14-06-2009, 12:28 AM
IRIS SCANNING
A FULL SEMINAR REPORT
Submitted by
MARIYATH K.A
in partial fulfillment for the award of the degree
of
BACHELOR OF TECHNOLOGY
in
COMPUTER SCIENCE & ENGINEERING
SCHOOL OF ENGINEERING
COCHIN UNIVERSITY OF SCIENCE &
TECHNOLOGY,
KOCHI-682 022
NOVEMER 2008Page 2
DIVISION OF COMPUTER SCIENCE & ENGINEERING
SCHOOL OF ENGINEERING
COCHIN UNIVERSITY OF SCIENCE & TECHNOLOGY,
KOCHI-682022
Certificate
Certified that this is a bonafide record of the seminar work entitled
IRIS SCANNING
done by the following student
MARIYATH K.A
Of the VII th semester Computer Science and Engineering in the year
2008 in partial fulfillment of the requirements to the award of Degree
of Bachelor of Technology in Computer Science and Engineering of
Cochin University of Science and Technology
Mrs. Rahna P Muhammad
Dr. David Peter S
Seminar Guide
Head of the
Department
Date
age 3ABSTRACT
Iris Scanning is a method of biometric authentication that uses pattern
recognition technique based on high resolution of the irises of an individual
eye. Biometrics is an automated method of capturing a personâ„¢s unique
biological data that distinguishes him or her from another individual. Iris
recognition has emerged as one of the most powerful and accurate
identification techniques in the modern world. It has proven to be most fool
proof technique for the identification of individuals with out the use of cards.
PINâ„¢s and passwords. It facilitates automatic identification where by
electronic transactions or access to places, information or accounts are made
easier, quicker and more secure.Page 4
ACKNOWLEDGEMENT
At the outset, we thank the lord almighty for the grace, strength and hope to
make our endeavor a success
We express our deep felt gratitude to Dr. David Peter S, Head of the
Division, Computer science for his constant encouragement.
I profoundly grateful to Mrs. Rahna P Muhammad Lecturer ,Department
of Computer Science , my mentor and seminar guide for her valuable
Guidance support ,suggestions and encouragement
Further more I would like to thank all others especially our parents and
numerous friends. This seminar would not have been a success without the
inspiration, valuable suggestions and moral support from the through out its
course
MARIYATH K.APage 5
TABLE OF CONTENTS
CHAPTER TITLE
PAGENO
NO.
LIST OF FIGURES
iii
1. INTRODUCTION
1
1.1 Biometrics-future of identity
1
2. IRIS RECOGNITION
4
2.1 Anatomy, physiology and development of iris
4
2.2 Iris as a powerful identifier
6
2.3 History and development
7
2.4 Science behind the technology 8
2.4.1 Image acquisition
8
2.4.2 Iris localization
10
2.4.3 Pattern matching
12
2.5 Mathematical explanation
13
2.5.1 Accuracy 13
2.5.2 Decision environment
15
2.6 Comparison between genetically identified iris patterns
17
2.7 Uniqueness of iris code
17
2.7.1 Independence of bits across iris codes
17
2.8 Binomial distribution of iris code hamming
18
2.8.1 Distances
18
2.9 Commensurability of iris codes
20Page 6
2.10 Advantages
21
2.11 Disadvantages of using iris for recognition
21
2.12 Applications
22
2.13 Iris recognition issues
25
3. CONCLUSION
26
4. REFERENCES
27Page 7
LIST OF FIGURES
FIG NO
TITLE
PAGE NO
1.1 Topology of identification methods
2
1.2 Comparison between cost and accuracy 3
2.1 A typical iris
5
2.2 Block diagram of iris recognition
8
2.3 Image acquisition rings for automated iris recognition
9
2.4 Iris code
11
2.5 Decision environment
15
2.6 Independence of bits across iris codes
18
2.7 Binomial distribution of iris code hamming distances
19
2.8 Identifying the mystery woman
24Page 8
1. INTRODUCTION
In todayâ„¢s information age it is not difficult to collect data about an individual and
use that information to exercise control over the individual. Individuals generally do not
want others to have personal information about them unless they decide to reveal it. With
the rapid development of technology, it is more difficult to maintain the levels of privacy
citizens knew in the past. In this context, data security has become an inevitable feature.
