I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
1
6
,
No
.
3
,
J
u
n
e
20
2
6
,
p
p
.
1576
~
1
5
8
7
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
1
6
i
3
.
pp
1
5
7
6
-
1
5
8
7
1576
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Enha
ncing
sE
M
G
finger g
est
ure
re
co
g
nition us
ing
optimized
1D
-
co
nv
o
lutiona
l neural net
wo
rk
Da
niel
Su
t
o
po
P
a
m
un
g
k
a
s
,
Su
m
a
ntr
i K
.
Ris
a
nd
riy
a
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
r
i
n
g
,
P
o
l
i
t
e
k
n
i
k
N
e
g
e
r
i
B
a
t
a
m
,
B
a
t
a
m,
I
n
d
o
n
e
si
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Feb
1
8
,
2
0
2
6
R
ev
is
ed
Ma
r
3
1
,
2
0
2
6
Acc
ep
ted
Ap
r
2
6
,
2
0
2
6
Ro
b
u
s
t
a
n
d
p
re
c
ise
fi
n
g
e
r
g
e
s
t
u
re
re
c
o
g
n
i
ti
o
n
u
s
i
n
g
s
u
rfa
c
e
e
lec
t
r
o
m
y
o
g
ra
p
h
y
(sE
M
G
)
is
e
s
se
n
t
ia
l
f
o
r
d
e
v
e
lo
p
i
n
g
i
n
t
u
it
i
v
e
p
r
o
s
t
h
e
t
ic
c
o
n
t
r
o
l
s
y
s
tem
s
.
Ho
we
v
e
r,
sEM
G
sig
n
a
ls
a
re
in
h
e
re
n
tl
y
st
o
c
h
a
stic
a
n
d
n
o
n
-
sta
ti
o
n
a
ry
,
p
o
sin
g
sig
n
ifi
c
a
n
t
c
h
a
ll
e
n
g
e
s
f
o
r
h
ig
h
-
a
c
c
u
ra
c
y
c
las
sifica
ti
o
n
in
fi
n
e
-
g
ra
in
e
d
m
o
v
e
m
e
n
ts.
Th
is
st
u
d
y
p
r
o
p
o
se
s
a
n
o
p
t
imiz
e
d
1
D
c
o
n
v
o
lu
t
io
n
a
l
n
e
u
ra
l
n
e
two
rk
(1
D
-
CNN
)
fra
m
e
wo
rk
f
o
r
c
las
sify
in
g
2
0
d
isti
n
c
t
fin
e
-
g
ra
in
e
d
fi
n
g
e
r
g
e
stu
re
s
u
sin
g
ra
w
sEM
G
d
a
ta
f
ro
m
a
n
8
-
c
h
a
n
n
e
l
we
a
ra
b
le
M
y
o
Arm
b
a
n
d
se
n
s
o
r.
Un
li
k
e
trad
it
i
o
n
a
l
m
e
th
o
d
s
th
a
t
re
ly
o
n
m
a
n
u
a
l
fe
a
tu
re
e
n
g
i
n
e
e
rin
g
,
th
e
p
ro
p
o
se
d
1
D
-
CNN
p
e
rfo
rm
s
e
n
d
-
to
-
e
n
d
lea
rn
i
n
g
to
a
u
t
o
m
a
ti
c
a
ll
y
e
x
trac
t
tem
p
o
ra
l
fe
a
tu
re
s.
Th
e
re
se
a
rc
h
sp
e
c
ifi
c
a
ll
y
in
v
e
stig
a
tes
th
e
imp
a
c
t
o
f
tem
p
o
ra
l
wi
n
d
o
win
g
stra
teg
ies
,
ra
n
g
in
g
fro
m
4
0
0
to
7
5
0
m
s,
o
n
m
o
d
e
l
p
e
rfo
rm
a
n
c
e
.
Ex
p
e
rime
n
tal
re
su
lt
s
d
e
m
o
n
stra
te
th
a
t
th
e
o
p
ti
m
ize
d
1
D
-
CNN
a
c
h
iev
e
s
a
p
e
a
k
tes
t
a
c
c
u
ra
c
y
o
f
9
4
.
4
%
wit
h
a
5
5
0
m
s
win
d
o
w
siz
e
,
d
e
m
o
n
stra
ti
n
g
th
e
m
o
d
e
l’s
ro
b
u
stn
e
ss
a
c
ro
ss
c
o
m
p
lex
g
e
stu
re
c
las
se
s
a
n
d
sig
n
ifi
c
a
n
t
ly
o
u
tp
e
rf
o
rm
in
g
t
h
e
b
a
se
li
n
e
p
ri
n
c
ip
a
l
c
o
m
p
o
n
e
n
t
a
n
a
ly
sis
-
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
(
P
CA
-
S
VM
)
m
e
th
o
d
wh
ich
o
n
ly
a
tt
a
in
e
d
7
3
.
0
%
a
c
c
u
ra
c
y
.
Wh
i
le
t
h
e
m
o
d
e
l
a
c
h
ie
v
e
d
p
e
rfe
c
t
c
las
sifica
ti
o
n
(1
0
0
%
)
fo
r
i
n
d
e
x
,
m
id
d
le,
a
n
d
li
tt
le
fin
g
e
r
m
o
v
e
m
e
n
ts,
a
p
e
rfo
rm
a
n
c
e
d
r
o
p
wa
s
o
b
se
rv
e
d
i
n
th
u
m
b
re
c
o
g
n
it
i
o
n
(
5
0
%
)
d
u
e
to
m
u
sc
u
lar
c
ro
ss
talk
fr
o
m
d
e
e
p
e
r
a
n
a
to
m
ica
l
lay
e
rs.
T
h
e
se
fin
d
in
g
s
in
d
ica
te
th
a
t
th
e
in
teg
ra
ti
o
n
o
f
o
p
ti
m
ize
d
win
d
o
win
g
a
n
d
1
D
-
CNN
a
rc
h
it
e
c
tu
re
s
p
r
o
v
i
d
e
s
a
h
ig
h
l
y
re
li
a
b
le
so
lu
ti
o
n
f
o
r
c
o
m
p
lex
larg
e
-
sc
a
le
g
e
stu
re
r
e
c
o
g
n
it
i
o
n
,
o
ffe
rin
g
a
r
o
b
u
st
f
o
u
n
d
a
ti
o
n
fo
r
th
e
n
e
x
t
g
e
n
e
ra
ti
o
n
o
f
m
u
lt
i
-
f
u
n
c
ti
o
n
a
l
p
ro
sth
e
ti
c
h
a
n
d
s.
K
ey
w
o
r
d
s
:
Fifth
win
d
o
win
g
o
p
tim
izatio
n
First s
E
MG
Fo
u
r
th
h
u
m
an
m
ac
h
in
e
in
ter
ac
tio
n
Seco
n
d
1
D
-
C
NN
T
h
ir
d
h
a
n
d
g
estu
r
e
r
ec
o
g
n
itio
n
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
:
Dan
iel
Su
to
p
o
Pam
u
n
g
k
as
Dep
ar
tm
en
t o
f
E
lectr
ical
E
n
g
i
n
ee
r
in
g
,
Po
litek
n
ik
Neg
er
i Bat
am
J
l A
Yan
i,
B
atam
2
9
4
3
2
,
I
n
d
o
n
esia
E
m
ail: d
an
iel@
p
o
lib
atam
.
ac
.
i
d
1.
I
NT
RO
D
UCT
I
O
N
Hu
m
an
-
co
m
p
u
ter
in
ter
ac
tio
n
(
HC
I
)
h
as
ev
o
lv
ed
f
r
o
m
in
ter
f
ac
e
-
b
ased
tex
t
an
d
g
r
ap
h
ics
to
in
ter
ac
tio
n
-
b
ased
,
m
o
r
e
n
atu
r
al,
an
d
i
m
m
er
s
iv
e
[
1
]
.
Ha
n
d
g
estu
r
e
r
ec
o
g
n
itio
n
(
HGR)
r
ep
r
esen
ts
a
p
iv
o
tal
co
m
p
o
n
en
t
in
co
n
tem
p
o
r
ar
y
i
n
ter
ac
tio
n
tech
n
o
lo
g
y
.
T
h
is
is
b
ec
au
s
e
th
e
h
an
d
h
as
m
a
n
y
d
eg
r
ee
s
o
f
f
r
ee
d
o
m
(
Do
F),
allo
win
g
p
e
o
p
le
to
c
o
m
m
u
n
icate
in
c
o
m
p
lex
way
s
with
o
u
t
wo
r
d
s
an
d
to
m
o
v
e
th
in
g
s
with
g
r
ea
t
ac
cu
r
ac
y
.
HGR
im
p
lem
en
tatio
n
in
clu
d
es
a
wid
e
r
an
g
e
o
f
ap
p
licatio
n
s
,
f
r
o
m
co
n
tr
o
l
-
h
an
d
p
r
o
s
th
etics
[
2
]
,
[
3
]
,
an
d
n
ew
n
a
v
ig
atio
n
en
v
ir
o
n
m
en
ts
to
v
ir
tu
al
r
ea
lity
(
VR
)
a
n
d
a
u
g
m
en
te
d
r
ea
lity
(
AR
)
s
y
s
tem
s
[
4
]
,
[
5
]
,
a
n
d
to
u
ch
less
s
y
s
tem
co
n
tr
o
l f
o
r
m
ed
ical
eq
u
ip
m
e
n
t
[
6
]
,
[
7
]
.
All th
ese
m
ee
t stric
t ster
ilit
y
s
tan
d
ar
d
s
.
T
h
e
q
u
ality
o
f
th
e
d
ata
co
llect
ed
d
u
r
in
g
HGR
d
ev
elo
p
m
e
n
t
d
ep
en
d
s
o
n
t
h
e
s
en
s
o
r
s
ch
o
s
en
.
A
d
ata
-
b
ased
(
v
is
io
n
-
b
ased
)
m
eth
o
d
t
h
at
u
s
es
a
ca
m
er
a
d
ep
th
s
en
s
o
r
o
r
an
in
f
r
ar
ed
s
en
s
o
r
is
o
f
ten
lim
ited
b
y
lin
e
-
of
-
s
ig
h
t
is
s
u
es,
ch
an
g
in
g
lig
h
tin
g
co
n
d
itio
n
s
,
an
d
th
e
n
ee
d
f
o
r
s
u
b
s
tan
tial
p
r
o
ce
s
s
in
g
p
o
wer
to
h
an
d
le
h
ig
h
-
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
-
8
7
0
8
E
n
h
a
n
cin
g
s
E
MG
fin
g
er g
estu
r
e
r
ec
o
g
n
itio
n
u
s
in
g
o
p
timiz
ed
…
(
Da
n
iel
S
u
to
p
o
P
a
mu
n
g
k
a
s
)
1577
r
eso
lu
tio
n
im
ag
es
[
8
]
.
W
ea
r
ab
le
s
en
s
o
r
s
lik
e
in
e
r
tial
m
ea
s
u
r
em
en
t
u
n
its
(
I
MU
s
)
[
9
]
a
n
d
E
lectr
o
m
y
o
g
r
a
p
h
y
(
E
MG
)
,
o
n
th
e
o
th
e
r
h
an
d
,
a
r
e
b
etter
at
h
a
n
d
lin
g
en
v
ir
o
n
m
en
tal
ch
an
g
es
an
d
allo
w
p
e
o
p
le
to
m
o
v
e
m
o
r
e
ea
s
ily
[
1
0
]
.
E
lectr
o
m
y
o
g
r
ap
h
y
o
r
s
u
r
f
ac
e
elec
tr
o
m
y
o
g
r
ap
h
y
(
s
E
MG
)
is
a
tech
n
o
lo
g
y
th
at
r
ec
o
r
d
s
th
e
b
io
elec
tr
ic
ac
tiv
ity
o
f
m
u
s
cle
f
ib
er
s
d
u
r
in
g
co
n
tr
ac
tio
n
a
n
d
r
ela
x
atio
n
[
1
1
]
,
[
1
2
]
.
s
E
MG
ac
q
u
is
itio
n
is
n
o
n
-
in
v
asiv
e
a
n
d
a
g
o
o
d
f
it
f
o
r
wea
r
a
b
le
tech
n
o
lo
g
ies
lik
e
m
y
o
-
ar
m
b
an
d
s
[
1
3
]
,
[
1
4
]
.
B
ec
au
s
e
th
e
s
ig
n
al
o
r
ig
in
ates
in
th
e
m
o
t
o
r
co
r
tex
an
d
tr
av
els
th
r
o
u
g
h
th
e
s
p
in
al
n
er
v
es
to
t
h
e
m
o
to
r
u
n
its
o
f
m
u
s
cles,
s
E
MG
is
th
o
u
g
h
t
to
b
e
p
a
r
ticu
lar
l
y
im
p
o
r
tan
t
in
HGR
[
1
5
]
.
T
h
is
lets
th
e
HGR
s
y
s
tem
f
ig
u
r
e
o
u
t
wh
at
a
p
er
s
o
n
wan
ts
(
an
tic
ip
atio
n
)
ev
e
n
wh
en
th
ey
ar
e
m
o
v
in
g
ar
o
u
n
d
[
1
6
]
,
[
1
7
]
.
Fin
e
m
o
to
r
m
o
v
em
en
ts
ar
e
d
el
icate
in
th
is
co
n
tex
t;
th
e
m
ain
p
r
o
b
lem
s
lie
in
th
e
co
m
p
lex
a
n
ato
m
y
o
f
th
e
lo
wer
ar
m
[
1
8
]
,
[
1
9
]
.
T
h
e
m
u
s
cles
th
at
m
o
v
e
th
e
f
in
g
er
s
an
d
h
a
n
d
s
,
s
u
ch
as
f
lex
o
r
d
ig
i
to
r
u
m
s
u
p
e
r
f
icialis
(
FDS),
f
lex
o
r
d
ig
ito
r
u
m
p
r
o
f
u
n
d
u
s
(
FDP),
a
n
d
ex
ten
s
o
r
d
ig
i
to
r
u
m
,
ar
e
cl
o
s
e
to
ea
c
h
o
t
h
er
an
d
o
f
ten
o
v
er
lap
[
2
0
]
.
W
ith
eig
h
t
s
en
s
o
r
ch
an
n
els
th
at
wr
ap
a
r
o
u
n
d
th
e
ar
m
,
th
e
M
y
o
Ar
m
b
an
d
lets
th
e
u
s
er
s
en
d
s
p
atial
s
ig
n
als
f
r
o
m
th
eir
m
u
s
cles
s
i
m
u
ltan
eo
u
s
ly
.
Ho
wev
er
,
b
ec
a
u
s
e
s
m
all
m
u
s
cle
s
ar
e
s
o
clo
s
e
to
g
eth
er
,
th
ey
ca
n
“
cr
o
s
s
talk
,
”
m
ea
n
in
g
th
at
s
en
s
o
r
s
in
o
n
e
m
u
s
cle
ca
n
d
etec
t
a
ctiv
ity
in
a
n
ea
r
b
y
m
u
s
cle.
T
h
is
is
b
ec
au
s
e
E
MG
s
ig
n
als
f
r
o
m
m
o
v
in
g
d
if
f
er
e
n
t
f
in
g
er
s
,
s
u
ch
as
th
e
litt
le
f
in
g
er
an
d
th
e
th
u
m
b
,
lo
o
k
v
er
y
s
im
ilar
[
1
8
]
,
[
2
1
]
.
T
h
is
s
p
atial
in
ter
f
er
en
ce
,
k
n
o
wn
as
m
u
s
cu
lar
cr
o
s
s
-
talk
,
is
ex
ac
er
b
ate
d
in
f
in
e
-
g
r
ain
ed
task
s
in
v
o
lv
in
g
2
0
d
if
f
er
en
t
class
es.
C
o
n
v
en
tio
n
al
f
ilter
s
an
d
h
eu
r
is
tic
-
b
ased
th
r
esh
o
ld
in
g
ar
e
o
f
ten
in
s
u
f
f
icien
t
to
d
ec
o
u
p
le
th
ese
o
v
er
lap
p
in
g
n
eu
r
al
d
r
iv
es.
