Invisible to Cameras: New Patterns Defeat Surveillance

Bill Swearingen has developed computer-generated patterns designed to defeat surveillance cameras and license plate readers, a project he calls 'noRecognition.' After extensive testing, these patterns can now be applied to clothing and objects to prevent common surveillance systems from triggering detection alerts. The technology does not block video recording but scrambles the identification algorithms, making people and objects harder for law enforcement to track. Swearingen, a cybersecurity professional, was motivated by concerns about pervasive surveillance and the desire to protect individual privacy and the right to protest without fear of being tracked.

Swearingen's work builds on previous efforts to counter surveillance technology. He refined his patterns using a reinforcement learning model that iteratively improved its ability to fool various detection algorithms. His model successfully defeated eleven open-source algorithms, including those used by Flock license plate readers and Clearview AI. A public demonstration at the Def Con cybersecurity conference in Las Vegas on Friday successfully proved the pattern's effectiveness in obscuring a vehicle from a Flock camera. The project is now seeking funding through a crowdsourcing campaign to produce merchandise featuring these anti-surveillance patterns.