I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
38
,
No
.
2
,
Ma
y
20
25
,
p
p
.
1
086
~1
0
9
7
I
SS
N:
2
502
-
4
7
52
,
DOI
:
1
0
.
1
1
5
9
1
/ijee
cs
.v
38.
i
2
.
p
p
1
0
8
6
-
1
0
9
7
1086
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs
.
ia
esco
r
e.
co
m
A hybrid l
ea
rning
mo
del t
o
d
etec
t
c
a
rdio
v
a
scula
r disea
se from
elect
ro
ca
rdio
g
ra
m
G.
V.
Ra
j
y
a
L
a
k
s
hm
i
1
,
S.
K
r
is
hn
a
Ra
o
2
,
K.
Venk
a
t
a
Ra
o
1
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
S
y
st
e
ms E
n
g
i
n
e
e
r
i
n
g
,
A
n
d
h
r
a
U
n
i
v
e
r
si
t
y
,
V
i
sa
k
h
a
p
a
t
n
a
m,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
S
i
r
C
.
R
.
R
e
d
d
y
C
o
l
l
e
g
e
o
f
En
g
i
n
e
e
r
i
n
g
,
El
u
r
u
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ap
r
1
1
,
2
0
2
4
R
ev
is
ed
No
v
5
,
2
0
2
4
Acc
ep
ted
No
v
11
,
2
0
2
4
Ca
rd
io
v
a
sc
u
lar
d
ise
a
se
s
(CVD
s)
c
o
n
ti
n
u
e
to
b
e
th
e
wo
rld
’
s
m
o
st
s
ig
n
ifi
c
a
n
t
c
a
u
se
o
f
m
o
r
b
i
d
it
y
a
n
d
m
o
r
talit
y
.
Th
is
p
a
p
e
r
in
t
ro
d
u
c
e
s
a
u
n
i
q
u
e
h
y
b
ri
d
lea
rn
in
g
m
o
d
e
l
fo
r
CVD
d
e
tec
ti
o
n
u
sin
g
a
d
v
a
n
c
e
d
d
e
e
p
lea
rn
in
g
(DL
)
m
e
th
o
d
s.
T
h
e
p
ro
p
o
se
d
m
e
th
o
d
c
o
m
b
in
e
s
th
e
p
o
ten
t
fe
a
tu
re
e
x
trac
ti
o
n
p
o
we
rs
o
f
t
h
e
Eff
icie
n
tNe
t
p
re
-
tra
in
e
d
m
o
d
e
l
with
a
tt
e
n
ti
o
n
m
e
c
h
a
n
ism
s
a
n
d
g
ra
p
h
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
tw
o
rk
s
(G
CNs
)
fo
r
imp
ro
v
e
d
p
e
rfo
rm
a
n
c
e
.
F
irst,
ric
h
re
p
re
se
n
tatio
n
s
fro
m
c
a
rd
io
v
a
sc
u
lar
e
lec
tro
c
a
rd
io
g
ra
m
(
ECG
)
d
a
ta
e
x
trac
t
u
sin
g
t
h
e
Eff
icie
n
tNe
t
a
rc
h
i
tec
tu
re
a
s
a
fe
a
tu
re
e
x
trac
to
r
.
Us
in
g
a
larg
e
d
a
tas
e
t
o
f
c
a
rd
i
o
v
a
sc
u
lar
ECG
ima
g
e
s,
y
o
u
c
a
n
fin
e
-
t
u
n
e
th
e
p
re
-
train
e
d
Eff
icie
n
tNe
t
m
o
d
e
l
with
P
i
p
e
li
n
e
to
m
a
k
e
it
m
o
re
su
it
a
b
le
f
o
r
d
ise
a
se
id
e
n
ti
fica
ti
o
n
.
I
n
c
lu
d
in
g
a
tt
e
n
ti
o
n
tec
h
n
iq
u
e
s
th
a
t
a
ll
o
w
th
e
n
e
two
r
k
to
f
o
c
u
s
o
n
in
f
o
rm
a
ti
v
e
re
g
i
o
n
s
wi
th
i
n
th
e
in
p
u
t
,
ECG
ima
g
e
s
e
n
h
a
n
c
e
d
th
e
m
o
d
e
l
’
s
d
isc
rimin
a
ti
v
e
c
a
p
a
c
it
y
.
T
h
e
m
o
d
e
l
c
a
n
a
tt
e
n
d
to
th
e
sa
li
e
n
t
a
re
a
s
se
lec
ti
v
e
ly
li
n
k
e
d
wit
h
CVD
p
a
t
h
p
h
y
sio
l
o
g
y
t
h
ro
u
g
h
d
y
n
a
m
ic
a
tt
e
n
t
io
n
p
r
o
c
e
ss
e
s.
M
o
re
a
c
c
u
ra
te
p
re
d
ictio
n
s
re
su
lt
fro
m
th
is
a
tt
e
n
ti
o
n
-
b
a
se
d
r
e
fin
in
g
,
stre
n
g
th
e
n
in
g
t
h
e
m
o
d
e
l
’
s
a
b
il
it
y
to
id
e
n
t
ify
sig
n
ifi
c
a
n
t
p
a
tt
e
rn
s
s
u
g
g
e
sti
v
e
o
f
c
a
rd
i
o
v
a
sc
u
lar
p
r
o
b
lem
s.
G
CN
a
ims
to
li
n
k
t
h
e
n
a
t
u
ra
l
s
tru
c
tu
re
i
n
c
a
rd
io
v
a
sc
u
lar
d
a
ta.
It
c
a
n
e
ffi
c
ien
tl
y
c
a
p
t
u
re
c
o
m
p
lex
in
tera
c
ti
o
n
s
a
n
d
d
e
p
e
n
d
e
n
c
ies
a
m
o
n
g
v
a
rio
u
s
d
a
ta
p
iec
e
s
b
y
e
x
p
re
ss
in
g
m
e
d
ica
l
d
a
ta
a
s
g
ra
p
h
s,
wh
e
re
n
o
d
e
s
c
o
rre
sp
o
n
d
to
ima
g
e
re
g
io
n
s,
a
n
d
e
d
g
e
s
imp
ly
sp
a
ti
a
l
c
o
n
n
e
c
ti
o
n
s.
C
o
m
b
i
n
in
g
G
CN
in
t
o
t
h
e
p
r
o
p
o
se
d
h
y
b
rid
lea
rn
in
g
a
r
c
h
it
e
c
tu
re
fa
c
il
it
a
tes
e
x
trac
ti
n
g
c
o
n
tex
tu
a
l
i
n
fo
rm
a
ti
o
n
fr
o
m
l
o
c
a
l
a
n
d
g
lo
b
a
l
so
u
rc
e
s,
a
u
g
m
e
n
ti
n
g
t
h
e
m
o
d
e
l
’
s a
c
c
u
ra
c
y
.
K
ey
w
o
r
d
s
:
C
ar
d
io
v
ascu
lar
d
is
ea
s
es
Dee
p
lear
n
in
g
E
f
f
icien
tNet
Gr
ap
h
co
n
v
o
lu
tio
n
al
n
etwo
r
k
s
T
r
an
s
f
er
lear
n
i
n
g
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
:
G.
V.
R
ajy
a
L
ak
s
h
m
i
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
an
d
Sy
s
tem
s
E
n
g
i
n
ee
r
in
g
,
An
d
h
r
a
U
n
iv
er
s
ity
Vis
ak
h
ap
atn
am
,
I
n
d
ia
E
m
ail
:
g
v
r
lax
m
i@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
An
elec
tr
o
ca
r
d
io
g
r
am
(
E
C
G)
is
a
s
ig
n
al
-
b
ased
m
eth
o
d
th
at
d
etec
ts
h
ea
r
t
d
is
ea
s
e
in
its
ea
r
ly
s
tag
es.
E
C
G
p
lay
s
a
m
ajo
r
r
o
le
in
ca
r
d
io
v
ascu
lar
d
is
ea
s
e
(
C
VD
)
an
d
p
r
ed
icts
th
e
ab
n
o
r
m
alities
p
r
esen
t
in
th
e
p
atien
t
’
s
E
C
G
s
am
p
le.
No
wad
ay
s
,
s
u
d
d
e
n
h
ea
r
t
attac
k
s
ca
u
s
e
s
u
d
d
en
d
ea
th
s
,
s
o
it
is
v
e
r
y
im
p
o
r
tan
t
to
f
in
d
ab
n
o
r
m
alities
in
th
e
h
u
m
a
n
h
ea
r
t
b
ased
o
n
elec
tr
ical
ac
tiv
ity
u
s
in
g
elec
tr
o
ca
r
d
io
g
r
ap
h
y
[
1
]
.
C
o
n
v
e
r
tin
g
th
e
E
C
G
s
ig
n
als
to
im
ag
es
b
ec
o
m
es
m
o
r
e
co
m
p
le
x
to
f
i
n
d
th
e
C
VD
in
th
e
ea
r
ly
s
tag
es.
T
h
er
e
ar
e
m
an
y
r
ea
s
o
n
s
f
o
r
h
ea
r
t
f
ailu
r
e,
s
u
ch
as
co
r
o
n
ar
y
ar
te
r
y
d
is
ea
s
e,
b
lo
o
d
v
ess
el
d
is
o
r
d
er
s
,
an
d
ar
r
h
y
th
m
ias.
E
ar
ly
d
etec
tio
n
a
n
d
d
iag
n
o
s
is
o
f
C
VD
is
a
v
e
r
y
cr
u
cial
task
b
ec
a
u
s
e
o
f
u
n
p
r
ed
ic
tab
le
h
ea
r
t
attac
k
s
.
T
h
e
E
C
G
i
m
ag
es
ca
p
tu
r
e
th
e
elec
tr
ical
ac
tiv
ities
p
r
o
d
u
ce
d
b
y
h
ea
r
tb
ea
ts
[
2
]
.
E
x
p
er
ts
ca
n
d
etec
t
h
ea
r
t
ab
n
o
r
m
ali
ties
f
r
o
m
th
e
in
p
u
t
E
C
G
s
ig
n
als
b
ased
o
n
h
ea
r
tb
ea
ts
,
v
ein
b
lo
ck
ag
es,
an
d
ca
r
d
iac
d
is
o
r
d
er
s
[
3
]
.
Acc
o
r
d
i
n
g
to
th
e
W
o
r
ld
Hea
lth
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
h
yb
r
id
lea
r
n
in
g
mo
d
el
to
d
e
tect
ca
r
d
io
va
s
cu
la
r
d
is
ea
s
e
fr
o
m
elec
tr
o
ca
r
d
io
g
r
a
m
(
G.
V
.
R
a
jya
La
ksh
mi
)
1087
Or
g
an
izatio
n
(
W
HO)
,
m
an
y
p
eo
p
le
s
u
f
f
er
f
r
o
m
u
n
p
r
e
d
ictab
le
C
VD.
T
h
er
e
ar
e
m
an
y
ty
p
es
o
f
d
iag
n
o
s
is
s
y
s
tem
s
av
ailab
le
f
o
r
d
etec
tin
g
h
ea
r
t
d
is
o
r
d
e
r
s
.
Am
o
n
g
th
es
e,
m
ed
ical
im
ag
e
an
aly
s
is
p
lay
s
a
s
ig
n
if
ican
t
r
o
le.
