• 1 minute Read

Sign of the Times: Camo That Outwits Facebook’s Facial Recognition

Face-detection software is everywhere. Defeat it with some avant-garde makeup!

Sign of the Times: Camo That Outwits Facebook’s Facial Recognition

Adam Harvey‘s thesis project at New York University’s Interactive Telecommunication Program is very Philip K. Dick: in a world where there’s some computer-assisted lens watching you almost everywhere, use the inherent dumbness of facial recognition algorithms against them. How? With artfully-applied, avant-garde face makeup called CV Dazzle. It works like a charm against Facebook’s built-in face-detecting bots, too:

CV Dazzle is named after the ingenious warship-camouflage designs deployed during World War I. Rather than actually hiding military vessels, the bold, jagged paint jobs made it difficult for naval rangefinders to discern details about the ship’s size, heading, armament, and so forth. Or at least, that was the idea — unlike Harvey’s digital version, the original Dazzle’s effectiveness was never proven. Of course, CV Dazzle benefits from the fact that face-detection software is much stupider than WWI seamen: you only have to apply a few strokes of face-paint to confuse it, rather than coat your whole kisser.


Harvey’s collaboration with DIS Magazine outlines the do’s and don’ts of Dazzling: avoid enhancing the eyes (mascara and eyeliner=bad), obscure the area where your nose, brow, and forehead intersect (apparently this region figures heavily in face-detecting algorithms), and don’t apply so much that it looks like a full-fledged mask (then you’ll probably be flagged for other reasons).

Oh, and don’t forget to be fabulous while you’re at it! So far, there’s still no algorithm for detecting that.

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.