IJRIT International Journal of Research in Information Technology, Volume 2, Issue 4, April 2014, Pg: 620- 624

International Journal of Research in Information Technology (IJRIT) www.ijrit.com

ISSN 2001-5569

A Review: Study of Iris Recognition Using Feature Extraction & Edge Detection Yachna Kumari1, Rohini Sharma2 1

2

(Student) Department of Computer Science & Engineering, Maharishi Markandeshwar University Mullana, Ambala, India [email protected]

(Associate Professor) Department of Information Technology, Maharishi Markandeshwar University Mullana, Ambala, India [email protected]

ABSTRACT Biometric is identification of a human by their physiological & behavioral characteristics. There are many type of biometric method available such as finger print, face recognition, speech recognition, DNA & Iris recognition. Iris recognition is one of the secure & accurate due to Epigenetic traits which do not change genetically. This paper, analyses the Iris recognition method segmentation, normalization, feature extraction & matching. Feature extraction is done using Gabor filter which is appropriate for texture representation & a linear filter for extraction. Edge detection is a process to find the true edge with noise free iris image for better matching result. Comparison of various edge detection techniques Canny, Sobel & Prewitt are made on their Peak Signal to Noise Ratio (PSNR) & Mean Square Error (MSE). A result show Canny has more accuracy than sobel & prewitt. Keyword: Iris recognition, Feature extraction, Gabor filter, Edge detection technique, PSNR & MSE.

1. INTRODUCTION Biometric deals with the identification of an individual on the basis of their physiological & behavioral feature. Nowadays there are many method used for human identification which include face recognition, finger print, speech recognition, DNA & Iris recognition. Iris recognition is the human eye internal part which is well protected from the damage & thread. Iris is eye part in between cornea & retina. Iris pattern consist of crypt, freckle, coronas, furrows, ridges & blood vessel in it. Iris is unique due to epigenetic factor [2] that remains unchanged throughout adult lifetime. Identical twins doesn’t have same iris pattern & a person iris pattern vary from left to right eye. The algorithm for iris recognition was developed by John Daugman at Cambridge University. Human recognition contains the following method: Image acquisition to capture the iris of an eye in form of image. Localization of inner or outer boundaries of iris. Normalization is performed to get all images in proper format for processing. Feature extraction is the technique for extracting the iris template & pattern. Matching is done on new iris template & iris template in database. There is various technique of matching hamming distance, Weighted Euclidean distance, normalizes correlation

Yachna Kumari, IJRIT

620

IJRIT International Journal of Research in Information Technology, Volume 2, Issue 4, April 2014, Pg: 620- 624

For feature extraction Gabor filter is used which is a linear filter for texture representation. Edge detection is used to find real edges with noise free image for better matching. The operator of edge detection such as canny, sobel & prewitt are used in this study. The purpose of this study is to investigate the effect of Gabor filter on feature extraction & to compare the various edge detection operator canny, sobel, prewitt. The objective of this study is to get the PSNR & MSE value of iris pattern.

2. RELATED WORK Daugman first proposed an algorithm for iris recognition. Operators are used to detect center or boundaries of iris. His work is based on iris code. Iris feature extraction algorithm uses the 1-D Gabor filter to extract the unique iris codes which are then compared using Hamming distance. Tze Weng et. Al used a Haar wavelet decomposition method to analyze the pattern of human iris. In feature extraction module, unrolled iris image is filtered using high pass filter & low pass filter for four times to corresponding coefficients, recognition rate of 98.45% is achieved [11]. V. Saravanan used Gabor filter for feature extraction which possesses a null FAR & FRR value & can also be lowered. Matching is done using hamming distance & Euclidean distance is used in this paper for comparing the feature extraction method [5]. C. Sanchez & R. Sanchez in their research has given a study of feature extraction algorithm such as Gabor filter & Zero crossing representation discrete dyadic wavelet transform has tested using various pattern recognition method. Hamming distance shows good result with 99.6% of classification success [7]. G. Padmavathi et.al used the comparison of canny, sobel, prewitt & log in detecting edges in noisy IR image. Sobel operator with median filter performs well in detecting edge [3].

3. METHODOLOGY AND IMPLEMNETATION 3.1 Segmentation Segmentation is the process of dividing a digital image into multiple regions consists of set of pixels. Segmentation is used to locate objects & boundaries such as lines & curve etc. Segmentation is done to detect the inner & outer boundaries of iris. Segmentation is the most important & also difficult process in image processing. The quality of image processing heavily depends upon the quality of segmentation process. Segmentation is obtained either by unsupervised clustering or by considering a gradient in the texture. Iris segmentation is important because correct iris region is needed to generate the template for accurate matching. Segmentation in Iris Recognition is used to set boundaries between iris , sclera, pupil & eyelids of an eye. Several researches have been made by different authors for iris location & segmentation, Daugman[13] for detecting the limbic and pupil boundaries implements integro-differential operator.

