I
A
E
S
I
n
t
e
r
n
at
io
n
al
Jou
r
n
al
of
A
r
t
if
ic
ia
l
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
V
ol
.
13
, N
o.
4
,
D
e
c
e
m
be
r
2024
, pp.
4263
~
4272
I
S
S
N
:
2252
-
8938
,
D
O
I
:
10.11591/
ij
a
i.
v
13
.i
4
.pp
4263
-
4272
4263
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ij
ai
.
ia
e
s
c
or
e
.c
om
Im
p
r
ovi
n
g
b
ase
l
i
n
e
r
e
d
u
c
t
i
on
f
or
e
m
ot
i
on
r
e
c
ogn
i
t
i
on
b
ase
d
on
e
l
e
c
t
r
oe
n
c
e
p
h
al
ogr
am
si
gn
a
l
s
I
M
ad
e
A
gu
s
Wi
r
aw
an
, I
G
e
d
e
M
ah
e
n
d
r
a D
ar
m
aw
ig
u
n
a, I
d
a B
agu
s
N
yom
an
P
as
c
i
m
a
D
e
pa
r
t
m
e
nt
of
I
nf
or
m
a
t
i
c
s
E
ngi
ne
e
r
i
ng,
F
a
c
ul
t
y of
E
ngi
ne
e
r
i
ng a
nd V
oc
a
t
i
ona
l
, U
ni
ve
r
s
i
t
a
s
P
e
ndi
di
ka
n G
a
ne
s
ha
, B
a
l
i
, I
ndone
s
i
a
A
r
t
ic
le
I
n
f
o
A
B
S
T
R
A
C
T
A
r
ti
c
le
h
is
to
r
y
:
R
e
c
e
iv
e
d
J
a
n 12, 2024
R
e
vi
s
e
d
M
a
r
13, 2024
A
c
c
e
pt
e
d
M
a
r
21, 2024
The
baseline
reduction
method
has
been
widely
used
to
define
electroenceph
alogram
(EEG)
signal
patterns.
However,
because
the
b
aseline
signal
in
this
approach
contains
artifacts,
the
baseline
reduction
ap
proach
cannot
perform
optimally.
As
a
result,
decreasing
artifacts
in
the
b
aseline
signal
is
critical.
The
m
ean,
Gaussian,
and
Savitzky
-
Golay
filters
will
be
compared
in
this
study
to
minim
ize
artifacts
in
the
baseline
signal.
Three
secondary
datasets
are
utilized
to
evaluate
these
approach
es'
capacity
to
remove
artifacts.
These
three
strategies
are
also
tested
with
the
conv
olution
neural
network
classifi
cation
algorit
hms.
When
applied
to
the
dataset
for
emotion
analysis
using
physio
logical
signals
(
DEAP
)
and
a
datas
et
for
multimodal
resear
ch
of
affec
t,
personality
traits
,
and
mood
on
indi
viduals
and
groups
(
AMIGOS
)
datasets,
the
m
ean
filter
can
increase
b
aseline
reduction
performance
based
on
twenty
-
four
test
scenarios.
On
th
e
data
readiness
for
machine
learning
research
(
DREAMER
)
dataset,
howev
er,
the
Gaussian
filter
is
preferable.
The
relative
difference
approach
was
employed
in
this
study'
s
baseline
reduction
process
to
generate
EEG
signal
p
atterns
that
are
easy
to
recognize
throughout
the
classification
phase,
which
i
mpacts
increasin
g accur
acy.
K
e
y
w
o
r
d
s
:
B
a
s
e
li
ne
r
e
duc
ti
on
E
le
c
tr
oe
nc
e
pha
lo
gr
a
m
s
ig
na
l
E
m
ot
io
n r
e
c
ogni
ti
on
S
m
oot
hi
ng a
ppr
oa
c
h
This is an
open
acce
ss artic
le unde
r the
CC BY
-
SA
license.
C
or
r
e
s
pon
di
n
g A
u
th
or
:
I
M
a
de
A
gus
W
ir
a
w
a
n
D
e
pa
r
tm
e
nt
of
I
nf
or
m
a
ti
c
s
E
ngi
ne
e
r
in
g, F
a
c
ul
ty
of
E
ngi
ne
e
r
in
g
a
nd V
oc
a
ti
ona
l
U
ni
ve
r
s
it
a
s
P
e
ndi
di
ka
n G
a
ne
s
ha
U
da
ya
na
R
oa
d, S
in
ga
r
a
ja
,
B
a
li
, I
ndone
s
ia
E
m
a
il
:
im
a
de
.
a
gus
w
ir
a
w
a
n
@
undi
ks
ha
.
a
c
.i
d
1.
I
N
T
R
O
D
U
C
T
I
O
N
H
um
a
ns
'
e
m
ot
io
na
l
r
e
s
pon
s
e
s
to
in
te
r
pe
r
s
ona
l
c
om
m
uni
c
a
ti
on
a
r
e
c
r
it
ic
a
l
ps
yc
hol
ogi
c
a
l
pr
oc
e
s
s
e
s
[
1]
.
E
m
ot
io
na
l
r
e
s
pons
e
s
in
e
a
c
h
qua
dr
a
nt
d
e
s
c
r
ib
e
hum
a
n
p
e
r
f
or
m
a
nc
e
a
nd
m
e
nt
a
l
he
a
lt
h
[
2]
.
N
um
e
r
ous
a
tt
e
m
pt
s
ha
ve
be
e
n
m
a
de
to
im
pr
ove
th
e
a
c
c
ur
a
c
y
of
e
m
ot
io
n
r
e
c
ogni
ti
on
ba
s
e
d
on
e
le
c
tr
oe
nc
e
pha
lo
gr
a
m
(
E
E
G
)
s
ig
na
ls
,
s
uc
h
a
s
d
e
e
p
le
a
r
ni
ng,
m
a
c
hi
ne
l
e
a
r
ni
ng,
a
nd
da
ta
s
e
t
ba
la
nc
in
g
[
3]
–
[
6]
.
H
ow
e
ve
r
,
a
w
id
e
r
a
nge
of
ot
he
r
f
a
c
to
r
s
,
s
uc
h
a
s
e
a
c
h
pa
r
ti
c
ip
a
nt
'
s
c
ha
r
a
c
te
r
is
ti
c
s
,
le
ve
l
of
c
ogni
ti
ve
a
bi
li
ty
,
a
nd
s
e
x,
c
a
n
ge
ne
r
a
te
th
e
di
f
f
e
r
e
nc
e
in
E
E
G
s
ig
na
l
pa
tt
e
r
ns
.
T
hi
s
di
f
f
e
r
e
nc
e
in
E
E
G
s
ig
na
l
pa
tt
e
r
ns
a
f
f
e
c
ts
e
m
ot
io
n
r
e
c
ogni
ti
on
a
c
c
ur
a
c
y
[
7]
.
A
ba
s
e
li
ne
r
e
duc
ti
on
te
c
hni
que
ha
s
be
e
n
r
e
s
e
a
r
c
he
d
to
s
ol
ve
th
is
pr
obl
e
m
[
8]
.
B
y
us
in
g t
he
a
ve
r
a
ge
va
lu
e
of
t
he
ba
s
e
li
ne
s
ig
n
a
l
c
ha
r
a
c
te
r
is
ti
c
s
, t
h
e
ba
s
e
li
ne
r
e
duc
ti
on a
ppr
oa
c
h r
e
du
c
e
s
t
he
t
r
ia
l
s
ig
na
l
c
ha
r
a
c
te
r
is
ti
c
s
[
7]
–
[
9]
.
H
ow
e
ve
r
,
r
e
c
or
di
ng
ba
s
e
li
ne
E
E
G
s
ig
na
ls
f
r
e
e
f
r
om
in
te
r
na
l
a
nd
e
xt
e
r
na
l
di
s
r
upt
io
ns
is
c
ha
ll
e
ngi
ng
[
3]
,
[
10
]
,
[
11]
.
E
le
c
tr
ooc
ul
ogr
a
m
(
E
O
G
)
,
e
le
c
tr
oc
a
r
di
og
r
a
m
(
E
C
G
)
,
e
le
c
tr
om
yogr
a
m
(
E
M
G
)
,
a
s
w
e
ll
a
s
pa
r
ti
c
ip
a
nt
s
'
e
m
ot
io
na
l
r
e
a
c
ti
ons
w
he
n
th
e
y
a
r
e
in
a
c
a
lm
c
o
ndi
ti
on,
c
a
n
a
ll
c
a
us
e
in
te
r
na
l
di
s
tu
r
ba
nc
e
s
to
E
E
G
s
ig
na
ls
[
11]
,
[
12]
.
T
he
r
e
c
or
di
ng
of
th
e
ba
s
e
li
ne
E
E
G
s
ig
na
l
m
a
y
a
l
s
o
be
di
s
tu
r
be
d
by e
le
c
tr
ic
it
y
a
t
e
a
c
h
e
le
c
tr
ode
[
3]
,
[
10]
,
[
11
]
.
T
he
s
e
di
s
tu
r
ba
nc
e
s
m
a
y
pr
e
ve
nt
th
e
b
a
s
e
li
ne
r
e
duc
ti
on
s
tr
a
te
gy
f
r
om
f
unc
ti
oni
ng
a
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
13
, N
o.
4
,
D
e
c
e
m
be
r
2024
:
4263
-
4272
4264
it
s
be
s
t.
E
li
m
in
a
ti
ng
in
te
r
f
e
r
e
nc
e
f
r
om
th
e
ba
s
e
li
ne
E
E
G
r
e
a
di
ngs
is
c
r
uc
ia
l
f
or
s
ol
vi
ng
th
e
s
e
is
s
ue
s
.
T
he
s
m
oot
hi
ng
m
e
th
od
c
a
n
r
e
duc
e
th
e
di
s
tu
r
ba
nc
e
s
/
a
r
ti
f
a
c
ts
in
th
e
ba
s
e
li
ne
E
E
G
s
ig
na
l
by
m
a
ki
ng
th
e
a
m
pl
it
ude
s
m
oot
he
r
[
13]
.
N
um
e
r
ous
s
m
oot
hi
ng
a
ppr
oa
c
he
s
,
s
uc
h
a
s
th
e
m
e
a
n,
S
a
vi
tz
ky
-
G
ol
a
y,
a
nd
G
a
us
s
ia
n
f
il
te
r
s
,
w
e
r
e
us
e
d
to
r
e
duc
e
th
e
a
r
ti
f
a
c
t
in
th
e
pr
e
vi
ous
s
tu
dy.
T
h
e
s
e
th
r
e
e
s
m
oot
hi
ng
te
c
hni
que
s
ha
v
e
lo
w
e
r
m
e
a
n
s
qua
r
e
d
e
r
r
or
(
M
S
E
)
va
lu
e
s
a
nd
c
a
lc
ul
a
ti
on
ti
m
e
s
[
13]
.
I
n
e
a
r
li
e
r
r
e
s
e
a
r
c
h,
th
e
s
m
oot
hi
ng
m
e
th
od
w
a
s
not
a
ppl
ie
d t
o e
m
ot
io
n r
e
c
ogni
ti
on t
o i
m
pr
ove
ba
s
e
li
ne
r
e
duc
ti
on pe
r
f
or
m
a
nc
e
.
B
a
s
e
d
on
s
tu
di
e
s
f
r
om
pr
e
vi
ous
r
e
s
e
a
r
c
h,
two
c
ont
r
ib
ut
io
ns
a
r
e
pr
opos
e
d
in
th
is
r
e
s
e
a
r
c
h:
i
)
in
ve
s
ti
ga
te
s
th
e
e
f
f
e
c
t
s
of
a
ppl
yi
ng
th
r
e
e
s
m
oot
hi
ng
m
e
t
hods
to
th
e
ba
s
e
li
ne
s
ig
na
l
f
or
th
e
ba
s
e
li
ne
r
e
duc
ti
on
pr
oc
e
s
s
.
