(IJCSIS) International Journal of Computer Science and Information Security,Vol. 10, No. 3, March 2012
A critical issue in biometric systems is to protectthe template of a user which is typically stored in a database ora smart card. The fuzzy vault construct is a biometriccryptosystem that secures both the secret key and thebiometric template by binding them within a cryptographicframework. In , fuzzy vault scheme has been proposedbased on fingerprint minutiae. Since the fuzzy vault storesonly a transformed version of the template, aligning the queryfingerprint with the template is a challenging task. We extracthigh curvature points derived from the fingerprint orientationfield and use them as helper data to align the template andquery minutiae. The helper data itself do not leak anyinformation about the minutiae template, yet contain sufficientinformation to align the template and query fingerprintsaccurately. Further, we apply a minutiae matcher duringdecoding to account for nonlinear distortion and this leads tosignificant improvement in the genuine accept rate. Theperformance improvement can be achieved by using multiplefingerprint impressions during enrollment and verification.Because of the stability and uniqueness, fingerprint iswidely used in biometric identification. The matching methodis one of the most crucial technologies in the AutomatedFingerprint Identification System (AFIS). Whether twofingerprints are matched relies on the similarity measurebetween the effective features of them. There are mainly twokinds of features used in fingerprint matching: local featuresand global features. Two most prominent local ridgecharacteristics, called minutiae, are ridge ending and ridgebifurcation. Minutiae are the most widely used features in thematching process.The performance of Automated FingerprintIdentification System (AFIS) is highly defined by thesimilarity of effective features in fingerprints. Minutia is oneof the most widely used local features in fingerprint matching.In , proposes two global statistical features of fingerprintimage, including the mean ridge width and the normalizedquality estimation of the whole image, and proposed a novelfingerprint matching algorithm based on minutiae setscombined with the global statistical features. The algorithmproposed in this paper has the advantage of both local andglobal features in fingerprint matching. It can improve theaccuracy of similarity measure without increasing of time andmemory consuming.The non–linear distortion in the fingerprint imagesmakes it very difficult to handle matching as it changes thegeometrical position of the minutiae points. The regions, thatare affected, shift the geometry of the minutiae and hence posea potential threat to acceptance of a genuine match. Thedistortion is due to the pressure applied on the scanner, thestatic friction, the skin moisture, elasticity, and rotationaleffects, which occur during the acquisition. The level of distortion increases from the center towards the outer regions.The existing approaches for fingerprint matching are:minutiae–based, and correlation-based. The former has severaladvantages over the latter such as lower time complexity,better space complexity, less requirement of hardware etc.The uniqueness of a fingerprint is due to uniquepattern shown by the locations of the minutiae points–irregularities of a fingerprint–ridge endings, and bifurcations.A novel minutiae-based approach , has been proposed tomatch fingerprint images using similar structures. Distortionposes serious threats through altered geometry, increases falseminutiae, and hence makes it very difficult to find a perfectmatch. This algorithm divides fingerprint images into twoconcentric circular regions – inner and outer – based on thedegree of distortion. The algorithm assigns weight ages for aminutiae–pair match based on the region in which the pairexists.
The algorithm has two stages. In the first stage, theminutiae points are extracted, and in the second stage, thealigning and the matching of the fingerprint images are done.The algorithm is designed to reduce time taken in aligning,immediately after the calculation of the binary image.Recent advances in automated fingerprintidentification technology, coupled with the growing need forreliable person identification have resulted in an increased useof fingerprints in both government and civilian applicationssuch as border control, employment background checks, andsecure facility access. In , Quadratic differentials naturallydefine analytic orientation fields on planar surfaces. Thismethod proposed model orientation fields of fingerprints byspecifying quadratic differentials which is used for reliableperson identification. Models for all fingerprint classes such asarches, loops and whorls are laid out. These models areparameterized by few, geometrically interpretable parameterswhich are invariant under Euclidean motions. Potentialapplications of these models are the use of their parameters asindices of large fingerprint databases, as well as the definitionof intrinsic coordinates for single fingerprint images. Theaccuracy of models is still challenging task for arches.General characteristics of the fingerprint emerge asthe skin on the fingertip begins to differentiate. Fingerprintrecognition systems have the advantages of both ease of useand low cost.
Because among various biometric identifiers,such as face, signature, and voice, the fingerprint has one of the highest levels of distinctiveness and performance and it isthe most commonly used biometric modality. Haiyun Xu et.al., , proposed a novel method to represent minutiae set as afixed-length feature vector, which is invariant to translation,and in which rotation and scaling become translations, so thatthey can be easily compensated for recognition. Thesecharacteristics enable the combination of fingerprintrecognition systems with template protection schemes thatrequire a fixed-length feature vector. This method introducesthe concept of algorithms for two representation methods: thelocation-based spectral minutiae representation and theorientation-based spectral minutiae representation. Bothalgorithms are evaluated using two correlation-based spectralminutiae matching algorithms. The performance can beimproved by using a fusion scheme and singular points
Thespectral minutiae representation overcomes the drawbacks of the minutiae sets, thus broadening the application of minutiae-based algorithms. The minutiae extractor is not reliable itaffects the efficiency of spectral minutiae representation.