During its re: MARS 2019 conference keynote today, Amazon announced StyleSnap, an AI-powered feature arriving in the Amazon App that helps users buy clothes they found in images. The system requires an input image, such as a screenshot, that features a particular fashion ‘look’ the user likes. StyleSnap uses this image to find clothing available for sale on the Amazon platform, paying close attention to the finer details.
Amazon’s mobile customers can access StyleSnap by opening the Amazon App, then tapping the ‘StyleSnap’ option located in the upper right-hand corner. The tool requires an image: it can be a picture you snapped of an outfit you saw in a magazine while in a waiting room, a photo of a sign you passed while out walking, a screenshot you took from a TV show, or any number of other situations in which the specific clothing items remain a mystery.
The AI powering StyleSnap analyzes the outfit featured in the image and then finds a selection of similar clothing items available for sale on Amazon. The company says its system looks at a variety of factors when choosing which items it presents, such as the price range, brand, and customer reviews.
Deep learning and computer vision make the system as reality, Amazon explains. Different settings and poses won’t impact the tool’s ability to find similar clothing items. Individual clothes are sorted into the most relevant categories — Amazon uses examples like flannel shirts and “fit-and-flair dresses.”
Online shopping has exploded in popularity with only two categories being the obvious exceptions: food and clothing. Though Amazon and Walmart are both working hard at normalizing the online grocery shopping experience, many consumers continue to hesitate when it comes to online clothing purchases. The huge variation in sizes, as well as low product quality and the dubious origins of many clothing products, make online clothes shopping a struggle.