I
n
m
e
d
ical
im
ag
e
an
al
y
s
is
,
E
C
G
in
ter
p
r
etatio
n
p
lay
s
a
v
ital
r
o
le
in
f
in
d
in
g
a
b
n
o
r
m
alities
in
th
e
h
u
m
an
h
ea
r
t.
C
o
m
b
in
in
g
im
ag
e
p
r
o
ce
s
s
in
g
te
ch
n
iq
u
es
with
m
ac
h
in
e
lear
n
in
g
(
ML
)
o
r
DL
m
o
d
els
im
p
r
o
v
ed
th
e
d
etec
tio
n
an
d
class
if
icatio
n
o
f
h
ea
r
t
d
is
ea
s
es.
T
h
e
p
r
ep
r
o
ce
s
s
in
g
m
et
h
o
d
s
ar
e
u
s
ed
to
r
em
o
v
e
th
e
n
o
is
e
f
r
o
m
th
e
i
n
p
u
t
im
ag
es
an
d
p
r
o
v
id
e
a
p
r
e
cise
im
ag
e
an
al
y
s
is
,
wh
ich
h
elp
s
g
iv
e
t
h
e
s
tatu
s
o
f
th
e
E
C
G
im
ag
es.
T
h
e
p
r
e
p
r
o
ce
s
s
in
g
s
tep
im
p
r
o
v
es
th
e
q
u
ality
o
f
th
e
im
a
g
es
an
d
s
h
o
ws
th
e
ac
c
u
r
ate
n
o
r
m
al
an
d
ab
n
o
r
m
al
s
p
ik
es
in
th
e
g
iv
en
E
C
G
im
ag
es.
T
h
e
E
C
G
s
ig
n
al
s
am
p
les
a
r
e
co
n
v
er
ted
f
r
o
m
s
ig
n
als
to
E
C
G
im
ag
es
to
f
in
d
th
e
ac
cu
r
ate
ab
n
o
r
m
al
r
eg
i
o
n
s
in
t
h
e
in
p
u
t
im
ag
es
[
4
]
,
[
5
]
.
Fig
u
r
e
1
s
h
o
ws
th
e
s
am
p
le
E
C
G
s
ig
n
al
im
ag
e.
I
n
r
ec
e
n
t
y
ea
r
s
,
th
e
d
e
v
elo
p
m
en
t
o
f
DL
,
a
b
r
an
c
h
o
f
M
L
an
d
ar
tific
ial
in
tellig
en
ce
(
AI
)
,
h
as
r
aised
f
r
esh
h
o
p
es f
o
r
tr
a
n
s
f
o
r
m
in
g
h
ea
lth
ca
r
e
p
r
o
ce
d
u
r
es,
esp
ec
ially
in
C
VD
id
en
tific
atio
n
.
Dee
p
lear
n
in
g
(
DL
)
m
eth
o
d
s
,
m
o
d
elled
af
ter
th
e
h
u
m
a
n
b
r
ain
’
s
n
eu
r
al
n
etwo
r
k
s
,
c
an
a
n
aly
ze
lar
g
e
v
o
lu
m
es
o
f
d
at
a
v
er
y
q
u
ick
ly
a
n
d
ac
cu
r
ately
,
p
r
o
v
id
in
g
p
r
e
v
io
u
s
ly
u
n
h
ea
r
d
-
o
f
in
s
ig
h
ts
in
t
o
i
n
tr
icate
m
ed
ical
is
s
u
es
[
5
]
.
T
h
e
p
r
im
a
r
y
f
o
cu
s
o
f
th
is
p
ap
er
was
th
e
d
ev
elo
p
m
en
t
o
f
E
C
G
im
ag
e
-
b
ased
ca
r
d
io
v
ascu
lar
illn
e
s
s
d
etec
tio
n
an
d
d
ia
g
n
o
s
is
.
R
esear
ch
er
s
an
d
m
e
d
ical
p
r
ac
titi
o
n
er
s
ca
n
in
v
esti
g
ate
n
o
v
el
m
eth
o
d
s
f
o
r
ea
r
ly
d
iag
n
o
s
is
,
r
i
s
k
ass
ess
m
en
t,
an
d
in
d
iv
id
u
alize
d
th
er
a
p
y
p
lan
s
f
o
r
p
atien
ts
with
C
VDs
b
y
u
tili
zin
g
DL
alg
o
r
ith
m
s
[
6
]
,
[
7
]
.
A
n
e
x
ten
s
iv
e
an
aly
s
is
o
f
cu
r
r
en
t
d
ev
elo
p
m
en
ts
,
o
b
s
tacle
s
,
an
d
p
o
ten
tial
o
p
p
o
r
tu
n
ities
,
th
is
s
tu
d
y
s
ee
k
s
to
clar
if
y
th
e
r
ev
o
lu
tio
n
a
r
y
p
o
ten
tial o
f
DL
in
ad
d
r
ess
in
g
th
e
w
o
r
ld
wid
e
b
u
r
d
en
o
f
ca
r
d
io
v
ascu
lar
illn
ess
es
[
8
]
.
Fig
u
r
e
1
.
E
C
G
s
am
p
le
im
ag
e
f
r
o
m
d
ataset
T
h
o
u
g
h
m
an
y
ap
p
r
o
ac
h
es
h
a
v
e
b
ee
n
d
ev
elo
p
e
d
f
o
r
C
VD
d
etec
tio
n
u
s
in
g
E
C
G
im
ag
es,
p
r
e
d
ictin
g
C
VD
ea
r
ly
o
n
b
ec
o
m
es
m
o
r
e
co
m
p
le
x
f
o
r
E
C
G
s
ig
n
als
as
s
h
o
wn
in
Fig
u
r
e
1
.
C
o
n
v
er
tin
g
th
e
E
C
G
s
ig
n
als
in
to
th
e
im
ag
e
b
ec
o
m
es
m
o
r
e
co
m
p
licated
wh
en
tr
an
s
latin
g
th
em
with
ac
cu
r
ate
p
atter
n
s
.
I
d
en
tify
in
g
s
lig
h
t
an
o
m
alies
in
th
e
wav
ef
o
r
m
th
at
m
ay
o
cc
u
r
f
r
o
m
m
ild
v
ar
iat
io
n
s
is
ted
io
u
s
.
E
C
G
s
ig
n
als
ar
e
ea
s
ily
em
b
ed
d
ed
with
n
o
is
e,
s
u
ch
as
b
aseli
n
e
wan
d
e
r
,
p
o
wer
lin
e
in
ter
f
er
en
ce
,
an
d
m
o
tio
n
ar
tifa
c
ts
,
wh
ich
im
p
o
s
e
im
p
er
ce
p
tib
le
in
f
lu
e
n
ce
o
n
t
h
e
E
C
G
im
ag
e
q
u
ality
th
at,
in
tu
r
n
,
af
f
ec
ts
d
etec
tio
n
alg
o
r
ith
m
ef
f
icien
c
y
.
Alth
o
u
g
h
th
e
ap
p
licatio
n
o
f
DL
,
esp
ec
ially
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
etwo
r
k
s
(
C
NNs
)
,
en
ab
les au
to
m
ated
f
ea
tu
r
e
ex
tr
ac
tio
n
,
it
is
o
p
en
to
d
eb
ate
wh
eth
er
h
an
d
c
r
af
ted
f
ea
tu
r
es
(
in
clu
d
in
g
R
-
R
in
ter
v
als
an
d
QR
S
co
m
p
lex
es)
o
r
m
o
r
e
d
ata
-
d
r
iv
en
o
n
es
p
r
o
v
id
e
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
.
T
h
e
f
in
d
in
g
is
th
at
b
alan
ce
is
s
tr
ict
with
E
C
G
s
ig
n
al
im
ag
es
.
T
o
o
v
er
c
o
m
e
th
is
,
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
,
g
r
ap
h
co
n
v
o
lu
tio
n
al
n
etwo
r
k
s
(
GC
Ns
)
,
was
in
tr
o
d
u
ce
d
to
b
alan
ce
th
e
is
s
u
es in
ex
is
tin
g
m
o
d
els.
Key
f
ac
to
r
s
o
f
t
h
is
w
o
r
k
−
Dete
ctin
g
C
VD
u
s
in
g
E
C
G
is
v
er
y
d
i
f
f
icu
lt b
ec
au
s
e
im
a
g
es m
ay
lo
s
e
q
u
ality
at
th
e
tim
e
o
f
co
n
v
er
s
io
n
.
−
E
ar
ly
p
r
e
d
ictio
n
o
f
C
VD
h
elp
s
th
e
ex
p
er
ts
to
r
ed
u
ce
t
h
e
d
ea
t
h
r
ate.
−
Ou
r
r
esear
ch
h
as
h
ar
n
ess
ed
th
e
p
o
wer
o
f
a
n
ef
f
ec
tiv
e
p
r
e
-
tr
ain
ed
m
o
d
el,
E
f
f
icie
n
tNet,
to
a
cc
u
r
ately
p
r
o
ce
s
s
th
e
in
tr
icate
p
atter
n
s
with
in
E
C
G
im
ag
es.
T
h
is
tec
h
n
o
lo
g
ical
a
d
v
an
ce
m
e
n
t
is
a
s
ig
n
if
ican
t
s
tep
f
o
r
war
d
i
n
o
u
r
q
u
est to
d
etec
t
C
VD
ea
r
ly
.
−
E
f
f
ec
tiv
e
p
r
e
-
p
r
o
ce
s
s
in
g
tech
n
iq
u
es
co
m
b
in
e
d
with
v
ar
io
u
s
n
o
is
e
f
ilter
s
,
s
u
ch
as
th
e
Sav
itzk
y
-
Go
lay
f
ilter
in
g
ap
p
r
o
ac
h
a
n
d
R
-
p
ea
k
s
in
E
C
G
im
ag
e
co
v
ar
ian
ce
p
r
ed
ictio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
38
,
No
.
2
,
Ma
y
20
25
:
1
0
8
6
-
1
0
9
7
1088
−
Ou
r
r
esear
ch
h
as
d
ev
elo
p
ed
a
r
o
b
u
s
t
Pip
elin
e
th
at
ef
f
icien
tl
y
tr
an
s
f
er
s
th
e
tr
ai
n
in
g
p
atter
n
s
to
th
e
test
in
g
m
o
d
el,
k
n
o
wn
as
th
e
GC
N
w
it
h
atten
tio
n
m
o
d
el
as
s
h
o
wn
i
n
Fig
u
r
e
2
.
T
h
is
p
r
o
ce
s
s
is
a
k
e
y
co
m
p
o
n
en
t
i
n
o
u
r
ea
r
l
y
p
r
e
d
ictio
n
o
f
C
VD.
−
T
h
e
atten
tio
n
m
ec
h
an
is
m
h
e
lp
s
im
p
r
o
v
e
t
h
e
ea
r
ly
p
r
ed
i
ctio
n
o
f
C
VD,
wh
ich
s
ig
n
i
f
ican
tly
im
p
ac
ts
o
u
tco
m
es.
T
h
e
o
r
g
an
izatio
n
o
f
wo
r
k
is
a
s
f
o
llo
ws:
s
ec
tio
n
2
liter
atu
r
e
s
u
r
v
ey
o
f
v
ar
io
u
s
e
x
is
tin
g
m
o
d
els
with
r
esear
ch
g
ap
s
an
d
its
p
er
f
o
r
m
an
ce
s
.