Conventional methods of identification based on possession of ID cards or exclusive
knowledge like social security number or a password are not altogether reliable. ID cards
can be almost lost, forged or misplaced: passwords can be forgotten. Such that an
unauthorized user may be able to break into an account with little effort. So it is need to
ensure denial of access to classified data by unauthorized persons. Biometric technology
has now become a viable alternative to traditional identification systems because of its
tremendous accuracy and speed. Biometric system automatically verifies or recognizes
the identity of a living person based on physiological or behavioral characteristics. Since
the persons to be identified should be physically present at the point of identification,
biometric techniques gives high security for the sensitive information stored in
mainframes or to avoid fraudulent use of ATMs.This paper explores the concept of Iris
recognition which is one of the most popular biometric techniques. This technology finds
applications in diverse fields.
1.1 Biometrics-future of identity
Biometric dates back to ancient Egyptians who measured people to identify them.
Biometric devices have three primary components.
1. Automated mechanism that scans and captures a digital or analog image of a
living personal characteristic
2. Compression, processing, storage and comparison of image with a stored data.
3. Interfaces with application systems.
A biometric system can be divided into two stages: the enrolment module and the
identification module. The enrolment module is responsible for training the system toPage 9
identity a given person. During an enrolment stage, a biometric sensor scans the personâ„¢s
physiognomy to create a digital representation. A feature extractor processes the
representation to generate a more compact and expressive representation called a
template. For an iris image these include the various visible characteristics of the iris such
as contraction, Furrows, pits, rings etc. The template for each user is stored in a biometric
system database. The identification module is responsible for recognizing the person.
During the identification stage, the biometric sensor captures the characteristics of the
person to be identified and converts it into the same digital format as the template. The
resulting template is fed to the feature matcher, which compares it against the stored
template to determine whether the two templates match.
The identification can be in the form of verification, authenticating a claimed
identity or recognition, determining the identity of a person from a database of known
persons. In a verification system, when the captured characteristic and the stored template
of the claimed identity are the same, the system concludes that the claimed identity is
correct. In a recognition system, when the captured characteristic and one of the stored
templates are the same, the system identifies the person with matching template.
Fig 1.1 Topology of identification methods
Biometrics encompasses both physiological and behavioral characteristics. A
physiological characteristic is a relatively stable physical feature such as finger print, iris
pattern, retina pattern or a Facial feature. A behavioral trait in identification is a personâ„¢s
AUTOMATED
BIOMETRICS
PHYSIOLOGICAL
BEHAVIORAL
FACE
FINGER
HAND
EYE
SIGNATURE
VOICE
KEY
STROKEPage 10
signature, keyboard typing pattern or a speech pattern. The degree of interpersonal
variation is smaller in a physical characteristic than in a behavioral one. For example, the
personâ„¢s iris pattern is same always but the signature is influenced by physiological
characteristics.
Disadvantages
Even though conventional methods of identification are indeed inadequate, the
biometric technology is not as pervasive and wide spread as many of us expect it to be.
One of the primary reasons is performance. Issues affecting performance include
accuracy, cost, integrity etc.
Accuracy
Even if a legitimate biometric characteristic is presented to a biometric system,
correct authentication cannot be guaranteed. This could be because of sensor noise,
limitations of processing methods, and the variability in both biometric characteristic as
well as its presentation.
Cost
Cost is tied to accuracy; many applications like logging on to a pc are sensitive to
additional cost of including biometric technology.
Fig 1.2. Comparison between cost and accuracy
2. IRIS RECOGNITION
HAND
SIGNATURE
IRIS
FACE
FINGER
VOICE
COST
ACCURACY
RETINAPage 11
Iris identification technology is a tremendously accurate biometric. Iris
recognition leverages the unique features of the human iris to provide an unmatched
identification technology. So accurate are the algorithms used in iris recognition that the
entire planet could be enrolled in an iris database with only a small chance of false
acceptance or false rejection. The technology addresses the FTE (Failure to Enroll)
problems which lessen the effectiveness of other biometrics. Only the iris recognition
technology can be used effectively and efficiently in large scale identification
implementations. The tremendous accuracy of iris recognition allows it, in many ways,
to stand apart from other biometric technologies.
2.1 Anatomy ,physiology and development of the iris
The word IRIS dates from classical times (a rainbow). The iris is a Protective
internal organ of the eye. It is easily visible from yards away as a colored disk, behind the
clear protective window of the cornea, surrounded by the white tissue of the eye. It is the
only internal organ of the body normally visible externally. It is a thin diaphragm
stretching across the anterior portion of the eye and supported by lens. This support gives
it the shape of a truncated cone in three dimensions. At its base the eye is attached to the
eyeâ„¢s ciliary body. At the opposite end it opens into a pupil. The cornea and the aqueous
humor in front of the iris protect it from scratches and dirt, the iris is installed in its own
casing. It is a multi layered structure. It has a pigmented layer, which forms a coloring
that surrounds the pupil of the eye. One feature of this pupil is that it dilates or contracts
in accordance with variation in light intensity.
