IMPLEMENTATION OF

HAMMING-CUT-MATCHING ALOGORITHM

IN IRIS RECOGINISITION

DEARTMENT OF COMPUTER SCIENCE & ENGINEERING

SASI INSTITUTE OF TECHNOLOGY
TADEPALLIGUDEM

PAPER PREPARED BY:

Udaya Kumar Gandham e-mail:[email protected]

Vidya Sagar LVR e-mail:[email protected]

Contact-no: 9391201984
ABSTRACT
BIOMETRIC SYSTEMS

This paper deals with the basics of The determination, measuring,

iris, its properties and how it adds some and codification of the unique
advantageous features to recognize the characteristic traits that each of us is
correct person. In this paper we discuss born with is known as the science of
the details regarding the information biometrics. Various forms of computer-
about how the iris is located, to based biometrics for personal
distinguish it from other parts of the eye, authentication have been around for the
how the scanner scans the whole pattern past twenty years, but not until recently
of the iris while enrolling and matching have some reached maturity and a
and how the scanned patterns are quality/reliability that has enabled their
converted into 256 bytes of data so that widespread application. In the past,
it can be stored in the database. We hand geometry enjoyed the advantage of
compare the iris codes of the current very small templates (codes containing
person who wants to access the database the biometric data), but with modern
and gives the matched results to the user computers this is no longer the main
accordingly. issue and iris based solutions are steadily
As the iris recognition technology gaining ground.
produces very low false rate when Retinal, iris, and fingerprint
compared to the other biometrics results recognition are mature technologies with
it is very preferable in many systems the most reliable performance. Of the
such as airports, banks, defense, etc.., three methods, iris recognition is the
where the security plays one of the least intrusive (unwanted involvement) with
major role. But in the case of fields greater accuracy.
where the database is huge, the In addition to reliable
comparison time is very high. performance some of the other
This paper includes the advantages of using biometrics are:
implementation of HAMMING-CUT- High security: It is based on physical
MATCHING algorithm which reduces characteristics, which cannot be lost or
the comparison time for matching the stolen.
iris code with database so that we can Certainty/accountability: A specific
use iris recognition in case of huge person, not just a holder of a
databases like voting system. As we token or somebody who knows a
have used hand geometry for verifying PIN/password, has been authenticated.
the identified person, it adds the security Users need not worry about someone
to the whole system. Finally this paper using their token or PIN code without
contains the applications of the their knowledge.
combined features of the recognition of Ease of administration: The problems
iris and verification of hand geometry in of handling forgotten PINs/passwords
airport. and lost/stolen keys or access cards are
eliminated, the benefit is a lot of time
and resources saved.
 Figure.1 Showing the process of the biometrics system

IDENTIFICATION USING IRIS in shades of brown, blue and green, no

RECOGNITION two irises are alike, not even within the
same individual or identical twins.
Iris recognition is one of the
biometric systems which utilize iris PROPERTIES:
patterns as a method of gathering unique
information about an individual. It is Glasses and contact lenses, even colored
considered to be one of the most reliable ones, do not interfere with the process.
biometrics with some of the lowest false In addition, recent medical advances
rejection and false acceptance rates and such as refractive surgery; cataract
so it is less intrusive. surgery and cornea transplants do not
IRIS: The iris, the colored change the iris' characteristics. In fact, it
portion of the eye, is approximately is impossible to modify the iris without
11mm (.433 inches) in diameter and risking blindness. And even a blind
consists of several layers and distinct person can participate. As long as a
features such as furrows, ridges, coronas, sightless eye has an iris, that eye can be
crypts, rings which controls the amount identified by iris recognition.
of light that enters into the eye. Varying
FINDING IRIS IN AN IMAGE: the other hand, does not require such
nearness.
An iris has a mesh-like texture to
it, with numerous overlays and patterns The iris-recognition software
that can measured by the computer. The uses about 256 “degrees of freedom" or
camera such as CCD having a high “points of reference” to search the data
resolution can be set at a distance of four for a match. The iris is found by using an
inches (10 centimeters) to 40 inches (one integrodifferential operator (1), which
meter), depending on the scanning determines the inner and outer
environment. When iris recognition is boundaries of the iris's colored
used for logging on to a personal pigmentation. Not all of the iris is used:
computer or checking in at an airport, a portion of the top, as well as 45˚ of the
people need to be somewhat closer to the bottom, are unused using the masking
camera. An automatic cash machine, on bits to account for eyelids and camera-
light reflections as shown in figure 2.

