A cancelable multi-biometric template generation algorithm based on bloom filter

Abstract

For the security issue of multiple biometric templates in current multi-biometric systems, this paper proposes a cancelable multi-biometric template generation algorithm based on Bloom filter. Our algorithm uses the XOR operation to fuse the grouped fingerprint binary features and the face binary features into one template at the feature level, then transforms the fusion template based on the irreversibility of the Bloom filter. The cancelability and diversity of the fusion template can be achieved by updating the random matrix. Finally, a traversal matching method is used to calculate the matching score in the encryption domain. The experimental results show that our algorithm can ensure the reliability of the identity authentication and improve the security of the multi-biometric template.