h
a
n
d
l
i
n
g
-
C
o
mp
l
i
c
a
t
e
t
o
e
x
t
e
n
d
s,
su
c
h
a
s
a
d
d
mo
r
e
mo
t
i
o
n
s
-
D
r
a
w
b
a
c
k
o
f
r
e
c
o
g
n
i
t
i
o
n
a
b
o
u
t
s
a
m
e
mo
v
e
m
e
n
t
s
O
u
r
me
t
h
o
d
-
Ea
sy
t
o
e
x
t
e
n
d
a
s
mo
v
e
me
n
t
s a
s re
-
t
r
a
i
n
mo
d
u
l
e
s
-
M
i
n
i
mal
d
r
a
w
b
a
c
k
o
f
r
e
c
o
g
n
i
t
i
o
n
m
o
r
e
t
h
a
n
p
r
e
v
i
o
u
s
me
t
h
o
d
-
Ea
sy
t
o
u
se
o
u
r
sy
st
e
m
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
201
9
:
5
2
7
0
-
5
2
7
6
5274
Fig
u
r
e
4
.
Mo
d
el
o
f
f
ir
s
t tr
ain
i
n
g
Fig
u
r
e
5
.
Mo
d
el
o
f
s
ec
o
n
d
tr
ain
in
g
T
ab
le
4
.
Su
m
m
ar
y
o
f
co
m
p
ar
i
s
o
n
s
a
n
d
ac
cu
r
ac
ies (
u
n
it
: %)
M
e
t
h
o
d
A
c
c
u
r
a
c
y
F
o
u
r
i
e
r
d
e
scri
p
t
o
r
b
a
se
d
me
t
h
o
d
5
0
.
4
G
E
I
b
a
se
d
me
t
h
o
d
6
5
.
1
O
u
r
me
t
h
o
d
9
5
.
9
4.
CO
NCLU
SI
O
N
R
ec
en
t
l
y
,
m
a
n
y
r
esear
c
h
er
s
h
av
e
s
tu
d
ied
o
n
th
e
ac
tiv
it
y
r
ec
o
g
n
itio
n
[
2
1
-
2
5
]
.
T
h
er
e
ar
e
s
e
v
er
al
t
y
p
es
o
f
HAR
s
y
s
te
m
s
.
Fo
r
ex
a
m
p
le,
HA
R
s
y
s
te
m
ca
n
b
e
u
s
ed
f
o
r
s
p
o
r
t
ac
tiv
it
ies
,
d
aily
l
if
e
ac
ti
v
it
ie
s
in
th
e
h
o
s
p
ital
u
s
e,
p
atie
n
t
m
o
n
ito
r
in
g
a
f
ter
s
u
r
g
er
y
,
ca
r
e
f
o
r
eld
er
p
eo
p
le
an
d
etc
.
I
n
tellig
e
n
t
s
u
r
v
eilla
n
ce
s
y
s
te
m
s
ar
e
r
ec
o
g
n
ized
w
h
en
a
p
atien
t f
all
s
w
it
h
o
u
t p
eo
p
le
ar
o
u
n
d
h
i
m
.
I
n
th
i
s
s
tu
d
y
,
h
o
w
e
v
er
,
w
e
p
r
o
p
o
s
ed
th
is
ap
p
r
o
ac
h
to
s
o
lv
e
th
e
ab
o
v
e
m
en
tio
n
ed
p
r
o
b
lem
s
.
H
u
m
an
ac
tiv
it
y
h
as
s
ev
er
al
u
n
iq
u
e
c
h
ar
ac
ter
is
tics
t
h
at
d
o
n
o
t
r
eq
u
i
r
e
a
s
u
b
j
ec
t.
I
n
th
is
s
tu
d
y
,
h
u
m
an
m
o
tio
n
u
s
i
n
g
C
NN
is
a
n
al
y
ze
d
an
d
t
h
er
m
o
g
r
ap
h
y
-
b
ased
h
u
m
a
n
ac
tiv
it
y
r
ec
o
g
n
it
io
n
is
u
s
ed
.
