Biometrics: Eyes Don't Lie

Privacy of personal data is an illusion in today's complex society. With
only passwords, or Social Security Numbers as identity or security measures
every one is vulnerable to invasion of privacy or breach of security.
Traditional means of identification are easily compromised and anyone can use
this information to assume another's identity.
Sensitive personal and corporate information can be assessed and even
criminal activities can be performed using another name.

India's civil aviation ministry is planning high tech methods such as
biometric identification and iris testing at airports, as part of identity
authentication measures. The ministry believes these measures would be more
accurate and less time-consuming.

Eye pattern recognition system provides a barrier to and virtually eliminates
fraudulent authentication and identity privacy and safety controls privileged
access or authorized entry to sensitive sites, data or material. In addition to
privacy protection, there are myriad of applications were iris recognition
technology can provide protection and security. This technology offers the
potential to unlock major business opportunities by providing high confidence
customer validation.

Of all the available biometric techniques, fingerprint identification has
been used the most. Common applications for fingerprint identification include
access control (buildings, ATMs, and computers are examples), forensics, and
driver-license registration. Systems using fingerprint ID systems usually employ
fast processing engines (RISC or DSP) and often, large databases (the FBI's
database contains over 70 mn fingerprints). To prevent someone from using a
fingerprint on a passive surface, such as a piece of sticky tape, newer systems
employ criteria beyond fingerprint matching.

Organizations use voice recognition for access to private areas such as
company intranets, computer networks, and financial databases. Unlike a
fingerprint ID system, which needs a special sensor, voice-recognition equipment
uses standard electronics components. A drawback of voice-recognition systems is
that they are of limited use in a noisy environment.

Pattern recognition is a widely used term for all kind of processes where
specific information is extracted from a large set of individual measurements.

Facial Recognition System

A facial recognition system is a computer-driven application for
automatically identifying a person from a digital image. It does that by
comparing selected facial features in the live image and a facial database.

Among the popular recognition algorithms are eigenface, fisherface, the
Hidden Markov model, and the neuronal motivated Dynamic Link Matching. A newly
emerging trend, claimed to achieve previously unseen accuracies, is
three-dimensional face recognition.

In order to verify the identity of an individual, a facial recognition system
goes through various steps. The first step involves capturing an image of the
individual's face with the use of a video camera. An algorithm called local
feature analysis (LFA) is then used to match distinctive features of the
individual's face to images contained in the system

These distinctive features are called nodal points. There are approximately
80 nodal points on the human face, such as the distance between the eyes, the
shape of jaw line, and the shape of the chin.

Generally, facial recognition systems account for slight variations in these
nodal points that may occur due to changes in facial expressions. After scanning
a database of images and finding a match, the facial recognition system rates
the possibility of the correct match, relying on the information it has
processed. Since many systems have the ability to recognize an individual's
face within seconds, a correct match may be determined while the individual is
still within a specific area.

In 2003, Identix Inc was named as one of the most reliable facial recognition
software companies. Since then, the company has further developed its product,
FaceIt®ARGUS. FaceIt® can provide face finding, face recognition, liveness and
image quality services.

Many facial recognition systems take advantage of Local Feature Analysis (LFA);
however, FaceIt is the only technology to utilize both LFA and skin biometrics,
uniqueness of skin texture.

Application Galore

Face recognition has a wide range of applications such as face-based video
indexing and browsing engines, multimedia management, human-computer
interaction, biometric identity authentication, and surveillance. Interest and
research activities in face recognition have increased significantly in the past
years. This is partly due to recent technology advances initially made by work
on eigenfaces and partly due to increased concerns in security. However, the
problem of face recognition remains a great challenge after several decades of
research.

Problems in Face Recognition

It is difficult for conventional methods to achieve high accuracy. The
appearance of a face is affected by extrinsic factors, including illumination,
pose and expression, as well as inaccuracies made in pre-processing stages such
as face alignment. Variations brought about by extrinsic factors make individual
face manifolds highly complex.

