“3D Face”: Biometric Template Protection for 3D Face Recognition

Abstract

In this paper we apply template protection to an authentication system based on 3D face data in order to protect the privacy of its users. We use the template protection system based on the helper data system (HDS). The experimental results performed on the FRGC v2.0 database demonstrate that the performance of the protected system is of the same order as the performance of the unprotected system. The protected system has a performance of a FAR ≈ 0.19% and a FRR ≈ 16% with a security level of 35 bits.

On the Performance of Helper Data Template Protection Schemes

“…A more detailed description of the HDS is provided in Section 2.3.1. An abstract overview of the HDS scheme as used in [33][34][35] is portrayed in Figure 1.5 and consists of two main parts: (i) Bit Extraction and (ii) Bit Protection part. In the enrolment phase, first a real-valued feature vector is extracted from each acquired enrolment sample by the Feature Extraction module.…”

“…bits with a smaller bit-error probability [33][34][35][42][43][44][45]. We limit the scope of our analysis to the simple binarization scheme, the reliable component selection (RCS) scheme [33][34][35], and the DROBA scheme [42].…”

“…However, the drawback of not having a score is that it is not possible to apply fusion at score-level. Therefore, published work on fusion with template protection are mainly focussed on fusion at feature-level or at decision-level [33,34,[46][47][48]. However, we show in Chapter 7 that by extending the PI reconstruction process with the derivation of a dissimilarity score, it is possible to apply fusion at score-level, given some limitations on the match and non-match regions that can be created.…”

“…In the verification phase, the key can be released by combining the protected template with a newly captured probe biometric sample. Examples of known key binding and release implementations are the Code-Offset Construction [40,64,65], Fuzzy Commitment Scheme (FCS) [36,[66][67][68][69][70][71], the Helper Data System (HDS) [33][34][35]72,73], Quantization Index Modulation (QIM) [48,74,75], and the Fuzzy Vault [46,[76][77][78][79][80][81][82][83][84][85][86]. The first three schemes are related to each other as portrayed in Figure 2.6.…”

“…This has stimulated research into methods for protecting the biometric data in order to mitigate these privacy concerns. Numerous methods such as the fuzzy commitment scheme [36], helper data system [33,34,48], fuzzy extractors [64,65], fuzzy vault [80,84] and cancelable biometrics [58] have been proposed for transforming the biometric data in such a way that the privacy is safeguarded. Several of these privacy or template protection techniques use some cryptographic primitives (e.g.…”