The human iris begins to form during the third month of gestation. The structures
creating its distinctive pattern are completed by the eighth month of gestation hut
pigmentation continues in the first years after birth. The layers of the iris have both
ectodermic and embryological origin, consisting of: a darkly pigmented epithelium,
pupillary dilator and sphincter muscles, heavily vascularized stroma and an anterior layer
chromataphores with a genetically determined density of melanin pigment granules. The
combined effect is a visible pattern displaying various distinct features such as archingPage 12
ligaments, crypts, ridges and zigzag collaratte. Iris color is determined mainly by the
density of the stroma and its melanin content, with blue irises resulting from an absence
of pigment: long wavelengths are penetrates and is absorbed by the pigment epithelium,
while shorter wavelengths are reflected and scattered by the stroma. The heritability and
ethnographic diversity of iris color have long been studied. But until the present research,
little attention had been paid to the achromatic pattern complexity and textural variability
of the iris among individuals.
A permanent visible characteristic of an iris is the trabecular mesh work, a tissue
which gives the appearance of dividing the iris in a radial fashion. Other visible
characteristics include the collagenous tissue of the stroma, ciliary processes, contraction
furrows, crypts, rings, a corona and pupillary frill coloration and sometimes freckle. The
striated anterior layer covering the trabecular mesh work creates the predominant texture
with visible light.
Fig 2.1. A Typical IrisPage 13
2.2 Iris as a powerful identifier
Iris is the focus of a relatively new means of biometric identification. The iris is
called the living password because of its unique, random features. It is always with you
and can not be stolen or faked. The iris of each eye is absolutely unique. The probability
that any two irises could be alike is one in 10 to 78
th
power †the entire human
population of the earth is roughly 5.8 billion. So no two irises are alike in their details,
even among identical twins. Even the left and right irises of a single person seem to be
highly distinct. Every iris has a highly detailed and unique texture that remains stable
over decades of life. Because of the texture, physiological nature and random generation
of an iris artificial duplication is virtually impossible.
The properties of the iris that enhance its suitability for use in high confidence
identification system are those following.
1. Extremely data rich physical structure about 400 identifying features
2. Genetic independence no two eyes are the same.
3. Stability over time.
4. Its inherent isolation and protection from the external environment.
5. The impossibility of surgically modifying it without unacceptable risk to vision.
6. Its physiological response to light, which provides one of several natural tests
against artifice.
7. The ease of registering its image at some distance forms a subject without
physical contact. unobtrusively and perhaps inconspicuously
8. It intrinsic polar geometry which imparts a natural co-ordinate system and an
origin of co-ordinates
9. The high levels of randomness in it pattern inter subject variability spanning
244 degrees of freedom - and an entropy of 32 bits square million of iris tissue.
2.3 History and developmentPage 14
The idea of using patterns for personal identification was originally proposed in
1936 by ophthalmologist Frank Burch. By the 1980â„¢s the idea had appeared in James
Bond films, but it still remained science fiction and conjecture. In 1987, two other
ophthalmologists Aram Safir and Leonard Flom patented this idea and in 1987 they asked
John Daugman to try to create actual algorithms for this iris recognition. These
algorithms which Daugman patented in 1994 are the basis for all current iris recognition
systems and products.
Daugman algorithms are owned by Iridian technologies and the process is
licensed to several other Companies who serve as System integrators and developers of
special platforms exploiting iris recognition in recent years several products have been
developed for acquiring its images over a range of distances and in a variety of
applications. One active imaging system developed in 1996 by licensee Sensar deployed
special cameras in bank ATM to capture IRIS images at a distance of up to 1 meter. This
active imaging system was installed in cash machines both by NCR Corps and by
Diebold Corp in successful public trials in several countries during I997 to 1999. a new
and smaller imaging device is the low cost Panasonic Authenticam digital camera for
handheld, desktop, e-commerce and other information security applications. Ticket less
air travel, check-in and security procedures based on iris recognition kiosks in airports
have been developed by eye ticket. Companies in several, countries are now using
Daughmanâ„¢s algorithms in a variety of products.
2.4 Science behind the technologyPage 15
The design and implementation of a system for automated iris recognition can be
subdivided in to 3.
