Vous êtes sur la page 1sur 13

image processing :

An efficient algorithm for iris pattern


recognition using 2D Gabor wavelet
transformation in Matlab

ABSTRACT:

Wavelet analysis have received significant attention because their multi-resolution


decomposition allows efficient image analysis. It is widely used for varied applications such as
noise reduction, and data compression, etc. In this paper we have introduced and applied the
concept of 2 dimensional Gabor wavelet transform to Biometric Iris recognition system. The
application of this transform in encoding the iris image for pattern recognition proves to
achieve increased accuracy and processing speed compared to other methods. With a strong
scientific approach and mathematical background we have developed an algorithm to facilitate
the implementation of this method under the platforms of MATLAB.
IMAGES – An introduction:
A dictionary defines image as a “reproduction or representation of the form of a person or
thing”. The inherent association of a human with the visual senses, predisposes one to
conceive an image as a stimulus on the retina of the eye, in which case the mechanism of
optics govern the image formation resulting in continuos range, multi-tone images.

A digital image can be defined to be a numerical representation of an object or more


strictly to be sampled, quantized function of two dimensions which has been generated by
optical means, sampled in an equally spaced rectangular grid pattern, and quantized in equal
intervals of graylevel.

WHY BIOMETRICS:
The word is crying out for the simpler access controls to personal authentication systems
and it looks like biometrics may be the answer. Instead of carrying bunch of keys, all those
access cards or passwords you carry around with you, your body can be used to uniquely
identify you. Furthermore, when biometrics measures are applied in combination with other
controls, such as access cards or passwords, the reliability of authentication controls takes a
giant step forward.

BIOMETRICS-AN OVERVIEW:
Biometrics is best defined as measurable physiological and/or behavioral
characteristics that can be utilized to verify the identity of an indivisual. They include the
following:
 Iris scanning
 Facial recognition
 Fingerprint verification
 Hand geometry
 Retinal scanning
 Signature verification
 Voice verification

ADVANTAGES OF THE IRIS IDENTIFICATION:


 Highly protected internal organ of the eye.
 Iris patterns possess a high degree of randomness.
 Variability: 244 degrees of freedom.
 Entropy: 3.2 bits per square millimetre.
 Uniqueness: set by combinatorial complexity.
 Patterns apparently stable throughout life.

IRIS – An introduction:
The iris is a colored ring that surrounds the pupil and contains easily visible yet
complex and distinct combinations of corona, pits, filaments, crypts, striations, radial furrows
and more.
The iris is called the “Living password” because of its unique, random features. It’s
always with you and can’t be stolen or faked. As such it makes an excellent biometrics
identifier.

SCANNING MECHANISM:
Iris scanning is undoubtedly the less intrusive of the eye related biometrics. It utilizes a
fairly conventional CCD camera element and requires no intimate contact between user and
reader.
The iris scanning procedures as used on humans is simple and painless. The person stands
a foot or so away from the camera and looks into the scanning device. In iris scanning, the eye
is illuminated by the light emitting diodes that surround the camera. The diodes emit in the
visible light spectrum. The scanner is not a laser-rectinal scanner so there are no laser hazards.
It scans a a high-definition photograph of the person’s eyes (i.e., some of the characteristics of
the iris tissue, such as rings, furrows and freckles are scanned). It then analyzes 266 different
points of data (512-byte biometrics template) from the trabecular meshwork of the iris.
The scanned pattern is then digitized and compared to previously recorded patterns.
Identification is achieved in two seconds and verification in three.

ALGORITHM
Wavelet transforms have received significant attention because their multiresolution
decomposition allows efficient image analysis, noise reduction and data compression. Using
Discrete wavelet transform (DWT), the procedures of terrain segmentation and speckle noise
reduction can be effectively combined into a single process. We have applied Gabor Wavelet
transform in our algorithm to achieve iris image pattern recognition.
We have divided our entire Algorithm into three parts

1) IMAGE TO CODE CONVERSION ALGORITHM


2) CODE COMPARISON ALGORITHM
3) AUTHENTICATION ALGORITHM
1) IMAGE TO CODE CONVERSION ALGORITHM
Input:
Iris Image from a CCD camera and SSN of the user.
Output:
Iris code for Iris Image.
ALGORITHM:
1) Test the resolution of the captured image.
2) Convert the RGB image to the monochrome format.
3) Locate the Iris in the image.
4) Encode the iris pattern to the corresponding grayscale by using circular grid.
5) Each isolated iris pattern is then demodulated to extract its phase information using
quadrature 2D Gabor wavelets.
6) A 512-byte bar code is generated based on the distribution of black and white areas of the
iris inside the grid.
7) Store the iris code and the SSN in the source folder.