Speed is also an important issue for many applications. To achieve high
accuracy, the recognition should be performed based on intrinsic properties, and
the algorithms should be able to deal with unfavorable influences due to
extrinsic factors and mis-alignment.

Iris Recognition

The iris is the pigmented tissue in the eyeball that surrounds the pupil,
and consists of the muscles that adjust the size of the pupil. This pattern is
unique to each individual, is fixed about two years after birth, and remains
unchanged for the rest of one's life.

Iris recognition is a promising approach for confirming identity on a network
In contrast to systems up to now that used ID cards and passwords, systems that
use iris recognition reference each individual's unique iris code, so it is
possible to maintain an extremely high level of security. Since it is also
possible to basically check one's identity without making contact, there are
no unpleasant feelings like there are when giving fingerprints.

Two types of eye scanning technologies are in use-retinal and iris. No
other biometric offers as much information as our eyes. The potential of
eye-scan identification has been known for decades. Both the iris and the retina
are distinctive, and remain stable over time. Retinal scanning was first
conceived in the 1970s, but the invasiveness and costs associated with imaging
the retina prohibited widespread market acceptance. The concept of iris
recognition was patented in 1986, and the software algorithms to realize the
technology emerged in the mid nineties. Iris recognition is less invasive, less
expensive, and highly accurate.

How It Works: An iris-recognition
algorithm first has to identify the approximately concentric circular outer
boundaries of the iris and the pupil in a photo of an eye. The set of pixels
covering only the iris is then transformed into a bit pattern that preserves the
information that is essential for a statistically meaningful comparison between
two iris images.

The mathematical methods used resemble those of modern lossy compression
algorithms for photographic images. In the case of general
IrisCode, a Garbor wavelet transform is used in order to extract the
spatial frequency range that contains a good best signal-to-noise ratio
considering the focus quality of available cameras. The results are a set of
complex numbers that carry local amplitude and phase information for the iris
image. In IrisCode, all amplitude information is discarded, and the resulting
2,048 bits that represent an iris consist only of the complex sign bits of the
representation of the iris image. Discarding the amplitude information ensures
that the IrisCode remains largely unaffected by changes in illumination and iris
color, which contributes significantly to the long-term stability of the code.
To verify an IrisCode, its Hamming distance to a previously recorded IrisCode
has to be below a suitable selected threshold.

The Challenges: A practical
problem of iris recognition is that the iris is usually partially covered by
eyelids and eyelashes. In order to reduce the false-reject risk in such cases,
additional algorithms are needed to identify the locations of eye lids and eye
lashes, and exclude the bits in the resulting code from the comparison
operation.

Like with most other biometric identification technology, a still not
satisfactorily solved problem with iris recognition is the problem of “live
tissue verification”. The reliability of any biometric identification depends
on ensuring that the signal acquired and compared has actually been recorded
from a live body part of the person to be identified, and is not a manufactured
template. Many commercially available iris recognition systems are easily fooled
by presenting a high-quality photograph of a face instead of a real face, which
makes such devices unsuitable for unsupervised applications, such as door
access-control systems. The problems of live tissue verification is less of a
concern in supervised applications (eg immigration control), where a human
operator supervises the process of taking the picture.

The Road Ahead

Biometric technologies are becoming a part of everyday life as an innovative
way to control access to buildings and information. Researchers are developing
devices that will recognize a person's smell, walk and even DNA.

Iris-recognition technology is one of the more popular options because of its
accuracy and convenience. LG Electronics has installed the iris devices in
banks, airports and other high-security areas. Last summer, the Nine Zero hotel
in Boston became the first hotel to use it. By 2008, sales of biometric
technologies are expected to reach $4.6 bn, up from $1 bn last year. Without a
doubt, they will become a part of everyday life in the near future.

Amardeep Gupta


maildqindia@cybermedia.co.in

The author is head, Department of Computer Science, DAV College, Amritsar