What if you could “feel” the way clothing fit before buying it online?
As the COVID-19 pandemic continues to fuel online shopping demands, a research team from the University of Saskatchewan is developing a way to solve the growing problem of clothing returns with an avatar that “feels” how clothes fit digitally.
Computer science professor Raymond Spiteri, and his team have developed personalized avatars that can try on clothes as people shop online and detect how the clothing will actually fit the shopper.
Here’s how it works:
First, the user uploads a couple of photos of themself in tighter-fitting clothing. Data such as their height, weight, sex, and measurements is also needed to create a personalized avatar, Spiteri said. This process helps analyze how the clothes will look and fit depending on the shopper’s body type.
The next phase involves the actual garment.
“We represent the garment as a mesh of particles as well. And we can use physics to sort of tell how bendy and how stretchy the garment is, and how it’s then going to interact with your avatar,” he said.
Next, a comfort map or heat map is shown to display how the clothing actually fits. For example, if the shirt is too tight underneath the arms the heat map will display red around that area. Based on that, shoppers can determine if they need to size up or down.
While most shopping avatars involve a Photoshop-like process, Spiteri and the research team are creating a more realistic experience that mimics actually trying on clothes in-person.
The software is also able to identify the type of material in the clothing. For now, the project only involves t-shirts because it’s a simple article of clothing to work on, but there are plans to expand into other articles of clothing and shoes, according to Spiteri.
He also said the researchers are also planning on implementing a social support network so that buyers can talk about how different garments fit.
“The idea could give them not just the confidence to buy, but actually have them be satisfied with what they get and not have to return [anything].”
According to a 2019 CNBC article, 30 to 40 per cent of online clothing and shoe purchases are returned. In the same article, David Sobie, co-founder and chief executive officer of Happy Returns said, “Shoppers return 5 to 10 per cent of what they purchase in store but 15 to 40 per cent of what they buy online.”
This shopping tool could solve some of the return issues that retailers have seen in the past year and half.
The next stage for the University of Saskatchewan research group is working with retailers to get the technology out for general public use. Spiteri said the group is still in the “convincing” stage, as the software is still in development.
“The retailers are obviously not going to jump on board. We definitely have a way to go still in terms of getting the feedback from the customers and adjusting it and then having it integrated within the various retailers websites.”
A shopping avatar that provides an in-person dressing room experience could be a game changer for the online shopping world, and Spiteri and the project are being recognized because of this.
This work is one of 16 research projects that have collectively earned Spiteri the Mitacs Award for Exceptional Leadership — Professor, awarded by Mitacs. Mitacs is a national innovation organization that solves business challenges with research solutions from academic institutions.
The award will be presented at a hybrid ceremony on November 23, at the National Arts Centre in Ottawa.