Enhancing biometric key of finger vein based on fuzzy vault

• Main Author: Wan Nourul Akmal binti Ab Aziz
• Format: Thesis
• Language: English
• Published: Universiti Sains Islam Malaysia 2015
• Subjects: Fuzzy Vault scheme; Biometrics; Authentication; Fuzzy vault biometric system
• Online Access: http://ddms.usim.edu.my/handle/123456789/8889

Fuzzy Vault scheme is very suitable for biometric based key generation where biometrics and authentication are bound together. Although many previous researches applied fingerprint, signature, retina, and iris to fuzzy vault biometric system, it has been identified that the existing biometric used has weakness and lack in security, as False Rejection Rate, FRR achieved still has the probability to reject genuine biometric query even though in a case by Uludag, et al. (2005) they obtained the results with 0.0% of False Acceptance Rate, FAR. Finger vein biometric has been used as it proposes a high accuracy rate and privacy if compared with other biometrics since they are hidden inside the body results in difficulty of forgery or theft. In this thesis, three contributions has been introduced to overcome the issue of FRR and FAR rates. The first contribution is the finger vein minutiae points has been extracted with two combination methods which are the Maximum Curvature Points in Image Profile and FingerPrint Application methods. The second contribution is an approach of biometric fuzzy vault by testing it using biometric finger vein. The third contribution is the design of polynomial hardening which has been added to the existing Fuzzy Vault in order to add some complexity to the vault. Based on the tests and analysis, it has been proved that by using finger vein for biometric fuzzy vault with some modifications, the error rates and security level of biometric fuzzy vault have been improved whereby FRR has been minimized to 0.005 by using finger vein as well as maintaining the 0.0% of FAR.