Improving accuracy of decoding process with pore-based fingerprint fuzzy vault in biometric cryptosystem

Abstract

Biometric cryptosystem is a template protection used to secure the cryptographic key using the biometric features. However, the major issues in biometric is irreplaceable, which the compromise biometric template database poses a serious threat to the biometric-cryptosystem. Fuzzy vault is used to protect the biometric templates and secret key simultaneously by binding them within a cryptographic framework. Current research of fingerprint fuzzy vault used fingerprint minutiae point as fingerprint template. Minutiae points pose some problem such as elastic distortion. This problem affects the accuracy of decoding process in fingerprint fuzzy vault. This project proposes to improve the decoding process using pore-based fingerprint fuzzy vault in biometric cryptosystem. The preliminary stage of this project identifies the classification of fingerprint, method of extracting pores and previous research on fingerprint fuzzy vault scheme. Then, pores extraction method and fingerprint fuzzy vault are been analyzed. Later, the method of improving decoding process using pore-based fingerprint fuzzy vault is designed. The results show that Genuine Acceptance Rate (GAR) is 0.8 and False Acceptance Rate (FAR) is 0, which achieved good accuracy. The experiments show that a Level 3 feature, which is pores, carry significant discriminatory information.