• 06.29.12

Computers Scan A Crowd, Gauging Its Mood

“Mood Meter” uses real-time facial recognition to create a live interactive display of how entire rooms and buildings are feeling.

Computers Scan A Crowd, Gauging Its Mood
Unfortunately, the real thing is just a wee bit ugly at this point.

Everyone knows what it feels like to walk into a lobby or onto a subway platform and instantly feel the “vibe” of the place–happy, sad, mad, desperate, relaxed. What are we intuiting so clearly and instantly about the mood of an entire group of people? “Mood Meter,” an installation created by MIT Media Lab, takes this ineffable experience and makes it tangible by displaying the mood of an entire public space as a real-time infographic.


Created by Javier Hernandez Rivera and M. Ehsan Hoque, Mood Meter uses cameras and facial recognition technology to log how many people in a room are smiling from moment to moment. The information is then parsed by a laptop and re-displayed as a live video feed of the room with cartoon faces superimposed over all of the inhabitants. Smiling faces are rendered as (duh) green smiley faces, while any other expressions get covered up by a yellow “neutral” face with a straight line for a mouth. It looks kind of like security camera footage remixed by John Baldessari.

Pretty cool–but what’s the point? After all, our ability to instantly sense and interpret the subtlest microexpressions on other humans’ faces has been honed by millions of years of evolution already, so what good are some primitive-in-comparison computer vision algorithms? For one thing, Mood Meter isn’t limited to one point of view like a person is. Because the system’s “intuition” comes from data, it can display the emotional profile of whole buildings by interpreting the visual signatures from many rooms and camera feeds once. (Rivera and Hoque did exactly that during the celebration of MIT’s 150th anniversary last April. Presumably there were a lot of green smiley faces on the display.)

But what’s more interesting is how the display itself can actually help change the attitudes of the people it’s watching, by creating a real-time feedback loop. “We could observe thousands of people interacting with the system, and while some people enjoyed testing the system with their friends, other people used it as a self-reflection tool to regulate their mood,” Rivera tells Co.Design. “At the end of the study we surveyed 300 people and found that people felt better when interacting with the system and when seeing other interacting with it.”

What if something like Mood Meter were installed in places where tensions tend to run high, like train stations, post offices or DMVs? The simple fact of seeing a sea of glum straight-mouthed faces on a screen could nudge people to start smiling more, if only just to see the display change. Research shows that smiling (even when you don’t feel happy) actually puts you in a better mood–which would contribute to the feedback loop.

Even those dorky cartoon faces contribute to Mood Meter’s “nudging” effect. Rivera and Hoque originally designed the faces into the system as a privacy measure, but Rivera admits that “we also wanted to encourage people to smile! When showing the neutral icon, people felt compelled to modify their face so they could get the smiling one.” Scores of surveillance cameras are already recording us in public places without our consent. If Mood Meter-like visualizations could be generated from some of that footage, at least we’d get to feel better about being watched.

[via BostInno]


About the author

John Pavlus is a writer and filmmaker focusing on science, tech, and design topics. His writing has appeared in Wired, New York, Scientific American, Technology Review, BBC Future, and other outlets.