T
h
e
m
ain
p
r
o
b
lem
w
a
s
to
r
eso
lv
e
t
h
e
p
e
r
ce
p
tio
n
o
f
d
a
y
ti
m
e
a
n
d
n
i
g
h
tti
m
e
h
u
m
a
n
ac
tiv
itie
s
i
n
t
h
is
s
t
u
d
y
t
h
at
c
o
u
ld
n
o
t
r
ec
o
g
n
ize
p
r
ev
io
u
s
s
t
u
d
ies.
Vi
s
io
n
-
b
ase
d
a
w
ar
e
n
ess
is
u
s
ef
u
l
f
o
r
ac
ti
v
it
y
r
ec
o
g
n
itio
n
s
y
s
te
m
s
.
Ho
w
e
v
er
,
t
h
e
t
h
er
m
al
ca
m
er
a
is
o
p
er
atin
g
d
u
r
in
g
t
h
e
d
ay
a
n
d
n
ig
h
t
d
escr
ib
ed
a
b
o
v
e
.
W
e
ca
n
s
ee
m
a
n
y
s
tu
d
ies
th
at
h
a
v
e
b
ee
n
r
esear
ch
ed
an
d
d
ev
elo
p
ed
b
y
m
et
h
o
d
s
u
s
ed
i
n
H
A
R
s
y
s
te
m
s
.
Ou
r
r
esu
lts
w
er
e
v
er
y
g
o
o
d
an
d
w
o
r
k
ed
w
ell
af
ter
w
e
u
s
ed
t
h
e
C
NN
m
et
h
o
d
in
th
e
H
AR
s
y
s
te
m
.
T
h
e
r
esu
lt
is
s
h
o
w
n
a
s
9
5
.
9
%,
an
d
is
m
o
r
e
r
ec
o
g
n
izab
le
th
an
t
h
e
o
th
er
m
et
h
o
d
s
.
W
e
p
lan
to
in
cr
ea
s
e
th
e
n
u
m
b
er
o
f
ac
tiv
ities
f
o
r
th
e
ex
p
er
i
m
en
t
in
t
h
e
f
u
t
u
r
e.
W
e
w
il
l
also
ad
d
m
o
r
e
ca
m
er
a
t
y
p
e
s
th
a
t
w
ill
al
lo
w
u
s
to
i
n
cr
ea
s
e
o
u
r
d
ata
s
ets
i
n
o
th
er
d
i
m
en
s
io
n
s
.
W
e
w
ill
also
b
e
p
r
esen
ti
n
g
r
esear
ch
o
n
p
eo
p
le
w
h
o
o
w
n
p
er
s
o
n
al
i
te
m
s
s
u
c
h
as
th
eir
w
a
llets
,
b
ag
s
an
d
ce
ll
p
h
o
n
es.
W
e
also
p
lan
to
s
tu
d
y
th
e
ac
cu
r
ac
y
an
d
la
y
o
u
t o
f
r
ea
l
-
ti
m
e
H
AR
s
y
s
te
m
s
.
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
is
w
o
r
k
w
a
s
s
u
p
p
o
r
ted
b
y
th
e
r
esear
ch
g
r
a
n
t o
f
P
aich
a
i U
n
iv
er
s
it
y
in
2
0
1
9
.
RE
F
E
R
E
NC
E
S
[1
]
J.
S
c
h
m
id
h
u
b
e
r,
“
De
e
p
lea
rn
in
g
i
n
n
e
u
ra
l
n
e
tw
o
rk
s: A
n
o
v
e
rv
ie
w
,
”
Ne
u
ra
l
n
e
two
rk
s
,
v
o
l.
6
1
,
p
p
.
8
5
-
117
,
2
0
1
5
.
[2
]
M
.
I.
Jo
r
d
a
n
a
n
d
T
.
M
.
M
it
c
h
e
ll
,
“
M
a
c
h
i
n
e
lea
rn
i
n
g
:
T
re
n
d
s,
p
e
rsp
e
c
ti
v
e
s,
a
n
d
p
ro
s
p
e
c
ts
,
”
S
c
ien
c
e
,
v
o
l.
3
4
9
,
n
o
.
6
2
4
5
,
p
p
.
2
5
5
-
2
6
0
,
2
0
1
5
.
[3
]
C.
H.
Hw
a
n
g
,
K.
C.
S
in
a
n
d
H.
K.
