john Gill technology header image

An Introduction to Biometrics - Fingerprint Recognition


What is fingerprint recognition?

Fingerprint recognition consists of comparing a print of the characteristics of a fingertip or a template of that print with a stored template or print.

How does it work?
A fingerprint consists of the features and details of a fingertip. There are three major fingerprint features: the arch, loop and whorl. Each finger has at least one major feature. Loops are lines that enter and exit on the same side of the print. Arches are lines that start on one side of the print, rise into hills and then exit on the other side of the print. Whorls are circles that do not exit on either side of the print. The smaller or minor features (or minutiae) consist of the position of ridge ends (ridges are the lines that flow in various patterns across fingerprints) and of ridge bifurcations (the point where ridges split in two). There are between 50 and 200 such minor features on every finger [1]. Fingerprint matching done on the basis of the three major features is called pattern matching while the more microscopic approach is called minutiae matching. Other features may be used for matching, but patterns and minutiae are the main ones [2].

  1. Acquire a sample
    Enrolment and acquisition can be done by sensors reading the tip of the finger directly and in real-time. A fingerprint scan contains a lot of information but scanners normally focus only on getting an image of the information that is essential for matching. The quality of the sensed fingerprint image is of key importance for the performance of the system. Given the small area of the fingertip, its detailed minutiae and its continuous use in everyday life (e.g. cuts, bruises, aging, weather conditions), poor image quality is a major concern in fingerprint applications.
    There are three types of live scanners:
    • Optical devices using a light source and lens to capture the fingerprint with a camera;
    • Solid-state sensors or silicon sensors appearing on the market in the mid-1990s to address the shortcomings of the early optical sensors [3];
    • And others, such as acoustic sensors that use acoustic signals to detect fingerprint details.
  2. Extracting features
    Getting a high quality image of the fingerprint is very important for accurate fingerprint recognition, but also feature extraction plays a crucial role. It consists of converting the fingerprint image into a usable and comparable format that does not require lots of storage space. The format or template is a compressed version of the fingerprint characteristics. Several approaches to automatic minutiae extraction exist, but most of these methods transform fingerprint images into binary images. This means that only the coordinates of the minutiae (30 or 40) are stored, reducing it to a few hundreds of bytes [4].
    Feature extraction is also needed because even a very precise fingerprint image will have distortions and false minutiae that need to be filtered out. For example, an algorithm may search the image and eliminate one of two adjacent minutiae, as minutiae are very rarely adjacent. Anomalies can also be caused by scars, sweat, or dirt. The algorithms used for feature extraction filter the image to eliminate the distortions and would-be minutiae [5].
  3. Comparing Templates
    The identification or verification process follows the same steps as the enrolment process with the addition of matching. It compares the template of the live image with a database of enrolled templates (identification), or with a single enrolled template (authentication).



  4. Declaring a Match
    The comparison between the sensed fingerprint image or template against records in a database or a chip usually yields a matching score quantifying the similarity between the two representations. If the score is higher than a certain threshold, a match is declared, i.e. belonging to the same finger(s). The decision of a match or non-match can be automated but it depends also on whether matching is done for identification or verification purposes.
    With identification applications, automated decision-making is possible when conditions are ideal. In the case of the Federal Bureau of Investigation (FBI) for instance, this means that fingerprint cards can be matched automatically when both enrolment and acquisition were done by law enforcement staff. But with latent prints (eg. collected at a crime scene), and prints with a lower quality image, the automated process is less reliable. Automated systems imitate the way human fingerprint experts work but the problem is that these systems can not have observed the many underlying information-rich features an expert is able to detect visually. Automatic systems are however, reliable, rapid, consistent and cost effective when matching conditions are good, but their level of sophistication cannot rival that of a well-trained fingerprint expert. Therefore, for instance a fingerprint expert can overrule an automated match [6].
    Verification applications, especially mainstream commercial fingerprint verification may be, to a certain extent, less accurate because the issues at stake are different (e.g. identifying criminals), but also because verification consists of 1-1 matching. Verification may use less information from a fingerprint compared to forensic scientists identifying a fingerprint. The former seems to be more like a possible, "close-enough correlation" of similarities. Also, because of background interference (dirt, scratches, light, etc.), and no human supervision, the quality of fingerprint images is lower. The result is a "best" matching score which would not be feasible for law enforcement [7].

Applications
Fingerprint identification of criminals for law enforcement continues to be one of the major applications domains for this technology. Another large scale application in Europe is EURODAC for asylum requests. In New York, fingerprints are used to prevent fraudulent enrolment for benefits. Using fingerprint recognition to secure physical access is another popular application. Moreover, embedding of fingerprint readers in electronic devices opens up a whole range of digital applications that are based on online authentication. Finally, decisions have been taken for the future integration of fingerprints (with other biometrics) on travel documents and passports.

The Integrated Automated Fingerprint Identification System, more commonly known as IAFIS, is one of the largest biometric database in the world. It is a US national fingerprint and criminal history system maintained by the FBI. It contains the fingerprints and corresponding criminal history information for more than 47 million subjects in the Criminal Master File. The fingerprints and corresponding criminal history information are submitted voluntarily by state, local, and federal law enforcement agencies. The IAFIS provides automated fingerprint search capabilities, electronic image storage, and electronic exchange of fingerprints and responses, 24 hours a day, 365 days a year [8]. In Europe, there is no such database. Criminal fingerprint databases are under control of national criminal authorities. The UK for instance, has a national automated fingerprint identification system (NAFI) containing more than four million records.

