International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 5, May 2014 ISSN 2319 - 4847
1
2
3
4
1,2
Assist. Prof. P.R.E.C., Loni
3,4
Lecturer, P.R.E.C., Loni
Abstract
This paper is with authentication of people based on iris recognition. This is widely developing technology. The presented biometric method is compared with a database for identification, or verification. The process is based on the unique nature and extreme richness of the human iris. Iris recognition is feasible for use in differentiating between identical twins. Though the iris color and the overall statistical quality of the iris texture may be dependent on genetic factors, the textural details are independent and uncorrelated for genetically identical iris pairs. For determining the recognition performance of the system one databases of digitized grayscale eye images were used. The iris recognition system consists of an automatic segmentation system that is based on the Hough transform, and is able to localize the circular iris and pupil region, occluding eyelids and eyelashes, and reflections. The extracted iris region was then normalized into a rectangular block with constant dimensions to account for imaging inconsistencies. The Classification is done using 2D Log-Gabor Filter. The accuracy is obtained 99.96% with FAR and
FRR is 0.00% and 0.0750% respectively.
Keywords: Iris images, 2D Log-Gabor Filter, Circular Hough transform (CHT), Random Hough transform (RHT).
1.
I NTRODUCTION
Biometric methods of personal identification are based on physical or behavioral characteristics, such as fingerprints, face, palm prints, hand geometry, retina and iris etc. The iris is the colored portion of the eye that surrounds the pupil. Its combination of pits, striations, filaments, rings, dark spots and freckles makes it a very accurate means of biometric identification [1], [7]. A typical iris recognition system often includes iris capture, preprocessing, feature extraction and feature matching. In iris recognition algorithm, preprocessing and feature extraction are two key processes. Iris preprocessing, including localization, segmentation, normalization and enhancement, is a basic step in iris identification algorithm. Iris feature extraction is the most important step in iris recognition, which determines directly the value of iris characteristics in actual application. These characteristics make it very attractive for use as a biometric for identifying individuals. Image processing techniques can be employed to extract the unique iris pattern from a digitised image of the eye, and encode it into a biometric template, which can be stored in a database.
Figure 1 Human Eye
This biometric template contains an objective mathematical representation of the unique information stored in the iris, and allows comparisons to be made between templates. When a subject wishes to be identified by iris recognition system, their eye is first photographed, and then a template created for their iris region. This template is then compared with the other templates stored in a database until either a matching template is found and the subject is identified, or no match is found and the subject remains unidentified.
2.
M ETHODOLOGY
2.1
Process Flow Diagram
The system is to be composed of a number of sub-systems, which correspond to each stage of iris recognition. These stages are segmentation – locating the iris region in an eye image, normalization – creating a dimensionally consistent representation of the iris region, and feature encoding – creating a template containing only the most discriminating features of the iris. The input to the system will be an eye image, and the output will be an iris template, which will provide a mathematical representation of the iris region.
Volume 3, Issue 5, May 2014 Page 278

International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 5, May 2014 ISSN 2319 - 4847
Figure2. Iris Image processing steps
The first stage will be to develop an algorithm to automatically segment the iris region from an eye image. This will require research into many different techniques such as circular Hough transform, and active contour models .Following this, the next stage will be to normalize the iris region in order to counteract imaging inconsistencies such as pupil dilation. An implementation of Daugman’s polar representation [2] will be used for this purpose, as this is the most documented method for iris normalization. Once a normalized iris pattern has been obtained, it will be convolved with
2D Log-Gabor Filter in order to extract features.
2.2 Hough Transform
To find circle basic two steps are used.
1) Edge Detection
2) Circle Detection.
The Hough transform is a standard computer vision algorithm that can be used to determine the parameters of simple geometric objects, such as lines and circles, present in an image. The circular Hough transform can be employed to deduce the radius and centre coordinates of the pupil and iris regions. An automatic segmentation algorithm based on the circular Hough transform is employed by Wildes P. [3]. Firstly, an edge map is generated by calculating the first derivatives of intensity values in an eye image and then thresholding the result. From the edge map, votes are cast in
Hough space for the parameters of circles passing through each edge point. These parameters are the centre coordinates x c and y c and radius r c which are able to define any circle according to the equation. x  x
0 , 1
 r cos  y  y
0 , 1
 r sin 
(1)
A pair of edge pixels can be mapped into parameter space pixels set ( p ,  ) . We can define the pixels set in parameter space as: x
2 c
 y
2 c
 r
2 c
 o
(2)
Figure3. a) Inner & Outer Circle b) Pupil Detection c) Segmented Iris Image
A maximum point in the Hough space will correspond to the radius and centre coordinates of the circle best defined by the edge points. Wildes [3] also make use of the parabolic Hough transform to detect the eyelids approximating the upper and lower eyelids.
2.3 Transformation from Cartesian to Polar Coordinates
A normalization process is implemented to compensate for size variations due to the possible changes in the camerato-face distance and to facilitate the feature extraction process by transforming the iris area represented by polar coordinate system into Cartesian coordinate system. The formula of the iris region from Cartesian coordinates to the normalized non-concentric polar representation is modeled as:
Volume 3, Issue 5, May 2014 Page 279