Ju
n
g
,
“
No
S
QL
Da
tab
a
se
De
sig
n
Us
in
g
UML
c
o
n
c
e
p
tu
a
l
Da
ta
M
o
d
e
l
Ba
se
d
o
n
P
e
ter
C
h
e
n
’s
F
ra
m
e
w
o
rk
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
A
p
p
li
e
d
E
n
g
i
n
e
e
rin
g
Res
e
a
rc
h
,
v
o
l
.
1
2
,
n
o
.
1
0
,
p
p
.
6
3
2
-
636
,
2
0
1
7
.
[4
]
D.
C.
Cires
a
n
,
U.
M
e
ier
U,
a
n
d
L
.
M
.
G
a
m
b
a
rd
e
ll
a
,
“
De
e
p
,
b
ig
,
sim
p
le
n
e
u
ra
l
n
e
ts
f
o
r
h
a
n
d
w
rit
ten
d
ig
it
re
c
o
g
n
it
io
n
,
”
Ne
u
ra
l
c
o
mp
u
ta
t
io
n
,
v
o
l.
2
2
,
n
o
.
1
2
,
p
p
.
3
2
0
7
-
3
2
2
0
,
2
0
1
0
.
[5
]
R.
P
o
p
p
e
,
"
A
su
rv
e
y
o
n
v
isio
n
-
b
a
se
d
h
u
m
a
n
a
c
ti
o
n
re
c
o
g
n
it
i
o
n
,
"
Ima
g
e
a
n
d
v
isio
n
c
o
mp
u
ti
n
g
,
v
o
l.
2
8
,
n
o
.
6
,
p
p
.
9
7
6
-
9
9
0
,
2
0
1
0
,
.
[6
]
J.
W
a
n
g
,
M
.
S
h
e
,
S
.
Na
h
a
v
a
n
d
i
a
n
d
A
.
Ko
u
z
a
n
i,
"
A
re
v
i
e
w
o
f
v
isio
n
-
b
a
se
d
g
a
it
re
c
o
g
n
it
io
n
m
e
t
h
o
d
s
f
o
r
h
u
m
a
n
id
e
n
ti
f
ica
ti
o
n
,
"
Dig
it
a
l
Ima
g
e
Co
mp
u
ti
n
g
:
T
e
c
h
n
i
q
u
e
s
a
n
d
Ap
p
li
c
a
ti
o
n
s (
DICTA
)
,
p
p
.
3
2
0
-
3
2
7
,
2
0
1
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
Hu
ma
n
a
ctivity
r
ec
o
g
n
itio
n
b
y
u
s
in
g
co
n
v
o
lu
tio
n
a
l
n
eu
r
a
l
n
e
tw
o
r
k
(
Ha
n
kil
K
im
)
5275
[7
]
P
.
K.
P
is
h
a
ra
d
y
a
n
d
S
.
M
a
rti
n
S
a
e
rb
e
c
k
,
"
Re
c
e
n
t
m
e
th
o
d
s
a
n
d
d
a
t
a
b
a
se
s
in
v
isio
n
-
b
a
se
d
h
a
n
d
g
e
stu
re
re
c
o
g
n
it
i
o
n
:
A
re
v
ie
w
,
"
Co
mp
u
ter
Vi
sio
n
a
n
d
Ima
g
e
Un
d
e
rs
ta
n
d
in
g
,
v
o
l
.
1
4
1
,
p
p
.
1
5
2
-
1
6
5
,
2
0
1
5
.
[8
]
D.
W
e
in
lan
d
,
R.
Re
m
i
Ro
n
f
a
rd
a
n
d
B.
Ed
m
o
n
d
,
"
A
su
rv
e
y
o
f
v
isio
n
-
b
a
se
d
m
e
th
o
d
s
f
o
r
a
c
ti
o
n
re
p
re
se
n
tatio
n
,
se
g
m
e
n
tatio
n
a
n
d
re
c
o
g
n
i
ti
o
n
,
"
C
o
mp
u
ter
v
isi
o
n
a
n
d
im
a
g
e
u
n
d
e
rs
ta
n
d
in
g
,
v
o
l
.
1
1
5
,
n
o
.
2
,
p
p
.
2
2
4
-
241
,
2
0
1
1
.
