A small but useful material about the features of the types of fingerprint sensors available on the market.
Fingerprint sensors today have gone beyond the premium segment of smartphones; additional hardware protection technology can be implemented even in relatively inexpensive mid-range devices. Since its introduction to the market, the technology has undergone significant evolutionary changes, so here is an overview of the fingerprint sensors on the market, indicating the differences between them.
Optical scanners
The oldest way to capture and compare fingerprints. As the name suggests, the technology is based on an optical image, essentially photography, and uses special algorithms to identify unique patterns on a surface, such as bumps or unique markings, by analyzing the lightest and darkest areas in the image.
By analogy with cameras in smartphones, such sensors have a specific resolution, the higher it is, the finer details will be available for processing by the scanner, which will increase the level of protection. However, such sensors receive more contrasting images than a conventional camera. They usually include a large number of diodes per inch to display details more clearly up close. The scanner is in the dark when the finger is scanned, so optical scanners also have 'on board' LEDs that act like a flash during scanning. Such an internal arrangement will give the smartphone additional millimeters of thickness and negatively affect the final form factor.
The main disadvantage of optical scanners is their unreliability. With their help, only a two-dimensional image is obtained; such a scanner can be 'deceived' with another good quality image or with an artificially created print from it. You should not trust this type of scanner, it is not secure enough to protect the most important information.
Today, fingerprint sensors in smartphones come in a variety of shapes and sizes, but they don't have optical scanners. By analogy with the beginning of the spread of resistive touch screens, optical scanners today can be found only in the most inexpensive hardware solutions. The need for increased security has led to the unanimous transition of smartphones to capacitor scanners.
Capacitor scanners
The most common type of fingerprint sensors. And again, the name gives out the main component, if you, of course, know a little about electronics – a capacitor. Instead of creating a traditional image of a fingerprint, capacitor scanners use arrays of tiny capacitors to collect information about the fingerprint. Connecting capacitors capable of retaining an electrical charge to the conductive board allows them to be used to read the details of the print. The charge in the capacitors will change slightly when your finger touches the board, and at the same time the air gap will leave the charge relatively unchanged. To track the changes, the integration circuit of the operational amplifier is used, subsequently the changes can be recorded with a signal converter from analog to digital.
After scanning, the digital information can be analyzed for distinctive and unique characteristics of the fingerprint, which can be saved for later comparison. Such a sensor is much more difficult to 'trick' than an optical one. The results cannot be reproduced in the image and are very difficult to counterfeit with any artificial print: different materials will cause different changes in the capacitor charge. The only security risk could come from the possibility of a software or hardware breach.
By creating a sufficiently large array of such capacitors (hundreds, if not thousands of capacitors in one scanner), it is possible to obtain an image of the bumps and grooves of a fingerprint in high detail using only electrical signals. By analogy with optical sensors, more capacitors will give a higher scanner resolution and increase protection to a certain level.
Because of the more components in the circuit, capacitor scanners can be more expensive. Some early versions attempted to trim the number of capacitors needed by using 'swipe' scanners, which received information from fewer capacitor elements by rapidly updating results as you swipe your finger over the sensor. The method was quite sophisticated and often required several attempts for a successful scan. Fortunately, today a simpler scheme of operation of the sensor is common: a simple press and hold is enough.
Ultrasound scanners
The latest fingerprint technology, first introduced in the Le Max Pro smartphone. Qualcomm and Sense ID technology played an important role in it. To actually collect the details about the print, the hardware platform includes an ultrasonic transmitter and receiver. An ultrasonic pulse is transmitted through a finger placed on the scanner. It is partially absorbed, partially transferred back to the sensor, depending on the bumps, pores and other details unique to each print.
No microphone is provided to pick up the returning signal; instead, a sensor is used that can read mechanical stress to calculate the intensity of the returned signal at different locations on the sensor. Scanning over a longer period of time allows additional information to be read, which in turn can provide a detailed 3D model of the scanned print. The 3D nature of the technology makes it an even safer alternative to capacitor scanners.
Algorithms and Cryptography
Most fingerprint sensors are based on very similar principles, but additional components and software can play a major role in differentiating products in terms of performance and functionality available to consumers.
The physical scanner is accompanied by a dedicated microcircuit that interprets the scanned information and transmits it in the required format to the smartphone processor. Different manufacturers use slightly different algorithms for identifying key characteristics of a fingerprint in terms of speed and accuracy.
Typically, these algorithms 'look for' where the bumps and lines end or where the bump splits in two. Collectively, these and other distinctive features are referred to as a fingerprint template or detailed fingerprint capture protocol. If several of these features match in a scanned print, the print will be counted as matched. Rather than comparing the whole fingerprint each time, comparing template features reduces the amount of processing power required to identify a fingerprint, avoids smearing errors, and also allows you to scan an off-center finger or just a portion of the fingerprint.
Undoubtedly, such information should be safely stored on the device and kept away from code that could compromise it. Instead of uploading user information to the network, ARM processors can securely store it on a dedicated physical chip using their TrustZone-based Trusted Execution Environment (TEE) technology. . This secure vault is also used for other cryptographic processes and communicates directly with secure hardware components such as a fingerprint sensor to prevent any eavesdropping attempts by software. Approved non-personal information such as a password can only be accessed by applications using the TEE client API.
Sample authentication scheme
A similar solution from Qualcomm is built into the Secure MSM architecture, Apple calls such a project 'Secure Enclave', but they are all based on the same principle – storing information on a separate part of the processor that cannot be accessed by applications running in a normal operating system environment. The Fast Identity Online (FIDO) alliance has developed strong cryptographic protocols that allow these hardware-protected zones to be used for authentication between hardware and services without a password. Therefore, you can enter the site or online store using your fingerprint, and your personal information will not leave the smartphone. This is achieved by transferring digital keys to the server, rather than biometric information.
Fingerprint sensors have become a fairly secure alternative to remembering countless passwords and usernames, and the further development of secure mobile payment systems means these scanners will become more prevalent and essential security tools in the future.
Original material by Rob Trigs
Elir: biometric identification in gadgets is just starting to gain momentum and a fingerprint sensor is just the first step. It's nice that manufacturers are thinking about security, improving existing technologies and complementing them with hardware solutions. From the material I understood the reason for the not entirely successful, in my opinion, implementation of the fingerprint scanner in my SGN 4, perhaps the only inconvenience in use. But more or less reliably hidden from prying eyes. In any case, the presence of such technology in the device is now very desirable, at least that's how it's presented. Even omitting marketing bravado, it can be noted that biometric identification has great potential and development in this direction will be active. Is a fingerprint scanner a must-have or just another consumer gimmick?