International Journal for Scientific Research,
Год журнала:
2024,
Номер
3(2), С. 247 - 266
Опубликована: Фев. 29, 2024
The
recent
pandemic
caused
by
Severe
Acute
Respiratory
Syndrome
Coronavirus
2
(SARS-CoV-2)
has
highlighted
the
importance
of
early
detection
infections,
especially
when
RT-PCR
testing
equipment
is
scarce.
This
study
introduces
a
machine
learning
algorithm
using
CT
scan
imaging
for
rapid
COVID-19
identification.
algorithm,
designed
as
computer-aided
model,
analyzed
536
images
(32x32
pixels)
categorized
into
infected
and
non-infected
groups.
model
preprocesses
Prewitt
filter
discrete
cosine
transform,
then
extracts
features
through
various
statistical
methods
histogram
oriented
gradients
(HOG).
Out
32
features,
29
showed
high
significance
(p-value
<
0.05),
effectively
distinguishing
normal
abnormal
cases.
These
were
classified
support
vector
(SVM)
k-nearest
neighbor
(KNN)
methods.
Performance
metrics
like
sensitivity,
specificity,
accuracy
used
to
evaluate
classifiers.
results
that
classifiers
KNN-1,
KNN-3,
KNN-5,
SVM-Linear
could
distinguish
between
perfectly
(100%)
it
was
applied
proposed
on
tested
ROIs
images.
Also,
SVM-RBF
had
less
performance
than
other
with
98.38%
but
still
at
high-performance
level.
indicate
physicians
can
utilize
an
assisted
tool
detecting
COVID-19.
Electronics,
Год журнала:
2022,
Номер
11(18), С. 2896 - 2896
Опубликована: Сен. 13, 2022
Deep
learning
is
a
convenient
method
for
doctors
to
classify
pulmonary
diseases
such
as
COVID-19,
viral
pneumonia,
bacterial
and
tuberculosis.
However,
task
requires
dataset
including
samples
of
all
these
more
effective
network
capture
the
features
images
accurately.
In
this
paper,
we
propose
five-classification
disease
model,
pre-processing
input
data,
feature
extraction,
classifier.
The
main
points
model
are
follows.
Firstly,
present
new
named
RED-CNN
which
based
on
CNN
architecture
constructed
using
RED
block.
block
composed
Res2Net
module,
ECA
Double
BlazeBlock
capable
extracting
detailed
information,
providing
cross-channel
enhancing
extraction
global
information
with
strong
capability.
Secondly,
by
merging
two
selected
datasets,
Curated
Chest
X-Ray
Image
Dataset
COVID-19
tuberculosis
(TB)
chest
X-ray
database,
five
types
data:
normal,
order
assess
efficiency
proposed
series
experiments
were
carried
out
5-fold
cross
validation,
results
accuracy,
precision,
recall,
F1
value,
Jaccard
scores
91.796%,
92.062%,
91.892%,
86.176%,
respectively.
Our
algorithm
performs
better
than
other
classification
algorithms.
International Journal for Scientific Research,
Год журнала:
2024,
Номер
3(2), С. 247 - 266
Опубликована: Фев. 29, 2024
The
recent
pandemic
caused
by
Severe
Acute
Respiratory
Syndrome
Coronavirus
2
(SARS-CoV-2)
has
highlighted
the
importance
of
early
detection
infections,
especially
when
RT-PCR
testing
equipment
is
scarce.
This
study
introduces
a
machine
learning
algorithm
using
CT
scan
imaging
for
rapid
COVID-19
identification.
algorithm,
designed
as
computer-aided
model,
analyzed
536
images
(32x32
pixels)
categorized
into
infected
and
non-infected
groups.
model
preprocesses
Prewitt
filter
discrete
cosine
transform,
then
extracts
features
through
various
statistical
methods
histogram
oriented
gradients
(HOG).
Out
32
features,
29
showed
high
significance
(p-value
<
0.05),
effectively
distinguishing
normal
abnormal
cases.
These
were
classified
support
vector
(SVM)
k-nearest
neighbor
(KNN)
methods.
Performance
metrics
like
sensitivity,
specificity,
accuracy
used
to
evaluate
classifiers.
results
that
classifiers
KNN-1,
KNN-3,
KNN-5,
SVM-Linear
could
distinguish
between
perfectly
(100%)
it
was
applied
proposed
on
tested
ROIs
images.
Also,
SVM-RBF
had
less
performance
than
other
with
98.38%
but
still
at
high-performance
level.
indicate
physicians
can
utilize
an
assisted
tool
detecting
COVID-19.