Fig.1: Eye Image with boundaries

Yachna Kumari, IJRIT

621

IJRIT International Journal of Research in Information Technology, Volume 2, Issue 4, April 2014, Pg: 620- 624

3.2 Image normalization Once the iris region is segmented, the next step is to normalize iris region to compensate for their variation. As iris is capture in different size with varying distance, due to illusion variations the radial size of a pupil may change. The resulting deformation of the iris texture will affect the performance of feature extraction or matching stage in image processing. Normalization is a process used to unwrapping the iris and convert it from Cartesian to polar coordinates.

Fig.2: Normalized iris image Normalization is done using Daugman’s Rubber sheet model. Rubber sheet model remap each pixel in localized iris region from the Cartesian coordinates to polar coordinates.

Fig.3: Unwrapping the iris

3.3 Feature extraction The iris pattern in a normalized iris image was represented by a group of feature vectors for pattern comparison (matching). This representation was built by calculating a set of feature values to form a feature vector, which was achieved by using a 1-D Gabor filter. 1-D Gabor filter technique extract the unique set of features from the iris & then store them. 1-D Gabor filter is used because Gabor filter is complex, which gives real & imaginary part. To understand the concept of Gabor filter it is necessary to understand Gabor Wavelet [6]. G(x, y) = S(x, y)Wr(x, y)……………………………………………………………………………….(1) Complex carrier in the form: S(x, y) = ej(2π(uo x+voy)=P)………………………………………………………………………………………………………………………(2) Extract real & imaginary part: Real Part:. Yachna Kumari, IJRIT

622

IJRIT International Journal of Research in Information Technology, Volume 2, Issue 4, April 2014, Pg: 620- 624

Re(s(x,y)) = cos(2π(uox + voy) +P……………………………………………………………………….…..(3) Imaginary Part: Im(s(x, y)) = sin(2π(uox+ voy) +P……………………………………………………………………….…..(4) Uo: Frequency of horizontal sinusoidal. Vo: Frequency of vertical sinusoidal. P: Arbitrary phase shift.

3.4 Edge Detection Edge detection provides a number of derivative operators which significantly detect the local changes of intensity in an image. Detection occur on the boundary between two different region of iris features. The purpose of edge detection is to significantly reduce the amount of data in an image, it returns a binary image containing 1’s where edge is found & 0’s else where. The operators used for edge detection are Sobel, Prewitt & Canny. The most powerful edge detection method is canny method because it uses two different thresholds in detecting between strong & weak edge. According to [4] it is observed that canny operator is best suited to extract most of the edges to generate the iris code for comparison. Sobel is a discrete differentiation operator, which compute gradient of image intensity. The result of sobel operator is corresponds to gradient vector and convolving the Iris feature with small, separable & filter in horizontal & vertical direction for high frequency variation in iris feature. Prewitt operator is almost the same with sobel operator, provided that the operator is not divided by 2. It is discrete differentiation operator & computes the gradient of the image intensity t each point. [4] Comparison of the PSNR value of canny, sobel & prewitt for edge detection are given as: PSNR value for S I 001 L04jpg-polarjpg Before

After Canny

24.93

9.123

Prewitt

Sobel

18.5

18.5

Table1: Comparison of PSNR for edge detection [5] The experiments have been implemented using human eye image from CASAI database. The experiment shows in Table1 PSNR value after edge detection of canny operator is less than the sobel, prewitt. Thus the canny operator is more reliable & produce accurate performance compared to other two operators.

3.5 Matching Comparison of the bit template generated is done to check if the two irises belong to the same person. Calculation of Hamming Distance (HD) is done for the comparison. HD is a fractional measure of dissimilarity of two binary template. Hamming distance is calculated between two template one resulting from processing & other one stored in database This code comparison uses the iris code data and the noisy mask data.

Yachna Kumari, IJRIT

623

IJRIT International Journal of Research in Information Technology, Volume 2, Issue 4, April 2014, Pg: 620- 624

Hamming Distance =

      ! "# $%!"&

4. CONCLUSION In this paper study of iris recognition is made based on 1-D Gabor filter which is used to extract the unique feature from iris pattern known as iris code. Edge detection methods canny, sobel, prewitt are used for detecting the edges between iris boundaries. PSNR value of edge detection method is calculated. Based on the result we conclude that canny shows less value as compare to sobel & prewitt. Thus Canny edge detection shows higher accuracy in iris recognition system.