[9
]
W
.
K.
W
o
n
g
,
Z
.
Y.
C
h
e
w
,
C.
K.
L
o
o
a
n
d
W
.
S
.
L
im
,
"
A
n
e
ffe
c
ti
v
e
tres
p
a
ss
e
r
d
e
tec
ti
o
n
sy
ste
m
u
sin
g
th
e
rm
a
l
c
a
m
e
ra
,
”
in
Co
mp
u
ter
Res
e
a
rc
h
a
n
d
De
v
e
lo
p
me
n
t,
2
0
1
0
S
e
c
o
n
d
In
t
e
rn
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
,
p
p
.
7
0
2
-
7
0
6
.
[1
0
]
W.
K.
W
o
n
g
,
P
.
N
.
T
a
n
,
C.
K.
L
o
o
a
n
d
W
.
S
.
L
im
,
"
A
n
e
ff
e
c
ti
v
e
su
rv
e
il
lan
c
e
s
y
ste
m
u
sin
g
th
e
rm
a
l
c
a
m
e
ra
,
"
in
S
ig
n
a
l
Acq
u
isit
io
n
a
n
d
Pro
c
e
ss
in
g
,
ICS
A
P
2
0
0
9
,
p
p
.
1
3
-
17
,
2
0
0
9
.
[1
1
]
A
.
S
h
a
rif
Ra
z
a
v
i
a
n
,
H.
A
z
izp
o
u
r
,
J.
S
u
ll
iv
a
n
a
n
d
S
.
Ca
rlsso
n
,
"
CNN
fea
tu
re
s
o
ff
-
t
h
e
-
sh
e
lf
:
a
n
a
st
o
u
n
d
i
n
g
b
a
se
li
n
e
fo
r
re
c
o
g
n
it
io
n
,
"
i
n
P
r
o
c
e
e
d
in
g
s
o
f
th
e
IEE
E
c
o
n
f
e
re
n
c
e
o
n
c
o
m
p
u
ter
v
isio
n
a
n
d
p
a
tt
e
r
n
re
c
o
g
n
it
io
n
w
o
rk
sh
o
p
s
,
p
p
.
8
0
6
-
813
,
2
0
1
4
.
[1
2
]
P
.
Ro
b
i
n
so
n
,
"
T
h
e
CNN
e
ffe
c
t:
Th
e
m
y
th
o
f
n
e
w
s,
f
o
re
ig
n
p
o
li
c
y
a
n
d
i
n
terv
e
n
ti
o
n
,
"
Ro
u
tl
e
d
g
e
,
2
0
0
5
.
[1
3
]
M
.
M
a
tsu
g
u
,
K.
M
o
ri,
Y.
M
it
a
ri
a
n
d
Y.
Ka
n
e
d
a
,
"
S
u
b
jec
t
in
d
e
p
e
n
d
e
n
t
f
a
c
ial
e
x
p
re
ss
io
n
re
c
o
g
n
it
io
n
w
it
h
ro
b
u
st
f
a
c
e
d
e
tec
ti
o
n
u
sin
g
a
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
,
"
Ne
u
ra
l
Ne
two
rk
s
,
v
o
l.
1
6
,
p
p
.
5
5
5
-
5
5
9
,
20
03
.
[1
4
]
J.Y.
Kim
,
S
.
W
.
He
o
a
n
d
J.T
.
L
i
m
,
“
A
L
ice
n
se
P
l;
a
te
Re
c
o
g
n
it
io
n
A
lg
o
rit
h
m
u
sin
g
M
u
lt
i
-
S
tag
e
N
e
u
ra
l
Ne
tw
o
rk
f
o
r
A
u
to
m
o
b
il
e
Blac
k
-
Bo
x
I
m
a
g
e
,
”
J
o
u
rn
a
l
o
f
th
e
Ko
re
a
In
f
o
rm
a
ti
o
n
a
n
d
Co
mm
u
n
ica
ti
o
n
En
g
in
e
e
rin
g
,
v
o
l.
2
2
,
n
o
.
1
,
p
p
.
4
4
-
48
,
2
0
1
8
.
[1
5
]
S
.
Do
d
g
e
a
n
d
L
.