Sectio
n
3
d
is
cu
s
s
ed
ab
o
u
t
th
e
m
eth
o
d
o
l
o
g
y
o
f
th
is
w
o
r
k
b
y
ex
p
lain
in
g
th
e
p
r
e
-
tr
ain
ed
m
o
d
el,
p
r
e
-
p
r
o
ce
s
s
in
g
tech
n
i
q
u
es,
a
n
d
a
tten
tio
n
m
ec
h
an
is
m
s
.
Sectio
n
4
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
GC
N
with
atten
tio
n
m
o
d
el
.
Sectio
n
5
r
esu
lts
an
d
d
is
cu
s
s
io
n
s
ex
p
lain
ed
th
e
ex
is
tin
g
m
o
d
e
l
p
er
f
o
r
m
an
ce
s
,
p
r
e
-
tr
ain
ed
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
s
an
d
p
r
o
p
o
s
ed
m
o
d
el
’
s
p
er
f
o
r
m
a
n
ce
s
.
Sectio
n
6
co
n
clu
s
io
n
an
d
f
u
t
u
r
e
wo
r
k
.
Fig
u
r
e
2
.
S
y
s
tem
a
r
c
h
itectu
r
e
2.
L
I
T
E
R
AT
U
RE
SU
RVE
Y
L
u
et
a
l.
[
9
]
p
r
o
p
o
s
ed
a
n
ew
m
o
d
el
th
at
p
r
ed
icts
C
V
D
with
an
ac
cu
r
ate
d
is
ea
s
e
r
ate.
T
h
e
r
ec
o
n
s
tr
u
ctio
n
er
r
o
r
is
u
s
ed
f
o
r
th
e
d
ep
th
o
f
th
e
n
etwo
r
k
an
d
s
p
ec
if
ies
th
e
tr
ain
in
g
an
d
ad
v
an
ce
d
o
p
tim
iza
tio
n
is
co
m
b
in
ed
.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
m
ain
tain
s
s
tab
le
o
u
tco
m
es
f
o
r
th
e
two
d
atasets
.
Fo
r
t
h
e
f
ir
s
t
d
ataset,
th
e
ac
cu
r
ac
y
is
9
1
.
2
6
%
an
d
8
9
.
7
8
%,
r
esp
ec
tiv
ely
.
J
in
et
a
l.
[
1
0
]
p
r
o
p
o
s
ed
a
m
o
d
el
th
at
p
r
ed
icts
h
ea
r
t
f
ailu
r
e
b
ased
o
n
th
e
p
atien
ts
’
h
ea
lth
co
n
d
itio
n
s
.
I
n
t
h
is
wo
r
k
,
th
e
o
n
e
-
h
o
t
en
c
o
d
in
g
co
m
b
in
ed
with
wo
r
d
v
ec
to
r
s
is
u
s
ed
to
d
ia
g
n
o
s
e
h
ea
r
t
f
ailu
r
es
u
tili
zin
g
th
e
d
e
f
au
lt
f
ac
to
r
s
o
f
an
l
o
n
g
-
s
h
o
r
t
ter
m
m
o
d
el
(
L
STM
)
m
o
d
el.
R
esu
lts
s
h
o
w
th
at
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
is
h
ig
h
ly
ac
cu
r
ate
b
ased
o
n
th
e
h
ea
r
t
f
ailu
r
e
r
is
k
s
.
Hab
ib
et
a
l.
[
1
1
]
in
tr
o
d
u
ce
d
a
n
ew
m
o
d
el
th
at
p
r
ed
icts
h
ea
r
t
f
ailu
r
es
an
d
r
ec
o
m
m
en
d
s
ac
cu
r
ate
m
ed
icin
es
to
p
atien
ts
,
wh
ich
h
elp
s
in
m
ak
i
n
g
b
etter
d
ec
is
io
n
-
m
ak
i
n
g
.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
also
f
in
d
s
th
e
r
elatio
n
s
h
ip
s
am
o
n
g
s
ev
er
al
m
ed
ical
d
ata
an
d
an
aly
ze
s
h
e
ar
t
f
ailu
r
es
in
th
e
ea
r
ly
s
tag
es.
T
h
e
co
m
p
ar
is
o
n
b
etwe
en
v
a
r
io
u
s
m
o
d
els
s
h
o
ws
th
e
d
etec
tio
n
o
f
h
ea
r
t
f
ailu
r
es
with
m
ed
icin
e
r
ec
o
m
m
en
d
atio
n
s
.
Ho
s
s
ain
et
a
l.
[
1
2
]
d
is
cu
s
s
ed
v
ar
io
u
s
AI
-
b
ased
m
o
d
els
ap
p
lied
to
h
ea
r
t
d
is
ea
s
e
p
r
ed
ictio
n
.
T
h
e
attr
ib
u
tes
th
at
h
elp
th
em
f
in
d
th
e
ac
cu
r
ate
f
ea
tu
r
es
u
s
in
g
th
e
co
r
r
elatio
n
-
b
ased
f
ea
tu
r
e
s
u
b
s
et
s
elec
tio
n
te
ch
n
iq
u
e
with
b
est
f
ir
s
t
s
ea
r
c
h
.
Fro
m
th
e
c
o
m
p
a
r
is
o
n
,
it
is
id
en
tifie
d
th
at
th
e
p
r
o
p
o
s
ed
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
(
MLP
)
o
b
tain
s
9
0
.
1
2
%
ac
cu
r
ac
y
f
o
r
two
d
atasets
th
at
ar
e
u
s
ed
f
o
r
h
ea
r
t
d
is
ea
s
e
p
r
e
d
ictio
n
.
B
izo
p
o
u
lo
s
an
d
Ko
u
ts
o
u
r
is
[
1
3
]
d
is
cu
s
s
ed
s
ev
er
al
DL
m
o
d
els
th
at
d
etec
t
h
ea
r
t
d
is
ea
s
es
b
ased
o
n
th
e
p
atien
t
h
ea
lth
d
ata,
s
ig
n
als,
an
d
im
ag
e
ty
p
e
o
f
d
at
a.
DL
alg
o
r
ith
m
s
h
av
e
h
ig
h
a
cc
u
r
ac
y
in
d
eter
m
i
n
in
g
ca
r
d
io
lo
g
y
ab
n
o
r
m
alities
.
Kir
an
y
az
et
a
l.
[
1
4
]
p
r
esen
ted
th
e
f
ast
an
d
r
o
b
u
s
t
E
C
G
class
if
icatio
n
an
d
m
o
n
ito
r
in
g
s
y
s
tem
th
at
im
p
lem
en
ts
th
e
C
NN,
wh
ich
u
s
e
s
th
e
tw
o
s
ig
n
if
ican
t
m
ajo
r
b
lo
ck
s
u
s
in
g
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
class
if
icatio
n
o
f
E
C
G
s
ig
n
als.
Fin
ally
,
th
e
p
r
o
p
o
s
ed
m
o
d
el
o
b
tain
ed
b
etter
r
esu
lts
.
Z
h
an
g
et
a
l.
[
1
5
]
in
tr
o
d
u
ce
d
a
n
ew
m
o
d
el
th
at
u
s
es
s
ig
n
al
p
r
o
ce
s
s
in
g
to
ex
tr
ac
t
co
r
r
ec
t
f
ea
tu
r
es
f
r
o
m
in
p
u
t
d
ata.
W
e
ap
p
ly
a
u
n
iq
u
e
wav
elet
d
o
m
ai
n
m
u
ltire
s
o
lu
tio
n
-
b
ased
C
NN
to
ex
tr
a
ct
r
el
iab
le
f
ea
tu
r
es
f
r
o
m
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
h
yb
r
id
lea
r
n
in
g
mo
d
el
to
d
e
tect
ca
r
d
io
va
s
cu
la
r
d
is
ea
s
e
fr
o
m
elec
tr
o
ca
r
d
io
g
r
a
m
(
G.
V
.
R
a
jya
La
ksh
mi
)
1089
th
e
in
p
u
t
s
am
p
le.
T
h
e
s
eg
m
en
tatio
n
m
o
d
el
is
a
p
p
lied
to
t
h
e
o
u
tco
m
e
o
f
f
ea
tu
r
e
e
x
tr
ac
tio
n
o
u
tp
u
t
to
s
eg
m
en
t
th
e
m
o
d
el
with
b
etter
s
eg
m
e
n
tatio
n
tech
n
iq
u
es.
Fin
ally
,
t
h
e
1
-
D
-
C
NN
au
to
m
atica
lly
ex
tr
ac
ted
th
e
in
ter
n
al
h
ier
ar
c
h
ical
f
ea
tu
r
es
an
d
o
b
ta
in
ed
a
class
if
icatio
n
ac
cu
r
ac
y
o
f
9
4
.
3
%,
wh
ich
is
h
ig
h
c
o
m
p
ar
ed
with
o
th
er
m
o
d
els.
Ali
et
a
l.
[
1
6
]
p
r
o
p
o
s
ed
th
e
s
tack
ed
-
b
ased
SVM
m
o
d
el
th
at
ef
f
ec
tiv
ely
p
r
ed
i
cts
h
ea
r
t
f
ailu
r
es.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
is
th
e
co
m
b
in
ed
m
o
d
el
th
at
in
teg
r
at
es
th
e
HGSA,
wh
ich
is
m
o
r
e
ca
p
ab
le
o
f
s
h
o
win
g
th
e
p
r
ac
tical
an
al
y
s
is
o
f
h
ea
r
t
f
ailu
r
e.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
o
b
tain
s
a
n
ac
cu
r
ac
y
o
f
9
2
.
3
4
%,
wh
ich
is
b
etter
co
m
p
ar
ed
with
th
e
o
th
er
s
ix
m
o
d
els.
Kh
an
et
a
l.
[
1
7
]
in
tr
o
d
u
ce
d
th
e
m
o
d
if
ied
d
ee
p
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
(
MD
C
NN
)
ap
p
r
o
ac
h
th
at
m
o
n
ito
r
s
h
ea
r
t
ab
n
o
r
m
alities
f
r
o
m
t
h
e
E
C
G
d
ataset
.
T
h
e
MD
C
NN
is
m
ain
ly
u
s
ed
to
class
if
y
th
e
d
at
a
b
elo
n
g
in
g
to
s
en
s
o
r
d
ata,
w
h
ich
is
n
o
r
m
al
an
d
a
b
n
o
r
m
al.
Fin
ally
,
MD
C
NN
’
s
ac
cu
r
ac
y
is
9
8
.
3
%,
wh
ich
is
b
etter
th
an
o
th
e
r
m
o
d
els.
Kh
an
et
a
l.
[
1
8
]
p
r
o
p
o
s
ed
t
h
e
in
teg
r
ated
elec
tr
o
n
ic
c
o
n
t
r
o
l
ce
n
tr
e
(
I
E
C
C
)
,
in
teg
r
ated
with
th
e
SHA
-
5
1
2
alg
o
r
ith
m
,
en
s
u
r
in
g
d
ata
in
teg
r
ity
.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
is
in
teg
r
ated
with
an
ad
v
an
ce
d
s
ec
r
et
k
ey
,
en
h
an
cin
g
th
e
s
y
s
tem
’
s
s
ec
u
r
ity
.
T
h
e
co
r
r
elatio
n
v
al
u
e
o
f
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
i
s
0
.
0
4
5
,
w
h
ich
is
n
ea
r
er
to
ze
r
o
;
th
is
r
e
p
r
esen
t
s
th
e
s
tr
en
g
th
o
f
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
.
