Biometrics

So far, the fascinating field of biometrics was either witnessed in Sci-Fi

movies, or used by defense establishments, large banks, etc. The world around

used to remain content with authenticating a person's identity with user names

and passwords. However, things are changing.

Biometric technology is becoming more affordable, and thereby becoming more

feasible for enterprises to adopt. The most popular one till date is fingerprint

scanning, which is finding applications in various segments, like retail, BPOs,

KPOs, banks, hospitals, etc. But work is on to make other biometric technologies

more accessible to the corporate world as well. That may not happen in the

immediate future, but we'll definitely hear about more biometric implementations

in the coming years. Let's look at some of the currently available and emerging

biometric techniques.

Iris scans



As the name suggests, an Iris scan uses features in the iris of human eye to

authenticate a person. Iris patterns are complex and at least 200 points on iris

can be used for comparison. The iris scan device usually has an infrared enabled

camera and requires an individual to be placed 2-3 feet away(depending upon the

capability of device) in such a way that the individual can see his/her own

eye's reflection on the device. The camera then captures an image of the iris

which is kept in database for matching or verification. As iris patterns are

unique and different in left and right eye, chances of false matching are low.

According to vendors, eyeglasses and contact lenses don't cause any problems

for iris authentication. They remain constant over a person's lifetime except in

case of an eye disease or injury.

Blood vessel patterns at the back of the retina form a distinct pattern that varies from person to person. This lies at the heart of Iris scans

Facial recognition



As the name suggests, this technology analyzes characteristics of a person's

face which don't change easily over a particular period of time such as distance

between the eyes, sides of mouth, area around the cheek bone etc. Initial 2D

face recognition systems were fooled by masks or face molds etc, but newer

devices require a user to blink or smile for verifyication-a process known as

testing for life. Facial recognition is also used for surveillance.

Voice recognition



This finds applications beyond user authentication, and goes beyond desktop

voice recognition software one hears of. Many industries have started using it

for automated customer query management on the phone. Instead of pushing keys in

a menu driven Interactive Voice Response system, a user has to give voice

commands. Unlike other biometric fields, a voice recognition system just

requires a microphone to operate, which is commonly available. For user

authentication, various aspects of an individual's voice are captured, like

resonant and harmonic frequencies of the vocal tract, etc. Earlier voice

recognition systems could be rendered useless by using a person's voice

recording to gain access. These are countered by combining voice recognition

with speech verification.

One commercial product that provides this is Nuance SpeechSecure. It can work

with IVR systems and is language independent. Its text independent feature

requires a caller to read out the password specified at the time of enrolment

and some other speech data which can be in any language. In this case speech

data is longer than the normal text dependent speaker verification but requires

less support from callers. The voice recognition technique is increasingly being

deployed in places where a large number of people need to be authenticated on a

regular basis. Banks have shown a lot of interest in this technology for

authencating customers over the phone. Likewise, other service industries like

railways, airlines, etc will also find this useful.

Another unique way of establishing identity. Vein patterns both at the back of the hand and in the palm vary in individuals and can be traced using infrared radiation

Hand geometry scans



Identification based on a hand's geometry is also becoming popular. It's

more beneficial in areas where systems like fingerprints, palm prints, etc fail.

For eg, industries like oil refineries, coal mines and traditional manufacturing

companies where dirt, oil, grease and other kinds of debris reduce the

feasibility of finger or palm print matching. Hand geometry detectors have been

used for access control in Olympics since 1984. The 1992 Barcelona Games marked

the first time that biometric access control was expanded from the Olympic

Village to the competitive venues. 1996 Olympic Games used hand geometry

scanning to control access to the Olympic Village, tracking 65,000 people.

Further security overlays have been implemented to control access to national

and press offices at the Olympics.

A hand geometry reader comprises of a CCD or CMOS imaging device that

measures fingers and their combinations or the entire hand in two or three

dimensions. The parameters taken into account include: length and curvature of

fingers; thickness and width of fingers and palm; knuckle shape; distance

between joints, hand surface area, bone structure and translucency. On an

average 90 different parameters are measured over 31,000 points in a full hand

scan that ignores any scars, grease or dirt on the surface. The technology is

less intrusive than iris and retinal scans, that involve passing light through

as sensitive an organ as eyes, and comes with an inherent ease of use, thus

reducing user error rates. Another convenient feature of such a kind of scan is

its extremely small template size that allows for faster processing and at the

same time can be accommodated on a variety of devices including smart cards. The

flip side of using these scanners is their relatively large size and their

comparatively higher cost. Moreover, injuries to hands and debilitating diseases

such as arthritis can render them useless.

Biometrics goes beyond the surface



It's time now to go beyond the skin, deep into the complicated arterial and

venous networks of the body to identify an individual. As an added security

measure, banks are installing security devices that scrutinize veins in an

individual's palm to establish a person's identity. The system uses a palm vein

detector to check out the number of veins in a person's palm, the location of

points at which they cross each other and their size. This is accomplished by

using infrared to scan for hemoglobin in the blood flowing through the veins.

De-oxygenated hemoglobin shows up in black while the surrounding hand area is

lighter in color. This is quite similar to the technique used to capture retinal

patterns. The information thus collected is matched against a database that

stores such information about a customer, to establish his identity. A good

thing about this technology is that it uses infrared radiation to scan a

person's palm, thus making it a non-contact technique as compared to finger

scans, which involves finger contact with the scanning device and hence is

unpopular because of hygiene concerns. The biometric template can also be stored

on a multi-purpose smart card that can be carried around by the individual for

accessing personal accounts.