pioneering generative AI in interior design
Collaboration awarded $730,000 in grant funding to build software like no other
Saskatchewan Polytechnic’s Digital Integration Centre of Excellence (DICE) and Metric Design have joined forces in an exciting collaboration to transform the interior design
industry. This homegrown AI project seeks to create cutting-edge software utilizing
digital twin and generative AI technology, revolutionizing how interior design work
is approached and executed.
“I’ve been searching for ways to automate and streamline certain elements of the interior
design process, which are often cumbersome, time-consuming and subject to error,”
says Tamara Bowman, owner and creative director at Metric Design. “While there are
project management and time-tracking tools available in the industry, none focus on
accelerating the product specification and procurement process. Our goal in working
with DICE is to build software to streamline and simplify this part of the interior
design process. As far as we know, there is no other interior design software like
this.”
With the support of DICE, Metric Design has secured significant funding to embark
on two digital projects that will be run in parallel starting this fall. The team
has been awarded multiple Applied Research and Development (ARD) grants from the Natural
Sciences and Engineering Research Council of Canada (NSERC) and Mitacs. The large
ARD grant amounts to $150,000 from NSERC and $75,000 from Mitacs per year for three
years, focusing on machine learning and generative AI systems. Additionally, a smaller
ARD grant totalling $40,000 from NSERC and $15,000 from Mitacs for up to one-year
concentrates on building software that will allow Metric Design to create a digital
twin of spaces being designed. This impressive funding totals $730,000, with Mitacs
funding supporting two student interns for the three-year AI project and one student
intern for the one-year digital twin project.
Metric Design, based in Saskatoon and renowned for its residential and commercial
interior design and renovation services, is leveraging this collaboration to create
interior design software using digital twin and generative AI technology. “There is
already software available that can scan a room and provide precise measurements.
Using digital twin technology, we will be able to scan or measure a space, then create
a digital twin of the room that allows us to initiate the design process quickly and
efficiently,” shares Bowman.
A digital twin is a virtual representation of an object or system designed to accurately
reflect the physical object. “The digital twin project will focus on the physical
layout of the room being designed, enabling us to adjust windows and walls without
the need for redesigning in a CAD/CAM program,” says Dr. Cyril Coupal, DICE research
manager.
“The digital twin will accurately replicate the client’s space, ensuring the virtual
version has the same dimensions. From this we can add materials, adjust lighting and
play around a bit with the space to accurately represent the true effects of the outcome.
The second project involves using machine learning to train an AI model to incorporate
the client’s and designer’s preferences into the design process, enabling us to use
the results to auto-select products and furniture from various vendors that will fit
the space and achieve the desired look and feel,” says Bowman. “With millions of products
available, narrowing down the options can be overwhelming. This software will also
help address that interior design challenge for both a consumer and a designer.”
By leveraging digital twin technology and generative AI, Metric Design seeks to provide
clients with multiple virtually designed options, enabling customers to visualize
and customize their spaces with ease and precision. “The new software will enable
us to make real-time adjustments within the platform to align with the client’s budget,
aesthetic and desired atmosphere. “We have a huge opportunity building this software
with DICE. I’m an interior design professional and business owner, not a tech expert.
Software can be a scary game. I feel so much more confident and secure with the opportunity
to work with a knowledgeable, local partner. I know what I’d like as a result and
I’m excited to work with DICE to get there.”
Coupal adds, “In a past life I did drafting for residential construction projects.
It’s exciting to use that experience to develop this digital platform. I’ll be leading
the DICE team to develop machine learning models that AI will use to generate the
interior design options for clients. Over the next three years we will be entering
millions of data records to train the AI system.”
The early stages of interior design often mean navigating through ambiguity, inaccuracy
and inconsistency, which can lead to costly corrections. Designers often manually
create digital prototype models for clients but this process is time-consuming, labour-intensive
and costly. The generative AI approach will speed up design efforts, facilitate better
communication of themes and concepts, and reduce or eliminate potential errors in
the final product by removing human error with manually measuring space and specifying
furniture.
“Another benefit of the program is that it will provide the client with a list of
materials, furniture options and their prices,” says Coupal. “It takes the guesswork
out of the interior design process and the costs involved.”
Metric Design’s end goal is using this technology to offer their virtual design services
nationwide. “Our new software will reduce the per-project hours, making high-quality
interior design more affordable and achievable for many,” says Bowman. “We don’t want
to remove the artist from the process, we want to give them tools to be able to do
a better job.”
“A lot of people choose not to work with a designer because it can be expensive and
the result is unknown. This new technology will make our services more affordable
and accessible,” adds Bowman. “This is an amazing opportunity and at the end of the
day, Metric Design is able to keep all of the intellectual property. That is incredible.”
Organizations partnering with DICE gain access to exceptional computational facilities,
faculty expertise, student researchers, research and development funding and a vast
network of connections. DICE, funded by NSERC and Innovation Saskatchewan, is Saskatchewan’s
first Technology Access Centre (TAC) and specializes in applied research related to
data manipulation, analysis, storage, transmission, and display. Learn more at saskpolytech.ca/dice.
link