Biometrics

Biometrics (including biometrics - from Ancient Greek βίος bios "life" and μέτρον métron " measure, scale " ) deals with measurements on living things and the requisite measurement and evaluation.

Depending on the application, there are different detail definitions. Christoph Bernoulli used in 1841 as one of the first scientists to the concept of biometrics in a very literal interpretation of the measurement and statistical analysis of the human life.

The concept of biometrics has the two facets of the biometric statistics and the biometric recognition method, which are separated in practice.

In biometric statistics it comes to the development and application of statistical methods for evaluation of measurements of all kinds of living beings. It is used extensively by all walks of life sciences. Pioneer of the scientific method was Karl Pearson ( 1857-1936 ). In this context, biometrics is used as a synonym for Biostatistics.

As a recognition method to put an early biometrics for personal identification. So Alphonse Bertillon developed in 1879 later called Bertillonage system for identification, the body length measurements based on 11 ( anthropometry ). 1892 Francis Galton laid the scientific foundation for the use of the fingerprint ( fingerprinting ).

Today, one defines biometrics in the field of personal recognition as automated recognition of individuals based on their behavioral and biological characteristics.

Further applications of biometrics, for example, automated disease diagnostics.

Biometrics lives on the interaction of the disciplines of life sciences, statistics, mathematics and computer science. Only today's information technology makes it possible to cope with the high computational power requirements of conventional biometric techniques.

• 4.1 Biometric statistics
• 4.2 Biometric recognition method

Biometric statistics

Biometrics as development and application of statistical methods in the context of empirical studies on living organisms is used to generate scientific knowledge, decision-making and economic optimization of products. Here are some examples:

• Biology
• Epidemiology: study of causes of disease, distribution channels and environmental influences, such as to support an effective health policy and disease prevention
• Forestry
• Genetics: study of the genetic components of diseases for better prevention and increase the chances of recovery
• Agriculture: fodder development and optimization; Plant breeding, yield optimization as a function of environmental parameters
• Medicine: identification of risk factors for certain diseases; Clinical studies in advance of marketing authorizations for determining the effects and side effects, assessment of risk-benefit ratio
• Insurance Mathematics: Calculation and prediction of the relevant parameters for life insurers, for example, the mortality table.
• Veterinary Medicine: degradation behavior of medicinal products; Research into the causes of disease, distribution channels and environmental influences

Biometric recognition method

Biometric recognition methods have experienced in recent years an enormous upswing. Technological progress allows the rapid measurement of biological characteristics and their evaluation at a reasonable cost and high quality in an increasing degree. The use of biometrics is a promising approach to solve the unsolved problem of many security concepts: How to connect identities and their associated rights, including the right identity having physical persons?

Founded in Australia in 2001 Biometrics Institute has the statutory duty to promote the responsible use of biometric technologies.

Biometric characteristics

The use of biometrics for automated recognition of individuals, it is important to find individual biometric behavioral or physical characteristics, among other things, to characterized by the following properties:

• Uniqueness: The measured value of the characteristic is different for every person possible
• Constance: The measured value does not depend on the age of the person or the time of measurement
• Measurability: There should exist a well -defined measure, for which there are suitable sensors
• Universality: The characteristic occurs in as many people.

Biometric characteristics are often distinguished in active / passive, verhaltens-/physiologiebasiert or dynamic / static. Among the long-term stable behavioral characteristics include the voice, the hand or signature, the typing and the transition dynamics. Physiological characteristics are long-term stable, for example, fingerprint, iris, or hand geometry. This distinction is largely accepted, but there are border areas. So most behavioral biometric characteristics are influenced by the physiology, such as the voice by the voice apparatus of man.

The biometric characteristics can be used, inter alia:

• DNA ( mobile DNA test, genetic fingerprinting)
• Fingerprint (finger line image )
• Transition style (English automatic gait recognition)
• Facial geometry
• Hand geometry
• Hand line structure
• Hand vein structure
• Iris ( iris )
• Body odor
• Body size ( anthropometry )
• Lip movement, usually associated with voice recognition (tone )
• Nail bed pattern
• Ear form
• Retina ( eye)
• Voice (not to be confused with voice recognition )
• Tip behavior on keyboards (English keystroke dynamics)
• Signature ( static, dynamic, and handwriting)
• Dental impression

Implementation and operation

A biometric recognition system consists mainly of the following components sensor (primary ), feature extraction and feature comparison together. What kinds of sensors are used depends very much on the biometric characteristic. Thus, a video camera is suitable for most characteristics; for fingerprint recognition, other imaging methods are considered. The sensor component supplies as a result of a biometric sample. The feature extraction removed using complex algorithms, all supplied by the sensor information that does not meet the required characteristic properties and, as a result of the biometric features. The feature comparator finally calculates a comparison value ( score ) between the stored biometric template in the learning phase and the current supplied by the feature extraction data set. Exceeds or falls below the comparison value into an ( adjustable ) threshold, the detection is considered successful.