W
hi
le
pr
e
vi
ou
s
r
e
s
e
a
r
c
h
h
a
s
e
xpl
or
e
d
th
e
im
pa
c
t
of
a
ppl
yi
ng
ba
s
e
li
ne
r
e
duc
ti
on,
it
ha
s
not
s
tu
di
e
d
th
e
e
f
f
e
c
t
of
ba
s
e
li
ne
r
e
duc
ti
on
w
it
h
a
ba
s
e
li
ne
s
ig
na
l
th
a
t
ha
s
be
e
n
s
m
oot
he
d;
ii
)
de
te
r
m
in
e
s
th
e
be
s
t
s
m
oot
hi
ng me
th
od ba
s
e
d on the
a
c
c
ur
a
c
y va
lu
e
s
f
or
r
e
c
ogni
z
in
g f
our
a
nd t
w
o
e
m
ot
io
n c
la
s
s
e
s
. T
he
s
m
oot
he
d
ba
s
e
li
ne
E
E
G
s
ig
na
l
is
e
xpe
c
te
d
to
m
a
xi
m
iz
e
th
e
ba
s
e
li
ne
r
e
duc
ti
on
s
tr
a
te
gy,
r
e
s
ul
ti
ng
in
hi
gh
a
c
c
ur
a
c
y
in
e
m
ot
io
n r
e
c
ogni
ti
on.
2.
M
E
T
H
O
D
T
he
e
f
f
e
c
ti
ve
ne
s
s
of
th
e
ba
s
e
li
ne
r
e
duc
ti
on
a
ppr
oa
c
h
w
a
s
opt
i
m
iz
e
d
by
c
hoos
in
g
th
e
be
s
t
s
m
oot
hi
ng
m
e
th
ods
.
T
he
r
e
la
ti
ve
di
f
f
e
r
e
nc
e
m
e
th
od
w
a
s
th
e
f
ounda
ti
on
f
o
r
th
e
ba
s
e
li
ne
r
e
duc
ti
on
pr
oc
e
dur
e
us
e
d
in
th
is
in
ve
s
ti
ga
ti
on
[
8]
.
F
ur
th
e
r
m
or
e
,
th
e
c
la
s
s
if
ic
a
ti
on
s
tr
a
te
gy
in
th
is
w
or
k
us
e
d
th
e
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
ks
(
C
N
N
)
m
e
th
od,
w
hi
le
th
e
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
a
nd
r
e
pr
e
s
e
nt
a
ti
on
pr
oc
e
s
s
e
s
us
e
d
th
e
di
f
f
e
r
e
nt
ia
l
e
nt
r
opy
(
D
E
)
a
nd
3D
c
ube
m
e
th
ods
[
7]
,
[
8]
.
T
hi
s
s
ta
ge
in
vol
ve
s
s
e
gm
e
nt
a
ti
on,
e
li
m
in
a
ti
ng
a
r
ti
f
a
c
ts
f
r
om
th
e
ba
s
e
li
ne
E
E
G
da
ta
,
a
nd
de
c
om
pos
it
io
n
.
B
a
s
e
d
on
th
e
c
ont
r
ib
ut
io
ns
pr
opos
e
d
i
n
th
is
r
e
s
e
a
r
c
h
(
or
a
nge
r
e
c
ta
ngl
e
)
,
th
e
pr
oc
e
s
s
of
a
n e
m
ot
io
n r
e
c
ogni
ti
on mode
l
ba
s
e
d on E
E
G
s
ig
na
ls
c
a
n be
de
s
c
r
ib
e
d a
s
s
how
n i
n F
ig
ur
e
1.
F
ig
ur
e
1
. E
m
ot
io
n
r
e
c
ogni
ti
on ba
s
e
d on E
E
G
s
ig
na
l
pr
oc
e
s
s
2.1
.
E
m
ot
io
n
d
at
as
e
t
P
r
e
pa
r
in
g
a
n
E
E
G
s
ig
na
l
-
ba
s
e
d
e
m
ot
io
n
da
ta
s
e
t
is
e
s
s
e
nt
ia
l
to
te
s
t
th
e
pr
opos
e
d
c
ont
r
ib
ut
io
n
in
th
is
r
e
s
e
a
r
c
h. T
hi
s
r
e
s
e
a
r
c
h us
e
s
t
hr
e
e
publi
c
da
ta
s
e
ts
:
th
e
da
ta
s
e
t
f
or
e
m
ot
io
n a
na
ly
s
is
us
in
g phys
io
lo
gi
c
a
l
s
ig
na
ls
(
D
E
A
P
)
,
da
ta
r
e
a
di
ne
s
s
f
or
m
a
c
hi
ne
le
a
r
ni
ng
r
e
s
e
a
r
c
h
(
D
R
E
A
M
E
R
)
,
a
nd
a
da
ta
s
e
t
f
or
m
ul
ti
m
oda
l
r
e
s
e
a
r
c
h
of
a
f
f
e
c
t,
pe
r
s
ona
li
ty
tr
a
it
s
a
nd
m
ood
on
in
di
vi
dua
ls
a
nd
gr
oups
(
A
M
I
G
O
S
)
.
T
he
s
e
th
r
e
e
da
t
a
s
e
t
s
ha
ve
di
f
f
e
r
e
nt
da
ta
c
ha
r
a
c
te
r
is
ti
c
s
,
s
uc
h
a
s
th
e
num
be
r
of
c
h
a
nn
e
ls
us
e
d,
E
E
G
e
qui
pm
e
nt
,
pa
r
ti
c
ip
a
nt
s
,
a
nd
r
e
c
or
di
ng
s
c
e
na
r
io
s
.
S
o,
by
u
s
in
g
th
e
s
e
th
r
e
e
d
a
ta
s
e
ts
,
it
i
s
ho
pe
d
th
a
t
w
e
w
il
l
ge
t
a
n
a
ppr
opr
ia
te
s
m
oot
hi
ng
m
e
th
od
to
im
pr
ove
ba
s
e
li
ne
r
e
duc
ti
on
pe
r
f
or
m
a
nc
e
a
nd
i
m
pa
c
t
in
c
r
e
a
s
in
g
th
e
a
c
c
ur
a
c
y
of
e
m
ot
io
n
r
e
c
ogni
ti
on.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
I
m
pr
ov
in
g bas
e
li
ne
r
e
duc
ti
on f
or
e
m
ot
io
n
r
e
c
ogni
ti
on bas
e
d on
…
(
I
M
ade
A
gus
W
ir
aw
an)
4265
2.2
.
S
e
gm
e
n
t
at
io
n
T
he
E
E
G
s
ig
n
a
l
i
s
di
vi
d
e
d
i
nt
o
t
w
o pi
e
c
e
s
th
r
o
ughou
t
th
e
s
e
g
m
e
nt
a
ti
on
pr
o
c
e
s
s
:
a
b
a
s
e
li
n
e
a
nd a
tr
i
a
l
s
ig
n
a
l.
T
he
ba
s
e
li
ne
s
ig
na
l
w
a
s
a
n
E
E
G
r
e
a
di
ng
r
e
pr
e
s
e
nt
in
g
th
e
s
ubj
e
c
ts
'
ne
u
tr
a
l
s
ta
t
e
[
14]
,
[
15]
.
T
h
e
b
a
s
e
li
n
e
s
ig
n
a
l
w
a
s
c
a
pt
ur
e
d
w
h
e
n
in
di
vi
du
a
l
s
w
e
r
e
c
a
lm
a
n
d
f
r
e
e
of
i
nt
e
r
na
l
a
nd
e
xt
e
r
n
a
l
a
r
ti
f
a
c
t
s
.
T
h
e
E
E
G
s
ig
na
l
us
e
d
in
th
e
tr
ia
l
s
ig
na
l
i
ndi
c
a
te
d
th
e
s
ub
je
c
ts
'
e
m
ot
io
na
l
r
e
s
p
on
s
e
s
.
T
he
ba
s
e
li
ne
s
ig
n
a
ls
in
th
e
D
E
A
P
d
a
ta
s
e
t
w
e
r
e
pr
e
s
e
nt
f
or
th
e
f
ir
s
t
th
r
e
e
s
e
c
ond
s
,
w
hi
le
th
e
tr
ia
l
s
ig
n
a
ls
w
e
r
e
pr
e
s
e
nt
f
or
f
our
th
r
ough
s
ix
ty
-
t
hr
e
e
s
e
c
ond
s
.
I
n
D
R
E
A
M
E
R
a
n
d
A
M
I
G
O
S
,
th
e
b
a
s
e
li
ne
s
ig
na
l
la
s
t
s
f
or
th
e
f
ir
s
t
f
iv
e
s
e
c
on
ds
,
w
hi
l
e
th
e
tr
i
a
l
s
ig
n
a
l
la
s
t
s
f
or
th
e
f
ol
lo
w
in
g s
ix
s
e
c
on
ds
[
16]
–
[
18]
.
O
n
D
E
A
P
,
D
R
E
A
M
E
R
, a
nd
A
M
I
G
O
S
,
th
e
E
E
G
s
ig
n
a
l
d
a
ta
w
a
s
s
a
m
pl
e
d
a
t
128
H
z
e
ve
r
y
s
e
c
o
nd.
T
h
e
r
e
w
e
r
e
32
c
ha
nne
l
s
i
n
th
e
D
E
A
P
da
ta
s
e
t.
T
he
r
e
a
r
e
14
c
ha
nne
l
s
in
th
e
D
R
E
A
M
E
R
a
n
d
A
M
I
G
O
S
da
t
a
s
e
t
s
[
16]
–
[
18]
.
T
h
e
D
E
A
P
da
t
a
s
e
t'
s
F
p1 c
ha
nne
l'
s
th
r
e
e
-
s
e
c
o
nd
ba
s
e
li
ne
s
ig
n
a
l
da
ta
is
s
how
n
i
n
T
a
bl
e
1
.
T
he
th
r
e
e
s
e
c
ond
s
of
ba
s
e
li
n
e
s
ig
n
a
ls
a
r
e
di
s
pl
a
ye
d
i
n
T
a
bl
e
1.
T
he
D
E
A
P
da
t
a
c
ol
le
c
ti
on
tr
ia
l
s
ig
n
a
l
s
e
gm
e
nt
r
uns
f
r
om
f
our
t
o
s
i
xt
y
-
th
r
e
e
s
e
c
o
nds
(
3
85 H
z
t
o
8064
H
z
)
.
T
a
bl
e
1
.
T
he
D
E
A
P
d
a
ta
s
e
t'
s
b
a
s
e
li
ne
s
ig
na
l
s
e
gm
e
nt
s
da
ta
t
a
ke
s
t
hr
e
e
s
e
c
ond
s
S
a
m
pl
i
ng r
a
t
e
(
H
z
)
B
a
s
e
l
i
ne
s
i
gna
l
(
)
1
1
…………
…………
384
384
2.3
.
P
r
oc
e
s
s
o
f
s
m
oot
h
in
g t
h
e
b
as
e
li
n
e
s
ig
n
al
T
hi
s
pr
oc
e
dur
e
i
s
th
e
s
tu
dy'
s
c
r
it
ic
a
l
c
ont
r
ib
ut
io
n.
T
hi
s
s
tu
dy
w
il
l
e
xa
m
in
e
th
r
e
e
s
m
oot
hi
ng
m
e
th
ods
to
e
li
m
in
a
te
in
te
r
f
e
r
e
nc
e
in
th
e
ba
s
e
li
ne
s
ig
na
l:
th
e
m
e
a
n
f
il
te
r
,
th
e
S
a
vi
tz
ky
-
G
ol
a
y
f
il
te
r
,
a
nd
th
e
G
a
us
s
ia
n
f
il
te
r
[
19]
.
