Sometime around 1831, a new word entered the American lexicon. That word, biometrics, is the subject of this month's column. As originally introduced, biometrics, a noun used with a singular verb, meant the statistical analysis of biological observations and phenomena.
This would be perfectly fine for our purposes if Plant Services were a magazine dedicated to bringing you the absolute best in information concerning chlorophyll-laden life forms. Unfortunately, the only plants we know about feature quite a bit more steel and concrete and rather less of that green, oxygen-releasing miracle that makes life as we know it possible.
Recently, however, the word has taken on another high-tech definition. Biometrics also now refers to technology that is supposed to be capable of automatically identifying an individual on the basis of inherent biological characteristics, such as face, voice, fingerprint and iris patterns or other factors unique to that singular person. Look at it as a macro version of a DNA match.
The ostensible reason given for the development of this technology is a higher degree of safety and security. Relying on documents, passports and photos, you see, isn't considered a totally foolproof way to verify a person's identity. Biometrics seeks to furnish absolutely positive identification. After all, one can forge a paper token of legitimacy, but surely nobody can fake an eyeball or fingerprint. Or can they?
In an attempt to learn more about this subject that may well be coming to a manufacturing plant near and dear to your hearts, join me on another dive into the chaos we affectionately call the Web in search of zero-cost, non-commercial, registration-free resources about biometrics.
Whenever one plows into a new field, it's a good idea to understand the terminology and jargon that grow there. For that, you need a glossary, and the good folks at the UK-based Association for Biometrics know your needs well. Click your way over to http://www.afb.org.uk/docs/glossary.htm, where you'll find a long list of applicable words and their definitions. It's not particularly user-friendly, as the word and its definition are set in the same font with no visual element separating entries. But, hey, it's free.
The big picture
The Pattern Recognition and Image Processing Lab in the Department of Computer Science and Engineering, Michigan State University offers a bit more detail about some of the biometric technologies at http://biometrics.cse.msu.edu/info.html in a document titled An Overview of Biometrics. When you get to the bottom of the page, click on one of the technologies for an expanded explanation of how things are supposed to work in the biometric world in which we'll be living.
Novell, Inc. posts all manner of information for software developers who use the company's products. Among the many articles is one that explains the basic concepts behind biometric technologies. You can find Overview of Biometrics at http://developer.novell.com/research/apnotes/2001/july/01/a0107013.htm.
Human beings are hardwired to use the face as a primary means of recognition. Your mug is, after all, the most distinguishing characteristic you can display to the world at large. A body is a body is a body, except when it looks at you and smiles, as in the photos on the wall in your local post office.
For your reading pleasure, BRTRC, a defense contractor based in Fairfax, Va. has posted Face Recognition 101: A Brief Primer by Duane M. Blackburn. The document doesn't explain the nitty-gritty of how the technology is supposed to work. Rather, it explores the functional testing that qualifies a given face recognition technology as useful. To learn more, face your mouse in the general direction of http://www.frvt.org/DLs/FR101.pdf and click the button. By the way, I was unable to determine what BRTRC stands for. Spooky, no?
Chances are, you're familiar with the art and science of monitoring machine vibration as a diagnostic tool to achieve optimum asset reliability. If so, it's not that big a mental leap to get a grasp on voice recognition. At the University of Pennsylvania School of Engineering and Applied Science, four scholars, Kohanski, Lipski, Tannir and Yeung, constructed a digital voice recognition program using commercially available software and the obviously necessary microphones and other audio hardware. Their lab report, Development of a Voice Recognition Program, shows how they did it, and the intellectually curious out there in readerland will enjoy it. Order that mouse of yours to go to www.seas.upenn.edu/courses/belab/LabProjects/2001/be310s01t2.doc to get the full story.
You've got at least one eye chock full of your identity. The iris is the colored ring around the pupil, which is the hole through which light enters the eyeball. The immutable patterns in the iris form the basis for identification. Dr. John G. Daugman of the Computer Laboratory of University of Cambridge is credited with being the inventor of iris scan technology. His original paper on the subject, High Confidence Visual Recognition of Persons by a Test of Statistical Independence, was published in November, 1993 and is available on the Web. Take a minute to watch as your mouse ambles over to http://www.cl.cam.ac.uk/users/jgd1000/PAMI93.pdf for the goods. The document is a bit of a heavy read, but I thought it had some historical significance.