Auto-WCEBleedGen Version V1 and V2: Challenge, Datasets and Evaluation
Misa Hub,
Palak Handa,
Divyansh Nautiyal
и другие.
Authorea (Authorea),
Год журнала:
2024,
Номер
unknown
Опубликована: Март 10, 2024
In
this
document,
we
provide
an
overview
of
the
Auto-WCEBleedGen
Version
V1
and
V2.
The
challenge
was
organized
virtually
by
MISAHUB
(Medical
Imaging
Signal
Analysis)
in
collaboration
with
8th
International
CVIP
2023
(Conference
on
Computer
Vision
Image
Processing)
from
August
15-November
11,
2023.
V2
is
being
Язык: Английский
VCE-AnomalyNet: A New Dataset Fueling AI Precision in Anomaly Detection for Video Capsule Endoscopy ⋆
Authorea (Authorea),
Год журнала:
2024,
Номер
unknown
Опубликована: Апрель 23, 2024
Video
capsule
endoscopy
(VCE)
is
a
minimally
invasive
diagnostic
technique
that
helps
in
the
detection
of
various
anomalies
like
polyps,
ulcers,
aphthae,
etc,
within
intestinal
lumen.
Due
to
high
no.
frames
VCE
and
low
doctor-to-patient
ratio
across
globe,
inspection
time
about
2-4
hours.
Research
has
shown
Artificial
Intelligence
(AI)
potential
decrease
reading
improve
upon
false-positive
rates.
However,
lack
AI
data
big
hindrance
it.
To
address
this
issue,
we
present
VCE-AnomalyNet
Dataset,
new
dataset
fueling
precision
anomaly
for
VCE.
The
comprises
108,832
accurately
labeled
with
bounding
box
annotations
YOLO
(You
Only
Look
Once)
format.
These
have
been
compiled
from
multiple
open-source
datasets,
aiming
support
research
automatic
available
at
Dataset
(zenodo.org)
.
Язык: Английский
Localization and Semantic Segmentation of Polyp in an Effort of Early Diagnosis of Colorectal Cancer from Wireless Capsule Endoscopy Images
S Jothiraj,
Jayanthy Anavai Kandaswami
Опубликована: Ноя. 25, 2022
Cancer
is
characterized
by
the
fast
growth
of
aberrant
cells
that
affect
adjacent
tissues.
Colorectal
cancer
could
be
diagnosed
in
early
stage
identifying
predecessor
polyp
are
initially
innocuous.
Endoscopy
aids
identification
real
time
monitoring.
Polyps
obscured
mucosa
surrounds
it
lumen
colon,
making
visual
differentiation
from
difficult
for
physicians
thereby
increasing
miss
rates.
Recognizing
colorectal
polyps
challenging
as
varies
widely
characteristics
representing
its
features.
With
emergence
deep
learning
techniques
especially
convolutional
neural
network
an
effort
was
made
to
detect
and
segment
polyps.
U-net
architecture
with
capability
learn
features
images
proposed
our
paper
semantic
segmentation
where
colon
localized.
Polyp
Kvasir
dataset
obtained
using
wireless
capsule
endoscopy
provides
advantage
viewing
entire
gastrointestinal
tract
used
this
paper.
The
framework
evaluated
assessing
performance
optimized.
predicted
results
produced
accuracy,
precision,
sensitivity
(recall),
IoU
f1
score
93.14%,
98.08%,
95.55%,
95.75%
98.01%
respectively.
Язык: Английский
Open-Source Datasets for Colonoscopy Polyps and Its AI-Enabled Techniques
Lecture notes in networks and systems,
Год журнала:
2023,
Номер
unknown, С. 63 - 76
Опубликована: Янв. 1, 2023
Язык: Английский
CNN Architecture-Based Image Retrieval of Colonoscopy Polyp Frames
Palak Handa,
Rishita Anand Sachdeva,
Nidhi Goel
и другие.
Lecture notes on data engineering and communications technologies,
Год журнала:
2023,
Номер
unknown, С. 15 - 23
Опубликована: Янв. 1, 2023
Язык: Английский
Wireless Capsule Endoscopy using Localization Techniques over IMU Sensor and Side-wall Cameras
Опубликована: Дек. 16, 2022
Predicting
the
performance
of
wireless
capsule
endoscopy
in
human
ileum
has
been
a
challenging
topic
for
over
decade.
Considering
its
compact,
coiled,
and
elongated
shape,
this
makes
sense.
This
paper
suggests
sensor-lens
hybrid
as
solution
to
these
issues
through
multisensory-aided
WCE
localization.
The
success
connection
is
quantified
here
by
RSSI.
It
recommended
use
Siamese
Caps
Net
camera-based
end
result
that
accurate
estimates
are
possible
thanks
method.
One
novel
approach
involves
verifying
Receiver's
new
location
based
on
Round
Trip
Time,
Communication
Delay,
Received
Signal
Strength
Indication,
Distance,
Last
Known
Coordinates.
Matlab
R2019b
then
used
calculate
results.
results
demonstrate
suggested
method
outperforms
state-of-the-art
methods
terms
Localization
Accuracy,
Standardized
Root
Mean
Square
Error,
Average
Translation
Error.
Язык: Английский