Feature extraction for a Slepian-Wolf biometric system using LDPC codes

Abstract

We present an information-theoretically secure biometric storage system using graph-based error correcting codes in a Slepian-Wolf coding framework. Our architecture is motivated by the noisy nature of personal biometrics and the requirement to provide security without storing the true biometric at the device. The principal difficulty is that real biometric signals, such as fingerprints, do not obey the i.i.d. or ergodic statistics that are required for the underlying typicality properties in the Slepian-Wolf coding framework. To meet this challenge, we propose to transform the biometric data into binary feature vectors that are i.i.d. Bernoulli (0.5), independent across different users, and related within the same user through a BSC-p channel with small p less-than 0.5. Since this is a standard channel model for LDPC codes, the feature vectors are now suitable for LDPC syndrome coding. The syndromes serve as secure biometrics for access control. Experiments on a fingerprint database demonstrate that the system is information-theoretically secure, and achieves very low false accept rates and low reject rates. IEEE International Symposium on Information Theory, 2008This work may not be copied or reproduced in whole or in part for any commercial purpose. Permission to copy in whole or in part without payment of fee is granted for nonprofit educational and research purposes provided that all such whole or partial copies include the following: a notice that such copying is by permission of Mitsubishi Electric Research Laboratories, Inc.; an acknowledgment of the authors and individual contributions to the work; and all applicable portions of the copyright notice. Copying, reproduction, or republishing for any other purpose shall require a license with payment of fee to Mitsubishi Electric Research Laboratories, Inc. All rights reserved. Abstract-We present an information-theoretically secure biometric storage system using graph-based error correcting codes in a Slepian-Wolf coding framework. Our architecture is motivated by the noisy nature of personal biometrics and the requirement to provide security without storing the true biometric at the device. The principal difficulty is that real biometric signals, such as fingerprints, do not obey the i.i.d. or ergodic statistics that are required for the underlying typicality properties in the SlepianWolf coding framework. To meet this challenge, we propose to transform the biometric data into binary feature vectors that are i.i.d. Bernoulli(0.5), independent across different users, and related within the same user through a BSC-p channel with small p < 0.5. Since this is a standard channel model for LDPC codes, the feature vectors are now suitable for LDPC syndrome coding. The syndromes serve as secure biometrics for access control. Experiments on a fingerprint database demonstrate that the system is information-theoretically secure, and achieves very low false accept rates and low false reject rates.

Privacy and security of features extracted from minutiae aggregates

“…As a first step in eliminating inferior features, those cuboids that have large overlap with other cuboids are discarded, because these would generate highly correlated features. The following procedure is used for this as in [1]: (1) Compute the relative overlap, i.e., the ratio of volume of intersection to the volume of union, for all pairs of cuboids. (2) Select the pair having highest relative overlap.…”

“…Note that, since m ≫ n, this system of equations is heavily underdetermined. The above problem can be transformed to a linear system of inequalities by replacing all the wi's by τ δ (see (1)) and then solving 2 for a integer-valued solution for ri. If there are unsatisfied constraints, some of those can be satisfied by appropriately selecting the corresponding wi.…”

“…Our overall strategy is based on extracting a binary vector from a minutiae map, applying an error correcting code to this vector and storing its syndrome as a secure biometric [1]. During authentication, the system accepts a probe biometric and attempts to recover the original biometric with the help of the syndrome.…”

“…Naturally, security and privacy depends on the way in which the features are extracted and secured and the type of side information available to the attacker. In, [1], binary features were extracted from the number of minutiae present in randomly chosen cuboidal patches in the (X, Y , Θ) space occupied by the minutia. The present work differs from [1] in three fundamental ways which will be detailed in the sequel: First, each cuboid now generates a richer feature set from which a larger number of bits can be extracted and those with the highest discriminability are used for matching.…”

A Novel Biometric Cryptosystem using LDPC and SHA based Iris Recognition

“…To illustrate how LDPC codes properties influence the performance of the system, we compared with RS codes in this paper, which are two of the most commonly used ECCs in iris systems with secure template storage. Sutcu et al [4] and Nagar et al [5] developed secure biometric systems based on LDPC codes and fingerprints. Wuet al [6] proposed a biometric cryptosystem using the iris and RS codes.…”