Basic concepts and principles of biometric technology

To master a subject, the first thing to do is to have a proficient grasp of its basic concepts and basic principles, and then repeatedly think about its connotation and extension. It is not a simple matter to explain a concept clearly. It may not only use the concepts of this subject, but also restore it to outside the subject. For example, if we study economics, then to explain the interest theory clearly, it is not enough to simply say "interest is the reward obtained by the owner of funds for lending money", but also to restore it to outside the economic terms. For example, Fisher's interest theory "not only ignores inflation, risk, and transaction costs, but more importantly, it ignores currency. He believes that in a society without currency, interest still exists when goods are exchanged for goods. The existence of interest does not require currency, but requires a market, and the exchange of goods is the market." This returns to the origin of the concept - the market. When explaining the market, it is necessary to restore it to outside economics. Of course, after crossing the boundaries of economics, we do not need to explain it anymore, because it has exceeded the scope of economics.

Although I personally think that it is very questionable whether biometrics can become a subject, I personally think that it is still very necessary to master the relevant knowledge and understand its basic concepts.

The following is a preliminary list of ten basic concepts related to biometrics. It is inevitable that some are omitted, and the explanation here is difficult to reach a truly in-depth level. I just hope to arouse your interest in thinking:

Biometrics (Biometric Authentication):

Biometrics (Biometric Identification Technology): It is the use of human biological characteristics for identity authentication. It is based on the two points that (1) human biological characteristics are different and (2) they can be measured or automatically identified and verified. Human biological characteristics include physiological characteristics or behavioral patterns. Physiological characteristics include hand shape, fingerprints, face shape, iris, retina, pulse, auricle, etc. Behavioral characteristics include signature, voice, keystroke strength, etc.

Question: 1. Should biometrics be called person identification or human body identification? 2. Do different biometric technologies have more similarities or differences? Should these technologies be completely independent?

Fingerprint identification:

Fingerprint identification is a technology that uses human fingerprint characteristics to authenticate human identity. Among all biometric technologies, fingerprint technology is the most mature and widely used.

Question: 1. Is fingerprint identification technology really mature? How should those who cannot collect fingerprints be included in the relevant system? 2. In addition to fingerprint recognition, what recognition technology has the most promising development prospects?

Algorithm

Algorithm is a computer software term, which mainly refers to the specific steps and methods required to complete a task. That is to say, given the initial state or input data, after a limited number of operations by the computer program, the required result can be obtained. Algorithms often contain repeated steps and some comparisons or logical judgments. If an algorithm is defective or not suitable for a problem, executing this algorithm will not solve the problem.

The algorithm mentioned in the field of biometrics is essentially the application of software algorithms in this field. Taking the fingerprint recognition algorithm as an example, its core algorithms include fingerprint matching algorithm, fuzzy fingerprint image processing algorithm, fingerprint feature classification, positioning, extraction algorithm, and fingerprint splicing algorithm.

Question: 1. Should the algorithm be the core of biometrics? 2. Chinese companies frequently win good rankings in fingerprint algorithm competitions. Does it mean that we have reached the world level in this regard?

Fingerprint Sensor In general, a sensor is a device that converts the measured value into an electrical signal output that has a certain relationship with it and is easy to process with a certain accuracy for the purpose of measurement. The fingerprint sensor is a device that measures the electrical signal of the finger. At present, there are different types of fingerprint sensors, such as optical and crystal capacitive. Question: What are the main technical challenges of fingerprint sensors for Chinese companies? Identity authentication (identity recognition) is simply a way of authenticating people who want to obtain a certain identity through some means or methods. For example, you can use your ID card or account number/password to authenticate whether you are a legitimate user of this system. Question: How many identity authentication technologies are there at present? What are their classification standards? AFIS (Automatic Fingerprint Identification System) AFIS (Automated Fingerprint Idenification System) refers to the computer processing of input fingerprint images to achieve fingerprint classification, positioning, extraction of morphology and detail features, and then comparison and identification of fingerprints based on the extracted features. Question: AFIS traditionally refers to fingerprint recognition systems specifically used in the police field. Is this term still necessary today? Information security In general, information security is to ensure the effectiveness and security of electronic information. Information security involves the confidentiality, integrity, availability, and controllability of information. Confidentiality is to resist passive attacks from opponents and ensure that information is not leaked to unauthorized persons. Integrity is to resist the active attack of the opponent and prevent the information from being tampered with without authorization. Availability is to ensure that information and information systems are indeed used by authorized users. Controllability is to implement security monitoring of information and information systems. Question: 1. Should the application focus of biometric technology be in the field of information security? 2. In the future, should information security technology lead biometric technology or should biometrics lead information security technology? Encryption Encryption technology mainly refers to a technical means used to protect data from leakage. It mainly converts important data into garbled code (encryption) for transmission, and then restores (decrypts) by the same or different means after reaching the destination. Encryption technology includes two elements: algorithm and key. The algorithm is the step of combining ordinary text (or understandable information) with a string of numbers (key) to produce incomprehensible ciphertext. The key is an algorithm used to encode and decode data. Question: 1. Is the file encryption function necessary for fingerprint recognition applications? 2. Does its existence itself indicate that the simple fingerprint recognition application is too simple to meet user needs? Fingerprint attendance The meaning of fingerprint attendance is very clear: use fingerprints instead of cards to record employee attendance. The reason why fingerprint attendance is listed as one of the ten basic concepts of biometrics is because of its importance. It is said that fingerprint attendance has occupied 60% of the entire Chinese fingerprint recognition market, and fingerprint recognition accounts for 90% of the market share. The fingerprint attendance system is a management system that uses fingerprint readers and computer systems to realize attendance registration and attendance management. Question: Why is fingerprint attendance the most successful biometric application? What aspects of user needs does it meet? True rejection rate and false recognition rate Because the computer only involves some limited information of the fingerprint when processing fingerprints, and the comparison algorithm is not an exact match, the result cannot be guaranteed to be 100% accurate. The important measurement indicator for specific applications of fingerprint recognition systems is the recognition rate. It mainly consists of two parts, the rejection rate (FRR) and the false positive rate (FAR). We can adjust these two values ​​according to different uses. FRR and FAR are inversely proportional. Express this number as 0-1.0 or as a percentage. ROC (Receiver Operating Curve)-curve gives the relationship between FAR and FRR.







































Questions: 1. Should a standard be established for FAR and FER? 2. How many biometric products currently have real FER and FAR values? Who will verify their authenticity?