A Minnesota startup was hard at work on the Oscars red carpet Sunday evening, and even though you probably didn't see them, their technology was in just about every picture.
During Oscar night, Duluth-based Runway Manhattan debuted FashionBrain, a platform that uses artificial intelligence to correctly identify fashion imagery. Runway Manhattan partnered with Reuters News Agency, which let their entire Oscars red carpet feed run through FashionBrain.
FashionBrain can currently tag around 400 terms related to apparel, hair and makeup from a fashion vocabulary of approximately 30,000 keywords developed by Runway Manhattan. These keywords include everything from "red dress" and "pink lipstick" to "dangling earrings" and "hair behind ears."
Runway Manhattan worked with BrainCreators, a Dutch AI startup, to develop FashionBrain. BrainCreators trained the AI utilizing a 500,000-photo database created by Runway Manhattan, which the Duluth company pieced together manually over the course of five years.
The platform's overall goal is to make it easier for fashion publications, brands, retailers, PR agencies and others easily search for specific styles.
"One of the most disruptive challenges in fashion publishing right now is the explosion in the sheer amount of imagery available," Runway Manhattan co-founder Markus Mueller said in a release. "Even if editors only gave one second of attention to each photo, it would take over a day to review all the relevant content, and one second is not enough time to make sound editorial decisions."
Mueller added that a platform like FashionBrain could do the same work in just a few seconds, allowing editors and writers to spend less time identifying trends and more time writing about them.
The red carpet has inspired technology in the past. Jennifer Lopez's infamous dress at the 2000 Grammy Awards played a significant role in the creation of Google Images. She wore a nearly sheer green Versace gown that plunged to her navel and was held together by fashion tape. Needless to say, everyone wanted to see it.
Years later, Eric Scmidt, executive chairman of Google, wrote: "At the time, it was the most popular search query we had ever seen. But we had no surefire way of getting users exactly what they wanted: J.Lo wearing that dress."
And thus, Google Image Search was born.
Although FashionBrain made its debut at the 2019 Oscars, Runway Manhattan has been in business for several years. The company got its start in Duluth in 2014, when its founders began tagging red carpet imagery in order to make fashion photos more easily discoverable. Since then, Runway Manhattan's five-person tech team has sorted hundreds of thousands of images, and its database is used by publications including Glamour and ELLE Italia.
The company has raised around $2 million in equity funding, Mueller told Minne Inno.
Going forward, Runway Manhattan will use the newly-launched FashionBrain for all of its services, which includes image tagging, content adaptation and story ideation based on live feedback from events.
"FashionBrain is the technological lead forward that the industry needs right now," Mueller said. "It solves not only a previously unsolved problem, but what seemed to be an increasingly unsolvable problem."