IRADA: integrated reinforcement learning and deep learning algorithm for attack detection in wireless sensor networks
Multimedia Tools and Applications,
Год журнала:
2024,
Номер
83(28), С. 71559 - 71578
Опубликована: Фев. 8, 2024
Язык: Английский
Deep Learning and Transfer Learning Methods to Effectively Diagnose Cervical Cancer from Liquid-Based Cytology Pap Smear Images
International Journal of Online and Biomedical Engineering (iJOE),
Год журнала:
2023,
Номер
19(04), С. 77 - 93
Опубликована: Апрель 3, 2023
As
cervical
cancer
is
considered
one
of
the
leading
causes
death
for
women
globally,
different
screening
techniques
have
emerged.
Papanicolaou
technique
generates
high
numbers
false
negatives
due
to
only
testing
20%
a
sample,
liquid-based
cytology
was
developed
test
100%
sample
and
improve
accuracy.
However,
as
larger
size
has
made
it
difficult
detect
lesion
images
through
microscope,
studies
looked
ways
intelligently
analyze
sample.
The
aim
this
study
develop
an
artificial
intelligence
image
recognition
system
that
detects
level
Pap
smears
under
Bethesda
classification
(NI/LSIEL/HSIEL/SCC).
For
purpose,
six
activities
were
carried
out:
dataset
selection,
data
augmentation,
optimization,
model
development,
evaluation
construction.
A
built
from
publicly
available
smear
passed
augmentation
algorithms
generated
total
2,676
images.
Two
models,
ResNet50V2
ResNet101V2,
Deep
Learning
Transfer
protocols.
showed
obtained
better
performance,
where
HSIL
SCC
type
precision
0.98
achieved
accuracy
0.97.
Finally,
based
on
its
performance
validated.
Язык: Английский
Data Security Mechanisms, Approaches, and Challenges for e-Health Smart Systems
International Journal of Online and Biomedical Engineering (iJOE),
Год журнала:
2023,
Номер
19(02), С. 42 - 66
Опубликована: Фев. 16, 2023
In
the
new
era,
trend
of
using
wearable
devices
and
smart
accessories
gained
considerable
popularity
become
a
necessity
utility
for
human
life
due
to
their
major
role
keep
monitoring
health
conditions
providing
healthcare
services.
The
combination
IoT
networks
with
edge
computing
paradigms
develops
an
intelligent
e-health
system
that
aims
monitor
different
real-time
scenarios.
deployment
exposes
several
challenges
regarding
security
privacy
aspects,
particularly
in
case
dealing
enormous
quantity
medical
data
risk
presented
by
exchanging
operations
external
entities.
this
paper
comprehensive
presentation
covered
basic
topics
layers
thus
advantages
limitations
terms
existing
has
been
mentioned,
subsequently,
adapted
exposed
cyber
through
traditional
systems
data,
discussion
blockchain
technology
come
over
application
opportunities,
where
approach
efficiently
ensure
transactions
network,
addition,
overview
outlined
main
research
works
related
technology.
Therefore,
study
diverse
reveals
framework
solutions
system’s
layers,
furthermore,
uncovering
benefits
integrating
technologies
such
as
Machine
Learning
(supervised,
unsupervised
types),
Deep
Learning,
Reinforcement
well
introducing
comparison
analysis
multiple
AI
algorithm
models
based
on
efficiency
future
purposes
provide
meets
patient
needs.
end
review
highlighted
further
directions
actual
open
e-Health
limitations.
Язык: Английский
Hybrid stacked autoencoder with dwarf mongoose optimization for Phishing attack detection in internet of things environment
Alexandria Engineering Journal,
Год журнала:
2024,
Номер
106, С. 164 - 171
Опубликована: Июль 9, 2024
Язык: Английский
Enhancing data security using Laplacian of Gaussian and Chacha20 encryption algorithm
Journal of Intelligent Systems,
Год журнала:
2024,
Номер
33(1)
Опубликована: Янв. 1, 2024
Abstract
Concealing
sensitive
information
inside
multimedia
is
very
interesting
in
secure
communication
due
to
its
wide
application.
This
study
discusses
the
different
difficulties
involved
embedding
ciphertext
video
frame
without
affecting
imperceptibility
and
strength
resistance
against
cyberattacks.