Figure.2 Differentiating iris from the other parts of the eye

ENCODING BY 2D WAVELET comparison is required the same process
DEMODULTION: is followed but instead of storing the
record it is compared to all the Iris Code
Each isolated iris pattern is then records stored in the database. The
demodulated to extract its phase comparison also doesn’t actually
information using quadrature 2D Gabor compare the image of the iris but rather
wavelets. It amounts to a patch-wise compares the hexadecimal value
phase quantization of the iris pattern, by produced after the algorithms have been
identifying in which quadrant of the applied.
complex plane each resultant phasor lies
when a given area of the iris is projected
onto complex-valued 2D Gabor
wavelets. Such a phase quadrant coding Hamming Distance:
sequences. A desirable feature of the
phase code is that it is a cyclic, or grey In order to compare the stored Iris
code: in rotating between any adjacent Code record with an image just scanned,
phase quadrants, only a single bit a calculation of the Hamming Distance
changes, unlike a binary code in which is required. The Hamming Distance is a
two bits may change, making some measure of the variation between the Iris
errors arbitrarily more costly than others. Code record for the current iris and the
Altogether 2,048 such phase bits (256 Iris Code records stored in the database.
bytes) are computed for each iris, but in Each of the 2048 bits is compared
a major improvement over the earlier against each other, i.e. bit 1 from the
algorithms, now an equal number of current Iris Code and bit 1 from the
masking bits are also computed to stored Iris Code record are compared,
signify whether any iris region is then bit 2 and so on.
obscured by eyelids, contains any
eyelash occlusions, specular reflections,
boundary artifacts of hard contact lenses,
or poor signal-to-noise ratio and thus
should be ignored in the demodulation
code as artifact. Thus an iris pattern is
converted into a sequence of phasor bits
which can be stored in the iris
recognition software and then can
compared for its identification.

STORING AND COMPARING THE

IMAGE:

This record is then stored in a

database for future comparison. When a
 Figure.3 showing the distribution of
hamming distance

Any bits that don’t match are

assigned a value of one and bits that do
match a value of zero. Once all the bits
have been compared, the number of non-
matching bits is divided by the total
number of bits to produce a two-digit
figure of how the two Iris Code records
Hamming False Accept False Reject
differ. For example a Hamming Distance
Distance Probability Probability
of 0.20 means that the two Iris Code
differ by 20%.
.28 1 in 1012 1 in 11,400

.29 1 in 1011 1 in 22,700

.30 1 in 6.2 billion 1 in 46,000

.31 1 in 665 million 1 in 95,000

False Reject Rate (FRR) occurs
when the biometric measurement taken .32 1 in 81 million 1 in 201,000
from the live subject fails to match the
template stored in the biometric system .33 1 in 11 million 1 in 433,000
False Accept Rate (FAR) occurs
.34 1 in 1.7 million 1 in 950,000
when the measurement taken from the
live subject is so close to another
.342 1 in 1.2 million 1 in 1.2 million
subject’s template that a correct match
will be declared by mistake. The point at .35 1 in 295,000 1 in 2.12 million
which the FRR and the FAR are equal is
known as the Crossover Error Rate .36 1 in 57,000 1 in 4.84 million
(CER). The lower the CER, the more
reliable and accurate the system. During .37 1 in 12,300 1 in 11.3 million
recognition mode, this comparison has to
occur between the Iris Code record from
the live subject and every Iris Code Enrolment in an iris recognition
stored in the database before the live system is normally quite fast. The actual
subject is rejected. capturing and testing of the image,
administrative requirements and training
The following table shows the of the subject can usually be
probabilities of false accept and false accomplished in a couple of minutes.
reject with iris recognition technology: Subjects who wear glasses should
remove their glasses during the initial
enrolment in a recognition system to
ensure that the best image is captured
without any reflection from the lenses in
the glasses. Enrolment in an iris
recognition system is normally quite
fast.
The actual capturing and testing
of the image, administrative
requirements and training of the subject
can usually be accomplished in a couple
of minutes. Subjects who wear glasses
should remove their glasses during the
initial enrolment in a recognition system Figure.4 Showing the iris scanner
to ensure that the best image is captured
without any reflection from the lenses in
the glasses. Contact lenses, on the other ADVANTAGES OF IRIS
hand, do not need to be removed as they RECOGNITION:
sit flush with the eye and hence have no
reflections to impede the initial scan.  As the morphogenesis of the
After the initial enrolment most users are iris that occurred during the
able to go through subsequent scanning month
without any additional instruction or of seven remain stable
assistance. Those who wear glasses no throughout life it never changes with
longer have to remove them after initial the time.
enrolment and wearing clear or colored  Uniqueness will be maintained
contact lenses pose no problems. Note between even the multi-birth
that the same eye used during enrolment children and so it is very
must be used during subsequent difficult to forge
comparisons.  There will be no effect on our
eyes since there is no direct
The comparison of a live subject Iris contact.
Code record with all the Iris Code  The accurateness of the
records in the database may seem like a scanning technology is great
large amount of data to process, in with
reality it normally only takes a few error rates being very low.
seconds. This comparison speed is  Scalability and speed is high
obviously affected by the speed of the when compared to other
system processor the database is running systems.
on and the size of the database itself.
VERIFICATION USING HAND
The proximity a user needs to be to GEOMETRY :
the scanning system is usually dependant
on the lens in use and the illumination. This biometric approach uses the
For example, systems scanning at the geometric form of the hand for
desktop PC level can operate with the confirming an individual’s identity.
subject seventeen to nineteen inches Because human hands are not unique,
from the unit. IRIS SCANNER: specific features must be combined to
assure dynamic verification.
The process of hand geometry
includes scanning the features such as
finger curves, thickness and length; implemented hamming-cut-matching
height and width of the back of the hand; algorithm to reduce the comparison time.
the distance between the joints and
overall bone structure. In normal process the iris code
To register in a hand-scan system a is compared with iris code in the
hand is placed on a reader’s covered flat database and the hamming distance is
surface. This placement is positioned by calculated by comparing the whole 2048
five guides or pins that correctly situate bits. If the hamming distance is within
the hand for the cameras. A succession some specified limit the person is given
of cameras captures 3-D pictures of the permission for access otherwise it rejects
sides and back of the hand. The the person’s identity. So, most of the
attainment of the hand-scan is a fast and time is spent for comparing the records
simple process. The hand-scan device in the database itself.
can process the 3-D images in 5 seconds
or less and the hand verification usually But we are comparing the iris
takes less than 1 second. The image codes bit by bit and simultaneously
capturing and verification software and checking whether the hamming distance
hardware can easily be integrated within exceeds the given limit or not. If it
standalone units. Hand-scan applications exceeds the limit, then comparison with
that include a large number of access the particular record is stopped and next
points and users can be centrally record is taken for comparison and so
administered, eliminating the need for on. If both the iris code is going to be
individuals to register on each device. similar then it will proceed with same
After the iris recognition for the record by comparing the next bits. This
identification of the authorized person, reduces the comparison time because we
the hand scanner scans the geometrical are not comparing all the bits present in
structure of the hand and gets the data to database. The hamming-cut-matching-
compare with the template of the algorithm is as follows.
enrolled persons and verifies the
identified person which increases the HAMMING-CUT-MATCHING
security in our implementation. ALGORITHM

