Memory-Efficient Fuzzy Fingerprint Vault based on the Geometric Hashing

Abstract

One of the solutions to the auto-alignment problem in the fuzzy fingerprint vault exploited the idea of the geometric hashing technique. Although this solution can provide higher verification accuracy, it requires more memory space due to the large size of the hash table. In this paper, we propose an approach to reduce the size of the hash table by using the time-memory tradeoff without sacrificing the verification accuracy. That is, instead of generating the full hash table at the enrollment phase, our approach generates the enrollment hash table "on-the-fly" at the verification phase. The size of the hash table can be reduced further by selecting the basis set carefully. Based on the experimental results, we confirm that the proposed approach can reduce both the static and the dynamic memory requirements without sacrificing both the verification accuracy and the security level.