T
he
be
s
t
of
th
e
th
r
e
e
a
ppr
oa
c
he
s
w
a
s
c
ho
s
e
n
b
a
s
e
d
on
th
e
r
e
s
ul
ti
ng
e
m
ot
io
n
r
e
c
ogni
ti
on
a
c
c
ur
a
c
y
va
lu
e
.
T
he
s
m
oot
hi
ng pr
oc
e
dur
e
f
or
t
he
m
e
a
n
f
il
te
r
m
e
th
od be
g
in
s
f
r
om
t
he
f
ir
s
t
s
a
m
pl
in
g
r
a
te
i
n t
he
ba
s
e
li
ne
s
ig
na
l
da
ta
(
1 H
z
)
.
I
n 1 s
e
c
ond, the
E
E
G
s
ig
na
l
w
i
ll
pr
oduc
e
12
8 s
a
m
pl
in
g va
lu
e
s
. T
he
s
m
oot
hi
ng pr
oc
e
s
s
f
or
th
e
F
p1 c
ha
nne
l
in
t
he
D
E
A
P
da
ta
s
e
t
is
i
ll
us
tr
a
te
d i
n
T
a
bl
e
2.
T
a
bl
e
2.
I
ll
us
tr
a
ti
on of
t
he
s
m
oot
hi
ng pr
oc
e
s
s
f
or
t
he
D
E
A
P
da
t
a
s
e
t
us
in
g t
he
m
e
a
n f
il
te
r
m
e
th
od on the
ba
s
e
li
ne
s
ig
n
a
l
S
a
m
pl
i
ng r
a
t
e
(
H
z
)
B
a
s
e
l
i
ne
s
i
gna
l
(
x
i
)
M
e
a
n f
i
l
t
e
r
(
z
i
)
1
1
1
=
0
+
1
+
2
2
∗
1
+
1
………
………
…………………………
384
384
384
=
383
+
384
+
0
2
∗
1
+
1
I
n
a
ddi
ti
on
to
a
n
a
ly
z
in
g
t
he
m
e
a
n
f
il
te
r
m
e
th
o
d
f
or
s
m
oot
hi
n
g
th
e
ba
s
e
li
ne
s
ig
n
a
l,
t
hi
s
s
t
udy
a
ls
o
s
tu
di
e
s
th
e
S
a
v
it
z
ky
G
ol
a
y
f
il
te
r
a
ppr
oa
c
h. T
he
s
m
o
ot
hi
ng
pr
o
c
e
s
s
f
or
th
e
F
p
1
c
h
a
nne
l
i
n
th
e
D
E
A
P
da
t
a
s
e
t
is
il
lu
s
tr
a
te
d
in
T
a
bl
e
3
.
F
in
a
ll
y
,
a
s
in
di
c
a
t
e
d
in
T
a
bl
e
4,
t
hi
s
s
tu
d
y
in
ve
s
ti
g
a
te
d
th
e
G
a
us
s
i
a
n
f
il
te
r
a
ppr
oa
c
h
f
or
s
m
oot
hi
ng
th
e
b
a
s
e
li
n
e
s
ig
n
a
l.
T
he
s
m
oot
hi
ng pr
oc
e
s
s
f
or
th
e
F
p1 c
ha
n
ne
l
in
th
e
D
E
A
P
d
a
ta
s
e
t
i
s
il
lu
s
tr
a
t
e
d
in
T
a
bl
e
4.
A
c
c
or
di
ng
to
T
a
bl
e
s
2
t
o
4
,
th
e
D
E
A
P
d
a
ta
s
e
t'
s
b
a
s
e
l
in
e
s
ig
n
a
l
s
m
oo
th
e
d
z
j
f
or
th
r
e
e
s
e
c
ond
s
h
a
s
a
384
s
a
m
pl
i
ng
r
a
t
e
.
T
h
e
z
j
va
l
ue
s
i
n t
he
D
R
E
A
M
E
R
a
nd
A
M
I
G
O
S
da
ta
s
e
ts
a
r
e
640
s
a
m
p
le
r
a
te
s
or
5
s
e
c
o
nds
.
T
a
bl
e
3.
D
e
pi
c
t
s
t
he
s
m
oot
hi
ng pr
oc
e
dur
e
f
or
t
he
D
E
A
P
da
ta
s
e
t
us
in
g t
he
S
a
vi
tz
ky G
ol
a
y f
il
te
r
t
e
c
hni
que
on
th
e
ba
s
e
li
ne
s
ig
na
l
S
a
m
pl
i
ng r
a
t
e
B
a
s
e
l
i
ne
s
i
gna
l
(
)
S
a
vi
t
z
ky G
ol
a
y f
i
l
t
e
r
(
)
1
1
1
=
1
35
(
−
3
−
2
+
12
−
1
+
17
+
12
−
1
−
3
+
2
)
…
….
……………………………………………………
384
384
384
=
1
35
(
−
3
−
2
+
12
−
1
+
17
+
12
−
1
−
3
+
2
)
T
a
bl
e
4.
T
he
s
m
oot
hi
ng pr
oc
e
dur
e
f
or
t
he
D
E
A
P
da
ta
s
e
t
is
i
ll
us
tr
a
te
d us
in
g t
he
G
a
us
s
ia
n f
il
te
r
m
e
th
od on the
ba
s
e
li
ne
s
ig
n
a
l
S
a
m
pl
i
ng r
a
t
e
(
H
z
)
B
a
s
e
l
i
ne
s
i
gna
l
(
)
G
a
us
s
i
a
n f
i
l
t
e
r
(
z
i
)
1
1
1
=
1
√
2
2
−
(
1
−
)
2
2
2
…….
……..
……………………
384
384
384
=
1
√
2
2
−
(
384
−
)
2
2
2
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
13
, N
o.
4
,
D
e
c
e
m
be
r
2024
:
4263
-
4272
4266
2.4
.
D
e
c
om
p
os
it
io
n
T
hi
s
pr
oc
e
s
s
w
a
s
c
a
r
r
ie
d
out
f
or
a
ll
c
ha
nne
ls
f
or
bot
h
ba
s
e
li
n
e
a
nd
tr
ia
l
E
E
G
s
ig
na
ls
[
7]
.
T
a
bl
e
5
il
lu
s
tr
a
te
s
th
e
de
c
om
pos
it
io
n
pr
oc
e
s
s
of
th
e
E
E
G
ba
s
e
li
ne
s
ig
na
l
f
or
th
e
F
p1
c
ha
nne
l
.
I
n
th
e
D
E
A
P
da
ta
s
e
t,
3
s
e
c
onds
of
ba
s
e
li
ne
s
ig
na
ls
(
384
H
z
s
a
m
pl
in
g
r
a
te
)
w
e
r
e
de
c
om
pos
e
d
in
to
f
our
f
r
e
que
nc
y
ba
nds
f
or
e
a
c
h
c
ha
nne
l
[
20]
.
T
h
e
tr
ia
l
s
ig
n
a
l
de
c
om
po
s
it
io
n
pr
oc
e
s
s
w
a
s
a
ls
o c
a
r
r
ie
d
out
on
th
e
F
p1 c
ha
nne
l,
a
s
il
lu
s
tr
a
te
d
in
T
a
bl
e
6.
I
n
T
a
bl
e
6,
60 s
e
c
ond
s
of
tr
ia
l
s
ig
na
ls
(
7,680
H
z
s
a
m
p
li
ng
r
a
te
)
a
r
e
de
c
om
pos
e
d
in
to
f
our
f
r
e
que
nc
y
ba
nds
a
nd e
a
c
h c
ha
nne
l
in
one
e
xpe
r
im
e
nt
(
f
r
om
40 e
xpe
r
im
e
nt
s
i
n t
he
D
E
A
P
da
ta
s
e
t)
.
T
a
bl
e
5
.
F
r
e
que
nc
y ba
nd d
e
c
om
pos
it
io
n pr
oc
e
s
s
f
or
ba
s
e
li
ne
s
ig
na
ls
on t
he
D
E
A
P
da
ta
s
e
t
F
r
e
que
nc
y ba
nd
B
as
e
l
i
ne
s
i
gna
l
1
…..
384
T
he
t
a
z
1
Th
e
t
a
…..
z
384
Th
e
t
a
A
l
pha
z
1
Al
p
h
a
…..
z
384
Al
p
h
a
B
e
t
a
z
1
B
e
t
a
…..
z
384
B
e
t
a
G
a
m
m
a
z
1
G
a
m
m
a
…..
z
384
G
a
m
m
a
T
a
bl
e
6
.
D
e
c
om
pos
it
io
n pr
oc
e
s
s
f
or
60 s
e
c
onds
of
t
r
ia
l
s
ig
na
ls
on t
he
D
E
A
P
da
ta
s
e
t
F
r
e
que
nc
y ba
nd
T
r
i
a
l
s
i
gna
l
385
…..
8064
T
he
t
a
x
385
Th
e
t
a
…..
x
8064
Th
e
t
a
A
l
pha
x
385
Al
p
h
a
…..
x
8064
Al
p
h
a
B
e
t
a
x
385
B
e
t
a
…..
x
8064
B
e
t
a
G
a
m
m
a
x
385
G
a
m
m
a
…..
x
8064
G
a
m
m
a
2.5
.
F
e
at
u
r
e
e
xt
r
ac
t
io
n
T
he
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
pr
oc
e
s
s
is
c
a
r
r
ie
d
out
to
obt
a
in
r
e
le
va
nt
E
E
G
s
ig
na
l
f
e
a
tu
r
e
s
.
T
he
f
e
a
tu
r
e
e
xt
r
a
c
ti
on pr
oc
e
s
s
i
s
pe
r
f
or
m
e
d e
ve
r
y s
e
c
ond (
128
H
z
s
a
m
pl
in
g r
a
te
)
i
n
e
a
c
h f
r
e
que
nc
y ba
nd f
o
r
ba
s
e
li
ne
a
nd
tr
ia
l
s
ig
na
ls
.
T
he
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
pr
oc
e
s
s
in
th
is
r
e
s
e
a
r
c
h
us
e
s
th
e
D
E
m
e
th
od
[
7
]
,
[
8
]
.
B
a
s
e
d
on
th
is
f
or
m
ul
a
t
he
f
e
a
tu
r
e
e
xt
r
a
c
ti
on pr
oc
e
s
s
f
or
t
he
ba
s
e
li
ne
s
ig
n
a
l
on
c
ha
nne
l
F
p1 c
a
n be
i
ll
us
tr
a
te
d i
n
T
a
bl
e
7.
T
a
bl
e
7
.
I
ll
us
tr
a
ti
on of
t
he
f
e
a
tu
r
e
e
xt
r
a
c
ti
on pr
oc
e
s
s
f
or
ba
s
e
li
ne
s
ig
na
ls
on t
he
D
E
A
P
da
ta
s
e
t
F
r
e
que
nc
y ba
nd
S
a
m
pl
i
ng r
a
t
e
(
H
z
)
F
e
a
t
ur
e
va
l
ue
s
S
e
c
ond
T
he
t
a
z
1
Th
e
t
a
…
z
128
Th
e
t
a
h
1
(
Th
e
t
a
)
1
A
l
pha
z
1
Al
p
h
a
…
z
128
Al
p
h
a
h
1
(
Al
p
h
a
)
1
B
e
t
a
z
1
B
e
t
a
…
z
128
B
e
t
a
h
1
(
B
e
t
a
)
1
G
a
m
m
a
z
1
G
a
m
m
a
…
z
128
G
a
m
m
a
h
1
(
G
a
m
m
a
)
1
T
he
t
a
z
129
Th
e
t
a
…
z
256
Th
e
t
a
h
2
(
Th
e
t
a
)
2
A
l
pha
z
129
Al
p
h
a
…
z
256
Al
p
h
a
h
2
(
Al
p
h
a
)
2
B
e
t
a
z
129
B
e
t
a
…
z
256
B
e
t
a
h
2
(
B
e
t
a
)
2
G
a
m
m
a
z
129
G
a
m
m
a
…
z
256
G
a
m
m
a
h
2
(
G
a
m
m
a
)
2
T
he
t
a
z
257
Th
e
t
a
…
z
384
Th
e
t
a
h
3
(
Th
e
t
a
)
3
A
l
pha
z
257
Al
p
h
a
…
z
384
Al
p
h
a
h
3
(
Al
p
h
a
)
3
B
e
t
a
z
257
B
e
t
a
…
z
384
B
e
t
a
h
3
(
B
e
t
a
)
3
G
a
m
m
a
z
257
G
a
m
m
a
…
z
384
G
a
m
m
a
h
3
(
G
a
m
m
a
)
3
B
a
s
e
d
on
T
a
bl
e
7,
th
e
b
a
s
e
li
ne
s
ig
na
l
w
il
l
pr
oduc
e
th
r
e
e
D
E
f
e
a
tu
r
e
s
va
lu
e
s
f
or
e
a
c
h
c
ha
nne
l
a
nd
f
r
e
que
nc
y
ba
nd
in
one
e
xpe
r
im
e
nt
.
