Automated fingerprint identification system

from Wikipedia, the free encyclopedia
A fingerprint scanner for an automated fingerprint identification system on a notebook.
Use of a mobile fingerprint scanner ( Handheld Interagency Identification Detection Equipment ) by the US Army in Iraq, 2008.
An adjudication workstation.

An automated fingerprint identification system ( AFIS ) is an information processing system of biometrics for the automated comparison of fingerprints with all or a subset of fingerprints of a (possibly very large) stored reference inventory of fingerprints.

Components

The most important elements of an AFIS system are

  • Algorithms for preprocessing recorded fingerprint images as a preliminary stage for feature extraction
  • Algorithms for extracting features from the recording of a fingerprint
  • the data structures for storing the extracted features of fingerprints (template) and the database for storing them
  • Algorithms for searching stored fingerprint templates with a high degree of correspondence to a given fingerprint template
  • a user interface for fingerprint capture, fingerprint search and manual processing of imprecise search results (adjudication)

Fingerprint scanners are often used as peripheral devices to AFIS in order to record fingerprints - but they can also be recorded differently - strictly speaking, fingerprint scanners are therefore not part of an AFIS.

Quality features

AFIS systems differ greatly, depending on the selection and coordination of the algorithms, with regard to type 1 errors (false positive - to False Acceptance Rate - called FAR for short) and type 2 errors ( false negative - to false rejection rate - called FRR for short ) Error rates - also with regard to their rotational invariance (ie the recognition of the correspondence between two fingerprints if one of the two recordings is "twisted") and independence from reference points (such as the center of the finger).

Essential specific quality features of an AFIS (in addition to the quality features that apply to all biometric systems ) are, in addition to the error rates mentioned, depending on the application, the above-mentioned invariances, the search speed, especially in the case of large reference databases, and the robustness of the search quality with regard to poor quality fingerprint recordings.

use cases

Essentially, all use cases of an AFIS system can be divided into two categories:

Fingerprint verification

In fingerprint verification (also called 1: 1 match), a given set of fingerprints is checked against a known limited set of fingerprints for a match, e.g. B. determine whether the details of a person with regard to their identity are actually correct: A recorded fingerprint is compared with the stored fingerprint of the person, for example to give a person access to a high-security area.

Fingerprint identification

In the case of fingerprint identification (also called 1: N match), a given set of fingerprints is compared with all stored fingerprints of a reference database for a match in order to e.g. B. determine whether the fingerprints of an individual are already in the reference inventory: A recorded fingerprint is compared with all stored fingerprints of the BKA to determine whether the person u. U. was involved in criminal offenses.

With fingerprint identification, a further distinction can be made in cases in which the individual cooperates (or can cooperate) in taking fingerprints - in this case, images of all 10 fingers are usually available in good quality.

In the second case, there may only be partial fingerprints of individual fingers (e.g. accident victim). If the quality is limited, as in the case mentioned, special search algorithms are required, which are usually considerably more time-consuming and resource-intensive than search algorithms based on high-quality fingerprint recordings. While searches based on high-quality recordings usually deliver search results after seconds or minutes, the above Search runs sometimes take days until a set of hits is identified, which then usually has to be further restricted by a highly specialized technician (this process is often referred to as "adjudication"). For example, Los Angeles Airport (LAX) operates a fingerprint adjudication department that deals exclusively with so-called adjudication.

Commercial systems / open source systems

There are numerous providers of AFIS systems around the world - since many of these systems now also contain algorithms for processing / recognizing faces. Sometimes also referred to as Automated Biometric Identification System, abbreviated ABIS. Large providers are u. a .:

In addition, there are also some open source implementations of AFIS systems or only system parts:

Areas of application

Police identification systems

The first automated fingerprint identification system (AFIS) of the German police was put into operation in 1993 and resulted in the time to capture unique features of the fingerprints taking only three minutes (previously 90), incomplete prints can be evaluated and matches are found more frequently. At the beginning of 2004 around 3,000,000 fingerprints are said to have been stored, with around 1,400 being added every day. In order to digitize the impressions, they are recorded with a video camera and then processed by special programs. A police officer only needs to specify the basic shape and, if necessary, correct the features identified by the program. The storage takes place centrally at the BKA , the processing decentrally at the LKA . On January 15, 2003, an AFIS-based, pan-European fingerprint identification system, EURODAC (European Dactyloscopy), was introduced. This includes u. a. Asylum seekers and illegal immigrants, provided they are over 14 years of age.