Ka
ra
m
,
"
Un
d
e
rs
ta
n
d
in
g
h
o
w
im
a
g
e
q
u
a
li
ty
a
ff
e
c
ts
d
e
e
p
n
e
u
ra
l
n
e
tw
o
rk
s
,
"
in
,
Qu
a
li
ty
o
f
M
u
lt
im
e
d
ia E
x
p
e
rien
c
e
(Qo
M
E
X
)
,
2
0
1
6
Ei
g
h
t
h
In
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
,
p
p
.
1
-
6
,
2
0
1
6
.
[1
6
]
Z.
L
iu
,
X
.
L
i,
P
.
L
u
o
,
CC.
L
o
y
a
n
d
X
.
T
a
n
g
,
"
S
e
ma
n
ti
c
i
ma
g
e
se
g
me
n
ta
t
io
n
v
ia
d
e
e
p
p
a
rs
in
g
n
e
two
rk
,
"
in
P
r
o
c
e
e
d
in
g
s o
f
th
e
IE
EE
I
n
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
Co
m
p
u
ter V
isi
o
n
,
p
p
.
1
3
7
7
-
1
3
8
5
,
2
0
1
5
.
[1
7
]
A
.
C.
Na
jarro
a
n
d
S
.
M
.
Kim
,
“
No
n
li
n
e
a
r
Co
m
p
e
n
sa
ti
o
n
Us
in
g
A
rti
f
icia
l
Ne
u
ra
l
N
e
t
w
o
rk
in
Ra
d
io
-
o
v
e
r
-
F
i
b
e
r
S
y
st
e
m
,
”
J
o
u
rn
a
l
o
f
I
n
fo
rm
a
ti
o
n
a
n
d
Co
mm
u
n
ica
t
io
n
Co
n
v
e
rg
e
n
c
e
En
g
i
n
e
e
rin
g
,
v
o
l.
1
6
,
n
o
.
1
,
p
p
.
6
-
11
,
2
0
1
8
.
[1
8
]
H.K.
Ju
n
g
,
J.Y.
Kim
a
n
d
H,K.
Ju
n
g
,
“
Co
n
v
o
lu
t
io
n
Ne
u
ra
l
N
e
tw
o
rk
(CN
N)
b
a
se
d
I,
a
g
e
P
r
o
c
e
ss
in
g
S
y
ste
m
,
”
J
o
u
rn
a
l
o
f
In
fo
rm
a
ti
o
n
a
n
d
C
o
m
mu
n
ica
t
io
n
Co
n
v
e
rg
e
n
c
e
En
g
in
e
e
rin
g
,
v
o
l.
1
6
,
n
o
.
3
,
p
p
.
1
6
0
-
1
6
5
,
2
0
1
8
.
[1
9
]
F
.
P
e
d
re
g
o
sa
,
G
.
V
a
ro
q
u
a
u
x
,
A
.
G
ra
m
f
o
rt,
e
t
a
l,
“
S
c
ik
it
-
lea
rn
:
M
a
c
h
in
e
lea
rn
i
n
g
in
P
y
th
o
n
,
”
J
o
u
r
n
a
l
o
f
M
a
c
h
i
n
e
L
e
a
rn
in
g
Res
e
a
rc
h
,
v
o
l.
1
2
,
n
o
.
1
,
p
p
.
2
8
2
5
-
2
8
3
0
,
2
0
1
1
.
[2
0
]
L
.
C.
Jia
o
,
S
.
Y.
Ya
n
g
,
F
.
Li
u
,
e
t
a
l,
“
S
e
v
e
n
ty
y
e
a
rs
b
e
y
o
n
d
n
e
u
ra
l
n
e
tw
o
rk
s:
re
tro
sp
e
c
t
a
n
d
p
r
o
sp
e
c
t
,
”
Ch
in
e
se
J
o
u
r
n
a
l
o
f
C
o
n
p
u
ter
s
,
v
o
l.
3
9
,
n
o
.
8
,
p
p
.
1
6
9
7
-
1
7
1
6
,
2
0
1
6
.
[2
1
]
K.
S
im
o
n
y
a
n
a
n
d
A
.
