Secure sketch for biometric templates

Qiming Li, Yagiz Sutcu, Nasir Memon

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

There have been active discussions on how to derive a consistent cryptographic key from noisy data such as biometric templates, with the help of some extra information called a sketch. It is desirable that the sketch reveals little information about the biometric templates even in the worst case (i.e., the entropy loss should be low). The main difficulty is that many biometric templates are represented as points in continuous domains with unknown distributions, whereas known results either work only in discrete domains, or lack rigorous analysis on the entropy loss. A general approach to handle points in continuous domains is to quantize (discretize) the points and apply a known sketch scheme in the discrete domain. However, it can be difficult to analyze the entropy loss due to quantization and to find the "optimal" quantizer. In this paper, instead of trying to solve these problems directly, we propose to examine the relative entropy loss of any given scheme, which bounds the number of additional bits we could have extracted if we used the optimal parameters. We give a general scheme and show that the relative entropy loss due to suboptimal discretization is at most (nlog3), where n is the number of points, and the bound is tight. We further illustrate how our scheme can be applied to real biometric data by giving a concrete scheme for face biometrics.

Original languageEnglish (US)
Title of host publicationAdvances in Cryptology - ASIACRYPT 2006 - 12th International Conference on the Theory and Application of Cryptology and Information Security, Proceedings
Pages99-113
Number of pages15
Volume4284 LNCS
DOIs
StatePublished - 2006
Event12th International Conference on the Theory and Application of Cryptology and Information Security, ASIACRYPT 2006 - Shanghai, China
Duration: Dec 3 2006Dec 7 2006

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4284 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other12th International Conference on the Theory and Application of Cryptology and Information Security, ASIACRYPT 2006
CountryChina
CityShanghai
Period12/3/0612/7/06

Fingerprint

Entropy Loss
Biometrics
Template
Entropy
Relative Entropy
Noisy Data
Optimal Parameter
Quantization
Discretization
Face
Concretes
Unknown

Keywords

  • Biometric template
  • Continuous domain
  • Secure sketch

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Li, Q., Sutcu, Y., & Memon, N. (2006). Secure sketch for biometric templates. In Advances in Cryptology - ASIACRYPT 2006 - 12th International Conference on the Theory and Application of Cryptology and Information Security, Proceedings (Vol. 4284 LNCS, pp. 99-113). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4284 LNCS). https://doi.org/10.1007/11935230_7

Secure sketch for biometric templates. / Li, Qiming; Sutcu, Yagiz; Memon, Nasir.

Advances in Cryptology - ASIACRYPT 2006 - 12th International Conference on the Theory and Application of Cryptology and Information Security, Proceedings. Vol. 4284 LNCS 2006. p. 99-113 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4284 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Li, Q, Sutcu, Y & Memon, N 2006, Secure sketch for biometric templates. in Advances in Cryptology - ASIACRYPT 2006 - 12th International Conference on the Theory and Application of Cryptology and Information Security, Proceedings. vol. 4284 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4284 LNCS, pp. 99-113, 12th International Conference on the Theory and Application of Cryptology and Information Security, ASIACRYPT 2006, Shanghai, China, 12/3/06. https://doi.org/10.1007/11935230_7
Li Q, Sutcu Y, Memon N. Secure sketch for biometric templates. In Advances in Cryptology - ASIACRYPT 2006 - 12th International Conference on the Theory and Application of Cryptology and Information Security, Proceedings. Vol. 4284 LNCS. 2006. p. 99-113. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/11935230_7
Li, Qiming ; Sutcu, Yagiz ; Memon, Nasir. / Secure sketch for biometric templates. Advances in Cryptology - ASIACRYPT 2006 - 12th International Conference on the Theory and Application of Cryptology and Information Security, Proceedings. Vol. 4284 LNCS 2006. pp. 99-113 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{cf995160717a4a8a98a1ee014af240a6,
title = "Secure sketch for biometric templates",
abstract = "There have been active discussions on how to derive a consistent cryptographic key from noisy data such as biometric templates, with the help of some extra information called a sketch. It is desirable that the sketch reveals little information about the biometric templates even in the worst case (i.e., the entropy loss should be low). The main difficulty is that many biometric templates are represented as points in continuous domains with unknown distributions, whereas known results either work only in discrete domains, or lack rigorous analysis on the entropy loss. A general approach to handle points in continuous domains is to quantize (discretize) the points and apply a known sketch scheme in the discrete domain. However, it can be difficult to analyze the entropy loss due to quantization and to find the {"}optimal{"} quantizer. In this paper, instead of trying to solve these problems directly, we propose to examine the relative entropy loss of any given scheme, which bounds the number of additional bits we could have extracted if we used the optimal parameters. We give a general scheme and show that the relative entropy loss due to suboptimal discretization is at most (nlog3), where n is the number of points, and the bound is tight. We further illustrate how our scheme can be applied to real biometric data by giving a concrete scheme for face biometrics.",
keywords = "Biometric template, Continuous domain, Secure sketch",
author = "Qiming Li and Yagiz Sutcu and Nasir Memon",
year = "2006",
doi = "10.1007/11935230_7",
language = "English (US)",
isbn = "3540494758",
volume = "4284 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "99--113",
booktitle = "Advances in Cryptology - ASIACRYPT 2006 - 12th International Conference on the Theory and Application of Cryptology and Information Security, Proceedings",

