Biometrics






Biometrics

Every person has a set of unique physiological, behavioral, and morphological characteristics that can be examined and quantified. Biometrics is the use of these characteristics to provide positive personal identification. Fingerprints and signatures have been used for many years to prove an individual's identity, but individuals can be identified in many other ways as well. Computerized biometric identification systems examine a particular trait and use that information to decide whether you have the right to enter a building, unlock a secured area, or access a system.

Biometric systems are available today that examine fingerprints, handprints, retina patterns, iris patterns, voice patterns, signatures, and keystroke patterns. Devices have also been proposed for such traits as body weight, footprints, signatures, lipprints, wrist vein patterns, brainwaves, skin oil characteristics, facial geometry, and weight/gait patterns.

Although the human body is intrinsically difficult to measure and quantify accurately, biometric devices are generally reliable (especially when sensible thresholds are established for determining, for example, how closely a signature needs to match to be called "identical"). For best results, use biometric devices in a two-factor authentication system, in conjunction with another authentication measure such as a password.

Of the devices currently on the market, only fingerprint, handprint, and retina pattern systems are properly classified as biometric systems, because they test actual physical characteristics. Voice, signature, and keystroke systems are more properly classified as behavioral systems, because they test patterns of physiology or behavior.

The typical biometric identification system obtains data from youfor example, a handprint, a retina pattern, an iris scan, or a voice pattern. It then converts that analog signal into a digital representation and compares that representation to the many "templates" stored in the system. These templates are obtained when you are originally enrolled in the system. For example, in a signature verification system, you're required to sign your name several times to allow the system to sample traits and construct a template for later comparisons.

Biometrics have shown great promise, and in fact have long history, but they are not yet extremely popular. Users have expressed reluctance to submit to such surveillance for all but the most crucial operations. False positives and false negatives (either allowing an unauthorized person in or preventing an authorized person from entering) remain a concern. Metrics involving the allowable ratio of false positives to negatives are part of the purchase process. Potential buyers seem to be mainly deterred by cost and by fears that unauthorized users will either impersonate authorized users or somehow bypass the devices altogether. Some recent experiments with gelatin hand and fingerprints, for instance, have indicated that biometrics alone is not adequate for personnel identification.

This is not to say the technology does not have great promise. As biometric devices gain more of a foothold in the market, it's expected that useful techniques will follow, which may make such systems desirable. One example is a program that keeps track of unsuccessful attempts to gain access and then stores the characteristics (e.g., fingerprints) of the unsuccessful intruder so that person can be tracked down at a later time.

Another feature sends a distress signal if the system determines that you (an authorized user) are being coerced into helping an unauthorized individual to gain access. To make this work, you need an agreed-upon signal. For example, with a fingerprint system, you could signal that you needed help by pressing the left index finger, not the expected right index finger, on the glass plate of the scanner. The intruder would be unlikely to notice this subtlety. The software could be programmed to allow entry, but to alert your organization's security forces that you'd been forced to help an intruder gain entry.

There are a lot of tradeoffs associated with biometric systems. Because such systems are new and because many people just don't like being measured, there's quite a bit of personal resistance to using them. Although most of these methods are quite effective from a technical point of view, they may prove completely ineffective if people reject them and if they're seen as being intrusive, time-consuming, or even dangerous.

Some biometric methods are viewed as being quite threatening. Retina identification systems tend to be the most frightening; despite reassurances, people fear that the system will run amuck and blind them. That's quite an obstacle to acceptance! Other systems, because they're more familiar and less threatening, don't meet with the same degree of resistance. Signature systems, for example, are well-accepted because people are accustomed to having their signatures verified during banking and credit card transactions.

Surveys indicate that biometric devices, in order of effectiveness, rank as follows (most secure to least secure):

  • Retina pattern

  • Fingerprint

  • Handprint

  • Voice pattern

  • Keystroke pattern

  • Signature

In order of social acceptance, the order is practically the opposite:

  • Keystroke pattern

  • Signature

  • Voice pattern

  • Handprint

  • Fingerprint

  • Retina pattern

Trade organizations for developers of biometric products are addressing technical issues associated with biometric devices and are also working on increasing public acceptance of these devices.

Retina Patterns

Everybody has a unique retinal vascular pattern. Unlike a fingerprint, the pattern of blood vessels in the retinal tissue can't be recorded or even photographed with ordinary equipment. Retina pattern verification systems examine the unique characteristics of an individual's retina and use that information to determine whether the individual should be allowed access.

A retina pattern verification system uses an infrared beam to scan your retina, measuring the intensity of light as it is reflected from different points and producing a digital profile of the blood vessel patterns in the retina. The system allows access only if your retina pattern sufficiently matches those of the retina pattern stored for you in the system. The newer systems also perform iris and pupil measurements. Hand-held devices are being developed for workstation access.

Retina systems are very reliable. Their ability to work properly is affected only by very serious injuries and a few rare diseases. They have been used successfully in national laboratories, office buildings, and prisons, but they are not well-accepted as access devices. Of all of the biometric systems, retina systems seem to be the most threatening because of the public's fear that scanners will blind or otherwise injure them.

Iris Scans

Rather than examining the retina at the back of the eye, an Iris scan looks at the colored part of the front of the eye. This is much easier to image, and can be very accurate for identification. Iris scans may provide a feasible biometric where retina scans still meet resistance.

Fingerprints

Everybody has a unique set of fingerprints. Fingerprint verification systems examine the unique characteristics of your fingerprints and use that information to determine whether or not you should be allowed access.