C
o
n
s
eq
u
en
tly
,
th
er
e
is
an
u
r
g
en
t
n
ee
d
f
o
r
a
co
m
p
u
tatio
n
al
f
r
am
ewo
r
k
th
at
ca
n
au
to
n
o
m
o
u
s
ly
lear
n
d
ee
p
laten
t
r
ep
r
esen
tatio
n
s
f
r
o
m
r
a
w
d
ata,
e
f
f
ec
tiv
ely
“
lear
n
in
g
”
to
i
g
n
o
r
e
th
e
n
o
is
e
an
d
f
o
c
u
s
o
n
th
e
u
n
iq
u
e
tem
p
o
r
al
s
ig
n
atu
r
es o
f
ea
ch
f
i
n
g
er
'
s
m
o
to
r
u
n
it r
ec
r
u
itm
en
t.
T
h
is
s
tu
d
y
e
x
am
in
es
u
s
in
g
s
E
MG
b
ec
au
s
e
it
ca
n
p
r
o
v
id
e
in
f
o
r
m
atio
n
ab
o
u
t
“
in
ten
tio
n
.
”
A
t
2
0
Hz
o
n
th
e
My
o
Ar
m
b
an
d
,
ea
ch
ch
a
n
n
el
g
iv
es
th
e
u
s
er
a
lo
t
o
f
i
n
f
o
r
m
atio
n
.
T
o
ac
co
u
n
t
f
o
r
t
h
e
v
ar
iab
ilit
y
i
n
th
e
s
E
MG
s
ig
n
al,
wh
ich
is
n
o
n
-
s
tatio
n
ar
y
an
d
s
to
ch
asti
c,
th
e
u
s
er
n
ee
d
s
a
p
r
ec
is
e
p
r
o
ce
s
s
in
g
win
d
o
w
an
d
p
r
ec
is
e
tim
in
g
f
o
r
ca
tch
ch
a
r
ac
ter
is
tics
th
at
d
if
f
er
ac
r
o
s
s
th
e
2
0
f
in
e
-
g
r
ain
e
d
f
in
g
er
g
estu
r
es
o
f
f
in
g
er
m
o
v
em
en
t
(
5
f
in
g
er
s
)
with
o
p
en
an
d
cl
o
s
e
ac
tio
n
s
[
2
2
]
.
T
o
m
a
k
e
p
r
o
s
th
etics
th
at
ar
e
m
o
r
e
r
esp
o
n
s
iv
e
an
d
ac
cu
r
ate,
th
e
u
s
er
n
ee
d
s
to
k
n
o
w
a
lo
t a
b
o
u
t
h
o
w
E
MG
s
ig
n
als s
h
o
w
m
u
s
c
le
co
n
tr
ac
tio
n
in
th
e
lo
wer
a
r
m
[
2
3
]
.
His
to
r
ically
,
E
MG
-
b
ased
h
an
d
m
o
v
em
e
n
t
p
atter
n
r
ec
o
g
n
it
io
n
h
as
r
elied
o
n
tr
ad
itio
n
al
m
ac
h
in
e
lear
n
in
g
(
ML
)
m
eth
o
d
s
th
at
n
ec
ess
itate
ex
ten
s
iv
e
f
ea
tu
r
e
en
g
in
ee
r
in
g
[
2
4
]
.
L
i
n
ea
r
d
i
s
cr
im
in
an
t
a
n
aly
s
is
(
L
DA)
[
2
5
]
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
[
2
6
]
,
an
d
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(k
-
NN)
[
2
7
]
ar
e
all
ex
am
p
les
o
f
co
n
v
en
tio
n
al
m
ac
h
in
e
lear
n
in
g
(
ML
)
alg
o
r
ith
m
s
th
at
h
av
e
b
ec
o
m
e
s
tan
d
ar
d
in
th
e
last
ten
y
ea
r
s
b
ec
au
s
e
th
ey
ar
e
ea
s
y
to
u
s
e.
L
DA
m
eth
o
d
is
o
f
ten
ch
o
s
en
b
ec
au
s
e
it
i
s
ea
s
y
to
u
s
e
an
d
wo
r
k
s
q
u
ick
ly
.
Ho
wev
er
,
it
d
o
es
n
o
t
wo
r
k
as
well
wh
en
th
er
e
ar
e
m
an
y
m
o
v
e
m
en
t
class
es
o
r
wh
en
th
e
d
ata
is
h
i
g
h
ly
v
a
r
iab
le
[
2
8
]
.
SVMs
wo
r
k
v
er
y
well,
esp
ec
ially
o
n
s
m
all
o
r
lo
w
-
d
im
en
s
io
n
al
d
atasets
.
Ho
wev
er
,
ch
o
o
s
in
g
th
e
r
ig
h
t
k
er
n
el
an
d
tu
n
in
g
h
y
p
e
r
p
ar
am
eter
s
ca
n
b
e
d
if
f
icu
lt
a
n
d
tim
e
-
co
n
s
u
m
in
g
[
2
9
]
.
k
-
NN
an
d
Naiv
e
B
ay
es
alg
o
r
ith
m
s
ar
e
ea
s
y
to
u
s
e
b
u
t
ar
e
s
u
s
ce
p
tib
le
to
n
o
is
e
a
n
d
r
eq
u
ir
e
ca
r
ef
u
l
f
e
atu
r
e
p
r
e
p
r
o
ce
s
s
in
g
[
3
0
]
.
Alth
o
u
g
h
f
u
n
d
am
en
tal,
s
h
allo
w
ANNs
g
en
er
ally
r
eq
u
ir
e
m
an
u
al
e
x
tr
ac
tio
n
o
f
tim
e
-
d
o
m
ain
(
T
D)
o
r
f
r
e
q
u
en
c
y
-
d
o
m
ain
(
FD)
f
ea
tu
r
e
s
to
ac
h
iev
e
o
p
tim
al
ac
cu
r
ac
y
[
3
1
]
,
[
3
2
]
.
A
p
r
in
cip
al
co
n
s
tr
a
in
t
o
f
th
ese
co
n
v
en
tio
n
al
m
et
h
o
d
o
lo
g
ies
is
th
eir
r
elian
ce
o
n
m
an
u
al
f
ea
tu
r
e
e
x
tr
ac
tio
n
o
r
f
ea
tu
r
e
en
g
i
n
ee
r
in
g
[
3
3
]
.
I
n
s
u
f
f
icien
t
f
ea
tu
r
e
en
g
in
ee
r
in
g
h
am
p
er
s
class
if
icatio
n
p
er
f
o
r
m
an
ce
,
esp
ec
ially
f
o
r
i
n
tr
icate
f
in
g
er
m
o
v
em
en
ts
c
h
ar
ac
ter
ized
b
y
s
u
b
s
tan
tial
cr
o
s
s
talk
am
o
n
g
m
u
s
cle
s
ig
n
als.
Desp
i
te
th
eir
wid
esp
r
ea
d
u
s
e,
tr
ad
itio
n
al
ML
alg
o
r
ith
m
s
s
u
f
f
er
f
r
o
m
a
s
ig
n
if
ican
t
'
in
f
o
r
m
atio
n
b
o
ttlen
ec
k
'
d
u
e
to
th
eir
r
elian
ce
o
n
s
tatio
n
ar
ity
ass
u
m
p
tio
n
s
.
s
E
MG
s
ig
n
als,
p
ar
ticu
lar
ly
d
u
r
i
n
g
m
u
lti
-
f
in
g
er
m
o
v
e
m
en
ts
,
ex
h
i
b
it
n
o
n
-
lin
ea
r
an
d
d
y
n
am
ic
ch
ar
ac
ter
is
tics
th
at
s
tatic
f
ea
tu
r
e
s
ets
(
lik
e
MA
V
o
r
R
MS)
f
ail
to
ca
p
tu
r
e
en
tire
ly
.
Fu
r
th
er
m
o
r
e,
th
e
m
a
n
u
al
s
elec
tio
n
o
f
f
ea
tu
r
es
o
f
ten
in
tr
o
d
u
ce
s
h
u
m
an
b
ias,
wh
er
e
ce
r
tain
f
r
eq
u
en
cy
co
m
p
o
n
en
ts
o
r
tem
p
o
r
al
s
h
if
ts
ar
e
o
v
er
lo
o
k
ed
,
lead
in
g
to
d
eg
r
ad
ed
p
er
f
o
r
m
an
c
e
wh
en
th
e
n
u
m
b
er
o
f
g
estu
r
e
cl
ass
es in
cr
ea
s
e
s
.
A
lo
t
o
f
p
r
ep
r
o
ce
s
s
in
g
is
n
ee
d
ed
f
o
r
tr
ad
itio
n
al
ML
m
eth
o
d
s
.
Featu
r
e
en
g
in
ee
r
in
g
(
m
an
u
al
f
ea
tu
r
e
ex
tr
ac
tio
n
)
.
T
h
e
r
esear
ch
er
m
u
s
t
d
eter
m
in
e
th
e
m
o
s
t
r
e
p
r
esen
tativ
e
f
ea
tu
r
es
f
r
o
m
th
e
tim
e
d
o
m
ain
,
in
clu
d
in
g
m
ea
n
ab
s
o
lu
te
v
alu
e
(
MA
V)
,
r
o
o
t
m
ea
n
s
q
u
a
r
e
(
R
MS)
,
an
d
ze
r
o
c
r
o
s
s
in
g
(
Z
C
)
,
o
r
f
r
o
m
th
e
f
r
e
q
u
en
cy
d
o
m
ain
u
tili
zin
g
th
e
f
ast
Fo
u
r
ier
tr
an
s
f
o
r
m
(
FF
T
)
[
8
]
.
T
h
is
p
r
o
ce
s
s
is
"h
an
d
cr
af
ted
,
"
m
ea
n
in
g
th
at
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
f
ea
t
u
r
es
d
ep
en
d
s
h
ea
v
ily
o
n
th
e
s
k
ill
o
f
t
h
e
p
e
r
s
o
n
p
er
f
o
r
m
in
g
i
t
an
d
o
f
ten
m
is
s
es
co
m
p
lex
tem
p
o
r
al
r
elatio
n
s
h
ip
s
in
th
e
r
aw
E
MG
d
ata.
Als
o
,
d
im
en
s
io
n
-
r
e
d
u
ctio
n
tech
n
iq
u
es su
ch
as
p
r
in
cip
al
co
m
p
o
n
en
t
an
aly
s
is
(
PC
A)
ar
e
o
f
ten
u
s
ed
to
s
im
p
lif
y
co
m
p
lex
d
ata
[
3
4
]
.
Ho
wev
e
r
,
th
ey
r
is
k
lo
s
in
g
im
p
o
r
tan
t
in
f
o
r
m
atio
n
ab
o
u
t m
o
v
em
en
ts
,
s
u
ch
as th
o
s
e
v
er
y
s
im
ilar
to
f
in
g
er
m
o
v
em
e
n
ts
.
Dee
p
l
ea
r
n
in
g
(
DL
)
h
as
b
ec
o
m
e
a
m
o
r
e
e
f
f
ec
tiv
e
m
eth
o
d
o
v
er
th
e
p
ast
ten
y
ea
r
s
,
e
n
ab
lin
g
th
e
au
to
m
atic
lear
n
i
n
g
o
f
f
ea
tu
r
es
f
r
o
m
r
aw
d
ata
o
r
with
m
in
i
m
al
p
r
ep
r
o
ce
s
s
in
g
[
2
4
]
,
[
3
5
]
.
DL
s
p
ec
if
ically
th
e
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
etwo
r
k
(
C
NN)
ar
ch
itectu
r
e,
r
ep
r
esen
ts
a
p
ar
ad
ig
m
s
h
if
t
th
r
o
u
g
h
au
to
m
atic
f
ea
tu
r
e
lear
n
in
g
[
3
6
]
–
[
3
8
]
,
an
d
a
1
D
C
NN
is
u
tili
ze
d
in
th
e
s
tu
d
y
.
C
NNs
ar
e
v
er
y
g
o
o
d
at
p
r
o
c
ess
in
g
b
o
th
s
p
atial
(
im
ag
e)
a
n
d
tem
p
o
r
al
(
tim
e
s
er
ies)
d
ata
[
3
9
]
.
T
h
is
a
r
ch
ite
ctu
r
e
lear
n
s
h
ier
ar
c
h
ical
r
ep
r
esen
tatio
n
s
d
ir
ec
tly
f
r
o
m
th
e
r
aw
s
ig
n
al
(
en
d
-
to
-
e
n
d
)
.
L
a
y
er
c
o
n
v
o
lu
tio
n
f
u
n
cti
o
n
s
s
er
v
e
as
au
to
m
atic
f
ilter
s
th
at
d
etec
t
tem
p
o
r
al
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
1
6
,
No
.
3
,
J
u
n
e
20
2
6
:
1
5
7
6
-
1
5
8
7
1578
p
atter
n
s
with
o
u
t
m
an
u
al
in
ter
v
en
tio
n
.
T
h
e
o
n
e
-
d
im
en
s
io
n
al
(
1
D)
C
NN
ar
ch
itectu
r
e
is
p
ar
ticu
lar
ly
u
s
ef
u
l
f
o
r
E
MG
d
ata,
wh
ich
co
n
s
is
ts
o
f
m
u
ltich
an
n
el
tim
e
-
s
er
ies
s
ig
n
als.
I
n
co
n
tr
ast
to
SVM,
wh
ic
h
ex
h
ib
its
h
ig
h
b
ias
wh
en
co
n
f
r
o
n
ted
with
2
0
class
es
with
o
v
er
lap
p
in
g
f
ea
tu
r
e
s
,
C
NN
ca
n
estab
lis
h
a
m
o
r
e
in
tr
icate
d
ec
is
io
n
b
o
u
n
d
ar
y
u
s
in
g
m
illi
o
n
s
o
f
o
p
tim
ized
p
ar
am
eter
s
.
Als
o
,
tr
ad
itio
n
al
ML
m
eth
o
d
s
ar
e
o
f
ten
s
tatic
an
d
h
ar
d
to
ad
ap
t
to
th
e
n
o
r
m
al
s
ig
n
a
l
d
r
if
t
in
b
io
lo
g
ical
E
MG
d
ata.
Dee
p
l
ea
r
n
in
g
,
o
n
th
e
o
th
er
h
an
d
,
ca
n
g
en
e
r
alize
b
etter
th
an
k
s
to
tech
n
iq
u
es
lik
e
d
r
o
p
o
u
t
an
d
b
atch
n
o
r
m
aliza
t
io
n
.
I
n
th
is
r
esear
ch
,
th
e
s
wit
ch
f
r
o
m
SVM
(
wh
ich
ac
h
iev
ed
o
n
ly
ab
o
u
t
7
3
%
ac
cu
r
ac
y
)
[
2
6
]
to
a
1
D
C
NN
is
u
s
ed
to
lev
er
a
g
e
th
e
m
o
d
el's
ab
ilit
y
to
d
etec
t
m
icr
o
-
tem
p
o
r
al
p
atter
n
s
th
at
m
an
u
al
f
ea
tu
r
e
s
tatis
tics
ca
n
n
o
t.
T
h
is
is
v
er
y
im
p
o
r
tan
t
f
o
r
d
is
tin
g
u
is
h
in
g
th
e
"o
p
en
in
g
"
a
n
d
"c
lo
s
in
g
"
ac
tio
n
s
o
f
th
e
f
i
n
g
er
s
,
wh
ich
ar
e
co
n
tr
o
lled
b
y
o
v
er
lap
p
in
g
m
u
s
cles.
T
h
e
1
D
C
NN
ar
ch
itectu
r
e
is
b
ased
o
n
ef
f
icien
t
co
m
p
u
tatio
n
an
d
its
u
s
ef
u
ln
ess
f
o
r
tim
e
-
s
er
ies
d
ata.
Un
lik
e
ar
ch
itectu
r
e
,
r
ec
u
r
r
e
n
t
n
eu
r
al
n
etwo
r
k
(
R
NN)
o
r
L
STM
,
wh
ich
o
f
ten
n
ee
d
lo
n
g
er
tr
ain
in
g
an
d
a
r
e
m
o
r
e
lik
ely
to
r
u
n
in
t
o
th
e
v
an
is
h
in
g
g
r
a
d
ien
t
p
r
o
b
lem
,
1
D
C
NN
s
ca
n
p
r
o
ce
s
s
d
ata
m
o
r
e
q
u
ic
k
ly
an
d
i
n
p
ar
allel
th
r
o
u
g
h
tem
p
o
r
al
co
n
v
o
lu
tio
n
.