T
h
e
p
r
o
p
o
s
ed
I
E
C
C
s
h
o
ws
a
h
ig
h
p
er
f
o
r
m
an
ce
co
m
p
ar
ed
with
R
SA
an
d
E
C
C
m
o
d
els.
I
s
h
a
q
et
a
l.
[
1
9
]
p
r
o
p
o
s
ed
a
m
o
d
el
th
at
f
o
cu
s
ed
o
n
f
in
d
in
g
t
h
e
r
ich
f
ea
t
u
r
es
th
at
ar
e
m
o
r
e
e
f
f
ec
tiv
e
f
o
r
class
if
icatio
n
.
T
h
e
co
m
b
in
e
d
ap
p
r
o
ac
h
was
u
s
ed
to
class
if
y
th
e
h
ea
r
t
s
am
p
les
an
d
o
b
tain
ed
a
n
ac
cu
r
ac
y
o
f
0
.
9
3
%.
Fit
r
iy
an
i
et
a
l.
[
2
0
]
p
r
o
p
o
s
ed
th
e
h
ea
t
p
u
m
p
d
esig
n
m
o
d
el
(
HPDM)
m
o
d
e
l
th
at
co
n
s
is
ts
o
f
Den
s
ity
-
b
ased
s
p
atial
clu
s
ter
in
g
o
f
ap
p
licatio
n
s
with
n
o
is
e
(
DB
SC
AN
)
,
wh
ich
r
em
o
v
es
th
e
n
o
is
e
r
e
g
ar
d
in
g
o
u
tlier
s
.
T
h
e
SMOT
E
-
E
NN
m
ain
ly
f
o
cu
s
ed
o
n
tr
ain
i
n
g
th
e
d
is
tr
ib
u
ted
d
ata
u
s
in
g
XGBo
o
s
t
to
p
r
ed
ict
h
ea
r
t
d
is
ea
s
e.
T
h
e
ex
p
er
im
en
tal
r
esu
lts
s
h
o
w
t
h
at
th
e
two
d
atasets
u
s
ed
f
o
r
e
x
p
er
im
e
n
tal
an
aly
s
is
h
av
e
an
ac
cu
r
ac
y
o
f
9
5
.
9
% a
n
d
9
8
.
4
%.
Do
r
n
ala
[
2
1
]
p
r
o
p
o
s
ed
m
u
lti
-
m
o
d
el
clo
u
d
s
er
v
ices
th
at
o
b
t
ain
ac
cu
r
ate
o
u
tc
o
m
es
o
n
h
ea
l
th
ca
r
e
d
ata
p
er
f
o
r
m
ed
in
th
e
clo
u
d
p
latf
o
r
m
.
I
n
th
is
co
n
tex
t,
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
is
ap
p
lied
to
clo
u
d
h
ea
lth
ca
r
e
d
ata.
W
aq
ar
et
a
l.
[
2
2
]
p
r
o
p
o
s
ed
t
h
e
SMOT
E
th
at
d
etec
ts
ca
r
d
iac
d
is
ea
s
es
u
s
in
g
p
atien
t
h
ea
lth
c
ar
e
d
ata.
T
h
e
co
s
t
-
ef
f
ec
tiv
e
ap
p
r
o
ac
h
p
r
e
d
icts
h
ea
r
t
d
is
ea
s
es
in
th
e
ea
r
ly
s
tag
es.
T
h
e
q
u
an
titativ
e
an
aly
s
is
s
h
o
ws
th
at
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
im
p
r
o
v
e
s
th
e
h
ig
h
ac
cu
r
ac
y
.
B
ad
er
-
El
-
Den
et
a
l.
[
2
3
]
in
tr
o
d
u
ce
d
th
e
en
s
em
b
le
class
if
icatio
n
m
o
d
el
th
at
f
i
n
d
s
h
ea
r
t
d
is
ea
s
es
b
ased
o
n
p
atie
n
t
d
ata.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
a
ch
is
th
e
b
iased
R
F
m
o
d
el
co
m
b
in
ed
with
KNN
to
f
in
d
th
e
m
alicio
u
s
in
f
o
r
m
at
io
n
f
r
o
m
th
e
d
ataset.
T
h
e
r
es
u
lts
s
h
o
w
th
at
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
is
h
ig
h
ly
a
c
cu
r
ate
co
m
p
a
r
ed
with
o
th
er
m
o
d
els.
R
ath
et
a
l.
[
6
]
p
r
o
p
o
s
ed
th
e
e
n
s
em
b
le
m
o
d
el,
wh
ich
is
a
co
m
b
in
atio
n
o
f
L
STM
an
d
GA
N
m
o
d
els.
T
h
ese
m
o
d
els
we
r
e
ap
p
lied
t
o
two
d
atasets
,
MI
T
-
B
I
H
an
d
PTB
-
E
C
G.
Fo
r
th
e
f
ir
s
t
d
ata
s
et,
th
e
ac
cu
r
ac
y
is
0
.
9
9
2
%,
an
d
f
o
r
th
e
s
ec
o
n
d
d
a
taset,
it
is
0
.
9
9
4
.
I
s
in
et
a
l.
[
2
4
]
d
is
cu
s
s
ed
v
ar
i
o
u
s
DL
alg
o
r
ith
m
s
th
at
d
etec
t
an
d
class
if
y
ca
r
d
iac
d
is
ea
s
es in
th
e
ea
r
ly
s
tag
es.
T
h
e
d
is
ea
s
e
d
ete
ctio
n
r
ate
is
9
2
.
8
%,
wh
ich
is
h
ig
h
.
B
ag
h
d
ad
i
et
a
l.
[
2
5
]
p
r
o
p
o
s
ed
th
e
a
d
v
an
ce
d
a
n
d
n
o
v
el
ap
p
r
o
ac
h
th
at
ef
f
ec
tiv
el
y
f
i
n
d
s
th
e
ac
cu
r
ate
h
ea
r
t
d
is
ea
s
es
b
y
o
b
tain
s
th
e
ac
cu
r
ac
y
o
f
9
0
.
3
4
%
an
d
F1
-
s
co
r
e
o
f
9
2
.
4
%.
Z
ian
i
et
a
l.
[
2
6
]
p
r
o
p
o
s
ed
u
s
in
g
FECG
to
d
etec
t
ab
n
o
r
m
al
f
et
al
E
C
G.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
co
m
b
i
n
es
C
NN
with
I
C
A,
SVD,
an
d
NM
F.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
s
h
o
ws
h
ig
h
p
e
r
f
o
r
m
an
ce
i
n
r
ea
l
-
tim
e
ap
p
licatio
n
s
.
Z
ian
i
et
a
l.
[
2
7
]
p
r
o
p
o
s
ed
a
n
o
v
el
ap
p
r
o
ac
h
th
at
s
o
lv
es
v
a
r
io
u
s
i
s
s
u
es
in
f
etal
E
C
G
f
in
d
in
g
s
.
Z
ian
i
et
a
l.
[
2
8
]
in
tr
o
d
u
ce
d
t
h
e
tim
e
-
s
ca
le
-
b
ased
ap
p
r
o
ac
h
th
at
co
m
b
i
n
ed
FECG
an
d
ME
C
G
to
f
in
d
th
e
SNR
an
d
FR
PD
A.
Z
ian
i
et
a
l.
[
2
9
]
p
r
o
p
o
s
ed
a
n
o
v
el
ap
p
r
o
ac
h
th
at
co
n
s
is
ts
o
f
SVD
an
d
I
C
A
th
at
im
p
r
o
v
es th
e
p
e
r
f
o
r
m
a
n
ce
.
3.
M
E
T
H
O
D
3
.
1
.
P
re
-
t
ra
ined m
o
del Eff
ic
ient
Net
C
NNs
with
th
e
E
f
f
icien
tNet
ar
ch
itectu
r
e
ar
e
well
-
k
n
o
wn
f
o
r
th
eir
ef
f
icac
y
a
n
d
ef
f
icien
cy
in
im
ag
e
ca
teg
o
r
izatio
n
a
p
p
licatio
n
s
.
I
ts
h
ig
h
p
er
f
o
r
m
an
ce
an
d
s
ca
lab
ilit
y
h
av
e
led
to
wid
esp
r
ea
d
ad
o
p
tio
n
in
v
ar
io
u
s
s
ec
to
r
s
.
B
a
s
ed
o
n
E
C
G
p
ictu
r
es
p
r
e
-
tr
ain
ed
f
o
r
ca
r
d
io
v
ascu
lar
d
is
o
r
d
er
s
,
E
f
f
ic
ien
tNet
m
ay
h
elp
with
au
to
m
ated
d
iag
n
o
s
is
an
d
r
is
k
ass
e
s
s
m
en
t.
Hea
r
t
f
ai
lu
r
e,
m
y
o
ca
r
d
ial
in
f
ar
ctio
n
,
an
d
ar
r
h
y
th
m
ias
ca
n
all
b
e
d
iag
n
o
s
ed
with
th
e
h
elp
o
f
E
C
G
im
ag
es,
wh
ich
p
r
o
v
id
es
r
eq
u
ir
ed
d
ata
r
eg
ar
d
in
g
elec
tr
ic
ac
tiv
ities
o
f
th
e
h
ea
r
t.
R
esear
ch
er
s
an
d
p
h
y
s
ician
s
c
an
p
r
o
f
it
f
r
o
m
th
e
f
ea
t
u
r
es
g
ai
n
ed
f
r
o
m
lar
g
e
-
s
ca
le
p
ictu
r
e
d
atasets
b
y
u
tili
zin
g
p
r
e
-
tr
ain
ed
m
o
d
els
lik
e
E
f
f
icien
tNet.
T
h
is
ca
n
ass
is
t
in
en
h
an
cin
g
th
e
ac
cu
r
ac
y
a
n
d
d
ep
en
d
ab
ilit
y
o
f
au
to
m
ated
E
C
G
in
ter
p
r
etatio
n
.
T
h
e
eq
u
ip
m
e
n
t
r
eq
u
ir
em
e
n
t
s
ar
e
co
n
s
tan
tly
r
is
in
g
b
ec
au
s
e
o
f
th
e
in
cr
ea
s
in
g
r
eso
lu
tio
n
o
f
th
e
in
p
u
t
im
ag
e
.
I
n
th
is
in
s
tan
ce
,
th
e
E
f
f
icien
tNet
-
B
0
n
etwo
r
k
was
ch
o
s
en
a
s
th
e
clas
s
if
icatio
n
m
o
d
el
b
ased
o
n
th
e
f
ea
tu
r
es
f
o
u
n
d
i
n
th
e
2
D
im
ag
es
o
f
th
e
ca
r
d
iac
s
lices
an
d
th
e
h
ar
d
w
ar
e
ca
p
ab
ilit
ies
o
f
th
e
av
ailab
le
ap
p
ar
atu
s
.
A
n
etwo
r
k
in
p
u
t
im
ag
e
r
eso
lu
tio
n
o
f
2
2
4
×
2
2
4
is
n
ee
d
ed
f
o
r
E
f
f
icien
tNet
-
B
0
.
T
h
e
r
eq
u
ir
em
e
n
ts
ar
e
s
atis
f
ied
w
h
en
th
e
wav
ef
o
r
m
im
ag
e
is
co
n
v
er
ted
to
an
im
ag
e
wit
h
a
r
eso
lu
tio
n
o
f
2
2
4
×2
2
4
.
