| Abstrakt: |
A biometric template is a digital reference of distinct characteristics that are extracted from a biometric sample. The feature set obtained during enrollment is stored in the system database as a template . Disparate passwords, when biometric templates are compromised, it is not feasible for a valid user to rescind his biometric identifiers and change to another set of uncompromised identifiers. Due to this irrevocable nature of biometric data, attacks against the stored templates constitute a major security and privacy peril in a biometric system. The primary aim of this paper is to propose a novel technique for security of templates by using fuzzy vault implementation. The first proposed approach is used to extract the feature point from finger knuckle print (FKP). This approach uses an image processing techniques to extract the biometric feature from the user's finger knuckle print. The second approach is one of unimodal fuzzy vault implementation. Fuzzy vault is a cryptographic framework that binds the biometric template with a uniformly random key to build a secure sketch of the template. Only the secure sketch is stored and if the original template is uniformly random, it is not feasible to retrieve either the template or the key without any knowledge of the user's biometric data. The performances of the proposed approaches are evaluated using the experimental observation. The simplicity and efficiency of the proposed methods make it readily applicable in various security fields. The performance of the proposed fuzzy vault implementation has been evaluated for Hong Kong polytechnic university FKP database. In this fuzzy vault implementation for biometric evaluation the performance metrics used are Failure-to-Capture Rate (FTCR), Genuine Accept Rate (GAR) and False Accept Rate (FAR) Performance analysis is carried out in two phases. In the initial phase, the secret key size is varied and the results are obtained. In phase two, the effect of adding salt and pepper noise is considered. Noise is added to analyze the strength of the fuzzy vault implementation. [ABSTRACT FROM AUTHOR] |