There is however, since January 2003, also a large central fingerprint database in the European Union, but for another purpose. It aims at preventing duplication of asylum requests in the EU Member states. EURODAC is an EU wide database (AFIS) set up to check the fingerprints of asylum seekers against the records of other EU countries. After one year of operation, an evaluation report on EURODAC highlighted satisfactory results in terms of efficiency, quality of service and cost effectiveness. The EURODAC central unit has been operating continuously. Within one year, it processed almost 250,000 fingerprints of asylum seekers. It detected 17,287 cases of multiple application (a same person having already made an asylum application in another country), which represents 7% of the total number of cases processed.

In addition to asylum requests, also illegal immigrants are identified. Almost 17,000 fingerprints of people in an illegal situation were detected and about 8,000 fingerprints related to attempts to cross borders illegally. The evaluation report also states that there were no data protection problems raised by the Member States' national data protection authorities regarding EURODAC operations [9].

The state of New York has over 900,000 people enrolled in a system which tracks entitlement to social services and protects against fraud known as “double dipping”, i.e. enrolling for a benefit under multiple names (OECD, 2004: 23). Fingerprint scanning is also being used to arrange secure access to schools and schools premises such as cafeterias and libraries. Finally, with the embedding of fingerprint scanners in electronic devices, online authentication (replacement of passwords, PINs, etc) becomes possible for a whole range of applications including electronic payments.

Finally, at EU level, the Council of European Ministers adopted the Regulation on mandatory facial images and fingerprints in EU passports at its meeting in Brussels on 13 December 2004. This Regulation applies to passports and travel documents issued by Member States (excluding Ireland, the UK and Denmark). After the Regulation is published in the Official Journal passports issued will have to contain a facial image within 18 months, and fingerprints within three years. Also a Committee will be set up by the European Commission with representatives from 22 Member States to decide on the details such as how many fingerprints are to be taken, the equipment needed and the costs [10].


Recommendations

Making the fingerprint reader easier to locate

Improving the accessibility of the fingerprint reader design

Providing instructions in an accessible format

For example, if the scan is not successful: "This scan was not successful, please move your finger to the right."

or

"This scan was not successful, please hold your finger still on the reader."

Informing the user that the reader is waiting for him/her to take action

Catering for users who do not require audio instructions (e.g. those who have good vision, or those who are familiar with the process)

Reference: Identification of Accessibility Issues for Visually Impaired Users of Biometric Technologies: Fingerprint Readers


Standards
Published standards:

  • ANSI INCITS 377-2004: Information Technology-Finger Pattern Based Interchange Format
  • ANSI INCITS 378-2004: Information Technology-Finger Minutiae Format for Data Interchange
  • ANSI INCITS 381-2004: Information Technology-Finger Image Based Interchange Format
  • ANSI/NIST-ITL 1-2000: Information Systems-Data Format for the Interchange of Fingerprint, Facial, & Tattoo (SMT) Information
  • ISO/IEC 19794-4:2005: Information technology - Biometric data interchange formats - Part 4: Finger image data
  • ISO/IEC FCD 19794-2: Information Technology-Biometric Data Interchange Formats-Part 2: Finger Minutiae Data

Standards under development:

  • INCITS PN-1703-D: Information Technology- Conformance Testing Methodology for the Finger Minutiae Interchange Format
  • INCITS PN-1705-D: Information Technology- Conformance Testing Methodology for the Finger Image Data Interchange Format
  • ISO/IEC CD 19794-3: Information Technology-Biometric Data Interchange Formats-Part 3: Finger Pattern Spectral Data
  • ISO/IEC WD 19794-8: Information Technology-Biometric Data Interchange Formats-Part 8: Finger Pattern Skeletal Data


Further information


[1] OECD, 2004: 21
[2] O’Gorman, L Fingerprint verification, in: A. Jain, Personal identification in a networked society, Dordrecht: Kluwer Academic Publishers. R. Bolle & S. Pankanti (eds.), 1999. pp. 43-64.
[3] Shortcomings are mainly size and cost. The chip sensors comprise an array of sensing elements (each pixel is a sensor) that image the fingerprint. Solid-state sensors have on-chip conversion (analogue to digital) so that a digital image can be generated. There are mainly two types of solid state sensors. Capacitive sensors are most prevalent and use electric field strengths for distant measurement of fingerprint ridges and valleys. Temperature sensors measure the temperature difference of a finger between the skin ridges and the air (valleys).
[4] Mainguet, J-F. & Pégulu, M. & Harris, J-B. Fingerprint recognition based on silicon chips, Future Generation Computer Systems, 16, 2000, pp 403-415.
[5]www.biometricgroup.com/reports/public/reports/finger-scan_extraction.html
[6] Jain, A. & Pankati, S. Automated Fingerprint Identification and Imaging Systems, in Lee, H. & Gaensslen, R. (eds.), Advances in Fingerprint Technology, 2nd Edition, Elsevier Science, 2001.
[7]onin.com
[8]www.fbi.gov/hq/cjisd/iafis.htm
[9] EURODAC detects 7% of multiple asylum applications during its first year of activity; Press release by the European Commission, Reference IP/04/581, 05/05/2004.
[10] EU Council Regulation on standards for security features and biometrics in passports, 15152/04, 10 December 2004.


Acknowledgement

The information contained in this section was collected from the following source:

Kind permission was given by Bioscrypt Inc, Cogent Systems Inc, Deltabit Oy and Lenova for the use of the above pictures.

 

 



John Gill Technology Limited Footer
John Gill Technology Limited Footer