International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volu...
Upcoming SlideShare
Loading in …5
×

A comparison of multiple wavelet algorithms for iris recognition 2

712 views
648 views

Published on

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
712
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
47
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

A comparison of multiple wavelet algorithms for iris recognition 2

  1. 1. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME386A COMPARISON OF MULTIPLE WAVELET ALGORITHMS FOR IRISRECOGNITIONSayeesh1, Dr. Nagaratna P. Hegde 21Asst. Professor, Dept. of CS & E, Alva’s Institute of Engg. & Tech., Shobhavana Campus,Mijar, Moodbidri – 574 225, Karnataka State.2Professor, Dept. of CS & E, Vasavi College of Engg., Ibrahimbhag,Hyderabad – 31, Andhra Pradesh.ABSTRACTPersonal identification has become the need of modern day life. The identificationmust be fast, automatic and foolproof. Biometrics has emerged as a strong alternative toidentify a person compared to the traditional ways. Also biometric identification can be madefast, automatic and is already foolproof. Among other biometrics, Iris recognition hasemerged as a strong way of identifying any person. Iris recognition is one of the newerbiometric technologies used for personal identification. It is one of the most reliable andwidely used biometric techniques available. In general, a typical iris recognition methodincludes capturing iris images, testing iris liveness, image segmentation, and imagerecognition using traditional and statistical methods. Each method has its own strengths andlimitations. In this paper, we compare the performance of various wavelets for Irisrecognition like complex wavelet transform, Gabor wavelet, and discrete wavelet transform.Keywords- Iris recognition, complex wavelets, Gabor wavelets, discrete wavelet transform.I. INTRODUCTIONThe demand for security systems is increasing day by day. Rigorous search for differentverification and identification techniques is the need of the day. Most traditional methods ofsecurity require a person to possess some type of physical possession, such as a key, or toknow certain information, such as a password. These techniques are not as secure asorganizations may desire. In recent years, the increasing capabilities of computers haveallowed more sophisticated and intelligent personal identification methods. BiometricINTERNATIONAL JOURNAL OF COMPUTER ENGINEERING& TECHNOLOGY (IJCET)ISSN 0976 – 6367(Print)ISSN 0976 – 6375(Online)Volume 4, Issue 2, March – April (2013), pp. 386-395© IAEME: www.iaeme.com/ijcet.aspJournal Impact Factor (2013): 6.1302 (Calculated by GISI)www.jifactor.comIJCET© I A E M E
  2. 2. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME387techniques, which use uniquely identifiable physical or behavioral characteristics to identifyindividuals, are one such method. Commonly used biometric features are the face,fingerprints, voice, DNA, retina, and the iris. Iris recognition is regarded as the most reliableand accurate biometric identification system available. Iris recognition is a biometric-basedmethod of identification. This method has many advantages, such as unique, stability, can becollected, nonaggressive, etc. The iris recognitions error rate is the lowest in most biometricidentification method. Now many research organizations at home and abroad spend a lot oftime and energy to do research of iris recognition. The human iris is an annular part betweenthe pupil (generally appearing black in image) and white sclera has an extraordinarystructure. The iris begins to form in the third month of gestation and structures creating itspattern are largely complete by the eight months, although pigment accretion can continue inthe first postnatal years. The word iris is generally used to denote the colored Portion of theeye. It is a complex structure comprising muscle, connective tissues and blood vessels. Theimage of a human iris thus constitutes a plausible biometric signature for establishing orconfirming personal identity. Further properties of the iris that makes it superior to fingerprints for automatic identification systems include, among others, the difficulty of surgicallymodifying its texture without risk, its inherent protection and isolation from the physicalenvironment, and its easily monitored physiological response to light. Additional technicaladvantages over fingerprints for automatic recognition systems include the ease of registeringthe iris optically without physical contact beside the fact that its intrinsic polar geometry doesmake the process of feature extraction easier. It involves using photographs of a person’seye(s) to determine the identity of the individual. The iris contains unique features, such asstripes, freckles, coronas, etc., collectively referred to as the texture of the iris. This texture isanalyzed and compared to a database of images to obtain a match. The probability of a falsematch is close to zero, which makes iris recognition a very reliable method of personalidentification.This paper discusses performance of different wavelet based algorithm for iris imageenhancement, noise reduction, feature extraction, and matching.II. RELATED LITERATUREA biometric system provides the automatic recognition of an individual based onsome unique feature or characteristic possessed by the individual. This section describes theoverview of the iris recognition system, theoretical background about wavelets, andprinciples of iris recognition.A. Overview of the Iris Recognition systemImage processing techniques can be employed to extract the unique iris pattern from adigitized image of the eye and encode it into the biometric template, which can be stored indatabase. This biometric template contains an objective mathematical representation of theunique information stored in the iris, and allows the comparisons made between templates.When a person wishes to be identified by an iris recognition system, their eye is firstphotographed and then template is created for their iris region. This template is thencompared with the template stored in a database, until either a matching template is foundand a subject is identified, or no match is found and subject remains unidentified.Human iris recognition process is basically divided into four steps.