F
ur
th
e
r
,
th
e
pr
oc
e
s
s
of
e
xt
r
a
c
ti
ng
e
xpe
r
im
e
nt
a
l
s
ig
na
l
f
e
a
tu
r
e
s
on
th
e
F
p1
c
ha
nne
l
c
a
n
be
il
lu
s
tr
a
te
d
in
T
a
bl
e
8.
B
a
s
e
d
on
T
a
bl
e
8,
th
e
tr
i
a
l
s
ig
na
l
w
il
l
pr
oduc
e
60
D
E
f
e
a
tu
r
e
s
f
or
e
a
c
h
c
ha
nne
l
a
nd f
r
e
que
nc
y ba
nd i
n one
e
xpe
r
im
e
nt
.
T
a
bl
e
8
.
I
ll
us
tr
a
ti
on of
t
he
f
e
a
tu
r
e
e
xt
r
a
c
ti
on pr
oc
e
s
s
f
or
t
r
ia
l
s
i
gna
ls
on t
he
D
E
A
P
da
ta
s
e
t
F
r
e
que
nc
y ba
nd
S
a
m
pl
i
ng r
a
t
e
(
H
z
)
F
e
a
t
ur
e
va
l
ue
s
S
e
c
ond
T
he
t
a
385
ℎ
…
..
513
ℎ
ℎ
4
(
ℎ
)
4
A
l
pha
385
ℎ
…
..
513
ℎ
ℎ
4
(
ℎ
)
4
B
e
t
a
385
…
..
513
ℎ
4
(
)
4
G
am
m
a
385
…
..
513
ℎ
4
(
)
4
…….
…….
…
..
…….
…….
…….
T
he
t
a
7
9
3
6
ℎ
…
..
8064
ℎ
ℎ
63
(
ℎ
)
63
A
l
pha
7
9
3
6
ℎ
…
..
8064
ℎ
ℎ
63
(
ℎ
)
63
B
e
t
a
7
9
3
6
…
..
8064
ℎ
63
(
)
63
G
am
m
a
7
9
3
6
…
..
8064
ℎ
63
(
)
63
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
I
m
pr
ov
in
g bas
e
li
ne
r
e
duc
ti
on f
or
e
m
ot
io
n
r
e
c
ogni
ti
on bas
e
d on
…
(
I
M
ade
A
gus
W
ir
aw
an)
4267
2.6
.
B
as
e
li
n
e
r
e
d
u
c
t
io
n
T
he
ba
s
e
li
n
e
r
e
duc
ti
on
pr
oc
e
s
s
c
ont
in
ue
s
a
f
te
r
th
e
ba
s
e
li
ne
a
nd
tr
ia
l
s
ig
na
l
f
e
a
tu
r
e
va
lu
e
s
a
r
e
obt
a
in
e
d.
T
hi
s
pr
oc
e
s
s
a
im
s
to
pr
oduc
e
f
e
a
tu
r
e
va
lu
e
s
f
r
om
t
r
ia
l
s
ig
na
ls
th
a
t
c
a
n
c
ha
r
a
c
te
r
iz
e
pa
r
ti
c
ip
a
nt
s
'
e
m
ot
io
na
l
r
e
a
c
ti
ons
s
o
e
m
ot
io
ns
c
a
n
be
c
la
s
s
if
ie
d
m
or
e
a
c
c
ur
a
te
ly
[
7]
,
[
21]
.
T
hi
s
r
e
s
e
a
r
c
h
w
il
l
e
xa
m
in
e
th
e
r
e
la
ti
ve
di
f
f
e
r
e
nc
e
m
e
th
od
f
or
th
e
ba
s
e
li
ne
r
e
duc
ti
on
pr
o
c
e
s
s
.
T
hi
s
m
e
th
od
is
m
os
t
a
ppr
opr
ia
te
f
or
c
ha
r
a
c
te
r
iz
in
g
di
f
f
e
r
e
nt
E
E
G
s
ig
na
l
pa
tt
e
r
ns
f
r
om
e
a
c
h
pa
r
ti
c
ip
a
nt
.
T
he
f
or
m
ul
a
f
or
th
e
r
e
la
ti
ve
di
f
f
e
r
e
nc
e
m
e
th
od c
a
n be
pr
e
s
e
nt
e
d i
n
(
1
)
[
7]
–
[
9]
:
(
)
=
(
(
)
(
)
(
1)
T
he
a
ve
r
a
ge
va
lu
e
of
th
e
DE
(
B
a
s
e
M
e
a
n
)
f
e
a
tu
r
e
on
th
e
ba
s
e
li
ne
s
ig
na
l
is
c
a
lc
ul
a
te
d
a
s
a
n
in
it
ia
l
s
ta
ge
.
T
h
e
f
ol
lo
w
in
g
is
a
n
il
lu
s
tr
a
ti
on
of
th
e
c
a
l
c
ul
a
ti
on
pr
oc
e
s
s
f
or
th
e
a
ve
r
a
ge
D
E
f
e
a
tu
r
e
va
lu
e
f
or
e
a
c
h
f
r
e
que
nc
y
ba
nd i
n one
E
E
G
c
ha
nne
l
on D
E
A
P
:
_
ℎ
(
ℎ
)
=
(
ℎ
1
(
ℎ
)
+
ℎ
2
(
ℎ
)
+
ℎ
3
(
ℎ
)
)
3
_
ℎ
(
ℎ
)
=
(
ℎ
1
(
ℎ
)
+
ℎ
2
(
ℎ
)
+
ℎ
3
(
ℎ
)
)
3
_
ℎ
(
)
=
(
ℎ
1
(
)
+
ℎ
2
(
)
+
ℎ
3
(
)
)
3
_
ℎ
(
)
=
(
ℎ
1
(
)
+
ℎ
2
(
)
+
ℎ
3
(
)
)
3
T
he
m
e
a
n
va
lu
e
obt
a
in
e
d
is
th
e
n
u
s
e
d
f
or
th
e
b
a
s
e
li
ne
r
e
duc
ti
on
pr
oc
e
s
s
f
or
th
e
tr
ia
l
s
ig
na
l.
T
a
bl
e
9
pr
e
s
e
nt
s
th
e
ba
s
e
li
ne
r
e
duc
ti
on
pr
oc
e
s
s
us
in
g
th
e
r
e
la
ti
ve
di
f
f
e
r
e
nc
e
m
e
t
hod
f
or
one
c
ha
nne
l
in
one
e
xpe
r
im
e
nt
on
th
e
D
E
A
P
da
ta
s
e
t.
T
a
bl
e
9
.
I
ll
us
tr
a
ti
o
n
of
ba
s
e
li
ne
r
e
du
c
ti
on
c
a
l
c
u
la
t
io
ns
u
s
i
ng
th
e
r
e
l
a
ti
ve
di
f
f
e
r
e
n
c
e
m
e
t
hod
o
n t
he
D
E
A
P
d
a
ta
s
e
t
F
r
e
que
nc
y
I
l
l
us
t
r
a
t
i
on of
t
he
c
a
l
c
ul
a
t
i
on of
t
he
r
e
l
a
t
i
ve
di
f
f
e
r
e
nc
e
m
e
t
hod
T
he
t
a
_
ℎ
4
(
ℎ
)
=
ℎ
4
(
ℎ
)
_
ℎ
(
ℎ
)
T
he
t
a
_
ℎ
63
(
ℎ
)
=
ℎ
63
(
ℎ
)
_
ℎ
(
ℎ
)
A
l
pha
_
ℎ
4
(
ℎ
)
=
ℎ
4
(
ℎ
)
_
ℎ
(
ℎ
)
A
l
pha
_
ℎ
63
(
ℎ
)
=
ℎ
63
(
ℎ
)
_
ℎ
(
ℎ
)
B
e
t
a
_
ℎ
4
(
)
=
ℎ
4
(
)
_
ℎ
(
)
B
e
t
a
_
ℎ
63
(
)
=
ℎ
63
(
)
_
ℎ
(
)
G
am
m
a
_
ℎ
4
(
)
=
ℎ
4
(
)
_
ℎ
(
)
G
am
m
a
_
ℎ
63
(
)
=
ℎ
63
(
)
_
ℎ
(
)
2.7
.
F
e
at
u
r
e
’
s
r
e
p
r
e
s
e
n
t
at
io
n
T
he
f
e
a
tu
r
e
va
lu
e
s
of
th
e
tr
ia
l
s
ig
na
l
r
e
duc
e
d
in
th
e
ba
s
e
li
ne
r
e
duc
ti
on
pr
oc
e
s
s
a
r
e
r
e
pr
e
s
e
nt
e
d
u
s
in
g
th
e
3D
c
ube
m
e
th
od
.
F
e
a
tu
r
e
r
e
pr
e
s
e
nt
a
ti
on
u
s
in
g
th
e
3D
c
u
be
is
ba
s
e
d
on
th
e
in
te
r
na
ti
ona
l
s
ys
te
m
10
-
20
s
ta
nda
r
d
f
or
c
ha
nne
l
pl
a
c
e
m
e
nt
on
th
e
s
c
a
lp
.
T
h
e
in
te
r
na
ti
ona
l
10
-
20
s
ys
te
m
m
a
ps
th
e
s
pa
ti
a
l
in
f
or
m
a
ti
on
be
twe
e
n
c
ha
nn
e
ls
r
e
pr
e
s
e
nt
in
g
E
E
G
s
ig
na
l
f
e
a
tu
r
e
s
to
a
9
×
9
m
a
tr
ix
.
T
he
9×
9
m
a
tr
ix
de
s
c
r
ib
e
s
th
e
po
s
it
io
n
of
c
ha
nne
l
pl
a
c
e
m
e
nt
i
n t
he
he
a
d. F
ig
ur
e
2 r
e
pr
e
s
e
nt
s
D
E
f
e
a
tu
r
e
s
m
a
ppe
d i
nt
o a
9×
9 m
a
tr
ix
[
3]
,
[
7]
.
T
he
va
lu
e
s
of
_
ℎ
4
(
ℎ
)
,
_
ℎ
4
(
ℎ
)
,
_
ℎ
4
(
)
,
a
nd
_
ℎ
4
(
)
or
c
ha
nne
ls
F
p1
to
O2
a
r
e
pr
e
s
e
nt
e
d
in
f
our
9
×
9
m
a
tr
ix
.
T
he
9
×
9
m
a
tr
ix
r
e
pr
e
s
e
nt
s
th
e
ta
,
a
lp
ha
,
be
ta
,
a
nd
ga
m
m
a
f
r
e
que
nc
ie
s
.
T
he
n,
th
e
s
e
f
our
m
a
tr
ic
e
s
a
r
e
c
om
bi
ne
d
in
to
one
c
ube
c
a
ll
e
d
th
e
3D
c
ube
.
