New t-shirt design aimed at confusing and improving AI-based CCTV systems

If you enjoy peering into kaleidoscopes or “getting the most out of your experience” at jam band gigs, chances are you’ve seen patterns like the one Northeastern University computer Xue Lin created. with researchers from IBM and the Massachusetts Institute of Technology.

But, unlike those recreational experiences, that pixelated mix of blue, yellow, purple, and green slapped on the front of a T-shirt has a practical application: It might be enough to trick video surveillance into knowing if it’s even a person wearing it.

Lin calls it a “Contradictory T-Shirt” because that’s the technical name for data designed to trick systems that use artificial intelligence to quickly detect patterns in large datasets, like someone wearing a hat. red in a crowd of people.

Editor’s Note: You’ll need to click on the Globe article to see the t-shirt design for yourself.

In this case, the conflicting data is the patterns on the shirt. Some AI systems find these patterns so confusing that they cannot recognize the wearer as a human.

Let’s hope no one wears the shirt at InfoComm 2020 next June in Las Vegas when Solutions360 President Brad Dempsey says he hopes to have an AI solution tailored to AV enterprises ready for prime time. .

The truth is that the shirt will not protect the wearer from facial recognition software, such as Amazon’s Rekognition system. But it could provide a defense against other types of surveillance, such as object recognition systems that identify people based on clothing and body shape.

A person wearing Lin’s T-shirt might not be recognized as human at all.

“When a person wears it, it will just disappear,” Lin said. The Boston Globe.

How the contradictory t-shirt works

The contradictory T-shirt works because AI vision systems don’t see things the way humans do. All they see are pixels: points of light they’ve learned to interpret as cars, mailboxes, or human bodies. Add enough unexpected pixels to the mix and the machines don’t know what they’re looking at.

Lin and a Northeastern graduate student, Kaidi Xu, designed the shirt to fool a common object recognition program called Yolo, an open source program that anyone can download and test.

The researchers found that when showing images of people wearing the shirt, Yolo failed to identify the wearers as people 63% of the time. Lin has not tested his shirt against other object recognition programs, such as the Avigilon system.

Since all of these programs work on the same general principles, it’s likely that they, too, will have trouble seeing the shirt – and the person wearing it.

Lin tells The Globe that she is not trying to help defeat these systems; she tries to make them even better, until they’re nearly impossible to fool.

“We’re trying to explore the vulnerability of these neural networks,” Lin said, “and hopefully we can fix that.”

I have a huge collection of T-shirts, probably enough to wear a different one every day for at least six months, but I’ll be adding this one to my collection as soon as possible. I’m all for keeping a low profile and playing with technology. And you?

About Catharine C. Bean

Check Also

How to print a t-shirt design at home using Rare Custom

T-shirts are a simple way to show off your personality and style. They’re also a …