Fingerprint Image Enhancement: Iterative Fast Fourier Transform Algorithm and Performance Evaluation
Fingerprint Image Enhancement: Iterative Fast Fourier Transform Algorithm and Performance Evaluation
Fingerprint Image Enhancement: Iterative Fast Fourier Transform Algorithm and Performance Evaluation
1. Introduction
Biometric provides automated method to identify a person based on physiological or
behavioural characteristics. The unique features measured are face, fingerprints, hand
geometry, handwriting, iris, retina, vein, and voice. Biometric technologies are playing vital
role to provide highly secure identification and personal verification methods. As the level of
security breaches and transaction fraud increases, the role of highly secure identification is
becoming apparent. Every biometric method uses some aspect of an individual's physical or
behavioural attributes as a means of authenticating the individual's identity. According to the
Chikkerur, today, the most pervasive biometric in use is fingerprint, a physical biometric [4].
Cole had given the concept that fingerprints have a unique pattern of ridges and furrows [6].
Fingerprint readers use this uniqueness to generate a pattern based on areas as line print, lines
merge, fork or loop like the round whorl, etc. One can find that the finger prints rarely use
the full print for identification. This pattern is stored in a database either in a remote
computer or in the device itself. When a person scans a print, this device compares the pattern
generated by the print with one in the database to make a positive identification.
A. K. Jain discussed that there are various phases of fingerprint identification as image
loading, image enhancement, normalization, thinning, minutia marking and minutia
extraction [13]. Fingerprint image quality is a vital issue to measure the performance of
11
2. Previous Work
This section summarizes the literature on fingerprint, palm and iris identification,
elaborated with various normalization and enhancement techniques used for quality
improvement of identification system. Roger Clarke identified need of identification and
various factors required for human identification [5]. According to this paper various
identification measures as names, codes, knowledge-based, token-based and biometrics had
taken into account. Hong emphasizes that Gabor filter has optimal joint resolution in both
spatial and frequency domains [9]. Kansaei used the concept of statistics to calculate the
Dominant Ridge Direction (DRD) for each 16*16 block. After deriving the DRD, they rotate
the block image according to the dominant orientation for image projection to enhance the
local ridge veins [21]. Miao proposed the method of minutiae extraction by tracing ridges on
gray image [23]. Hong used Gabor filters as band pass filters to remove the noise and prevent
true ridge/valley structures because the filter has both frequency selective and orientation
selective properties [9]. They used the fact that a fingerprint is made of white followed by
black lines of bounded number of pixels in time domain. They use the genetic algorithm to
generate black and white lines of different widths, and then translate the enhancement
technique into the optimization problem while getting the best match with the original
fingerprint. Hong introduced the method of Gabor filtering for fingerprint image [10]. N.K.
Ratha presented data hiding concept using wavelet transformation and WSQ compression
techniques including the use of filters, with the assumption of secure image capture device.
They demonstrated the coherence enhancing diffusion filtering method and shown the
implementation of Min Avg. Co-relation Energy (MACE) filters for authentication while
convolution (random no. generation and add it to the training images) provide irrevocability.
Yang-Han Pang proposed the concept of cancelable palmprint authentication system [35].
This work is designed to transform palmprint image into a lower dimensional compact feature
representation. Cancelable verification key is generated with the help of random number
generation. To generate the key wavelet based palmprint features are hashed with a set of
pseudo random data. Eric Kukula dealt with the various factors i.e. technical as well as
environmental (operational), current system characteristics of Hand Geometry software
12
installing in the Purdue University [22]. A performance analysis in terms of False Rejection
Rate (FRR) and False Acceptance Rate (FAR) had shown in order to measure the efficiency
of fingerprint identification system. As a result they got the fact that 93% people liked it, 98%
interested to use it because of easiness and 87% preferred it than existing card-based. They
demonstrated that core and Delta points are global features, represent intrinsic points of
reference that can be used to align two prints and reduce computation complexity of the
matcher. They provided an approach which reduces complexity from O (n3) or O (n2) to O
(n). Where n represent number of minutia. They used Fast Fourier Transformation for image
enhancement. For singular point detection various steps are as image pre processing, complex
filtering, and post-processing. They demonstrated that biometrics is a method of enhancing
the security and privacy of biometric authentication. Instead of using your real finger
transformation used. Transformation means mapping the original finger into some different
parameters and it can be done by various techniques as Cartesian Transformation, Radial
Transformation, and Functional Transformation. Eric Kukula presented various factors i.e.
technical as well as environmental (operational), current system characteristics of hand
geometry software installing in the Purdue University and had done a performance analysis in
terms of FRR and FAR [22]. As a result they got the fact that 93% people are linked it, 98%
interested to use it because of easiness and 87% preferred it than existing card-based. Salil
Prabhakar and fellow researchers reviewed various applications in the area of biometrics [26].