Algorithm in Detail:
Testing the resolution:
The image captured by a CCD camera (480 x 640) should be converted to RGB to
monochrome format. It is then resolved to a minimum of 70 pixels in iris radius for the rich
details of iris patterns. Now a days 100 to 140 pixels can be resolved easily.

Locating Iris in the Image:


To locate the iris in the image the center coordinates and radius of both the iris
and the pupil are determined using the operator explained below. A CCD camera captures the
image of the eye. Some imaging platforms deployed a wide-angle camera for coarse
localization of eyes in faces, to steer the optics of a narrow-angle pan/tilt camera that acquires
higher resolution images of eyes. In these trails, most imaging was done without active pan/tilt
camera optics, but instead exploited visual feedback via a mirror or video image to enable
cooperating subjects to position their own eyes within the fields of view of a single narrow-
angle camera.

Focus assessment is performed in real-time (faster than video frame rate) by measuring
the total high-frequency power in the 2D Fourier spectrum of each frame, and seeking to
maximize this quantity either by moving an active lens or by providing audio feedback to
subjects to adjust their range appropriately. Images passing a minimum focus criterion were
then analyzed to find the iris, with precise localization of its boundaries using a coarse-to-fine
strategy terminating in single-pixel precision estimates of the center coordinates and radius of
both the iris and the pupil. Although the results of the iris search greatly constrain the pupil
search, concentricity of these boundaries cannot be assumed. Very often the pupil center is
nasal, and inferior, to the iris center. Its radius can range from 0.1 to 0.8 of the iris radius.
Thus, all three parameters defining the pupillary circle must be estimated separately from those
of the iris. A differential operator for determining these parameters is:

Where I (x,y) is an image such as fig 1 containing an eye. The operator searches over
the image domain (x,y) for the maximum in the blurred partial derivative with respect to
increasing radius r, of the normalized contour integral of I (x, y) along a circular arc ds of

radius ‘ r’ and center co-ordinates ( Xo, Yo ) The symbol * denotes convolution and is

a smoothing function such as Gaussian of scale σ. The complete operator behaves in effect as

a circular edge detector, blurred at a scale set by σ 8, which searches iteratively for a
maximum contour integral derivative with increasing radius at successively finer scales of

analysis through the three parameter space of center coordinates and radius ( Xo, Yo, r)
defining a path of contour integration.

The operator in (1) serves to find both the pupillary boundary and the outer (limbus)
boundary of the iris, although the initial search for the limbus also incorporates evidence of an
interior pupil to improve its robustness since the limbic boundary itself usually has extremely
soft contrast when long wavelength NIR illumination is used. Once the coarse-to-fine iterative
searches for both these boundaries have reached single pixel precision, then a similar approach
have to detecting curvilinear edges is used to localize both upper and lower eyelid boundaries.
The path of contour integration in (1) is changed from circular to arcuate, with spline
parameters fitted by standard statistical estimation methods to describe optimally the available
evidence for each eyelid boundary. The result of all these localization operations is the
isolation of iris tissue from other image regions, as illustrated in fig 1 by graphical overlay on
the eye.

Iris Feature Encoding by 2D Gabor Wavelet Demodulation:

Each isolated iris pattern is then demodulated to extract its phase information using
quadrature 2D Gabor wavelets (Daugman 1985, 1988, 1994).This encoding process is
illustrated in Fig 2. It amounts to a patch-wise phase quantization of the iris pattern, by
identifying in which quadrant of the complex plane each resultant phasor lies when a given
area of the iris is projected onto complex-valued 2D Gabor wavelets:

Where h{Re, Im} can be regarded as a complex-valued bit whose, real and imaginary
parts are either 1 or 0 depending on the sign of the 2D integral; is the raw iris image in
a dimensionless polar coordinate system that is size- and translation-invariant, and which also

corrects for pupil dilation as explained in a later section; d and e are the multi-scale
2D wavelet size parameters, spanning an 8-fold range from 0.15mm to 1.2mm on the iris;
f is wavelet frequency, spanning 3 octaves in inverse proportion to ; and
represent the polar coordinates of each region of iris for which the phasor coordinates
are computed. Such a phase quadrant coding sequence is illustrated for one iris by
the bit stream shown graphically in Fig1.

