(Journal of Security Engineering),

8 5

2011 10

New Algorithm Biometric-Based Iris Pattern Recognition

System: Basis of Identity Authentication and Verification
Eun-suk Cho1), Ronnie D. Caytiles2), Seok-soo Kim3)

Abstract
This study presents a new algorithm for biometric-based iris recognition system. The proposed iris
identification algorithm consists of four major fundamental steps: image processing; image localization; IER
extraction; and image pattern recognition. An image of an individuals eye is processed into an 8-bit
grayscale BMP image until the boundaries of the iris are detected. The IER (Iris Effective Region) feature
is then extracted using the feature extraction algorithm. This IER feature is a 12x8 BMP image pattern
which contains the gray values of the iris and will be used for pattern matching. The IER feature extracted
is then used as input for the pattern recognition algorithm wherein the linear correlation coefficients are
analyzed and will be used as the basis for the identification and verification of an individual.
Keywords : Iris Recognition, Image Processing, Image Localization, IER Extraction, Pattern Recognition,
Linear Correlation Coefficient

1. Introduction
Biometric-based technologies have recently gained much consideration in the identification and verification of
an individuals identity. Individuals identity is verified based on physiological characteristics such as face,
fingerprints, hand geometry, iris, retina, and vein. Among these physiological characteristics, biometric
identification and verification based on iris patterns has attracted the attention of biometric-based identification
and verification research and development community. Iris patterns become interesting as an alternative approach
to reliable visual recognition of persons.
Iris is the annular ring between the pupil and the sclera of the eye. Individuals irides patterns are unique
Received(August 05, 2011), Review request(August 06, 2011), Review Result(1st: August 21, 2011, 2nd: August 30, 2011)
Accepted(October 31, 2011)
1

306-791 Department of Multimedia Engineering, Hannam University

email: eunsukk@empal.com

306-791 Department of Multimedia Engineering, Hannam University

email: gnsnfcknrs@yahoo.com

(Corresponding Author) 306-791 Department of Multimedia Engineering, Hannam University, Professor

email: sskim0123@naver.com

585

New Algorithm Biometric-Based Iris Pattern Recognition System: Basis of Identity Authentication and Verification

and structurally distinct although its coloration and structure is genetically linked and identical yet it has unique
characteristics that allows it to be used for accurate recognition of individuals identity. The variations in the
gray level intensity values of the irises distinguish two individuals. A difference exists between identical twins
and even between left and right eye of the same person.
The iris is highly protected, non-invasive and ideal for handling applications requiring management of large
user groups. Although small and sometimes problematic to image, the iris has the great mathematical advantage
that its pattern variability among different persons is enormous. In addition, as an internal yet externally visible
organ of the eye, the iris is well protected from the environment and stable over time. As a planar object, its
image is relatively insensitive to angle of illumination and changes in viewing angle but both fine
transformations and non-fine pattern distortion caused by pupillary dilation can be readily reversible. The ease
of localizing eyes in faces and the distinctive annular shape of the iris facilitate reliable and precise isolation of
this feature and the creation of a size-invariant representation.
In this study, new algorithm biometric-based iris recognition pattern system as a method of biometric
authentication, uses pattern-recognition techniques based on high-resolution images of the irides of an
individual's eye. It involves four main phases: image processing; image localization; IER extraction; and image
pattern recognition. Furthermore, this paper explains how the algorithms work and presents new data on the
statistical properties and singularity of iris patterns. Algorithms developed for recognizing persons by their iris
patterns have been tested with no false matches in several million comparison tests.
Iris recognition is achieved by comparison of results with stored patterns. Recognition will be based on the
results and analysis of the computation of the linear correlation coefficient. The biometrics information of an
individual who claims certain identity is compared with the biometrics on the record that represent the identity
that this individual claims. The comparison result determines whether the identity claims shall be accepted or
rejected.
Hence, this proposed new algorithm biometric-based iris recognition pattern system for iris pattern
identification for iris images would give a better performance for biometric authentication or identification. The
accuracy of iris recognition systems is proven to be much higher compared to other types of biometric systems
characteristics like fingerprint, handprint and voiceprint.