OUR IMPLEMENTATION TO START THE PROCESS

REDUCE THE COMPARISON WHILE(FOR ALL RECORDS IN DATA
TIME: BASE)
HAM=0, I=0, K=0
The comparison speed of a WHILE ( I<2047)
READ A[I] ,D[K]
person’s iris code with iris database is
obviously affected by the speed of the /* A is the current iris code bit ,D is the
system processor and the size of the Database iris code bit */
database itself. If the iris recognition
technology is used in large database IF ( A[I] < > D[K])
system like identifying the citizenship of THEN
a person, then the time taken for HAM->HAM+1
comparing his iris code with the IF( HAM = MAX )
database will be very high. So we have
 /* HAM is the hamming EXIT PROGRAM
match MAX is the maximum limit for hamming X: TRANSFER THE CONTROL TO
Distance */
THE NEXT DATABASE RECORD
THEN GOTO X; END WHILE
ELSE I->I+1, K->K+1 PRINT ( “YOU ARE NOT OUR
AUTHORISED PERSON”);
ENDIF STOP THE PROCESS
ENDIF
END WHILE
PRINT ( “YOR ARE THE CORRECT
PERSON”);

ADVANTAGES: drive. The databases will not be shared

1. The comparison time for iris with other organizations.
code matching is greatly It is designed to allow only
reduced by our hamming-cut- authorized people around the airport
matching algorithm. terminal in certain areas. But there is a
2. We have IMPLEMENTED chance that an unauthorized person had
THIS HAMMING-CUT- time to dash in behind an authorized
MATCHING ALGORITHM person once the iris-recognition allowed
IN A SIMPLE FORMAT IN access -- a practice it termed
C. "piggybacking" or "tail-gating."

IMPLEMENTATION IN AIRPORT To avoid this problem we can have

SECURITY: two doorways. The first door opens upon
iris recognition and hand geometry and
At an airport, for example, a then closes behind the person, leaving
person's name and airline information them in a small portal. The next door
accompanies the iris code, but to protect will only open upon confirming there is
people's privacy no further data -- a just one person in the portal. Both the
Social Security number is collected. The identification and verification can be
stored file is only 512 bytes with a implemented along with our added
resolution of 640 x 480, allowing for security in airport.
massive storage on a computer's hard
 Figure.5 Showing the iris recognition in airport

CONCLUSION: speed required to minimize

user frustration when
 Passwords, token cards and accessing company systems.
PINs are all risks to the  Users no longer have to
security of an organization worry about remembering
due to human nature. Our passwords and system
inability to remember administrators no longer need
complex passwords and to worry about the never-
tendency to write these down ending problem of users
along with losing token cards disclosing passwords or
or forgetting PINs all having weak passwords that
contribute to the possible are easily cracked.
breakdown in security for an  As according to our
organization. implementation the
 The uniqueness of the iris comparison time of the iris
and low probability of a false code with the iris database is
acceptance or false rejection very much less from the
all contribute to the benefits current system.
of using iris recognition
technology. If a two-factor authentication system is
 It provides an accurate and implemented, iris recognition with a
secure method of hand scanner for the verification, then
authenticating users onto the strength of authentication increases
company systems, is a non- and provides another part to defense in
intrusive method and has the depth for the airportr