What Really Happens When You Get Your Fingerprints Scanned

The use of fingerprints as a form of identification dates back to the reign of Hammurabi in ancient Babylon. Nearly 4,000 years later, fingerprints are still a vital form of identity authentication, but a bit has changed in the intervening years. Instead of leaving our fingerprint on a contract, we mostly use them to unlock our phones and occasionally verify our identity at the bank or airport. How these fingerprint scans work can vary widely, and there's even some variability in how our digitized fingerprints are analyzed and compared.

When it comes to unlocking your smartphone with a fingerprint, one of three techniques is generally used. The most common is optical fingerprint scanning, in which your phone takes a closeup picture of your fingerprint. Another common technique is capacitive scanning, which uses a surface (similar to a touchpad) to record the patterns in electrical charge caused by the ridges and valleys in your fingerprints. Finally there are ultrasonic scanners which use sound waves to create a sort of 3D echogram of your fingerprint and even your pores.

Each method has its own benefits and drawbacks, with some, like optical and capacitive, being more cost-effective, and others, like ultrasonic, being more reliable under more diverse conditions. It's also important to remember that none of them is immune from being hacked. But once the fingerprints are recorded, what happens on the back end to match two fingerprints?

The act of using your fingerprint to unlock your phone begins with a process called "enrollment" where your phone takes multiple scans of your fingerprint to build an internal template. This template isn't an idealized image of your fingerprint; instead, it records the details of your fingerprint. On the most basic level, fingerprints can be divided into classes, but computers are looking at smaller details, like where a ridge splits or ends. These details are called minutiae.

When your phone takes a scan of your fingerprint, it first converts it into a black-and-white image before isolating the minutiae data. Because you're not likely to take a perfect scan of your finger every time, matching algorithms don't look for a perfect match between fingerprints. Instead, they look at clusters of minutiae. If enough clusters of minutiae match, they likely belong to the same finger.

The future of fingerprint analysis might not rely entirely on minutiae analysis. A team of researchers used AI to uncover patterns common between all of a person's fingerprints. In other words, given a print from a single finger, they were able to predict if another print from a different finger belonging to the same individual. Don't expect this kind of technology to show up on your phone just yet, though — just know that it's in the pipeline.