T
h
e
p
r
im
a
r
y
o
b
jec
tiv
e
o
f
th
e
E
f
f
icien
tNet
was
a
co
m
p
o
u
n
d
s
ca
lin
g
tech
n
i
q
u
e
th
at
s
ca
les
th
e
n
etwo
r
k
ep
t
h
,
wid
th
,
a
n
d
r
eso
lu
tio
n
f
o
r
th
e
E
C
G
im
ag
es
eq
u
ally
.
Fig
u
r
e
3
ex
p
lain
s
th
e
o
v
er
all
lay
er
s
p
r
esen
t
in
th
e
E
f
f
icien
tNet
m
o
d
el
th
at
tr
ain
s
o
n
f
in
d
in
g
th
e
p
atter
n
s
i
n
th
e
E
C
G
im
ag
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
38
,
No
.
2
,
Ma
y
20
25
:
1
0
8
6
-
1
0
9
7
1090
Fig
u
r
e
3
.
E
f
f
icien
tNE
T
ar
ch
ite
ctu
r
e
d
iag
r
a
m
De
p
th
:
d
=
α
∅
(
1
)
W
id
th
:
w
=
β
∅
(
2
)
R
e
s
ol
ution
:
r
=
γ
∅
(
3
)
d
is
th
e
n
etwo
r
k
d
ep
th
(
n
u
m
b
er
o
f
lay
er
s
)
.
w
is
th
e
wid
th
m
u
ltip
lier
.
r
is
th
e
r
eso
lu
tio
n
m
u
ltip
lier
.
∝
,
β
,
a
n
d
γ
ar
e
h
y
p
er
p
a
r
am
eter
s
th
at
co
n
tr
o
l h
o
w
m
u
ch
to
s
ca
le
th
e
d
ep
th
,
wid
th
,
an
d
r
eso
lu
tio
n
r
esp
e
ctiv
ely
.
∅
is
a
co
mp
o
u
n
d
co
efficien
t th
a
t c
o
n
tr
o
ls
o
ve
r
a
ll mo
d
el
s
iz
e.
3
.
2
.
P
re
-
pro
ce
s
s
ing
t
ec
hn
iq
ues
E
CG
im
a
g
es
3
.
2
.
1
.
Sa
v
it
z
k
y
-
g
o
la
y
f
ilte
ring
a
pp
ro
a
ch
E
C
G
s
ig
n
als
ar
e
o
n
e
o
f
th
e
b
asic
ca
r
d
io
lo
g
y
test
s
r
o
u
tin
ely
p
er
f
o
r
m
ed
to
d
etec
t
h
ea
r
t
p
r
o
b
lem
s
,
b
u
t
we
s
o
m
etim
es
f
ac
e
s
ev
er
al
ch
allen
g
es
d
u
e
to
n
o
is
e
an
d
ar
tifa
cts.
T
h
e
p
u
r
p
o
s
e
o
f
th
ese
m
eth
o
d
s
is
to
im
p
r
o
v
e
s
ig
n
al
q
u
alit
y
b
y
f
ilter
in
g
o
u
t
n
o
is
e,
ar
tifa
cts,
an
d
in
ter
f
er
en
ce
s
at
th
e
s
am
e
tim
e
th
at
it
r
etain
s
r
elev
an
t
f
ea
tu
r
es,
allo
win
g
f
o
r
co
r
r
ec
t
d
iag
n
o
s
is
o
r
an
aly
s
es.
E
C
G
im
ag
es
ar
e
u
s
u
ally
g
e
n
er
ated
f
r
o
m
E
C
G
s
ig
n
als
f
o
r
v
is
u
aliza
tio
n
an
d
an
aly
s
is
;
in
t
h
is
r
eg
ar
d
,
f
ilter
i
n
g
tech
n
iq
u
es
o
p
tim
ize
im
a
g
e
q
u
ality
,
en
h
a
n
cin
g
clar
ity
wh
ile
m
in
im
izin
g
d
is
to
r
tio
n
s
.
A
n
o
u
tlin
e
o
f
E
C
G
an
aly
s
is
to
d
iag
n
o
s
e
d
if
f
er
e
n
t
h
ea
r
t
d
is
ea
s
es
is
g
iv
en
in
s
ec
tio
n
3
b
ased
o
n
th
e
im
p
o
r
tan
ce
o
f
f
i
lter
in
g
tech
n
iq
u
es
as
a
ty
p
e
o
f
im
ag
e
p
r
o
ce
s
s
in
g
f
o
r
s
i
g
n
al
p
r
o
ce
s
s
in
g
th
at
h
as
b
ee
n
u
s
u
ally
ap
p
lied
in
p
r
e
v
io
u
s
y
ea
r
s
.
I
t
m
ay
ad
d
r
ess
th
e
is
s
u
es
o
f
n
o
is
e
an
d
ar
tifa
cts
in
E
C
G
s
ig
n
als
o
r
im
ag
es
th
at
ca
n
co
r
r
u
p
t
im
p
o
r
tan
t
in
f
o
r
m
atio
n
an
d
d
ec
r
ea
s
e
th
e
r
eliab
ilit
y
o
f
d
iag
n
o
s
tic
in
ter
p
r
etatio
n
s
.
T
h
er
e
m
ay
b
e
th
e
s
am
e
in
tr
o
d
u
ctio
n
o
f
p
r
e
p
r
o
ce
s
s
in
g
s
tep
s
lead
in
g
to
a
n
ec
ess
ity
f
o
r
ex
ten
s
iv
e
f
ilter
in
g
as
well
s
in
ce
th
ese
ar
e
p
r
er
e
q
u
is
ites
en
h
a
n
cin
g
th
e
r
eliab
ilit
y
o
f
E
C
G
-
b
ased
d
ia
g
n
o
s
tic
s
y
s
tem
s
.
T
h
e
f
ilter
in
g
tech
n
iq
u
es
in
E
C
G
im
ag
e
p
r
o
ce
s
s
in
g
p
av
e
th
e
way
to
war
d
s
ex
p
lain
in
g
d
if
f
er
en
t
m
eth
o
d
s
an
d
m
eth
o
d
o
l
o
g
ies
im
p
lem
en
ted
f
o
r
im
p
r
o
v
in
g
th
e
q
u
ality
o
f
E
C
G
im
ag
es
to
g
et
ac
cu
r
ate
an
aly
s
is
an
d
d
iag
n
o
s
is
.
I
t
em
p
h
asizes
th
e
n
ec
ess
ity
o
f
f
ilter
in
g
as
a
p
r
ep
r
o
ce
s
s
in
g
s
tep
an
d
s
ets
th
e
to
n
e
f
o
r
f
u
r
th
er
d
is
cu
s
s
io
n
o
n
d
if
f
er
en
t
ty
p
es
o
f
f
ilter
s
an
d
th
eir
ap
p
licatio
n
s
in
E
C
G
s
ig
n
al
p
r
o
ce
s
s
in
g
.
T
h
is
p
ap
er
ap
p
lied
t
h
e
Sav
itzk
y
-
Go
lay
Fil
ter
in
g
m
eth
o
d
to
elim
in
ate
n
o
is
e
with
in
E
C
G
im
ag
es.
A
v
al
u
ab
le
tech
n
iq
u
e
i
n
s
m
o
o
th
in
g
d
ata
is
f
itti
n
g
a
p
o
ly
n
o
m
ial
to
th
e
s
m
all
s
u
b
s
ets o
f
v
alu
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
h
yb
r
id
lea
r
n
in
g
mo
d
el
to
d
e
tect
ca
r
d
io
va
s
cu
la
r
d
is
ea
s
e
fr
o
m
elec
tr
o
ca
r
d
io
g
r
a
m
(
G.
V
.
R
a
jya
La
ksh
mi
)
1091
Th
e
Sav
itzk
y
-
Go
la
y
f
ilter
co
ef
f
icien
ts
ca
n
b
e
ca
lcu
lated
u
s
in
g
th
e
f
o
llo
win
g
eq
u
atio
n
s
:
Fo
r
ca
lcu
latin
g
th
e
s
m
o
o
th
in
g
co
ef
f
icien
ts
:
A
=
X
(
X
T
X
)
−
1
C
(
4
)
X
is
a
m
atr
ix
co
n
tain
in
g
th
e
p
o
wer
s
o
f
th
e
in
teg
er
s
eq
u
en
ce
f
r
o
m
-
(N
-
1
)
/2
to
(
N
-
1
)
/
2
,
wh
e
r
e
N
is
th
e
win
d
o
w
s
ize.
C
is
th
e
d
if
f
er
en
tiatio
n
m
atr
ix
,
wh
ich
d
ep
e
n
d
s
o
n
th
e
d
esire
d
d
er
iv
ativ
e
o
r
d
er
a
n
d
th
e
p
o
ly
n
o
m
ial
o
r
d
er
.
Fo
r
ca
lcu
latin
g
th
e
d
if
f
er
e
n
tiatio
n
co
ef
f
icien
ts
:
B
=
(
X
T
X
)
−
1
X
T
D
(
5
)
D
is
a
m
atr
ix
co
n
tain
in
g
th
e
p
o
wer
s
o
f
th
e
in
teg
er
s
eq
u
en
ce
f
r
o
m
-
(N
-
1
)
/2
to
(
N
-
1
)
/2
,
r
aised
to
th
e
d
esire
d
d
er
iv
ativ
e
o
r
d
er
.
On
ce
y
o
u
h
a
v
e
o
b
tai
n
ed
th
e
co
ef
f
icien
ts
(
A
f
o
r
s
m
o
o
th
in
g
an
d
B
f
o
r
d
if
f
er
en
tiatio
n
)
,
y
o
u
ca
n
p
e
r
f
o
r
m
th
e
co
n
v
o
l
u
tio
n
o
p
er
atio
n
u
s
in
g
th
ese
co
ef
f
icien
ts
an
d
th
e
i
n
p
u
t
s
ig
n
al
to
o
b
tain
th
e
f
ilter
ed
s
ig
n
al.
3
.
2
.
2
.
Det
ec
t
ing
R
-
pea
k
s
in E
CG
im
a
g
es
Dete
ctin
g
R
-
p
ea
k
s
in
E
C
G
s
ig
n
als
i
s
a
f
u
n
d
am
en
tal
task
in
b
io
m
ed
ical
s
ig
n
al
p
r
o
ce
s
s
in
g
,
p
ar
ticu
lar
ly
in
an
aly
zin
g
ca
r
d
iac
ac
tiv
ity
.
T
h
e
E
C
G
wav
ef
o
r
m
’
s
g
r
ea
test
p
ea
k
,
k
n
o
wn
as
t
h
e
R
-
p
ea
k
,
d
en
o
tes
th
e
d
ep
o
lar
izatio
n
o
f
th
e
h
ea
r
t
’
s
v
en
tr
icles.
Acc
u
r
ate
d
etec
t
io
n
o
f
R
-
p
ea
k
s
is
cr
u
cial
f
o
r
d
iag
n
o
s
in
g
v
ar
io
u
s
ca
r
d
iac
ab
n
o
r
m
alities
an
d
m
o
n
ito
r
in
g
h
ea
r
t
h
ea
lth
.
T
h
e
p
r
o
ce
s
s
o
f
R
-
p
ea
k
d
etec
tio
n
in
v
o
lv
es
an
aly
zin
g
th
e
E
C
G
s
ig
n
al
to
l
o
ca
te
th
e
p
r
o
m
in
en
t
p
ea
k
s
co
r
r
esp
o
n
d
in
g
t
o
th
e
R
-
wav
es.