In the " learning phase ", the enrollment, the biometric feature data encoded as a reference pattern in digital form can be stored. The next contact with the biometric system a current sample is recorded and compared with the reference sample (template ). The system then decides whether the similarity of the two samples is sufficiently high and thus, for example, may be made an entry or not.

The most important species are the detection and verification of the ID. During verification, the person needs to be verified the system first their name or user ID announce. After the biometric system decides whether the person belongs to the corresponding reference feature data set or not. In the identification of the person to be recognized reveals only her biometric characteristic, the system determines it by comparison with the reference feature records all users the associated name or user ID.

Performance criteria

As supplied by the biometric sensor samples are subject to strong statistical fluctuations, there may be occasional false positives. The reliability of the identification or verification is mainly assessed according to two criteria: after admission rate Unauthorized and after the rejection rate Authorized:

• False Acceptance Rate (FAR ) = admission rate Untitled
• False Rejection Rate (FRR ) = rejection rate Authorized

Both rates depend in opposite directions from the decision threshold from: Although a higher selected threshold reduces the FAR, but simultaneously increases the FRR and vice versa. Therefore, results such as the sole indication of FAR without associated FRR no sense. In a FRR of 10 %, the ( verification ) FAR with good biometric systems reach values ​​of 0.1 % depending on the feature to < 0.000001 %.

While the FAR is a constant for a given decision threshold Verfikationssystemen, it grows in identification systems with the number of stored reference data records. Approximation, there is the resulting total FAR from the multiplication of the underlying verification FAR with the number of records. This is the reason why only highly distinctive characteristics such as iris and ten fingerprints permit reliable identification over large databases with millions of entries.

Finally, describing the

• Falschenrolmentrate ( FER ) = rate of unsuccessful enrollments

The fact that not every biometric characteristic for each person at any time is available in sufficient quality. The FER depends not only on the particular constitution of the biometric characteristic, it is like the other error rates also affected by the performance of the technology and the participation of enrolten subject.

In general, the error rates can be described theoretically not calculate, but are to be determined in elaborate statistical studies. Here, the effort increases with decreasing error rates inversely proportionally. Process for performance testing and evaluation for biometric systems describes the ISO / IEC 19795th

In biometric systems as well as the detection time plays a major role. In addition to the safety and reliability of the user acceptance and usability are ( usability ) when assessing a biometric system are decisive criteria.

Applications

Biometric recognition methods are used almost anywhere where the identity of a person directly or indirectly plays a role. However, not necessarily all applications successfully. It is important that the application and the capabilities of a specific biometric characteristic match. The most common methods are the verification with map / card ID and the pure, wherein the user is authenticated only using the biometric characteristic. Although the latter is very comfortable, but provides with number of users increases demands on the biometric characteristic (FAR ), the computing power and data protection, and is not usually suitable for safety- critical areas. When using a card, the biometric reference data can be stored on a chip or printed on the map as a 2D bar code. There are also systems that use the card only as a pointer for the reference data set stored in a database.

Automated fingerprint identification systems (AFIS ) support the Daktyloskopen when comparing crime scene latents with the stored or abzunehmenden fingerprints of criminals or suspects. While the manual evaluation of fingerprints in Germany since 1903 is one of the best tools of the criminal police investigation, found the first computer- based methods in the 1980s in the U.S. and in Germany in 1993 the entrance to the investigative work.

PC Fingerprint Logon: With the advent of cost-effective semiconductor fingerprint sensors from about 1998, established the first products on the market that replaced the password logon to the PC or to the corporate network through a fingerprint recognition or supplemented. Although such systems previously could only be enforced by the professional field is to be expected in the future that most notebooks will be equipped as standard with even cheaper strip sensors. ( Strip sensors require the user an active movement over the sensor. ) The main argument is called by eliminating forgotten passwords cost savings.

Biometric passports and identity cards: Based on the international standard 9303 ICAO issued in Germany since November 1, 2005, only passports with an integrated chip on which a digital photograph is stored as the biometric sample. Since November 2007, the fingerprints are captured. Biometric passports are distinguished by the following characteristics: any staff reductions at the border control by higher clearance rate, assist in determining the membership of passport and holders, high cost, the passport holder shall bear, as well as unexplained data protection situation when using the biometric data from foreign countries. In Switzerland, the electronic recording of biometric characteristics in the pass is voluntary. From 1 November 2010, the German ID cards are provided with biometric features. Of these, the partial exception of the provisional passports (indicated by a green envelope ), the children's passports or the provisional identity cards. Although these are not integrated chip, but still put a biometric photo ahead. In children, more deviations are allowed on the photographs and fingerprints can optionally be delivered only from the age of 6.

