A Novel Computer-Aided Approach for Predicting COVID-19 Severity Using Hyperparameters in ResNet50v2 from X-ray Images
International Journal of experimental research and review,
Год журнала:
2024,
Номер
42, С. 120 - 132
Опубликована: Авг. 30, 2024
This
research
has
been
globally
impacted
by
COVID-19
virus,
which
was
a
very
uncommon,
highly
contagious
&
dangerous
respiratory
illness
demanding
early
detection
for
effective
containment
and
further
spread.
In
this
research,
we
proposed
an
innovative
methodology
that
utilizes
images
of
X-rays
at
stage.
By
employing
convolution
neural
network,
enhance
the
accuracy
performance
via
using
ResNet50v2
hyperparameter.
The
achieves
remarkable
with
average
99.12%.
surpasses
other
available
models
based
on
different
deep
learning
like
VGG,
Xception
DenseNet
COVID
identification
help
X-ray
images.
scans
are
now
preferably
used
modality
COVID-19,
given
its
widespread
utilization
effectiveness.
However,
manual
treatment
examination
is
challenging,
specifically
in
field
facing
limitation
skilled
medical
staff.
Utilization
demonstrated
significant
potential
results
automating
diagnosis
timely
films.
suggested
architecture
developed
prediction
analysis
cases
It
firmly
believes
study
holds
alleviating
workload
frontline
radiologists,
expediting
patient
treatment,
facilitating
pandemic
control
efforts.
Язык: Английский
Automatic ECG Arrhythmia Recognition using ANN and CNN
International Journal of experimental research and review,
Год журнала:
2024,
Номер
45(Spl Vol), С. 01 - 14
Опубликована: Ноя. 30, 2024
Present
research
highlights
the
need
for
more
patient-oriented
monitoring
systems
cardiac
health,
especially
in
aftermath
of
COVID-19.
The
study
introduces
a
contactless
and
affordable
ECG
device
capable
recording
heart
arrhythmias
remote
monitoring,
which
is
vital
managing
rising
incidence
untimely
attacks.
Two
deep
learning
algorithms
have
been
developed
to
design
system:
RCANN
(Real-time
Compressed
Artificial
Neural
Network)
RCCNN
Convolutional
Network),
respectively,
based
on
ANN
CNN.
These
methods
are
designed
classify
analyze
three
different
forms
datasets:
raw,
filtere
filtered
+
compressed
signals.
were
this
identify
most
suitable
type
dataset
that
can
be
utilized
regular/remote
monitoring.
This
data
prepared
using
online
signals
from
Physionet(ONLINE)
real-time
Arduino
sensor
device.
Performance
analysed
basis
accuracy,
sensitivity,
specificity
F1
score
all
kinds
databases
both
RCANN.
For
raw
data,
accuracy
99.2%,
sensitivity
99.7%,
F1-Score
99.2%.
RCCNN,
93.2%,
91.5%,
95.1%,
93.5%
Filtered
Data,
97.7%,
95.9%,
99.4%,
97.6%.
90.5%,
85.8%,
96.4%,
90.9%
96.6%,
97.6%,
95.7%,
96.5%.
85.2%,
79.2%,
94.5%,
86.7%
performance
evaluation
shows
with
datasets
outperforms
other
approaches
telemonitoring
makes
it
promising
approach
individualized
health
management.
Язык: Английский
A Fusion Method for Detection and Classification of Diseases in Tomato Plants Using Swarm-based Deep Learning
International Journal of experimental research and review,
Год журнала:
2024,
Номер
45(Spl Vol), С. 135 - 152
Опубликована: Ноя. 30, 2024
Precise
identification
and
detection
of
ailments
in
tomato
plants
are
essential
for
preserving
crop
vitality
optimizing
agricultural
productivity.
This
promotes
the
use
methods
that
can
be
maintained
over
time
decreases
financial
losses
caused
by
plant
diseases.
Detecting
classifying
diseases
is
critical
ensuring
health
maximizing
Utilizing
advanced
computer
vision
techniques
this
purpose
enhances
precision
monitoring
health,
ultimately
leading
to
more
efficient
targeted
interventions.
research
work
presents
a
novel
framework
Tomato
Plant
Disease
Detection
Classification
(TPDDC)
using
fusion
swarm-based
deep-learning
techniques.
Our
approach
leverages
K-means
clustering
with
Grasshopper
Optimization
(GO)
segmenting
Regions
Interest
(ROI)
from
leaf
images,
followed
feature
extraction
optimization
Maximally
Stable
Extremal
(MSER)
GO.
The
optimized
features
then
classified
Convolutional
Neural
Network
(CNN).
proposed
TPDDC
model
was
evaluated
Village
Dataset,
encompassing
ten
different
Experimental
results
demonstrate
significant
improvements
classification
accuracy,
achieving
an
average
accuracy
97.6%
GO-based
compared
92.7%
without
These
underscore
effectiveness
integrating
deep
learning
robust
precise
disease
plants.
Язык: Английский