Information Geometry and Manifold Learning: A Novel Framework for Analyzing Alzheimer’s Disease MRI Data
Diagnostics,
Год журнала:
2025,
Номер
15(2), С. 153 - 153
Опубликована: Янв. 10, 2025
Background:
Alzheimer’s
disease
is
a
progressive
neurological
condition
marked
by
decline
in
cognitive
abilities.
Early
diagnosis
crucial
but
challenging
due
to
overlapping
symptoms
among
impairment
stages,
necessitating
non-invasive,
reliable
diagnostic
tools.
Methods:
We
applied
information
geometry
and
manifold
learning
analyze
grayscale
MRI
scans
classified
into
No
Impairment,
Very
Mild,
Moderate
Impairment.
Preprocessed
images
were
reduced
via
Principal
Component
Analysis
(retaining
95%
variance)
converted
statistical
manifolds
using
estimated
mean
vectors
covariance
matrices.
Geodesic
distances,
computed
with
the
Fisher
Information
metric,
quantified
class
differences.
Graph
Neural
Networks,
including
Convolutional
Networks
(GCN),
Attention
(GAT),
GraphSAGE,
utilized
categorize
levels
graph-based
representations
of
data.
Results:
Significant
differences
structures
observed,
increased
variability
stronger
feature
correlations
at
higher
levels.
distances
between
Impairment
Mild
(58.68,
p<0.001)
(58.28,
are
statistically
significant.
GCN
GraphSAGE
achieve
perfect
classification
accuracy
(precision,
recall,
F1-Score:
1.0),
correctly
identifying
all
instances
across
classes.
GAT
attains
an
overall
59.61%,
variable
performance
Conclusions:
Integrating
geometry,
learning,
GNNs
effectively
differentiates
AD
stages
from
The
strong
indicates
their
potential
assist
clinicians
early
identification
tracking
progression.
Язык: Английский
Automated Gluten Detection in Bread Images Using Convolutional Neural Networks
Applied Sciences,
Год журнала:
2025,
Номер
15(4), С. 1737 - 1737
Опубликована: Фев. 8, 2025
Celiac
disease
and
gluten
sensitivity
affect
a
significant
portion
of
the
population
require
adherence
to
gluten-free
diet.
Dining
in
social
settings,
such
as
family
events,
workplace
gatherings,
or
restaurants,
makes
it
difficult
ensure
that
certain
foods
are
gluten-free.
Despite
availability
portable
testing
devices,
these
instruments
have
high
costs,
disposable
capsules,
depend
on
user
preparation
technique,
cannot
analyze
an
entire
meal
detect
levels
below
legal
thresholds,
potentially
leading
inaccurate
results.
In
this
study,
we
propose
RGB
(Recognition
Gluten
Bread),
novel
deep
learning-based
method
for
automatically
detecting
bread
images.
is
decision-support
tool
help
individuals
with
celiac
make
informed
dietary
choices.
To
develop
method,
curated
annotated
three
unique
datasets
images
collected
from
Pinterest,
Instagram,
custom
dataset
containing
information
about
flour
types.
Fine-tuning
pre-trained
convolutional
neural
networks
(CNNs)
Pinterest
dataset,
our
best-performing
model,
ResNet50V2,
achieved
77%
accuracy
recall.
Transfer
learning
was
subsequently
applied
adapt
model
Instagram
resulting
78%
Finally,
further
fine-tuning
significantly
different
improved
performance,
achieving
86%,
precision
87%,
recall
F1-score
86%.
Our
analysis
revealed
performed
better
flours,
higher
scores
This
study
demonstrates
feasibility
image-based
detection
highlights
its
potential
provide
cost-effective
non-invasive
alternative
traditional
methods
by
allowing
receive
immediate
feedback
content
their
meals
through
simple
food
photography.
Язык: Английский
Emerging Technology
Gastrointestinal Endoscopy Clinics of North America,
Год журнала:
2025,
Номер
unknown
Опубликована: Апрель 1, 2025
Язык: Английский
Ulcerative Colitis, LAIR1 and TOX2 Expression, and Colorectal Cancer Deep Learning Image Classification Using Convolutional Neural Networks
Cancers,
Год журнала:
2024,
Номер
16(24), С. 4230 - 4230
Опубликована: Дек. 19, 2024
Background:
Ulcerative
colitis
is
a
chronic
inflammatory
bowel
disease
of
the
colon
mucosa
associated
with
higher
risk
colorectal
cancer.
Objective:
This
study
classified
hematoxylin
and
eosin
(H&E)
histological
images
ulcerative
colitis,
normal
colon,
cancer
using
artificial
intelligence
(deep
learning).
Methods:
A
convolutional
neural
network
(CNN)
was
designed
trained
to
classify
three
types
diagnosis,
including
35
cases
(n
=
9281
patches),
21
control
12,246),
18
63,725).
The
data
were
partitioned
into
training
(70%)
validation
sets
(10%)
for
network,
test
set
(20%)
performance
on
new
data.
CNNs
included
transfer
learning
from
ResNet-18,
comparison
other
CNN
models
performed.
Explainable
computer
vision
used
Grad-CAM
technique,
additional
LAIR1
TOX2
immunohistochemistry
performed
in
analyze
immune
microenvironment.
Results:
Conventional
clinicopathological
analysis
showed
that
steroid-requiring
characterized
by
endoscopic
Baron
histologic
Geboes
scores
expression
lamina
propria,
but
lower
isolated
lymphoid
follicles
(all
p
values
<
0.05)
compared
mesalazine-responsive
colitis.
classification
accuracy
99.1%
99.8%
cancer,
control.
heatmap
confirmed
which
regions
most
important.
also
differentiated
between
based
H&E,
LAIR1,
staining.
Additional
10
(adenocarcinoma)
correctly
classified.
Conclusions:
are
especially
suited
image
conditions
such
as
cancer;
relevant
immuno-oncology
markers
Язык: Английский