}

TY - GEN

T1 - Secure sketch for biometric templates

AU - Li, Qiming

AU - Sutcu, Yagiz

AU - Memon, Nasir

PY - 2006

Y1 - 2006

N2 - There have been active discussions on how to derive a consistent cryptographic key from noisy data such as biometric templates, with the help of some extra information called a sketch. It is desirable that the sketch reveals little information about the biometric templates even in the worst case (i.e., the entropy loss should be low). The main difficulty is that many biometric templates are represented as points in continuous domains with unknown distributions, whereas known results either work only in discrete domains, or lack rigorous analysis on the entropy loss. A general approach to handle points in continuous domains is to quantize (discretize) the points and apply a known sketch scheme in the discrete domain. However, it can be difficult to analyze the entropy loss due to quantization and to find the "optimal" quantizer. In this paper, instead of trying to solve these problems directly, we propose to examine the relative entropy loss of any given scheme, which bounds the number of additional bits we could have extracted if we used the optimal parameters. We give a general scheme and show that the relative entropy loss due to suboptimal discretization is at most (nlog3), where n is the number of points, and the bound is tight. We further illustrate how our scheme can be applied to real biometric data by giving a concrete scheme for face biometrics.

AB - There have been active discussions on how to derive a consistent cryptographic key from noisy data such as biometric templates, with the help of some extra information called a sketch. It is desirable that the sketch reveals little information about the biometric templates even in the worst case (i.e., the entropy loss should be low). The main difficulty is that many biometric templates are represented as points in continuous domains with unknown distributions, whereas known results either work only in discrete domains, or lack rigorous analysis on the entropy loss. A general approach to handle points in continuous domains is to quantize (discretize) the points and apply a known sketch scheme in the discrete domain. However, it can be difficult to analyze the entropy loss due to quantization and to find the "optimal" quantizer. In this paper, instead of trying to solve these problems directly, we propose to examine the relative entropy loss of any given scheme, which bounds the number of additional bits we could have extracted if we used the optimal parameters. We give a general scheme and show that the relative entropy loss due to suboptimal discretization is at most (nlog3), where n is the number of points, and the bound is tight. We further illustrate how our scheme can be applied to real biometric data by giving a concrete scheme for face biometrics.

KW - Biometric template

KW - Continuous domain

KW - Secure sketch

UR - http://www.scopus.com/inward/record.url?scp=77649250983&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77649250983&partnerID=8YFLogxK

U2 - 10.1007/11935230_7

DO - 10.1007/11935230_7

M3 - Conference contribution

SN - 3540494758

SN - 9783540494751

VL - 4284 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 99

EP - 113

BT - Advances in Cryptology - ASIACRYPT 2006 - 12th International Conference on the Theory and Application of Cryptology and Information Security, Proceedings

ER -