When it comes to disguises, silly mustaches and fake noses won’t cut it anymore.
As facial recognition capabilities grow more sophisticated, cameras and algorithms can to do more with less. Even grainy images, like those you might find on a gas station surveillance camera, can hold enough information to match a face to a database. But there are ways to hide.
Your face is garnering a lot of interest these days. Police departments use facial recognition systems to identify criminals. Facebook knows your friends’ faces. Facial recognition is being incorporated into billboards to display ads based on the sex of the person looking at them.
In the not-so-distant future, your face might replace your wallet—a smile will serve as your identification card and credit card. Amazon plans to eliminate the checkout line at its new brick-and-mortar grocery store concept, Amazon Go, in Seattle. How? According to the company’s website:
Our checkout-free shopping experience is made possible by the same types of technologies used in self-driving cars: computer vision, sensor fusion, and deep learning. Our Just Walk Out Technology automatically detects when products are taken from or returned to the shelves and keeps track of them in a virtual cart.
How computer vision will work at Amazon Go isn’t exactly clear, but your face may play an important role in the shopping experience. Right now, the store is only open to Amazon employees, but it is expected to open to the public sometime in 2017.
Given all this attention, there may come a day when we want to avoid this kind of computer recognition.
Adam Harvey, a Berlin-based artist, is developing a line of clothing and accessories aimed at disrupting facial recognition software by fighting fire with fire. His forthcoming HyperFace project is a set of wearable patterns that overloads facial recognition software with images of faces to distract from the real person hiding behind it, exploiting weaknesses in the technology.
Faces In A Crowd
A facial recognition system keys in on dominant features and parses them into numeric sequences that are calculated according to the parameters of the algorithm. By crunching the numbers, it can to determine whether it’s “seeing” a human being or not, and who that face belongs to. Some algorithms need no more than 100 pixels—2.5 percent of an Instagram photo—to identify 78 relevant facial characteristics, said Harvey in a 2016 talk at the Chaos Communication Congress.
Harvey’s designs are collages that mimic basic facial features, sending a barrage of information that obscures a real face. Theoretically, worn as a shirt, scarf or shawl, patterns (pictured above) should protect your identity from nosy algorithms.
HyperFace is an extension of Harvey’s NYU thesis project to thwart face detectors with makeup and hair gel. In addition to hairstyles that covered the face, concealing the “T-zone”—the area around the bridge of the nose and eyes—seems to be most important.
Other designers have proposed similar concepts, such as Swiss graphic artist Simone C. Niquille. Her Realface Glamouflage collection is a series of shirts with images of celebrities meant to confuse facial recognition software in the same way. A different, less aesthetic, concept relies on glasses with near-infrared LED’s to blind cameras and hide your face.
These designs might be the opening salvos in an arms race between those who wish to scan your face and those hoping to remain hidden.
Still ahead are the discussions about ethics and boundaries that will shape a society where technology can pick anyone out of a crowd.