PeerJ Computer Science,
Год журнала:
2024,
Номер
10, С. e2290 - e2290
Опубликована: Сен. 9, 2024
The
adoption
and
integration
of
the
Internet
Things
(IoT)
have
become
essential
for
advancement
many
industries,
unlocking
purposeful
connections
between
objects.
However,
surge
in
IoT
has
also
made
it
a
prime
target
malicious
attacks.
Consequently,
ensuring
security
systems
ecosystems
emerged
as
crucial
research
area.
Notably,
advancements
addressing
these
threats
include
implementation
intrusion
detection
(IDS),
garnering
considerable
attention
within
community.
In
this
study,
aim
to
enhance
network
anomaly
detection,
we
present
novel
approach:
Deep
Neural
Decision
Forest-based
IDS
(DNDF-IDS).
DNDF-IDS
incorporates
an
improved
decision
forest
model
coupled
with
neural
networks
achieve
heightened
accuracy
(ACC).
Employing
four
distinct
feature
selection
methods
separately,
namely
principal
component
analysis
(PCA),
LASSO
regression
(LR),
SelectKBest,
Random
Forest
Feature
Importance
(RFFI),
our
objective
is
streamline
training
prediction
processes,
overall
performance,
identify
most
correlated
features.
Evaluation
on
three
diverse
datasets
(NSL-KDD,
CICIDS2017,
UNSW-NB15)
reveals
impressive
ACC
values
ranging
from
94.09%
98.84%,
depending
dataset
method.
achieves
remarkable
time
0.1
ms
per
record.
Comparative
analyses
other
recent
random
Convolutional
Networks
(CNN)
based
models
indicate
that
performs
similarly
or
even
outperforms
them
certain
instances,
particularly
when
utilizing
top
10
One
key
advantage
lies
its
ability
make
accurate
predictions
only
few
features,
showcasing
efficient
utilization
computational
resources.
Frontiers in Plant Science,
Год журнала:
2025,
Номер
15
Опубликована: Янв. 13, 2025
Introduction
Weeds
are
a
major
factor
affecting
crop
yield
and
quality.
Accurate
identification
localization
of
crops
weeds
essential
for
achieving
automated
weed
management
in
precision
agriculture,
especially
given
the
challenges
recognition
accuracy
real-time
processing
complex
field
environments.
To
address
this
issue,
paper
proposes
an
efficient
crop-weed
segmentation
model
based
on
improved
UNet
architecture
attention
mechanisms
to
enhance
both
speed.
Methods
The
adopts
encoder-decoder
structure
UNet,
utilizing
MaxViT
(Multi-Axis
Vision
Transformer)
as
encoder
capture
global
local
features
within
images.
Additionally,
CBAM
(Convolutional
Block
Attention
Module)
is
incorporated
into
decoder
multi-scale
feature
fusion
module,
adaptively
adjusting
map
weights
enable
focus
more
accurately
edges
textures
weeds.
Results
discussion
Experimental
results
show
that
proposed
achieved
84.28%
mIoU
88.59%
mPA
sugar
beet
dataset,
representing
improvements
3.08%
3.15%
over
baseline
model,
respectively,
outperforming
mainstream
models
such
FCN,
PSPNet,
SegFormer,
DeepLabv3+,
HRNet.
Moreover,
model’s
inference
time
only
0.0559
seconds,
reducing
computational
overhead
while
maintaining
high
accuracy.
Its
performance
sunflower
dataset
further
verifies
generalizability
robustness.
This
study,
therefore,
provides
accurate
solution
segmentation,
laying
foundation
future
research
identification.
International Journal of Robotics and Automation Technology,
Год журнала:
2024,
Номер
11, С. 1 - 12
Опубликована: Май 22, 2024
Abstract:
This
work
aims
to
test
the
performance
of
you
only
look
once
version
8
(YOLOv8)
model
for
problem
drone
detection.
Drones
are
very
slightly
regulated
and
standards
need
be
established.
With
a
robust
system
detecting
drones
possibilities
regulating
their
usage
becoming
realistic.
Five
different
sizes
were
tested
determine
best
architecture
size
this
problem.
The
results
indicate
high
across
all
models
that
each
is
used
specific
case.
Smaller
suited
lightweight
approaches
where
some
false
identification
tolerable,
while
largest
with
stationary
systems
require
precision.
Soil Use and Management,
Год журнала:
2025,
Номер
41(1)
Опубликована: Янв. 1, 2025
Abstract
Slope
stability
is
a
critical
factor
in
ensuring
the
safety
and
longevity
of
infrastructure,
especially
areas
prone
to
landslides
soil
erosion.
Traditional
methods
slope
assessment,
while
widely
used,
often
struggle
provide
accurate
results
when
applied
Technosols—soils
modified
by
human
activities
composed
waste
materials.
This
study
proposes
novel
approach
that
combines
artificial
intelligence
techniques
improve
precision
predictions
these
complex
types.
The
method
utilizes
model
based
on
neural
networks,
trained
large
dataset
factors.
Unlike
conventional
techniques,
proposed
integrates
multiple
environmental
material
properties
more
assessment
compared
other
models.
model's
performance
demonstrated
R
2
values
.999975
for
test
datasets,
which
significantly
better
than
similar
work
statistical
analysis.
Moreover,
incorporating
Shapley
Additive
Explanations
(SHAP),
we
clear
understanding
impact
various
parameters
stability.
findings
suggest
machine
learning‐based
offers
reliable
tool
evaluation
Technosols,
making
it
valuable
addition
field.