They proposed the technique to generate an irrevocable cryptographic key. This technique
used bifurcation points and ending points for feature extraction. Vince Thomas and fellow
researchers represented iris texture feature for gender identification. Various steps as data
acquisition, segmentation, feature extraction, geometric features, texture features were
included [32]. They had been proposed iris coding method based on Discrete Cosine
Transform (DCT). This work demonstrated the use of novel patch encoding methods in
capturing iris texture information. To model for matching and non- matching probability
statistical analysis used in order to predict worst case equal error rates. In this paper authors
proposed the concept that in case of iris pattern stolen or any other activity which force to
compromise with the security of the iris identification system, cancellable iris biometric has
been used. It can be achieved by the transformation of original iris pattern. They presented
four techniques as gray-combo, bin-combo, gray-salt and bin-salt and proposed an on-line
biometric authentication based on signature and for this purpose non-invertible transformation
has been generated and convolution transformation has been used [8].
Jinyu Zuo and friends described that in case of Iris pattern stolen or any other activity
which force to compromise with the security of the iris identification system, cancelable iris
biometric has been used. It can be achieved by the transformation of original iris pattern and
four techniques as gray-combo, bin-combo and gray salt, bin-salt [39]. They had given the
approach in which hand is divided into different parts as parts related with fingers and
confidence of each finger has analyzed by support vector data description. After the
measurement of confidence of each finger consider the finger which has highest confidence
and if this value is higher than threshold value then raise the confidence of other fingers also.
3. Proposed Algorithm
Proposed fingerprint enhancement algorithm is based on Iterative Fast Fourier
transformation and used to enhance the intensity of pixels of the fingerprint images with
maintaining its frequency and orientation selective properties. Various steps of this algorithm
are shown as follows-
13
14
discussed above if we apply the orientation property of fingerprint the problem can be
alleviated largely.
15
16
FFT Algorithm
Goodness Index(Without
Enhancement)
0.42
0.40
0.30
0.27
0.23
0.21
0.19
0.17
0.15
0.13
0.11
0.10
0.08
0.06
Goodness Index(GI) is used to measure the performance of proposed algorithm. The goodness Index
can be calculated with the following formula given by N. Ratha et.al.:
[
Where p represents paired minutiae. Minutia fa and fb are called as paired minutiae if fa is located in a
tolerance box centered around fb. a depicts the missing minutiae that is not captured by minutiae extraction
algorithm. b represents spurious minutiae that is detected by minutiae extraction algorithm but not come
within the boundary of tolerance box of any minutiae. qi represents quality factor of the ith window and ti
depicts the number of true minutiae in the ith window. GI considers both missing and spurious minutiae
and calculate performance on the basis of ratio of these minutiae with true minutiae. The greater the value
of GI, better the minutiae extraction algorithm. The maximum value of GI may be 1 that means no
missing or spurious minutia and this is an ideal situation. Table 1 show that the value of GI is increasing
after applying the proposed enhancement algorithm. That derives the improved accuracy of algorithm.
17
8. Conclusion
We have designed a fingerprint enhancement algorithm which can increase the degree of clarity of
ridges and valleys. Since high quality fingerprint image acquired by using an adaptive fingerprint image
enhancement method is critical to the quality of any fingerprint identification system, a fingerprint
enhancement method based on iterative Fast Fourier Transformation has been designed in this paper and
comparative analysis with the existing method, i.e., Fast Fourier Transformation has been shown with the
help of graph. The performance of the algorithm is evaluated using the goodness index of minutia
extraction process. Experiments on a public domain fingerprint database (i.e., FVC 2006) demonstrates
that the use of minutia descriptors leads to an order of magnitude reduction in the false accept rate without
significantly affecting the genuine accept rate. Based on the observation of good quality rolled images,
the ridge and valleys intervals of each image are considered in order to select the Region of Interest (ROI)
for effective enhancement. Experimental results show that our algorithm improved the goodness index as
well as matching performance of the FIS. The algorithm also dealt with the broken ridges/false minutia
problem and removes them from further processing. If these issues cannot entertain, these can reduce the
performance of minutia extraction process.