1. image acquisition
2. iris localization and
3. Pattern matching
Fig 2.2 Block Diagram of Iris Recognition
2.4.1 Image acquisition
The iris recognition process begins with video-based image acquisition a process
which deals with the capturing of a high quality image of the iris while remaining non-
invasive to the human operator. There are 3 important requisites for this process
a) It is desirable to acquire images of the iris with sufficient resolution and
sharpness to support recognition
b) It is important to have good contrast in the interior iris pattern without
restoring to a level of illumination that annoys the Operator, that is adequate
intensity of source constrained by operators comfort with brightness.
c) These images must be well framed without unduly constraining the operator.
The widely used recognition system is the daugmen system which captures
images with the iris diameter typically between 100 and 200 pixels from a
distance of 15, 46 cm using a 330 mm lens.
Operator
Image
Acquisition
Image
Iris
Localization
Digitizing
Image
Numeric
Code
Pattern
Matching
Bar Codes Stored
In Computer
DatabasePage 16
Fig 2.3 Image acquisition rings for automated iris recognition
Image acquisition is performed as follows. It uses LED based point light sources
in conjunction with a wide angle camera no more than 3 feet from the subjectâ„¢s eye. By
carefully positioning the light source below the operator, reflection of point source can be
avoided in the imaged iris. The system makes use of light, which is visible to human eye.
Infrared illumination can also be employed. This System requires the operator to self
position his eye in front of the camera. It provides the operator with a live video feed
back via beam splitter. This allows the operator to see what the camera is capturing and
to adjust his position. Once a series of images of sufficient quality is acquired, it is
automatically forwarded for subsequent processing.
2.4.2 Iris localization
Image acquisition of iris can be expected to yield an image containing only the
iris. Rather image acquisition will capture the iris as part of a larger image that also
contains data derived from the surrounding eye region. Prior to performing iris pattern
CAMERA
LENS
EY
E
LIGHT
LCD DISPLAY
FRAME
GRABBER
BEAM SPLITTERPage 17
matching it is important to localize that portion of the image that corresponds to iris. Iris
localization is a process that delimits the iris from the rest of the acquired image. After
the camera situates the eye, the Daugmanâ„¢s algorithm narrows in from the right and left
of the iris to locate its outer edge. This horizontal approach accounts for obstruction
caused by the eyelids. It simultaneously locates the inner edge of the iris, excluding the
lower 90
0
because of inherent moisture and lighting issues.
Conversion of an iris image into a numeric code that can be easily manipulated is
essential to its use. This process developed by John Daugman. Permits efficient
comparison of irises. Upon the location of the iris, an iris code is computed based on the
information from a set of Gabor wavelets. The Gabor wavelet is a powerful tool to make
iris recognition practical. These wavelets are specialized filter banks that extract
information from a signal at a variety of locations and scales. The filters are members of
a family of functions developed by Dennis Gabor, that optimizes the resolution in both
spatial and frequency domains. The 2-D Gabor wavelets filter and map segments of iris
into hundreds of vectors. The wavelets of various sizes assign values drawn from the
orientation and spatial frequency of select areas, bluntly referred to as the what of the
sub-image, along with the position of these areas, bluntly referred to as the where. The
what and where are used to form the Iris Code. Not all of iris is used: a portion of the
top, as well as 45
0
of the bottom is unused to account for eyelids and cameraâ€light
reflections. The iris Code is calculated using 8 circular bands that have been adjusted to
the iris and pupil boundaries.
Iris recognition technology converts the visible characteristics of the iris into a
512 byte Iris Code, a template stored for future verification attempts. 5l2 bytes is a fairly
compact size for a biometric template, but the quantity of information derived from the
iris is massive. From the iris 11 mm diameters, Dr. Daugmanâ„¢s algorithms provide 3.4
bits of data per square mm. This density of information is such that each iris can be said
to have 266 unique spots, as opposed to 13- 60 for traditional biometric technology.
This 266 measurement is cited in all iris recognition literature, after allowing for the
algorithms for relative functions and for characteristics inherent to most human eyes. Dr.Page 18
Daugman concludes that 173 independent binary degrees of freedom can be extracted
from his algorithm-and exceptionally large number fur a biometric, for future
identification, the database will not be comparing images of iris, but rather hexadecimal
representations of data returned by wavelet filtering and mapping. The Iris Code for an
iris is generated within one second. Iris Code record is immediately encrypted and cannot
be reverse engineered.