The use of fingerprints to identify people dates from the 1800s. In the past, manual methods were used to classify and cross-check fingerprints according to certain patterns of ridges and whorlsin particular, detailed features of the print called minutiae. A fingerprint may have up to 150 of these minutiae. In the late 1960s, the FBI automated its system for cross-checking fingerprints, and all fingerprint checking was converted to automated systems by 1983.

A fingerprint system works like this: you place one finger on a glass plate. Light flashes inside the machine, reflects off the fingerprint, and is captured by a scanner, which transmits the fingerprint information to the computer for analysis. The fingerprint system digitizes the ridges and other characteristics of the fingerprint and compares these characteristics against the fingerprint templates stored in the system (or, in more primitive systems, against a print on a card that you carry). The system allows access only if your fingerprint sufficiently matches the template.

The more sophisticated fingerprint verification systems also perform a three-dimensional analysis of the fingerprint including infrared mechanisms for ensuring that a pulse is present. This means that an intruder can't gain entry by presenting a mold of an authorized user's finger or, worse still, an authorized finger that's no longer attached to its owner (a particularly grisly type of hacking!).

Because fingerprinting has historically been used as a law-enforcement tool, fingerprint systems are pretty well-accepted by potential users of such systemsparticularly in criminal justice organizations, in the military, in high-security organizations such as defense plants, and, increasingly, in banks. They have several disadvantages. They are slower than certain other types of biometric systems. In addition, their ability to work properly depends on the condition of the fingers being presented. Burns or other physical problems can affect the system's ability to match fingerprints, as can any substance (e.g., dust, perspiration, grease, glue) on fingers. And as mentioned, gelatin coatings can allow someone to "forge" a fingerprint, and then discard or eat the evidence once past the barrier.

Handprints

Everybody has unique handprints. Handprint or hand geometry verification systems examine the unique measurements of your hand and use that information to determine whether you should be allowed access.

With a handprint verification system, you place your hand on a reader, aligning all of your fingers along narrow grooves with glass between. A sensor beneath the plate scans the fingers, recording light intensity from an overhead light, and measuring fingers from tip to palm to within 1/10,000 of an inch. The information is digitized and compared against a handprint template stored for you in the system. The system allows access only if your handprint sufficiently matches that of the stored template.

The older handprint systems examined finger length and the thickness and curve of the webbing between fingers. The newer hand geometry systems examine a whole set of topographical characteristics, such as the depth of the skin creases in the palm.

Very few handprint systems are in use today, though the technology is pretty well-accepted because it's not considered to be as intrusive as other types of biometric systems. Handprint systems are said to be less reliable than fingerprint systems. Like fingerprint systems, their ability to work properly depends on the physical condition of the hand. Injuries, swelling, or the presence of rings, or even nail polish, on your fingers may affect the system's ability to match a handprint.

Voice Patterns

Everybody has a unique vocal and acoustic pattern. Voice verification systems examine the unique characteristics of your voice. Some systems also examine your own phonetic and linguistic patterns and use that information to determine whether you should be allowed access.

With a voice verification system, you speak a particular phrase. The system converts the acoustic strength of a speaker's voice into component frequencies and analyzes how they're distributed. The system compares your voice to a stored voiceprint. The voiceprint is a "voice signature" constructed by sampling, digitizing, and storing several repetitions of a particular phrase. The system allows access only if your voice signature sufficiently matches those of the stored voiceprint.

Voice systems are fairly well-accepted (they are viewed as being nonthreatening) in financial organizations such as banks (particularly vaults), credit card authorization centers, and certain types of ATMs. Their ability to work properly depends to some extent on the physical condition of the larynx. Respiratory diseases, injuries, stress, and background noises may affect the system's ability to match a voiceprint.

Keystrokes

Everybody has a unique pattern or rhythm of typing. Keystroke verification systems examine the unique characteristics of your keystrokes (your own electronic signature) and use that information to determine whether you should be allowed access.

With a keystroke system, you must type until the system can construct a reliable template of your keyboard rhythm. Once a template is available, the system will be able to examine the speed and timing of your typing during the login process, and compare it to the keystroke template stored for you. The system will allow access only if your keystroke patterns sufficiently match those of the stored template.

Because keystroke verification may be built into the ordinary login process and doesn't require a separate verification cycle, it may eventually win wide acceptance.

Certain keystroke systems are passive systems that continuously sample your keystrokes. The goal is to determine whether in fact, the person who logged onto your system under your account (presumably that would be you) remains in that position, or whether an intruder has somehow supplanted you to gain access. Because such systems can be used to perform surveillance of your work habits (What are you typing? At what rate?), they raise privacy issues.

Signature and Writing Patterns

Everybody has a unique signature and signature-writing pattern. Signature verification systems examine the unique characteristics of your signature, and the way in which you write your signature, and use that information to determine whether you should be allowed access.

With a signature verification system, you sign your name, using a biometric pen, typically attached by a cable to a workstation. The pen, or the pad on which you write, converts your signature into a set of electrical signals that store the dynamics of the signing process (e.g., changes in pressure as you press down lightly on one stroke and more forcefully on another). The system compares the signature to a signature template stored for you. It may also analyze various timing characteristics, such as pen-in-air movements, that are unique to you and that are much more difficult to forge than the actual static signature on a page. The system allows access only if your signature and related characteristics sufficiently match those of the stored template.

Signature systems are a very well-accepted type of biometric system because people are accustomed to having their signatures scrutinized. Such systems are also much cheaper than many of the other biometric systems described in this section.



 Python   SQL   Java   php   Perl 
 game development   web development   internet   *nix   graphics   hardware 
 telecommunications   C++ 
 Flash   Active Directory   Windows