Figuring out size and cut of clothes through a website can suck the fun out of shopping online, but Revery.ai is developing a tool that leverages computer vision and artificial intelligence to create a better online dressing room experience.
Under the tutelage of University of Illinois Center for Computer Science advisrr David Forsyth, a team consisting of Ph.D. students Kedan Li, Jeffrey Zhang and Min Jin Chong, is creating what they consider to be the first tool using existing catalog images to process at a scale of over a million garments weekly, something previous versions of virtual dressing rooms had difficulty doing, Li told TechCrunch.
Revery.ai co-founders Jeffrey Zhang, Min Jin Chong and Kedan Li. Image Credits: Revery.ai
California-based Revery is part of Y Combinator’s summer 2021 cohort gearing up to complete the program later this month. YC has backed the company with $125,000. Li said the company already has a two-year runway, but wants to raise a $1.5 million seed round to help it grow faster and appear more mature to large retailers.
Before Revery, Li was working on another startup in the personalized email space, but was challenged in making it work due to free versions of already large legacy players. While looking around for areas where there would be less monopoly and more ability to monetize technology, he became interested in fashion. He worked with a different adviser to get a wardrobe collection going, but that idea fizzled out.
The team found its stride working with Forsyth and making several iterations on the technology in order to target business-to-business customers, who already had the images on their websites and the users, but wanted the computer vision aspect.
Unlike its competitors that use 3D modeling or take an image and manually clean it up to superimpose on a model, Revery is using deep learning and computer vision so that the clothing drapes better and users can also customize their clothing model to look more like them using skin tone, hair styles and poses. It is also fully automated, can work with millions of SKUs and be up and running with a customer in a matter of weeks.
Its virtual dressing room product is now live on many fashion e-commerce platforms, including Zalora-Global Fashion Group, one of the largest fashion companies in Southeast Asia, Li said.
Revery.ai landing page. Image Credits: Revery.ai
“It’s amazing how good of results we are getting,” he added. “Customers are reporting strong conversion rates, something like three to five times, which they had never seen before. We released an A/B test for Zalora and saw a 380% increase. We are super excited to move forward and deploy our technology on all of their platforms.”
This technology comes at a time when online shopping jumped last year as a result of the pandemic. Just in the U.S., the e-commerce fashion industry made up 29.5% of fashion retail sales in 2020, and the market’s value is expected to reach $100 billion this year.
Revery is already in talks with over 40 retailers that are “putting this on their roadmap to win in the online race,” Li said.
Over the next year, the company is focusing on getting more adoption and going live with more clients. To differentiate itself from competitors continuing to come online, Li wants to invest body type capabilities, something retailers are asking for. This type of technology is challenging, he said, due to there not being much in the way of diversified body shape models available.
He expects the company will have to collect proprietary data itself so that Revery can offer the ability for users to create their own avatar so that they can see how the clothes look.
“We might actually be seeing the beginning of the tide and have the right product to serve the need,” he added.
Article From & Read More ( Virtual dressing room startup Revery.ai applying computer vision to the fashion industry - TechCrunch )https://ift.tt/2UGkCUt
Fashion
Bagikan Berita Ini
0 Response to "Virtual dressing room startup Revery.ai applying computer vision to the fashion industry - TechCrunch"
Post a Comment