Fig 2.4 Iris code
2.4.3 Pattern matching
When a live iris is presented for comparison, the iris pattern is processed and
encoded into 512 byte Iris Code. The Iris Code derived from this process is comparedPage 19
with previously generated Iris Code. This process is called pattern matching. Pattern
matching evaluates the goodness of match between the newly acquired iris pattern and
the candidateâ„¢s data base entry. Based on this goodness of match final decision is taken
whether acquired data does or doesnâ„¢t come from the same iris as does the database entry.
Pattern matching is performed as follows. Using integer XOR logic in a single
clock cycle, a long vector of each to iris code can be XORed to generate a new integer.
Each of whose bits represent mismatch between the vectors being compared. The total
number of 1s represents the total number of mismatches between the two binary codes.
The difference between the two recodes is expressed as a fraction of mismatched bits
termed as hamming distance. For two identical Iris Codes, the hamming distance is Zero.
For perfectly unmatched Iris Codes, the hamming distance is 1. Thus iris patterns are
compared. The entire process i.e. recognition process takes about 2 seconds. A key
differentiator for iris recognition is its ability to perform identification using a one to
many search of a database, with no limitation on the number of iris code records
contained there in.Page 20
2.5 Mathematical explanation
An Iris Code is constructed by demodulation of the iris pattern. This process
uses complex-valued 2D Gabor wavelets to extract the structure of the iris as a sequence
of phasors, whose phase angles are quantized to set the bits in the first code.
This process is performed in a doublyâ€dimensionless polar co-ordinate system
that is invariant to the size of the iris, and also invariant to the dilation diameter of the
pupil within the iris.
The demodulating wavelets are parameterized with four degrees-of-freedom:
Size, orientation and two positional co-ordinates. They span several octaves in size, in
order to extract iris structure at many different scales of analysis. Because the information
extracted from the iris is inherently described in terms of phase, it is insensitive to
contrast, camera gain and illumination level. The phase description is very compact,
requiring only 256 bytes to represent each iris pattern. These 2D wavelets are optimal
encoders under the inherent Heisenbergâ€Weyl uncertainty relation for extraction of
information in conjoint spatial-spectral representations.
The recognition of irises by their recodes is based upon the failure of a test of
statistical independence. Any given Iris Code is statistically guaranteed to pass a test of
independence against any Iris Code computed from a different eye; but it will uniquely
fail the same test against the eye from which it was composed. Thus the key to iris
recognition is the failure ofâ„¢ a test of statistical independence.
2.5.1 Accuracy
The Iris Code constructed from these Complex measurements provides such a
tremendous wealth of data that iris recognition offers level of accuracy orders of
magnitude higher than biometrics. Some statistical representations of the accuracy
follow
age 21¢ The odds of two different irises returning a 75% match (i.e. having Hamming
Distance of 0.25): 1 in 10
16
.
¢ Equal Error Rate (the point at which the likelihood of a false accept and false
reject are the same): 1 in 12 million.
¢ The odds of two different irises returning identical Iris Codes: 1 in 10
52
Other numerical derivations demonstrate the unique robustness of these
algorithms. A personâ„¢s right and left eyes have a statistically insignificant increase in
similarity: 0.0048 on a 0.5 mean. This serves to demonstrate the hypothesis that iris shape
and characteristic are phenotype - not entirely; determined by genetic structure. The
algorithm can also account for the iris: even if 2/3 of the iris were completely obscured,
accurate measure of the remaining third would result in an equal error rate of 1 in
100000.
Iris recognition can also accounts for those ongoing changes to the eye and iris
which are defining aspects of living tissue. The pupilâ„¢s expansion and contraction, a
constant process separate from its response to light, skews and stretches the iris. The
algorithms account for such alteration after having located at the boundaries of the iris.
Dr. Daugman draws the analogy to a Ëœhomogenous rubber sheet which, despite its
distortion retains certain consistent qualities. Regardless of the size of the iris at any
given time, the algorithm draws on the same amount or data, and its resultant Iris Code is
stored as a 512 byte template. A question asked of all biometrics is there is then ability to
determine fraudulent samples. Iris recognition can account for this in several ways the
detection of pupillary changes, reflections from the cornea detection of contact lenses
atop the cornea and use of infrared illumination to determine the state of the sample eye
tissue.Page 22
2.5.2 Decision Environment
Fig 2.5 Decision environment
The performance of any biometric identification scheme is characterized by its
Decision Environmentâ„¢. This is a graph superimposing the two fundamental histograms
of similarity that the test generates: one when comparing biometric measurements from
the SAME person (different times, environments, or conditions), and the other when
comparing measurements from DIFFERENT persons. When the biometric template of a
presenting person is compared to a previously enrolled database of templates toPage 23
determine the Persons identity, a criterion threshold (which may be adaptive) is applied to
each similarly score. Because this determines whether any two templates are deemed to
be same or different, the two fundamental distributions should ideally be well
separated as any overlap between them causes decision errors.