Au
to
m
atic
f
ea
tu
r
e
e
x
tr
ac
tio
n
:
1
D
C
NNs
u
s
e
co
n
v
o
lu
tio
n
al
lay
e
r
s
to
au
to
m
atica
lly
f
in
d
t
h
e
m
o
s
t
d
i
s
cr
im
in
ativ
e
tem
p
o
r
al
p
atter
n
s
(
k
er
n
els)
in
E
MG
s
ig
n
als.
T
h
is
r
ed
u
ce
s
th
e
n
ee
d
f
o
r
m
an
u
al
f
ea
tu
r
e
en
g
i
n
ee
r
in
g
,
wh
ich
ca
n
b
e
er
r
o
r
-
p
r
o
n
e
[
4
0
]
.
Ma
n
ag
in
g
tem
p
o
r
al
d
e
p
en
d
en
cies
:
1
D
k
er
n
els
n
atu
r
ally
ca
p
tu
r
e
tim
e
-
b
ased
d
ep
en
d
e
n
cies
an
d
s
h
o
r
t
-
ter
m
ch
an
g
es
ass
o
ciate
d
with
m
u
s
cle
m
o
v
em
en
t
[
4
1
]
.
B
etter
p
er
f
o
r
m
a
n
ce
:
R
ec
en
t
r
e
s
ea
r
ch
s
h
o
ws
th
at
1
D
C
NNs
ar
e
m
o
r
e
ac
cu
r
ate
a
n
d
s
tab
le
t
h
an
tr
ad
itio
n
al
ML
m
eth
o
d
s
,
esp
ec
ially
wh
en
th
e
s
u
b
ject
o
r
en
v
ir
o
n
m
en
t
c
h
an
g
es
[
4
2
]
.
T
h
is
m
ak
es
it
a
g
o
o
d
ch
o
ice
f
o
r
a
n
ap
p
licatio
n
-
in
tr
o
d
u
ctio
n
g
estu
r
e
th
at
r
eq
u
ir
es
lo
w
laten
c
y
.
1
D
C
NN
ef
f
ec
tiv
ely
ex
tr
ac
ts
s
p
atial
in
ter
s
en
s
o
r
y
(
8
My
o
ch
an
n
els)
an
d
tem
p
o
r
al
f
ea
tu
r
es,
m
ak
in
g
it
well
-
s
u
ited
f
o
r
p
r
o
ce
s
s
in
g
co
m
p
lex
m
o
tio
n
d
atasets
o
f
f
in
g
er
s
.
T
h
e
s
tu
d
y
'
s
n
o
v
elty
p
r
im
ar
il
y
r
esid
es
in
th
r
ee
ess
en
tial
asp
ec
ts
:
f
ir
s
tly
,
th
e
class
if
ica
tio
n
o
f
2
0
d
is
tin
ct
f
in
e
-
g
r
ai
n
ed
g
estu
r
es;
in
co
n
tr
ast
to
m
o
s
t
HGR
s
tu
d
ies
th
at
u
tili
ze
o
n
ly
5
-
7
g
estu
r
es,
th
is
s
tu
d
y
em
p
lo
y
s
h
an
d
r
o
u
g
h
n
ess
.
I
n
v
esti
g
ate
th
e
clas
s
if
icatio
n
p
r
ec
is
io
n
f
o
r
ea
ch
f
in
g
er
in
d
iv
i
d
u
a
lly
(
o
p
en
an
d
clo
s
e
ac
tio
n
s
)
.
Seco
n
d
,
th
e
s
tu
d
y
f
o
u
n
d
th
at
a
5
5
0
m
s
win
d
o
w
is
b
est
f
o
r
b
alan
ci
n
g
r
eso
lu
ti
o
n
in
f
o
r
m
atio
n
an
d
m
o
d
el
s
tab
ilit
y
.
T
h
is
win
d
o
w
is
b
etter
th
an
th
e
s
tan
d
a
r
d
6
0
0
m
s
win
d
o
w.
T
h
ir
d
,
in
co
r
p
o
r
atio
n
o
f
a
lear
n
in
g
r
ate
s
ch
ed
u
l
er
in
to
E
MG
Da
ta:
E
x
ec
u
tio
n
o
f
a
r
ate
-
ad
a
p
t
iv
e
lear
n
in
g
s
ch
e
d
u
le
d
em
o
n
s
tr
ated
s
u
b
s
tan
tial
im
p
r
o
v
em
e
n
t in
m
o
d
el
co
n
v
er
g
en
ce
s
tab
ilit
y
o
n
v
ar
iab
le
b
i
o
-
s
ig
n
al
d
ata,
ac
h
iev
i
n
g
a
test
a
cc
u
r
ac
y
o
f
9
4
.
4
%.
T
h
e
m
a
in
g
o
al
o
f
th
is
p
r
o
je
ct
i
s
t
o
d
ev
el
o
p
a
n
d
d
e
p
l
o
y
a
1
D
C
NN
a
r
ch
ite
ct
u
r
e
t
h
at
r
el
ia
b
l
y
r
e
co
g
n
iz
es
p
at
te
r
n
s
o
f
f
i
n
g
er
o
p
e
n
i
n
g
a
n
d
cl
o
s
i
n
g
m
o
t
io
n
b
y
e
x
t
r
a
cti
n
g
f
ea
tu
r
es
f
r
o
m
r
aw
s
E
MG
d
ata
.
T
h
e
g
o
a
l
o
f
t
h
is
r
es
ea
r
c
h
is
t
o
b
u
i
ld
a
m
o
d
el
t
h
at
a
u
t
o
m
ati
ca
l
ly
r
e
co
g
n
i
ze
s
2
0
d
i
f
f
er
en
t
f
i
n
g
e
r
m
o
v
e
m
e
n
ts
u
s
in
g
a
n
e
n
d
-
to
-
e
n
d
lea
r
n
i
n
g
a
p
p
r
o
ac
h
.
T
h
is
w
o
u
l
d
r
ed
u
ce
t
h
e
n
ee
d
f
o
r
m
a
n
u
al
f
e
atu
r
e
ex
tr
ac
t
io
n
,
a
co
m
m
o
n
p
r
ac
ti
ce
i
n
tr
ad
iti
o
n
al
m
et
h
o
d
s
.
T
h
is
s
tu
d
y
als
o
s
o
u
g
h
t
t
o
i
d
e
n
ti
f
y
t
h
e
o
p
ti
m
al
i
n
p
u
t
p
a
r
a
m
et
e
r
s
b
y
c
o
m
p
ar
in
g
t
h
e
c
lass
i
f
i
ca
t
io
n
p
e
r
f
o
r
m
a
n
c
e
o
f
w
in
d
o
w
s
iz
es
o
f
6
0
0
m
s
a
n
d
5
5
0
m
s
.
T
h
is
w
as
d
o
n
e
to
m
a
k
e
s
u
r
e
t
h
e
c
o
n
t
r
o
l
w
o
r
k
e
d
p
e
r
f
ec
t
ly
.
T
h
is
r
esear
ch
aim
s
to
e
v
alu
ate
s
y
s
tem
s
tab
ilit
y
b
y
im
p
l
em
en
tin
g
a
lear
n
in
g
r
ate
s
ch
ed
u
ler
an
d
an
aly
zin
g
its
im
p
ac
t
o
n
th
e
v
alid
atio
n
ac
cu
r
ac
y
cu
r
v
e
d
u
r
in
g
m
o
d
el
d
ev
elo
p
m
en
t.
T
h
is
is
cr
u
cial
f
o
r
m
ed
ical
r
o
b
o
ts
b
ec
au
s
e
th
ey
n
ee
d
to
en
s
u
r
e
th
at
co
n
tr
o
l
r
em
ain
s
co
n
s
is
ten
t
an
d
wo
r
k
s
.
T
h
is
s
tu
d
y
aim
s
to
p
r
o
v
id
e
em
p
ir
ical
ev
id
e
n
ce
o
f
th
e
s
u
p
er
io
r
ity
o
f
d
e
ep
lea
r
n
in
g
tech
n
iq
u
es
o
v
e
r
tr
a
d
itio
n
al
m
ac
h
in
e
lear
n
i
n
g
alg
o
r
ith
m
s
,
s
u
ch
as
L
DA
a
n
d
SVM.
T
h
is
r
esear
ch
s
ee
k
s
t
o
ad
v
an
ce
th
e
d
ev
elo
p
m
en
t
o
f
m
o
r
e
r
esp
o
n
s
iv
e
a
n
d
in
tu
itiv
e
p
r
o
s
th
etic
h
a
n
d
tech
n
o
lo
g
y
b
y
d
em
o
n
s
tr
atin
g
s
ig
n
if
ican
t
ac
cu
r
ac
y
im
p
r
o
v
em
e
n
ts
o
v
er
tr
a
d
itio
n
al
PC
A
-
SV
M
m
eth
o
d
o
lo
g
ies.
T
h
is
ar
ticle
will
b
e
s
tr
u
ctu
r
ed
as
f
o
llo
ws:
s
ec
tio
n
1
is
i
n
tr
o
d
u
ctio
n
.
I
t
d
escr
ib
es
th
e
b
ac
k
g
r
o
u
n
d
o
f
th
e
s
tu
d
y
,
th
e
u
r
g
e
n
cy
o
f
u
s
in
g
s
E
MG
s
ig
n
als
f
o
r
p
r
o
s
th
etic
co
n
tr
o
l,
th
e
lim
itatio
n
s
o
f
t
r
ad
itio
n
al
m
eth
o
d
s
,
an
d
th
e
n
o
v
elty
a
n
d
p
u
r
p
o
s
e
o
f
th
e
1
D
C
NN
ar
ch
itectu
r
e.
T
h
is
is
f
o
llo
wed
b
y
s
ec
tio
n
2
:
m
ater
ials
an
d
m
eth
o
d
s
,
wh
ich
ex
p
lain
in
d
e
tail
th
e
d
ata
-
ac
q
u
is
itio
n
p
r
o
ce
d
u
r
e
u
s
in
g
th
e
My
o
Ar
m
b
an
d
,
th
e
ex
p
er
im
en
tal
p
r
o
to
co
l
f
o
r
2
0
f
in
g
er
m
o
v
em
en
ts
,
s
ig
n
al
p
r
ep
r
o
ce
s
s
in
g
tech
n
iq
u
es,
in
clu
d
in
g
win
d
o
win
g
an
d
s
eg
m
en
tatio
n
,
an
d
th
e
d
esig
n
o
f
t
h
e
p
r
o
p
o
s
ed
1
D
C
NN
s
y
s
tem
ar
ch
itectu
r
e.
T
h
e
f
ir
s
t
p
ar
t
is
f
o
llo
wed
b
y
a
c
o
m
p
r
e
h
en
s
iv
e
an
aly
s
is
o
f
th
e
e
x
p
er
im
e
n
tal
r
esu
lts
in
th
e
r
esu
lts
an
d
d
is
cu
s
s
io
n
s
ec
tio
n
an
d
f
in
ally
th
e
c
o
n
clu
s
io
n
an
d
f
u
tu
r
e
wo
r
k
s
s
ec
tio
n
,
wh
er
e
th
is
s
ec
tio
n
s
u
m
m
ar
izes
th
e
m
ain
f
in
d
in
g
s
o
f
th
e
s
tu
d
y
r
eg
ar
d
in
g
t
h
e
ef
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
an
d
p
r
o
v
i
d
es
s
tr
ateg
ic
d
ir
ec
tio
n
s
f
o
r
th
e
d
ev
elo
p
m
en
t
o
f
m
o
r
e
p
r
ec
is
e
p
r
o
s
th
etic
c
o
n
tr
o
l
s
y
s
tem
s
in
th
e
f
u
tu
r
e.
2.
M
E
T
H
O
D
T
h
is
ch
ap
ter
d
elin
ea
tes
s
y
s
te
m
atic
r
esear
ch
,
en
co
m
p
ass
in
g
th
e
d
ata
ac
q
u
is
itio
n
p
r
o
ce
d
u
r
e,
s
ig
n
al
p
r
ep
r
o
ce
s
s
in
g
,
s
ig
n
al
d
etec
tio
n
,
th
e
im
p
lem
en
tatio
n
o
f
a
1
D
-
C
NN
ar
ch
itectu
r
e,
an
d
th
e
o
p
tim
izatio
n
s
tr
ateg
y
em
p
lo
y
ed
t
o
en
h
an
ce
2
0
-
class
class
if
icatio
n
ac
r
o
s
s
2
0
g
estu
r
e
class
es.
T
h
e
d
ata
co
llectio
n
m
eth
o
d
s
in
s
tu
d
ies
ar
e
g
r
o
u
n
d
e
d
in
d
ev
elo
p
m
en
t
p
r
o
to
co
ls
p
r
ev
io
u
s
ly
v
alid
ated
b
y
Dela
et
a
l.
[
2
6
]
.
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
-
8
7
0
8
E
n
h
a
n
cin
g
s
E
MG
fin
g
er g
estu
r
e
r
ec
o
g
n
itio
n
u
s
in
g
o
p
timiz
ed
…
(
Da
n
iel
S
u
to
p
o
P
a
mu
n
g
k
a
s
)
1579
T
h
e
My
o
Ar
m
b
an
d
(
T
h
alm
ic
L
ab
s
I
n
c.
)
is
u
s
ed
to
ac
q
u
ir
e
t
h
e
s
E
MG
s
ig
n
al.
T
h
is
d
ev
ice
is
ch
o
s
en
b
ec
au
s
e
it
is
wir
ele
s
s
,
n
o
n
-
in
v
asiv
e,
an
d
ea
s
y
to
u
s
e
f
o
r
w
ea
r
ab
le
ap
p
licatio
n
s
[
4
3
]
.
T
h
e
My
o
Ar
m
b
an
d
h
as
eig
h
t
s
E
MG
s
en
s
o
r
ch
a
n
n
els
with
d
r
y
elec
tr
o
d
es
wr
a
p
p
ed
ar
o
u
n
d
th
e
s
u
b
ject'
s
lo
wer
ar
m
.
T
h
e
s
ig
n
als
we
co
llected
h
av
e
8
-
b
it
r
eso
lu
tio
n
an
d
a
s
am
p
lin
g
r
ate
o
f
2
0
0
Hz,
an
d
th
ey
ar
e
s
en
t
th
r
o
u
g
h
a
B
lu
eto
o
th
lo
w
en
er
g
y
(
B
L
E
)
m
o
d
u
le.
Data
a
cq
u
is
itio
n
f
o
r
t
h
is
p
ilo
t
s
tu
d
y
was
p
er
f
o
r
m
ed
b
y
a
s
in
g
le
h
e
alth
y
m
ale
s
u
b
ject
(
ag
ed
2
5
)
t
o
estab
lis
h
a
b
a
s
elin
e
p
er
f
o
r
m
an
ce
f
o
r
th
e
p
r
o
p
o
s
ed
1
D
-
C
NN
ar
ch
itect
u
r
e
ac
r
o
s
s
a
h
ig
h
-
d
im
en
s
io
n
al
g
estu
r
e
s
p
ac
e.
W
h
ile
f
o
cu
s
in
g
o
n
a
s
in
g
le
p
ar
ticip
an
t,
th
is
s
tu
d
y
p
r
io
r
itizes
th
e
tech
n
ical
f
ea
s
ib
ilit
y
o
f
class
if
y
in
g
2
0
d
i
s
tin
ct
f
in
e
-
g
r
ai
n
ed
f
in
g
e
r
m
o
v
em
en
ts
,
wh
ich
is
s
ig
n
if
ican
tly
h
ig
h
er
th
an
th
e
5
-
7
g
estu
r
es
ty
p
ically
f
o
u
n
d
in
e
x
is
tin
g
liter
atu
r
e
[
4
4
]
.
