Design Transformation Pathways for AI-Generated Images in Chinese Traditional Architecture
Electronics,
Год журнала:
2025,
Номер
14(2), С. 282 - 282
Опубликована: Янв. 12, 2025
This
study
introduces
a
design
transformation
model
for
AI-generated
Chinese
traditional
architectural
images
(SD
Lora&Canny)
based
on
Stable
Diffusion
(SD).
By
integrating
parameterization
techniques
such
as
Low-Rank
Adaptation
(Lora)
and
edge
detection
algorithms
(Canny),
the
achieves
precise
restoration
of
form,
color
elements,
decorative
symbols
in
architecture.
Using
Beijing
Drum
Tower
experimental
subject,
statistical
analysis
software
(SPSS
V28.0)
was
employed
to
conduct
quantitative
evaluation
comparative
generated
by
DALL-E,
MidJourney,
SD,
SD
Lora&Canny
models.
The
results
demonstrate
that
significantly
outperforms
generation
tools
accuracy
visual
fidelity.
Finally,
this
applied
create
digital
cultural
product
AR
Bell
Fridge
Magnet,
showcasing
its
practical
application
creation
verifying
innovative
potential
preservation
transmission
Язык: Английский
Integrating Multimodal Generative AI and Blockchain for Enhancing Generative Design in the Early Phase of Architectural Design Process
Опубликована: Июль 17, 2024
AI
advances
integrate
generative
design
tools
in
architecture,
providing
architects
with
sophisticated
options.
It
enables
the
creation
of
intricate,
high-performing
projects
by
exploring
diverse
possibilities
and
algorithms.
Generative
empower
to
create
better-performing,
sustainable,
efficient
solutions
explore
possibilities.
This
paper
leverages
multimodal
enhance
creativity
combining
textual
visual
inputs.
Blockchain
technology
converts
metadata
into
NFTs,
ensuring
secure,
authentic,
traceable
data
storage.
The
framework
addresses
ownership,
legal
adherence,
client-architect
collaboration
is
entirely
scalable
for
digital
authentication.
research
exemplifies
pragmatic
fusion
blockchain
applied
architectural
more
transparent,
effective
results.
study
provides
a
strategy
that
uses
technologies
achieve
an
creative
workflow
early
stages
design.
Язык: Английский
Developing an Urban Landscape Fumigation Service Robot: A Machine-Learned, Gen-AI-Based Design Trade Study
Applied Sciences,
Год журнала:
2025,
Номер
15(4), С. 2061 - 2061
Опубликована: Фев. 16, 2025
Generative
AI
(Gen-AI)
revolutionizes
design
by
leveraging
machine
learning
to
generate
innovative
solutions.
It
analyzes
data
identify
patterns,
creates
tailored
designs,
enhances
creativity,
and
allows
designers
explore
complex
possibilities
for
diverse
industries.
This
study
uses
a
Gen-AI
generation
process
develop
an
urban
landscape
fumigation
service
robot.
proposes
machine-learned
multimodal
feedback-based
variational
autoencoder
(MMF-VAE)
model
that
incorporates
readily
available
spraying
robot
dataset
includes
considerations
from
various
research
efforts
ensure
real-time
deployability.
The
objective
is
demonstrate
the
effectiveness
of
data-driven
approaches
in
generating
specifications
with
targeted
requirements
autonomous
navigation,
precision
spraying,
extended
runtime.
comprises
three
stages:
(1)
parameter
fixation,
emphasizing
functionality-based
aesthetic-based
specifications;
(2)
specification
using
proposed
MMF-VAE
without
dataset;
(3)
development
based
on
generated
specifications.
A
comparative
analysis
evaluated
impact
dataset-driven
generation.
proved
more
feasible
optimized
real-world
deployment
integration
inputs
iterative
feedback
refinement.
prototype
was
then
constructed
model’s
parametric
constraints
tested
actual
scenarios
validate
operational
viability.
highlights
transformative
potential
robotic
workflows.
Язык: Английский
Multi-Objective Optimization Design for Cold-Region Office Buildings Balancing Outdoor Thermal Comfort and Building Energy Consumption
Energies,
Год журнала:
2024,
Номер
18(1), С. 62 - 62
Опубликована: Дек. 27, 2024
Performance
parameters
and
generative
design
applications
have
redefined
the
human–machine
collaborative
relationship,
challenging
traditional
architectural
paradigms
guiding
process
toward
a
performance-based
transformation.
This
study
proposes
multi-objective
optimization
(MOO)
approach
based
on
performance
simulation,
utilizing
Grasshopper-EvoMass
platform.
The
Non-dominated
Sorting
Genetic
Algorithm
II
(NSGA-II)
is
applied
to
coordinate
two
metrics—outdoor
thermal
comfort
building
energy
loads—for
of
design.
results
indicate
that
(1)
workflow
established.
Compared
baseline
design,
optimized
form
shows
significant
improvement
in
performance.
Pareto
optimal
solutions,
under
2022
meteorological
conditions,
demonstrate
an
annual
efficiency
16.55%,
outdoor
neutrality
ratio
increases
by
1.11%.
These
suggest
effectively
balances
loads
comfort.
(2)
A
total
1500
solutions
were
generated,
from
which
16
selected
through
front
method.
resulting
layouts
provide
multiple
feasible
configurations
for
early-stage
phase.
Язык: Английский