Metalenz’s Futuristic Smartphone Camera for Stronger Biometrics

Facial recognition on your phone might seem secure—until a mask or 3D-printed face lets someone else unlock it. In our latest interview, we got hands-on with Metalenz, the company using polarisation-sensitive optics to deliver what may become the industry standard for deepfake-proof face ID.

Standard face ID systems rely on intensity images: they see light, shape, and colour, but not what the light actually does. And that’s a problem. Today’s face recognition can be tricked by masks, glasses, or images, all of which can mimic the shape of a real face. That’s where Metalenz comes in.

How PolarID Compares to FaceID

Their PolarID camera doesn’t just look at your face—it reads the polarisation of the light reflecting off it. This unique signal can’t be faked. Lars Johnsson from Metalenz uses his identical twin counterpart ‘Larry’ to demonstrate the difference between polarisation signals between a real person and a silicon mask. Pictured is him having a sip :P.

Polarisation tells the sensor both the shape and the material properties of what it’s seeing. Human skin, with its sweat, blood vessels, and subdermal layers, polarises light differently from plastic or paper. That’s the secret to building deepfake-proof face ID.

Nanostructural Optical Sensor

Light waves aren’t just random—they have direction. When they hit and reflect off a surface, the angle of polarisation changes depending on the material. Most image sensors ignore this. But Metalenz uses metasurface optics: nanostructured silicon that can sort light based on polarisation before it even hits the sensor. That allows them to extract this hidden layer of information and feed it into a machine learning model trained to detect real, live human faces.

Why does this matter? Because attackers are getting smarter. Biometric spoofing isn’t just a Hollywood trope. From bank apps to secure emails and corporate systems, your face is becoming the new password. Deepfake-proof face ID means those systems can finally tell the difference between a real user and a high-quality fake.

Lens Construction

What’s most impressive is that this breakthrough tech from Metalenz, spun out of Harvard, fits into a selfie camera module that costs under $5. These flat lenses are constructed using a similar process to how silicon wafers are constructed. It can integrate into phones, laptops, access control terminals, and even AR glasses—without taking up more space or battery than existing sensors.

And it’s not some distant vision. The first devices featuring deepfake-proof face ID are expected to launch as early as 2026. So whether you’re securing your smartphone or an unmanned gate, Metalenz’s polarisation-based camera is poised to redefine biometric security for the next generation.