A desirable feature of the phase code portrayed in Fig 2 is that it is a cyclic, or gray code:
in rotating between any adjacent quadrants, only a single bit changes, unlike a binary code in
which two bits may change, making some errors arbitrarily more costly than others. Altogether
2,048 such phase bits (256 bytes) are computed for each iris, but in a major improvement over
the earlier (Daugman 1993) algorithm, now an equal number of masking bits are also
computed to signify whether any iris region is 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.
Only phase information is used for recognizing irises because amplitude information is not
very discriminating, and it depends upon extraneous factors such as imaging contrast,
illumination, and camera gain. The phase bit settings which code the sequence of projection
quadrants as shown in fig 2 capture the information of wavelet zero-crossing, as is clear from
the sign operator in (2). The extraction of phase has the further advantage that phase angle are
assigned regardless of how low the image contrast may be. Its phase bit stream has statistical
properties such as run lengths similar to those of the code for the properly focused eye image
in Fig 1. The benefit which arises from the fact that phase bits are set also for a poorly
focussed images as shown here, even if based only on random CCD noise, is that different
poorly focused irises never become confused with each other when their phase codes are
compared. By contrast, images of different faces look increasingly alike when poorly resolved,
and may be confused with each other by appearance-based recognition algorithms.

2) CODE COMPARISON ALGORITHM


Input:
Current Iris code & SSN

Output:
Match in terms of hamming distance.

Algorithm:
1) Compare the Iris code (gray code) of the test image with that of the considered image in
the source folder.
2) The difference between the codes is expressed as Hamming distance.
3) Compute the Discrete Cosine Transform (DCT) of the Iris code
4) Find the Probability density function (PDF).
5) Compare the PDF of the test image with that of the considered image in the source
folder.

Algorithm in Detail:
The Current Iris code is compared with the iris code stored in the source folder
corresponding to the given SSN. The difference between the codes is expressed as a fraction of
mismatched bits termed as Hamming distance. The key to iris recognition is the failure of a
test of statistical independence, which involves so many degrees-of freedom that this test is
virtually guaranteed to be passed whenever the phase codes for two different eyes are
compared, but to be uniquely failed when any eye's phase code is compared with another
version of itself.

The test of statistical independence is implemented by the simple Boolean Exclusive-


OR operator (XOR) applied to the 2,048 bit phase vectors that encode any two iris patterns,
masked (AND' ed) by both of their corresponding mask bit vectors to prevent non-iris artifacts
from influencing iris comparisons. The XOR operator ‘ j ’ detects disagreement between any
corresponding pair of bits, while the AND operator ‘ k ’ ensures that the compared bits are
both deemed to have been uncorrupted by eyelashes, eyelids, specular reflections, or other
noise. The norms of the resultant bit vector and of the AND'ed mask vectors are then measured
in order to compute a fractional Hamming Distance (HD) as the measure of the dissimilarity
between any two irises, whose two phase code bit vectors are denoted {codeA, codeB} and
whose mask bit vectors are denoted {maskA, maskB}:

Hamming Distance:

