Electronics,
Journal Year:
2024,
Volume and Issue:
14(1), P. 62 - 62
Published: Dec. 27, 2024
The
accurate
detection
and
localization
of
polyps
during
endoscopic
examinations
are
critical
for
early
disease
diagnosis
cancer
prevention.
However,
the
presence
artifacts
noise,
along
with
high
similarity
between
surrounding
tissues
in
color,
shape,
texture
complicates
polyp
video
frames.
To
tackle
these
challenges,
we
deployed
multivariate
regression
analysis
to
refine
model
introduced
a
Noise-Suppressing
Perception
Network
(NSPNet)
designed
enhanced
performance.
NSPNet
leverages
wavelet
transform
enhance
model’s
resistance
noise
while
improving
multi-frame
collaborative
strategy
dynamic
videos,
efficiently
utilizing
temporal
information
strengthen
features
across
Specifically,
High-Low
Frequency
Feature
Fusion
(HFLF)
framework,
which
allows
capture
high-frequency
details
more
effectively.
Additionally,
an
improved
STFT-LSTM
Polyp
Detection
(SLPD)
module
that
utilizes
from
sequences
feature
fusion
environments.
Lastly,
integrated
Image
Augmentation
(IAPD)
improve
performance
on
unseen
data
through
preprocessing
enhancement
strategies.
Extensive
experiments
demonstrate
outperforms
nine
SOTA
methods
four
datasets
key
metrics,
including
F1Score
recall.
Remote Sensing,
Journal Year:
2024,
Volume and Issue:
16(10), P. 1712 - 1712
Published: May 11, 2024
Landslide
disasters
have
garnered
significant
attention
due
to
their
extensive
devastating
impact,
leading
a
growing
emphasis
on
the
prompt
and
precise
identification
detection
of
landslides
as
prominent
area
research.
Previous
research
has
primarily
relied
human–computer
interactions
visual
interpretation
from
remote
sensing
identify
landslides.
However,
these
methods
are
time-consuming,
labor-intensive,
subjective,
low
level
accuracy
in
extracting
data.
An
essential
task
deep
learning,
semantic
segmentation,
been
crucial
automated
image
recognition
tasks
because
its
end-to-end
pixel-level
classification
capability.
In
this
study,
mitigate
disadvantages
existing
landslide
methods,
we
propose
multiscale
segment
network
(MsASNet)
that
acquires
different
scales
features,
designs
an
encoder–decoder
structure
strengthen
boundary,
combines
channel
mechanism
feature
extraction
The
MsASNet
model
exhibited
average
95.13%
test
set
Bijie’s
dataset,
mean
91.45%
Chongqing’s
90.17%
Tianshui‘s
signifying
ability
extract
information
efficiently
accurately
real
time.
Our
proposed
may
be
used
efforts
toward
prevention
control
geological
disasters.
Wiley Interdisciplinary Reviews Data Mining and Knowledge Discovery,
Journal Year:
2025,
Volume and Issue:
15(1)
Published: Jan. 20, 2025
ABSTRACT
Medical
image
segmentation
is
prerequisite
in
computer‐aided
diagnosis.
As
the
field
experiences
tremendous
paradigm
changes
since
introduction
of
foundation
models,
technicality
deep
medical
model
no
longer
a
privilege
limited
to
computer
science
researchers.
A
comprehensive
educational
resource
suitable
for
researchers
broad,
different
backgrounds
such
as
biomedical
and
medicine,
needed.
This
review
strategically
covers
evolving
trends
that
happens
fundamental
components
emerging
multimodal
datasets,
updates
on
learning
libraries,
classical‐to‐contemporary
development
models
latest
challenges
with
focus
enhancing
interpretability
generalizability
model.
Last,
conclusion
section
highlights
future
worth
further
attention
investigations.
Bioengineering,
Journal Year:
2024,
Volume and Issue:
11(6), P. 545 - 545
Published: May 27, 2024
Automatically
segmenting
polyps
from
colonoscopy
videos
is
crucial
for
developing
computer-assisted
diagnostic
systems
colorectal
cancer.