T
h
is
ap
p
r
o
ac
h
p
r
o
v
id
es
a
r
ig
o
r
o
u
s
in
iti
al
v
alid
atio
n
o
f
th
e
m
o
d
el'
s
f
ea
tu
r
e
ex
tr
ac
tio
n
ca
p
ab
ilit
ies b
ef
o
r
e
f
u
tu
r
e
ex
p
an
s
io
n
to
a
b
r
o
a
d
er
d
e
m
o
g
r
a
p
h
ic.
T
h
e
s
u
b
ject
is
s
ea
ted
,
with
th
e
r
ig
h
t
ar
m
r
esti
n
g
o
n
th
e
to
p
o
f
th
e
tab
le
to
m
in
im
ize
th
e
b
u
r
d
e
n
o
n
s
tatic
m
u
s
cles
an
d
ar
tef
ac
t
m
o
v
em
en
t.
E
ig
h
t
My
o
Ar
m
b
a
n
d
elec
tr
o
d
es
ar
e
p
o
s
itio
n
ed
o
n
t
h
e
f
lex
o
r
d
ig
ito
r
u
m
s
u
p
er
f
icialis
m
u
s
cle
an
d
th
e
e
x
ten
s
o
r
m
u
s
cles a
r
o
u
n
d
th
e
lo
wer
ar
m
.
T
h
er
e
a
r
e
2
0
d
if
f
er
e
n
t
way
s
f
o
r
t
h
e
f
i
n
g
er
s
t
o
m
o
v
e,
ea
c
h
with
two
m
ain
ac
tio
n
s
:
o
p
en
in
g
a
n
d
clo
s
in
g
.
T
h
e
d
ata
co
llectio
n
p
r
o
to
co
l
r
eq
u
ir
e
d
th
e
s
u
b
ject
t
o
p
er
f
o
r
m
ea
ch
f
in
g
er
m
o
v
em
e
n
t
f
o
r
a
d
u
r
atio
n
o
f
5
s
ec
o
n
d
s
.
T
o
k
ee
p
m
u
s
cles
f
r
o
m
g
ettin
g
tire
d
,
th
e
r
ela
x
atio
n
p
h
ase
(
r
est)
last
s
f
iv
e
s
ec
o
n
d
s
b
etwe
en
m
o
v
em
en
ts
.
T
h
e
m
o
v
em
en
t
is
r
ep
ea
ted
u
p
to
1
0
tim
es
p
er
cy
cle
to
en
s
u
r
e
s
u
f
f
icien
t
d
ata
f
o
r
tr
ain
in
g
th
e
n
etwo
r
k
(
d
ee
p
n
eu
r
al
n
etwo
r
k
)
.
T
h
e
My
o
Ar
m
b
a
n
d
'
s
r
aw
d
ata
n
ee
d
s
to
b
e
clea
n
ed
an
d
s
p
lit
u
p
b
ef
o
r
e
it
ca
n
b
e
u
s
ed
in
th
e
1
D
C
NN
m
o
d
el.
T
o
ca
p
tu
r
e
th
e
tem
p
o
r
al
d
y
n
am
ics
o
f
E
MG
s
ig
n
als,
a
s
lid
in
g
win
d
o
w
tech
n
iq
u
e
[
4
5
]
was
u
s
ed
.
B
ased
o
n
em
p
ir
ical
r
esu
lts
,
th
e
win
d
o
w
s
ize
is
s
et
to
5
5
0
m
s
.
At
2
0
0
Hz,
with
a
s
am
p
le
win
d
o
w
o
f
1
1
0
p
o
in
ts
p
er
ch
an
n
el,
r
esu
ltin
g
in
a
n
in
p
u
t
m
atr
ix
o
f
s
ize
1
1
0
×
8
.
A
s
lid
in
g
win
d
o
w
tech
n
i
q
u
e
with
a
5
0
%
o
v
er
lap
was
im
p
lem
en
ted
to
s
ig
n
if
ican
tly
ex
p
an
d
th
e
d
ataset
s
ize,
ad
d
r
ess
in
g
th
e
d
ata
-
in
ten
s
iv
e
r
e
q
u
ir
em
en
ts
o
f
d
ee
p
lear
n
in
g
.
Alth
o
u
g
h
th
e
in
itial
r
ec
o
r
d
i
n
g
co
n
s
is
ted
o
f
1
0
c
o
n
tr
o
lled
r
ep
etitio
n
s
p
er
cy
cl
e
f
o
r
ea
c
h
o
f
th
e
2
0
class
es,
th
e
win
d
o
win
g
an
d
o
v
e
r
lap
p
r
o
ce
s
s
au
g
m
en
ted
th
e
d
ataset
in
to
th
o
u
s
an
d
s
o
f
i
n
d
iv
id
u
al
s
eg
m
en
ts
.
T
h
is
s
tr
ateg
y
en
s
u
r
es
th
at
th
e
1
D
-
C
NN
is
ex
p
o
s
ed
to
s
u
f
f
icien
t
tem
p
o
r
al
v
ar
iatio
n
s
,
th
e
r
eb
y
en
h
an
cin
g
th
e
m
o
d
el'
s
g
en
er
aliza
tio
n
an
d
p
r
e
v
en
tin
g
o
v
er
f
itti
n
g
d
esp
ite
th
e
f
o
cu
s
ed
n
u
m
b
er
o
f
p
ar
ticip
an
t
s
.
Kee
p
in
m
i
n
d
t
h
at
th
e
s
E
MG
s
ig
n
al
is
h
ig
h
l
y
d
e
p
en
d
e
n
t
o
n
t
h
e
s
u
b
ject'
s
s
k
in
an
d
co
n
tr
ac
tio
n
s
tr
en
g
th
,
an
d
n
o
r
m
aliza
tio
n
is
p
er
f
o
r
m
ed
u
s
in
g
Z
-
s
co
r
e
Stan
d
ar
d
izatio
n
.
T
h
is
m
eth
o
d
tr
a
n
s
f
o
r
m
s
th
e
d
ata
s
o
th
at
it
h
as
a
m
ea
n
µ
o
f
0
an
d
a
s
tan
d
ar
d
d
e
v
iatio
n
σ
o
f
1
.
T
h
is
h
as
b
ee
n
s
h
o
wn
to
s
p
ee
d
u
p
co
n
v
er
g
e
n
ce
in
alg
o
r
ith
m
s
th
at
u
s
e
g
r
ad
ien
t d
escen
t
[
4
6
]
,
[
4
7
]
.
=
−
µ
(
1
)
w
h
er
e
is
s
co
r
e,
µ
is
m
ea
n
an
d
is
s
tan
d
ar
d
d
ev
iatio
n
.
T
h
is
s
tu
d
y
s
u
g
g
ests
s
u
b
s
titu
tin
g
tr
ad
itio
n
al
PC
A
an
d
SVM
with
1
D
C
NN
s
,
wh
ich
ex
ce
l
at
au
to
m
atica
lly
ex
tr
ac
tin
g
f
ea
tu
r
es
f
r
o
m
tim
e
-
s
er
ies
d
ata
with
o
u
t
r
eq
u
ir
in
g
i
n
tr
icate
h
an
d
cr
af
ted
f
ea
tu
r
es
[
4
7
]
.
T
h
e
1
D
C
NN
m
o
d
el
ar
ch
itectu
r
e
d
ev
elo
p
ed
in
th
is
s
tu
d
y
co
m
p
r
is
es
a
s
er
ies
o
f
in
teg
r
ated
lay
er
s
f
o
r
au
to
m
atic
tem
p
o
r
al
f
ea
t
u
r
e
ex
t
r
ac
tio
n
.
T
h
e
n
etwo
r
k
s
tr
u
ctu
r
e
b
eg
i
n
s
with
a
1
D
c
o
n
v
o
l
u
tio
n
al
lay
e
r
t
h
at
u
s
es
a
tem
p
o
r
al
co
n
v
o
l
u
tio
n
f
ilter
to
ex
tr
ac
t
lo
ca
l
f
ea
tu
r
e
p
atter
n
s
f
r
o
m
t
h
e
r
aw
s
E
MG
s
ig
n
al,
with
t
h
e
ap
p
licatio
n
o
f
a
r
ec
tifie
d
lin
ea
r
u
n
it
(
R
eL
U)
a
ctiv
atio
n
f
u
n
ctio
n
to
in
tr
o
d
u
c
e
n
o
n
-
lin
ea
r
ity
in
to
th
e
m
o
d
el
[
4
8
]
.
T
o
m
ain
tain
th
e
s
tab
ilit
y
o
f
th
e
tr
ain
in
g
p
r
o
ce
s
s
o
n
f
lu
ctu
atin
g
B
io
s
ig
n
al
d
ata,
a
b
atch
n
o
r
m
aliza
tio
n
lay
er
is
ad
d
ed
im
m
ed
iately
af
ter
th
e
co
n
v
o
l
u
tio
n
p
r
o
ce
s
s
to
n
o
r
m
alize
th
e
ac
tiv
atio
n
s
[
4
9
]
.
Nex
t,
a
p
o
o
lin
g
lay
er
with
th
e
Ma
x
Po
o
lin
g
1
D
m
ec
h
a
n
is
m
is
u
s
ed
t
o
r
ed
u
ce
th
e
s
p
atial
d
i
m
en
s
io
n
ality
wh
ile
p
r
eser
v
in
g
th
e
m
o
s
t
d
o
m
in
an
t
f
ea
tu
r
es,
m
ak
in
g
th
e
m
o
d
el
m
o
r
e
r
o
b
u
s
t to
s
m
all
s
ig
n
al
s
h
if
ts
[
5
0
]
.
T
o
m
itig
ate
th
e
r
is
k
o
f
o
v
er
f
i
ttin
g
,
a
d
r
o
p
o
u
t
la
y
er
is
ap
p
lied
with
a
r
ate
o
f
0
.
5
,
w
h
ich
wo
r
k
s
b
y
r
an
d
o
m
l
y
tu
r
n
i
n
g
o
f
f
n
e
u
r
o
n
s
d
u
r
in
g
th
e
tr
ai
n
in
g
p
h
ase
[
5
1
]
.
T
h
e
ex
tr
ac
ted
h
ig
h
-
le
v
el
f
ea
tu
r
es
ar
e
th
e
n
in
teg
r
ated
u
s
in
g
a
f
u
lly
co
n
n
e
cted
lay
er
with
3
2
n
eu
r
o
n
s
.
I
n
th
e
f
in
al
s
tag
e,
a
So
f
tMa
x
o
u
t
p
u
t
lay
er
is
u
s
ed
to
g
en
er
ate
a
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
th
at
class
if
ies
th
e
d
ata
in
t
o
2
0
d
is
cr
ete
f
in
g
er
m
o
v
em
en
t
class
e
s
[
5
2
]
.
T
h
is
o
v
er
all
co
n
f
i
g
u
r
atio
n
en
ab
les
th
e
m
o
d
el
to
r
ec
o
g
n
ize
co
m
p
lex
m
o
tio
n
p
atter
n
s
with
g
r
ea
ter
p
r
ec
is
io
n
th
an
tr
ad
itio
n
al
m
eth
o
d
s
.
T
h
e
m
o
d
el
tr
ain
in
g
p
r
o
ce
s
s
was
ca
r
r
ied
o
u
t
u
s
in
g
th
e
Ad
am
o
p
tim
izatio
n
alg
o
r
ith
m
[
5
3
]
.
An
in
itial
lear
n
in
g
r
ate
o
f
0
.
0
0
1
is
s
et
f
o
r
th
is
alg
o
r
ith
m
.
T
o
im
p
r
o
v
e
m
o
d
el
p
e
r
f
o
r
m
an
ce
a
n
d
s
tab
i
lity
,
two
ad
d
itio
n
al
s
tr
ateg
ies
wer
e
im
p
lem
en
ted
:
th
e
lear
n
in
g
r
ate
s
ch
ed
u
ler
an
d
ea
r
ly
s
to
p
p
in
g
.
T
h
e
f
ir
s
t
s
tr
ateg
y
is
d
y
n
a
m
ically
ad
ju
s
tin
g
th
e
lear
n
in
g
r
ate
b
as
ed
o
n
m
o
n
ito
r
e
d
v
ali
d
atio
n
m
etr
ics.
T
h
is
m
eth
o
d
allo
win
g
t
h
e
m
o
d
el
to
r
ea
c
h
th
e
g
lo
b
al
m
in
i
m
u
m
with
a
h
ig
h
er
le
v
el
o
f
p
r
ec
is
io
n
a
n
d
a
cc
u
r
ac
y
.
T
h
er
e
is
a
s
ec
o
n
d
s
tr
ateg
y
,
wh
ich
is
to
h
av
e
ea
r
ly
s
to
p
p
in
g
.
T
h
is
s
tr
ateg
y
au
to
m
atica
lly
s
to
p
s
th
e
tr
ain
in
g
p
r
o
ce
s
s
if
th
e
v
alid
atio
n
lo
s
s
d
o
es
n
o
t
d
ec
r
ea
s
e
f
o
r
7
co
n
s
ec
u
tiv
e
ep
o
ch
s
o
r
a
p
atien
ce
p
e
r
io
d
.
T
h
e
im
p
lem
en
tatio
n
o
f
th
is
m
ec
h
an
is
m
is
cr
u
cial
t
o
p
r
ev
en
t
o
v
er
f
itti
n
g
an
d
en
s
u
r
e
th
at
t
h
e
s
av
ed
m
o
d
el
p
ar
am
eter
s
r
ep
r
esen
t
th
e
r
e
s
u
lts
with
th
e
b
est
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
1
6
,
No
.
3
,
J
u
n
e
20
2
6
:
1
5
7
6
-
1
5
8
7
1580
g
en
er
aliza
tio
n
ab
ilit
y
o
n
th
e
t
est
d
ata.
T
h
e
p
r
o
p
o
s
ed
1
D
-
C
NN
ar
ch
itectu
r
e
is
d
esig
n
ed
to
p
er
f
o
r
m
en
d
-
to
-
e
n
d
f
ea
tu
r
e
ex
tr
ac
tio
n
f
r
o
m
r
aw
s
E
MG
s
ig
n
als.
As d
etailed
in
T
ab
le
1
.
T
ab
le
1
.
Ar
c
h
itectu
r
e
an
d
lay
e
r
p
ar
am
eter
s
o
f
1
D
-
C
NN
a
lg
o
r
ith
m
La
y
e
r
(
T
y
p
e
)
O
u
t
p
u
t
S
h
a
p
e
K
e
r
n
e
l
S
i
z
e
P
a
r
a
me
t
e
r
s/
N
o
t
e
I
n
p
u
t
La
y
e
r
(
5
5
0
,
8
)
-
8
-
c
h
a
n
n
e
l
sE
M
G
i
n
p
u
t
C
o
n
v
1
D
(
5
4
8
,
6
4
)
3
F
e
a
t
u
r
e
ma
p
p
i
n
g
B
a
t
c
h
n
o
r
ma
l
i
z
a
t
i
o
n
(
5
4
8
,
6
4
)
-
R
e
g
u
l
a
r
i
z
a
t
i
o
n
R
e
LU
(
A
c
t
i
v
a
t
i
o
n
)
(
5
4
8
,
6
4
)
-
N
o
n
-
l
i
n
e
a
r
i
t
y
M
a
x
P
o
o
l
i
n
g
1
D
(
2
7
4
,
6
4
)
2
D
o
w
n
sa
mp
l
i
n
g
C
o
n
v
1
D
(
2
7
2
,
1
2
8
)
3
H
i
g
h
-
l
e
v
e
l
f
e
a
t
u
r
e
s
R
e
LU
(
A
c
t
i
v
a
t
i
o
n
)
(
2
7
2
,
1
2
8
)
-
N
o
n
-
l
i
n
e
a
r
i
t
y
G
l
o
b
a
l
A
v
g
P
o
o
l
i
n
g
1
D
-
1
2
8
-
P
a
r
a
me
t
e
r
r
e
d
u
c
t
i
o
n
D
r
o
p
o
u
t
-
1
2
8
-
R
a
t
e
:
0
.
