Enhanced fingerprint fuzzy vault based on distortion invariant minutiae structures

Abstract

The fuzzy vault is a promising cryptosystem for the protection of fingerprint templates. However it requires the alignment of the query template to the enrolled one. Most existing implementations either rely on the extraction of the core which is not a guaranteed task, or use publicly available helper data which may leak information about the protected minutiae. We propose a new alignment approach for the fingerprint fuzzy vault based on distortion invariant minutiae features. Experimental results show improvements in genuine and false accept rates over previous implementations.