Existing
automatic
polyp
segmentation
methods
often
struggle
to
fulfill
the
real-time
demands
of
clinical
applications
due
their
substantial
parameter
count
and
computational
load,
especially
those
based
on
Transformer
architectures.
To
tackle
these
challenges,
a
novel
lightweight
long-range
context
fusion
network,
named
LightCF-Net,
proposed
in
this
paper.
This
network
attempts
model
spatial
dependencies
while
maintaining
performance,
better
distinguish
background
noise
thus
improve
accuracy.
A
Fusion
Attention
Encoder
(FAEncoder)
designed
which
integrates
Large
Kernel
(LKA)
channel
attention
mechanisms
extract
deep
representational
features
unearth
dependencies.
Furthermore,
newly
Visual
Mamba
module
(VAM)
added
skip
connections,
modeling
encoder-extracted
reducing
interference
through
mechanism.
Finally,
Pyramid
Split
(PSA)
used
bottleneck
layer
richer
multi-scale
contextual
features.
The
method
was
thoroughly
evaluated
four
renowned
datasets:
Kvasir-SEG,
CVC-ClinicDB,
BKAI-IGH,
ETIS.
Experimental
findings
demonstrate
that
delivers
higher
accuracy
less
time,
consistently
outperforming
most
advanced
networks.
Bioengineering,
Journal Year:
2024,
Volume and Issue:
11(10), P. 959 - 959
Published: Sept. 25, 2024
Colorectal
cancer
remains
a
leading
cause
of
cancer-related
deaths
worldwide,
with
early
detection
and
removal
polyps
being
critical
in
preventing
disease
progression.
Automated
polyp
segmentation,
particularly
colonoscopy
images,
is
challenging
task
due
to
the
variability
appearance
low
contrast
between
surrounding
tissues.
In
this
work,
we
propose
an
edge-enhanced
network
(EENet)
designed
address
these
challenges
by
integrating
two
novel
modules:
covariance
attention
(CEEA)
cross-scale
edge
enhancement
(CSEE)
modules.
The
CEEA
module
leverages
covariance-based
enhance
boundary
detection,
while
CSEE
bridges
multi-scale
features
preserve
fine-grained
details.
To
further
improve
accuracy
introduce
hybrid
loss
function
that
combines
cross-entropy
edge-aware
loss.
Extensive
experiments
show
EENet
achieves
Dice
score
0.9208
IoU
0.8664
on
Kvasir-SEG
dataset,
surpassing
state-of-the-art
models
such
as
Polyp-PVT
PraNet.
Furthermore,
it
records
0.9316
0.8817
CVC-ClinicDB
demonstrating
its
strong
potential
for
clinical
application
segmentation.
Ablation
studies
validate
contribution
Semantic
segmentation
is
a
significant
and
demanding
work
in
computer
vision
it
has
gained
more
attention
worldwide.
This
article
delivers
an
in-depth
analysis
of
vision-based
semantic
approaches
for
3D
point
cloud
data.
investigates
the
emergence
development
both
domestically
internationally.
It
also
outlines
historical
evolution
various
branches
emphasizing
recent
advancements
driven
by
deep
learning
techniques.
Despite
notable
progress,
challenges
persist,
including
handling
variability
object
shapes
sizes,
computational
costs,
robustness
against
different
conditions.
survey
aims
to
evaluate
synthesize
current
research,
identifying
strengths
weaknesses
traditional
modern
methods,
highlighting
potential
future
research
directions.
The
study
offers
valuable
information
on
implementation
performance
presenting
comprehensive
methodologies,
datasets,
evaluation
metrics
guiding
researchers
towards
suitable
techniques
several
applications.
