April 2007 — News
Print this article | Email this articleClick here to receive your FREE subscription to T.H.E. Journal
Biometrics in K-12: Ban or Buy? (Part 1)
What are biometrics?
Biometrics are measurable physical and behavioral characteristics that can be used for verification or identification of an individual. Physical biometrics are gathered from analyzing fingertip patterns; measuring facial characteristics or the shape of a hand; and analyzing features of the iris or retina in an eye, vein patterns, and DNA for genetic makeup. Body odors, ear lobes, sweat pores, and lips have also been investigated to characterize individuals. Behavioral biometrics include vocal behavior, signature dynamics, keystrokes, and a person's gait (Biometrics 101, n.d.).
How does the technology work?
Biometrics rely on pattern recognition systems. Most are based on collecting a set of minutiae or points from an individual, determined by location, orientation, and minutia type associated with the biometric (Jain, 2004). Hardware, application software, and a reader or scanning device are needed to capture biometric data, convert it to digital form, and then store it as a template in a database to be used for comparisons to future captures from those devices.
Data from fingerprints and hand geometry appear to be the most commonly collected in schools, and a couple of school districts have experimented with iris recognition. Fingerprint technology relies on capturing a set of minutiae from "where a friction ridge begins, terminates, or splits into two or more ridges" (Biometrics Glossary, n.d.). The latest hand geometry technology relies on capturing 3D images of the size and shape of the hand. Templates are made from measurements such as length, width, thickness, and surface area (Ingersoll Rand Recognition Systems, FAQ, 2006). Iris recognition works by capturing an image of the iris from an individual glancing into the aperture of a digital camera from about three to 10 inches away. The iris is analyzed for its unique patterns, which are then converted into a 512 byte code (The Point Group, n.d.).
Sagem Morpho Inc. briefly describes the biometric scanning process used in its MorphoTouch product. Five pictures, worth viewing, quickly illustrate how the numeric template (fig. 1) is created from the scanned fingerprint image. One or two fingers, generally the index finger of each hand, are pressed flat on the scanning surface to capture points from the center of the finger, as opposed to capturing points from all of a finger by rolling it from nail to nail. For children under 18 the image is standardized and resized before processing. Up to 40 or so uniquely identifying points from the swirls and arcs of the scanned finger(s) are captured, then the original image is discarded. The captured points are converted to numbers ("just like a social security number") and stored as a template to be later used for verifying or identifying an individual stored in the database. The original fingerprint image can't be recreated from this template. The company notes differences between this type of fingerprint scanning and that used for forensic and criminal investigations to help ensure parents and students that the biometric templates are not like those required by law enforcement agencies for identification purposes.
The process of verification involves authenticating a claimed identity (you are who you say you are), a direct one to one approach comparing the verification template, and a recapture of the fingerprint, to the enrollment template, the one created the first time. Identification is a one to many approach to find out who someone is by submitting a biometric sample and comparing it to all the stored templates in the database (Biometrics Glossary, n.d.). Each time a scan is made, the template created differs from the one captured and stored the first time. The minutiae collected also depend on the nature of the sensor on the reader.
The point is, when the templates are compared, the system determines if a sufficient number of matches exist to verify or identify. That's the challenge, and that's where where error rates come in. The performance relies on the database, which depends on the quality of the original images captured and how well each original template was created. The condition of a finger (e.g., cuts and bruises, too wet or dry, too cold or hot) affects the quality. False accepts (i.e., verifying an imposter), false rejects (i.e., failing to verify a legitimate individual), and failure to enroll are primary concerns (Jain, 2004).
