Iris Recognition: By Jennifer Grigsby

Figure 1

I. Introduction

Iris recognition is a type of biometrics that recognizes a person by analyzing the random patterns of their iris. The iris is the colored part of the eye: an internal organ (muscle) in the eye that regulates the size of the pupil, controlling how much light gets in. The coloration and structure of an iris are genetically linked, but the detailed patterns are not. Two irides can be genetically identical, but everybody’s has a different structure.

HOW IT WORKS

Before the actual iris recognition can happen, several steps need to take place: location of the iris using landmark features, imaging, feature isolation, and extracting. Modern iris cameras use infrared light for the iris scan so it doesn’t harm the subject or make them uncomfortable. During imaging, a 2D Gabor wavelet filters and maps the segments of the iris into phasors (vectors). The phasors include information on the orientation, spatial frequency, and position. This info is used to map the IrisCode. An IrisCode describes iris patterns by using phase information collected in the phasors. Two IrisCodes are compared to perform the iris recognition. Hamming Distance (HD) is used to measure how closely the two match up statistically. If the HD is under a certain number, then the two IrisCodes are from the same iris. So for iris recognition to match up two irides, it has to fail a test of statistical significance (the HD has to be too low to be significant).

Figure 2: a localized iris

II. Discovery (Timeline)

1936: Frank Burch (ophthamologist) proposed concept of using iris patterns to recognize people

1985: Drs. Leonard Flow and Aran Safir proposed the concept that no two irides are the same

1987: Iris identification patented

1993: Defense Nuclear agency began testing to develop a prototype unit (completed in 1995 by Flom, Safir, and Dr. John Daugman)

1994: Daugman’s algorithms for automated iris recognition were patented

1995: first commercial products available

2005: broad patent for the concept of iris recognition expired, enabling other companies to use their own algorithms to make iris recognition devices

Patent on IrisCodes (developed by Daugman) expires in 2011

III. Biography of Investigator

Dr. John G. Daugman teaches neural computing, information theory, and computer vision classes at University of Cambridge. He is best known for developing IrisCode, the algorithm for iris recognition. After earning his B.A. and PhD degrees at Harvard University, he taught there. He co-founded Iridian Technologies, and is also a board member. Awards that Daugman has received include the following: the Presidential Young Investigator Award of the National Science Foundation, the Information Technology Award and Medal of the British Computer Society, and the Order of the British Empire from Her Majesty Queen Elizabeth II.

Figure 3: an example of an IrisCode

IV. Impact on the World/Humanity

It is a convenient means of access (especially for disabled people) because it’s fast, automatic, hands-free, and reliable. The majority of identification today is done by showing documentation like passports or ID cards, or entering a password, but fraud is pretty much unavoidable with methods like these. When iris recognition is configured and used properly, there’s no room for error in terms of identifying the wrong person.

IRIS RECOGNITION IN ACTION

Figure 4

- United Arab Emirates Homeland Security Border Control

- Schiphole Airport, Netherlands (passport-free immigration)

- United Kingdom: IRIS (Iris Recognition Immigration System)

- A number of U.S. and Canadian airports

- Google (uses it to access to their datacenters)

- and much more!

Figure 5: The IriScan model 2100 is a device that scans irides

V. Journal Article Review

As of January 2004, when this journal article was written, iris recognition had been tested several million times without producing any false matches. The phase information gathered by phasors is so complex and particular that identification can be made very confidently. An advantage of iris recognition over other types of biometrics is its ability to be used in “identification mode”, meaning an entire database can be searched and come up with the result, versus “verification” in which one individual is compared to another and either passes or fails (many comparisons versus few comparisons).

VI. References

Daugman, J. (n.d.). Iris Recognition. The International Center for Disability Resources on the Internet. Retrieved May 1, 2011, from http://www.icdri.org/biometrics/iris_biometrics.htm

Iris Recognition. (2006, August 7). National Science and Technology Council, Subcommittee on Biometrics. Retrieved May 1, 2011, from www.biometrics.gov/Documents/irisrec.pdf

John Daugman. (n.d.). American Scientist Online. Retrieved May 1, 2011, from http://www.americanscientist.org/authors/detail/john-daugman

Pictures:

Daugman., J. (n.d.). Iris recognition. The Computer Laboratory. Retrieved May 1, 2011, from http://www.cl.cam.ac.uk/~jgd1000/iris_recognition.html

Examples of IrisCodes. (n.d.). The Computer Laboratory. Retrieved May 1, 2011, from http://www.cl.cam.ac.uk/~jgd1000/examples.html

File:IriScan model 2100 iris scanner 1.jpg - Wikipedia, the free encyclopedia. (n.d.). Wikipedia, the free encyclopedia. Retrieved May 1, 2011, from http://en.wikipedia.org/wiki/File:IriScan_model_2100_iris_scanner_1.jpg

Journal Article:

Daugman, J. (n.d.). IEEE Xplore - How iris recognition works. IEEE Xplore - Home. Retrieved May 1, 2011, from http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=1262028