Season tickets: For non-transferable season tickets offers itself to the use of biometric identification, to prevent disclosure to unauthorized parties. The Hanover Zoo is for this purpose for several years successfully a face recognition system. Other applications, mostly based on fingerprint, are becoming increasingly common in gyms, solariums and spas.

Physical access: For access to particularly sensitive areas, conventional authentication methods supplemented by biometrics. Examples are face detection in entryways to smart card development areas, fingerprint recognition in nuclear power plant areas and iris recognition in the baby care unit of a hospital Berchtesgaden. In Japan, the hand vein recognition enjoys great popularity.

Pay by fingerprint: More and more stores offer their registered ordinary customers the opportunity to take to pay with a debit card fingerprint where payment is made by direct debit. Features: the customer needs them cash or a card; there are data protection laws similar problems as with discount card systems.

Detection of asylum seekers: from asylum seekers are recognized when they enter the EU the prints of all 10 fingers. With the help of the Eurodac central database can then be determined whether an asylum seeker has already been rejected from another EU country.

Casinos occasionally use biometrics (usually facial recognition and fingerprint) to prevent gambling addicts from entering. Players who know themselves that they are at times addictive, may voluntarily deposit with the casino their biometric data in order to protect themselves in this way prior to the exercise of their addictive behavior.

Doubtful applications: the use of biometrics makes sense only if the biometric characteristic can meet the specific requirements of an application. For example, with extremely high error rates ( FRR) to be expected when trying to identify construction workers for the purpose of attendance control with fingerprint systems on site. Reason is the pollution and temporary wear of the finger lines. A fully automated search of the target population by facial recognition to conventional surveillance cameras fails generally to the insufficient recognition rate caused by one for the identification to poor image quality and setting a low probability of occurrence of the searched. Experts strongly discourage applications from, may mislead the attacker, fingers cut off by authorized persons (examples: immobilizer or ATM in the identification mode ).

Safety aspects

In applications where an incorrect verification or identification may cause damage, is not only a sufficiently low false acceptance rate (FAR ) is important. Since biometric characteristics can copy as mechanical pattern or as a record, depending on the application and characteristic also ensure that the biometric recognition system is able to distinguish facsimile of originals and former dismissed if necessary. This is particularly important because a biometric characteristic usually can not be replaced as a password.

To solve this problem differently exist powerful method for automated copy detection. The defense of misuse separate parts of the body however, is made by methods of liveness detection. Copy and live recognition come but for cost reasons, usually only at high security requirements in question. Other methods rely on the combination of several characteristics, the combination with conventional authentication methods or manual monitoring to detect intrusion attempts. Most simple biometric systems for small nominal height are not currently equipped with a copy or liveness detection, which results in detail to criticisms of biometrics. Most documented attempts at forgery go but from conscious and left behind in good quality latent fingerprints. Scientific studies, such as the degree of risk in real life seem not currently exist.

Privacy Policy

Biometric authentication systems are subject to the legal protection of privacy in general. For privacy the following properties are important:

• Biometric characteristics can be well used as unique identifiers, more or less. On this basis, an abuse is possible, as we know from e- mail addresses, U.S. Social Security numbers or credit card numbers. If the biometric characteristics in different applications (eg, payment systems, access control ) used, they may also allow assignment of persons on these applications of time ( cross- matching) and the determination of profiles. Furthermore, the use of biometrics to monitor (eg the whereabouts or behavior ) is possible up to the state abuse.
• Biometric characteristics can be not like change passwords or cryptographic keys or " call back ".
• Biometric characteristics are usually not perfect secrets, but can be detected from recordings or tracks without the knowledge and consent of the owner.
• Biometric systems can be only imperfectly protect against facsimile.
• Biometric characteristics can further information, eg on gender, ethnic origin, physical condition or health status included.

For all these reasons, compliance with data protection principles is essential. These include, in the case of biometrics:

• Avoid risks: access -protected and encrypted storage of the biometric reference data as possible under the full power of disposal of the biometric subject. Alternatively Biometric Template Protection can be used to prevent the misuse of the stored reference data.
• Elimination of sample information, which are not necessary for the detection ( diseases etc.).
• Limited to applications where no harm to the biometric subject arises when the biometric data from falling into the wrong hands.
• Voluntary application for biometric test subject and ability to be able to non-discriminatory use other authentication types.