Design of Iris Recognition System Using Reverse Biorthogonal Wavelet for UBIRIS Database

232 Design of Iris Recognition System Using Reverse Biorthogonal Wavelet for UBIRIS Database Shivani 1, Er. Pooja kaushik 2, Er. Yuvraj Sharma 3 1 M.Tech Final Year Student, 2,3 Asstt. Professor of Electronics and Communication Engineering, 1,2,3 Department of Electronics and Communication Engineering, 1,2,3 Maharishi Markandeshwar University, Mullana, Ambala, Haryana, INDIA. ABSTRACT In recent years, iris recognition has become a popular recognition system as compare to the existing systems. It is an important biometric method for human identification with high accuracy. It is the most reliable and accurate biometric identification system available today. The most popular biometric types are: signature, face, iris, finger prints, hand and voice. Among all these biometric techniques, iris recognition is one of the accurate due to its high reliability for personal identification. A number of techniques are available for iris recognition system like Morlet Wavelet, Reverse Biorthogonal Wavelet, K-D Tree Method and Biorthogonal Wavelet etc. In this paper Reverse Biorthogonal Wavelet is used for Iris Recognition system and using this technique Hamming distance between iris codes is calculated to measure the difference between two iris images in the database. Iris recognition is then performed by matching the iris pair with the minimum Hamming distance. The iris recognition system consists of four steps, (1) Normalization, (2) Segmentation, (3) Feature Extraction, (4) Matching. Fig.1 shows the structure of the human eye and their other parts such as pupil, sclera and eyelids are present along with the iris. 1.1. THE HUMAN IRIS The iris is a thin circular diaphragm, which lies between the cornea and the lens of the human eye [2]. The image of the human eye is shown in figure1. The pupil is surrounded by the area that is called iris. The function of the iris is used to control the amount of light is to be entered into the eye. The average diameter of the iris is 12 mm, and the pupil size can vary from 1% to 8% of the iris diameter [2]. Iris Upper Eyelid Keywords: Iris Recognition and Personal Identification, UBiris database. I. INTRODUCTION The iris recognition is the most accurate technology available today. It has been a popular area of research in the past decade. It plays the major role to identify the person with the help of huge database. In 1987, Flom and safir [12] obtained the first patent for an automated iris recognition process, and a few years later Daugman published a method that is the basis of near all of the currently deployed systems. This research presents a design of iris recognition system using reverse Biorthogonal wavelet for UBIRIS.v1 database. It uses the reverse Biorthogonal wavelet [4] for feature extraction. The iris recognition rate is improved by relatively accurately locating the iris as well as pupil and eliminating the signal noise in an eye image. Pupil Lower eyelid Scelra Fig. 1 The human eye The unique patterns in the human iris are formed by 1 month of birth, and remain unchanged throughout one s lifetime. Iris is a thin, circular structure in the eye. It controls the diameter and size of the pupils and thus the amount of light reaching the retina. 1.2. PUPIL The pupil is the black opening in the centre of the eye. Light enters through the pupil and goes through the lens, which focuses the image on the retina. The size of the pupil is controlled by muscles. When more light is required, the pupil expands as per the need. In bright light, the pupil becomes smaller. The function of the iris

233 is to control the amount of light entering through the pupil. The larger the pupil, the more light can enter into the pupil [3]. 1.3. ADVANTAGES OF IRIS RECOGNITION SYSTEM Uniqueness: The iris pattern is unique as compared to the other pattern because of texture details in iris image, such as freckles, coronas, stripes and furrows. Even twins do not have the same iris pattern. They have totally different iris details. It is the most reliable and informative biometric pattern. Stability: The iris texture is to be formed during gestation and the main structures of iris are shaped after 1 months of birth. It has also been shown that the iris is essentially stable across ones s lifetime. Scalability: Image of the iris region can be normalized into rectangle regions so that binary feature codes of fixed length can be extracted. Therefore iris recognition is used for large-scale