T
h
is
is
ty
p
ica
lly
ac
h
iev
e
d
u
s
i
ng
s
ig
n
al
p
r
o
ce
s
s
in
g
tech
n
iq
u
es,
m
ath
em
atica
l
al
g
o
r
ith
m
s
,
an
d
m
ac
h
i
n
e
lear
n
in
g
m
eth
o
d
s
.
Var
io
u
s
alg
o
r
ith
m
s
h
av
e
b
ee
n
d
ev
el
o
p
ed
o
v
er
t
h
e
y
ea
r
s
t
o
a
u
to
m
ate
th
is
p
r
o
ce
s
s
,
r
an
g
in
g
f
r
o
m
s
im
p
le
th
r
esh
o
ld
-
b
ased
m
eth
o
d
s
to
m
o
r
e
s
o
p
h
is
ticated
ap
p
r
o
a
ch
es
in
v
o
lv
in
g
w
av
elet
tr
an
s
f
o
r
m
s
,
tem
p
late
m
atch
in
g
,
an
d
n
eu
r
al
n
etwo
r
k
s
.
T
h
e
im
p
o
r
tan
c
e
o
f
ac
cu
r
ate
R
-
p
ea
k
d
etec
tio
n
ca
n
n
o
t
b
e
o
v
e
r
s
tated
,
as
it
f
o
r
m
s
th
e
b
asis
f
o
r
m
an
y
s
u
b
s
eq
u
en
t
an
aly
s
es,
s
u
ch
as
h
ea
r
t
r
ate
v
ar
iab
ilit
y
an
aly
s
is
,
ar
r
h
y
t
h
m
ia
d
etec
tio
n
,
an
d
ass
ess
in
g
c
ar
d
iac
f
u
n
ctio
n
.
Mo
r
eo
v
er
,
with
th
e
ad
v
en
t
o
f
wea
r
ab
le
E
C
G
m
o
n
ito
r
in
g
d
ev
ices
an
d
telem
ed
icin
e,
au
to
m
ated
R
-
p
ea
k
d
etec
tio
n
alg
o
r
ith
m
s
p
lay
a
cr
u
cial
r
o
le
in
p
r
o
v
id
in
g
r
ea
l
-
tim
e
f
ee
d
b
ac
k
o
n
h
ea
r
t
h
ea
lth
an
d
f
ac
ilit
atin
g
r
em
o
te
p
atien
t m
o
n
ito
r
in
g
.
Sev
er
al
al
g
o
r
ith
m
s
h
a
v
e
b
ee
n
d
ev
el
o
p
ed
f
o
r
th
is
p
u
r
p
o
s
e,
a
n
d
m
a
n
y
o
f
th
em
ar
e
b
ased
o
n
s
p
ec
if
ic
m
ath
em
atica
l
eq
u
atio
n
s
o
r
s
ig
n
al
p
r
o
ce
s
s
in
g
tech
n
iq
u
es.
O
n
e
co
m
m
o
n
ly
u
s
ed
m
eth
o
d
is
th
e
Pan
-
T
o
m
p
k
in
s
alg
o
r
ith
m
,
wh
ich
in
v
o
lv
es
s
ev
er
al
s
tep
s
in
clu
d
in
g
b
an
d
p
a
s
s
f
ilter
in
g
,
d
if
f
er
en
tiatio
n
,
s
q
u
ar
in
g
,
in
teg
r
atio
n
,
an
d
th
r
esh
o
l
d
in
g
.
T
h
e
k
e
y
eq
u
atio
n
s
an
d
s
tep
s
in
v
o
lv
e
d
in
th
e
Pan
-
T
o
m
p
k
in
s
alg
o
r
ith
m
:
3
.
2
.
3
.
At
t
ent
io
n
m
ec
ha
nis
m
Hea
r
t
d
is
o
r
d
er
s
r
ep
r
esen
t
a
s
u
b
s
tan
tial
g
lo
b
al
ca
u
s
e
o
f
d
ea
th
,
an
d
ef
f
ec
tiv
e
ca
r
e
an
d
th
e
av
o
id
an
ce
o
f
u
n
f
av
o
u
r
ab
le
co
n
s
eq
u
en
ce
s
d
ep
en
d
h
ea
v
ily
o
n
ea
r
ly
d
etec
ti
o
n
.
E
C
G
im
ag
i
n
g
is
a
f
r
e
q
u
en
tly
u
s
ed
d
iag
n
o
s
tic
tech
n
iq
u
e
f
o
r
ev
alu
atin
g
h
ea
r
t
h
ea
lth
b
y
m
o
n
ito
r
in
g
th
e
elec
tr
ical
ac
tiv
ity
o
f
th
e
h
ea
r
t.
T
h
e
atten
tio
n
m
ec
h
an
is
m
,
in
s
p
ir
ed
b
y
h
o
w
p
eo
p
le
co
n
ce
n
tr
at
e
o
n
p
e
r
tin
e
n
t
in
f
o
r
m
atio
n
wh
e
n
p
r
o
ce
s
s
in
g
d
ata,
is
a
p
o
ten
t
s
tr
ateg
y
in
th
is
f
ield
.
B
y
s
el
ec
tiv
ely
f
o
cu
s
in
g
o
n
s
ig
n
if
ican
t
p
o
r
tio
n
s
o
f
th
e
in
p
u
t
d
a
ta
an
d
d
is
m
is
s
in
g
u
n
n
ec
ess
ar
y
in
f
o
r
m
atio
n
,
t
h
e
atten
tio
n
m
ec
h
a
n
is
m
en
a
b
les
th
e
m
o
d
el
to
p
er
f
o
r
m
m
o
r
e
ac
cu
r
at
e
an
d
p
r
ac
tical
an
aly
s
is
.
R
esear
ch
er
s
wan
t
t
o
im
p
r
o
v
e
p
atien
t
o
u
tco
m
es
b
y
im
p
r
o
v
in
g
ca
r
d
iac
d
is
ea
s
e
id
en
tific
atio
n
an
d
d
iag
n
o
s
is
b
y
a
p
p
ly
in
g
atten
tio
n
p
r
o
ce
s
s
es
to
E
C
G
im
ag
e
an
a
ly
s
is
.
T
h
is
r
esear
ch
d
escr
ib
es
th
e
u
s
e
o
f
atten
ti
o
n
m
ec
h
an
is
m
s
in
itiated
b
y
th
e
GC
N
to
d
iag
n
o
s
e
h
ea
r
t
d
is
ea
s
e
f
r
o
m
E
C
G
im
ag
es.
T
h
e
at
ten
tio
n
m
ec
h
an
is
m
h
elp
s
a
n
o
v
el
s
tr
ateg
y
f
o
r
u
s
in
g
atten
tio
n
p
r
o
ce
s
s
es
in
E
C
G
im
ag
e
p
r
o
ce
s
s
in
g
.
T
h
e
p
r
im
a
r
y
g
o
al
o
f
th
is
wo
r
k
is
to
cr
ea
te
a
u
to
m
ated
AI
tech
n
o
lo
g
ies
th
at
will
en
h
an
c
e
th
e
p
r
ec
is
io
n
an
d
e
f
f
icac
y
o
f
h
ea
r
t
d
is
ea
s
e
d
etec
tio
n
,
th
er
eb
y
h
elp
in
g
b
o
th
p
atien
ts
an
d
m
ed
ical
p
r
o
f
ess
io
n
als.
4.
P
RO
P
O
SE
D
M
E
T
H
O
DO
L
O
G
Y:
G
CN
WI
T
H
AT
T
E
N
SI
O
N
M
O
D
E
L
C
VDs
ar
e
a
m
ajo
r
g
lo
b
al
ca
u
s
e
o
f
d
ea
th
a
n
d
p
lace
a
h
e
av
y
s
tr
ain
o
n
h
ea
lth
ca
r
e
s
y
s
t
em
s
.
E
ar
ly
d
etec
tio
n
a
n
d
p
r
ec
is
e
d
ia
g
n
o
s
is
ar
e
ess
en
tial
f
o
r
C
VD
to
b
e
m
an
ag
e
d
a
n
d
tr
ea
ted
ef
f
ec
ti
v
ely
.
DL
m
eth
o
d
s
,
p
ar
ticu
lar
ly
GC
Ns,
to
an
aly
ze
E
C
G
d
ata
to
ass
is
t
in
d
iag
n
o
s
in
g
an
d
p
r
o
g
n
o
s
is
ca
r
d
io
v
ascu
lar
d
is
o
r
d
er
s
h
av
e
g
ain
ed
p
o
p
u
la
r
ity
in
r
ec
en
t
y
e
ar
s
.
B
ec
au
s
e
GC
Ns
ar
e
a
ty
p
e
o
f
n
eu
r
al
n
etwo
r
k
th
at
o
n
ly
wo
r
k
s
with
g
r
ap
h
-
s
tr
u
ctu
r
ed
d
ata,
th
ey
ar
e
well
-
s
u
ite
d
f
o
r
task
s
in
v
o
lv
in
g
r
elatio
n
s
h
ip
s
an
d
co
r
r
elatio
n
s
b
etwe
en
th
e
p
o
in
ts
in
th
e
d
ata
(
lik
e
r
e
p
licatin
g
co
n
d
u
ctiv
ities
as
s
ee
n
in
E
C
G
r
e
ad
in
g
s
)
.
I
n
th
is
p
ap
er
,
we
ai
m
to
in
v
esti
g
ate
th
e
ef
f
icac
y
o
f
GC
N
f
o
r
C
VD
b
a
s
ed
o
n
E
C
G
im
ag
es.
E
C
G
s
i
g
n
als
ca
n
b
e
r
ep
r
esen
ted
as
g
r
ap
h
s
;
in
o
u
r
s
tu
d
y
,
n
o
d
es
co
r
r
esp
o
n
d
to
t
h
e
d
ata
p
o
in
ts
o
f
a
n
E
C
G
s
ig
n
al,
a
n
d
ed
g
es
r
ep
r
esen
t
tem
p
o
r
al
d
e
p
en
d
en
cies
b
etwe
en
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
38
,
No
.
2
,
Ma
y
20
25
:
1
0
8
6
-
1
0
9
7
1092
th
ese
d
ata
p
o
in
ts
,
wh
ic
h
in
d
icate
s
th
at
GC
Ns
ar
e
s
u
itab
le
m
o
d
els
f
o
r
ex
tr
ac
tin
g
r
elev
a
n
t
f
ea
tu
r
es
ass
o
ciate
d
with
ca
r
d
iac
a
r
r
h
y
t
h
m
ia
d
ete
ctio
n
.
B
elo
w
ar
e
th
e
s
tep
s
t
o
d
etec
t
a
b
n
o
r
m
al
co
n
d
itio
n
s
o
n
a
g
i
v
en
s
et
o
f
im
ag
es.
Step
1
:
r
ep
r
esen
tatio
n
o
f
d
ata
−
Su
p
p
o
s
e
th
at
A
is
th
e
in
p
u
t
f
ea
tu
r
e
m
atr
ix
f
o
r
E
C
G
im
ag
es,
with
ea
ch
r
o
w
b
ein
g
a
s
a
m
p
le
an
d
ea
ch
co
lu
m
n
a
f
ea
tu
r
e.
−
Ass
u
m
e
th
at
th
e
ad
jace
n
cy
m
atr
ix
B
illu
s
tr
ate
s
th
e
lin
k
s
b
e
twee
n
v
ar
io
u
s
E
C
G
s
am
p
les.
