IEEE Access,
Journal Year:
2021,
Volume and Issue:
9, P. 147512 - 147526
Published: Jan. 1, 2021
In
this
era
of
COVID19,
proper
diagnosis
and
treatment
for
pneumonia
are
very
important.
Chest
X-Ray
(CXR)
image
analysis
plays
a
vital
role
in
the
reliable
pneumonia.
An
experienced
radiologist
is
required
this.
However,
even
an
radiographer,
it
quite
difficult
time-consuming
to
diagnose
due
fuzziness
CXR
images.
Also,
identification
can
be
erroneous
involvement
human
judgment.
Hence,
authentic
automated
system
play
important
here.
cutting-edge
technology,
deep
learning
(DL)
highly
used
every
sector.
There
several
existing
methods
but
they
have
accuracy
problems.
study,
automatic
detection
has
been
proposed
by
applying
extreme
machine
(ELM)
on
Kaggle
images
(Pneumonia).
Three
models
studied:
classification
using
(ELM),
ELM
with
hybrid
convolutional
neural
network
-
principle
component
(CNN-PCA)
based
feature
extraction
(ECP),
ECP
which
contrast-enhanced
contrast
limited
adaptive
histogram
equalization
(CLAHE).
Among
these
three
methods,
final
model
provides
optimistic
result.
It
achieves
recall
score
98%
98.32%
multiclass
classification.
On
other
hand,
binary
100%
99.83%
accuracy.
The
method
also
outperforms
methods.
outcome
compared
benchmarks
that
include
accuracy,
precision,
recall,
etc.
Sensors,
Journal Year:
2022,
Volume and Issue:
22(3), P. 807 - 807
Published: Jan. 21, 2022
After
lung
cancer,
breast
cancer
is
the
second
leading
cause
of
death
in
women.
If
detected
early,
mortality
rates
women
can
be
reduced.
Because
manual
diagnosis
takes
a
long
time,
an
automated
system
required
for
early
detection.
This
paper
proposes
new
framework
classification
from
ultrasound
images
that
employs
deep
learning
and
fusion
best
selected
features.
The
proposed
divided
into
five
major
steps:
(i)
data
augmentation
performed
to
increase
size
original
dataset
better
Convolutional
Neural
Network
(CNN)
models;
(ii)
pre-trained
DarkNet-53
model
considered
output
layer
modified
based
on
augmented
classes;
(iii)
trained
using
transfer
features
are
extracted
global
average
pooling
layer;
(iv)
two
improved
optimization
algorithms
known
as
reformed
differential
evaluation
(RDE)
gray
wolf
(RGW);
(v)
fused
probability-based
serial
approach
classified
machine
algorithms.
experiment
was
conducted
Breast
Ultrasound
Images
(BUSI)
dataset,
accuracy
99.1%.
When
compared
with
recent
techniques,
outperforms
them.
Computational and Mathematical Methods in Medicine,
Journal Year:
2021,
Volume and Issue:
2021, P. 1 - 12
Published: July 15, 2021
In
a
general
computational
context
for
biomedical
data
analysis,
DNA
sequence
classification
is
crucial
challenge.
Several
machine
learning
techniques
have
used
to
complete
this
task
in
recent
years
successfully.
Identification
and
of
viruses
are
essential
avoid
an
outbreak
like
COVID-19.
Regardless,
the
feature
selection
process
remains
most
challenging
aspect
issue.
The
commonly
representations
worsen
case
high
dimensionality,
sequences
lack
explicit
features.
It
also
helps
detecting
effect
drug
design.
days,
deep
(DL)
models
can
automatically
extract
features
from
input.
work,
we
employed
CNN,
CNN-LSTM,
CNN-Bidirectional
LSTM
architectures
using
Label
-mer
encoding
classification.
evaluated
on
different
metrics.
From
experimental
results,
CNN
with
id="M2">
offers
accuracy
93.16%
93.13%,
respectively,
testing
data.