personal identification applications. Security: The iris recognition is more secure simply because of unique pattern. Human iris has many special and physiological characteristics. This technology is very secure as compare to the other. and pupil. This step of segmentation helps to separating the iris part from the eye image. The inner and outer boundaries of iris are localized by finding the edge image using the canny edge detector. In this recognition system canny edge detector have three steps: finding the gradient, non-maximum suppression and the hysteresis thresholding. By the use of canny edge detector, it reduces the pixels on the circle boundary. But with the use of Hough transform, there is no change by the absence of few pixels. It means that successful localization boundary can be obtained. Edge detection is used to detect the boundaries of the iris and the pupil. Daugman [21] used the Integro-differential operator to detect the boundaries and the radii. This Daugman theory of edge detector behaves as a circular edge detector by searching the gradient image along the boundary of circle of increasing radii. Then, maximum sum of all the circles is calculated and is used to find the circle center and radii. The use of Hough transform in another way is to detect the parameters of the object, and in this case to find the circles in the edge image. That means every edge pixel have different radii. The Hough transform is used to calculate the outer boundary of the iris using the whole image and then is used to calculate the pupil boundary. II. METHODOLOGY The algorithm can be divided into four steps: Segmentation, Normalization, Feature Extraction and Matching. The block diagram given below: (a) (b) (c) Fig. 3 Segmented outer and inner area of iris (a) Localized outer and inner Area of iris (b) Localized of pupil (c) Localized of iris boundary. 2.2 NORMALIZATION Fig. 2 Block Diagram of a iris recognition system 2.1 SEGMENTATION The initial stage of iris recognition system is to acquire an image called image acquisition. The second step is to convert the acquired image into gray scale image. In this paper digitized grayscale image is used. In a grayscale image, each pixel contains the single value that indicates the grayscale intensity, which may vary from to 255. This paper uses the part of the eye to carry the information with the help of iris. Iris lies between sclera In order to remove the pupil; Dugman [18] rubber-sheet method was used to unwrap the iris image. The main aim of normalization is to convert the iris image from cartesian coordinate (x,y) to polar coordinate (r, ). This paper uses the polar coordinate transform to carry on the normalization, because the inner and outer circle are not cocentric [18]. Each iris image was normalized and unwrapped to multiple rows and columns. This process cuts the iris image at one point along the radial direction and streches it into a rectangular shape [18]. (a) (b)

234 = (1) (c) (d) Fig. 4 The normalized iris images of different eyes. 2.3 FEATURE EXTRACTION The iris pattern in a normalized iris image was repersented by a group of feature vectors for pattern comprasion (matching). This repersentation was built by calculating a set of feature values to form a feature vector, which was achieved by using a reverse biorthogonal wavelet [2]. Transforming the input data into the set of features is called feature extraction. The use of reverse biorthogonal wavelet is used to extract feature form the iris region. The main aim of feature extraction is used to convert image(data) into the binary form. The waveform of reverse biorthognal is very simple i.e. in the form of -1 to +1. Reverse biorthogonal wavelet uses both time and frequency domain that s why this wavelet creates very simple wave forms.wavelet function is always equal to the window function therefore, it solves the problem between the time resolution and frequency resolution [11]. Time and frequency analysis are characteristics of wavelet transform. The wavelet transform coefficient can reflect the signal or function. Then accuracy is to be improved. 