A Dual-Level Cancelable Framework for Palmprint Verification and Hack-Proof Data Storage

TL;DR

The paper tackles the privacy risk of palmprint template leakage by introducing a Dual-Level Cancelable Palmprint Verification (DCPV) framework that couples a first-level cancelable template produced by a Hashing Competition Palmprint Network with a second-level Negative Database (NDB) for secure data storage. A matrix-based matching approach enables efficient verification against NDB-derived negatives, while an NP-hardness result supports irreversibility of the stored data. The authors provide a thorough security analysis, implement a trainable, binarized feature transformer with a supervised contrastive loss, and demonstrate robust performance across public palmprint datasets with minimal accuracy loss. This dual-protection strategy is particularly suitable for cloud-based palmprint verification, offering strong unlinkability and resistance to database breaches without compromising practical utility.

Abstract

In recent years, palmprints have been widely used for individual verification. The rich privacy information in palmprint data necessitates its protection to ensure security and privacy without sacrificing system performance. Existing systems often use cancelable technologies to protect templates, but these technologies ignore the potential risk of data leakage. Upon breaching the system and gaining access to the stored database, a hacker could easily manipulate the stored templates, compromising the security of the verification system. To address this issue, we propose a dual-level cancelable palmprint verification framework in this paper. Specifically, the raw template is initially encrypted using a competition hashing network with a first-level token, facilitating the end-to-end generation of cancelable templates. Different from previous works, the protected template undergoes further encryption to differentiate the second-level protected template from the first-level one. The system specifically creates a negative database (NDB) with the second-level token for dual-level protection during the enrollment stage. Reversing the NDB is NP-hard and a fine-grained algorithm for NDB generation is introduced to manage the noise and specified bits. During the verification stage, we propose an NDB matching algorithm based on matrix operation to accelerate the matching process of previous NDB methods caused by dictionary-based matching rules. This approach circumvents the need to store templates identical to those utilized for verification, reducing the risk of potential data leakage. Extensive experiments conducted on public palmprint datasets have confirmed the effectiveness and generality of the proposed framework. Upon acceptance of the paper, the code will be accessible at https://github.com/Deep-Imaging-Group/NPR.

Figures (8)

• Figure 1: (a) Traditional Cancellable Palmprint Verification System. (b) Our Proposed Dual-Level Cancelable Palmprint Verification System.
• Figure 2: The attack pipeline of the malicious verification
• Figure 3: The overview of the proposed methods. The proposed DCPV is a dual-level protection framework. In the first-level protection, the model trained with a hybrid loss would generate a binarized cancelable template with the first-level token. Then, the system would store the set of the negative form of the first-level template. In this way, for the matching pairs, the query and stored templates are different in the verification stage. Hence, DCPV could alleviate the potential risk of unauthorized access by adversaries to the stored database.
• Figure 4: The ROC curves of the proposed method on different databases. (a)-(b) denotes the IITD, and PolyU results.
• Figure 5: The matching distributions of the proposed method on different databases. (a)-(f) denotes the IITD, PolyU, Red, Green, Blue, and NIR results.