Algorithms,
Journal Year:
2025,
Volume and Issue:
18(1), P. 42 - 42
Published: Jan. 12, 2025
The
colonoscopy
procedure
heavily
relies
on
the
operator’s
expertise,
underscoring
importance
of
automated
polyp
segmentation
techniques
in
enhancing
efficiency
and
accuracy
colorectal
cancer
diagnosis.
Nevertheless,
achieving
precise
remains
a
significant
challenge
due
to
high
visual
similarity
between
polyps
their
backgrounds,
blurred
boundaries,
complex
localization.
To
address
these
challenges,
Multi-scale
Selective
Edge-Aware
Network
has
been
proposed
facilitate
segmentation.
model
consists
three
key
components:
(1)
an
Edge
Feature
Extractor
(EFE)
that
captures
edge
features
with
precision
during
initial
encoding
phase,
(2)
Cross-layer
Context
Fusion
(CCF)
block
designed
extract
integrate
multi-scale
contextual
information
from
diverse
receptive
fields,
(3)
Aware
(SEA)
module
enhances
sensitivity
high-frequency
details
decoding
thereby
improving
preservation
accuracy.
effectiveness
our
rigorously
validated
Kvasir-SEG,
Kvasir-Sessile,
BKAI
datasets,
mean
Dice
scores
91.92%,
82.10%,
92.24%,
respectively,
test
sets.
Journal of Imaging,
Journal Year:
2025,
Volume and Issue:
11(2), P. 55 - 55
Published: Feb. 13, 2025
Segmentation
neural
networks
are
widely
used
in
medical
imaging
to
identify
anomalies
that
may
impact
patient
health.
Despite
their
effectiveness,
these
face
significant
challenges,
including
the
need
for
extensive
annotated
data,
time-consuming
manual
segmentation
processes
and
restricted
data
access
due
privacy
concerns.
In
contrast,
classification
networks,
similar
capture
essential
parameters
identifying
objects
during
training.
This
paper
leverages
this
characteristic,
combined
with
explainable
artificial
intelligence
(XAI)
techniques,
address
challenges
of
segmentation.
By
adapting
tasks,
proposed
approach
reduces
dependency
on
To
demonstrate
concept,
Medical
Decathlon
'Brain
Tumours'
dataset
was
utilised.
A
ResNet
network
trained,
XAI
tools
were
applied
generate
segmentation-like
outputs.
Our
findings
reveal
GuidedBackprop
is
among
most
efficient
effective
methods,
producing
heatmaps
closely
resemble
masks
by
accurately
highlighting
entirety
target
object.
BioMedInformatics,
Journal Year:
2025,
Volume and Issue:
5(1), P. 10 - 10
Published: Feb. 14, 2025
Background:
In
clinical
practice,
identifying
the
location
and
extent
of
tumors
lesions
is
crucial
for
disease
diagnosis
treatment.
Artificial
intelligence,
particularly
deep
neural
networks,
offers
precise
automated
segmentation,
yet
limited
data
high
computational
demands
often
hinder
its
application.
Transfer
learning
helps
mitigate
these
challenges
by
significantly
reducing
costs,
although
applying
models
can
still
be
resource
intensive.
This
study
aims
to
present
flexible
computationally
efficient
architecture
that
leverages
transfer
delivers
highly
accurate
results
across
various
medical
imaging
problems.
Methods:
We
evaluated
three
datasets
with
varying
similarities
ImageNet:
ISIC
2018
(skin
lesions),
CBIS-DDSM
(breast
masses),
Shenzhen
Montgomery
CXR
Set
(lung
segmentation).
An
ablation
on
tested
pre-trained
backbones,
architectures,
loss
functions.
Results:
The
optimal
configuration—DeepLabV3+
a
ResNet50
backbone
Log-Cosh
Dice
loss—was
validated
remaining
datasets,
achieving
state-of-the-art
results.
Conclusion:
Computationally
simpler
architectures
deliver
robust
performance
without
extensive
resources,
establishing
DeepLabV3+
as
baseline
future
studies.
domain,
enhancing
quality
more
critical
improving
segmentation
accuracy
than
increasing
model
complexity.