The
main
goal
of
this
suggest
a
novel
technique
conceal
encrypted
data
frames
securely.
done
by
combining
Laplacian
Gaussian
(LoG)
edge
detection
algorithm
Chacha20
encryption
algorithm.
LoG
facilitates
detect
suitable
location
for
concealing
data.
While
ChaCha20
used
encrypt
before
them
these
locations,
which
augmenting
security.
proposed
method
involves
sequence
steps
involving
detecting
edges
frames,
determining
data,
encrypting
XORing
it
with
key
generated
using
algorithm,
determined
edges,
then
reconstruct
rebuild
that
concealed
keeping
lowest
level
visual
distortion.
experimental
results
showed
two
approaches
provide
fast,
robust,
method,
can
be
seen
evaluating
system
mean
square
error,
peak
signal-to-noise
ratio,
correlation,
number
pixels
change
rate,
unified
average
changing
intensity,
entropy,
evaluation
metrics
excellent
results.
suggests
strong
embed
employed
communication,
copyright
protection,
authentication.
Merging
algorithms
produce
perfect
both
security
perception
quality,
provides
means
farther
achievement
techniques.
Язык: Английский
A Comprehensive Review of Cyber-Attacks Targeting IoT Systems and Their Security Measures
International Journal of Safety and Security Engineering,
Год журнала:
2024,
Номер
14(4), С. 1073 - 1086
Опубликована: Авг. 30, 2024
The
Internet
of
Things
(IoT)
represents
the
backbone
current
and
future
technologies.The
main
objective
IoT
is
to
make
human
life
easier
by
automating
most
daily
jobs.The
endless
web
connections
entices
opponents
utilize
IoT's
weaknesses.For
that
reason,
this
technological
innovation
faces
a
few
serious
safety
confidentiality
problems.These
problems
are
actual
motivation
research.This
paper
reviews
latest
research
possible
types
attacks
can
affect
systems
including
exploration
infrastructure.Various
cybersecurity
threats,
network,
application,
physical
attacks,
aim
compromise
discussed.Moreover,
regarding
attack
types,
we
performed
statistical
analysis
using
Excel
for
percentage
found
DDoS
common
with
21%.In
addition,
comparing
Deep
Learning
(DL)
accuracy
measures
traditional
methods,
DL
methods
achieved
an
more
than
98%,
so
they
better
effective
in
detecting
classifying
due
their
high
accuracy.However,
rarely
do
researchers
focus
on
computational
complexity.Finally,
highlights
some
statistics
Python
language
negative
impact
network
traffic.
Язык: Английский
Distributed Multiclass Cyberattack Detection Using Golden Jackal Optimization With Deep Learning Model for Securing IoT Networks
IEEE Access,
Год журнала:
2024,
Номер
12, С. 132434 - 132443
Опубликована: Янв. 1, 2024
Язык: Английский
Attack Detection in Wireless Sensor Networks using Novel Artificial Intelligence Algorithm
Опубликована: Окт. 11, 2023
Язык: Английский
Analysis and study of IDSs for cybersecurity vulnerability detection and prevention: A Comprehensive Review
Lijin Shaji,
Suji Pramila R
2022 13th International Conference on Computing Communication and Networking Technologies (ICCCNT),
Год журнала:
2023,
Номер
unknown
Опубликована: Июль 6, 2023
Smart
environments
are
worried
about
private
keys
and
security
problems.
To
prevent
attacks,
which
take
benefit
of
a
few
these
intrusion
detection
systems
(IDSs)
vulnerabilities,
were
modeled.
In
the
age
Internet
Things
(IoT),
data
generated
by
connected
objects
presents
fundamentals
for
big
analytics
that
could
be
used
to
recognize
anomalies
patterns
in
data.
Numerous
cyber
use
IDSs
detect
intrusions
employ
variety
approaches
architectures.
this
work,
several
research
papers
have
been
utilized
based
on
diverse
detecting
preventing
vulnerabilities.
Moreover,
work
is
reviewed
classifying
them
preceding
works.
addition,
gaps
problems
recognized
previous
works
stated;
thus,
researchers
can
bring
elucidation
improve
their
future.
The
literature
analyzed
features,
such
as
performance
evaluation
measures,
datasets
used.
Язык: Английский