On fuzzy syndrome hashing with LDPC coding

TL;DR

This paper analyzes fuzzy syndrome hashing with LDPC codes, highlighting its security benefits and designing suitable codes for biometric authentication that tolerate data variations.

Contribution

It provides a detailed analysis of fuzzy syndrome hashing's security and introduces LDPC code design tailored for this application.

Findings

Offers strong protection against information leakage

Demonstrates advantages over fuzzy commitment schemes

Provides specific LDPC code design guidelines

Abstract

The last decades have seen a growing interest in hash functions that allow some sort of tolerance, e.g. for the purpose of biometric authentication. Among these, the syndrome fuzzy hashing construction allows to securely store biometric data and to perform user authentication without the need of sharing any secret key. This paper analyzes this model, showing that it offers a suitable protection against information leakage and several advantages with respect to similar solutions, such as the fuzzy commitment scheme. Furthermore, the design and characterization of LDPC codes to be used for this purpose is addressed.