Zi
ss
e
r
m
a
n
,
"
V
e
ry
d
e
e
p
c
o
n
v
o
lu
ti
o
n
a
l
n
e
tw
o
rk
s
f
o
r
lar
g
e
-
sc
a
le
i
m
a
g
e
re
c
o
g
n
it
io
n
,
"
a
rXiv p
re
p
ri
n
t
a
rX
iv:1
4
0
9
.
1
5
5
6
,
2
0
1
4
.
[2
2
]
M
.
He
n
rio
n
,
“
P
ro
p
a
g
a
ti
n
g
u
n
c
e
rt
a
in
ty
in
Ba
y
e
sia
n
n
e
tw
o
rk
s
b
y
p
ro
b
a
b
i
li
stic
lo
g
ic
sa
m
p
li
n
g
,
”
M
a
c
h
in
e
In
tell
ig
e
n
c
e
a
n
d
Pa
tt
e
rn
Rec
o
g
n
it
io
n
,
v
o
l.
5
,
p
p
.
1
4
9
-
1
6
3
,
1
9
8
8
.
[2
3
]
G
.
J.
J.
Bu
rg
a
n
d
P
.
J.
F
.
G
ro
e
n
e
n
,
“
G
e
n
S
V
M
:
A
g
e
n
e
ra
li
z
e
d
m
u
lt
icla
ss
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
,
”
J
o
u
rn
a
l
o
f
M
a
c
h
in
e
L
e
a
r
n
in
g
Res
e
a
rc
h
,
v
o
l.
1
7
,
n
o
.
1
,
p
p
.
1
-
41
,
2
0
1
6
.
[2
4
]
D.
Ch
e
n
,
“
F
a
u
lt
Clas
sif
ica
ti
o
n
Re
se
a
rc
h
o
f
A
n
a
lo
g
El
e
c
tro
n
ic
Circu
it
s
Ba
se
d
o
n
S
u
p
p
o
rt
Ve
c
to
r
M
a
c
h
in
e
,
”
Ch
e
mic
a
l
En
g
i
n
e
e
rin
g
T
ra
n
s
a
c
ti
o
n
s
,
v
o
l.
5
1
,
p
p
.
1
3
3
3
-
1
3
3
8
,
2
0
1
6
.
[2
5
]
C.
M
a
,
C.
Ch
e
n
,
Q.
L
iu
,
e
t
a
l,
“
S
o
u
n
d
Qu
a
l
it
y
Ev
a
lu
a
ti
o
n
o
f
th
e
In
terio
r
No
ise
o
f
P
u
re
El
e
c
tri
c
Ve
h
icle
Ba
se
d
o
n
Ne
u
ra
l
Ne
tw
o
rk
M
o
d
e
l
,
”
IEE
E
T
r
a
n
sa
c
ti
o
n
s
o
n
In
d
u
stri
a
l
El
e
c
tro
n
ics
,
v
o
l.
6
4
,
n
o
.
1
2
,
p
p
.
9
4
4
2
-
9
4
5
0
,
2
0
1
7
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
H
a
n
k
il
K
i
m
,
h
e
re
c
e
i
v
e
d
th
e
B.
S
.
,
M
.
S
.
d
e
g
re
e
s
f
ro
m
th
e
De
p
a
rt
m
e
n
t
o
f
El
e
c
tro
n
ic
En
g
in
e
e
rin
g
o
f
Ha
n
b
a
t
Na
ti
o
n
a
l
U
n
iv
e
rsity
,
Ko
re
a
,
in
2
0
0
6
,
a
n
d
2
0
1
1
,
a
n
d
th
e
P
h
.
D.
d
e
g
re
e
in
2
0
1
5
f
ro
m
th
e
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
En
g
in
e
e
rin
g
o
f
P
a
ich
a
i
Un
iv
e
rsit
y
,
Ko
re
a
.
S
in
c
e
2
0
0
6
,
h
e
h
a
s
w
o
rk
e
d
in
t
h
e
De
p
a
rtm
e
n
t
o
f
M
u
sic
&
S
o
u
n
d
T
e
c
h
n
o
l
o
g
y
a
t
Ko
re
a
Un
iv
e
r
sity
o
f
M
e
d
ia
A
rts
a
s
a
p
ro
f
e
ss
o
r.