One metric for decidability, or decisionâ€making power, is d. This is defined as
the separation between the means of two distributions, divided by the square root of their
average variance. One advantage of using dâ„¢ for comparing, the decision-making power
of biometrics is the fact that it does not depend on any choice about the decision
threshold used. Which of course may vary from liberal to conservative when selecting the
trade-off between the False Accept Rate (FAR) and False Reject Rate (FRR)? The dâ„¢
metric is a measure of the inherent degree to which any decrease in one error rate must be
paid for by an increase in the other error rate, when decision thresholds are varied. It
reflects the intrinsic separability of the two distributions.
Decidability metrics such as dâ„¢ can he applied regardless of what measure of
similarity a biometric uses. In the particular case of iris recognition, the similarly measure
is a Hamming distance: the fraction of bits in two iris Codes that disagree. The
distribution on the left in the graph shows the result when different images of the same
eye are compared: typically about 10% of the bits may differ. But when Iris Codes from
different eyes are compared: With rations to look for and retain the best match. The
distribution on the right is the result. The fraction of disagreeing bits is very tightly
packed around 45%. Because of the narrowness of this right-hand distribution, which
belongs to the family of binomial extreme-value distributions, it is possible to make
identification decisions with astronomical levels of confidence. For example, the odds of
two different irises agreeing just by chance in more than 75 of their Iris Code bits, is only
one in 10-to-the- 14
th
power. These extremely low probabilities of getting a False Match
enable the iris recognition algorithms to search through extremely large databases, even
of a national or planetary scale, without confusing one Iris Code for another despite so
many error opportunities.Page 24
2.6 Comparison between genetically identical iris patterns
Although the striking visual similarity of identical twins reveals the genetic
penetrance of overall facial appearance, a comparison of genetically identical irises
reveals just the opposite for iris patterns: the iris texture is an epigenetic phenotypic
feature, not a genotypic feature. A convenient source of genetically identical irises is the
right and left pair from any given person. Such pairs have the same genetic relationship
as the four irises of two identical twins, or indeed in the probable future, the 2N irises of
N human clones. Eye color of course has high genetic penetrance, as does the overall
statistical quality of the iris texture, but the textural details are uncorrelated and
independent even in genetically identical pairs. So performance is not limited by the birth
rate of identical twins or the existence of genetic relationships.
2.7 Uniqueness of iris code
2.7.1 Independence of bits across iris codes
It is important to establish and to measure the amount of independent variation
both within an iris and between different irises. There are correlations within an iris
because local structure is self-predicting; for example, furrows tend to propagate
themselves radially. Such self-correlations limit the number of degrees of freedom within
irises. But even more important is the question of whether systematic correlations exist
between different irises.Page 25
Fig 2.6. Independence of bits across iris codes
This probability distribution suggests that they do not. It plots the probability that
bits in different positions within the Iris Code are set to 1, for a randomly sampled
population of different Iris Codes. The fact that this distribution hovers near 0.5 indicates
that all Iris Code bits are equally likely to be 0 or 1. This property makes Iris Codes
maximum entropy codes in a bit-wise sense. The fact that this distribution is uniform
indicates that different irises do not systematically share any common structure. For
example, if most irises had a furrow or crypt in the 12-oâ„¢clock position, then the plot
shown here would not be flat. The recognition of persons by their Iris Codes is based
Upon the failure of a test of statistical independence. The plot shown here illustrates why
any given Iris Code is statistically guaranteed to pass a test of independence against
any Iris Code computed from a different eye.
2.8 Binomial distribution of iris code hamming
2.8.1 Distances
The Histogram given below shows the outcomes of 2,307,025 comparisons
between different pairs of irises. For each pair comparison, the percentage of their Iris
Code bits that disagreed was computed and tallied as a fraction. Because of the zero-
mean property of the wavelet demodulators, the computed coding bits are equally likelyPage 26
to be 1 or 0. Thus when any corresponding bits of two different Iris Codes are compared,
each of the four combinations (00), (01), (10), (11) has equal probability. In two of these
cases the bits agree, and in the other two they disagree. Therefore one would expect on
average 50% of the bits between two different Iris Codes to agree by chance. The above
histogram presenting comparisons between 2.3 million different pairings of irises shows a
mean fraction of 0.499 of their Iris Code bits agreeing by chance.
