Pretty soon, marketers will be able to autonomously classify and target your social identity, using little more than an Instagram photo.
Researchers at the University of California San Diego have developed an algorithm that essentially scans photos of individuals to determine which subculture they belong to. It's like something out of Minority Report.
Specifically, the program examines an array of physical characteristics, such as hair style, jewelry, attire, tattoos, even the texture of clothing, while also scanning the environment of a given photo—such as a dance club, formal event, or bar.
With this data, the algorithm can then "automatically classify" (we prefer the term "judge") subjects into the following labels: Biker, Country, Goth, Heavy-Metal, Hip-Hop, Hipster, Raver, Surfer. It’s not difficult to imagine how such information could be used to create a kind of “consumer taxonomy” for marketers.
"Little work has been done to automatically classify images of people into social categories," the report reads. "We tackle this problem by analyzing pictures of groups of individuals and creating models to represent them. We capture the features that distinguish each subculture and show promising results for automatic classification."
Alarmingly dubbed "Visual Recognition of Urban Tribes," the concept is still in an early phase—with a success rate that’s less than 50 percent (what, exactly, determines "success," we don’t know). But the computer scientists working on the project are pretty clear about the possibilities:
"This points to an excellent opportunity for computer vision to interact with other fields, including marketing strategies and psychological sociology."
Some of the categories are a bit archaic ("heavy-metal," really?), and we can’t help but note the irony in using label-eschewing fashion styles to, well, label. But as a 21st century person you’d be silly not to expect this. Facebook alone receives more than 300 million photos per day, in addition to the wealth of personal information users willingly offer to social networks. Why wouldn’t marketers try to exploit that data?
You can’t be incensed about something like this if you’re the type who Instagrams 18 photos of your breakfast and then "likes" Exxon-Mobil on Facebook.
But, ah shucks, now I’m the one who’s labeling.