The launch of the new iPhone 5S has brought fingerprint scanning back to the forefront of security technology, but this isn’t the first time that we’ve talked about fingerprint security and smartphones. Remember that the fact that the Motorola Atrix shipped with a similar piece of technology all the way back in 2011.
Anyway, let’s take a look at how the technology works. There’s two main ways that you can digitally capture your fingerprints, either via an optical scanner or by using clever capacitor circuits.
If you wanted to compare fingerprints, the obvious place to start might be with a picture. That’s essentially how optical scanners work, as at the center of an optical fingerprint scanner is the same charge couple device used in digital cameras.
A charge couple device is a light sensor system that consists of light sensitive diodes called photosites, which are responsible for generating electrical signals when they detect light. Different levels of light produce different levels of charge in each photosite, and each photosite diode is a single pixel of the completed image.
The number of these little photosites on the sensor will determine the resolution of the image generated, which in turn determines how accurately the scanner can differentiate between fingerprints. In other words, the higher the PPI sensor the higher the level of security.
For a point of reference, the iPhone 5S’ sensor is 500ppi squeezed onto a tiny chip, which is a higher concentration of pixels than the HD screen on the Galaxy S4 (441ppi).
But in order to detect any levels of light there has to be a light source, after all sunlight won’t be any good once your finger is covering the sensor. So arrays of LEDs are used to illuminate the finger, which then reflects light back to the sensor where it can be recorded. The ridges and curves will all reflect light slightly differently, allowing for minute details to be picked up and recorded. Interestingly, CCD systems actually generate an inverted image, with darker areas representing more reflected light. But this isn’t really important for scanning fingerprints as we won’t be seeing the image.
The different levels of charge in each pixel is then converted into a digital value by an analog-to-digital converter, which allows the processor to make sense of the image data. These values can then be stored and compared when a user re-scans their finger to unlock their device.
There’s also plenty of additional post-processing that goes on in order to determine the accuracy and clarity of the image before it’s compared to any stored fingerprint, but we’ll talk about that a little later.
The second type of fingerprint scanner uses capacitors and op-amp circuits to generate electrical signals, rather than light. These capacitance scanners require a number of separate electronic components just to scan a small area, and aren’t quite so widely used due to the expense involved in building them.
This scanner, as the name implies, is based on capacitor circuits. Capacitance measures the ability of an entity to store an electrical charge, and these scanners rely on the use of a fingerprint to alter the capacitance of the circuit.
We won’t go into too much detail about the electronics, but there are a couple of important things to note in order to understand how this works.
Firstly the amplifier. Amplifiers work by adjusting an output signal based on the two input signals. The amplifier in the scanner is arranged to create an integrator circuit using a capacitor formed from two conductor plates separated by an insulating layer.
For those not familiar with calculus, integration means that the op-amp circuit will generate an output voltage proportional to the magnitude and duration that an input voltage signal has deviated from the default volts. So when the capacitance of the input is changed the charge will also change, producing a slight current and voltage that can be measured by the integrator circuit.
Here’s an example of how it works. At the start the reset button is pressed, shorting the capacitors. Once released the capacitor charges up and balances the circuit, giving us a point of reference for any change. When a finger is placed in close proximity to the conducting plates (capacitor) it alters the capacitance of the circuit, which then adjusts the input and therefore also the output voltage of the amplifier. The closer the finger is to the conductor the more the capacitance is affected. If you imagine this applied multiple times across an entire finger, you will end up with thousands of different voltages for each peak and valley of the fingerprint. In the end this produces a picture which is quite similar to an optical scanner.
Whilst this might seem like a lot more hassle and more expensive than an optical camera, it is a lot more secure as it can’t be fooled so easily by fakes. Optical scanners can be fooled, as it aims to only capture the light and dark areas produced by a fingerprint, whilst a capacitance scanner requires a real fingerprint shape and real ridges and grooves in order to create the correct electrical signal.
On top of all that hardware, there’s plenty of additional software processing to be done as well. Complex algorithms are used to analyse and recognize certain patterns and features of a person’s fingerprint for comparison, rather than trying to match two fingerprints exactly.
This is done by looking for key features on the fingerprint, such as where ridges and lines end, or where a ridge splits in two. Collectively, these and other distinctive features are sometimes called minutiae or typical. If a scanned fingerprint matches several of these minutiae then it will be considered a match. This helps reduce the amount of processing power required to identify each fingerprint, helps avoid errors if the scanned fingerprint is smudged, and also allows the finger to placed off-centre or be identified with only a partial print.
Although the security offered by fingerprint scanners might be somewhat questionable, it’s an interesting technology that we could well see more of in the smartphone market in the near future. Especially with the drive towards combining more of our personal and financial information to single accounts, tying all of this to your biometric data seems like a logical step.