Adobe Sneaks: Changing outfits in eCommerce with AI

As more people are staying home, online activity is seeing a surge. Online sales of apparel for instance (driven in part by promotions) have increased 34 per cent between March 12 and April 11 (per aggregated data from Adobe Analytics). And while categories like pants have fallen 13 per cent, more comfortable clothes in the pyjama category have risen 143 per cent.

For years, apparel brands have had to make themselves more accessible beyond the physical store. Initial forays into eCommerce have evolved into more sophisticated uses of AI and data, as well as content. This move was unilateral, spanning fast fashion to luxury retailers. Nike as an example, used a set of apps (e.g. SNKRS) to drive affinity for the brand online. In the case of Prada, a 2019 collaboration with Adobe focused on using data and AI to get a better grasp of what consumers needed online.

The digital storefront had become table stakes. For many now, it is a means of survival. And once the world returns to a more normalised state and consumer spending resumes, we expect that brands will continue investing in their digital strategy. We also believe that they will use technology in new ways, to be more efficient while meeting consumer demand.

We are showcasing “Project Clothes Swap” in Adobe Experience Manager, to show how AI can take different outfits and move them around on different models online. To illustrate how this works, imagine a brand that has a repository of models that have been photographed in the past. When new styles come in, only the clothes have to be photographed. Via Adobe Sensei, the AI will automatically form clothing on a model and deliver all the variations a brand needs.

While we don’t expect physical photoshoots to go away, there will be instances where the time and cost savings can be compelling – especially when the right technology is available. The underlying AI in Project Clothes Swap, a patented method called “SieveNet,” was created to remove distortions that are characteristic of other techniques. It aligns with Adobe’s long history of working with graphics and AI that has been trained to understand the nuances in composition, textures and the like.

Project Clothes Swap can also play a helping hand when it comes to promoting diversity in the images we see online. A recent study has shown that over 75 percent of female shoppers prefer brands that feature a variety of ethnicities in their ads. In the same way that one model can showcase different outfits, the reverse is possible as well.

As with many AI technologies, the use cases can be broadened and customized depending on the brand. On the consumer side for instance, Project Clothes Swap can be leveraged in scenarios where shoppers want to upload their own photo and see how a piece of clothing looks on them. The same goes for putting different pieces of clothing and accessories together one on model, which the AI can help place accurately.

by Eric Matisoff, analytics & data science evangelist, Adobe

Leave a Comment

Your email address will not be published. Required fields are marked *

Related Posts