Analysis of the principles of biometric technology

The so-called biometric technology refers to the technology that uses each person's unique biological characteristics to verify the user's identity, also known as biometric authentication. "In theory, biometric authentication is the most reliable way to authenticate identity because it directly uses a person's physical characteristics to represent each person's digital identity. Different people have different biological characteristics, so it is almost impossible to be counterfeited and copied." Biometric technology is mainly a technology that authenticates identity through measurable biological characteristics such as body or behavior; and biometrics refer to unique physiological characteristics or behaviors that can be measured or automatically identified and verified. Shao Yu introduced that biometrics are divided into two categories: physical characteristics and behavioral characteristics. Physical characteristics include: fingerprints, palm shape, retina, iris, human body odor, face shape, blood vessels in the hand, bones and DNA, etc.; behavioral characteristics include: signature, voice, walking gait, etc.

Next, focusing on the general principles and analysis of biometric technology, Shao Yu will dissect and analyze the six most common and mature biometric technologies one by one.

1. Fingerprint recognition technology

Fingerprints are innate physical characteristics of human beings. After about 14 years old, everyone's fingerprints have been fixed. Fingerprints are "fixed" and will not change as people continue to grow. Fingerprints are also "unique". Two different people will not have the same fingerprints. Since the famous British anthropologists Gudmund and Francher published an experimental report in 1888, proving this characteristic with scientific methods, countries around the world have successively set up fingerprint storage cabinets. After more than 60 years, no completely identical fingerprint cards have been found.

As early as 7000-6000 BC, the ancient Assyrians and Chinese were aware of the characteristics of fingerprints and used fingerprints as a symbol of identity. In the mid-19th century, the scientific study of fingerprints began. Computers were used to automatically identify fingerprints in the 1960s, and the Automatic Fingerprint Identification System (AFIS) for criminal investigations gradually began to be widely used around the world. At that time, fingerprint collection generally used ink printing. After 1980, with the invention of personal computers and optical fingerprint collectors, fingerprint recognition technology began to enter some non-judicial fields, such as resident ID cards. After 1990, the emergence of cheap fingerprint collectors and computing equipment solved the problem of fast and accurate matching algorithms, and fingerprint recognition technology moved towards individual-based applications. Since then, people have studied fingerprint recognition technology and have a further understanding of the nature of fingerprints. Now, fingerprint recognition has developed to the stage of liveness recognition technology and has been widely used in various fields such as security, attendance, and computer information identity authentication.

Fingerprint recognition mainly includes the processes of capture, calculation, transmission, and verification. The principles are as follows:

① Capture: First, use the fingerprint scanner to capture the fingerprint pattern and other related features. The newly obtained image has a lot of noise. This is mainly caused by work and environment; for example, the finger is dirty, the finger has knife wounds, scars, marks, dryness, moisture or tearing, etc. Then perform image enhancement, reduce noise, and enhance the contrast of ridges and valleys. And design a suitable and matching filter and appropriate threshold for the operations involved in processing fingerprint images.

② Calculation (algorithm): Use program calculation and statistics to find out the relevant feature points of the fingerprint that have the so-called "different for each person and unchanged for life" and digitize them. The digitized data naturally still has the fingerprint's "different for each person and unchanged for life" feature.

③Transmit: The digitized fingerprint features are transmitted on the computer using various methods. Regardless of the transmission method or encryption and decryption method used, the unique characteristics are still retained.

④ Verify: The transmitted data is then processed by a program to verify its similarity with the comparison data in the database. If the statistical similarity reaches a certain level, it means that this is the fingerprint data transmitted by the person himself.

Therefore, as long as the above principles are met, there is no loss in any conversion, and the comparison value is above a certain level, it can be confirmed to be the person's fingerprint.

Depicted below is the process of using a biometric system.

(1) Capturing selected biometric features;

(2) Processing biometrics, extracting and registering biometric templates;

(3) storing templates in a local repository, a central repository, or in a portable token such as a smart card;

(4) Scan selected biometrics on site;

(5) Processing biometrics and extracting biometric templates;

(6) The scanned biometric template is matched with the stored biometric template;

(7) Providing a matching score for the application;

(8) Record relevant security audit clues of system usage.

It is scientifically proven that these lines are different in patterns, breakpoints and intersections, and are unique and permanent. This is an important scientific basis for supporting fingerprint biometrics. Fingerprint recognition usually uses the comparative recognition method, which is to use a fingerprint collector to convert the characteristic points of the finger (breakpoints, bifurcation points, isolated points) into an encrypted binary file that can be recognized by the computer and stored in the computer. During each identity authentication, as long as the newly collected fingerprint is compared with the fingerprint file in the computer, the identity authentication of anyone can be completed in a very short time.

Fingerprint recognition technology has the advantages of being easy to obtain, highly practical, easy to increase reliability, fast and convenient fingerprint scanning, and wide application. Taking Yaliute's products as an example, the fingerprint recognition technology product series includes: fingerprint door locks, fingerprint attendance machines, fingerprint safes, fingerprint mice, fingerprint collectors, fingerprint USB flash drives, fingerprint hard drives, network fingerprint authentication platform systems, etc.