  3. 3. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME388i) Localization: Inner and outer boundaries of the iris are extracted.ii) Normalization: Iris of different people may be of different size. For the same person, the sizemay vary because of changes in the illumination and other factors. So, normalization isperformed to get all the images in a standard form suitable for the processing.iii) Feature extraction: Iris provides abundant texture information; a feature vector is formedwhich consists of the ordered sequence of features extracted from the various representations ofthe iris images.iv) Matching: Feature vectors are classified through euclidean Distance.B. WaveletsAddison describes a wavelet as a mathematical function used to divide a given function ora continuous-time signal into different frequency components and study each component with aresolution that matches its scale. The wavelets are scaled and are the translated copies (known as“daughter wavelets”) of a finite-length or fast-decaying oscillating waveform (known as the“mother wavelet”). Wavelet transforms have advantages over traditional Fourier transforms forrepresenting functions that have discontinuities and sharp peaks, and for accuratelydeconstructing and reconstructing finite, non-periodic, and/or non-stationary signals. Theseunderlying characteristics make wavelets applicable for creating the feature vector that isnecessary in the iris recognition algorithm.C. Iris Recognition Algorithms and PrinciplesMany algorithms have been developed for the iris recognition system. The waveletfunctions or wavelet analysis is a recent solution for overcoming the shortcomings in imageprocessing, which is crucial for iris recognition. Nabti and Bouridane proposed a novelsegmentation method based on wavelet maxima and a special Gabor filter bank for featureextraction, which obtains an efficient recognition with an accuracy of 99.43%. The steps are asfollows: the multi-scale edge detection method is used for iris image processing, the extraction offeatures from an iris-polarized image using the proposed Gabor filter bank, and matching withHamming distance for identification and recognition. Narote et al. proposed a new algorithm foriris recognition based on the Dual Tree Complex Wavelet Transform. The Dual Tree ComplexWavelet Transform (DTCWT) provides three significant advantages: they have reduced shiftsensitivity with low redundancy, improved directionality, and explicit phase information.Experimental results show that the above algorithm based on DTCWT is nearly 25 times fasterthan that of Narote. Also, the authentication using DTCWT demonstrates that the approach ispromising in terms of improving iris-based identification.III. IRIS RECOGNITION SYSTEMA typical Iris Recognition system basically consists of following main modules as shownbelow,Figure 1. Iris Recognition SystemImageAcquisitionEyeImageImageSegmentationImageNormalizationFeatureExtraction&MatchingIrisFeaturedatabaseRecognition
  4. 4. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME389A. Image AcquisitionIt is to capture a sequence of iris images from the person concerned using aspecifically designed sensor. Since, the iris is fairly small and exhibits abundant featuresunder infrared lighting, capturing iris images of high quality is one of the major challengesfor practical applications. While designing an image acquisition apparatus the factors thatmust be taken into consideration is the lighting system, the positioning systems and physicalcapture system.B. Iris SegmentationThe next stage of iris recognition is to isolate actual iris region in an eye image, theeyelids and eyelashes normally occlude the upper and lower parts of the iris region. Also,specular reflections can occur within the iris region corrupting the iris pattern. A technique isrequired to isolate and exclude these artifacts as well as locating the circular iris regionC. Iris NormalizationThe normalization process will produce iris regions, which have the same constantdimensions, so that two photographs of the same iris under different conditions will havecharacteristic features at the same spatial location. Another point of note is that the pupilregion is not always concentric within the iris region, and is usually slightly nasal. This mustbe taken into account while trying to normalize the ‘doughnut’ shaped iris region to haveconstant dimensions. The rubber sheet model takes into account pupil dilation and sizeinconsistencies in order to produce a normalized representation with constant dimensions.IV.WAVELET BASED ALGORITHMSWavelet transforms are used to extract the feature of normalized iris image, waveletcoefficients vectors are used as a feature for iris recognition, four types of waveletcoefficients e.g. vertical, horizontal, approximate and detail can be used, here simple Harrwavelet is used,Figure 2: (a) Horizontal (b) Vertical (c) Approximate and(d) Detail coefficients of Haar wavelet transform for iris templateWavelet transform has three main disadvantages, Shift sensitivity, Poor directionality andAbsence of phase information, these disadvantages can be overcome by complex wavelet.
