SEM Image Quality Assessment Based on Intuitive Morphology and Deep Semantic Features
IEEE Access,
Journal Year:
2022,
Volume and Issue:
10, P. 111377 - 111388
Published: Jan. 1, 2022
The
widespread
use
of
scanning
electron
microscopy
(SEM)
has
increased
the
requirements
for
SEM
image
quality.
images
obtained
by
beam
feedback
have
more
complex
texture
features
than
natural
optical
imaging,
and
this
condition
results
in
poor
performance
algorithms
used
assessing
quality
on
datasets,meanwhile,the
field
assessment(IQA)
is
mostly
aimed
at
specific
distortion
types.
In
order
to
solve
above
two
problems,to
address
rich
texture,
few
edges,
extreme
sensitivity
degree
images,
we
propose
a
semantic
IQA
(TSIQA)
method
based
sparse
mask
information
entropy
increase.
First,
construct
neural
network
containing
module
(SMM),
which
extract
intuitive
spatial
channel
domains.
Simultaneously,
growth
attention
(IGA)
introduced
into
SMM
detect
difference
between
current
past
extracting
deep
information.
assessment
experiments
datasets
show
that
compared
with
state-of-the-art
methods,
including
popular
no-reference
techniques
adapted
SEM-IQA,
TSIQA
superiority
typical
criteria.
Language: Английский
Research on distortion quality evaluation of computer network shared image based on visual sensitivity
International Journal of Wireless and Mobile Computing,
Journal Year:
2023,
Volume and Issue:
24(1), P. 27 - 27
Published: Jan. 1, 2023
Shared
image
distortion
will
affect
the
user's
experience,
and
then
damage
people's
life
entertainment
experience.In
view
of
this,
this
research
starts
with
evaluation
classification
network
shared
quality,
improves
quality
algorithm
combined
sensitive
characteristics
human
vision,
verifies
its
performance
superiority
through
comparative
experiments.The
results
show
that
some
improved
reference
algorithms
reaches
highest
values,
which
are
0.7923,
0.3224,
0.7931
0.8213,
respectively.The
non-reference
achieves
values
positive
indicators
in
comparison
0.487
0.287,
respectively,
while
lowest
value
negative
is
0.902.It
can
be
seen
conforms
to
eyes,
has
high
computational
efficiency
broad
application
prospects.
Language: Английский