2.4 MATCHING The last step of iris recognition is matching the image with the stored image or database. UBIRIS.v1 database is to be used for this iris recognition system. It contains the 765 grayscale images with 18 unique eye 5 and different images of each unique eye. For the comparison of the two iris codes, the hamming distance algorithm is employed. Since the iris region contains features with very high degrees of freedom, and each iris produces a bit-pattern which is independent to that produced by another iris, whereas the codes produced by the same iris would be similar. If two bits patterns are completely independent, then the ideal Hamming distance between the two patterns will be equal to.5. It happens because independent bit pattern are completely random. Therefore, half of the bits will agree and half will disagree between the two patterns. The Hamming distance is the matching metric employed by Daugman [21], and calculation of the Hamming distance is taken only in bits that are generated from the actual iris region. The Hamming distance will be defined as follows: III. Where and are the two bit wise template to compare and N is the number of bits represented by each templates. In the present case, the HD is.445 which signifies that if the hamming distance between the two templates is below.445 than both the irises are of same eye and if the HD value falls above.445, it signifies that both the irises are from different eye. RESULT Inner-class and inter-class hamming distance: The main objective of iris recognition system is to determine the value of hamming distance which can separate innerclass and inter-class iris image. Hamming distance between the iris images of same person is known as inner-class and hamming distance between the iris patterns of different persons is known as inter-class. A much better iris recognition system tries to achieve the overlapping of the hamming distance of two classes that should as small as possible, because results depend upon the separation value. When the hamming distance value of two iris pattern is less than the separation point, then results shows that the code of two iris is belongs to a same person. Otherwise, the code of two iris is belongs to a different person. The overlap between the two distributions determines the error rate [2]. For the innerclass irises, the right side area of separation point creates the false rejection rate. For the inter-class iris, the left side area of the separation point creates the false acceptance rate. 5 45 4 35 3 25 2 15 1 5 inner-class Mean=.4 inter-class Mean=.54.1.2.3.4.5.6.7.8.9 1 Fig. 5 Distribution of Inner-class and Inter-class

235 12 1 8 6 4 2.1.2.3.4.5.6.7.8.9 1 Fig. 6 Distribution of Inner-class 12 Fig. 5 shows the distribution of inner-class and interclass hamming distance. This result clearly shows that the separation between inter-class and inner-class hamming distance. This result indicates that it was still very much possible to clearly separate the iris codes of different people by using this technique. In fig. 7 and 8 graph shows the distribution of inner-class and interclass hamming distance. This graph is shown in the form of bar. The graph is the plot between hamming distance and the number of images TABLE 1. EXPERIMENTAL RESULTS Time 12 (seconds) Seperation.445 point Inner.4 Mean Inter.54 Mean 1 8 6 4 2 -.2.2.4.6.8 1 1.2 5 45 4 35 3 25 2 15 1 5 Fig. 7 of Inner-class -.2.2.4.6.8 1 1.2 Fig. 8 of Inter-class IV. CONCLUSION This paper presents an iris recognition system using reverse biorthogonal wavelet for UBIRIS.v1 database, which was tested using database of grayscale eye images that is a set of human eye images from UBIRIS.v1 database is used in the experiments. From UBIRIS.v1 database, 5 different images of 5 persons are taken(25 sample of the iris). Firstly, segmented algorithm was presented, which would localize the iris region and the segmentation was achieved through the use of the circular hough transform for localizing the iris and pupil region. It also uses the canny edge detection for wide range of edges. Next, the normalization is used to convert the cartesian coordinate into the polar coordinate. This was achieved by implementing a version of daugman s rubber sheet model, which is unwrapped into a rectangular block with polar dimensions. Finally in the feature extraction, the reverse biothogonal wavelet is used to extract the feature of the human iris. This decomposition obtained the coefficients and converted into binary codes and is to be used for calculation of hamming distance for matching purpose. The results show that the average time for both search and matching is approximately 12 seconds. Seperation point between inner and inter class is.445. The innerclass mean of the Hamming distance is.4, while the inter-class mean of the Hamming distance is.54. The Proposed algorithm exactly finds iris regions with higher reliability, greater efficiency and in small time duration.

