Walmart is releasing an app that uses AI to let you try on clothes at home. It’s part of a growing effort to use computers to take the guesswork out of online apparel shopping. “While the virtual try-on trend has been growing in fashion for some years, rudimentary solutions allowed customers to see very roughly how a garment would look on their body, but not how it would fit,” Vadim Rogovskiy, CEO of 3DLOOK, a company which produces virtual try-on software, said in an email interview. “Now, solutions are using AI to calculate customers’ precise body measurements, so customers can discover whether it’s likely to fit their unique body, offering a convenient, cost-efficient and environmentally-friendly way to try on clothing online.”
Virtual Fitting Rooms
One frustration of online shopping for clothes is understanding how an item will look on you before you buy. But Walmart says it has the answer to this problem with its virtual fitting room. The technology, called Zeekit, is available on the Walmart app and Walmart.com. You start with the Choose My Model experience, which allows you to choose from 50 models between 5'2" – 6'0" in height and sizes XS – XXXL. To understand how an item will look on them, customers can refer to the model who best represents their height, body shape, and skin tone. “Zeekit was built with a vision to provide every person the chance to see themselves in any item of clothing found online, and that is a vision we share,” Denise Incandela, executive vice president of apparel and private brands at Walmart US, wrote in the news release. AI solutions could also help reduce waste. Online retail sales have grown by an average of 18.6 percent annually for the past five years and are expected to climb by another 33 percent by 2025. This growth coincides with an increase in ‘bracketing,’ where customers purchase items in multiple sizes, styles, and colors before returning those that don’t fit right or look good, Rogovskiy said. “Essentially, with no way to try before they buy, almost two-thirds of shoppers are using free return policies to transform their homes into personal fitting rooms,” he added.
AI for Clothes
Retailers use AI software to understand your body shape. The software typically combines 3D scans of many people to predict your profile, either from a couple of images or from a few measurements, Jim Downing, CEO of Metail, a fashion technology company that makes virtual fitting room tech, told Lifewire in an email interview. AI could also predict whether you’re likely to keep a garment in a particular size option, Downing said. “This area will often use a predicted body shape and your fit preference as inputs and bring that together with historical transaction data from a retail site and metadata about specific garments to predict which size you’re most likely to keep,” he added. Determining what the garment will look like on you is another area where AI is used, Downing said. The latest augmented reality (AR) technology also utilizes AI to track your body as you move around and changes a model of the garment to match it, so it looks like you’re wearing the clothes. There’s still room for improvement in the virtual dressing room, though. Most of the current AI solutions manipulate images to make it look like the clothes are being worn by your avatar or a model who’s close to your body shape. Because they aren’t based on the underlying garments’ cutting patterns and fabric properties, they’re not accurate enough to show you exactly how a specific garment size option would fit, Downing explained. The growing adoption of 3D CAD software among fashion brands could eventually make the process even more accurate. Metail is working on a technique called EcoShot, which allows brands to create pattern-accurate model photography of 3D garments on real models. “These are typically used during the design process when brands are revising and selecting designs, but are being increasingly used in e-commerce to create diverse model photography in a scalable way,” Downing said. In the future, AI could even serve as an online sales assistant, Rogovskiy suggested, “offering customers tailored style recommendations within their virtual fitting room, based on their unique body size and shape.”