Fig 2.7 .B inomial distribution of iris code hamming distances
The standard deviation of this distribution, 0.032, reveals the effective number of
independent bits (binary degrees of freedom) when Iris Codes are compared. Because of
correlations within irises and within computed Iris Codes, the number of degrees of
freedom is considerably smaller than the number of bits computed. But even correlated
Bernoulli trials (coin tosses) generate binomial distributions; the effect of theirPage 27
correlations is equivalent to reducing the effective number of Bernoulli trials. For
comparisons between different pairs of Iris Codes, the distribution shown above
corresponds to that for the fraction of "heads" that one would get in runs of 244 tosses of
a fair coin. This is a binomial distribution, with parameters p=q=0.5 and N=244 Bernoulli
trials (coin tosses). The solid curve in the above histogram is a plot of such a binomial
probability distribution. It gives an extremely exact fit to the observed distribution, as
may be seen by comparing the solid curve to the data histogram.
The above two aspects show that Hamming Distance comparisons between
different Iris Codes are binomially distributed, with 244 degrees of freedom. The
important corollary of this conclusion is that the tails of such distributions are dominated
by factorial combinatorial factors, which attenuate at astronomic rates. This property
makes it extremely improbable that two different Iris Codes might happen to agree just
by chance in, say, more than 2/3rds of their bits (making a Hamming Distance below 0.33
in the above plot). The confidence levels against such an occurrence are the reason why
iris recognition can afford to search extremely large databases, even on a national scale,
with negligible probability of making even a single false match.
2.9 Commensurability of iris codes
A critical feature of this coding approach is the achievement of commensurability
among iris codes, by mapping all irises into a representation having universal format and
constant length, regardless of the apparent amount of iris detail. In the absence of
commensurability among the codes, one would be faced with the inevitable problem of
comparing long codes with short codes, showing partial agreement and partial
disagreement in their lists of features. It is not obvious mathematically how one would
make objective decisions and compute confidence levels on a rigorous basis in such a
situation. This difficulty has hampered efforts to automate reliably the recognition of
fingerprints. Commensurability facilitates and objectifies the code comparison process, as
well as the computation of confidence.Page 28
2.10 Advantages
¢
Highly protected, internal organ of the eye
¢
Externally visible; patterns imaged from a distance
¢
Iris patterns possess a high degree of randomness
o
Variability: 244 degrees-of-freedom
o
Entropy: 3.2 bits per square-millimeter
o
Uniqueness: set by combinatorial complexity
¢
Changing pupil size confirms natural physiology
¢
Pre-natal morphogenesis (7th month of gestation)
¢
Limited genetic penetrance of iris patterns
¢
Patterns apparently stable throughout life
¢
Encoding and decision-making are tractable
¢
Image analysis and encoding time: 1 second
¢
Decidability index (d-prime): d' = 7.3 to 11.4
¢
Search speed: 100,000 Iris Codes per second
2.11 Disadvantages of using iris for identification
¢ Small target (1 cm) to acquire from a distance (1m)
¢ Located behind a curved, wet, reflecting surface
¢ Obscured by eyelashes, lenses, reflections
¢ Partially occluded by eyelids, often drooping
¢ Deforms non-elastically as pupil changes size
¢ Illumination should not be visible or bright
¢ Some negative connotationsPage 29
2.12 Applications
Iris-based identification and verification technology has gained acceptance in a
number of different areas. Application of iris recognition technology can he limited only
by imagination. The important applications are those following:
¢ ATM™s and iris recognition: in U.S many banks incorporated iris recognition
technology into ATMâ„¢s for the purpose of controlling access to oneâ„¢s bank
accounts. After enrolling once (a 30 second process), the customer need only
approach the ATM, follow the instruction to look at the camera, and be
recognized within 2-4 seconds. The benefits of such a system are that the
customer who chooses to use bankâ„¢s ATM with iris recognition will have a
quicker, more secure transaction.
¢ Tracking Prisoner Movement: The exceptionally high levels of accuracy
provided by iris recognition technology broadens its applicability in high risk,
high-security installations. Iris scan has implemented their devices with great
success in prisons in Pennsylvania and Florida. By this any prison transfer or
release is authorized through biometric identification. Such devices greatly ease
logistical and staffing problems.