T
he
f
ir
s
t
3D
c
ube
r
e
pr
e
s
e
nt
s
1
s
e
c
ond
of
E
E
G
s
ig
na
l
f
e
a
tu
r
e
da
ta
f
or
a
ll
c
ha
nne
ls
a
nd f
our
f
r
e
que
nc
y ba
nds
, a
s
s
how
n i
n F
ig
ur
e
2.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
13
, N
o.
4
,
D
e
c
e
m
be
r
2024
:
4263
-
4272
4268
F
ig
ur
e
1
. I
ll
us
tr
a
ti
on of
f
e
a
tu
r
e
r
e
pr
e
s
e
nt
a
ti
on on c
ha
nne
ls
F
p1
to
O
2 f
or
e
a
c
h f
r
e
que
nc
y ba
nd us
in
g 3D
c
ube
G
a
m
m
a
,
be
ta
,
a
lp
ha
,
a
nd
th
e
ta
f
r
e
que
nc
ie
s
a
r
e
r
e
pr
e
s
e
nt
e
d
th
r
ough
a
9×
9
m
a
tr
ix
w
it
h
r
e
d,
y
e
ll
ow
,
gr
e
e
n,
a
nd
bl
ue
c
ol
or
s
.
T
he
D
E
A
P
a
nd
D
R
E
A
M
E
R
da
t
a
s
e
t
s
pr
oduc
e
2
,
400
a
nd
3
,
728
3D
c
ube
s
f
or
e
a
c
h
pa
r
ti
c
ip
a
nt
.
F
ur
th
e
r
m
or
e
,
one
pa
r
ti
c
ip
a
nt
ge
ne
r
a
te
d
6
,
192
3D
C
ube
s
f
or
lo
ng
e
xpe
r
im
e
nt
s
on
th
e
A
M
I
G
O
S
da
ta
s
e
t
a
nd 1
,
394 3D
c
ube
s
f
or
s
hor
t
e
xpe
r
im
e
nt
s
.
2.8
.
C
la
s
s
if
i
c
at
io
n
T
he
s
e
th
r
e
e
s
m
oot
hi
ng
a
ppr
oa
c
h
e
s
a
r
e
a
ppl
ie
d
in
th
e
C
N
N
m
e
th
od,
to
a
s
s
e
s
s
th
e
s
u
c
c
e
s
s
of
th
e
ba
s
e
li
ne
r
e
duc
ti
on
a
ppr
oa
c
h
[
7]
,
[
8]
,
[
22]
.
F
ig
ur
e
3
pr
e
s
e
nt
s
th
e
a
r
c
hi
te
c
tu
r
e
of
th
e
C
N
N
m
e
th
od
u
s
e
d
to
r
e
c
ogni
z
e
f
our
c
la
s
s
e
s
of
e
m
ot
io
ns
,
na
m
e
ly
hi
gh
a
r
ous
a
l
pos
it
iv
e
va
le
nc
e
(
H
A
P
V
)
,
hi
gh
a
r
ous
a
l
ne
ga
ti
ve
va
le
nc
e
(
H
A
N
V
)
,
lo
w
a
r
ous
a
l
po
s
i
ti
ve
va
le
n
c
e
(
L
A
P
V
)
,
a
nd
lo
w
a
r
ous
a
l
n
e
ga
ti
ve
va
l
e
nc
e
(
L
A
N
V
)
.
T
he
C
N
N
a
r
c
hi
te
c
tu
r
e
i
n F
ig
ur
e
3 us
e
s
i
nput
da
ta
f
r
om
da
ta
r
e
pr
e
s
e
nt
e
d i
n a
3D
c
ube
. T
hi
s
da
ta
i
s
t
he
n c
onvolve
d
f
our
ti
m
e
s
to
pr
oduc
e
f
our
f
e
a
tu
r
e
m
a
ps
.
T
he
f
ir
s
t
f
e
a
tu
r
e
m
a
p
m
e
a
s
ur
e
s
9
×
9
×
64,
th
e
s
e
c
ond
f
e
a
tu
r
e
m
a
p
m
e
a
s
ur
e
s
9
×
9
×
128,
th
e
th
ir
d
f
e
a
tu
r
e
m
a
p
m
e
a
s
ur
e
s
9
×
9
×
256,
a
nd
th
e
f
our
th
f
e
a
tu
r
e
m
a
p
m
e
a
s
ur
e
s
9
×
9
×
64.
T
he
v
a
lu
e
s
in
th
e
f
our
th
f
e
a
tu
r
e
m
a
p
a
r
e
r
e
s
h
a
pe
d
vi
a
th
e
f
la
tt
e
ni
ng
pr
oc
e
s
s
,
r
e
s
ul
ti
ng
in
a
s
c
a
la
r
va
lu
e
of
5
,
184
(
9
×
9
×
64)
.
T
hi
s
s
c
a
la
r
va
lu
e
is
us
e
d
a
s
in
put
da
ta
a
t
th
e
in
put
la
ye
r
node
.
N
e
xt
,
th
is
va
lu
e
is
f
ul
ly
c
onne
c
te
d
to
th
e
hi
dde
n
la
ye
r
w
it
h
1
,
024
node
s
.
F
in
a
ll
y,
th
e
hi
dde
n
node
is
f
ul
ly
c
onne
c
te
d
to
th
e
out
put
la
ye
r
.
I
n
F
ig
ur
e
3,
th
e
out
put
la
ye
r
is
f
our
node
s
.
T
he
s
e
f
our
node
s
r
e
pr
e
s
e
nt
th
e
f
our
r
e
c
ogni
z
e
d
c
la
s
s
e
s
of
e
m
ot
io
ns
.
A
pa
r
t
f
r
om
th
e
f
our
e
m
ot
io
n
c
la
s
s
e
s
,
th
is
r
e
s
e
a
r
c
h
w
il
l
a
ls
o
r
e
c
ogni
z
e
two
e
m
ot
io
n
c
la
s
s
e
s
,
hi
gh
a
nd
lo
w
,
f
or
th
e
e
m
ot
io
n’
s
a
r
ous
a
l
a
nd
va
le
nc
e
.
T
he
a
da
m
opt
im
iz
e
r
m
e
th
od
upda
te
s
th
e
w
e
ig
ht
va
lu
e
s
,
a
nd
L
2
r
e
gul
a
r
iz
a
ti
on
c
a
lc
ul
a
te
s
th
e
e
r
r
or
va
lu
e
s
.
I
n
a
ddi
ti
on,
th
e
le
a
r
ni
ng
r
a
te
,
ba
tc
h,
a
nd e
poc
h
pa
r
a
m
e
t
e
r
va
lu
e
s
a
r
e
s
e
t
a
t
0.0001, 128, a
nd 50, r
e
s
p
e
c
ti
ve
ly
.
2.9. As
s
e
s
s
m
e
n
t
m
od
e
l
an
d
p
ar
am
e
t
e
r
s
I
n
th
is
r
e
s
e
a
r
c
h,
m
ode
l
e
va
lu
a
ti
on
us
e
s
k
-
f
o
ld
c
r
os
s
-
va
li
da
ti
on,
w
he
r
e
th
e
k
va
lu
e
is
s
e
t
a
t
10
.
T
hr
ough
k
-
f
ol
d
c
r
os
s
-
va
li
da
ti
on,
th
e
pa
r
a
m
e
te
r
e
v
a
lu
a
ti
on
p
r
oc
e
s
s
c
a
n
be
c
a
r
r
ie
d
out
.
S
om
e
e
va
lu
a
ti
on
pa
r
a
m
e
te
r
s
a
r
e
a
c
c
ur
a
c
y,
e
r
r
or
,
pr
e
c
is
io
n,
r
e
c
a
ll
,
a
nd
F
1.
B
a
s
e
d
on
th
e
pr
e
c
is
io
n
a
nd
r
e
c
a
ll
va
lu
e
s
r
e
s
ul
ti
ng
f
r
om
t
he
t
w
o e
m
ot
io
n c
la
s
s
e
s
a
nd t
h
e
f
our
e
m
ot
io
n c
la
s
s
e
s
, t
he
F
1 va
lu
e
c
a
n be
c
a
lc
ul
a
te
d.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
I
m
pr
ov
in
g bas
e
li
ne
r
e
duc
ti
on f
or
e
m
ot
io
n
r
e
c
ogni
ti
on bas
e
d on
…
(
I
M
ade
A
gus
W
ir
aw
an)
4269
F
ig
ur
e
2
. C
N
N
a
r
c
hi
te
c
tu
r
e
3.
R
E
S
U
L
T
S
A
N
D
D
I
S
C
U
S
S
I
O
N
T
hi
s
s
tu
dy
w
il
l
e
m
pl
oy
th
r
e
e
da
ta
s
e
ts
,
D
E
A
P
,
D
R
E
A
M
E
R
,
a
nd
A
M
I
G
O
S
,
to
a
s
s
e
s
s
th
e
de
pe
nda
bi
li
ty
of
th
e
ba
s
e
li
ne
r
e
duc
ti
on
pr
oc
e
dur
e
on
va
r
io
us
da
ta
f
e
a
tu
r
e
s
.
T
a
bl
e
10
s
how
s
th
e
a
c
c
ur
a
c
y
va
lu
e
of
e
m
ot
io
n
r
e
c
ogni
ti
on
in
th
e
th
r
e
e
D
E
A
P
,
D
R
E
A
M
E
R
,
a
nd
A
M
I
G
O
S
da
ta
s
e
ts
.
A
c
c
or
di
ng
to
th
e
r
e
s
ul
ts
of
th
e
te
s
ts
,
th
e
G
a
us
s
ia
n
a
ppr
oa
c
h
de
li
ve
r
s
good
a
c
c
ur
a
c
y
onl
y
in
th
e
D
R
E
A
M
E
R
da
ta
s
e
t
f
or
th
e
two
a
nd
f
our
e
m
ot
io
n
c
la
s
s
e
s
.
O
n
th
e
ot
he
r
ha
nd,
th
e
m
e
a
n
f
il
te
r
a
ppr
oa
c
h
c
a
n
ge
ne
r
a
te
e
xc
e
ll
e
nt
a
c
c
ur
a
c
y
f
or
c
a
te
gor
iz
in
g
two
a
nd
f
our
e
m
ot
io
n
c
la
s
s
e
s
in
th
e
D
E
A
P
a
nd
A
M
I
G
O
S
da
ta
s
e
ts
.
T
he
s
ig
na
l
a
m
pl
it
ude
s
of
th
e
th
r
e
e
da
ta
s
e
ts
w
il
l
be
c
om
pa
r
e
d t
o de
te
r
m
in
e
t
he
s
our
c
e
of
t
hi
s
di
f
f
e
r
e
nc
e
.
T
a
bl
e
10
.
T
he
out
c
om
e
s
of
t
hr
e
e
s
m
oot
hi
ng me
th
ods
t
e
s
t
s
on t
h
r
e
e
da
ta
s
e
t
s
No
S
m
oot
hi
ng
E
m
ot
i
on
c
l
a
s
s
D
E
A
P
D
R
E
A
M
E
R
A
M
I
G
O
S
1
D
e
f
a
ul
t
F
our
c
l
a
s
s
79
.
83
81
.
01
94
.
06
2
G
a
us
s
i
a
n
F
our
c
l
a
s
s
72
.
77
92
.
88
93
.
72
3
M
e
a
n
F
our
c
l
a
s
s
95
.
48
79
.
79
97
.
61
4
S
a
vi
t
z
ky G
ol
a
y
F
our
c
l
a
s
s
79
.
22
81
.
82
94
.
48
5
D
e
f
a
ul
t
A
r
ous
a
l
92
.
70
92
.
53
94
.
06
6
G
a
us
s
i
a
n
A
r
ous
a
l
91
.
69
96
.
94
92
.
84
7
M
e
a
n
A
r
ous
a
l
98
.
58
92
.
06
94
.
53
8
S
a
vi
t
z
ky G
ol
a
y
A
r
ous
a
l
93
.
14
92
.
60
94
.