References
[1] G. B. Amayeh, Improving Hand Based Verification Through Online Finger Template Update Based on Fused
Confidences, IEEE Transactions on Pattern Recognition, (2009), pp. 978-984.
[2] T. F. Ayers, Modeling ngerprints:Components for thetask of synthesis, Proceeding of Introduction to Modeling and
Simulationon Biometric Technologies, Calagary, (2004), pp. 155-162.
[3] R. C. Bolle, Biometrics perils and patches, Transaction of Pattern Recognition, (2002), pp. 2727-2738.
[4] S. G. Chikkerur, Minutia Verification in Fingerprint Images Using Steerable Wedge Filters, IEEE Transactions on Pattern
Recognition and Machine Intelligence, (2005), pp. 45-51.
[5] R. Clarke, Human Identification in Information Systems: Management Challenges and Public Policy Issues, Information
Technology & People, (1994), pp. 6-37.
[6] S. A. Cole, R. McNally and K. Jordan, Truth Machine: the Contentious History of DNA Fingerprinting, Chicago: The
University of Chicago Press, (1995).
[7] A. M. Gerez, Systematic methods for the computation of the directional fields and singular points of fingerprints, IEEE
Transactions on Pattern Analysis and Machine Intelligence, (2002), pp. 905-919.
[8] G. J. Guo, Iris Extraction Based on Intensity Gradient and Texture Difference, IEEE workshop on Applications of
Computer Vision, (2008), pp. 211-216.
[9] L. W. Hong, Fingerprint Image Enhancement: Algorithm and Performance Evaluation, Transactions on Pattern Analysis
and Machine Intelligence, (1998), pp. 777-789.
[10] L. Hong, Y. Wan and A. K. Jain, Fingerprint Image Enhancement: Algorithm and Performance Evaluation, Transactions
on Pattern Analysis and Machine Intelligence, vol. 20, no. 8, (1998), pp. 777-789.
[11] S. G. Chikkerur, Minutia Verification in Fingerprint Images Using Steerable Wedge Filters, IEEE Transactions on Pattern
Recognition and Machine Intelligence, (2005), pp. 45-51.
[12] L. W. Hong, Fingerprint image enhancement:Algorithm and performance evaluation, IEEE Transaction of Pattern
Analysis and Machine Intelligence, (2000), pp. 670-781.
[13] A. H. Jain, Biometrics:Promisingfrontiers for emerging identication market, ACM Communications, (2000), pp. 91-98.
[14] A. H. Jain, On-line Fingerprint Verification, IEEE Transaction on Pattern Analysis and Machine Intelligence, (2002), pp.
302-314.
[15] A. K. Jain, An Identity Authentication System using Fingerprints, Proceedings of the IEEE, (1997), pp. 1365-1388.
[16] A. K. Jain, S. Prabhakar, L. Hong and S. Pankanti, Filterbank-based fingerprint matching, IEEE Transaction on Pattern
Recognition and Machine Intelligence, (2000), pp. 846-859.
[17] A. N. Jain, Score Normalization in Multimodal Biometric Systems, The Journal of Pattern Recognition, (2005), pp. 22702285.
[18] A. P. Jain, Biometrics: A Grand Challenge, Proceedings of International Conference on Pattern Recognition, (2004), pp.
935-942.
[19] A. P. Jain, FingerCode: A lterbank for ngerprint representation and matching, Proceeding of IEEE Conference CVPR,
(1999), pp. 187-193.
[20] A. R. Jain, A deformable model for ngerprint matching, Transaction of Pattern Recognition, (2005), pp. 95-103.
[21] M. B. Kansaei, Fingerprint Feature Enhancement Using Block-Direction on Reconstructed Images, International
Conference on Information, Communications and Signal Processing, (1997), pp. 721-725.
18
[22] E. E. Kukula, Implementation of Hend Geometry at Purdue University's Recreational Center: An Analysis of User
Perspectives and System Performance, (2005).
[23] D. M. Maio, FVC2002: Fingerprint verication competition, IEEE Pattern Analysis Machine Intelligence, (2008), pp.