I
t
ca
n
d
ep
ict
th
e
tem
p
o
r
al
co
r
r
elatio
n
s
b
etwe
en
E
C
G
s
ig
n
als in
th
is
s
ce
n
ar
io
.
Step
2
: G
C
N
l
ay
er
−
T
h
e
eq
u
atio
n
th
at
f
o
llo
ws
is
u
s
ed
to
ca
lc
u
late
th
e
o
u
tp
u
t
o
f
a
s
in
g
le
GC
N
lay
er
g
iv
en
t
h
e
in
p
u
t
f
ea
t
u
r
e
m
atr
ix
A
an
d
a
d
jace
n
cy
m
atr
i
x
B
:
A
(
l
+
1
)
=
σ
(
D
̂
−
1
2
X
̂
Y
̂
−
1
2
A
(
l
)
W
(
l
)
)
(
5
)
A
(
l
)
→
Featu
r
e
m
atr
ix
at
th
e
la
y
er
l.
W
(
l
)
→
W
eig
h
t m
atr
ix
o
f
GC
N
lay
er
.
A
̂
=
A
+
I
→
Ad
jace
n
cy
m
atr
ix
with
e
x
ten
d
ed
s
elf
-
co
n
n
ec
tio
n
.
D
̂
is
th
e
d
eg
r
ee
m
atr
ix
o
f
A
̂
.
r
ep
r
esen
ts
th
e
ac
tiv
atio
n
f
u
n
ct
io
n
s
u
ch
as R
eL
U.
Step
3
:
f
in
al
lay
er
a
n
d
p
r
ed
icti
o
n
I
n
th
is
wo
r
k
,
t
h
e
class
if
icatio
n
task
s
ca
n
b
e
p
er
f
o
r
m
ed
u
s
in
g
th
e
o
u
t
p
u
t
o
f
t
h
e
f
in
al
G
C
N
lay
er
.
A
s
ig
m
o
id
ac
tiv
atio
n
f
u
n
ctio
n
co
u
ld
b
e
u
s
ed
f
o
r
b
in
a
r
y
ca
teg
o
r
izatio
n
(
n
o
r
m
al
v
s
.
ab
n
o
r
m
a
l)
:
Y
=
σ
(
H
(
L
)
W
(
L
)
)
(
6
)
Y
th
e
p
r
ed
icted
o
u
tp
u
t,
H
(
L
)
is
th
e
o
u
tp
u
t
o
f
last
GC
N
lay
er
,
an
d
W
(
L
)
is
th
e
weig
h
t m
atr
ix
.
Step
4
:
lo
s
s
f
u
n
ctio
n
T
h
e
cr
o
s
s
-
en
tr
o
p
y
lo
s
s
f
u
n
ctio
n
is
u
s
ed
f
o
r
b
in
ar
y
class
if
icati
o
n
,
an
d
it is
r
ep
r
esen
ted
as:
ℒ
=
−
1
N
∑
[
y
i
l
og
(
y
̂
i
)
+
(
1
−
y
i
)
l
og
(
1
−
y
̂
i
)
]
N
i
=
1
(
7
)
wh
er
e
N
is
th
e
to
tal
s
am
p
les,
y
i
is
th
e
ac
tu
al
lab
el,
y
̂
i
is
th
e
p
r
ed
icted
lab
el.
Step
5
:
o
p
tim
izatio
n
Usi
n
g
g
r
ad
ien
t d
escen
t o
r
its
v
ar
iatio
n
s
,
lik
e
Ad
am
o
r
R
MSp
r
o
p
,
u
p
d
ate
th
e
GC
N
’
s
weig
h
ts
:
W
(
l
)
←
W
(
l
)
−
∝
∂
ℒ
∂
W
(
l
)
(
8
)
∝
R
ep
r
esen
t th
e
lear
n
in
g
r
ate.
Step
6
:
th
e
f
in
al
s
tep
I
n
th
e
f
in
al
s
tep
,
all
th
e
lay
er
s
ar
e
ag
g
r
eg
ate
d
in
to
f
in
al
lay
e
r
s
an
d
p
r
ed
ict
th
e
r
esu
lt
(
class
if
icatio
n
)
b
ased
o
n
p
atter
n
s
id
en
tifie
d
.
5.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
NS
T
h
e
ex
p
e
r
im
en
tal
an
al
y
s
is
is
m
ain
ly
f
o
c
u
s
ed
o
n
d
e
v
elo
p
in
g
th
e
p
r
o
p
o
s
ed
m
o
d
el
u
s
in
g
th
e
Py
th
o
n
lan
g
u
ag
e
with
ad
v
an
c
ed
lib
r
ar
ies.
T
h
e
E
C
G
im
ag
e
d
ataset
i
s
p
r
o
ce
s
s
ed
b
y
u
s
in
g
1
6
GB
R
A
M,
an
d
a
1
T
B
h
ar
d
d
r
iv
e
is
r
eq
u
ir
ed
.
I
n
th
is
s
ec
tio
n
,
th
e
co
m
p
ar
is
o
n
b
etwe
en
v
ar
io
u
s
DL
alg
o
r
ith
m
s
is
also
d
is
cu
s
s
ed
,
an
aly
zin
g
th
e
p
er
f
o
r
m
a
n
ce
in
ter
m
s
o
f
g
iv
en
p
ar
am
eter
s
.
I
n
th
is
s
ec
tio
n
,
th
e
co
m
p
ar
is
o
n
s
b
etwe
en
v
ar
io
u
s
alg
o
r
ith
m
s
th
at
ap
p
lied
o
n
E
C
G
s
ig
n
al
im
ag
es.
T
h
e
p
er
f
o
r
m
a
n
ce
m
etr
ic
s
s
h
o
ws th
e
s
tr
en
g
th
o
f
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
.
A
c
c
ura
c
y
(
ACC
)
=
TP
+
TN
TP
+
FP
+
TN
+
FN
Pr
e
c
ision
(
Pr
e
)
=
TP
TP
+
FP
Se
n
s
itivit
y
(
S
n
)
=
TP
TP
+
FN
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
h
yb
r
id
lea
r
n
in
g
mo
d
el
to
d
e
tect
ca
r
d
io
va
s
cu
la
r
d
is
ea
s
e
fr
o
m
elec
tr
o
ca
r
d
io
g
r
a
m
(
G.
V
.
R
a
jya
La
ksh
mi
)
1093
Sp
e
c
ifi
c
ity
(
S
p
)
=
TN
TN
+
FP
F1
−
Score
=
2
∗
(
Pr
ecis
i
o
n
∗
Recal
l
)
(
Pr
ecis
i
o
n
+
Recal
l
)
Fig
u
r
e
4
s
h
o
ws
th
e
co
u
n
t
v
alu
es
o
f
th
e
E
C
G
s
am
p
les
o
b
tain
ed
b
y
Naïv
e
B
ay
es
(
NB
)
.
B
ased
o
n
p
ast
k
n
o
wled
g
e
o
f
p
o
ten
tial
ev
en
t
-
r
elate
d
co
n
d
itio
n
s
,
NB
d
escr
i
b
es
th
e
p
r
o
b
a
b
ilit
y
o
f
an
ev
en
t.
I
t
ca
lcu
lates
ea
ch
class
’
s
lik
elih
o
o
d
b
ased
o
n
t
h
e
in
p
u
t
f
ea
tu
r
es
a
n
d
s
elec
ts
th
e
class
with
th
e
m
o
s
t
s
ig
n
if
ican
t
p
r
o
b
ab
ilit
y
.
Giv
en
th
e
class
lab
el,
th
e
NB
class
if
ier
ass
u
m
es
th
at
ev
er
y
f
ea
tu
r
e
is
in
d
ep
e
n
d
en
t
o
f
ev
er
y
o
th
er
f
ea
tu
r
e.
T
h
is
ass
u
m
p
tio
n
f
r
eq
u
e
n
tly
n
ee
d
s
to
b
e
co
r
r
ec
ted
in
r
ea
l
-
wo
r
ld
d
ata,
wh
ich
co
u
ld
r
esu
lt
in
a
lo
s
s
o
f
ac
cu
r
ac
y
.
Fin
ally
,
th
e
NB
o
b
tain
ed
th
e
lo
w
v
alu
es
f
o
r
p
r
ed
ictin
g
o
u
tc
o
m
es.
Fig
u
r
e
5
s
h
o
ws
th
e
p
er
f
o
r
m
an
ce
o
b
tain
ed
b
y
im
p
lem
en
tin
g
th
e
k
-
n
ea
r
s
t
n
eig
h
b
o
r
(
KNN
)
with
E
C
G
s
am
p
le
im
ag
es.
Her
e,
th
e
T
N
ac
h
iev
ed
th
e
h
ig
h
co
u
n
t
v
alu
es
o
f
4
1
1
an
d
T
P
ac
h
iev
ed
th
e
3
2
5
c
o
u
n
t
v
alu
es.
FP
an
d
FN
s
h
o
ws
th
e
lo
w
v
alu
es.
T
h
ese
co
u
n
t
v
alu
es
ar
e
b
ased
o
n
o
b
tain
ed
ac
tu
al
r
esu
lts
.
Fig
u
r
e
6
d
escr
ib
es
th
e
p
er
f
o
r
m
an
ce
o
f
SVM
in
ter
m
s
o
f
a
ctu
al
an
d
p
r
ed
icte
d
v
alu
es.
T
h
e
T
P,
FP
,
FN
s
h
o
ws
th
e
lo
w
v
a
lu
es
an
d
T
N
ac
h
iev
ed
th
e
h
i
g
h
co
u
n
t
v
alu
e
th
at
o
b
tain
s
th
e
h
ig
h
ac
cu
r
ac
y
.
Fig
u
r
e
4
.
C
o
u
n
t
v
alu
es o
b
tain
ed
b
y
u
s
in
g
NB
Fig
u
r
e
5
.
C
o
u
n
t
v
alu
es o
b
tain
ed
b
y
u
s
in
g
KNN
T
ab
le
1
in
itializes
th
e
p
er
f
o
r
m
an
ce
o
f
v
a
r
io
u
s
ML
alg
o
r
i
th
m
s
th
at
p
e
r
f
o
r
m
th
e
class
if
icatio
n
o
f
n
o
r
m
al
ab
n
o
r
m
al
im
ag
es.
T
h
e
ac
cu
r
ac
y
o
f
SVM
is
0
.
8
9
wh
ich
is
h
ig
h
co
m
p
ar
e
with
o
th
er
alg
o
r
ith
m
s
.
Fig
u
r
e
7
r
e
p
r
esen
ts
th
e
o
v
e
r
all
p
er
f
o
r
m
an
ce
o
f
ML
Alg
o
r
ith
m
s
.
T
ab
le
2
s
h
o
ws
th
e
co
m
p
ar
ati
v
e
p
er
f
o
r
m
a
n
ce
b
ased
o
n
th
e
p
r
e
-
tr
ai
n
ed
v
alu
es.
T
h
e
p
er
f
o
r
m
an
ce
o
f
E
f
f
icien
tNE
T
ac
h
iev
e
d
th
e
h
ig
h
v
alu
es
co
m
p
ar
e
with
o
th
er
e
x
is
tin
g
m
o
d
els
VGG1
6
an
d
R
E
SNET
.
Hig
h
p
er
f
o
r
m
an
ce
in
itializes
to
f
in
d
th
e
ac
cu
r
ate
p
atter
n
s
.