05
9
D
e
f
a
ul
t
V
a
l
e
nc
e
91
.
73
89
.
96
96
.
70
10
G
a
us
s
i
a
n
V
a
l
e
nc
e
90
.
48
96
.
05
96
.
10
11
M
e
a
n
V
a
l
e
nc
e
98
.
07
88
.
90
97
.
30
12
S
a
vi
t
z
ky G
ol
a
y
V
a
l
e
nc
e
92
.
83
90
.
23
96
.
40
F
ig
ur
e
4
de
pi
c
ts
th
e
ba
s
e
li
ne
s
ig
na
l
p
a
tt
e
r
n
f
or
th
e
e
ig
ht
h
pa
r
ti
c
ip
a
nt
in
th
e
e
ig
ht
h
e
xp
e
r
im
e
nt
be
f
or
e
a
nd
a
f
te
r
th
e
s
m
oot
hi
ng
ope
r
a
ti
on.
F
ig
ur
e
5
de
pi
c
ts
th
e
ba
s
e
li
n
e
s
ig
na
l
pa
tt
e
r
n
f
or
th
e
f
ir
s
t
pa
r
ti
c
ip
a
nt
in
th
e
f
ir
s
t
e
xpe
r
im
e
nt
be
f
or
e
a
nd
a
f
te
r
th
e
s
m
oot
hi
ng
te
c
hni
que
.
F
ig
ur
e
6
de
pi
c
ts
th
e
1
-
s
e
c
ond
r
a
w
ba
s
e
li
ne
s
ig
na
l
f
r
om
th
e
D
R
E
A
M
E
R
da
ta
s
e
t
f
or
th
e
te
nt
h
in
di
vi
dua
l
in
th
e
t
e
nt
h
e
xpe
r
im
e
nt
.
T
he
a
m
pl
it
ude
of
th
e
E
E
G
s
ig
na
l
s
houl
d
id
e
a
ll
y
be
in
th
e
100
µV
r
a
nge
[
23
]
,
[
24
]
.
N
on
e
th
e
le
s
s
,
w
he
n
c
a
pt
ur
in
g
th
e
ba
s
e
li
ne
s
ig
na
l,
di
f
f
e
r
e
nt
a
r
ti
f
a
c
ts
c
a
n
in
te
r
ve
ne
to
r
a
is
e
th
e
s
ig
na
l
a
m
pl
it
ude
.
B
a
s
e
d
on
a
c
om
pa
r
is
on
of
th
e
ba
s
e
li
ne
s
ig
na
l
pa
tt
e
r
ns
i
n t
he
t
hr
e
e
da
ta
s
e
t
s
(
F
ig
ur
e
s
4
to
6
)
, i
t
w
a
s
di
s
c
ove
r
e
d
th
a
t
th
e
a
m
pl
it
ude
of
t
he
ba
s
e
li
ne
s
ig
na
l
in
t
he
D
R
E
A
M
E
R
da
ta
s
e
t
is
gr
e
a
te
r
th
a
n
th
e
a
m
pl
it
ude
of
th
e
ba
s
e
li
ne
s
ig
na
l
in
th
e
D
E
A
P
a
nd
A
M
I
G
O
S
da
ta
s
e
ts
.
A
s
a
r
e
s
ul
t,
u
s
in
g
th
e
G
a
us
s
ia
n
f
il
te
r
m
e
th
od
c
a
n
r
e
duc
e
th
e
a
m
pl
it
ude
of
th
e
E
E
G
s
ig
na
l
to
a
n
id
e
a
l
v
a
lu
e
w
hi
le
pr
oduc
in
g
m
or
e
a
c
c
ur
a
c
y
th
a
n
us
in
g
th
e
m
e
a
n
f
il
te
r
a
ppr
oa
c
h.
T
he
G
a
us
s
ia
n
f
il
te
r
a
ppr
oa
c
h,
on
th
e
ot
he
r
ha
nd, a
lt
e
r
s
t
he
E
E
G
s
ig
na
l
pa
tt
e
r
n.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
13
, N
o.
4
,
D
e
c
e
m
be
r
2024
:
4263
-
4272
4270
F
ig
ur
e
3
. A
M
I
G
O
S
c
ha
nne
l
A
F
3 r
a
w
E
E
G
s
ig
na
l
da
ta
F
ig
ur
e
4
. D
E
A
P
c
ha
nne
l
F
p1 r
a
w
E
E
G
s
ig
na
l
da
ta
F
ig
ur
e
5
. D
R
E
A
M
E
R
'
s
r
a
w
E
E
G
s
ig
na
l
da
ta
f
or
c
ha
nne
l
A
F
3
T
he
ba
s
e
li
ne
r
e
duc
ti
on
s
tr
a
te
gy
ha
s
be
e
n
us
e
d
in
m
a
ny
s
tu
di
e
s
.
H
ow
e
ve
r
,
th
e
ba
s
e
li
ne
s
ig
na
l
e
m
pl
oye
d
in
th
is
s
tu
dy
s
ti
ll
c
ont
a
in
s
a
r
ti
f
a
c
ts
.
T
hi
s
phe
no
m
e
non
pr
e
ve
nt
s
th
e
b
a
s
e
li
ne
r
e
duc
ti
on
f
r
om
pe
r
f
or
m
in
g
opt
im
a
ll
y.
S
om
e
r
e
s
e
a
r
c
he
r
s
a
tt
e
m
pt
to
im
pr
ove
a
c
c
ur
a
c
y
by
a
ppl
yi
ng
a
de
e
p
le
a
r
ni
ng
a
ppr
oa
c
h
to
th
e
c
la
s
s
if
ic
a
ti
on
pr
oc
e
s
s
,
s
u
c
h
a
s
C
N
N
[
7]
,
c
a
ps
ul
e
n
e
two
r
k
[
22]
c
om
bi
na
ti
on
C
N
N
+
l
ong
s
hor
t
te
r
m
m
e
m
or
y
(
L
S
T
M
)
[
25]
,
a
nd de
e
p f
or
e
s
t
[
3
]
, but
t
hi
s
c
la
s
s
if
ic
a
ti
on me
th
od i
s
not
ye
t
c
a
pa
bl
e
of
r
e
c
ogni
z
in
g t
he
E
E
G
s
ig
na
l
pa
tt
e
r
n
th
a
t
is
f
or
m
e
d.
T
h
e
ba
s
e
li
ne
s
ig
na
l
s
m
oot
hi
ng
pr
oc
e
dur
e
c
a
n
im
pr
ove
th
e
ba
s
e
li
ne
r
e
duc
ti
on
a
ppr
oa
c
h.
T
he
im
pr
ove
d
ba
s
e
li
ne
r
e
duc
ti
on
m
e
th
od
ol
ogy
c
a
n
pr
ovi
de
E
E
G
s
ig
na
l
pa
tt
e
r
ns
e
a
s
il
y
r
e
c
ogni
z
e
d
by
c
la
s
s
if
ic
a
ti
on
s
ys
t
e
m
s
.
A
lt
hough
th
is
s
tu
dy
a
nd
r
e
s
e
a
r
c
he
r
s
[
8]
us
e
d
D
E
,
r
e
la
ti
ve
di
f
f
e
r
e
nc
e
,
3D
c
ube
,
a
nd
C
N
N
m
e
th
ods
f
or
e
a
c
h
pr
oc
e
s
s
,
th
e
ba
s
e
li
ne
s
i
gna
l
th
a
t
ha
d
b
e
e
n
s
m
oot
he
d
us
in
g
th
e
m
e
a
n
f
il
te
r
m
e
th
od
w
a
s
pr
ove
n
to
be
a
bl
e
to
im
pr
ove
th
e
pe
r
f
or
m
a
nc
e
of
th
e
ba
s
e
li
ne
r
e
duc
ti
on
pr
oc
e
s
s
.
T
he
s
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
I
m
pr
ov
in
g bas
e
li
ne
r
e
duc
ti
on f
or
e
m
ot
io
n
r
e
c
ogni
ti
on bas
e
d on
…
(
I
M
ade
A
gus
W
ir
aw
an)
4271
r
e
s
ul
ts
a
r
e
pr
ove
n
by
in
c
r
e
a
s
e
d
r
e
c
ogni
ti
on
a
c
c
ur
a
c
y
va
lu
e
s
f
or
two
a
nd
f
our
e
m
ot
io
n
c
la
s
s
e
s
in
th
e
D
E
A
P
a
nd
A
M
I
G
O
S
da
ta
s
e
ts
.
T
he
m
e
a
n
f
il
te
r
m
e
th
od
pr
opos
e
d
in
th
is
s
tu
dy
c
a
n
r
e
duc
e
ba
s
e
li
ne
s
ig
n
a
ls
c
ont
a
m
in
a
te
d
w
it
h
a
r
ti
f
a
c
ts
c
om
p
a
r
e
d
to
th
e
ot
he
r
two
s
m
oot
hi
ng
m
e
th
ods
.
A
lt
hough
th
e
m
e
a
n
f
il
te
r
a
ppr
oa
c
h
c
a
n
r
e
duc
e
th
e
a
m
pl
it
ude
of
th
e
b
a
s
e
li
ne
s
ig
na
l
a
nd
i
m
pr
ove
a
c
c
ur
a
c
y,
it
c
a
n
onl
y
b
e
us
e
d
on
E
E
G
s
ig
na
ls
w
it
h a
n a
m
pl
it
ude
of
200 µV.
M
e
a
nw
hi
le
, E
E
G
s
ig
na
ls
c
a
n only be
us
e
d i
n a
G
a
us
s
ia
n f
il
te
r
a
ppr
oa
c
h
f
or
s
ig
na
l
a
m
pl
it
ude
s
gr
e
a
te
r
th
a
n
4
,
000
µV.
T
he
s
e
r
e
s
ul
ts
a
r
e
pr
ove
n
by
in
c
r
e
a
s
e
d
r
e
c
ogni
ti
on
a
c
c
ur
a
c
y
va
lu
e
s
f
or
th
e
D
R
E
A
M
E
R
da
ta
s
e
t'
s
two
a
nd
f
our
e
m
ot
io
n
c
la
s
s
e
s
.
T
hi
s
pr
opos
e
d
m
e
th
od
c
a
n
r
e
du
c
e
a
r
ti
f
a
c
ts
in
th
e
ba
s
e
li
ne
s
ig
na
l
w
it
hout
c
ha
ngi
ng
th
e
e
xi
s
ti
ng
E
E
G
s
i
gna
l
pa
tt
e
r
n.
T
hi
s
r
e
s
e
a
r
c
h
c
om
pr
e
he
ns
iv
e
ly
r
e
duc
e
d
a
r
ti
f
a
c
ts
in
th
e
ba
s
e
li
ne
s
ig
na
l
us
in
g
th
e
m
e
a
n
f
il
te
r
a
nd
G
a
us
s
ia
n
m
e
th
ods
.
H
ow
e
ve
r
,
in
-
de
pt
h
s
tu
di
e
s
m
a
y
be
ne
e
de
d
to
d
e
te
r
m
in
e
th
e
be
s
t
m
e
th
od
f
or
r
e
duc
in
g
a
r
ti
f
a
c
ts
in
th
e
th
r
e
e
da
ta
s
e
ts
.
I
n
a
ddi
ti
on,
f
ur
th
e
r
s
tu
di
e
s
a
r
e
e
s
s
e
nt
i
a
l
to
im
pr
ove
th
e
c
la
s
s
if
ic
a
ti
on
pr
oc
e
s
s
by
de
te
r
m
in
in
g
a
ppr
opr
ia
te
c
la
s
s
if
ic
a
ti
on
m
e
th
ods
.
U
s
in
g
th
e
C
N
N
m
e
th
od
in
th
is
r
e
s
e
a
r
c
h
c
a
n
r
e
s
ul
t
in
th
e
lo
s
s
of
s
pa
ti
a
l
in
f
o
r
m
a
ti
on
f
r
om
th
e
E
E
G
s
ig
na
l
[
21]
, [
22]
, [
26]
.