The denominator tallies the total number of phase bits that mattered in iris comparisons
after artifacts such as eyelashes and specular reflections were discounted, so the resulting HD
is a fractional measure of dissimilarity; 0 would represent a perfect match. The Boolean
operators ‘ j ’ and ‘ k ’ are applied in vector form to binary strings of length up to the word
length of the CPU, as a single machine instruction. Thus for example on an ordinary 32-bit
machine, any two integers between 0 and 4 billion can be XOR'ed in a single machine
instruction to generate a third such integer, each of whose bits in a binary expansion is the
XOR of the corresponding pair of bits of the original two integers. This implementation of (3)
in parallel 32-bit chunks enables extremely rapid comparisons of iris codes when searching
through a large database to find a match. On a 300 MHZ CPU, such exhaustive searches are
performed at a rate of about 100,000 irises per second. Because any given bit in the phase code
for an iris is equally likely to be 1 or 0, and because different irises are uncorrelated, the
expected proportion of agreeing bits between the codes for two different irises is HD=0.500.
The histogram in Fig 4 shows the distribution of HDs obtained from various samples of
images. The observed mean of HD was p=0.499 with standard deviation their

full distribution in Fig 4 corresponds to a binomial having degree-of-


freedom, as shown by the solid curve.

The extremely close fit of the theoretical binomial to the observed distribution is a
consequence of the fact that each comparison between two phase code bits from two different
irises is essentially a Bernoulli trial, albeit with correlations between successive "coin tosses".
In the phase code for any given iris , only small subsets of bits are mutually independent due to
the internal correlations, especially radial, within an iris.

Fig- 4 . Binomial distribution of Iris Hamming distances

3) AUTHENTICATION ALGORITHM
Input:
Hamming Distance (HD) from the comparison Algorithm.
Output:
Whether the person is authorized or not.
Algorithm:
1) If HD is zero or below the threshold value of 0.32 then the PDF of the test image and the
PDF of the corresponding image in the source folder are compared.
2) If HD is above 0.32 then the person is rejected and is not authorized.
3) If the PDFs are equal then the person is accepted to login.

Advantages of Gabor wavelet transform:


1) The iris recognition system requires an accuracy rate of almost 100% for pattern
recognition which can be achieved using Gabor wavelet transform.
2) Multi spectral analysis can be easily achieved.
3) This transform performs well even for the poorly focused eye images .
4) It can able to reconstruct bits of the demodulated phase sequence, set primarily by the
random CCD noise
5) The Harr Wavelet transform can be used for achieving still quicker processing of image
but the accuracy of the result is based on the clearness of the picture information we
encode.

VERIFICATION USING MATLAB:


MATLAB refers to ‘MATrix LABoratory’. It is used especially for mathematical
computation purposes. If we are considering a picture of a JPEG format, the picture as said
above will be represented as a matrix in MATLAB. When you are manipulating the same in
MATLAB, you are just going to manipulate the elements of the matrix, which wholly
represent that picture.
The group of images of the iris of the accessible person is first stored in a separate
folder. These images will be of RGB format. Using MATLAB, these RGB images are
converted to their respective monochrome formats, edge detected and their individual DCTs
and PDFs are taken. The RGB to monochrome conversion is handled as it will facilitate the
comparison and DCT helps for the image compression. Unlike textual data, we will most
likely not have a finite number of unique feature vectors that can occur in our data and
therefore a continuos PDF is calculated for the images. The test image is also manipulated as
said above. Now the manipulated test image is compared with the manipulated group. The
comparison can be made with least number of allowances. If the error is within the allowed
permit, the error is allowed and the person is permitted to access the system. If not, he will be
rejected.

ADVANTAGES OF MATLAB COMPARISON:


• Accuracy rate is much higher than conventional methods.
• False rejection rate is less.
• Conversion from RGB to grayscale, edge detection, DCT transformations, HD & PDF are
easily accomplished with MATLAB.
• The comparison will be very quick and ease.
• External interfacing will be easier.

Conclusion:
Thus an Iris Biometric system can establish an higher level of authenticity for security
systems with high degree of accuracy and reliability. Hence such a system can be designed to
protect the high security systems against hacking and cracking.

REFERENCES :

 An Iris Biometric for System public and personal use”


By Michael Negin et al. IEEE Computer Feb 2000 vol:33 n0:2.

 J.D.Daugman , “High-Confidence Visual Recognition Of Persons by a test of Statistical


Independence” ,
IEEE Transaction pattern matching and machine Intelligence, Nov 1993.

 “Understanding Iridology” by Dr.Bernard Jensen & Dr.Donald V. Bodeen.

J.Shapiro“Embedded image coding using Gabor wavelet coefficients”


IEEE Transaction on Signal Processing, Vol 41, Dec-1995.

============================================================
=======