One
of
the
main
challenges
rescue
operations
after
a
devastating
earthquake
is
timely
location
people
trapped
under
debris.
We
propose
system
that
exploits
smartphone
to
detect
presence
and
implicitly
interact
with
person
in
buildings.
It
leverages
phone
microphone
sound
waves
generated
by
human
breathing,
heartbeat,
movement.
analyzes
signals
on
itself
using
deep
learning.
A
server
collecting
results
can
support
search-and-rescue
or
trigger
further
actions,
such
as
an
emergency
call.
The
preliminary
evaluation
based
proof-of-concept
Android
app
demonstrate
accurate
detection
within
specific
range
smartphone.
Abstract
This
paper
presents
the
Coswara
dataset,
a
dataset
containing
diverse
set
of
respiratory
sounds
and
rich
meta-data,
recorded
between
April-2020
February-2022
from
2635
individuals
(1819
SARS-CoV-2
negative,
674
positive,
142
recovered
subjects).
The
contained
nine
sound
categories
associated
with
variants
breathing,
cough
speech.
metadata
demographic
information
age,
gender
geographic
location,
as
well
health
relating
to
symptoms,
pre-existing
ailments,
comorbidity
test
status.
Our
study
is
first
its
kind
manually
annotate
audio
quality
entire
(amounting
65
hours)
through
manual
listening.
summarizes
data
collection
procedure,
demographic,
symptoms
information.
A
COVID-19
classifier
based
on
bi-directional
long
short-term
(BLSTM)
architecture,
trained
evaluated
different
population
sub-groups
in
understand
bias/fairness
model.
enabled
analysis
impact
gender,
date
recording,
language
proficiency
detection
performance.
JMIR mhealth and uhealth,
Год журнала:
2023,
Номер
12, С. e44406 - e44406
Опубликована: Авг. 18, 2023
In
the
modern
world,
mobile
apps
are
essential
for
human
advancement,
and
pandemic
control
is
no
exception.
The
use
of
technology
detection
diagnosis
COVID-19
has
been
subject
numerous
investigations,
although
thorough
analysis
prevention
conducted
using
apps,
creating
a
gap.
Sensors,
Год журнала:
2024,
Номер
24(4), С. 1173 - 1173
Опубликована: Фев. 10, 2024
Respiratory
diseases
represent
a
significant
global
burden,
necessitating
efficient
diagnostic
methods
for
timely
intervention.
Digital
biomarkers
based
on
audio,
acoustics,
and
sound
from
the
upper
lower
respiratory
system,
as
well
voice,
have
emerged
valuable
indicators
of
functionality.
Recent
advancements
in
machine
learning
(ML)
algorithms
offer
promising
avenues
identification
diagnosis
through
analysis
processing
such
audio-based
biomarkers.
An
ever-increasing
number
studies
employ
ML
techniques
to
extract
meaningful
information
audio
Beyond
disease
identification,
these
explore
diverse
aspects
recognition
cough
sounds
amidst
environmental
noise,
detect
symptoms
like
wheezes
crackles,
voice/speech
evaluation
human
voice
abnormalities.
To
provide
more
in-depth
analysis,
this
review
examines
75
relevant
across
three
distinct
areas
concern
diseases’
symptoms:
(a)
detection,
(b)
(c)
diagnostics
speech.
Furthermore,
publicly
available
datasets
commonly
utilized
domain
are
presented.
It
is
observed
that
research
trends
influenced
by
pandemic,
with
surge
COVID-19
diagnosis,
mobile
data
acquisition,
remote
systems.
Informatics in Medicine Unlocked,
Год журнала:
2022,
Номер
32, С. 101049 - 101049
Опубликована: Янв. 1, 2022
The
goal
of
this
paper
is
to
classify
the
various
cough
and
breath
sounds
COVID-19
artefacts
in
signals
from
dynamic
real-life
environments.
main
reason
for
choosing
than
other
common
symptoms
detect
patients
comfort
their
homes,
so
that
they
do
not
overload
Medicare
system
therefore
unwittingly
spread
disease
by
regularly
monitoring
themselves.
presented
model
includes
two
phases.
first
phase
sound-to-image
transformation,
which
improved
Mel-scale
spectrogram
approach.
second
consists
extraction
features
classification
using
nine
deep
transfer
models
(ResNet18/34/50/100/101,
GoogLeNet,
SqueezeNet,
MobileNetv2,
NasNetmobile).
dataset
contains
information
data
almost
1600
people
(1185
Male
415
Female)
all
over
world.
Our
most
accurate,
its
accuracy
99.2%
according
SGDM
optimizer.
good
enough
a
large
set
labelled
may
be
used
check
possibility
generalization.
results
demonstrate
ResNet18
best
stable
classifying
tones
restricted
dataset,
with
sensitivity
98.3%
specificity
97.8%.