2. Related Works
For the past years, iris recognition is considered as one of the most accurate and reliable method of
biometric identification. The accuracy of iris recognition systems is proven to be much higher compared to
other biometric features since iris patterns have rich textures. In this section we briefly review a sample of the
iris recognition literatures:
586

(Journal of Security Engineering),

8 5

2011 10

John G. Daugman, Ph.D. pioneered the formulation of iris recognition algorithms required for image
acquisition and one-to-many matching and become the basis of almost all currently commercially deployed
iris-recognition systems. His algorithm locates the iris boundaries using an integrodifferential operator that finds
the circles in the image where the intensity is changing most rapidly with respect to changes in the radius.
Once located, the iris image is converted to a Cartesian form by projecting it to onto a dimensionless
pseudo-polar coordinate system. The iris features are encoded and a signature is created using a 2-D
complex-valued Gabor filter. Finally, two images are said to be independent if their fractional Hamming
distance (Hd) is above a certain threshold, otherwise they are a match. Hd is equal to the number of

mismatching bits divided by number of compared bits [1 3].

Christel-Loc Tisse et al examined a new iris recognition system that implements (i) gradient decomposed
Hough transform / integral-differential operators combination for iris localization and (ii) the analytic image
concept (2D Hilbert transform) to extract pertinent information from iris texture [4, 5]
Li Ma et al proposed a new approach for personal identification based on iris recognition. They have
detailed the steps of iris recognition, including image preprocessing, feature extraction and classifier design.
Their proposed algorithm uses a bank of Gabor filters to capture both local and global iris characteristics to
form a fixed length feature vector. Iris matching is based on the weighted Euclidean distance between the two
corresponding iris vectors [6].
Zhuoshi Wei et al presented a framework to synthesize large realistic iris databases, providing an alternative
to iris database collection. Firstly, iris patch is used as a basic element to characterize visual primitive of iris
texture, and patch-based sampling is applied to create an iris prototype [7].
Wildes locates the iris boundaries by using Hough transform method and a voting procedure. The Wildes
algorithm compares two images by geometrically warping one image, via shifting and rotation, until it best fit
with the other image, in the sense of minimizing mean square distance. An application of normalized
correlation and Fishers linear discriminant is its basis for matching [8].
Boles & Boashash locates the pupil center using an edge detection method, records grey level values on
virtual concentric circles, and then constructs the zero-crossing representation on these virtual circles based on a
one-dimensional dyadic wavelet transform. Two dissimilarity functions for the purposes of matching were
created, one using every point of the representation and the other using only the zero crossing points. The
algorithm has been tested successfully on a small database of iris images, with and without noise [9].
Lim, et al., in proposed alternative approaches to both feature extraction and matching. For feature extraction
the use of the Gabor Transform and the Haar Wavelet Transform is compared, and their results indicates that
the Haar Transform is better. The matching process uses an LVQ competitive learning neural network, which is
optimized by a careful selection of initial weight vectors [10].

587

New Algorithm Biometric-Based Iris Pattern Recognition System: Basis of Identity Authentication and Verification

3. Linear Correlation Coefficient

Linear correlation coefficient [23] measures the strength and the direction of a linear relationship between
two variables. It is denoted by r and defined as

r +1.

Where, r varies from -1 to +1. So, -1

If x and y have a strong positive linear correlation, r is close to +1 which is called Positive Correlation and
if x and y have a strong negative linear correlation, r is close to -1 which is called Negative Correlation. If
there is no linear correlation or a weak linear correlation, r is close to 0. A value near zero means that there
is a random, nonlinear relationship between the two variables. A correlation greater than 0.8 is generally
described as strong, whereas a correlation less than 0.5 is generally described as weak. These values can vary
based upon the "type" of data being examined. A study utilizing scientific data may require a stronger
correlation than a study using social science data.
Now, the second central moment is called the variance of x. It is denoted by is the standard
deviation of x which is the positive square root of the variance. Variance of x is defined by

Where,

is the mean of x,
N is the number of pairs of data.
Now, the variance of y is defined by-

588

(Journal of Security Engineering),

8 5

2011 10

is the mean of y,
N is the number of pairs of data,

is the standard deviation of y which is the positive square root of the variance.
Now, the second order mixed central moment is called the covariance of X and Y. It is denoted by
cov(X,Y) and is defined as-

Where,
is the mean of x ,
is the mean of y, N is the number of pairs of data.
Now, the correlation coefficient is-

Equation 1, 2, 3, 4 are used in this algorithm for calculation of correlation coefficient.

4. Proposed Work
The propose method for iris pattern recognition identification is divided into four main phases: image
589

New Algorithm Biometric-Based Iris Pattern Recognition System: Basis of Identity Authentication and Verification

processing; image localization; IER detection and extraction; and image recognition.

