The Faces of Drunk People Are Riddled with Hotspots and Shame


In the future, when all police are thermal-scanning robots and all private citizens are real people who maybe like to sneak a bottle of zinfandel and a sleeve of Thin Mints into a Thursday matinee every week and get a little wasted, the police bots will be able to tell all the lonely matinee-goers from the sober and productive citizens simply by glancing at them. That’s because, thanks to enlightening new research from the University of Patras in Greece, drunkenness can be detected by temperature changes in a person’s face and, as I’m sure I don’t have to tell you, police bots in the future will all have a single, lidless infrared eye, the better for peering into your imperfect mortal soul.

To determine that drinking leaves a slug-trail of hotspots all over a person’s face, researchers ordered 20 participants to drink an 11-ounce glass of beer every 20 minutes. All in all, participants had four beers, and after each beer, researchers took a sequence of infrared pictures of their faces. Two methods emerged for determining drunkenness, the first of which involved using an algorithm to compare the infrared photos to a database of photos of drunk and sober people. Since drunk people have hotspots on their faces, researchers could tell the drunk people from sober people pretty easily. The second method used a new algorithm to analyze the temperature differences on certain parts of the face, spotting drunk people by how hot their noses and mouths were compared to their foreheads.

Researchers then happily concluded that such technology could be used in airport scanners to immediately determine who could keep buying anxiety-deadening booze for transatlantic flights, and who was already so drunk that they’d have to hand over all of their cash to the closest Sbarro’s, where they could enjoy a reheated broccoli pizza and think about their poor life choices.

Drunks detected by thermal camera [LiveScience via CBS]

Image via Olga Sapegina/Shutterstock.

Inline Feedbacks
View all comments
Share Tweet Submit Pin