IMPROVING THE ACCURACY OF FUZZY VAULT SCHEME IN FINGERPRINT BIOMETRIC

Abstract

At present, authentication techniques using fingerprint biometrics have been widely used in various fields. This is because the authentication techniques using biometrics are safer and more comfortable than using traditional passwords. In order to realize this, a technique in the biometric cryptosystem is proposed in the research, called the fuzzy vault scheme. Although the fingerprint data in the form of minutiae can be protected with a fuzzy vault scheme compared to traditional authentication systems, it can reduce user convenience. Previous studies proposed a distance-based method in the fuzzy vault scheme. The distance-based method is proposed because it is no need to align and rotate the fingerprint image during registration or authentication. Then with the distance-based method also does not produce a helper data that can lead to information leakage that can be exploited by impostor. In the research, the distance-based method is proposed with several modifications, which are the minutiae filter and candidate points identification techniques. The previous method produced FRR 13.4375% and FAR 0.4515% and the proposed method produces FRR 8.9475% and FAR 0.3520%.