International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 5, May 2014 ISSN 2319 - 4847
I ( x ( r  ), x ( r ,  ) y ( r ,  )

 y ( r ,  ))
( 1 
( 1 
 I ( r ,  ) r ) x p
(  ) r ) y p
(  )

 rx i
(  ) ry i
(  )
(3)
(4)
(5)
Where I(x, y) is the iris region image, (x, y) are the original Cartesian coordinates, (r,

) are the corresponding normalized polar coordinates, and x p
, y p
and x i
, y i
are the coordinates of pupil and iris boundary along direction  .
Figure4.
a) Original mage b) Corresponding polar coordinates
3.
F EATURE E XTRACTION U SING 2D L OG -G ABOR F ILTER
In recent years, Gabor filter based methods have been widely used in computer vision, especially for texture analysis.
Gabor elementary functions are Gaussians modulated by sinusoidal functions. It is shown that the Functional form of
Gabor filters conforms closely to the receptive profiles of simple cortical cells, and Gabor filtering is an effective scheme for image representation.
A two-dimensional (2D) Log- Gabor filter can be represented by the following equation in the spatial domain
G ( u , v )  exp{

2
[log(
[log( u k
1
/
/ u u
0
0
)]
)]
2
2
} exp{

2  v
1
2
2 v
}
(6)
Where u
1
 u cos(  )  v sin(  ) , v
1
  sin(  )  v cos(  ),  is the orientation of 2D Log-Gabor filter, u
0
is the center frequency, k determines the bandwidth of the filter in the u
1
direction, &  v determines the bandwidth of the filter in the v
1
direction. With the help of Hamming distance we measure the how many bits are the same between two bit patterns.
3.1
Matching Algorithm
3.1.1
Hamming Distance Measurement
Hamming distance of two bit patterns, a decision can be made as to whether the two patterns were generated from different irises or from the same one. For the two masked binary templates A & B, HD can be calculated by
HD 
Where
1 i
N

 1
A i
 B i
& MS

 ( 1  HD ) (7)
N
N -the number of bits is represented by each template &  is XOR operation, HD gives the matching score
( MS

) for the iris template as given in Equation 5. For handling the rotation, templates are shifted left and right bitwise and a number of HD values are calculated from successive shifts [1] and the smallest value is used as the final HD to calculate the matching score. This bit-wise shifting in the horizontal direction corresponds to rotation of the original iris region at an angle given by the angular resolution used. It is capable of taking into account the misalignments in the normalized iris pattern caused by rotational differences during imaging. When the Hamming distance of two templates is calculated, rotational inconsistencies take in account. One template is shifted left and right bit-wise and a number of Hamming distance values are calculated from successive shifts. This bit-wise shifting in the horizontal direction corresponds to rotation of the original iris region by an angle given by the angular resolution used. If an angular resolution of 180 is used, each shift will correspond to a rotation of 2 degrees in the iris region. This method is suggested by Daugman [1], and corrects for misalignments in the normalized iris pattern caused by rotational differences during imaging. From the calculated Hamming distance values, only the lowest is taken, since this corresponds to the best match between two templates. The number of bits moved during each shift is given by two times the number of filters used, since each filter will generate two bits of information from one pixel of the normalized region. The actual number of shifts required to normalize rotational inconsistencies will be determined by the maximum angle difference between two images of the same eye, and one shift is defined as one shift to the left, followed by one shift to the right.
3.2 Degree of Freedom Calculation
Uniqueness was also be determined by measuring the number of degrees of freedom represented by the templates. This gives a measure of the complexity of iris patterns, and can be calculated by approximating the collection of inter-class
Hamming distance values as a binomial distribution. The number of degrees of freedom, DOF , can be calculated by:
Volume 3, Issue 5, May 2014 Page 280

International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 5, May 2014
DOF
Where, p ( 1  p )


2 p - Mean,  - Standard deviation
(8)
ISSN 2319 - 4847
The shifting also caused a reduction in the number of degrees of freedom, DOF . This reduction in DOF is an anomaly caused by a smaller standard deviation, which itself is caused by taking the lowest Hamming distance from 10 calculated values. This shows that, due to shifting, the distribution is not merely shifted towards the left, but the characteristics of the distribution are changed. Therefore the degrees of freedom formula is not very useful with shifting introduced, since it relies on the distribution approximating a binomial.
Table 1 : DOF Values calculated as follows
Sr. No.
1
Iris
Images
IMG 01 p
DOF Values
0.4490

