No More Needle in a Haystack

Imagine walking down the street, surrounded by surveillance cameras that can detect your every move, and feeling like a needle in a haystack. You're not alone. With the rise of facial recognition and license plate readers, it's becoming increasingly difficult to keep a low profile. But what if there was a way to make yourself invisible to these cameras?

Blocking Detection, Not Recording

Bill Swearingen has been working on a solution for over a year, and his project, noRecognition, is making waves in the cybersecurity community. Rather than blocking surveillance cameras from recording video footage, Swearingen's patterns scramble the camera's ability to identify objects, people, or faces, preventing detection alerts.

Swearingen's patterns don't block cameras from recording, but they do make it difficult for the cameras to identify what they're recording. This means that individuals can opt out of being tracked without having to worry about being caught on camera.

"Privacy is a fundamental right," Swearingen said in a call with TechCrunch. "I want to allow people to opt out of being tracked."

A Year of Research and Testing

Swearingen's journey began with a proof-of-concept test lab, where he incrementally defeated one open-source video camera detection algorithm after another. Over time, he refined his patterns by scaling up his tests with additional computer processing power.

His proof of concept evolved into a reinforcement learning model, which trained itself on which patterns work and which don't against specific camera algorithms. This allowed Swearingen to create patterns that could defeat multiple algorithms at once.

• Some of the 11 open-source detection algorithms that Swearingen's patterns can defeat include:
  • Flock license plate readers
  • Axon body-worn cameras
  • Cameras running Clearview AI
• Swearingen's model creates new patterns every minute, each batch mathematically better than the last.

Real-World Test

On Friday at the Def Con cybersecurity conference in Las Vegas, Swearingen ran his first real-world test. With help from Donut Media, the test involved covering a 2009 Toyota Yaris with one of Swearingen's newest patterns to see if the car would be invisible to detection by a Flock camera.

"We proved it was effective," said Swearingen, though the wheels were a challenge. The video of the demo will be out in the next few weeks, said Donut Media.

Next Steps

The noRecognition project has a crowdsourcing campaign to help fund the sale of early merchandise featuring the patterns, from T-shirts to hoodies, with the potential for pattern-printed skins for vehicles down the line. Swearingen's aim is for the patterns to be high-quality and resolution good enough to work from a distance, while also looking aesthetically fashionable.

He's keeping his strongest patterns off the internet to prevent the camera makers from defeating them, but the work is not yet done. His models are continuing to grind out new patterns.

"Every failure improves my model, and so [the patterns] keep getting better and better," he said.