The primary step in enrolling is capturing your biometric data for the very first time. With a fingerprint scanner, you would have pressed your finger against the sensor several times. What the system does not save is a picture of your finger. Rather, it identifies particular, quantifiable features and converts those features into a mathematical representation known as a template. That template is encrypted and saved in a database. It is vital that you enroll using good quality data, so that there is a solid template established to compare to in the future. If the initial data is low quality, the system will have difficulty recognizing you in the future. Once you have completed the enrollment process, the system is ready to verify or identify you.
There are two primary ways that a biometric system verifies or identifies an individual.
In verification (also referred to as 1:1 matching), you provide information identifying yourself (such as by entering a username) and then present your biometric data to the system. The system then compares the live scan of your biometric data with your stored template to determine whether you are who you claim to be. This is similar to how you unlock your smartphone. In identification (also referred to as 1:n matching), you do not tell the system what name you are claiming; instead, the system takes your live scan and compares it to every single template stored in its database to determine if there is a Match. This process requires more computing power than verification and typically occurs when a large amount of unidentified individuals need to be identified, such as when police use their databases to try and identify an unknown individual.
Fingeprint Scanners
As mentioned earlier, fingerprint scanners are probably the most well-known form of biometric technology. However, there are various kinds of fingerprint scanners, all of which employ different concepts to record the unique ridges and valley areas of your fingertips.
Optical scanners
Optical scanners were the first kind of scanner and are still commonly used today. Optical scanners usually rely on a light source that illuminates your finger and a camera or CCD (Charge-coupled Device) sensor that records a digital image of your fingerprint. Your fingerprint appears as darker ridges and lighter valleys. The software then processes this image to extract the unique features, such as bifurcations (where ridges meet), terminations (where ridges end), etc. These points are called minutia.
Capacitive scanners
Modern smart phones and many high-security systems use capacitive scanners. Capacitive scanners take advantage of a matrix of tiny capacitors to generate an image of your fingerprint. When your finger comes into contact with the surface, your ridges touch the sensors and therefore alter the capacitance at those points. The valleys, located farther away, produce a different level of capacitance. The scanner then maps these differences in capacitance to create a detailed “electric” image of your fingerprint. Because capacitive scanners rely on your human skin as opposed to simply recording an image of it, it provides more security than optical scanners.
Ultrasonic Scanners
Newer and less common are Ultrasonic Scanners, but they provide additional levels of security. Ultrasonic Scanners send out sound waves toward your finger and measure the reflections that rebound off of it. Various components of your fingerprint — ridges, valleys and even layers below the skin — will affect how far away and how loudly the sound waves echo back. Therefore, Ultrasonic Scanners can develop a three-dimensional map of your fingerprint, providing additional protection against spoofing attempts. Samsung has integrated ultrasonic scanning into some of their flagship models.
Extracting features and comparing templates
Irrespective of how the scanning was accomplished, the central concept is extracting those minute details. Algorithms analyze not only where these minute details occur but also what type (i.e., ending or branching) they are, and which direction they are oriented. A template may contain dozens of such minute details. After extraction of minute detail during verification, comparison is made between the minute details extracted from the new scan and those previously saved in the template. The system checks to see if enough minutiae points exist within a given percentage similarity range (for example 70-80%) before determining that there is a valid Match.
Beyond Fingerprints: additional types of biometric modality
Facial recognition
Facial recognition systems examine unique features of your face including the space between your eyes, the shape of your cheek bones and jaw line. Many early facial recognition systems used two-dimensional images, but modern versions of facial recognition use 3-d mapping, infrared dots (as seen in Apple’s Face ID), or thermal imaging to create a more secure version of your facial structure. Facial recognition systems identify key facial landmarks and transform them into a numeric value.
Iris and retinal scanning
The Iris (colored area around the pupil) contains an incredible intricate pattern of fibers, ridges, and furrows, that is consistent across a lifetime. To record this pattern, Iris scanners utilize near-infrared light to capture the patterns in your Iris and create a detailed template. Retinal scanning utilizes infrared light to capture the unique pattern of blood vessels behind your eye. Due to their great complexity and uniqueness, both Iris and retinal scanning are considered extremely accurate.
Voice recognition
Voice recognition examines both the physical attributes of your vocal cords as well as behavioral characteristics associated with speech (such as pitch, tone, speed). Voice recognition is somewhat unreliable since voice can change due to sickness, emotional states, or external noise sources. While voice recognition offers users an easy-to-use experience through virtual assistants for example, it lacks reliability for higher-level security purposes.
Biometric security concerns
Like any security mechanism, Biometrics offers advantages over traditional passwords however raises additional concerns regarding security and privacy issues. Since you cannot change your Iris or Fingerprints if they are compromised, protecting and encrypting templates securely is essential. Strong encryption and one-way hash algorithms ensure that biometric data cannot be rebuilt from stored templates. The european union’s gdpr considers biometric data to be “special categories of personal data” that require increased safeguards due to their sensitive nature.
Frequently asked questions
Can biometric systems be fooled or spoofed?
Despite their sophistication, no biometric system is completely fool-proof. Older optical fingerprint scanners could occasionally be tricked with high-resolution latent print pictures or fake fingers made from plastic materials. Modern systems employing capacitive or ultrasonic scanning techniques add liveness detection capabilities that differentiate between living tissue and fake material. For example, they might test for pulse rate, temperature changes, or micro-scale skin elasticity.
Are actual images of my fingerprint stored?
Generally no images of your actual fingerprint are stored. Biometric systems convert physical features into either mathmatical templates or digital codes, not reproducible images of your fingerprint or face. As an added layer of security, once created and stored, these templates are encrypted.
What does multi-factor authentication using Biometrics mean?
Combining at least two independent factors for authentication constitutes multi-factor authentication (MFA). In addition to utilizing Biometrics as one factor (i.e., a fingerprint scan) along with something you know (a pin) or something you possess (a security token) greatly enhances security by increasing the difficulty in compromising all factors simultaneously.
As demonstrated above biometric systems, specifically fingerprint scanners represent a valuable tool in verifying identities in our rapidly evolving digital environment. By converting our own unique biological identifiers into secure methods for accessing devices or authenticating ourselves, Biometrics represents a simple yet effective solution for users worldwide.
Sources
- Fingerprint Recognition Using Minutiae Extraction — National Center for Biotechnology Information (NCBI)
- Biometrics Program — National Institute of Standards and Technology (NIST)
- How Do Fingerprint Scanners Work? — Scientific American
- Face ID Security White Paper — Apple (PDF)
- Iris recognition — Wikipedia
