Feature Distribution of the Fingerprint Template Generated by the Geometric Hashing-Based Fuzzy Vault

Recently, a cryptographic construct, called fuzzy vault, has been proposed for crypto-biometric systems, and some implementations for fingerprint have been reported to protect the stored fingerprint template by hiding the fingerprint features, i.e., adding "chaff" minutiae into the original minutiae set. Also, solutions to the most challenging issue in applying the fuzzy vault to fingerprint(i.e., "align" the input fingerprint with the stored fingerprint template in the domain of the fuzzy vault) have been proposed. One of the solutions exploits the idea of the geometric hashingto solve the auto-alignment problem. In this paper, we analyze the feature distribution of the generated hash table as a stored fingerprint template and improve the security of the geometric hashing-based fuzzy vault by uniformly distributing the features in the stored fingerprint template. The goals of this addition make the feature distribution of the hash table for the chaff minutiae similar to that of the real minutiae as well as the feature distribution of the fingerprint image for both types of minutiae uniform. Based on the experimental results, we confirm that the proposed approach can perform the fingerprint verification more securely without a significant degradation of the verification accuracy.