[Paper Review] A Generalized Approach for Cancellable Template and Its Realization for Minutia Cylinder-Code

Xingbo Dong, Zhe Jin+1 more|arXiv (Cornell University)|Mar 2, 2022

Biometric Identification and Security27 references4 citations

TL;DR

This paper proposes gIoM, a generalized Index-of-Max hashing technique for cancellable biometric template protection, enabling secure, revocable, and non-invertible representation of unordered, variable-size biometric templates like fingerprint minutiae via Minutia Cylinder-Code (MCC). The method transforms MCC into an index-based feature space, ensuring computational infeasibility of template reconstruction while preserving recognition performance on FVC2002 and FVC2004 databases.

ABSTRACT

Hashing technology gains much attention in protecting the biometric template lately. For instance, Index-of-Max (IoM), a recent reported hashing technique, is a ranking-based locality sensitive hashing technique, which illustrates the feasibility to protect the ordered and fixed-length biometric template. However, biometric templates are not always in the form of ordered and fixed-length, rather it may be an unordered and variable size point set e.g. fingerprint minutiae, which restricts the usage of the traditional hashing technology. In this paper, we proposed a generalized version of IoM hashing namely gIoM, and therefore the unordered and variable size biometric template can be used. We demonstrate a realization using a well-known variable size feature vector, fingerprint Minutia Cylinder-Code (MCC). The gIoM transforms MCC into index domain to form indexing-based feature representation. Consequently, the inversion of MCC from the transformed representation is computational infeasible, thus to achieve non-invertibility while the performance is preserved. Public fingerprint databases FVC2002 and FVC2004 are employed for experiment as benchmark to demonstrate a fair comparison with other methods. Moreover, the security and privacy analysis suggest that gIoM meets the criteria of template protection: non-invertibility, revocability, and non-linkability.

Motivation & Objective

• To address the limitation of existing hashing techniques that require fixed-length, ordered biometric templates, which do not suit variable-size, unordered biometric data like fingerprint minutiae.
• To develop a generalized hashing approach capable of protecting unordered and variable-size biometric templates while maintaining recognition accuracy.
• To realize a cancellable template system based on Minutia Cylinder-Code (MCC) that ensures non-invertibility, revocability, and non-linkability.
• To evaluate the proposed method on standard fingerprint databases (FVC2002 and FVC2004) for fair comparison with existing methods.

Proposed method

• Propose gIoM, a generalized version of Index-of-Max (IoM) hashing, to handle unordered and variable-size biometric templates.
• Transform the Minutia Cylinder-Code (MCC), a variable-length feature vector representing fingerprint minutiae, into an index domain through gIoM for indexing-based representation.
• Apply a permutation-based transformation to the MCC feature vector before applying the gIoM operation to enhance randomness and security.
• Ensure non-invertibility by making the reverse mapping from the transformed index representation to the original MCC computationally infeasible.
• Use a ranking-based locality-sensitive hashing mechanism in the index domain to preserve discriminative information for recognition.
• Integrate the gIoM framework into a cancellable biometric template system that supports template revocation and prevents template linking across systems.

Experimental results

Research questions

• RQ1 Can a generalized hashing technique be designed to protect unordered and variable-size biometric templates such as fingerprint minutiae?
• RQ2 How can the non-invertibility of the transformed biometric template be mathematically and computationally ensured in the absence of fixed-length constraints?
• RQ3 To what extent does the gIoM-based representation preserve recognition performance compared to existing methods on standard fingerprint databases?
• RQ4 Does the proposed method satisfy core privacy properties: non-invertibility, revocability, and non-linkability?
• RQ5 Can the gIoM approach be effectively realized using the Minutia Cylinder-Code (MCC) representation without compromising accuracy or security?

Key findings

• The gIoM method successfully transforms variable-size Minutia Cylinder-Code (MCC) templates into a secure index-based representation, enabling cancellable template protection.
• The inverse mapping from the gIoM-transformed representation to the original MCC is computationally infeasible, ensuring strong non-invertibility.
• The method achieves competitive recognition performance on FVC2002 and FVC2004 databases, demonstrating preservation of biometric utility.
• Security and privacy analysis confirms that gIoM satisfies the three core criteria: non-invertibility, revocability, and non-linkability.
• The proposed approach enables effective protection of unordered and variable-size biometric templates, extending the applicability of hashing-based cancellable templates beyond fixed-length formats.