5
D
e
n
se
(
O
u
t
p
u
t
)
-
20
-
S
o
f
t
ma
x
(
2
0
c
l
a
ss
e
s)
T
h
e
s
p
ec
if
ic
co
n
f
ig
u
r
atio
n
o
f
th
e
1
D
-
C
NN
ar
ch
itectu
r
e
was
ch
o
s
en
to
b
alan
ce
co
m
p
u
tatio
n
al
ef
f
icien
cy
an
d
class
if
icatio
n
ac
cu
r
ac
y
.
T
h
e
B
atch
No
r
m
al
izatio
n
lay
er
s
wer
e
in
te
g
r
ate
d
to
s
tab
ilize
th
e
tr
ain
in
g
p
r
o
ce
s
s
ag
ain
s
t th
e
s
t
o
ch
asti
c
n
atu
r
e
o
f
s
E
MG
s
ig
n
als,
wh
ile
th
e
Dr
o
p
o
u
t r
ate
o
f
0
.
5
was stra
teg
ically
ap
p
lied
to
m
itig
ate
o
v
er
f
itti
n
g
d
u
r
i
n
g
h
ig
h
-
d
im
en
s
io
n
al
class
if
icatio
n
.
Fu
r
th
er
m
o
r
e,
th
e
in
clu
s
io
n
o
f
a
g
lo
b
a
l
av
er
ag
e
p
o
o
lin
g
(
GAP)
la
y
e
r
s
er
v
es
as
a
s
tr
u
ctu
r
al
r
e
g
u
lar
ize,
r
ed
u
ci
n
g
t
h
e
to
tal
p
a
r
am
eter
co
u
n
t
a
n
d
p
r
ev
en
tin
g
th
e
m
o
d
el
f
r
o
m
b
ec
o
m
in
g
o
v
e
r
ly
co
m
p
le
x
,
wh
ich
is
es
s
en
tial
f
o
r
p
o
ten
tial
r
ea
l
-
tim
e
p
r
o
s
th
etic
ap
p
licatio
n
s
.
T
h
e
d
ata
is
d
iv
id
ed
in
to
th
r
ee
p
ar
ts
:
7
0
%
f
o
r
tr
ain
in
g
,
1
5
%
f
o
r
v
alid
atio
n
,
an
d
1
5
%
f
o
r
test
in
g
.
T
h
e
ev
alu
atio
n
u
s
ed
th
e
m
etr
ics
A
cc
u
r
ac
y
,
Pre
cisi
o
n
,
R
ec
all,
an
d
F1
-
Sco
r
e
o
n
th
e
test
d
ata.
T
h
e
Usag
e
C
o
n
f
u
s
io
n
Ma
tr
ix
was
u
s
ed
to
b
r
ea
k
d
o
wn
m
o
d
el
p
er
f
o
r
m
an
ce
o
n
d
if
f
icu
lt
class
es,
wh
ich
wer
e
s
ep
ar
ated
b
y
m
u
s
cu
la
r
cr
o
s
s
talk
p
h
en
o
m
e
n
a
[
5
4
]
.
Fig
u
r
e
1
s
h
o
ws th
e
f
lo
wch
ar
t
o
f
t
h
e
m
eth
o
d
s
th
at
ar
e
u
s
ed
in
th
ese
ex
p
er
im
en
ts
.
Fig
u
r
e
1
.
Flo
w
ch
a
r
t o
f
t
h
e
m
e
th
o
d
s
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
-
8
7
0
8
E
n
h
a
n
cin
g
s
E
MG
fin
g
er g
estu
r
e
r
ec
o
g
n
itio
n
u
s
in
g
o
p
timiz
ed
…
(
Da
n
iel
S
u
to
p
o
P
a
mu
n
g
k
a
s
)
1581
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
ch
ap
ter
s
h
o
ws
th
e
r
esu
lts
o
f
an
ex
p
er
im
en
t
th
at
u
s
e
d
th
e
cu
r
r
en
t
1
D
C
NN
m
o
d
el
to
s
o
r
t
2
0
f
in
g
er
m
o
v
em
e
n
ts
.
T
h
e
an
aly
s
is
co
m
p
ar
ed
th
e
r
esu
lts
o
f
th
e
b
asic
m
eth
o
d
(
SVM)
an
d
ev
alu
ated
th
e
m
o
d
el'
s
s
tab
ilit
y
u
s
in
g
a
co
m
p
r
eh
en
s
iv
e
p
er
f
o
r
m
a
n
ce
m
etr
ic.
B
ef
o
r
e
th
e
class
if
icatio
n
p
r
o
ce
s
s
u
s
in
g
th
e
1
D
-
C
NN
ar
c
h
itectu
r
e,
th
e
r
aw
s
E
MG
s
ig
n
als
wer
e
an
aly
ze
d
to
en
s
u
r
e
d
ata
q
u
ality
.
T
h
e
s
E
MG
s
ig
n
als
o
b
tain
ed
f
r
o
m
th
e
My
o
Ar
m
b
an
d
a
r
e
in
h
er
en
tly
n
o
n
-
s
tatio
n
ar
y
an
d
h
a
v
e
h
ig
h
l
y
v
ar
iab
le
am
p
litu
d
es
ac
r
o
s
s
elec
tr
o
d
es.
Sig
n
al
v
is
u
aliza
tio
n
r
ev
ea
led
h
i
g
h
s
to
ch
asti
c
f
lu
ctu
atio
n
s
in
th
e
r
aw
d
ata,
wh
ich
,
if
lef
t
u
n
ad
d
r
ess
ed
,
co
u
ld
h
in
d
er
th
e
co
n
v
e
r
g
en
ce
o
f
th
e
d
ee
p
lear
n
in
g
m
o
d
el.
T
h
e
ap
p
licatio
n
o
f
n
o
r
m
aliza
tio
n
u
s
in
g
Z
-
s
co
r
e
s
tan
d
ar
d
iz
atio
n
s
u
cc
ess
f
u
lly
tr
an
s
f
o
r
m
e
d
th
e
d
ata
d
is
tr
ib
u
tio
n
to
h
av
e
a
m
ea
n
̅
o
f
ze
r
o
an
d
a
s
tan
d
a
r
d
d
e
v
iatio
n
̅
o
f
o
n
e
.
Vis
u
ally
,
th
is
p
r
o
ce
s
s
b
alan
ce
d
th
e
d
y
n
am
ic
r
an
g
e
(
am
p
litu
d
e
)
a
cr
o
s
s
th
e
eig
h
t
s
en
s
o
r
ch
an
n
els,
allo
win
g
th
e
1
D
-
C
NN
m
o
d
el
to
f
o
cu
s
o
n
ex
tr
ac
tin
g
tem
p
o
r
al
f
ea
t
u
r
es
a
n
d
m
u
s
cle
co
n
tr
ac
tio
n
p
atter
n
s
with
o
u
t
b
ein
g
af
f
ec
ted
b
y
d
i
f
f
er
en
ce
s
in
s
en
s
o
r
p
r
ess
u
r
e
s
tr
en
g
th
o
n
th
e
s
k
in
o
r
v
ar
ia
b
ilit
y
in
t
h
e
s
u
b
ject'
s
m
u
s
cle
s
tr
en
g
th
.
T
h
is
p
r
o
v
id
es
a
cr
u
cial
f
o
u
n
d
atio
n
f
o
r
th
e
co
n
v
o
lu
tio
n
al
f
ilter
to
au
to
m
atica
lly
d
etec
t
m
icr
o
-
tem
p
o
r
al
f
ea
tu
r
es
in
th
e
n
e
x
t
s
tag
e.
Fig
u
r
e
2
illu
s
tr
ates
th
e
co
m
p
ar
is
o
n
b
e
twee
n
th
e
r
aw
s
E
MG
s
ig
n
al
s
an
d
t
h
e
o
u
tp
u
t
o
f
th
e
Z
-
s
c
o
r
e
s
tan
d
a
r
d
izatio
n
p
r
o
ce
s
s
.
Fig
u
r
e
2
.
R
aw
s
E
MG
s
ig
n
als an
d
p
r
e
p
r
o
ce
s
s
ed
s
ig
n
als
As
d
em
o
n
s
tr
ated
i
n
T
a
b
le
2
an
d
v
is
u
alize
d
with
a
b
ar
c
h
ar
t
in
Fig
u
r
e
3
,
th
e
1
D
-
C
NN
m
o
d
el'
s
p
er
f
o
r
m
an
ce
is
h
ig
h
ly
s
en
s
itiv
e
to
th
e
tem
p
o
r
al
win
d
o
w
d
u
r
atio
n
.
T
h
e
h
ig
h
est
test
ac
cu
r
ac
y
o
f
0
.
9
4
4
(
9
4
.
4
%)
was
ac
h
iev
ed
with
a
win
d
o
w
s
ize
o
f
5
5
0
m
s
.
I
n
ter
esti
n
g
ly
,
p
er
f
o
r
m
a
n
ce
s
ig
n
if
ican
tly
d
eg
r
ad
ed
at
7
0
0
m
s
(
0
.
1
8
5
ac
cu
r
ac
y
)
,
s
u
g
g
esti
n
g
th
at
ex
ce
s
s
iv
ely
lo
n
g
win
d
o
ws
in
tr
o
d
u
ce
n
o
n
-
s
tatio
n
ar
y
n
o
is
e
th
at
in
ter
f
er
es
with
th
e
co
n
v
o
lu
tio
n
al
f
ilter
s
'
ab
ilit
y
to
ex
tr
ac
t
s
tab
le
m
o
t
o
r
u
n
it
ac
tiv
atio
n
p
atter
n
s
.
As
o
b
s
er
v
ed
in
Fig
u
r
e
3
,
th
e
ac
cu
r
ac
y
c
u
r
v
e
e
x
h
ib
its
a
n
o
n
-
lin
ea
r
r
elatio
n
s
h
ip
with
th
e
win
d
o
w
d
u
r
atio
n
.
T
h
e
s
h
ar
p
d
ec
lin
e
in
ac
cu
r
ac
y
at
4
5
0
m
s
a
n
d
7
0
0
m
s
in
d
ic
ates
th
at
th
e
1
D
-
C
NN
is
h
ig
h
ly
s
en
s
itiv
e
to
tem
p
o
r
al
b
o
u
n
d
ar
y
co
n
d
itio
n
s
.
A
win
d
o
w
th
at
is
to
o
s
h
o
r
t
(
b
elo
w
5
0
0
m
s
)
f
ails
to
p
r
o
v
id
e
en
o
u
g
h
tem
p
o
r
al
in
f
o
r
m
atio
n
f
o
r
th
e
k
er
n
els
to
d
etec
t
a
co
m
p
lete
f
ir
in
g
p
atter
n
,
wh
ile
a
win
d
o
w
ex
ce
ed
i
n
g
6
5
0
m
s
in
tr
o
d
u
ce
s
ex
ce
s
s
iv
e
n
o
n
-
s
tatio
n
ar
y
ar
tifa
cts
th
at
b
lu
r
th
e
d
is
tin
ct
io
n
b
etwe
en
g
estu
r
e
class
es.
T
h
er
ef
o
r
e,
5
5
0
m
s
was
s
elec
ted
as
th
e
o
p
tim
al
p
ar
am
eter
f
o
r
all
s
u
b
s
eq
u
e
n
t e
x
p
er
im
en
ts
.
T
h
e
s
h
if
t
f
r
o
m
a
6
0
0
m
s
win
d
o
w
(
0
.
9
1
%
ac
cu
r
ac
y
)
to
a
5
5
0
m
s
win
d
o
w
(
9
4
.
4
%
ac
cu
r
ac
y
)
in
d
icate
s
th
at
tem
p
o
r
al
r
eso
lu
tio
n
is
a
cr
itical
f
ac
to
r
in
s
E
MG
p
atte
r
n
r
ec
o
g
n
itio
n
.
A
5
5
0
m
s
d
u
r
atio
n
p
r
o
v
id
es
an
o
p
tim
al
b
alan
ce
;
it
is
lo
n
g
en
o
u
g
h
to
ca
p
t
u
r
e
th
e
ess
en
tial
m
o
to
r
u
n
it
f
ir
in
g
p
atter
n
s
wh
ile
r
em
ain
in
g
s
h
o
r
t
en
o
u
g
h
to
av
o
id
th
e
in
clu
s
io
n
o
f
e
x
ce
s
s
iv
e
tr
an
s
itio
n
n
o
is
e
a
n
d
s
to
ch
asti
c
f
lu
ctu
atio
n
s
th
at
o
f
ten
d
eg
r
a
d
e
d
ee
p
lear
n
in
g
p
e
r
f
o
r
m
an
ce
.
T
h
e
s
tab
ilit
y
o
f
th
e
m
o
d
el
d
u
r
in
g
th
e
tr
ain
in
g
p
h
ase
was
en
s
u
r
ed
th
r
o
u
g
h
t
h
e
im
p
lem
en
tatio
n
o
f
a
lear
n
in
g
r
ate
s
ch
ed
u
ler
an
d
ea
r
ly
s
to
p
p
in
g
[
5
5
]
,
[
5
6
]
.
Fig
u
r
e
4
illu
s
tr
ates
th
e
tr
ain
in
g
an
d
v
alid
atio
n
c
u
r
v
es.
W
h
ile
th
e
tr
ain
in
g
ac
cu
r
ac
y
r
e
ac
h
ed
9
6
%,
th
e
v
alid
ati
o
n
ac
c
u
r
ac
y
s
tab
ilized
at
ap
p
r
o
x
im
at
ely
7
4
%
as
s
h
o
wn
in
Fig
u
r
e
4
(
a)
.
T
h
is
g
ap
s
u
g
g
e
s
ts
a
d
eg
r
ee
o
f
o
v
er
f
itti
n
g
,
wh
ich
is
a
co
m
m
o
n
ch
allen
g
e
in
s
E
MG
-
b
ased
d
ee
p
lear
n
in
g
d
u
e
to
t
h
e
in
h
er
en
t
s
to
ch
asti
city
an
d
s
ig
n
al
-
to
-
n
o
is
e
r
atio
v
ar
ia
b
ilit
y
in
b
io
-
s
ig
n
als.
Ho
wev
er
,
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
1
6
,
No
.
3
,
J
u
n
e
20
2
6
:
1
5
7
6
-
1
5
8
7
1582
im
p
lem
en
tatio
n
o
f
a
lear
n
in
g
r
ate
s
ch
ed
u
ler
an
d
d
r
o
p
o
u
t
(
0
.
5
)
s
u
cc
ess
f
u
lly
p
r
ev
en
te
d
th
e
v
alid
atio
n
lo
s
s
f
r
o
m
d
iv
er
g
in
g
as
s
h
o
wn
in
Fig
u
r
e
4
(
b
)
,
en
s
u
r
in
g
th
at
th
e
m
o
d
el
m
ain
tain
s
its
ab
ilit
y
to
g
en
er
a
lize
ac
r
o
s
s
u
n
s
ee
n
test
in
g
d
ata
r
ath
er
th
a
n
m
er
ely
m
em
o
r
izin
g
th
e
tr
ain
in
g
s
et.
T
ab
le
2
.
T
h
e
o
v
e
r
all
ac
cu
r
ac
y
v
s
win
d
o
w
s
ize
W
i
n
d
o
w
S
i
z
e
(
ms)
Te
st
A
c
c
u
r
a
c
y
4
0
0
0
.
8
3
1
4
5
0
0
.
5
9
3
5
0
0
0
.
4
3
1
5
5
0
0
.
9
4
4
6
0
0
0
.
9
0
9
6
5
0
0
.
6
5
5
7
0
0
0
.
1
8
5
7
5
0
0
.
8
1
5
Fig
u
r
e
3
.
I
m
p
ac
t o
f
v
ar
y
in
g
te
m
p
o
r
al
win
d
o
w
d
u
r
atio
n
s
o
n
th
e
1
D
-
C
NN
class
if
icatio
n
ac
cu
r
ac
y
(
a
)
(
b
)
Fig
u
r
e
4
.