4.1 Image Processing

Image processing refers to any form of signal processing wherein the input is an image (a picture or video)
and the output could be either image or a set of characteristics or parameters related to the image. The
following algorithm signifies the steps of image processing in our approach:
Step 1: After an image of an individuals eye has been acquired with a Powerful Digital Camera, the image
must be resized to a 100*100 24-bit BMP image format.
Step 2: Open this 24-bit BMP file in binary mode.
Step 3: Create a new BMP file and copy the ImageInfo (first 54 byte) of the Header from the 24-bit BMP
file.
Step 4: Edit the Header by changing the file size, bit depth, and colors to 8-bit BMP format.
Step 5: Copy the values from the Color Table in a sample gray scale image to this newly created BMP file
at 54th byte.
Step 6: It is then transformed from RGB to an 8-bit grayscale image using the following formula:
blueValue = (0.299*redValue + 0.587*greenValue + 0.114*blueValue);
greenValue = (0.299*redValue + 0.587*greenValue + 0.114*blueValue);
redValue = (0.299*redValue + 0.587*greenValue + 0.114*blueValue);
grayValue = blueValue = greenValue = redValue;
Step 7: Write to new BMP file.
The output is a new 8-bit grayscale BMP image.

4.2 Image Localization

Iris localization aims to find the parameters, centers and radii, of the two iris boundaries, detecting the lower
and upper eyelid and isolating eyelashes. The boundaries of the image must be located before performing iris
pattern matching. In our approach, the iris localization algorithm includes both the inner boundary localization
and the outer boundary localization. The following algorithm signifies the steps of image localization on our
approach:
Step 1: Take the 8-Bit gray Scale Image produced from previous Algorithm as input and open this BMP
file in Binary Read mode.
Step 2: Detect the PUPIL Boundary and set the boundary pixels to 255 (white) using the following
pseudocode:
590

(Journal of Security Engineering),

8 5

2011 10

for(x=0;x<=originalImage.rows-1;x++){
for(y=0;y<=originalImage.cols-1;y++){
if (Grayvalue(x, y)==0) {
if((GrayValue(x,y-1)!=0)&& GrayValue(x,y+1)==0))){
GrayValue (x, y-1) = 255
}else if((GrayValue(x,y-1)==0)&&GrayValue(x,y+1)!=0))){
GrayValue (x, y+1) = 255
}else
No change in the GrayValue.
}
}
}
Step 3: Detect the IRIS Boundary and set the boundary pixels to 255 (white) using following pseudocode:
for(x=0;x<=originalImage.rows-1;x++){
for(y=0;y<=originalImage.cols-1;y++){
if(x>12 && x<70 && y>12 && y<50){
if(Grayvalue(x,y)<100 && Grayvalue(x,y)!=0){
if(((Grayvalue(x,y-1))-(GrayValue(x,y+1))>24 &&
(Grayvalue(x,y-1))>=100&&(Grayvalue(x,y-1))!=255 &&
(Grayvalue(x,y+1))<100 &&(Grayvalue(x,y+1))!=0 ){
GrayValue(x,y-1) = 255
}else
No change in the GrayValue
}else
No change in the GrayValue.
}else
Write to new BMP file
}
}
By considering the 8-bit BMP image, two-pass masking is used, namely, Horizontal and Vertical kernels.
The PlanarImage now passed through these masks or kernels. Resultant transformed image generates the distinct
marks for IRIS area; the process is called edge detection.

4.3 IER Extraction

591

New Algorithm Biometric-Based Iris Pattern Recognition System: Basis of Identity Authentication and Verification

After the detection of pupil boundary and the iris left boundary, we have to extract the effective region of
the iris. The effective region of the iris means a specific size (say 55) of matrix which is in the iris portion,
neither in sclera nor in pupil portion be included in this effective region. The iris effective region is extracted
using the following algorithm:
Step 1: Take the 8-Bit BMP Image produced from previous algorithm as input and open this BMP file in
binary Read Mode.
Step 2: Read the raster Data and store the raster data into a matrix of vectorSize. Where vectorSize =
filesize - (54+(4*256)).
Step 3: Then an 8*12 Iris Pattern is extracted from Edge Detected BMP using following pseudocode:
for(x=0;x<=originalImage.rows-1;x++) {
for(y=0;y<=originalImage.cols-1;y++) {
if( y<30&&x=((originalImage.rows/2)+4)&&GrayValue==255) {
for(i=0;i<8;i++) {
for(j=0;j<12;j++) {
*(edgeImage.data + (i * edgeImage.cols) + j) = *(originalImage.data + (x *
originalImage.cols) - (i * originalImage.cols) + (y + j));
Write to new BMP Image file
}
}
}
}
}
The 12x8 8-bit BMP image is the output of the algorithm. In this algorithm, we locate the middle row and
first column of the input image, and then we move 4 pixels upward and check the gray value of each pixel
until the 255 (white) values is reached. Then the pixels are read and stored the corresponding gray values into
an 8x12 matrix.
The output is a new BMP image of size 12x8 and a matrix of the same size which contains the gray
values.