T
h
ese
tr
ai
n
in
g
p
atter
n
s
h
elp
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
an
d
in
cr
ea
s
e
th
e
p
er
f
o
r
m
an
ce
with
th
e
ac
cu
r
ac
y
o
f
0
.
9
8
%
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
38
,
No
.
2
,
Ma
y
20
25
:
1
0
8
6
-
1
0
9
7
1094
Fig
u
r
e
6
.
C
o
u
n
t
v
alu
es o
b
tain
ed
b
y
u
s
in
g
SVM
Fig
u
r
e
8
d
escr
ib
es
th
e
p
er
f
o
r
m
an
ce
o
f
s
tate
-
of
-
ar
t
alg
o
r
ith
m
s
wh
ich
s
h
o
ws
th
e
h
ig
h
i
m
p
ac
t
o
n
test
in
g
.
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
p
r
e
-
tr
ain
e
d
m
o
d
els
was
ap
p
lied
o
n
tr
ain
in
g
E
C
G
s
i
g
n
al
im
ag
es
d
ata.
T
h
e
E
f
f
icien
tNE
T
s
h
o
ws
th
e
h
ig
h
ac
c
u
r
ac
y
o
f
0
.
9
8
%
an
d
v
ice
v
er
s
a.
I
t
in
d
icate
s
th
e
o
v
er
all
d
etec
tio
n
r
ate
is
h
ig
h
co
m
p
ar
e
with
e
x
is
tin
g
m
o
d
els.
T
ab
le
1
.
L
is
t o
f
a
l
g
o
r
ith
m
s
th
a
t p
er
f
o
r
m
th
e
class
if
icatio
n
b
ased
o
n
g
iv
e
n
p
ar
am
eter
s
A
l
g
o
r
i
t
h
ms
A
c
c
Pre
Sn
Sp
F1
-
s
c
o
r
e
N
B
[
2
9
]
0
.
5
0
0
.
5
4
0
.
5
0
0
.
5
0
0
.
5
2
K
N
N
[
2
9
]
0
.
7
8
0
.
7
8
0
.
7
4
0
.
8
2
0
.
7
6
S
V
M
0
.
8
9
0
.
6
0
0
.
6
2
0
.
9
3
0
.
6
1
T
ab
le
2
.
L
is
t o
f
a
l
g
o
r
ith
m
s
th
a
t p
er
f
o
r
m
th
e
class
if
icatio
n
b
ased
o
n
g
iv
e
n
p
ar
am
eter
s
A
l
g
o
r
i
t
h
ms
A
c
c
Pre
Sn
Sp
F1
-
s
c
o
r
e
V
G
G
1
6
0
.
6
9
0
.
7
0
0
.
7
4
0
.
6
1
0
.
6
2
R
ESN
ET
0
.
7
5
0
.
7
7
0
.
7
7
0
.
7
8
0
.
8
6
Ef
f
i
c
i
e
n
t
N
ET
0
.
9
8
0
.
9
9
0
.
9
8
0
.
9
8
0
.
9
8
Fig
u
r
e
7
.
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
ML
alg
o
r
ith
m
s
o
b
tain
ed
f
r
o
m
t
h
e
co
u
n
t v
alu
e
s
o
f
co
n
f
u
s
io
n
m
atr
ix
Fig
u
r
e
8
.
Per
f
o
r
m
an
c
e
co
m
p
ar
is
o
n
s
b
etwe
en
p
re
-
tr
ain
ed
m
o
d
els
Fig
u
r
e
9
s
h
o
ws
th
e
co
u
n
t
v
alu
es
o
b
tain
ed
b
y
u
s
in
g
th
e
ANN
.
Her
e,
th
e
T
P
v
alu
es
ar
e
h
ig
h
with
3
8
1
a
n
d
L
o
w
v
alu
es
ar
e
FN,
FP
,
an
d
T
N.
I
f
th
e
T
P
’
s
o
b
tain
ed
h
ig
h
v
alu
es
th
en
th
e
m
o
d
el
ac
c
u
r
ately
id
en
tifie
s
a
s
izab
le
p
er
ce
n
tag
e
o
f
th
e
p
o
s
itiv
e
o
cc
u
r
r
e
n
ce
s
in
th
e
d
atase
t.
I
t
is
u
s
u
ally
a
d
esira
b
le
r
es
u
lt,
p
ar
ticu
lar
ly
i
n
s
itu
atio
n
s
lik
e
h
ea
r
t d
is
ea
s
e
d
iag
n
o
s
is
,
wh
en
ac
cu
r
ately
r
ec
o
g
n
izin
g
p
o
s
itiv
e
ca
s
es is
cr
itic
al.
Fig
u
r
e
1
0
s
h
o
ws
th
e
co
u
n
t
v
alu
es
o
f
C
NN
wh
ich
s
h
o
ws
h
ig
h
f
o
r
T
N
’
s
.
Hig
h
er
T
Ns
r
ep
r
esen
t
th
e
m
o
d
el
’
s
s
p
ec
if
icity
–
its
ca
p
a
city
to
ac
cu
r
ately
d
etec
t
n
eg
ativ
e
ca
s
es
am
o
n
g
all
ac
tu
al
n
eg
ativ
e
in
s
tan
ce
s
.
A
lo
w
f
alse
p
o
s
itiv
e
r
ate
is
s
h
o
wn
b
y
h
ig
h
s
p
ec
if
icity
,
wh
ich
m
ea
n
s
th
e
m
o
d
el
is
less
lik
ely
to
in
ter
p
r
et
n
eg
ativ
e
ca
s
es
as
p
o
s
itiv
e
i
n
co
r
r
ec
tly
.
Fig
u
r
e
1
1
s
h
o
ws
th
e
h
ig
h
T
P
’
s
co
m
p
ar
ed
with
o
th
er
ex
is
tin
g
alg
o
r
ith
m
s
.
T
h
e
p
r
o
p
o
s
ed
GC
N
o
b
tain
ed
th
e
ac
cu
r
ac
y
o
f
0
.
8
6
%
wh
ich
is
h
ig
h
co
m
p
a
r
e
with
ex
is
tin
g
m
o
d
els
b
u
t
it
is
lo
w
in
r
ea
l
tim
e
s
ce
n
ar
io
s
h
o
wn
in
T
a
b
le
3
.
T
h
e
T
P
’
s
o
f
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
o
b
tain
ed
th
e
h
ig
h
v
alu
es
th
en
o
th
er
FP
,
FN
,
an
d
T
N
.
Fig
u
r
e
1
2
s
h
o
ws
th
e
q
u
a
n
titativ
e
p
er
f
o
r
m
a
n
ce
o
f
s
ev
er
al
alg
o
r
ith
m
s
u
s
ed
in
th
is
p
ap
er
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
h
yb
r
id
lea
r
n
in
g
mo
d
el
to
d
e
tect
ca
r
d
io
va
s
cu
la
r
d
is
ea
s
e
fr
o
m
elec
tr
o
ca
r
d
io
g
r
a
m
(
G.
V
.
R
a
jya
La
ksh
mi
)
1095
Fig
u
r
e
9
.
C
o
u
n
t
v
alu
es
o
b
tain
ed
b
y
u
s
in
g
ANN
Fig
u
r
e
1
0
.
C
o
u
n
t
v
alu
es
o
b
tai
n
ed
b
y
u
s
in
g
C
NN
T
ab
le
3
.
L
is
t o
f
DL
a
l
g
o
r
ith
m
s
th
at
p
er
f
o
r
m
th
e
class
if
icatio
n
b
ased
o
n
g
iv
en
p
ar
am
eter
s
A
l
g
o
r
i
t
h
ms
A
c
c
Pre
Sn
Sp
F1
-
s
c
o
r
e
ANN
0
.
6
4
0
.
6
8
0
.
7
1
0
.
5
5
0
.
6
9
C
N
N
0
.
7
8
0
.
7
2
0
.
6
9
0
.
8
2
0
.
7
1
G
C
N
0
.
8
6
0
.
8
8
0
.
9
3
0
.
6
3
0
.
9
0
Fig
u
r
e
1
1
.
C
o
u
n
t
v
alu
es
o
b
tai
n
ed
b
y
u
s
in
g
GC
N
Fig
u
r
e
1
2
.
Per
f
o
r
m
a
n
ce
co
m
p
ar
is
o
n
s
b
etwe
en
DL
m
o
d
els
6.
CO
NCLU
SI
O
N
Dete
ctin
g
C
VDs
f
r
o
m
E
C
G
s
ig
n
als
is
cr
u
cial
f
o
r
ea
r
ly
d
ia
g
n
o
s
is
an
d
in
ter
v
e
n
tio
n
.
I
n
th
is
s
tu
d
y
,
a
h
y
b
r
id
lear
n
in
g
m
o
d
el
co
m
b
in
in
g
atten
tio
n
m
ec
h
an
is
m
s
an
d
GC
N
s
wa
s
p
r
o
p
o
s
ed
f
o
r
th
is
task
.
T
h
e
r
esu
lts
o
f
th
e
s
tu
d
y
d
em
o
n
s
tr
ate
th
e
ef
f
e
ctiv
en
ess
o
f
th
e
h
y
b
r
i
d
lear
n
in
g
m
o
d
el
in
ac
cu
r
ately
d
etec
tin
g
C
VDs f
r
o
m
E
C
G
s
ig
n
als.
T
h
e
m
o
d
el
ca
n
f
o
c
u
s
o
n
r
elev
an
t
p
a
r
ts
o
f
th
e
E
C
G
s
ig
n
al,
ca
p
tu
r
in
g
im
p
o
r
tan
t
p
a
tter
n
s
an
d
f
ea
tu
r
es
in
d
icativ
e
o
f
ca
r
d
io
v
ascu
lar
ab
n
o
r
m
alities
.
Ad
d
itio
n
ally
,
t
h
e
in
teg
r
atio
n
o
f
GC
Ns
en
ab
les
th
e
m
o
d
el
to
ca
p
tu
r
e
th
e
c
o
m
p
le
x
r
elatio
n
s
h
ip
s
an
d
d
ep
en
d
en
cies
am
o
n
g
d
if
f
er
e
n
t
s
eg
m
en
ts
o
f
th
e
E
C
G
s
ig
n
al,
en
h
an
ci
n
g
its
ab
ilit
y
to
ex
tr
ac
t
m
ea
n
in
g
f
u
l
in
f
o
r
m
atio
n
f
o
r
d
is
ea
s
e
d
etec
tio
n
.
Fin
ally
,
th
e
h
y
b
r
id
lear
n
in
g
m
o
d
el
lev
er
ag
in
g
atten
tio
n
m
ec
h
an
is
m
s
an
d
GC
Ns
p
r
esen
t
s
a
p
r
o
m
is
in
g
ap
p
r
o
ac
h
f
o
r
th
e
d
etec
tio
n
o
f
C
VDs
f
r
o
m
E
C
G
s
ig
n
als.
Fu
tu
r
e
r
esear
c
h
m
ay
f
o
c
u
s
o
n
f
u
r
th
er
r
e
f
in
i
n
g
th
e
m
o
d
el
a
r
ch
itectu
r
e,
e
x
p
lo
r
in
g
a
d
d
itio
n
al
d
atasets
,
an
d
co
n
d
u
ctin
g
clin
ic
al
v
alid
atio
n
s
tu
d
ies to
f
ac
ilit
a
te
its
in
teg
r
atio
n
in
to
r
o
u
tin
e
c
lin
ical
p
r
ac
tice.
Evaluation Warning : The document was created with Spire.PDF for Python.