Sustainable Cities and Society,
Journal Year:
2021,
Volume and Issue:
72, P. 103079 - 103079
Published: June 9, 2021
A
sustainable
healthcare
focuses
on
enhancing
and
restoring
public
health
parameters
thereby
reducing
gloomy
impacts
social,
economic
environmental
elements
of
a
city.
Though
it
has
uplifted
health,
yet
the
rise
chronic
diseases
is
concern
in
cities.
In
this
work,
lung
cancer
detection
model
developed
to
integrate
Internet
Health
Things
(IoHT)
computational
intelligence,
causing
least
harm
environment.
IoHT
unit
retains
connectivity
continuously
generates
data
from
patients.
Heuristic
Greedy
Best
First
Search
(GBFS)
algorithm
used
select
most
relevant
attributes
upon
which
random
forest
applied
classify
differentiates
affected
patients
normal
ones
based
detected
symptoms.
It
observed
during
experiment
that
GBFS-Random
shows
promising
outcome.
While
an
optimal
accuracy
98.8
%
was
generated,
simultaneously,
latency
1.16
s
noted.
Specificity
sensitivity
recorded
with
proposed
are
97.5
97.8
%,
respectively.
The
mean
accuracy,
specificity,
sensitivity,
f-score
value
96.96
96.26
96.34
96.32
respectively,
over
various
types
datasets
implemented.
smart
intelligent
sustainable.
reduces
unnecessary
manual
overheads,
safe,
preserves
resources
human
resources,
assists
medical
professionals
quick
reliable
decision
making
diagnosis.
Sensors,
Journal Year:
2021,
Volume and Issue:
21(12), P. 4093 - 4093
Published: June 14, 2021
The
Internet
of
Medical
Things
(IoMT)
is
increasingly
being
used
for
healthcare
purposes.
IoMT
enables
many
sensors
to
collect
patient
data
from
various
locations
and
send
it
a
distributed
hospital
further
study.
provides
patients
with
variety
paid
programmes
help
them
keep
track
their
health
problems.
However,
the
current
system
services
are
expensive,
offloaded
in
network
insecure.
research
develops
new,
cost-effective
stable
framework
based
on
blockchain-enabled
fog
cloud.
study
aims
reduce
cost
application
as
they
processing
system.
devises
an
different
algorithm
techniques,
such
Blockchain-Enable
Smart-Contract
Cost-Efficient
Scheduling
Algorithm
Framework
(BECSAF)
schemes.
Blockchain
schemes
ensure
consistency
validation
symmetric
cryptography.
due
workflow
tasks
scheduled
other
nodes,
heterogeneous,
earliest
finish,
time-based
scheduling
deals
execution
under
deadlines.
Simulation
results
show
that
proposed
outperform
all
existing
baseline
approaches
terms
implementation
applications.
Diagnostics,
Journal Year:
2021,
Volume and Issue:
11(7), P. 1212 - 1212
Published: July 5, 2021
Breast
cancer
is
becoming
more
dangerous
by
the
day.
The
death
rate
in
developing
countries
rapidly
increasing.
As
a
result,
early
detection
of
breast
critical,
leading
to
lower
rate.
Several
researchers
have
worked
on
segmentation
and
classification
using
various
imaging
modalities.
ultrasonic
modality
one
most
cost-effective
techniques,
with
higher
sensitivity
for
diagnosis.
proposed
study
segments
lesion
images
Dilated
Semantic
Segmentation
Network
(Di-CNN)
combined
morphological
erosion
operation.
For
feature
extraction,
we
used
deep
neural
network
DenseNet201
transfer
learning.
We
propose
24-layer
CNN
that
uses
learning-based
extraction
further
validate
ensure
enriched
features
target
intensity.
To
classify
nodules,
vectors
obtained
from
were
fused
parallel
fusion.
methods
evaluated
10-fold
cross-validation
vector
combinations.
accuracy
CNN-activated
DenseNet201-activated
Support
Vector
Machine
(SVM)
classifier
was
90.11
percent
98.45
percent,
respectively.