REFERENCES [1] A. Jain, “An Introduction to Biometric recognition”, IEEE transactions on circuit & system for video technology vol. 14, pp 4-20, 2004. [2] J.Daugman & C. Downing, “Epigenetic randomness, complexity & singularity of human iris pattern”, Proceedings. Biological Sciences / The royal society, vol. 268, no. 1477, pp. 1737- 40, Aug.2001. [3] G. Padmavathi, P. Subhashini & P.K. Lavanya, “Performance evaluation of various edge detectors & filter for noisy IR image”, Sensors, signals, visualization, imaging, simulation and material. [4] Z.Z.Abidin, M.Manaf & A.S.Shibghtatullah, “Feature extraction from epigenetic traits using edge detection in iris recognition system”, ICSIPA, vol. 269, sept 2013 IEEE. [5] V.Saravanam & R.Sinduja, “Iris Authentication through Gabor Filter using DSP processor”, IEEE Conference on Information and Communication Technologies, vol.978, no.4763, June 2013. [6] Mohd. Tariq Khan, D. Arora & S. Shukla, “Feature Extraction through Iris image using 1-D Gabor Filter on different Iris Database”, IEEE 2013. [7] C. Sanchez – Avila & R. Sanchez – Reillo, “ Two different approaches for iris recognition using Gabor filters & multiscale Zero- crossing representation”, Pattern Recognition, pp. 231- 240, 2005. [8] S. Joshi, Bhavana. Desai & J.L. Kalyan, “Proposed Approach for Iris Recognition in security Based Applications”, International Journal of Science and Research, vol. 2, Feb 2013. [9] Priyam Gosh, M. Rajashekharababu, “ Authentication using Iris Recognition with Pallel Approach”, International Journal of Computer Science & Network Security, vol. 13, no. 5, 2013. [10] Yulin Si, Jiangyuan Mei & H.Gao, “Novel Approaches to Improve Robustness, Accuracy and Rapidity of Iris Recognition Systems”, IEEE Transaction on Industrial Informatics, vol.8, 2012. [11] Tze Weng Ng, T.L.Tay & Siak Wang Khor, “ Iris Recognition Using Haar Wavelet Decomposition”, ICSPS, 2010. [12] John Daugman, “New Method in Iris Recognition”, IEEE Transaction on System, vol.37, October 2007. [13] J. G. Daugman, “ Demodulation by complex-valued wavelets for stochastic pattern recognition”, International Journal of Wavelets, Multiresolution and information Processing, vol. 1, no. 1,pp. 1-17,2003

Yachna Kumari, IJRIT

624

A Review: Study of Iris Recognition Using Feature Extraction ... - IJRIT

INTRODUCTION. Biometric ... iris template in database. There is .... The experiments have been implemented using human eye image from CASAI database.

1009KB Sizes 0 Downloads 93 Views

Recommend Documents

Review on Fingerprint Recognition System Using Minutiae ... - IJRIT
IJRIT International Journal of Research in Information Technology, Volume 1, ... it a personal identification and ,thus have a number of disadvantages like tokens.

Affine Invariant Feature Extraction for Pattern Recognition
Nov 13, 2006 - I am also thankful and grateful to my thesis supervisor Dr. Syed Asif Mehmood Gilani for his continuous guidance and support that he extended during the course of this work. I had numerous discussions with him over the past year and he

Adaptive Extraction of Information Using Relaxation Labelling ... - IJRIT
Abstract-Internet forums are important services where users can request and exchange information ... 2.2 Web data extraction based on partial tree alignment:.

Learning a Selectivity-Invariance-Selectivity Feature Extraction ...
Since we are interested in modeling spatial features, we removed the DC component from the images and normalized them to unit norm before the learning of the features. We compute the norm of the images after. PCA-based whitening. Unlike the norm befo

Wall Follower Robot Using Fuzzy Logic: A Review - IJRIT
system that enables a mobile robot in moving through a corridor or following a .... The gain scheduling controller will be used before the FLC to control the error signal ... 2) computing the path winding number, 3) learning a combinatorial map,.

Wall Follower Robot Using Fuzzy Logic: A Review - IJRIT
and to enter open areas of the rooms [1,2]. In attempts .... Computer simulations are carried out to compare between a Knowledge Based Fuzzy Logic Controller.

A Random Field Model for Improved Feature Extraction ... - CiteSeerX
Center for Biometrics and Security Research & National Laboratory of Pattern Recognition. Institute of ... MRF) has been used for solving many image analysis prob- lems, including .... In this context, we also call G(C) outlier indicator field.

0123965497-Feature-Extraction-Processing-Computer-Edition.pdf ...
0123965497-Feature-Extraction-Processing-Computer-Edition.pdf. 0123965497-Feature-Extraction-Processing-Computer-Edition.pdf. Open. Extract. Open with.

feature extraction & image processing for computer vision.pdf ...
feature extraction & image processing for computer vision.pdf. feature extraction & image processing for computer vision.pdf. Open. Extract. Open with. Sign In.