Visible to the public "Using Honey Templates to Augment Hash Based Biometric Template Protection"Conflict Detection Enabled

Title"Using Honey Templates to Augment Hash Based Biometric Template Protection"
Publication TypeConference Paper
Year of Publication2015
AuthorsB. Yang, E. Martiri
Conference Name2015 IEEE 39th Annual Computer Software and Applications Conference
Date PublishedJuly
ISBN Number978-1-4673-6564-2
Accession Number15476068
Keywordsauthentication, biometric template database leakage detectability, biometric template protection, biometric template storage, biometric template verification, biometrics (access control), cryptographic hash calculation, cryptography, data privacy, data protection, database leakage, Databases, fuzzy commitment, fuzzy vault, hash, hash based biometric template protection, hash based BTPS, hashed password cracking, honey template, honey templates, honeyword, honeywords, learning (artificial intelligence), machine learning based protected template generation protocol, privacy leakage concern, pubcrawl, pubcrawl170105, secure sketch, Servers, Sugar

Hash based biometric template protection schemes (BTPS), such as fuzzy commitment, fuzzy vault, and secure sketch, address the privacy leakage concern on the plain biometric template storage in a database through using cryptographic hash calculation for template verification. However, cryptographic hashes have only computational security whose being cracked shall leak the biometric feature in these BTPS; and furthermore, existing BTPS are rarely able to detect during a verification process whether a probe template has been leaked from the database or not (i.e., being used by an imposter or a genuine user). In this paper we tailor the "honeywords" idea, which was proposed to detect the hashed password cracking, to enable the detectability of biometric template database leakage. However, unlike passwords, biometric features encoded in a template cannot be renewed after being cracked and thus not straightforwardly able to be protected by the honeyword idea. To enable the honeyword idea on biometrics, diversifiability (and thus renewability) is required on the biometric features. We propose to use BTPS for his purpose in this paper and present a machine learning based protected template generation protocol to ensure the best anonymity of the generated sugar template (from a user's genuine biometric feature) among other honey ones (from synthesized biometric features).

Citation Key7273375