With
98.9
accuracy,
version
SVM
outperformed
other
algorithms.
When
compared
recent
algorithms,
algorithm
achieves
better
diagnosis
Applied Artificial Intelligence,
Journal Year:
2021,
Volume and Issue:
35(15), P. 2157 - 2203
Published: Dec. 2, 2021
Breast
cancer
is
one
of
the
most
prevalent
types
that
plagues
females.
Mortality
from
breast
could
be
reduced
by
diagnosing
and
identifying
it
at
an
early
stage.
To
detect
cancer,
various
imaging
modalities
can
used,
such
as
mammography.
Computer-Aided
Detection/Diagnosis
(CAD)
systems
assist
expert
radiologist
to
diagnose
This
paper
introduces
findings
a
systematic
review
seeks
examine
state-of-the-art
CAD
for
detection.
based
on
118
publications
published
in
2018–2021
retrieved
major
scientific
publication
databases
while
using
rigorous
methodology
review.
We
provide
general
description
analysis
existing
use
machine
learning
methods
well
their
current
state
mammogram
image
classification
methods.
presents
all
stages
including
pre-processing,
segmentation,
feature
extraction,
selection,
classification.
identify
research
gaps
outline
recommendations
future
research.
may
helpful
both
clinicians,
who
diagnosis
researchers
find
knowledge
create
more
contributions
diagnostics.
Applied Sciences,
Journal Year:
2021,
Volume and Issue:
11(24), P. 12122 - 12122
Published: Dec. 20, 2021
Breast
cancer
detection
using
mammogram
images
at
an
early
stage
is
important
step
in
disease
diagnostics.
We
propose
a
new
method
for
the
classification
of
benign
or
malignant
breast
from
images.
Hybrid
thresholding
and
machine
learning
are
used
to
derive
region
interest
(ROI).
The
derived
ROI
then
separated
into
five
different
blocks.
wavelet
transform
applied
suppress
noise
each
produced
block
based
on
BayesShrink
soft
by
capturing
high
low
frequencies
within
sub-bands.
An
improved
fractal
dimension
(FD)
approach,
called
multi-FD
(M-FD),
proposed
extract
multiple
features
denoised
block.
number
extracted
reduced
genetic
algorithm.
Five
classifiers
trained
with
artificial
neural
network
(ANN)
classify
Lastly,
fusion
process
performed
results
blocks
obtain
final
decision.
approach
tested
evaluated
four
benchmark
image
datasets
(MIAS,
DDSM,
INbreast,
BCDR).
present
single-
double-dataset
evaluations.
Only
one
dataset
training
testing
single-dataset
evaluation,
whereas
two
(one
training,
testing)
evaluation.
experiment
show
that
yields
better
INbreast
whilst
obtained
remaining
outperforms
other
state-of-the-art
models
Mini-MIAS
dataset.
Cancers,
Journal Year:
2021,
Volume and Issue:
13(23), P. 6116 - 6116
Published: Dec. 4, 2021
Breast
cancer
is
now
the
most
frequently
diagnosed
in
women,
and
its
percentage
gradually
increasing.
Optimistically,
there
a
good
chance
of
recovery
from
breast
if
identified
treated
at
an
early
stage.
Therefore,
several
researchers
have
established
deep-learning-based
automated
methods
for
their
efficiency
accuracy
predicting
growth
cells
utilizing
medical
imaging
modalities.
As
yet,
few
review
studies
on
diagnosis
are
available
that
summarize
some
existing
studies.
However,
these
were
unable
to
address
emerging
architectures
modalities
diagnosis.
This
focuses
evolving
deep
learning
detection.
In
what
follows,
this
survey
presents
architectures,
analyzes
strengths
limitations
studies,
examines
used
datasets,
reviews
image
pre-processing
techniques.
Furthermore,
concrete
diverse
modalities,
performance
metrics
results,
challenges,
research
directions
future
presented.