4.4 Pattern Recognition

This test enables the comparison of two iris patterns. This algorithm is based in the linear correlation
coefficient. From this n number of IRIS patterns only single IRIS Pattern of an individual can be generated
by Statistical Analysis, which can be strongly used for pattern recognition of human identification.
592

(Journal of Security Engineering),

8 5

2011 10

Step 1: Take 1 training (input) iris pattern image of any individual and given identification number.
Step 2: Read all 10 iris pattern images from database against his/her identification number.
Step 3: Taking input pattern image with each stored iris pattern images in database, calculate correlation
coefficient(r) using the following psudocode:
for(i=0;i<row;i++){
for(j=0;j<col;j++){
x = matrix_database[i][j];
y = matrix_Input[i][j];
sum_x2 += x*x;
sum_y2 += y*y;
sum_xy += x*y;
sum_x += x;
sum_y += y;
}
}Sxx = sum_x2 - ((sum_x*sum_x)/n);
Sxy = sum_xy - ((sum_x*sum_y)/n);
Syy = sum_y2 - ((sum_y*sum_y)/n);
r = Sxy/(sqrt(Sxx*Syy));
Step 4: Ten correlation coefficient(r) values are generated and then calculate the average of these 10
r-values. If the average of the 10 r-values >= 0.8 then the individual with the given input iris
pattern is identified as a member of the organization. Otherwise the individual is unauthorized.

5. Experimental Results
In this thesis, iris recognition performance is evaluated using the False Acceptance Rate (FAR) and False
Rejection Rate (FRR).
The false acceptance rate, or FAR, is the measure of the likelihood that the biometric security system will
incorrectly accept an access attempt by an unauthorized user. A systems FAR typically is stated as the ratio of
the number of false acceptances divided by the number of identification attempts. FAR is defined as

The false rejection rate, or FRR, is the measure of the likelihood that the biometric security system will
incorrectly reject an access attempt by an authorized user. A systems FRR typically is stated as the ratio of
593

New Algorithm Biometric-Based Iris Pattern Recognition System: Basis of Identity Authentication and Verification

the number of false rejections divided by the number of identification attempts. FRR is defined as :

Output of this Algorithm is as follows:

[Table 1] For Authorized Persons.

Qr1
Qr2
Qr3
Qr4
Qr5
Qr6
Qr7
Qr8
Qr9
Qr10

Person1
0.92527
0.920889
0.901001
0.866415
0.813186
0.841426
0.944121
0.931877
0.925426
0.946985

Person2
0.887013
0.894657
0.597282
0.903323
0.892065
0.924572
0.967363
0.964236
0.920788
0.910433

Person3
0.90191
0.94282
0.950476
0.896841
0.923513
0.773495
0.900203
0.932124
0.932675
0.930373

Person4
0.817018
0.837767
0.842439
0.861058
0.916403
0.906519
0.908223
0.903906
0.781454
0.861127

Person5
0.771615
0.9443
0.698773
0.811424
0.884811
0.84561
0.731279
0.85349
0.792124
0.83753

Person2
-0.04326
-0.0285
0.006919
0.064761
0.009741
-0.04319
-0.06479
-0.07535
-0.09937
-0.09589

Person3
-0.43106
-0.37071
-0.35172
-0.39909
-0.34608
0.027949
-0.44331
-0.37269
-0.43547
-0.41842

Person4
0.019021
-0.03883
0.067881
-0.08786
-0.07962
-0.02023
-0.01245
0.005396
-0.04114
-0.00903

Person5
-0.04106
0.06212
-0.06451
-0.0627
-0.03233
0.095871
0.090123
0.095601
0.106594
0.073078

[Table 2] For Unauthorized Persons.

