The science of Biometrics is the development of automated systems that can
identify a person based on their personal attributes. These systems can measure
physical or behavioral attributes unique to an individual to determine their
identity.
Physical attribute recognition currently includes the following
methods:
- Face
- This is the method most like the one
people use to recognize one another. To “recognize” the face of an individual,
the system measures the size and distances and angles between such facial
landmarks as eyes, ears, corners of the mouth, chin, eyebrows, and nose. Since
it ignores things like facial hair and hair color, it is difficult to spoof by
traditional methods used to keep people from recognizing an individual.
Sunglasses, facial putty, and masks are the only methods with a chance of
working, as they actually change or hide the faces’ geometry. However, facial
recognition is constantly improving. A pair of twins named Michael and Alex
Bronstein has developed a 3D scanning system that can even tell the two of them
apart. It does require two cameras, but the system has the ability to compare
facial structures as they appear in different poses or light conditions,
variables that could distort a face seen as a two-dimensional image.
- Fingerprint
- Arguably the oldest form of
biometric identification, a fingerprint identification system compares the
pattern of ridges on a person’s fingertips with those stored in its memory. The
first systems used a camera to “see” the person’s fingerprint, but false
fingertips could defeat this method. The latest fingerprint sensors don't
actually look at your fingerprint; they measure the electrical capacitance
of the live-skin/dead-skin interface between the fingertip and the ridges
that make up the print. This means you can't fool it with an overlay anymore.
It also means that you would be faced with a window of opportunity after
cutting off someone's thumb, since it would eventually be unable to trigger the
sensor properly once it was completely "dead".
- Hand Geometry
- Measuring the shape of a
person’s hand is another relatively old method of detecting a person’s
identity. Not as accurate as a fingerprint scan, a hand-geometry scan is useful
for places that require a relatively high level of security but do not wish to
antagonize their workers with a high level of invasiveness. It is often used in places with a
potentially high level of illegal-alien labor where workers wouldn’t want their
fingerprints on file.
- Eye
- There are two ways to determine a person’s
identity using the eye. A retina scan uses the pattern of blood vessels on the
retina at the back of the eye to determine identity, while an iris scan uses
the complex web of fibers in a person’s iris. The hotter technology with eye recognition is iris scanning, which has
the advantage of being able to use simpler optics and a greater standoff
distance without losing accuracy, which lends a feeling to the user that it is
less invasive. In addition, the iris changes radically at death, making it very
difficult to fool by using a detached eye. An additional advantage of an
iris scanner is that it could be extended to use with animals like expensive
pets and such.
Behavioral attribute recognition uses a person’s actions to identify
that person. Methods include:
- Speech
- Voice recognition is more than just how
a person sounds when they speak, it is also how they form words, and the
rhythm and cadence of their sentences. It is only possible to fool with tape if you can
get the person to say the entire code phrase at once, as splicing the statement
together from separate words spoken by the person (as was used in the movie
Sneakers to spoof the voice-recognition system at their target) wouldn’t have
the rhythm of the sentence and the proper spacing between the words that would
exist if the sentence were spoken directly.
- Handwriting
- Recognizing the shape and form of
a person’s handwriting is probably the only method of identification older than
the fingerprint. In a biometric system, those attributes are coupled with the
rhythm and cadence of the writing itself, preventing someone who can merely
copy the way a person’s handwriting looks from fooling the system. This method
is being used more and more for credit-card verification, and most delivery
companies already use some form of this so that a person’s signature can be
transmitted to the originator to verify delivery, and prevent lawsuits from
those who may claim that their signature was forged.
- Body motion
- There are some systems that measure
aspects such as a person’s walk via pressure pads, or how they move on camera
down a hallway, how they punch the security code into a keypad, or other
noninvasive recognition method. These systems are useful in places where the
facility does not wish the person being identified to know that they are being
checked.
Of course, any system worth its salt always uses two- or three-stage
security anyway, with one or two biometric tests coupled with a password,
keycode, or physical pass such as a keycard/passkey/security dongle (or both.)