4.
C
O
N
C
L
U
S
I
O
N
O
ur
f
in
di
ngs
pr
ovi
de
c
onc
lu
s
iv
e
e
vi
de
n
c
e
th
a
t
th
e
s
m
oot
hi
ng
pr
oc
e
dur
e
us
in
g
th
e
m
e
a
n
f
il
te
r
a
nd
G
a
us
s
ia
n
f
il
te
r
a
ppr
oa
c
he
s
a
r
e
pr
ove
n
to
e
li
m
in
a
te
a
r
ti
f
a
c
ts
in
t
he
ba
s
e
li
ne
s
ig
na
l.
T
he
s
e
two
a
ppr
oa
c
h
e
s
c
a
n
in
c
r
e
a
s
e
ba
s
e
li
ne
r
e
duc
ti
on
pe
r
f
or
m
a
nc
e
by
s
m
oot
hi
ng
th
e
ba
s
e
li
ne
s
ig
na
l.
T
he
r
e
la
ti
ve
di
f
f
e
r
e
nc
e
a
ppr
oa
c
h
w
a
s
e
m
pl
oye
d
in
th
is
s
tu
dy'
s
ba
s
e
li
ne
r
e
duc
ti
on
pr
oc
e
s
s
to
ge
ne
r
a
te
E
E
G
s
ig
na
l
pa
tt
e
r
ns
th
a
t
a
r
e
e
a
s
y
to
r
e
c
ogni
z
e
th
r
oughout
th
e
c
la
s
s
if
ic
a
ti
on
pha
s
e
,
w
hi
c
h
im
pa
c
ts
in
c
r
e
a
s
in
g
a
c
c
ur
a
c
y.
T
he
C
N
N
m
e
th
od
f
or
c
la
s
s
if
ic
a
ti
on
a
c
hi
e
ve
s
th
e
be
s
t
a
c
c
ur
a
c
y
va
lu
e
by
in
tr
oduc
in
g
two
c
la
s
s
e
s
a
nd
f
our
c
la
s
s
e
s
of
e
m
ot
io
n
s
.
H
ow
e
ve
r
,
in
-
de
pt
h
s
tu
di
e
s
m
a
y
be
n
e
e
de
d
to
d
e
te
r
m
in
e
th
e
b
e
s
t
m
e
th
od
f
or
r
e
duc
in
g
a
r
ti
f
a
c
ts
in
th
e
th
r
e
e
da
ta
s
e
ts
.
I
n
a
ddi
ti
on,
f
ur
th
e
r
s
tu
di
e
s
a
r
e
e
s
s
e
nt
ia
l
to
im
pr
ove
th
e
c
la
s
s
if
ic
a
ti
on
pr
oc
e
s
s
by
de
te
r
m
in
in
g
a
ppr
opr
ia
te
c
la
s
s
if
ic
a
ti
on
m
e
th
ods
.
U
s
in
g
th
e
C
N
N
m
e
th
od
in
th
is
r
e
s
e
a
r
c
h
c
a
n
r
e
s
ul
t
in
th
e
lo
s
s
of
s
pa
ti
a
l
in
f
or
m
a
ti
on f
r
o
m
t
he
E
E
G
s
ig
na
l
.
A
C
K
N
O
WL
E
D
G
E
M
E
N
T
S
T
hi
s
w
or
k
is
f
ul
ly
s
uppor
te
d
by
F
undi
ng
f
or
th
e
2023
f
is
c
a
l
ye
a
r
,
N
o:
643/
U
N
48.16/L
T
/2
023,
f
r
om
th
e
I
ns
ti
tu
te
f
or
R
e
s
e
a
r
c
h a
nd C
om
m
uni
ty
S
e
r
vi
c
e
a
t
th
e
U
ni
ve
r
s
it
a
s
P
e
ndi
di
ka
n G
a
ne
s
ha
.
R
E
F
E
R
E
N
C
E
S
[
1]
H
.
C
ha
o,
L
.
D
ong,
Y
.
L
i
u,
a
nd
B
.
L
u,
“
E
m
ot
i
on
r
e
c
ogni
t
i
on
f
r
om
m
ul
t
i
ba
nd
E
E
G
s
i
gna
l
s
u
s
i
ng
c
a
p
s
ne
t
,”
Se
n
s
or
s
,
vol
.
19,
no.
9,
pp. 1
–
16,
2019, doi
:
10.3390/
s
19092212.
[
2]
L
.
S
hu
e
t
al
.
,
“
A
r
e
vi
e
w
of
e
m
ot
i
on
r
e
c
ogni
t
i
on
us
i
ng
phy
s
i
ol
ogi
c
a
l
s
i
gn
a
l
s
,”
Se
ns
o
r
s
,
vol
.
18,
no.
7,
pp.
1
–
41,
2018,
doi
:
10.3390/
s
18072074.
[
3]
J
.
C
he
ng
e
t
al
.
,
“
E
m
ot
i
on
r
e
c
ogni
t
i
on
f
r
om
m
ul
t
i
-
c
ha
nne
l
E
E
G
vi
a
de
e
p
f
or
e
s
t
,”
I
E
E
E
J
our
nal
of
B
i
om
e
di
c
al
and
H
e
al
t
h
I
nf
or
m
at
i
c
s
, vol
. 25, no. 2, pp. 453
–
464, 2021, doi
:
10.1109/
J
B
H
I
.2020.2995767.
[
4]
A
.
B
.
U
s
a
kl
i
,
“
I
m
pr
ove
m
e
nt
of
E
E
G
s
i
gna
l
a
c
qui
s
i
t
i
on:
a
n
e
l
e
c
t
r
i
c
a
l
a
s
p
e
c
t
f
or
s
t
a
t
e
of
t
he
a
r
t
of
f
r
ont
e
nd,”
C
om
put
at
i
onal
I
nt
e
l
l
i
ge
nc
e
and N
e
ur
os
c
i
e
nc
e
, vol
. 2010,
pp. 1
–
8,
2010, doi
:
10.1155/
2010/
63
0649.
[
5]
R
.
W
a
r
d
o
y
o,
I
.
M
.
A
.
W
i
r
a
w
a
n
,
a
n
d
I
.
G
.
A
.
P
r
a
d
i
p
t
a
,
“
O
v
e
r
s
a
m
p
l
i
ng
a
p
p
r
oa
c
h
u
s
i
n
g
r
a
d
i
us
-
S
M
O
T
E
f
o
r
i
m
b
a
l
a
n
c
e
e
l
e
c
t
r
oe
n
c
e
p
ha
l
o
g
r
a
p
h
y
da
t
a
s
e
t
s
,”
E
m
e
r
g
i
n
g
Sc
i
e
nc
e
J
o
ur
n
a
l
,
v
o
l
.
6
,
n
o.
2
,
p
p
.
38
2
–
3
9
8
,
2
02
2
,
do
i
:
1
0.
2
8
9
91
/
E
S
J
-
2022
-
06
-
02
-
01
3
.
[
6]
I
.
M
.
A
.
W
i
r
a
w
a
n,
R
.
W
a
r
doyo,
a
nd
D
.
L
e
l
ono,
“
T
he
c
ha
l
l
e
nge
s
of
e
m
ot
i
on
r
e
c
ogni
t
i
on
m
e
t
hods
ba
s
e
d
on
e
l
e
c
t
r
oe
nc
e
pha
l
ogr
a
m
s
i
gna
l
s
:
A
l
i
t
e
r
a
t
ur
e
r
e
vi
e
w
,”
I
nt
e
r
nat
i
onal
J
our
nal
of
E
l
e
c
t
r
i
c
al
and
C
om
put
e
r
E
ngi
ne
e
r
i
ng
,
vol
.
12,
no.
2,
pp.
1508
–
1519,
2022,
doi
:
10.11591/
i
j
e
c
e
.v12i
2.pp1508
-
1519.
[
7]
Y
.
Y
a
ng,
Q
.
W
u,
Y
.
F
u,
a
nd
X
.
C
he
n,
“
C
ont
i
nuous
c
onvol
ut
i
ona
l
ne
ur
a
l
ne
t
w
or
k
w
i
t
h
3D
i
nput
f
or
E
E
G
-
ba
s
e
d
e
m
ot
i
on
r
e
c
ogni
t
i
on,”
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on
N
e
u
r
al
I
nf
or
m
at
i
on
P
r
oc
e
s
s
i
ng
,
vo
l
.
11307,
pp.
433
–
443,
2018,
doi
:
10.1007/
978
-
3
-
030
-
04239
-
4_39.
[
8]
I
.
M
.
A
.
W
i
r
a
w
a
n,
R
.
W
a
r
doyo,
D
.
L
e
l
ono,
S
.
K
us
r
ohm
a
ni
a
h,
a
nd
S
.
A
s
r
or
i
,
“
C
om
pa
r
i
s
on
of
b
a
s
e
l
i
ne
r
e
du
c
t
i
on
m
e
t
hods
f
or
e
m
ot
i
on
r
e
c
ogni
t
i
on
ba
s
e
d
on
e
l
e
c
t
r
oe
nc
e
pha
l
ogr
a
m
s
i
gna
l
s
,”
2021
6t
h
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on
I
nf
or
m
at
i
c
s
and
C
om
put
i
ng,
I
C
I
C
, 2021,
pp. 1
-
7,
doi
:
10.1109/
I
C
I
C
54025.2021.9632948.
[
9]
D
.
L
e
l
ono,
H
.
N
ur
a
di
,
M
.
R
.
S
a
t
r
i
yo,
T
.
W
.
W
i
dodo,
A
.
D
ha
r
m
a
w
a
n,
a
nd
J
.
E
.
I
s
t
i
ya
nt
o,
“
C
om
pa
r
i
s
on
of
di
f
f
e
r
e
nc
e
,
r
e
l
a
t
i
ve
a
nd
f
r
a
c
t
i
ona
l
m
e
t
hods
f
or
c
l
a
s
s
i
f
i
c
a
t
i
on
of
t
he
bl
a
c
k
t
e
a
ba
s
e
d
on
e
l
e
c
t
r
oni
c
nos
e
,”
2019
I
nt
e
r
nat
i
onal
C
onf
e
r
e
n
c
e
on
C
o
m
put
e
r
E
ngi
ne
e
r
i
ng, N
e
t
w
or
k
, and I
nt
e
l
l
i
ge
nt
M
ul
t
i
m
e
di
a, C
E
N
I
M
, 2019,
pp. 1
-
7,
doi
:
10.1109/
C
E
N
I
M
48368.2019.8973308.
[
10]
K
. G
a
s
pe
r
, “
U
t
i
l
i
z
i
ng ne
ut
r
a
l
a
f
f
e
c
t
i
ve
s
t
a
t
e
s
i
n r
e
s
e
a
r
c
h:
t
he
or
y, a
s
s
e
s
s
m
e
nt
,
a
n
d r
e
c
om
m
e
nda
t
i
ons
,”
E
m
ot
i
on R
e
v
i
e
w
, vol
. 10, no.
3, pp. 255
–
266, 2018, doi
:
10.1177/
1754073918765660.
[
11]
K
.
G
a
s
p
e
r
,
L
.
A
.
S
pe
n
c
e
r
,
a
nd
D
.
H
u,
“
D
oe
s
ne
ut
r
a
l
a
f
f
e
c
t
e
xi
s
t
?
H
ow
c
ha
l
l
e
n
gi
ng
t
hr
e
e
be
l
i
e
f
s
a
bout
ne
ut
r
a
l
a
f
f
e
c
t
c
a
n
a
dva
n
ce
a
f
f
e
c
t
i
ve
r
e
s
e
a
r
c
h,”
F
r
ont
i
e
r
s
i
n P
s
y
c
hol
ogy
, vol
. 10, 2019,
pp. 1
-
11
,
doi
:
10.3389/
f
ps
yg.2019.02476.
