TN Note: Young technocrats in training are developing real-world systems to identify a person and then “read” his emotional state. Homeland Security, DARPA and the FBI are already hot on the trail of this technology. The obvious problem is that if the software errs, the error will still be attached to your profile. Stuff like this could mark you for the rest of your life.
Bill Hedgcock knows it sounds a little creepy.
Tucked into the white ceiling tiles, the ceiling camera he had installed at the Pappajohn Business Building at the University of Iowa scans the faces of all who pass under it and instantly calculates their moods — collecting readings for joy, frustration, confusion, fear, anger and sadness.
“We nicknamed it ‘the creepy study,’ because we just wanted to be out about it, just so everyone’s aware,” said Hedgcock, an associate professor of marketing. “It sounds creepy.”
The facial encoding data is part of a larger student research project underway in the Tippie College of Business that uses automated technology to read emotions by measuring slight movements in the facial muscles, such as a movement of the eyebrow or a widening of a lip.
Hedgcock believes the University of Iowa is the world’s only business school with this kind of real-time software that converts images of people’s faces into readings for different moods. If it works, experts see huge potential for facial encoding in the worlds of marketing, advertising and political campaigns.
But that’s no sure bet. Hedgcock makes it clear that the technology’s accuracy remains unproven — it’s one reason why his students have spent months picking it apart, getting a better understanding of the technology’s limitations.
If his students eventually deem the technology a stinker, it’s no big deal, Hedgcock says. But if the process is proven and adopted by marketing firms, his students will have a leg up when they go on the job hunt.
“If it does work, they walk out there and they know something that no one else knows how to do,” he says. “It’s not written in a textbook. It doesn’t exist anywhere.”
Hedgcock wanted to emphasize privacy with his students, so the camera doesn’t record video or images of the people it measures. It simply kicks out data measuring what it sees, and it sees a lot — millions of rows of data so far.
Students have used the mood measurements in a variety of ways. From the first-floor camera, they crunched the data to see if the weather affects mood and if mood affects food sales at the business school’s snack bar.
Do sunny skies make people happier? Do people purchase more caffeine when they’re down?
But students say the best use of the technology isn’t in a generic space such as a hallway on campus. People walking to and from classes or meetings aren’t exactly the most expressive.
Students say the technology is likely most effective with a more specific aim, such as judging reaction to a short advertisement or a political debate — occasions more likely to elicit an emotional response.
Students paired up with Frank N. Magid Associates, a market research firm with offices in Marion, Ia. The company is known for its work gauging viewer sentiment about local television stations and their on-air personalities.
While it may never fully replace more conventional methods of judging consumer preference such as surveys and dial tests, Magid leaders say they believe facial encoding could be a supplement to their current suite of measurement tools.
wpDiscuz