Qr1
Qr2
Qr3
Qr4
Qr5
Qr6
Qr7
Qr8
Qr9
Qr10

594

Person1
-0.00599
-0.01092
0.023669
-0.04637
-0.02285
-0.07197
-0.01049
0.004875
0.053236
-0.01722

(Journal of Security Engineering),

8 5

2011 10

From Table 1, we can see there are 0 falsely rejected incidents in 50 samples and from Table 2, there are
2 falsely accepted incidents in 50 samples in our experimental database results. i.e.
FAR (%)
FRR (%)

MMU1 Iris Database

2.0
0

6. Conclusion and Future Work

In this study, a new algorithm of iris identification was proposed for iris recognition. IER feature extraction
used for pattern matching was compared with the stored patterns within the database to ensure high quality
result in authentication. The biological characteristics of iris pattern use in statistical correlation coefficient for
iris pattern recognition were considered. It was found out that Statistical Estimation Theory can play an
immense role.
The reliability of identification, verification and authentication of an individuals identity is highly correlated
with the results of pattern matching based on the linear correlation coefficient of pattern recognition algorithm.
High correlation in pattern matching results to better verification and authentication of the individuals true
identity.
The new algorithm biometric-based iris recognition pattern system will result to a clear distinction between
an individuals true identity and a fraud one. It is deemed that iris pattern recognition system will become the
leading technology in identity verification and authentication. Significant progress in the iris recognition model
and industry has been observed, hence, further and continued research and testing is highly recommended in
diverse environments and configurations. Research issues based on iris localization, nonlinear normalization,
segmentation, handling noisy, degraded images, and large scale identification are highly recommended.
595

New Algorithm Biometric-Based Iris Pattern Recognition System: Basis of Identity Authentication and Verification

References

[1] J. G. Daugman, High confidence visual recognition of person by a test of statistical independence, IEEE
Trans. PAMI 15, 1148-1161 (1993).
[2] J. G. Daugman, "The importance of being random: statistical principles of iris recognition," Pattern.
Recognition 36, 279-291 (2003).
[3] J. G. Daugman, How iris recognition works, IEEE Trans. Circuits and Syst. for Video Tech. 14(1), 21-30
(2004).
[4] C. L. Tisse, L. Martin, L. Torres, M. Robert, "Person identification technique using human iris recognition"
ST Journal of System Research Current Issue 2003.
[5] C. Tisse, L. Martin , L. Torres, and M. Robert, Iris recognition system for person identification, PRIS
2002 : 186-199 (2002).
[6] L. Ma, T. Tan, Y. Wang, and D. Zhang, Personel identification based on iris texture analysis, IEEE
Trans. On Pattern Analysis and Machine Intelligence 25(12):1519-1533, (2003).
[7] Z. Wei, T. Tan and Z. Sun, "Synthesis of Large Realistic Iris Databases Using Patch-based Sampling" IEEE
2008.
[8] R. P. Wildes, Automated iris recognition: An emerging biometric technology, Proceedings of the IEEE 85
1348-1363 (1997).
[9] W. Boles, and B. Boashash, A human identification technique using images of the iris and wavelet
transform, IEEE Trans. Signal Proc. 4:1185-1188, (1998).
[10] S. Lim, K. Lee, O. Byeon, and T. Kim, Efficient iris recognition through improvement of feature vector
and classifier, ETRI J. 23(2):61-70, (2001).

596

(Journal of Security Engineering),

8 5

2011 10

Author
Eun-suk Cho
1994 Associate in Dept. of Nursing Science, Suwon Women Colleage.
2000 B.L. in Dept. of Law, Korea National Open University
2003 Ed. M. in Dept. of Education, Ajou University
Currently, Integrated Course for M.S. and Ph.D. in Dept. of Multimedia Engineering,
Hannam University.
Research interests : Hospital Security, Patients' privacy, U-healthcare system.

Ronnie D. Caytiles

1995
2000: Bachelor of Science in Computer Engineering, Western Institute of
Technology, Iloilo City, Philippines
Currently: Integrated Course for M.S. and Ph.D. in Multimedia Engineering, Hannam
University, Daejon, Korea.
Research Interests: Information Technology Security, U-Learning, Control and
Automation

Seoksoo Kim
Received a B.S. degree in computer engineering from Kyungnam University , Korea,
1989, and M.S. degree in Information engineering from Sungkyun-kwan
University, Korea, 1991 and Ph D. degree in Information engineering from
Sungkyun-kwan University, Korea, 2002. In 2003 he joined the faculty of
Hannam University, Korea where he is currently a professor in Department of
Multimedia. His research interests include Multimedia Communication systems,
Distance learning, Multimedia Authoring, Telemedicine, Multimedia Programming,
Computer Networking. Information Security. He is a Member of KCA, KICS,
KIMICS, KIPS, KMS, and DCS.

597

New Algorithm Biometric-Based Iris Pattern Recognition System: Basis of Identity Authentication and Verification

598