A cancelable multi-biometric system based on the feature-level fusion of fingerprint and finger vein

Increasing security concerns have triggered the proliferation of biometrics-based applications. However, traditional uni-biometric systems cannot trump emerging security attacks and show insufficient recognition accuracy. Hence, multi-biometrics that fuses two or more modalities has become a preferred solution. Existing multi-biometrics schemes suffer from feature incompatibility and privacy leakage issues due to the features being stored in plaintext with different data formats. In this paper, we propose a cancelable multi-biometrics (fingerprint and finger vein) feature-level fusion scheme, ensuring privacy-preserving ability and decent accuracy. To explore the most discriminative features, the minutiae-based fingerprint descriptors and the image-based finger vein features were extracted. Then, the one permutation hashing-based template protection algorithm was deployed to transform these two features into a uniform representation. Extensive experiments were conducted over six self-constructed multi-biometrics datasets, demonstrating the proposed approach’s high recognition performance. The security concerns in terms of renewability and unlinkability are also well-guaranteed.