  5. 5. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME390A. Complex WaveletsComplex Wavelets Transforms use complex valued filtering that decomposes thereal/complex signals into real and imaginary parts in transform domain. The real andimaginary coefficients are used to compute amplitude and phase information, just the type ofinformation needed to accurately describe the energy localization of oscillating functions.Here complex frequency B-spline wavelet is used for iris feature extraction A complexfrequency B-spline wavelet is defined byFigure 3: Complex Frequency B-Spline wavelet coefficients for iris templateB. Gabor WaveletThe main idea of this method is that: firstly we construct two-dimensional Gaborfilter, and we take it to filter these images, and after we get phase information, code it into2048 bits, i.e. 256 bytes. In image processing, a Gabor filter, named after Dennis Gabor, is alinear filter used for edge detection. Frequency and orientation representations of Gabor filterare similar to those of human visual system, and it has been found to be particularlyappropriate for texture representation and discrimination. In the spatial domain, a 2D Gaborfilter is a Gaussian kernel function modulated by a sinusoidal plane wave. The Gabor filtersare self-similar – all filters can be generated from one mother wavelet by dilation androtation.Its impulse response is defined by a harmonic function multiplied by a Gaussianfunction. Because of the multiplication-convolution property (Convolution theorem), theFourier transform of a Gabor filters impulse response is the convolution of the Fouriertransform of the harmonic function and the Fourier transform of the Gaussian function. Thefilter has a real and an imaginary component representing orthogonal directions. The twocomponents may be formed into a complex number or used individually.Gabor filters are directly related to Gabor wavelets, since they can be designed for anumber of dilations and rotations. However, in general, expansion is not applied for Gaborwavelets, since this requires computation of bi-orthogonal wavelets, which may be very time-consuming. Therefore, usually, a filter bank consisting of Gabor filters with various scalesand rotations is created. The filters are convolved with the signal, resulting in a so-calledGabor space. This process is closely related to processes in the primary visual cortex. Jonesand Palmer showed that the real part of the complex Gabor function is a good fit to thereceptive field weight functions found in simple cells in a cats striate cortex.