Finally,
shown
more
trustworthy
accurate
any
present
model.
Cough
study
promising
put
extrapolation
generalization
test.
Frontiers in Public Health,
Год журнала:
2023,
Номер
11
Опубликована: Июль 4, 2023
Background
Artificial
intelligence
(AI)
is
a
broad
outlet
of
computer
science
aimed
at
constructing
machines
capable
simulating
and
performing
tasks
usually
done
by
human
beings.
The
aim
this
scoping
review
to
map
existing
evidence
on
the
use
AI
in
delivery
medical
care.
Methods
We
searched
PubMed
Scopus
March
2022,
screened
identified
records
for
eligibility,
assessed
full
texts
potentially
eligible
publications,
extracted
data
from
included
studies
duplicate,
resolving
differences
through
discussion,
arbitration,
consensus.
then
conducted
narrative
synthesis
data.
Results
Several
methods
have
been
used
detect,
diagnose,
classify,
manage,
treat,
monitor
prognosis
various
health
issues.
These
models
conditions,
including
communicable
diseases,
non-communicable
mental
health.
Conclusions
Presently
available
shows
that
models,
predominantly
deep
learning,
machine
can
significantly
advance
care
regarding
detection,
diagnosis,
management,
monitoring
different
illnesses.
Scientific African,
Год журнала:
2023,
Номер
21, С. e01757 - e01757
Опубликована: Июнь 10, 2023
The
severe
acute
respiratory
syndrome
coronavirus
2
(SARS-CoV-2)
virus's
worldwide
pandemic
has
highlighted
the
urgent
need
for
reliable,
quick,
and
affordable
diagnostic
tests
comprehending
controlling
epidemic
by
tracking
world
population.
Given
how
crucial
it
is
to
monitor
manage
pandemic,
researchers
have
recently
concentrated
on
creating
quick
detection
techniques.
Although
PCR
still
preferred
clinical
test,
there
a
pressing
substitutes
that
are
sufficiently
rapid
cost-effective
provide
diagnosis
at
time
of
use.
creation
simple
POC
equipment
necessary
home
testing.
Our
review's
goal
an
overview
many
methods
utilized
identify
SARS-CoV
in
various
samples
utilizing
portable
devices,
as
well
any
potential
applications
smartphones
epidemiological
research
detection.
point
care
(POC)
employs
range
microfluidic
biosensors
based
smartphones,
including
molecular
sensors,
immunological
biosensors,
hybrid
imaging
biosensors.
For
example,
number
tools
been
created
COVID-19,
theories.
Integrated
devices
can
be
using
loop-mediated
isothermal
amplification,
which
combines
amplification
with
colorimetric
Electrochemical
approaches
regarded
substitute
optical
sensing
techniques
utilize
fluorescence
being
more
beneficial
Minimizing
simplicity
used
detection,
together
amplify
DNA
or
RNA
under
constant
temperature
conditions,
without
repeated
heating
cooling
cycles.
Many
virus
data
visualization,
making
these
user-friendly
broadly
distributed
throughout
nations.
Overall,
our
provides
review
different
novel,
non-invasive,
affordable,
efficient
identifying
COVID-19
contagious
infected
people
halting
disease's
transmission.
Frontiers in Digital Health,
Год журнала:
2025,
Номер
6
Опубликована: Янв. 28, 2025
Introduction
Artificial
intelligence
(AI)
models
trained
on
audio
data
may
have
the
potential
to
rapidly
perform
clinical
tasks,
enhancing
medical
decision-making
and
potentially
improving
outcomes
through
early
detection.
Existing
technologies
depend
limited
datasets
collected
with
expensive
recording
equipment
in
high-income
countries,
which
challenges
deployment
resource-constrained,
high-volume
settings
where
a
profound
impact
health
equity.
Methods
This
report
introduces
novel
protocol
for
collection
corresponding
application
that
captures
information
guided
questions.
Results
To
demonstrate
of
Voice
EHR
as
biomarker
health,
initial
experiments
quality
multiple
case
studies
are
presented
this
report.
Large
language
(LLMs)
were
used
compare
transcribed
(from
same
patients)
conventional
techniques
like
choice
Information
contained
samples
was
consistently
rated
equally
or
more
relevant
evaluation.
Discussion
The
HEAR
facilitates
an
electronic
record
(“Voice
EHR”)
contain
complex
biomarkers
from
voice/respiratory
features,
speech
patterns,
spoken
semantic
meaning
longitudinal
context–potentially
compensating
typical
limitations
unimodal
datasets.