0.0081
DOF (
3.5
2
IMG 02 0.4530 0.0062 6.43
3
IMG03 0.4528 0.0065 5.8
4
IMG 04 0.4378 0.0083 3.5
5
IMG 05 0.4424 0.0080 3.8
6
IMG 06 0.4512 0.0067 5.4
7
IMG 07 0.4593 0.0034 2.202
8
IMG 08 0.4220 0.0149 1.10
9
IMG 09 0.4530 0.0062 6.43
10
IMG 10 0.4490 0.0081 3.75
4. Results & Discussion
In this paper we also compared the accuracy and the false enrollment rate of the proposed iris verification algorithm with the algorithms proposed by Daugman [1] and Tan et al . [9].Result shows that the verification performance of the proposed iris recognition algorithm is evaluated assuming that there is 0% false enrollment rate. Figure shows the intra-class and inter-class distance distribution plot of Hamming Distance calculated by the proposed 2D log Gabor filter. The accuracy (Accuracy = 100 – (FAR+FRR)/2) of the iris verification system is 99.96%, with zero false acceptance rate. This shows that the proposed decision algorithm reduces the false rejection rate. Figure 5 shows that by increasing the number of training templates the FRR decreases and after 10 training samples it becomes approximately constant.
Table 2: Performance Comparison of FAR & FRR
Threshold Value FAR FRR Accuracy
0.20
0.25
0.00
0.00
0.97
0.87
99.51
99.56
0.30
0.35
0.40
0.45
0.50
0.00
0.00
0.00
0.65
1.00
0.67
0.40
0.0750
0.250
0.00
99.66
99.8
99.96
99.8
99.5
With the ‘CASIA’ data set perfect recognition is possible. With a separation point of 0.4 a false accept rate and false reject rate of 0.00% and 0.0750% respectively is achieved, which still allows for accurate recognition as shown in table
2.
Table 3: Performance Comparison of Accuracy
ALGORITHM FAR FRR ACCURACY
Daugman [1] 0.00 0.13 99.94
Tan et. Al [9]
Proposed system
0.00 1.03
0.00 0.0750
99.49
99.96
Volume 3, Issue 5, May 2014 Page 281

International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 5, May 2014 ISSN 2319 - 4847
In this paper we also compared the Different algorithm, accuracy and the false rejection rates of the proposed algorithm with the algorithms proposed by Daugman [1], and Tan [9]. Table 1 shows that time required for system is very less than other system. Table III shows the verification accuracy of the algorithms.
5.
C
ONCLUSION
The uniqueness of the iris and low probability of a false acceptance or false rejection all contribute to the benefits of using iris recognition technology. As the same image is taken as input and database, so the Hamming distance obtained is 0. Thus the templates are perfectly matched. The system performed with perfect recognition on a set of 75 eye images; however, tests on another set of 60 images. Experimental results show that the proposed system is used to reduce the false rejection cases with zero false acceptances rate. Also, the proposed system is compared with the existing algorithms.
References
[1] G. Daugman, “High Confidence Visual Recognition of Persons by a Test of Statistical Independence”, IEEE
Transactions on Pattern Analysis & Machine Intelligence. Vol. 15, No. 11, pp. 1148-1161, 1993.
[2] Daugman, “Recognizing People by their Iris Pattern”, Information Security Technical Report, Vol. l3 (1) pp 33-39,
1998.
[3] Wildes P. “Iris Recognition: An Emerging Biometric Technology”, Proceedings of IEEE, Vol. 85 (9) pp 1348-1363,
1997.
[4] Masek Libor, “Recognition of Human Iris Patterns for Biometric Identification”, School of Computer Science and
Software Engineering, The University of Western Australia. 2003.
[5] Zhen ping HU, Wang Chengfu, YU Lina, “Iris Location Improved Randomized Hough Transform” Yi Qi Yi Biao
Xue Bao, Vol. 24 (5) pp 477-479, 2003.
[6] Ali Ajdari, Karim Faez. “Fast Circle Detection Using Gradient Vectors”, Proceedings of 7nd Digital Image
Computering: Techniques and Applications, Vol. 1 pp 879-887, 2003.
[7] Chen Tehchuan Chung Kouliang, “An Efficient Randomized Algorithm for Detecting Circles”, Computer Vision and Image Understandings.
[8] L. Ma, T. Tan, Y. Wang and D. Zhang, “Efficient Iris Recognition by Characterizing Key Local variations”, IEEE
Transactions on Image Processing, Vol. 13, No. 6, 2004, pp. 739-750.
[9] T. Tan, L. Ma, Y. Wang and D. Zhang, “Personal Identification based on Iris Texture Analysis”, IEEE Transactions on Pattern Analysis and Machine Intelligence , Vol. 25, No. 12, 2003, pp. 1519-1533.
[10] L. Machala and J. Pospisil, “Alternatives of the Statistical Evaluation of the Human Iris Structure”, Proceedings of the SPIE, Vol. 4356, 2001, pp. 385-393.
[11] J. G. Daugman, “How iris recognition works,”IEEE Trans. Circuits and Syst. For Video Tech. 14(1), 21-30 (2004).
[12] John Daugman, “The importance of being random: statistical principles of iris Recognition”, Pattern recognition, pp.1-13.
[13] R. Wildes, etc., “A Machine-vision System for Iris Recognition”, Machine Vision and Applications, vol.9, pp.1-8,
1996.
[14] John Daugman, “New Method in Iris Recognition”, IEEE Transitions on Systems, Man &, Cybernetics Part B:
Cybernetics, Vol. 37, No. 5, October 2007.
Volume 3, Issue 5, May 2014 Page 282