Paroxysmal
atrial
fibrillation
(
PAF)
i
s
t
he
initial
phase
of
AF),
often
progressing
stealthily
to
the
chronic
stage
due
absence
noticeable
symptoms.
Hence,
timely
identification
PAF
is
pretty
necessary.
This
study
proposes
an
automated
machine
learning-based
detection
algorithm
utilizing
a
single-lead
electrocardiogram
signal.
A
total
25
features
are
extracted
from
1-minute
segments
and
optimal
feature
set,
selected
by
deploying
minimum
redundancy
maximum
relevance
algorithm,
used
train
decision
tree
(DT)
random
forest
(RF)
classifiers.
The
training
testing
stages
included
43
subjects,
subjectwise
10-fold
cross-validation
was
performed.
RF
outperforms
DT
classifier
chieving
91.94%
accuracy,
91.75%
sensitivity,
91.47%
F1
score.
higher
accuracy
using
shorter
ECG
remarks
significance
proposed
model
for
AF
monitoring.
IEEE Sensors Letters,
Journal Year:
2024,
Volume and Issue:
8(4), P. 1 - 4
Published: Feb. 29, 2024
Automatic
assessment
of
electrocardiogram
(ECG)
signal
quality
plays
a
vital
role
in
reducing
false
alarms
and
improving
the
trustworthiness
unobtrusive
health
monitoring
devices
under
noisy
ECG
recordings,
which
are
unavoidable
continuous
monitoring.
In
this
letter,
we
present
an
(ECG-SQA)
method
based
on
Fourier
magnitude
spectrum
as
input
to
1-D
convolutional
neural
network
(1-D
CNN)
with
optimal
hyperparameters
activation
function,
significantly
reduces
CNN
model
size
computational
load
resource-constrained
devices.
On
untrained
databases
including
single-lead
multilead
signals
having
different
kinds
P
waves,
QRS
complexes,
T
waves
(PQRST)
morphologies
various
noise
sources,
CNN-based
ECG-SQA
had
sensitivity
99.30%,
specificity
95.40%
for
three
convolution
layers,
dense
kernel
$3\times
1$
.
This
study
demonstrated
that
parameter
selection
can
reduce
resources
52%
852
kB
67697
parameters
compared
other
models.
Real-time
implementation
Raspberry
Pi
computing
shows
processing
time
is
124.4
notation="LaTeX">$\pm$
42.5
ms
checking
5
s
signal.
Frontiers in Physiology,
Journal Year:
2024,
Volume and Issue:
15
Published: April 9, 2024
Atrial
fibrillation
(AF)
is
the
most
common
cardiac
arrhythmia,
which
clinically
identified
with
irregular
and
rapid
heartbeat
rhythm.
AF
puts
a
patient
at
risk
of
forming
blood
clots,
can
eventually
lead
to
heart
failure,
stroke,
or
even
sudden
death.
Electrocardiography
(ECG),
involves
acquiring
bioelectrical
signals
from
body
surface
reflect
activity,
standard
procedure
for
detecting
AF.
However,
occurrence
often
intermittent,
costing
significant
amount
time
effort
medical
doctors
identify
episodes.
Moreover,
human
error
inevitable,
as
experienced
professionals
overlook
misinterpret
subtle
signs
As
such,
it
critical
importance
develop
an
advanced
analytical
model
that
automatically
interpret
ECG
provide
decision
support
diagnostics.