T
r
ain
in
g
a
n
d
v
alid
atio
n
p
er
f
o
r
m
a
n
ce
cu
r
v
es illu
s
tr
atin
g
(
a)
m
o
d
el
co
n
v
er
g
en
ce
an
d
(
b
)
s
tab
ilit
y
u
s
in
g
a
5
5
0
m
s
win
d
o
w
T
h
e
co
n
v
er
g
en
ce
o
b
s
er
v
ed
at
ep
o
ch
1
8
,
wh
er
e
th
e
ea
r
ly
s
to
p
p
in
g
m
ec
h
a
n
is
m
in
ter
v
en
ed
,
p
r
o
v
es
th
at
th
e
m
o
d
el
s
u
cc
ess
f
u
lly
lear
n
e
d
g
en
e
r
alize
d
f
ea
tu
r
es
r
ath
er
t
h
an
m
er
el
y
m
em
o
r
izin
g
t
h
e
tr
ain
in
g
d
ata.
T
h
e
u
s
e
o
f
b
atc
h
n
o
r
m
aliza
tio
n
f
u
r
th
er
co
n
tr
ib
u
ted
to
th
is
s
tab
ilit
y
b
y
n
o
r
m
alizin
g
ac
tiv
atio
n
s
ag
ain
s
t
th
e
n
o
n
-
s
tatio
n
ar
y
n
atu
r
e
o
f
r
aw
s
E
M
G
s
ig
n
als,
allo
win
g
f
o
r
f
aster
an
d
m
o
r
e
r
eliab
le
g
r
ad
ien
t
d
escen
t.
T
h
e
d
etailed
p
er
f
o
r
m
an
ce
o
f
t
h
e
1
D
-
C
NN
m
o
d
el
in
class
if
y
in
g
f
in
g
er
g
estu
r
es
is
ev
alu
ate
d
u
s
in
g
a
C
o
n
f
u
s
io
n
Ma
tr
ix
,
as
s
h
o
wn
in
Fig
u
r
e
5
.
T
ab
le
3
also
d
es
cr
ib
es
th
e
r
esu
lts
o
f
th
e
co
n
f
u
s
io
n
m
atr
ix
.
T
h
is
m
atr
ix
p
r
o
v
id
es
a
c
o
m
p
r
eh
en
s
iv
e
o
v
er
v
iew
o
f
th
e
d
is
tr
ib
u
tio
n
o
f
th
e
m
o
d
el'
s
p
r
ed
ictio
n
s
ag
ain
s
t
th
e
tr
u
e
lab
els o
f
th
e
test
in
g
d
ata.
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
-
8
7
0
8
E
n
h
a
n
cin
g
s
E
MG
fin
g
er g
estu
r
e
r
ec
o
g
n
itio
n
u
s
in
g
o
p
timiz
ed
…
(
Da
n
iel
S
u
to
p
o
P
a
mu
n
g
k
a
s
)
1583
Fig
u
r
e
5
.
C
o
n
f
u
s
io
n
m
atr
i
x
f
o
r
class
if
icatio
n
p
er
f
o
r
m
a
n
ce
T
ab
le
3
.
Su
m
m
a
r
y
ac
c
u
r
ac
y
f
o
r
ea
ch
f
in
g
er
g
r
o
u
p
F
i
n
g
e
r
G
r
o
u
p
A
c
c
u
r
a
c
y
C
l
a
ss
A
n
a
l
y
s
i
s E
r
r
o
r
I
n
d
e
x
1
0
0
%
P
e
r
f
e
c
t
C
l
a
ss
i
f
i
c
a
t
i
o
n
Li
t
t
l
e
F
i
n
g
e
r
1
0
0
%
P
e
r
f
e
c
t
C
l
a
ss
i
f
i
c
a
t
i
o
n
M
i
d
d
l
e
1
0
0
%
P
e
r
f
e
c
t
C
l
a
ss
i
f
i
c
a
t
i
o
n
R
i
n
g
8
3
.
3
%
1
sa
mp
l
e
mi
s
c
l
a
s
si
f
i
e
d
a
s
M
i
d
d
l
e
F
i
n
g
e
r
Th
u
mb
5
0
.
0
%
O
f
t
e
n
c
o
n
f
u
se
d
w
i
t
h
t
h
e
mi
d
d
l
e
f
i
n
g
e
r
a
n
d
t
h
e
r
i
n
g
f
i
n
g
e
r
B
ased
o
n
test
in
g
r
esu
lts
o
n
3
3
d
ata
s
am
p
les,
th
e
m
o
d
el
ac
h
iev
ed
an
ac
c
u
r
ac
y
o
f
8
7
.
9
%.
C
las
s
-
by
-
class
an
aly
s
is
r
ev
ea
led
th
e
f
o
llo
win
g
s
ig
n
if
ican
t
f
in
d
in
g
s
:
P
er
f
ec
t
c
lass
if
icatio
n
(
1
0
0
%
a
c
cu
r
ac
y
)
:
T
h
e
m
o
d
el
d
em
o
n
s
tr
ated
e
x
ce
llen
t
p
er
f
o
r
m
an
ce
with
p
e
r
f
ec
t
class
if
icatio
n
o
f
th
e
in
d
ex
f
in
g
er
(
5
/5
)
,
litt
le
f
in
g
er
(
5
/5
)
,
an
d
m
i
d
d
le
f
in
g
er
(
1
1
/1
1
)
.
T
h
i
s
d
em
o
n
s
tr
ates
th
at
th
e
1
D
-
C
NN
ar
ch
itectu
r
e
is
h
ig
h
l
y
e
f
f
e
ctiv
e
in
r
ec
o
g
n
izin
g
th
e
u
n
iq
u
e
co
n
tr
ac
tio
n
p
atter
n
s
o
f
th
e
f
lex
o
r
d
i
g
ito
r
u
m
s
u
p
er
f
icialis
an
d
e
x
ten
s
o
r
d
ig
i
to
r
u
m
m
u
s
cles
th
at
m
o
v
e
th
ese
th
r
ee
f
in
g
e
r
s
.
Min
o
r
m
is
class
if
icatio
n
(
r
in
g
f
in
g
e
r
)
:
T
h
e
r
in
g
f
in
g
er
ac
h
iev
ed
8
3
.
3
%
ac
cu
r
ac
y
,
with
5
o
u
t
o
f
6
s
am
p
les
co
r
r
ec
tly
class
if
ied
.
On
e
s
am
p
le
was
d
etec
ted
as
t
h
e
m
id
d
le
f
in
g
e
r
.
An
ato
m
icall
y
,
th
is
o
f
ten
o
cc
u
r
s
d
u
e
to
th
e
ten
d
o
n
attac
h
m
en
t
b
etwe
en
th
e
r
in
g
an
d
m
id
d
le
f
in
g
er
s
,
wh
ich
ca
u
s
es
th
e
s
E
M
G
s
ig
n
als
f
r
o
m
th
e
two
f
in
g
er
s
to
h
av
e
h
i
g
h
ly
s
im
ilar
p
atter
n
s
[
5
7
]
.
T
h
e
T
h
u
m
b
c
h
allen
g
e
(
5
0
%
a
cc
u
r
ac
y
)
: T
h
e
b
ig
g
est ch
allen
g
e
was f
o
u
n
d
with
th
e
th
u
m
b
,
wh
ich
o
n
ly
ac
h
iev
ed
5
0
%
ac
cu
r
ac
y
(
3
o
u
t
o
f
6
co
r
r
ec
t
s
am
p
les).
T
h
e
th
u
m
b
was
f
r
eq
u
en
tly
m
is
id
en
tifie
d
as
th
e
m
i
d
d
le
f
in
g
er
(
2
s
am
p
les)
o
r
th
e
r
in
g
f
in
g
er
(
1
s
am
p
le)
.
T
h
is
p
o
o
r
p
er
f
o
r
m
an
ce
is
s
cien
tific
ally
attr
ib
u
ted
to
th
e
p
h
en
o
m
en
o
n
o
f
m
u
s
cle
cr
o
s
s
talk
[
5
8
]
–
[
6
0
]
.
T
h
e
5
0
%
ac
cu
r
ac
y
o
b
s
er
v
ed
f
o
r
th
u
m
b
g
estu
r
es
is
s
cien
tific
ally
s
ig
n
if
ican
t.
An
ato
m
ically
,
t
h
e
m
u
s
cles
r
esp
o
n
s
ib
le
f
o
r
th
u
m
b
o
p
p
o
s
itio
n
,
s
u
c
h
as
th
e
o
p
p
o
n
en
ts
p
o
llicis
,
ar
e
lo
ca
ted
in
t
h
e
d
ee
p
er
la
y
er
s
o
f
t
h
e
f
o
r
ea
r
m
.
T
h
e
My
o
Ar
m
b
an
d
’
s
s
u
r
f
ac
e
elec
tr
o
d
e
s
p
r
im
ar
ily
ca
p
tu
r
e
s
u
p
er
f
icial
m
u
s
cle
ac
tiv
ity
,
le
ad
in
g
t
o
a
'
m
ask
in
g
ef
f
ec
t'
wh
er
e
th
e
s
tr
o
n
g
er
s
ig
n
als
f
r
o
m
t
h
e
f
lex
o
r
d
ig
ito
r
u
m
(
m
id
d
le
an
d
r
i
n
g
f
i
n
g
er
s
)
d
o
m
in
ate
th
e
r
ec
o
r
d
e
d
p
o
ten
tial
[
6
1
]
.
T
h
is
m
u
s
cu
lar
cr
o
s
s
talk
cr
ea
tes
h
ig
h
f
ea
t
u
r
e
s
im
ilar
ity
in
th
e
laten
t
s
p
ac
e,
m
ak
in
g
it
d
if
f
icu
lt
f
o
r
t
h
e
So
f
tMa
x
lay
er
to
estab
lis
h
a
clea
r
d
ec
is
io
n
b
o
u
n
d
ar
y
f
o
r
t
h
u
m
b
m
o
v
em
en
ts
[
6
1
]
.
T
h
is
p
h
en
o
m
en
o
n
is
clea
r
ly
r
ef
lecte
d
in
th
e
co
n
f
u
s
io
n
m
atr
ix
as
s
h
o
wn
in
Fig
u
r
e
5
,
wh
e
r
e
th
e
m
is
class
if
icatio
n
clu
s
ter
s
p
r
ed
o
m
in
a
n
tly
o
cc
u
r
b
etwe
en
th
e
th
u
m
b
an
d
t
h
e
f
in
g
e
r
s
co
n
tr
o
lled
b
y
th
e
s
u
p
er
f
icial
f
lex
o
r
m
u
s
cles.
As
a
r
esu
lt
,
th
e
b
io
elec
tr
ic
s
ig
n
als
ca
p
t
u
r
ed
b
y
th
e
My
o
Ar
m
b
an
d
'
s
s
u
r
f
ac
e
s
en
s
o
r
s
ar
e
o
f
ten
atten
u
ate
d
o
r
m
ask
ed
b
y
s
tr
o
n
g
er
s
u
p
er
f
icial
m
u
s
cle
ac
tiv
ity
[
6
2
]
,
[
6
3
]
.
Ov
er
all,
d
esp
ite
t
h
e
co
n
s
tr
ain
ts
o
n
th
u
m
b
d
etec
tio
n
,
th
e
m
o
d
el'
s
ab
ilit
y
to
class
if
y
th
e
m
ajo
r
ity
o
f
m
o
v
em
en
ts
with
v
er
y
h
ig
h
ac
cu
r
ac
y
(
ab
o
v
e
8
3
%
to
1
0
0
%).
T
h
is
in
d
icate
s
th
at
th
e
u
s
e
o
f
Z
-
s
co
r
e
n
o
r
m
aliza
tio
n
an
d
5
5
0
m
s
win
d
o
w
s
ize
o
p
tim
izatio
n
h
as
s
u
cc
ess
f
u
lly
m
in
im
ized
th
e
v
a
r
iab
ilit
y
o
f
s
E
MG
s
ig
n
als an
d
p
r
o
v
id
ed
s
tab
le
te
m
p
o
r
al
f
ea
t
u
r
es f
o
r
th
e
1
D
-
C
NN
m
o
d
el.
T
h
e
in
teg
r
atio
n
o
f
a
lear
n
i
n
g
r
ate
s
ch
ed
u
ler
an
d
ea
r
ly
s
to
p
p
i
n
g
s
u
cc
ess
f
u
lly
s
tab
ilized
th
e
v
alid
atio
n
cu
r
v
e,
p
r
ev
en
tin
g
th
e
m
o
d
el
f
r
o
m
o
v
e
r
-
ad
ju
s
tin
g
to
n
o
is
e
in
th
e
E
MG
s
ig
n
al
[
6
4
]
,
[
6
5
]
.
Mo
d
e
r
n
s
ig
n
al
p
r
o
ce
s
s
in
g
tech
n
i
q
u
es
an
d
ed
g
e
co
m
p
u
tin
g
c
o
n
s
id
er
atio
n
s
f
u
r
th
er
v
alid
ate
th
is
ap
p
r
o
ac
h
f
o
r
f
u
tu
r
e
wea
r
ab
le
d
ev
ices.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
1
6
,
No
.
3
,
J
u
n
e
20
2
6
:
1
5
7
6
-
1
5
8
7
1584
4.
CO
NCLU
SI
O
N
T
h
is
s
tu
d
y
h
as
s
u
cc
ess
f
u
lly
d
ev
elo
p
ed
a
n
o
p
tim
ized
1
D
c
o
n
v
o
lu
ti
o
n
al
n
eu
r
al
n
etwo
r
k
(
1
D
-
C
NN)
ar
ch
itectu
r
e
f
o
r
f
in
e
-
g
r
ain
e
d
f
in
g
er
g
estu
r
e
r
ec
o
g
n
itio
n
u
s
in
g
r
aw
s
E
MG
s
ig
n
als.
T
h
e
e
x
p
er
im
en
tal
r
esu
lts
d
em
o
n
s
tr
ate
th
at
th
e
p
r
o
p
o
s
ed
m
o
d
el
ac
h
ie
v
es
a
p
ea
k
class
if
icatio
n
ac
cu
r
ac
y
o
f
9
4
.
4
%
f
o
r
2
0
d
is
tin
ct
f
in
g
er
m
o
v
em
en
ts
.
T
h
is
p
er
f
o
r
m
a
n
c
e
u
n
d
er
s
co
r
es
th
e
s
u
p
er
io
r
ity
o
f
en
d
-
to
-
en
d
d
ee
p
lea
r
n
in
g
i
n
ca
p
tu
r
in
g
co
m
p
lex
tem
p
o
r
al
p
atter
n
s
d
ir
ec
tly
f
r
o
m
r
aw
b
io
elec
tr
ic
s
ig
n
als with
o
u
t th
e
n
ee
d
f
o
r
m
an
u
al
f
ea
tu
r
e
ex
tr
ac
tio
n
.
A
k
ey
f
i
n
d
in
g
o
f
th
is
r
esear
ch
is
th
e
s
ig
n
if
ica
n
t
im
p
a
ct
o
f
tem
p
o
r
al
win
d
o
win
g
o
n
m
o
d
el
p
er
f
o
r
m
an
ce
.
R
ed
u
cin
g
t
h
e
wi
n
d
o
w
s
ize
f
r
o
m
6
0
0
to
5
5
0
m
s
p
r
o
v
e
d
to
b
e
th
e
o
p
tim
al
s
tr
a
teg
y
,
p
r
o
v
id
i
n
g
t
h
e
n
ec
ess
ar
y
tem
p
o
r
al
r
eso
lu
tio
n
to
ca
p
tu
r
e
m
u
s
cle
co
n
tr
ac
tio
n
d
y
n
am
ics
wh
ile
m
ain
tain
in
g
h
ig
h
s
tab
ilit
y
.
T
h
e
im
p
lem
en
tatio
n
o
f
a
lear
n
in
g
r
ate
s
ch
ed
u
ler
an
d
ea
r
ly
s
to
p
p
in
g
f
u
r
th
er
en
s
u
r
e
d
th
e
m
o
d
el'
s
r
o
b
u
s
tn
ess
,
p
r
ev
en
tin
g
o
v
e
r
f
itti
n
g
d
esp
ite
th
e
s
to
ch
asti
c
n
atu
r
e
o
f
s
E
MG
d
ata.