  6. 6. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME391The Gabor space is very useful in image processing applications such as opticalcharacter recognition, iris recognition and fingerprint recognition. Relations betweenactivations for a specific spatial location are very distinctive between objects in an image.Furthermore, important activations can be extracted from the Gabor space in order to create asparse object representation.Local regions of an iris are projected onto quadrature 2-D Gabor wavelets using equation (1).Where is a complex-valued bit whose real and imaginary parts are either 1or0 (sign) depending on the sign of the 2-D integral; is the raw iris image in adimensionless polar coordinate system that is size and translation-invariant; α and β are themulti scale 2-D wavelet size parameters, spanning an 8-fold range from 0.15 mm to 1.2 mmon the iris; ω is wavelet frequency, spanning 3 octaves in inverse proportion to β;represents the polar coordinates of each region of iris for which the phasor coordinates h(Re,Im ), like figure 4.Equation (1) generates complex-valued projection coefficients whose real andimaginary parts specify the coordinates of a phasor in the complex plane. The angle of eachphasor is quantized to one of the four quadrants, setting two bits of phase information. Thisprocess is repeated all across the iris with many wavelet sizes, frequencies, and orientationsto extract 2,048 bits, i.e. 256 bytes. Such a phase quadrant coding sequence is illustrated forone iris by the bit stream shown graphically in Figure 4.After feature extraction of figure 1, get figure 6 & 7.Figure 4. Phase-Quadrant Demodulation Code
  7. 7. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME392Figure 5. Normalized Unwrapped IrisFigure 6. Real ComponentFigure 7. Imaginary ComponentC. The Discrete Wavelet TransformComputing wavelet coefficients at every possible scale is a fair amount of work, andit generates an awful lot of data. That is why we choose only a subset of scales and positionsat which to make our calculations. It turns out, rather remarkably, that if we choose scalesand positions based on powers of two so-called dyadic scales and positions then our analysiswill be much more efficient and just as accurate. We obtain such an analysis from the discretewavelet transform (DWT) given by (1).An efficient way to implement this scheme using filters was developed in 1988. Thisalgorithm is in fact a classical scheme known in the signal processing community as a two-channel sub band coder. This very practical filtering algorithm yields a fast wavelet transforma box into which a signal passes, and out of which wavelet coefficients quickly emerge. Let’sexamine this in more depth.Let,Both and can be expressed as linear combinations of double-resolution copies ofthemselves.
  8. 8. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME393Here in (2) in (3) the expansion coefficients are called scaling and wavelet vectors,respectively. They are the filter coefficients of fast wavelet transform (FWT),Approximate coefficientsHorizontal coefficientsVertical coefficientsDiagonal coefficientsHere is the original image whose DWT is to be computed.V. EXPERIMENTATION AND RESULTSA. Complex WaveletsThe iris templates are matched using different angles 210,240,280,320 and 350degrees and it is observed that as angles increases percentage of matching also increases thebetter match is observed at angle 350 which is 93.05%.Further by detecting eyelids andeyelashes the iris image is cropped and iris template is generated for matching purpose theresults obtained is better than previous results the matching score is 95.30%.Figure 8. Graph for angles verses matching percentage of iris imagesB. Gabor WaveletWe use images of eyes from 10 persons, and every person has six images of eyes. Thetop three images are used as test images and the next three images are used for trainingpurpose. We use the Daugman’s methods to iris regions segmentation and use Gabor waveletfor feature extraction. At last, in the identification stage we calculate Hamming distancebetween a test image & a training image. The smallest distance among them is expressed, thattest image belongs to this class. The recognition rate is 96.5%.C. The Discrete Wavelet Transform:The technique developed here uses all the frequency resolution planes of DiscreteWavelet Transform (DWT). These frequency planes provide abundant texture informationpresent in an iris at different resolutions. The accuracy is improved up to 98.98%. Withproposed method FAR and FRR is reduced up to 0.0071% and 1.0439% respectively.