His
c
u
rre
n
t
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
m
u
lt
im
e
d
ia
so
u
n
d
,
b
r
o
a
d
c
a
st
so
u
n
d
,
S
R(S
o
u
n
d
Re
in
f
o
rc
e
m
e
n
t),
S
o
u
n
d
En
g
in
e
e
ri
n
g
,
a
n
d
d
a
tab
a
se
.
S
u
n
g
o
c
k
Le
e
,
s
h
e
re
c
e
i
v
e
d
th
e
B.
S
.
d
e
g
re
e
f
ro
m
th
e
De
p
a
rt
m
e
n
t
o
f
En
g
li
sh
L
a
n
g
u
a
g
e
a
n
d
Cu
lt
u
re
a
t
Ha
n
y
a
n
g
Un
iv
e
rsit
y
,
Ko
re
a
,
in
2
0
0
8
,
t
h
e
M
.
S
.
d
e
g
re
e
f
ro
m
th
e
De
p
a
rtme
n
t
o
f
En
g
li
sh
L
a
n
g
u
a
g
e
Ed
u
c
a
ti
o
n
a
t
Ko
re
a
Un
iv
e
rsity
,
Ko
re
a
,
in
2
0
1
2
,
a
n
d
th
e
P
h
.
D.
d
e
g
re
e
f
ro
m
th
e
De
p
a
rt
m
e
n
t
o
f
Co
m
p
u
ter
En
g
in
e
e
rin
g
a
t
P
a
ich
a
i
Un
iv
e
rsity
,
Ko
re
a
in
2
0
1
5
.
S
i
n
c
e
2
0
1
2
,
s
h
e
h
a
s
w
o
rk
e
d
a
s
a
n
Ex
e
c
u
ti
v
e
Dire
c
to
r
a
t
Da
e
w
o
o
Vo
c
a
ti
o
n
a
l
Co
ll
e
g
e
.
He
r
c
u
rre
n
t
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
v
o
c
a
ti
o
n
a
l
e
d
u
c
a
ti
o
n
a
n
d
trai
n
in
g
(
V
ET
)
i
n
c
o
m
p
u
ter
sc
ien
c
e
,
im
a
g
e
p
ro
c
e
ss
in
g
,
m
a
c
h
in
e
lea
rn
i
n
g
,
b
ig
d
a
ta,
VR
a
n
d
Io
T
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
201
9
:
5
2
7
0
-
5
2
7
6
5276
H
o
e
k
y
u
n
g
J
u
n
g
,
h
e
re
c
e
iv
e
d
th
e
M
.
S
.
d
e
g
re
e
in
1
9
8
7
a
n
d
P
h
.
D.
d
e
g
re
e
in
1
9
9
3
f
ro
m
th
e
De
p
a
rtm
e
n
t
o
f
Co
m
p
u
ter
En
g
in
e
e
rin
g
o
f
Kw
a
n
g
w
o
o
n
U
n
iv
e
rsity
,
Ko
re
a
.
F
ro
m
1
9
9
4
t
o
1
9
9
5
,
h
e
w
o
rk
e
d
f
o
r
ET
RI
a
s
a
re
se
a
rc
h
e
r.
S
i
n
c
e
1
9
9
4
,
h
e
h
a
s
w
o
rk
e
d
i
n
t
h
e
De
p
a
rtm
e
n
t
o
f
Co
m
p
u
t
e
r
En
g
in
e
e
rin
g
a
t
P
a
ic
h
a
i
Un
iv
e
rsity
,
w
h
e
re
h
e
n
o
w
w
o
rk
s
a
s
a
p
ro
f
e
s
so
r.
His
c
u
rre
n
t
re
se
a
rc
h
i
n
tere
sts
in
c
lu
d
e
m
u
lt
im
e
d
ia
d
o
c
u
m
e
n
t
a
r
c
h
it
e
c
tu
re
m
o
d
e
li
n
g
,
in
f
o
rm
a
ti
o
n
p
r
o
c
e
ss
in
g
,
i
n
f
o
rm
a
ti
o
n
re
tri
e
v
a
l,
m
a
c
h
in
e
lea
rn
in
g
,
b
ig
d
a
ta,
a
n
d
Io
T
.
Evaluation Warning : The document was created with Spire.PDF for Python.