Big Data,
Год журнала:
2023,
Номер
12(2), С. 155 - 172
Опубликована: Июнь 8, 2023
Diabetic
foot
ulcer
(DFU)
is
a
problem
worldwide,
and
prevention
crucial.
The
image
segmentation
analysis
of
DFU
identification
plays
significant
role.
This
will
produce
different
the
same
idea,
incomplete,
imprecise,
other
problems.
To
address
these
issues,
method
through
internet
things
with
technique
virtual
sensing
for
semantically
similar
objects,
four
levels
range
(region-based,
edge-based,
image-based,
computer-aided
design-based
segmentation)
deeper
images
implemented.
In
this
study,
multimodal
compressed
object
co-segmentation
semantical
segmentation.
result
predicting
better
validity
reliability
assessment.
experimental
results
demonstrate
that
proposed
model
can
efficiently
perform
analysis,
lower
error
rate,
than
existing
methodologies.
findings
on
multiple-image
dataset
show
obtains
an
average
score
90.85%
89.03%
correspondingly
in
two
types
labeled
ratios
before
after
without
(i.e.,
25%
30%),
which
increase
10.91%
12.22%
over
previous
best
results.
live
studies,
our
system
improved
by
59.1%
compared
deep
segmentation-based
techniques
its
smart
improvements
contemporaries
are
15.06%,
23.94%,
45.41%,
respectively.
Proposed
range-based
achieves
interobserver
73.9%
positive
test
namely
likelihood
ratio
set
only
0.25
million
parameters
at
pace
data.
Electronics,
Год журнала:
2023,
Номер
12(5), С. 1167 - 1167
Опубликована: Фев. 28, 2023
COVID-19
(coronavirus
disease
2019)
is
a
new
viral
infection
that
widely
spread
worldwide.
Deep
learning
plays
an
important
role
in
images
diagnosis.
This
paper
reviews
the
recent
progress
of
deep
applications
from
five
aspects;
Firstly,
33
datasets
and
data
enhancement
methods
are
introduced;
Secondly,
classification
based
on
supervised
summarized
four
aspects
VGG,
ResNet,
DenseNet
Lightweight
Networks.
The
segmentation
attention
mechanism,
multiscale
residual
connectivity
dense
mechanism;
Thirdly,
application
semi-supervised
diagnosis
terms
consistency
regularization
self-training
methods.
Fourthly,
unsupervised
autoencoder
generative
adversarial
Moreover,
challenges
future
work
diagnostic
field
summarized.
latest
research
status
learning,
which
positive
significance
to
detection
COVID-19.
Journal of Computer Assisted Tomography,
Год журнала:
2024,
Номер
48(4), С. 652 - 662
Опубликована: Янв. 15, 2024
Artificial
intelligence
(AI)-assisted
medical
imaging
technology
is
a
new
research
area
of
great
interest
that
has
developed
rapidly
over
the
last
decade.
However,
there
been
no
bibliometric
analysis
published
studies
in
this
field.
The
present
review
focuses
on
AI-related
computed
tomography
Web
Science
database
and
uses
CiteSpace
VOSviewer
to
generate
knowledge
map
conduct
basic
information
analysis,
co-word
co-citation
analysis.
A
total
7265
documents
were
included
number
had
an
overall
upward
trend.
Scholars
from
United
States
China
have
made
outstanding
achievements,
general
lack
extensive
cooperation
In
recent
years,
areas
difficulty
optimization
upgrading
algorithms,
application
theoretical
models
practical
clinical
applications.
This
will
help
researchers
understand
developments,
interest,
frontiers
field
provide
reference
guidance
for
future
studies.
EAI Endorsed Transactions on Pervasive Health and Technology,
Год журнала:
2024,
Номер
10
Опубликована: Фев. 21, 2024
INTRODUCTION:
In
this
study,
we
explore
the
intricate
domain
of
Diabetic
Foot
Ulcers
(DFU)
through
development
a
comprehensive
framework
that
encompasses
diverse
operational
scenarios.
The
focus
lies
on
identification
and
classification
assessment
diabetic
foot
ulcers,
implementation
smart
health
management
strategies,
collection,
analysis,
intelligent
interpretation
data
related
to
ulcers.
introduces
an
innovative
approach
predicting
ulcers
their
key
characteristics,
offering
technical
solution
for
forecasting.
exploration
delves
into
various
computational
strategies
designed
analysis
tailored
patients
with
OBJECTIVES:
primary
objective
paper
is
present
forecasting
utilizing
analysis.
METHODS:
Techniques
derived
from
social
network
are
employed
conduct
research,
focusing
geared
towards
study
highlights
methodologies
addressing
unique
challenges
posed
by
central
emphasis
integration
Internet
Medical
Things
(IoMT)
in
prediction
strategies.
RESULTS:
main
results
include
proposal
IoMT-based
computing
covering
entire
spectrum
DFU
such
as
localization,
assessment,
management,
detection.
also
acknowledges
faced
previous
including
low
rates
elevated
false
alarm
rates,
proposes
automatic
recognition
approaches
leveraging
advanced
machine
learning
techniques
enhance
accuracy
efficacy.
CONCLUSION:
proposed
significant
advancement
associated
demonstrates
promise
improving
efficiency
ulcer
marking
positive
stride
overcoming
existing
limitations
research.
Big Data,
Год журнала:
2023,
Номер
12(2), С. 155 - 172
Опубликована: Июнь 8, 2023
Diabetic
foot
ulcer
(DFU)
is
a
problem
worldwide,
and
prevention
crucial.
The
image
segmentation
analysis
of
DFU
identification
plays
significant
role.
This
will
produce
different
the
same
idea,
incomplete,
imprecise,
other
problems.
To
address
these
issues,
method
through
internet
things
with
technique
virtual
sensing
for
semantically
similar
objects,
four
levels
range
(region-based,
edge-based,
image-based,
computer-aided
design-based
segmentation)
deeper
images
implemented.
In
this
study,
multimodal
compressed
object
co-segmentation
semantical
segmentation.
result
predicting
better
validity
reliability
assessment.
experimental
results
demonstrate
that
proposed
model
can
efficiently
perform
analysis,
lower
error
rate,
than
existing
methodologies.
findings
on
multiple-image
dataset
show
obtains
an
average
score
90.85%
89.03%
correspondingly
in
two
types
labeled
ratios
before
after
without
(i.e.,
25%
30%),
which
increase
10.91%
12.22%
over
previous
best
results.
live
studies,
our
system
improved
by
59.1%
compared
deep
segmentation-based
techniques
its
smart
improvements
contemporaries
are
15.06%,
23.94%,
45.41%,
respectively.
Proposed
range-based
achieves
interobserver
73.9%
positive
test
namely
likelihood
ratio
set
only
0.25
million
parameters
at
pace
data.