402-412.
[24] D. M. Monro, DCT Based Iris Recognition, IEEE Transactions on Pattern Analysis and Machine Intelligence, (2007), pp.
586-593.
[25] Q. J. Peng Wang, Modeling and Predicting Face Recognition System Performance Based on analysis of Similarity Scores,
IEEE Transactions on Pattern Analysis and Machine Intelligence, (2007), pp. 665-670.
[26] S. J. Prabhakar, Learning Fingerprint Minutiae Location and Type, Transactions of Pattern Recognition, (2003), pp. 18471857.
[27] N. C. Ratha, Cancelable Biometrics, IEEE Conference on Pattern Recognition, (2006), pp. 48-51.
[28] N. Ratha, S. Chen and A. K. Jain, Adaptive Flow Orientation Based Feature Extraction in Fingerprint Images, Pattern
Recognition, vol. 28, no.11, (1995), pp. 1657-1672.
[29] M. V. Savvides, Cancelable Biometric Filters for Face Recognition, Proceedings of the IEEE conference on Pattern
Recognition, (2004), pp. 51-54.
[30] C. R. Sharat, Impact of Singular Point Detection on Fingerprint Matching Performance, IEEE Transactions on Pattern
Recognition and Machine Intelligence, (2006), pp. 64-70.
[31] K. H. Takahashi, Generating Provably Secure Cancelable fingerprint Templates based on Correlation-invariant Random
Filtering, IEEE Transactions on Pattern Recognition, (2009), pp. 978-984.
[32] V. C. Thomas, Learning to predict gender from iris images, IEEE Transactions on Pattern Analysis and Machine
Intelligence, (2007), pp. 411-418.
[33] J. S. Thorton, A Bayesian Approach to Deformed Pattern Matching of Iris Images, IEEE Transactions on Pattern Analysis
and Machine Intelligence, (2007), pp. 596-606.
[34] S. W. Wang, Fingerprint enhancement in the singular point area, IEEE Signal Processing, (2004), pp. 16-19.
[35] H. J. Yang, Generating Secure Cancelable Fingerprint Templates using Local and Global Features, IEEE Transactions on
Pattern Recognition, (2009).
[36] J. L. Yang, A modied Gabor lter design Method for ngerprint image enhancement, Transaction of Pattern Recognition,
(2003), pp. 1805-1817.
[37] A. S. Yanushkevich, Simulation of biometric information:The new generation of biometric system, Proceeding of Int.
Workshop on Modeling and Simulation on Biometric Technologies, (2004), pp. 87-98.
[38] J. G. Zhou, A Model-Based Method for the Computation of Fingerprints Orientation Field, IEEE Transactions on Image
Processing, (2004), pp. 821-835.
[39] J. R. Zuo, Cancelable Iris Biometric, IEEE Conference on Pattern Recognition, (2008), pp. 244-248.
Authors
Dr. Ela Kumar obtained her B.E. in Electronics and communication from IIT
Roorkee in 1988 and M.Tech in Computer Science and Technology in 1990 from IIT
Roorkee. She obtained her Ph.D. degree in the area of Natural Language Processing
from Delhi University in 2003. She is actively engaged in Technical teaching and
research since last 23 years. She has published almost 58 research papers in
international journals, national journals, international conferences and national
conferences. She has authored four books and many course materials related to
computer science. She has supervised four doctoral research candidates. She is on the
reviewer panel of many international journals. She has been conferred Rashtriya
Gaurav Award and her bibliography was included in Asian Admirable achiever 2011.
Presently she is working as Professor and Dean, School of Information and
Communication Technology, Gautam Buddha University, Greater Noida, UP, India.
19
Er. Sandhya Tarar obtained her B.Tech degree from UPTU, Lucknow in
Computer Science and Engineering in 2005 and M.Tech from Aligarh Muslim
University, Aligarh in Computer Science and Engineering. She is pursuing Ph.D.
from Guatam Buddha University, Greater Noida since 2010. Her area of research is
Biometrics and Pattern Recognition. She has six years of teaching and research
experience in reputed institutions. She has 13 publications in various International
journals, International conferences and national conferences. She is on the review
panel of many International Conferences and Journals. She is the member of different
professional organizations as International Association of Computer Science and
Information Technology, International Association of Engineering, Computer society
of India and Indian Society of Technical Education. Currently working as
20