W
h
ile
th
e
m
o
d
el
ac
h
iev
ed
1
0
0
%
ac
cu
r
ac
y
f
o
r
th
e
in
d
ex
,
m
id
d
le,
an
d
litt
le
f
in
g
er
s
,
th
e
lo
wer
p
er
f
o
r
m
an
ce
o
b
s
er
v
ed
f
o
r
th
e
th
u
m
b
(
5
0
%)
h
ig
h
lig
h
ts
th
e
p
er
s
is
ten
t
ch
allen
g
e
o
f
m
u
s
cle
cr
o
s
s
talk
f
r
o
m
d
ee
p
er
an
ato
m
ical
lay
er
s
,
s
p
e
cif
ically
th
e
o
p
p
o
n
e
n
ts
p
o
llicis
,
wh
ich
ar
e
less
ac
ce
s
s
ib
le
to
s
u
r
f
ac
e
elec
tr
o
d
es.
T
h
e
m
o
d
el
ac
h
ie
v
ed
a
p
ea
k
test
ac
cu
r
ac
y
o
f
9
4
.
4
%
u
n
d
er
o
p
tim
al
co
n
d
itio
n
s
,
with
an
o
v
er
all
a
v
er
ag
e
ac
cu
r
ac
y
o
f
8
7
.
9
%
ac
r
o
s
s
all
s
p
ec
if
ic
test
s
s
h
o
wn
in
th
e
m
atr
ix
.
I
n
co
n
cl
u
s
io
n
,
th
e
o
p
tim
ized
1
D
-
C
NN
f
r
am
ewo
r
k
o
f
f
er
s
a
h
i
g
h
ly
ac
cu
r
ate
an
d
s
ca
lab
le
s
o
lu
tio
n
f
o
r
m
u
lticlas
s
g
estu
r
e
r
ec
o
g
n
it
io
n
,
p
a
v
in
g
t
h
e
way
f
o
r
m
o
r
e
in
tu
itiv
e
a
n
d
p
r
ec
is
e
co
n
tr
o
l sy
s
tem
s
in
n
ex
t
-
g
en
er
a
tio
n
p
r
o
s
th
etic
h
a
n
d
s
.
Fo
r
f
u
tu
r
e
d
ev
el
o
p
m
en
t
an
d
r
esear
ch
,
s
ev
er
al
s
tr
ateg
ic
n
ee
d
s
s
h
o
u
ld
b
e
ad
d
r
ess
ed
with
th
e
ap
p
r
o
p
r
iate
s
y
s
tem
.
T
h
e
p
r
im
ar
y
f
o
cu
s
s
h
o
u
ld
b
e
o
n
im
p
r
o
v
in
g
class
if
icatio
n
ac
cu
r
ac
y
f
o
r
f
in
g
er
a
n
d
th
u
m
b
,
wh
ich
ca
n
b
e
ac
h
iev
ed
b
y
ad
d
in
g
an
s
E
MG
s
en
s
o
r
o
v
e
r
th
e
t
h
en
ar
m
u
s
cle
o
r
b
y
co
m
b
i
n
in
g
f
r
eq
u
e
n
cy
-
d
o
m
ain
f
ea
tu
r
es
tailo
r
ed
to
d
if
f
icu
lt
class
es.
I
n
ad
d
itio
n
,
co
n
s
i
d
er
in
g
th
e
s
tu
d
y
,
th
is
is
s
till
d
o
n
e
o
f
f
lin
e
.
I
m
p
lem
en
tatio
n
o
f
th
e
m
o
d
el
f
o
r
r
ea
l
-
tim
e
in
-
s
y
s
tem
co
n
t
r
o
l
is
h
ig
h
ly
r
ec
o
m
m
e
n
d
ed
t
o
ac
h
iev
e
test
-
lev
el
laten
cy
an
d
r
esp
o
n
s
iv
en
ess
in
h
an
d
p
r
o
s
th
etics o
r
r
o
b
o
tics
.
Fu
r
th
er
ex
p
er
im
en
tatio
n
is
n
e
ce
s
s
ar
y
,
in
v
o
lv
in
g
ad
d
itio
n
al
s
u
b
jects
an
d
test
in
g
in
v
ar
io
u
s
p
o
s
itio
n
s
an
d
ar
m
p
o
s
tu
r
es,
to
en
h
an
ce
th
e
m
o
d
el'
s
ca
p
ac
ity
to
g
en
er
alize
to
ch
an
g
es
in
s
en
s
o
r
o
r
ien
tatio
n
an
d
m
u
s
cle
f
atig
u
e.
Ad
d
itio
n
ally
,
ex
p
lo
r
in
g
h
y
b
r
id
ar
c
h
itectu
r
es,
s
u
ch
as
C
NN
-
L
STM
,
is
ad
v
is
ab
le.
T
h
e
ab
ilit
y
to
ca
p
tu
r
e
tem
p
o
r
al
d
ep
en
d
en
cies
ca
n
m
o
r
e
ef
f
ec
tiv
ely
r
ed
u
ce
e
r
r
o
r
s
in
class
if
y
in
g
f
in
g
er
s
ex
h
ib
itin
g
p
atter
n
co
n
tr
ac
tio
n
s
ak
in
to
t
h
o
s
e
o
f
m
u
s
cles.
W
ith
th
e
d
ev
elo
p
m
e
n
ts
,
it
is
an
ticip
ated
th
at
th
e
s
y
s
tem
f
o
r
in
tr
o
d
u
cin
g
f
in
g
er
m
o
v
e
m
en
t
ca
n
e
v
o
lv
e
in
to
a
s
o
lu
tio
n
f
o
r
m
o
r
e
p
r
ec
is
e,
r
eliab
le,
an
d
r
ea
d
ily
a
v
ailab
le
h
ar
d
war
e
-
in
teg
r
ated
p
r
o
s
th
etics.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
is
r
esear
ch
was
s
u
p
p
o
r
ted
b
y
Po
litek
n
ik
Neg
er
i
B
atam
th
r
o
u
g
h
th
e
Pro
ject
B
ased
L
ea
r
n
in
g
(
PB
L
)
s
ch
em
e.
T
h
e
au
th
o
r
s
wo
u
ld
lik
e
to
ac
k
n
o
wled
g
e
th
e
f
in
a
n
cial
an
d
in
s
titu
tio
n
al
s
u
p
p
o
r
t
p
r
o
v
id
ed
b
y
th
e
in
s
titu
tio
n
f
o
r
th
e
c
o
m
p
letio
n
o
f
th
is
wo
r
k
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Dan
iel
S
u
to
p
o
Pam
u
n
g
k
as
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Su
m
an
tr
i K
.
R
is
an
d
r
iy
a
✓
✓
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict
o
f
in
ter
est.
T
h
e
a
u
th
o
r
s
d
ec
lar
e
th
at
th
ey
h
av
e
n
o
k
n
o
wn
co
m
p
etin
g
f
in
an
cial
in
ter
ests
o
r
p
er
s
o
n
al
r
elatio
n
s
h
ip
s
th
at
co
u
ld
h
av
e
ap
p
ea
r
ed
to
in
f
lu
en
ce
th
e
w
o
r
k
r
ep
o
r
ted
in
th
is
p
ap
er
.
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
-
8
7
0
8
E
n
h
a
n
cin
g
s
E
MG
fin
g
er g
estu
r
e
r
ec
o
g
n
itio
n
u
s
in
g
o
p
timiz
ed
…
(
Da
n
iel
S
u
to
p
o
P
a
mu
n
g
k
a
s
)
1585
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
in
g
s
o
f
th
is
s
tu
d
y
ar
e
av
ailab
le
f
r
o
m
th
e
co
r
r
esp
o
n
d
i
n
g
au
t
h
o
r
(
DSP),
u
p
o
n
r
ea
s
o
n
ab
le
r
eq
u
est.
RE
F
E
R
E
NC
E
S
[
1
]
R
.
R
o
d
r
i
g
u
e
s,
D
.
M
i
r
a
n
d
a
,
V
.
C
a
r
v
a
l
h
o
,
a
n
d
D
.
M
a
t
o
s,
“
D
e
si
g
n
a
n
d
d
e
v
e
l
o
p
men
t
o
f
a
n
E
M
G
u
p
p
e
r
LI
M
B
c
o
n
t
r
o
l
l
e
d
p
r
o
st
h
e
s
i
s:
a
p
r
e
l
i
mi
n
a
r
y
a
p
p
r
o
a
c
h
,
”
A
c
t
u
a
t
o
rs
,
v
o
l
.
1
4
,
n
o
.
5
,
p
.
2
1
9
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
a
c
t
1
4
0
5
0
2
1
9
.
[
2
]
H
.
Le
e
,
Y
.
L
e
e
,
L.
J
i
a
n
g
,
J.
C
h
e
n
,
W
.
L.
L
a
m,
a
n
d
Y
.
T.
Lo
,
“
A
d
v
a
n
c
e
s
i
n
m
y
o
e
l
e
c
t
r
i
c
n
e
u
r
o
p
r
o
st
h
e
t
i
c
s:
A
n
a
r
r
a
t
i
v
e
r
e
v
i
e
w
o
f
c
u
r
r
e
n
t
t
r
e
n
d
s,
c
h
a
l
l
e
n
g
e
s,
a
n
d
f
u
t
u
r
e
d
i
r
e
c
t
i
o
n
s,
”
Ad
v
a
n
c
e
d
T
e
c
h
n
o
l
o
g
y
in
N
e
u
r
o
sc
i
e
n
c
e
,
v
o
l
.
2
,
n
o
.
4
,
p
p
.
2
0
4
–
2
1
5
,
2
0
2
5
,
d
o
i
:
1
0
.
4
1
0
3
/
a
t
n
.
a
t
n
-
d
-
25
-
0
0
0
0
7
.
[
3
]
S
.
W
a
n
g
,
J
.
Zh
e
n
g
,
B
.
Zh
e
n
g
,
a
n
d
X
.
Ji
a
n
g
,
“
P
h
a
se
-
b
a
s
e
d
g
r
a
s
p
c
l
a
s
si
f
i
c
a
t
i
o
n
f
o
r
p
r
o
s
t
h
e
t
i
c
h
a
n
d
c
o
n
t
r
o
l
u
si
n
g
sE
M
G
,
”
Bi
o
s
e
n
so
rs
,
v
o
l
.
1
2
,
n
o
.
2
,
p
.
5
7
,
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
b
i
o
s
1
2
0
2
0
0
5
7
.
[
4
]
M
.
Li
n
a
r
d
a
k
i
s
,
I
.
V
a
r
l
a
m
i
s,
a
n
d
G
.
T.
P
a
p
a
d
o
p
o
u
l
o
s,
“
S
u
r
v
e
y
o
n
h
a
n
d
g
e
s
t
u
r
e
r
e
c
o
g
n
i
t
i
o
n
f
r
o
m
v
i
s
u
a
l
i
n
p
u
t
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
3
,
p
p
.
1
3
5
3
7
3
–
1
3
5
4
0
6
,
2
0
2
5
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
5
.
3
5
9
3
4
2
8
.
[
5
]
A
.
R
.
A
l
i
,
M
.
W
.
A
.
R
a
m
a
d
a
n
,
a
n
d
P
.
Z
o
me
t
a
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
f
o
r
EM
G
-
b
a
s
e
d
g
e
st
u
r
e
r
e
c
o
g
n
i
t
i
o
n
i
n
b
r
a
i
n
–
c
o
mp
u
t
e
r
i
n
t
e
r
f
a
c
e
s
a
n
d
h
u
ma
n
o
i
d
r
o
b
o
t
s,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
R
o
b
o
t
i
c
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
9
,
n
o
.
4
,
p
p
.
1
7
8
9
–
1
8
0
0
,
2
0
2
5
,
d
o
i
:
1
0
.
1
0
0
7
/
s4
1
3
1
5
-
0
2
5
-
0
0
4
6
3
-
1.
[
6
]
R
.
W
e
n
,
W
.
L
.
Ta
y
,
B
.
P
.
N
g
u
y
e
n
,
C
.
B
.
C
h
n
g
,
a
n
d
C
.
K
.
C
h
u
i
,
“
H
a
n
d
g
e
s
t
u
r
e
g
u
i
d
e
d
r
o
b
o
t
-
a
ss
i
st
e
d
s
u
r
g
e
r
y
b
a
se
d
o
n
a
d
i
r
e
c
t
a
u
g
me
n
t
e
d
r
e
a
l
i
t
y
i
n
t
e
r
f
a
c
e
,
”
C
o
m
p
u
t
e
r
M
e
t
h
o
d
s
a
n
d
Pr
o
g
r
a
m
s
i
n
Bi
o
m
e
d
i
c
i
n
e
,
v
o
l
.
1
1
6
,
n
o
.
2
,
p
p
.
6
8
–
8
0
,
2
0
1
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
m
p
b
.
2
0
1
3
.
1
2
.
0
1
8
.
[
7
]
F
.
L
a
g
a
n
à
e
t
a
l
.
,
“
D
e
v
e
l
o
p
m
e
n
t
o
f
a
n
i
n
t
e
g
r
a
t
e
d
s
y
st
e
m
o
f
sE
M
G
si
g
n
a
l
a
c
q
u
i
s
i
t
i
o
n
,
p
r
o
c
e
ss
i
n
g
,
a
n
d
a
n
a
l
y
s
i
s
w
i
t
h
A
I
t
e
c
h
n
i
q
u
e
s,
”
S
i
g
n
a
l
s
,
v
o
l
.
5
,
n
o
.
3
,
p
p
.
4
7
6
–
4
9
3
,
2
0
2
4
,
d
o
i
:
1
0
.
3
3
9
0
/
s
i
g
n
a
l
s
5
0
3
0
0
2
5
.
[
8
]
C
.
C
u
i
,
M
.
S
.
S
u
n
a
r
,
a
n
d
G
.
E
g
S
u
,
“
D
e
e
p
v
i
si
o
n
-
b
a
s
e
d
r
e
a
l
-
t
i
me
h
a
n
d
g
e
s
t
u
r
e
r
e
c
o
g
n
i
t
i
o
n
:
a
r
e
v
i
e
w
,
”
P
e
e
rJ
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
1
1
,
p
.
e
2
9
2
1
,
2
0
2
5
,
d
o
i
:
1
0
.
7
7
1
7
/
p
e
e
r
j
-
c
s.
2
9
2
1
.
[
9
]
A
.
D
a
h
i
y
a
,
D
.
W
a
d
h
w
a
,
R
.
K
a
t
t
i
,
a
n
d
L.
G
.
O
c
c
h
i
p
i
n
t
i
,
“
Ef
f
i
c
i
e
n
t
h
a
n
d
g
e
s
t
u
r
e
r
e
c
o
g
n
i
t
i
o
n
u
si
n
g
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
I
MU
-
b
a
s
e
d
w
e
a
r
a
b
l
e
d
e
v
i
c
e
,
”
I
E
EE
S
e
n
s
o
r
s Let
t
e
rs
,
v
o
l
.
8
,
n
o
.
1
2
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
LS
EN
S
.
2
0
2
4
.
3
5
0
1
5
8
6
.
[
1
0
]
D
.
K
u
mar
a
n
d
A
.
G
a
n
e
s
h
,
“
A
c
r
i
t
i
c
a
l
r
e
v
i
e
w
o
n
h
a
n
d
g
e
st
u
r
e
r
e
c
o
g
n
i
t
i
o
n
u
si
n
g
sE
M
G
:
C
h
a
l
l
e
n
g
e
s,
a
p
p
l
i
c
a
t
i
o
n
,
p
r
o
c
e
s
s
a
n
d
t
e
c
h
n
i
q
u
e
s,
”
J
o
u
rn
a
l
o
f
Ph
y
si
c
s:
C
o
n
f
e
re
n
c
e
S
e
ri
e
s
,
v
o
l
.
2
3
2
7
,
n
o
.
1
,
p
.
1
2
0
7
5
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
8
8
/
1
7
4
2
-
6
5
9
6
/
2
3
2
7
/
1
/
0
1
2
0
7
5
.
[
1
1
]
W
.
L
i
,
P
.
S
h
i
,
a
n
d
H
.
Y
u
,
“
G
e
s
t
u
r
e
r
e
c
o
g
n
i
t
i
o
n
u
s
i
n
g
su
r
f
a
c
e
e
l
e
c
t
r
o
my
o
g
r
a
p
h
y
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
p
r
o
s
t
h
e
s
e
s
h
a
n
d
:
st
a
t
e
-
of
-
t
h
e
-
a
r
t
,
c
h
a
l
l
e
n
g
e
s
,
a
n
d
f
u
t
u
r
e
,
”
Fr
o
n
t
i
e
rs
i
n
N
e
u
r
o
s
c
i
e
n
c
e
,
v
o
l
.