  9. 9. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME394VI. CONCLUSIONIn this paper, we compare the performance of various wavelets for Iris recognitionlike complex wavelet transform, Gabor wavelet, and discrete wavelet transform. Usingcomplex wavelet, different coefficient vectors are calculated. Minimum distance classifierwas used for final matching. The smaller the distance the more the images matched. It isobserved that for the complex wavelets the results obtain are good than the simple waveletbecause in complex wavelet we get both phase and angle also real and imaginary coefficients,so we can compare all these parameters for iris matching purpose.2D Gabor wavelets havethe highest recognition rate. Because iris is rotator, and 2D Gabor wavelets have rotationinvariance, it has the highest recognition rate. But 2D Gabor wavelets have highcomputational complexity, and need more time. Discrete wavelet transform used for irissignature formation gives better and reliable results.REFERENCES[1] Biometrics: Personal Identification in a Networked Society, A. Jain, R. Bolle and S.Pankanti, eds. Kluwer, 1999.[2] D. Zhang, Automated Biometrics: Technologies and Systems. Kluwer, 2000[3] Anil Jain. Introduction to Biometrics. Michigan State University, East Lansing, MI.[4] L. Ma, T, Yunhong Wang, and D. Zhang. Personal identification based on iris textureanalysis. IEEE Transactions on Pattern Analysis and Machine Intelligence, vol.25,no.12, 2003[5] John Daugman. Recognizing persons by their iris patterns. Cambridge University,Cambridge, UK.[6] J. Daugman, “Demodulation by Complex-Valued Wavelets for Stochastic PatternRecognition,” Int’l J. Wavelets, Multiresolution and Information Processing, vol. 1, no.1, pp. 1-17, 2003[7] W. Boles and B. Boashash, “A Human Identification Technique Using Images of theIris and Wavelet Transform,” IEEE Trans. Signal Processing, vol. 46, no. 4, pp. 1185-1188, 1998[8] R. Wildes, J. Asmuth, G. Green, S. Hsu, R. Kolczynski, J. Matey, and S. McBride, “AMachine-Vision System for Iris Recognition,” Machine Vision and Applications, vol. 9,pp. 1-8, 1996[9] S. Lim, K. Lee, O. Byeon, and T. Kim, “Efficient Iris Recognition throughImprovement of Feature Vector and Classifier,” ETRI J., vol. 23, no. 2, pp. 61-70, 2001[10] L. Ma, Y. Wang, and T. Tan, “Iris Recognition Based on Multichannel GaborFiltering,” Proc. Fifth Asian Conf. Computer Vision, vol. I, pp. 279-283, 2002[11] C. Tisse, L. Martin, L. Torres, and M. Robert, “Person Identification Technique UsingHuman Iris Recognition” Proc. Vision Interface, pp. 294-299, 2002[12] J. Daugman, “How Iris Recognition Works,” IEEE Transaction on Circuits and Systemfor Video Technology, vol. 14, no. 1,pp. 21–30, 2004.[13] J. Daugman,” High Confidence Visual Recognition of Persons by a Test of StatisticalIndependence,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol.15, no. 11, pp. 1148–1161, 1993.
  10. 10. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME395[14] L.Ma, T. Tan, Y. Wang and D. Zhang,” Efficient Iris Recognition by CharacterizingKey Local Variations,” IEEE Transactions on Image Processing, vol. 13, no. 6, pp.739–750, 2004.[15] K. Miyazawa, K. Ito, T. Aoki, K. Kobayashi, and H. Nakajima, “An efficient irisrecognition algorithm using phase-based image matching,” Proc. Int. IEEE Conf. onImage Processing, Vol. II, pp. 49–52, Sept. 2005.[16] Panchamkumar D Shukla,”Complex Wavelet Transforms and their applications” anM.Phil Thesis, University of Stratchlyde, Signal Processing Division, 2003.[17] S. Lim, K. Lee, O. Byeon, and T. Kim, “Efficient Iris Recognition throughImprovement of Feature Vector and Classifier”, ETRI Journal, Vol.23, No.2, June,2001.[18] J. Daugman, “High Confidence Visual Recognition of Persons by a Test of StatisticalIndependence”. IEEE Transl. on Pattern Analysis and Machine Intelligence, Vol.15,issue 11, 1993.[19] Makram Nabti and Bouridane, “An effective iris recognition system based on waveletmaxima and Gabor filter bank”, IEEE trans. on iris recognition, 2007.[20] Narote et al. “An iris recognition based on dual tree complex wavelet transform”. IEEEtrans. on iris recognition, 2007.[21] Institute of Automation Chinese Academy of Sciences. Database of CASIA iris image[EB/OL][22] L. Masek, “Recognition of Human Iris Patterns for Biometric Identification”, TheUniversity of Western California, 2003.[23] N. G. Kingsbury, “Image processing with complex wavelets,” Philos.Trans. R. Soc.London A, Math. Phys. Sci, vol. 357, no. 3, pp. 2543–2560, 1999.[24] Vijay M.Mane, GauravV. Chalkikar and Milind E. Rane, “Multiscale Iris RecognitionSystem”, International journal of Electronics and Communication Engineering &Technology (IJECET), Volume 3, Issue 1, 2012, pp. 317 - 324, ISSN Print: 0976- 6464,ISSN Online: 0976 –6472.[25] Darshana Mistry and Asim Banerjee, “Discrete Wavelet Transform using Matlab”,International journal of Computer Engineering & Technology (IJCET), Volume 4,Issue 2, 2012, pp. 252 - 259, ISSN Print: 0976 – 6367, ISSN Online: 0976 – 6375.

×