Enhancing Semantic Scene Segmentation for Indoor Autonomous Systems Using Advanced Attention-Supported Improved UNet
Research Square (Research Square),
Год журнала:
2024,
Номер
unknown
Опубликована: Июль 2, 2024
Abstract
This
paper
introduces
EFFB7-UNet,
an
advanced
semantic
segmentation
framework
tailored
for
Indoor
Autonomous
Vision
Systems
(IAVSs)
utilizing
the
U-Net
architecture.
The
employs
EfficientNetB4
as
its
encoder,
significantly
enhancing
feature
extraction.
It
integrates
a
spatial
and
channel
Squeeze-and-Excitation
(scSE)
attention
block,
emphasizing
critical
areas
features
to
refine
outcomes.
Comprehensive
evaluations
using
NYUv2
Dataset
various
augmented
datasets
were
conducted.
study
systematically
compares
EFFB7-UNet's
performance
with
multiple
encoders,
including
ResNet50,
ResNet101,
MobileNet
V2,
VGG16,
VGG19,
EfficientNets
B0-B6.
findings
reveal
that
EFFB7-UNet
not
only
surpasses
these
configurations
in
terms
of
accuracy
but
also
highlights
effectiveness
scSE
block
achieving
superior
results.
Without
depth
information,
achieves
12\%
improvement
mean
Intersection
over
Union
(mIOU).
notable
enhancement
adaptability
across
different
domains,
implying
substantial
progress
reliability
Intelligent
(IAVS)
technologies.
Язык: Английский
Fine-tuned depth-augmented U-Net for enhanced semantic segmentation in indoor autonomous vision systems
Journal of Real-Time Image Processing,
Год журнала:
2024,
Номер
22(1)
Опубликована: Дек. 6, 2024
Язык: Английский