Machine Graphics and Vision,
Journal Year:
2024,
Volume and Issue:
33(2), P. 77 - 90
Published: Dec. 23, 2024
Pupillometry
measures
pupil
size,
and
several
open-source
algorithms
are
available
to
analyse
pupillometry
data.
However,
only
a
few
studies
compared
these
algorithms'
accuracy
computational
resources.
This
study
aims
compare
the
of
computer
vision-based
(Swirski,
Starburst,
PuRe,
ElSe,
ExCuSe
algorithms)
machine
learning
algorithm,
DeepLabCut,
double-blinded
human
examiners
(gold-standard).
Training
DeepLabCut
with
different
architectures
variable
number
markers
(2-9
markers)
was
done
on
an
dataset.
The
duration
training
statistically
longer
for
ResNet152
model
MobileNet
model.
diameters
in
software
such
as
Swirski
were
from
measurements.
2
3
marker
models
closest
In
conclusion,
this
work
highlights
efficiency
lower
based
architecture
which
consumes
fewer
resources
is
more
accurate.
Technology and Health Care,
Journal Year:
2025,
Volume and Issue:
unknown
Published: Jan. 19, 2025
Background
Deficits
in
concentration
with
social
stimuli
are
more
common
children
affected
by
autism
spectrum
disorder
(ASD).
Developing
visual
attention
is
one
of
the
most
vital
elements
for
detecting
autism.
Eye
tracking
technology
a
potential
method
to
identify
an
early
biomarker
based
on
children's
abnormal
patterns.
Objective
retinal
scan
path
images
can
be
generated
eyeball
movement
during
time
watching
screen
and
capture
eye
projection
sequences,
which
helps
analyze
behavior
children.
The
Shi-Tomasi
corner
detection
methodology
uses
open
CV
corners
gaze
images.
Methods
In
proposed
ADET
model,
detection-based
vision
transformer
(CD-ViT)
technique
utilized
diagnose
at
stage.
Generally,
model
divides
input
into
patches,
fed
encoder
process.
fine-tuned
resolve
binary
classification
issues
once
features
extracted
via
remora
optimization.
Specifically,
acts
as
cornerstone
work
help
technique.
This
study
dataset
547
eye-tracking
both
non-autistic
Results
Experimental
results
show
that
suggested
frameworkachieves
better
accuracy
38.31%,
23.71%,
13.01%,
1.56%,
18.26%,
44.56%
than
RM3ASD,
MLP,
SVM,
CNN,
our
methods.
Conclusions
screening
strongly
suggests
it
used
assist
medical
professionals
providing
efficient
accurate
detection.
Frontiers in Psychiatry,
Journal Year:
2025,
Volume and Issue:
16
Published: April 14, 2025
Autism
is
a
serious
threat
to
an
individual’s
physical
and
mental
health.
Early
screening,
diagnosis,
intervention
can
effectively
reduce
the
level
of
deficits
in
individuals
with
autism.
However,
traditional
methods
rely
on
professionalism
psychiatrists
require
great
deal
time
effort,
resulting
large
proportion
autism
being
diagnosed
after
age
6.
Artificial
intelligence
(AI)
combined
machine
learning
used
improve
efficiency
early
young
children.
This
review
aims
summarize
AI-assisted
for
children
(infants,
toddlers,
preschoolers).
To
achieve
screening
diagnosis
children,
AI
have
built
predictive
models
automation
behavioral
analyzed
brain
imaging
genetic
data
break
barrier
established
intelligent
systems
mass
screening.
For
education
optimize
teaching
environment
provide
individualized
interventions,
constructed
monitoring
dynamic
tracking,
created
support
continuous
meet
diverse
needs
As
continues
develop,
further
research
needed
build
shared
database
autism,
generalize
migrate
effects
appearance
performance
AI-powered
robots,
failure
rates
costs
technologies.
Applied Sciences,
Journal Year:
2024,
Volume and Issue:
14(7), P. 3053 - 3053
Published: April 5, 2024
Autism
Spectrum
Disorder
is
known
to
cause
difficulties
in
social
interaction
and
communication,
as
well
repetitive
patterns
of
behavior,
interests,
or
hobbies.
These
challenges
can
significantly
affect
the
individual’s
daily
life.
Therefore,
it
crucial
identify
assess
children
with
early
benefit
long-term
health
children.
Unfortunately,
many
are
not
diagnosed
misdiagnosed,
which
means
they
miss
out
on
necessary
interventions.
Clinicians
other
experts
face
various
during
diagnostic
process.
Digital
tools
facilitate
diagnosis
effectively.
This
study
aimed
explore
use
machine
learning
techniques
a
dataset
collected
from
serious
game
designed
for
autism
investigate
how
these
assist
classification
make
clinical
process
more
efficient.
The
responses
were
gathered
who
participated
interactive
games
deployed
mobile
devices,
data
analyzed
using
types
neural
networks,
such
multilayer
perceptrons
constructed
networks.
performance
metrics
models,
including
error
rate,
precision,
recall,
reported,
comparative
experiments
revealed
that
network
integer
rule-based
networks
approach
was
superior.
Based
evaluation
metrics,
this
method
showed
lowest
rate
11.77%,
high
accuracy
0.75,
good
recall
0.66.
Thus,
be
an
effective
way
classify
both
typically
developed
Disorder.
Additionally,
used
automatic
screening
procedures
intelligent
system.
results
indicate
clinicians
could
enhance
conventional
methods
contribute
providing
better
care
individuals
autism.
Diagnostics,
Journal Year:
2024,
Volume and Issue:
14(10), P. 1047 - 1047
Published: May 18, 2024
One
of
the
most
challenging
problems
when
diagnosing
autism
spectrum
disorder
(ASD)
is
need
for
long
sets
data.
Collecting
data
during
such
periods
challenging,
particularly
dealing
with
children.
This
challenge
motivates
investigation
possible
classifiers
ASD
that
do
not
sets.
In
this
paper,
we
use
eye-tracking
covering
only
5
s
and
introduce
one
metric
able
to
distinguish
between
typically
developed
(TD)
gaze
patterns
based
on
short
time-series
compare
it
two
benchmarks,
using
traditional
metrics
state-of-the-art
AI
classifier.
Although
can
track
disorders
in
visual
attention
our
approach
a
substitute
medical
diagnosis,
find
newly
introduced
achieve
an
accuracy
93%
classifying
eye
trajectories
from
children
surpassing
both
benchmarks
while
needing
fewer
The
classification
method,
series,
performs
better
than
standard
at
level
best
even
these
are
trained
longer
time
series.
We
also
discuss
advantages
limitations
method
comparison
state
art:
besides
low
amount
data,
simple,
understandable,
straightforward
criterion
apply,
which
often
contrasts
“black
box”
methods.
2022 International Conference on Advances in Computing, Communication and Applied Informatics (ACCAI),
Journal Year:
2024,
Volume and Issue:
unknown, P. 1 - 8
