Imagine you’re scrolling through Instagram when you spot one of your favorite celebrities wearing a gorgeous jacket. You fall in love with it instantly and desperately scan the post for tags or outfit credits to no avail. In the past, you’d be left daydreaming about that enviable outfit, with no way to track down similar items to add to your own wardrobe. Enter EyeStyle, founded by Jie Feng, ‘17SEAS.
Using a single photo of your desired garment, Feng’s Chrome extension employs machine learning and visual search algorithms to comb through millions of online product databases, tracking down the exact product or similar alternatives available for online purchase.
Feng is currently in his final year of the Columbia Computer Science PhD program, with a particular focus on computer vision and machine learning. The concept of EyeStyle arose in part from his work for Amazon A9, the product search branch of the world’s most-used online retailer. “We have witnessed this great advancement with the introduction of deep learning — this development has given way to a new opportunity for these algorithms to be applied to commercial products,” says Feng.
Although Amazon has already implemented a visual search function, Feng notes that its utility is limited to “very distinctive objects” like a book cover or record sleeve. Because EyeStyle has a narrower scope, however, the tool is capable of matching user images and products with a greater much degree of precision.
With this focus on fashion, Feng hopes to “revolutionize the future of shopping” by creating an intuitive interface for empowered consumers. Currently, online shoppers are forced to scroll through endless product pages to find an item they want to buy, which can be tedious — a negative experience that leads to lower revenues for retailers. “Our vision is to simultaneously help shoppers find what they want and help retailers to better understand how they can build a stronger relationship with their customers,” Feng explains.
EyeStyle was recently accepted into NYC Media Lab’s Combine program, an effort on the part of New York-based faculty, entrepreneurs, and media executives to commercialize media technologies from universities. “I found Combine to be a really great resource,” says Feng. “With this program, we want to broaden our perspective by speaking with potential customers and coming to understand their needs.” Combine provides mentorship, community oversight, and a small amount of initial grant funding to its 2-5 person teams, all with the aim of helping its constituent startups succeed.
Feng views EyeStyle’s New York location as an asset, particularly because of its vibrant fashion community and tech-forward mentality. Next on the horizon, Feng aims to use the EyeStyle software to develop a retailer-facing product that would suggest similar styles based on items previously viewed by a customer, or even automatically populate item description information.
If you’re interested in spending less time searching for the perfect garment, consider adding EyeStyle to your browser today.