1
5
,
A
p
r
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
8
9
/
f
n
i
n
s.
2
0
2
1
.
6
2
1
8
8
5
.
[
1
2
]
M
.
S
.
K
a
r
i
s
e
t
a
l
.
,
“
En
s
e
m
b
l
e
v
o
t
i
n
g
r
e
g
r
e
ss
o
r
f
o
r
e
n
h
a
n
c
e
d
p
r
e
d
i
c
t
i
o
n
i
n
E
M
G
-
b
a
s
e
d
p
r
o
st
h
e
t
i
c
w
r
i
s
t
c
o
n
t
r
o
l
,
”
J
o
u
r
n
a
l
o
f
Ro
b
o
t
i
c
s
a
n
d
C
o
n
t
ro
l
(
J
RC
)
,
v
o
l
.
6
,
n
o
.
4
,
p
p
.
1
8
7
2
–
1
8
8
4
,
2
0
2
5
,
d
o
i
:
1
0
.
1
8
1
9
6
/
j
r
c
.
v
6
i
4
.
2
6
2
2
2
.
[
1
3
]
U
.
C
ô
t
é
-
A
l
l
a
r
d
e
t
a
l
.
,
“
D
e
e
p
l
e
a
r
n
i
n
g
f
o
r
e
l
e
c
t
r
o
m
y
o
g
r
a
p
h
i
c
h
a
n
d
g
e
st
u
r
e
s
i
g
n
a
l
c
l
a
ssi
f
i
c
a
t
i
o
n
u
si
n
g
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
,
”
I
EEE
T
ra
n
s
a
c
t
i
o
n
s
o
n
N
e
u
r
a
l
S
y
st
e
m
s
a
n
d
Re
h
a
b
i
l
i
t
a
t
i
o
n
E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
2
7
,
n
o
.
4
,
p
p
.
7
6
0
–
7
7
1
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
0
9
/
TN
S
R
E.
2
0
1
9
.
2
8
9
6
2
6
9
.
[
1
4
]
F
.
S
.
B
o
t
r
o
s
,
A
.
P
h
i
n
y
o
m
a
r
k
,
a
n
d
E.
J.
S
c
h
e
me
,
“
D
a
y
-
to
-
d
a
y
st
a
b
i
l
i
t
y
o
f
w
r
i
st
E
M
G
f
o
r
w
e
a
r
a
b
l
e
-
b
a
se
d
h
a
n
d
g
e
st
u
r
e
r
e
c
o
g
n
i
t
i
o
n
,
”
I
EEE
A
c
c
e
ss
,
v
o
l
.
1
0
,
p
p
.
1
2
5
9
4
2
–
1
2
5
9
5
4
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
2
.
3
2
2
5
7
6
1
.
[
1
5
]
I
.
K
a
r
a
c
a
n
a
n
d
K
.
S
.
T
ü
r
k
e
r
,
“
A
c
o
m
p
a
r
i
s
o
n
o
f
e
l
e
c
t
r
o
m
y
o
g
r
a
p
h
y
t
e
c
h
n
i
q
u
e
s:
s
u
r
f
a
c
e
v
e
r
su
s
i
n
t
r
a
m
u
sc
u
l
a
r
r
e
c
o
r
d
i
n
g
,
”
Eu
r
o
p
e
a
n
J
o
u
rn
a
l
o
f
Ap
p
l
i
e
d
Ph
y
si
o
l
o
g
y
,
v
o
l
.
1
2
5
,
n
o
.
1
,
p
p
.
7
–
2
3
,
2
0
2
5
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
4
2
1
-
0
2
4
-
0
5
6
4
0
-
x.
[
1
6
]
A
.
R
.
A
s
i
f
e
t
a
l
.
,
“
P
e
r
f
o
r
ma
n
c
e
e
v
a
l
u
a
t
i
o
n
o
f
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
f
o
r
h
a
n
d
g
e
st
u
r
e
r
e
c
o
g
n
i
t
i
o
n
u
s
i
n
g
EM
G
,
”
S
e
n
so
rs
(
S
w
i
t
zer
l
a
n
d
)
,
v
o
l
.
2
0
,
n
o
.
6
,
p
.
1
6
4
2
,
2
0
2
0
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
0
0
6
1
6
4
2
.
[
1
7
]
A
.
S
u
l
t
a
n
a
,
M
.
T.
I
.
O
p
u
,
M
.
S
.
A
l
a
m,
a
n
d
F
.
A
h
m
e
d
,
“
D
e
c
o
m
p
o
si
t
i
o
n
a
n
d
m
u
l
t
i
-
sca
l
e
a
n
a
l
y
si
s
o
f
s
u
r
f
a
c
e
e
l
e
c
t
r
o
m
y
o
g
r
a
p
h
i
c
si
g
n
a
l
f
o
r
f
i
n
g
e
r
m
o
v
e
me
n
t
s,”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
E
l
e
c
t
r
i
c
a
l
a
n
d
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
ri
n
g
(
I
J
EC
E)
,
v
o
l
.
1
5
,
n
o
.
5
,
p
.
4
5
9
3
,
2
0
2
5
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
c
e
.
v
1
5
i
5
.
p
p
4
5
9
3
-
4
6
0
4
.
[
1
8
]
W
.
C
a
e
s
a
r
e
n
d
r
a
,
T.
T
j
a
h
j
o
w
i
d
o
d
o
,
a
n
d
D
.
P
a
m
u
n
g
k
a
s
,
“
E
M
G
b
a
se
d
c
l
a
s
si
f
i
c
a
t
i
o
n
o
f
h
a
n
d
g
e
st
u
r
e
s
u
si
n
g
P
C
A
a
n
d
A
N
F
I
S
,
”
i
n
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
2
0
1
7
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
R
o
b
o
t
i
c
s
,
B
i
o
m
i
m
e
t
i
c
s,
a
n
d
I
n
t
e
l
l
i
g
e
n
t
C
o
m
p
u
t
a
t
i
o
n
a
l
S
y
st
e
m
s,
R
o
b
i
o
n
e
t
i
c
s
2
0
1
7
,
2
0
1
7
,
v
o
l
.
2
0
1
7
-
D
e
c
e
mb
e
r
,
p
p
.
1
8
–
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
R
O
B
I
O
N
ETI
C
S
.
2
0
1
7
.
8
2
0
3
4
3
0
.
[
1
9
]
A
.
R
.
S
o
b
i
n
o
v
a
n
d
S
.
J.
B
e
n
sma
i
a
,
“
Th
e
n
e
u
r
a
l
mec
h
a
n
i
sms
o
f
ma
n
u
a
l
d
e
x
t
e
r
i
t
y
,
”
N
a
t
u
re
R
e
v
i
e
w
s
N
e
u
r
o
sc
i
e
n
c
e
,
v
o
l
.
2
2
,
n
o
.
1
2
,
p
p
.
7
4
1
–
7
5
7
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
3
8
/
s
4
1
5
8
3
-
0
2
1
-
0
0
5
2
8
-
7.
[
2
0
]
B
.
E
.
L
u
n
g
a
n
d
B
.
B
u
r
n
s,
“
A
n
a
t
o
m
y
,
sh
o
u
l
d
e
r
a
n
d
u
p
p
e
r
l
i
m
b
,
h
a
n
d
f
l
e
x
o
r
d
i
g
i
t
o
r
u
m
p
r
o
f
u
n
d
u
s m
u
s
c
l
e
,
”
S
t
a
t
P
e
a
rl
s
,
2
0
1
9
.
[
2
1
]
M
.
R
.
K
.
K
a
d
a
v
a
t
h
,
M
.
N
a
s
o
r
,
a
n
d
A
.
I
mr
a
n
,
“
E
n
h
a
n
c
e
d
h
a
n
d
g
e
s
t
u
r
e
r
e
c
o
g
n
i
t
i
o
n
w
i
t
h
s
u
r
f
a
c
e
e
l
e
c
t
r
o
m
y
o
g
r
a
m
a
n
d
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
,
”
S
e
n
s
o
rs
,
v
o
l
.
2
4
,
n
o
.
1
6
,
2
0
2
4
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
4
1
6
5
2
3
1
.
[
2
2
]
A
.
S
h
a
r
ma,
I
.
S
h
a
r
ma
,
a
n
d
A
.
K
u
m
a
r
,
“
S
i
g
n
a
l
a
c
q
u
i
s
i
t
i
o
n
a
n
d
t
i
m
e
–
f
r
e
q
u
e
n
c
y
p
e
r
sp
e
c
t
i
v
e
o
f
E
M
G
s
i
g
n
a
l
-
b
a
sed
sy
st
e
ms
a
n
d
a
p
p
l
i
c
a
t
i
o
n
s,
”
I
ETE
T
e
c
h
n
i
c
a
l
R
e
v
i
e
w
(
I
n
s
t
i
t
u
t
i
o
n
o
f
El
e
c
t
ro
n
i
c
s
a
n
d
T
e
l
e
c
o
m
m
u
n
i
c
a
t
i
o
n
E
n
g
i
n
e
e
rs
,
I
n
d
i
a
)
,
v
o
l
.
4
1
,
n
o
.
4
,
p
p
.
4
6
6
–
4
8
5
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
8
0
/
0
2
5
6
4
6
0
2
.
2
0
2
3
.
2
2
6
5
8
9
7
.
[
2
3
]
A
.
K
.
Y
a
d
a
v
,
M
.
B
i
n
K
h
a
l
e
e
q
,
A
.
A
.
K
h
a
n
,
a
n
d
M
.
S
a
r
f
r
a
z
,
“
S
e
l
e
c
t
i
o
n
o
f
k
e
y
M
u
sc
l
e
s
a
n
d
g
e
s
t
u
r
e
s
f
o
r
r
e
l
i
a
b
l
e
c
o
n
t
r
o
l
i
n
E
M
G
-
b
a
s
e
d
w
r
i
st
p
r
o
s
t
h
e
t
i
c
s,
”
2
0
2
5
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
C
o
g
n
i
t
i
v
e
C
o
m
p
u
t
i
n
g
i
n
E
n
g
i
n
e
e
r
i
n
g
,
C
o
m
m
u
n
i
c
a
t
i
o
n
s,
S
c
i
e
n
c
e
s
a
n
d
Bi
o
m
e
d
i
c
a
l
H
e
a
l
t
h
I
n
f
o
rm
a
t
i
c
s,
I
C
3
E
C
S
BH
I
2
0
2
5
,
p
p
.
4
3
2
–
4
3
7
,
2
0
2
5
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
3
E
C
S
B
H
I
6
3
5
9
1
.
2
0
2
5
.
1
0
9
9
1
2
8
4
.
[
2
4
]
T.
Za
i
m,
S
.
A
b
d
e
l
-
H
a
d
i
,
R
.
M
a
h
m
o
u
d
,
A
.
K
h
a
n
d
a
k
a
r
,
S
.
M
.
R
a
k
h
t
a
l
a
,
a
n
d
M
.
E
.
H
.
C
h
o
w
d
h
u
r
y
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
-
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
-
b
a
s
e
d
my
o
e
l
e
c
t
r
i
c
c
o
n
t
r
o
l
s
y
st
e
m
f
o
r
u
p
p
e
r
LI
M
B
r
e
h
a
b
i
l
i
t
a
t
i
o
n
u
t
i
l
i
z
i
n
g
EEG
a
n
d
E
M
G
si
g
n
a
l
s
:
a
sy
s
t
e
ma
t
i
c
r
e
v
i
e
w
,
”
Bi
o
e
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
1
2
,
n
o
.
2
,
2
0
2
5
,
d
o
i
:
1
0
.
3
3
9
0
/
b
i
o
e
n
g
i
n
e
e
r
i
n
g
1
2
0
2
0
1
4
4
.
[
2
5
]
F
.
Ev
c
i
,
A
.
Ef
e
k
a
a
n
Ef
e
,
a
n
d
E.
I
.
K
o
n
u
k
se
v
e
n
,
“
A
c
o
mp
r
e
h
e
n
s
i
v
e
a
n
a
l
y
si
s
o
f
LD
A
,
S
V
M
,
a
n
d
n
e
u
r
a
l
n
e
t
w
o
r
k
a
l
g
o
r
i
t
h
ms
i
n
mu
l
t
i
c
l
a
ss
my
o
e
l
e
c
t
r
i
c
i
d
e
n
t
i
f
i
c
a
t
i
o
n
o
f
Li
m
b
mo
v
e
me
n
t
s
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
f
C
o
n
t
r
o
l
,
D
y
n
a
m
i
c
S
y
st
e
m
s,
a
n
d
R
o
b
o
t
i
c
s
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
1
5
9
/
c
d
sr
2
4
.
1
1
0
.
[
2
6
]
L.
D
e
l
a
,
D
.
S
u
t
o
p
o
,
S
.
K
u
r
n
i
a
w
a
n
,
T.
Tj
a
h
j
o
w
i
d
o
d
o
,
a
n
d
W
.
C
a
e
s
a
r
e
n
d
r
a
,
“
EM
G
b
a
s
e
d
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
h
a
n
d
g
e
st
u
r
e
u
s
i
n
g
P
C
A
a
n
d
S
V
M
,
”
L
e
c
t
u
re
N
o
t
e
s
i
n
El
e
c
t
r
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
8
9
8
,
p
p
.
4
5
9
–
4
7
7
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
981
-
19
-
1
8
0
4
-
9
_
3
5
.
[
2
7
]
P
.
S
a
h
u
,
B
.
K
.
S
i
n
g
h
,
a
n
d
N
.
N
i
r
a
l
a
,
“
O
p
t
i
m
i
z
e
d
k
-
n
e
a
r
e
s
t
n
e
i
g
h
b
o
r
s
f
o
r
c
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
p
r
o
s
t
h
e
t
i
c
h
a
n
d
m
o
v
e
m
e
n
t
s
u
si
n
g
e
l
e
c
t
r
o
m
y
o
g
r
a
p
h
y
s
i
g
n
a
l
,
”
E
n
g
i
n
e
e
r
i
n
g
A
p
p
l
i
c
a
t
i
o
n
s
o
f
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
1
3
3
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
n
g
a
p
p
a
i
.
2
0
2
4
.
1
0
8
3
9
0
.
[
2
8
]
M
.
Z
h
e
n
g
,
K
.
J
i
a
n
g
,
R
.
X
u
,
a
n
d
L.
Q
i
,
“
A
n
a
d
a
p
t
i
v
e
LD
A
o
p
t
i
ma
l
t
o
p
i
c
n
u
mb
e
r
s
e
l
e
c
t
i
o
n
m
e
t
h
o
d
i
n
n
e
w
s
t
o
p
i
c
i
d
e
n
t
i
f
i
c
a
t
i
o
n
,
”
I
EEE
A
c
c
e
ss
,
v
o
l
.
1
1
,
p
p
.
9
2
2
7
3
–
9
2
2
8
4
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
3
.
3
3
0
8
5
2
0
.
[
2
9
]
H
.
N
i
u
,
G
.
B
.
M
c
C
a
l
l
u
m,
A
.
B
.
C
h
a
n
g
,
K
.
K
h
a
n
,
a
n
d
S
.
A
z
a
m,
“
E
x
p
l
o
r
i
n
g
u
n
s
u
p
e
r
v
i
se
d
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
a
l
g
o
r
i
t
h
ms:
t
a
c
k
l
i
n
g
h
i
g
h
d
i
me
n
s
i
o
n
a
l
i
t
y
i
n
s
mal
l
d
a
t
a
s
e
t
s,
”
S
c
i
e
n
t
i
f
i
c
R
e
p
o
r
t
s
,
v
o
l
.
1
5
,
n
o
.
1
,
p
.
2
1
9
7
3
,
2
0
2
5
,
d
o
i
:
1
0
.
1
0
3
8
/
s4
1
5
9
8
-
0
2
5
-
0
7
7
2
5
-
9.
Evaluation Warning : The document was created with Spire.PDF for Python.