Fingerprinting for the automated fingerprint identification system IDENT for US-VISIT ( US Customs and Border Protection )

Access and control systems

Fingerprint scanner for an automated fingerprint identification system on a government building in Brazil

AFIS systems are of great importance when granting access to security areas - here, in addition to the identity of the person via the fingerprint comparison, appropriate sensors may also check whether the finger is being placed on by a living person ( life recognition ) or whether it is is a copy (e.g. wax print) of a fingerprint (copy recognition).

Electoral systems

So-called "voting cards" are currently being introduced in many countries to ensure orderly elections. Primarily, the "Voting Cards" are used to limit voters to actually existing persons and to clearly identify a voter when casting their vote - this ensures that each person entitled to vote can only cast his or her vote (s) once. Such systems have been implemented in Cameroon and Sierra Leone, for example.

Health systems

AFIS and ABIS systems are also gaining in importance in the health system. Examples are the use of AFIS systems in hospitals to regulate access to patients and electronic patient files , as well as the implementation of health cards (similar to the German electronic health card) with stored biometric characteristics in order to avoid drug abuse or multiple use of services.

Challenges

privacy

In the meantime, fingerprints are stored for a wide variety of purposes - these are subject to data protection regulations , because improper use can cause considerable damage to individual people, groups and, under certain circumstances, entire countries.

swell

Web links

credentials

  1. ↑ bsi.bund.de - Publication by the Federal Office for Information Security (BSI) Fingerprint Recognition - page 2ff - accessed on August 14, 2014.
  2. ↑ bsi.bund.de - Publication of the Federal Office for Security in Information Processing (BSI) - Fingerprint Recognition - page 2ff - accessed on August 14, 2014.
  3. ↑ bsi.bund.de - Publication of the Federal Office for Security in Information Processing (BSI) - Fingerprint Recognition - page 9ff - accessed on August 14, 2014.
  4. ↑ bsi.bund.de Biometric authentication with fingerprint recognition - 2.3 Performance of a biometric system - page 9ff - accessed on August 14, 2014.
  5. ↑ bsi.bund.de Biometric authentication with fingerprint recognition - 2.1 Introduction - page 6 - accessed on August 14, 2014.
  6. ↑ bsi.bund.de Biometric authentication with fingerprint recognition - 2.1 Introduction - page 6 - accessed on August 14, 2014.
  7. ↑ Fingerprint Administration and Adjudication Unit (FAAU) Overview ( Memento of April 5, 2015 in the Internet Archive ) accessed on August 14, 2014.
  8. ↑ sourceafis.angeloflogic.com SourceAFIS Project page
  9. ↑ sourceforge.net Fingerprint Recognition SDK Project page
  10. ↑ Facts and figures on AFIS. Federal Criminal Police Office (Germany) (BKA), 19. May 2003.
  11. ↑ Konrad Schober: European police cooperation. Heidelberg 2017, ISBN 978-3-8114-4258-0 , p. 229 (footnote 651 ).
  12. ↑ EURODAC - an EU-wide electronic system for the identification of asylum seekers. ( Memento of November 24, 2006 in the Internet Archive ) Presentation of EURODAC by the European Commission - Justice and Home Affairs
  13. ↑ Eurodac: A Solution Looking for a Problem? ( Memento of June 9, 2007 in the Internet Archive ) Study on the negotiation and implementation of the Eurodac regulation in the Council of Interior and Justice Ministers of the EU (in English), European Integration online Papers, Vol. 10 (2006)
  14. ↑ Fact Sheet: IDENT and IAFIS Interoperability , US-Visit
  15. ↑ Cameroon prepares for biometric voter system. ( Memento of May 4, 2014 in the Internet Archive ) accessed on August 13, 2014.
  16. ↑ sl.undp.org UNDP in Sierra Leone - New procedures contribute to credible elections, high voter turnout in Sierra Leone - accessed on August 13, 2014.
  17. ↑ fingerprint recognition trialled in Lewisham PCT. ( Memento from August 14, 2014 in the Internet Archive )