[
12]
X
. J
i
a
ng, G
. B
.
B
i
a
n,
a
nd Z
. T
i
a
n,
“
R
e
m
ova
l
of
a
r
t
i
f
a
c
t
s
f
r
om
E
E
G
s
i
gna
l
s
:
a
r
e
vi
e
w
,”
Se
ns
o
r
s
,
vol
. 19,
no. 5,
pp.
1
-
18
,
2019, doi
:
10.3390/
s
19050987.
[
13]
P
.
K
ow
a
l
s
ki
a
nd
R
.
S
m
yk,
“
R
e
vi
e
w
a
nd
c
om
pa
r
i
s
on
of
s
m
oot
hi
ng
a
l
gor
i
t
hm
s
f
or
one
-
di
m
e
ns
i
ona
l
da
t
a
noi
s
e
r
e
du
c
t
i
on,”
2018
I
nt
e
r
nat
i
onal
I
nt
e
r
di
s
c
i
pl
i
nar
y
P
hD
W
or
k
s
hop, I
I
P
hD
W
2018
, pp. 277
–
281, 2018, doi
:
10.1109/
I
I
P
H
D
W
.2018.8388373.
[
14]
K
.
G
.
G
opa
n,
N
. S
i
nha
,
a
n
d J
.
D
.
B
a
b
u,
“
S
t
a
t
i
s
t
i
c
a
l
f
e
a
t
u
r
e
a
na
l
ys
i
s
f
o
r
E
E
G
ba
s
e
l
i
ne
c
l
a
s
s
i
f
i
c
a
t
i
o
n:
e
ye
s
op
e
n vs
e
ye
s
c
l
os
e
d
,”
I
E
E
E
R
e
g
i
o
n
10
A
nnu
al
I
nt
e
r
na
t
i
on
al
C
on
f
e
r
e
nc
e
,
P
r
oc
e
e
di
ngs
/
T
E
N
C
O
N
,
pp
. 2
46
6
–
2
4
69,
20
17
, d
oi
:
10.
11
09
/
T
E
N
C
O
N
.20
16
.78
48
476
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
13
, N
o.
4
,
D
e
c
e
m
be
r
2024
:
4263
-
4272
4272
[
15]
N
.
Z
hua
ng,
Y
.
Z
e
ng,
K
.
Y
a
ng,
C
.
Z
ha
ng,
L
.
T
ong,
a
nd
B
.
Y
a
n,
“
I
nve
s
t
i
ga
t
i
ng
pa
t
t
e
r
ns
f
or
s
e
l
f
-
i
nduc
e
d
e
m
ot
i
on
r
e
c
ogni
t
i
on
f
r
om
E
E
G
s
i
gna
l
s
,”
Se
ns
or
s
, vol
. 18, no. 3,
pp. 1
-
22
,
2018, doi
:
10.3390/
s
18030841.
[
16]
S
.
K
oe
l
s
t
r
a
e
t
al
.
,
“
D
E
A
P
:
A
da
t
a
ba
s
e
f
or
e
m
ot
i
on
a
n
a
l
ys
i
s
;
u
s
i
ng
phy
s
i
o
l
ogi
c
a
l
s
i
gna
l
s
,
”
I
E
E
E
T
r
ans
ac
t
i
ons
on
A
f
f
e
c
t
i
v
e
C
om
put
i
ng
, vol
. 3, no. 1, pp. 18
–
31, 2012, doi
:
10.1109/
T
-
A
F
F
C
.2011.15.
[
17]
S
.
K
a
t
s
i
gi
a
nni
s
a
nd
N
.
R
a
m
z
a
n,
“
D
R
E
A
M
E
R
:
a
d
a
t
a
ba
s
e
f
or
e
m
ot
i
on
r
e
c
ogni
t
i
on
t
hr
ough
E
E
G
a
nd
E
C
G
s
i
gna
l
s
f
r
om
w
i
r
e
l
e
s
s
l
ow
-
c
os
t
of
f
-
t
he
-
s
he
l
f
de
vi
c
e
s
,”
I
E
E
E
J
our
nal
of
B
i
om
e
di
c
al
and
H
e
al
t
h
I
nf
or
m
at
i
c
s
,
vol
.
22,
no.
1,
pp.
98
–
107,
2018,
doi
:
10.1109/
J
B
H
I
.2017.2688239.
[
18]
J
.
A
.
M
.
-
C
or
r
e
a
,
M
.
K
.
A
ba
di
,
N
.
S
e
be
,
a
nd
I
.
P
a
t
r
a
s
,
“
A
M
I
G
O
S
:
a
da
t
a
s
e
t
f
or
a
f
f
e
c
t
,
pe
r
s
ona
l
i
t
y
a
nd
m
ood
r
e
s
e
a
r
c
h
o
n
i
ndi
vi
dua
l
s
a
nd
gr
oups
,”
I
E
E
E
T
r
ans
ac
t
i
ons
on
A
f
f
e
c
t
i
v
e
C
o
m
put
i
ng
,
vol
.
12,
no.
2,
pp.
479
–
493,
2021,
doi
:
10.1109/
T
A
F
F
C
.2018.2884461.
[
19]
A
.
K
.
-
S
t
e
r
ni
uk
e
t
al
.
,
“
C
om
pa
r
i
s
on
of
s
m
oot
hi
ng
f
i
l
t
e
r
s
i
n
a
na
l
ys
i
s
of
E
E
G
d
a
t
a
f
or
t
he
m
e
di
c
a
l
di
a
gnos
t
i
c
s
pur
pos
e
s
,”
Se
ns
or
s
,
vol
. 20, no. 3,
pp. 1
-
19
,
2020, doi
:
10.3390/
s
20030807.
[
20]
C
.
P
a
n,
C
.
S
hi
,
H
.
M
u,
J
.
L
i
,
a
nd
X
.
G
a
o,
“
E
E
G
-
ba
s
e
d
e
m
ot
i
on
r
e
c
ogni
t
i
on
us
i
ng
l
ogi
s
t
i
c
r
e
gr
e
s
s
i
on
w
i
t
h
ga
us
s
i
a
n
k
e
r
ne
l
a
n
d
l
a
pl
a
c
i
a
n
pr
i
or
a
nd
i
nve
s
t
i
ga
t
i
on
of
c
r
i
t
i
c
a
l
f
r
e
que
nc
y
ba
nd
s
,”
A
ppl
i
e
d
Sc
i
e
nc
e
s
,
vol
.
10,
no.
5,
pp.
1
-
25
,
2020,
doi
:
10.3390/
a
pp10051619.
[
21]
Y
.
L
i
u
e
t
al
.
,
“
M
ul
t
i
-
c
ha
nne
l
E
E
G
-
ba
s
e
d
e
m
ot
i
on
r
e
c
ogni
t
i
on
vi
a
a
m
ul
t
i
-
l
e
ve
l
f
e
a
t
ur
e
s
gui
de
d
c
a
ps
ul
e
ne
t
w
or
k,”
C
om
put
e
r
s
i
n
B
i
ol
ogy
and M
e
di
c
i
ne
, vol
. 123, 2020, doi
:
10.1016/
j
.c
om
pbi
om
e
d.2020.10392
7.
[
22]
I
.
M
.
A
.
W
i
r
a
w
a
n,
R
.
W
a
r
doyo,
D
.
L
e
l
ono,
a
nd
S
.
K
us
r
ohm
a
ni
a
h,
“
C
o
nt
i
nuous
c
a
ps
ul
e
ne
t
w
or
k
m
e
t
hod
f
or
i
m
pr
ovi
ng
e
l
e
c
t
r
oe
nc
e
pha
l
ogr
a
m
-
ba
s
e
d
e
m
ot
i
on
r
e
c
ogni
t
i
on,”
E
m
e
r
gi
ng
Sc
i
e
nc
e
J
ou
r
nal
,
vol
.
7,
no.
1,
pp.
116
–
134,
2023,
doi
:
10.28991/
E
S
J
-
2023
-
07
-
01
-
09.
[
23]
V
.
R
oy
a
nd
S
.
S
hukl
a
,
“
D
e
s
i
gni
ng
e
f
f
i
c
i
e
nt
bl
i
nd
s
our
c
e
s
e
pa
r
a
t
i
on
m
e
t
hods
f
o
r
E
E
G
m
ot
i
on
a
r
t
i
f
a
c
t
r
e
m
ova
l
ba
s
e
d
on
s
t
a
t
i
s
t
i
c
a
l
e
va
l
ua
t
i
on,”
W
i
r
e
l
e
s
s
P
e
r
s
onal
C
om
m
uni
c
at
i
ons
, vol
. 108, no. 3, pp. 1311
–
132
7, 2019, doi
:
10.1007/
s
11277
-
019
-
06470
-
3.
[
24]
M
. T
e
pl
a
n, “
F
unda
m
e
nt
a
l
s
of
E
E
G
m
e
a
s
ur
e
m
e
nt
,”
M
e
as
ur
e
m
e
nt
Sc
i
e
nc
e
R
e
v
i
e
w
, vol
. 2, pp. 1
–
11, 2002
.
[
25]
Y
.
Y
a
ng,
Q
.
W
u,
M
.
Q
i
u,
Y
.
W
a
ng,
a
nd
X
.
C
he
n,
“
E
m
o
t
i
on
r
e
c
ogni
t
i
on
f
r
om
m
ul
t
i
-
c
ha
nne
l
E
E
G
t
hr
ough
pa
r
a
l
l
e
l
c
onvol
ut
i
ona
l
r
e
c
ur
r
e
nt
ne
ur
a
l
ne
t
w
or
k,”
2018
I
nt
e
r
nat
i
onal
J
oi
nt
C
onf
e
r
e
nc
e
on
N
e
ur
al
N
e
t
w
o
r
k
s
(
I
J
C
N
N
)
,
2018,
pp.
1
-
7,
doi
:
10.1109/
I
J
C
N
N
.2018.8489331.
[
26]
S
.
S
a
bour
,
N
.
F
r
os
s
t
,
a
nd
G
.
E
.
H
i
nt
on,
“
D
yna
m
i
c
r
ou
t
i
ng
be
t
w
e
e
n
c
a
ps
ul
e
s
,”
A
dv
anc
e
s
i
n
N
e
u
r
al
I
nf
or
m
at
i
on
P
r
oc
e
s
s
i
n
g
Sy
s
t
e
m
s
, vol
. 2017, pp. 3857
–
3867, 2017.
B
I
O
G
R
A
P
H
I
E
S
O
F
A
U
T
H
O
R
S
I
Made
Agus
Wirawan
is
a
lecturer
in
the
Department
of
Informatics
Engineering, Faculty of Engineering
and Vocational Studies,
Universi
tas Pendidikan Ganesha,
Bali,
Indonesia.
His
areas
of
research
interest
include
machine
learnin
g,
artificial
intelligence,
deep
learning,
and
EEG
signals
.
He
can
be
c
ontacted
at
email:
imade.agusw
irawan@undiksha.ac.id
.
I
Gede
Mahendra
Darmawiguna
is
a
lecturer
in
the
Department
of
Informatics
Engineering, Faculty of Engineering
and
Vocational Stud
ies,
Universi
tas Pendidikan
Ganesha,
Bali,
Indonesia.
His
areas
of
research
interest
include
machine
learnin
g,
artificial
intelligence,
and deep l
earning
.
He can be contacted at email:
mahendra
.darmawig
una@
undiksha.ac
.id
.
Ida
Bagus
Nyoman
Pascima
is
a
lecturer
in
the
Department
of
Informatic
s
Engineering, Faculty of Engineering
and Vocational Studies,
Universi
tas Pendidikan Ganesha,
Bali,
Indonesia.
His
areas
of
research
interest
include
machine
learnin
g,
artificial
intell
igence,
and deep l
earning
.
He can be contacted at email:
gus.pascima@
undiksha.ac.id
.
Evaluation Warning : The document was created with Spire.PDF for Python.