236 V. REFERENCES [1] Shivani, Pooja Kaushik, Yuvraj Sharma, Review of Iris Recognition System, IJRIT International Journal of research in Information Technology, Vol. 2, April 214. [2] Steve Zhou and Junping Sun, A Novel Approach for Code Match in Iris Recognition,IEEE International Conference on Computer And Information Science, June 213. [3] Sukhwinder Singh, Ajay Jatav, A Closure Looks To Iris Recognition System, International Journal Of Scientific & Engineering Research, Vol. 3, March 213. [4] Zaheer Zainal Abidin, Mazani Manaf, Experimental Approach On Thresholding Using Reverse Biorthogonal wavelet decomposition For Eye Image IEEE,213. [5] Adams Wai-Kin Kong, Modeling IrisCode and Its Variants as Convex Polyhedral Cones and Its Security Implications IEEE Transactions On Image Processing, Vol. 22, March 213. [6] Bimi Jain, Dr.M.K.Gupta, Prof.JyotiBharti, Efficient Iris Recognition Algorithm Using Method Of Moments International Journal of Artificial Intelligence & Applications (IJAIA), Vol.3, September 212. [7] Javier Galballya, Arun Rossb, Marta Gomez- Barreroa, Julian Fierreza, Javier Ortega-Garciaa A New Vulnerability Of Iris Recognition Systems, 212 [8] B. Thiyaneswaran, S. Padma, Iris Recognition using Left and Right Iris Feature of the Human Eye for Bio-Metric Security System, International Journal of Computer Applications, Vol.5, July 212. [9] Yu Li, Zhou Xue Fast, Iris Boundary Location Based on Window Mapping Method Seventh International Conference on Computational Intelligence and Security, 211 [1] H.R. Gite, C.N. Mahender Iris Code Generation And Recognition International Journal Of Machine Intelligence Vol. 3, 211 [11] Zhonghua Lin Bibo, Iris Recognition Method Based on the Imaginary Coefficients of Morlet Wavelet Transform, Seventh International Conference on Fuzzy Systems and Knowledge Discovery 21. [12] Hugo Proenca, Silvio Filipe, The UBIRIS.v2: A Database of Visible Wavelength Iris Images Captured On-the-Move and At-a-Distance,IEEE Transactions on Pattern Analysis And Machine Intelligence, Vol. 32, Aug 21. [13] Zhonghua Lin, A Novel Iris Recognition Method Based on thenatural-open Eyes IEEE,21. [14] Fei Tan and Zhengming Li, Iris Localization Algorithm Based on Gray Distribution Features, IEEE, 21. [15] Lin Zhonghua, Lu Bibo, Adaptive Iris Recognition Method Based on the Marr Wavelet Transform Coefficients School of Computer Science and Technology Henan Polytechnic University,29. [16] Jing-Hui Li, New Algorithm of Iris Localization,World Congress on Computer Science and Information Engineering IEEE,29. [17] John Daugman, New Methods In Iris Recognition, IEEE Transactions On Systems, Vol. 37, October 27. [18] Mr. P.P.Chitte, Prof. J.G.Rana, IRIS Recognition System Using ICA, PCA, Daugman s Rubber Sheet Model Together International Journal of Computer Technology and Electronics Engineering (IJCTEE), Vol. 2, 27. [19] J.Lu and M.Xie, A new method for iris localization, IEEE, 27. [2] Padma Polash Paul, Md. Maruf Monwar, Human Iris Recognition for BiometricIdentification, IEEE, 27. [21] John Daugman, How Iris Recognition Works, IEEE Transactions on Circuits and Systems for Video Technology, Vol. 14, Jan 24. [22] Traian Dogaru and Lawrence Carin, Multiresolution Time-Domain Using CDF Biorthogonal Wavelets, IEEE Transactions On Microwave Theory And Techniques, Vol. 49, May 21 [23] Samuel S. Osofsky, Calculation Of Transient Sinusoidal Signal Amplitudes Using The Morlet Wavelet, IEEE Transactions On Signal Processing, Vol. 47, December 1999. [24] Shyh-Jier Huang, Cheng-Tao Hsieh and Ching-Lien Huang, Application of Morlet Wavelets to Supervise Power System Disturbances,IEEE Transactions on Power Delivery, Vol. 14, No. 1, January 1999. [25] Dong Wei, Jun Tian, Raymond O. Wells, Jr., and C. Sidney Burrus, A New Class of Biorthogonal Wavelet Systems for Image Transform Coding, IEEE Transactions. Shivani is a final year student of M.Tech (Electronics and communication engineering) at MM University, Mullana, she has received her B.Tech (Electronics and communication engineering) from Doaba Women Institute of Engineering and Technology(DWIET), Punjab, India in 28.