Applications of this type are well suited to iris recognition technology. First,
being fairly large, iris recognition physical security devices are easily integrated into the
mountable, sturdy apparatuses needed or access control, The technologyâ„¢s phenomenal
accuracy can be relied upon to prevent unauthorized release or transfer and to identify
repeat offenders re-entering prison under a different identity.Page 30
¢ Computer login: The iris as a living password.
¢ National Border Controls: The iris as a living password.
¢ Telephone call charging without cash, cards or PIN numbers.
¢ Ticket less air travel.
¢ Premises access control (home, office, laboratory etc.).
¢ Driving licenses and other personal certificates.
¢ Entitlements and benefits authentication.
¢ Forensics, birth certificates, tracking missing or wanted person
¢ Credit-card authentication.
¢ Automobile ignition and unlocking; anti-theft devices.
¢ Anti-terrorism (e.g.:†suspect Screening at airports)
¢ Secure financial transaction (e-commerce, banking).
¢ Internet security, control of access to privileged information.
¢ Biometricâ€key Cryptography for encrypting/decrypting messages.Page 31
Fig 2.8 .Identifying the mystery woman
Iris recognition system is also finding unexpected applications. The best know
example involved using iris recognition to confirm the identification of a mysterious
young afghan woman named Sharbat Gula originally photographed by Steve McCurry in
1984.Some 18 years later, McCurry photographed Sharbat Gula in Afghanistan .At the
behest of National Geographic, Dr.John Dougman,developer of the Iris recognition
system, then compared the irises in the photographs using his algorithms. He concluded
that the eyes were a match.Page 32
2.13 Iris Recognition: Issues
Every biometric technology has its own challenges. When reviewing test results,
it is essential to consider the environment and protocols of the test. Much industry testing
is performed in laboratory settings on images acquired in ideal conditions. Performance
in a real world application may result in very different performance as there is a learning
curve for would-be user of the system and not every candidate will enroll properly or
quickly the first time. There are some issues which affect the functionality and
applicability of iris recognition technology in particular.
The technology requires a certain amount of user interaction the enroller must
hold still in a certain spot, even if only momentarily. It would be very difficult to enroll
or identify a non-cooperative subject. The eye has to have a certain degree of lighting to
allow the camera to capture the iris; any unusual lighting situation may affect the ability
of the camera to acquire its subject. Lastly, as with any biometric, a backup plan must be
in place if the unit becomes inoperable. Network crashes, power failure, hardware and
software problems are but a few of the possible ways in which a biometric system would
become unusable. Since iris technology is designed to be an identification technology, the
fallback procedures may not be as fully developed as in a recognition schematic. Though
these issues do not reduce the exceptional effectiveness of iris recognition technology,
they must be kept in mind, should a company decide to implement on iris-based solution.Page 33
3
CONCLUSION
The technical performance capability of the iris recognition process far surpasses
that of any biometric technology now available. Iridian process is defined for rapid
exhaustive search for very large databases: distinctive capability required for
authentication today. The extremely low probabilities of getting a false match enable the
iris recognition algorithms to search through extremely large databases, even of a
national or planetary scale. As iris technology grows less expensive, it could very likely
unseat a large portion of the biometric industry, e-commerce included; its technological
superiority has already allowed it to make significant inroads into identification and
security venues which had been dominated by other biometrics. Iris-based biometric
technology has always been an exceptionally accurate one, and it may soon grow much
more prominent.Page 34
4. REFERENCES
1. Daugman J (1999) "Wavelet demodulation codes, statistical independence, and
pattern recognition." Institute of Mathematics and its Applications, Proc.2nd
IMA-IP. London: Albion, pp 244 - 260.
2. Daugman J (1999) "Biometric decision landscapes." Technical Report No TR482,
University of Cambridge Computer Laboratory.
3. Daugman J and Downing C J (1995) "Demodulation, predictive coding, and
spatial vision." Journal of the Optical Society of America A, vol. 12, no. 4, pp 641
- 660.
4. Daugman J (1993) "High confidence visual recognition of persons by a test of
statistical independence." IEEE Transactions on Pattern Analysis and Machine
Intelligence, vol. 15, no. 11, pp 1148 - 1160.
5. Daugman J (1985) "Uncertainty relation for resolution in space, spatial frequency,
and orientation optimized by two-dimensional visual cortical filters." Journal of
the Optical Society of America A, vol. 2, no. 7, pp 1160 - 1169.
WEB SITE
¢ http://cl.cam.ac.uk/jgd1000/
¢ http://en.wikipediawiki/Iris_recognition
