Multiplication-Free Biometric Recognition for Faster Processing under Encryption

Abstract

The cutting-edge biometric recognition systems extract distinctive feature vectors of biometric samples using deep neural networks to measure the amount of (dis-)similarity between two biometric samples. Studies have shown that personal information (e.g., health condition, ethnicity, etc.) can be inferred, and biometric samples can be reconstructed from those feature vectors, making their protection an urgent necessity. State-of-theart biometrics protection solutions are based on homomorphic encryption (HE) to perform recognition over encrypted feature vectors, hiding the features and their processing while releasing the outcome only. However, this comes at the cost of those solutions' efficiency due to the inefficiency of HE-based solutions with a large number of multiplications; for (dis-)similarity measures, this number is proportional to the vector's dimension. In this paper, we tackle the HE performance bottleneck by freeing the two common (dis-)similarity measures, the cosine similarity and the squared Euclidean distance, from multiplications. Assuming normalized feature vectors, our approach pre-computes and organizes those (dis-)similarity measures into lookup tables. This transforms their computation into simple table-lookups and summation only. We study quantization parameters for the values in the lookup tables and evaluate performances on both synthetic and facial feature vectors for which we achieve a recognition performance identical to the non-tabularized baseline systems. We then assess their efficiency under HE and record runtimes between 28.95ms and 59.35ms for the three security levels, demonstrating their enhanced speed.

Biometric Template Attacks and Recent Protection Mechanisms: A Survey

“…Results from different biometric datasets indicate HERS is more computationally efficient than HEFT but has less accuracy in terms of verification performance. Another computationally efficient approach is the multiplication-free strategy proposed by Bassit et al [161]. The authors overcome the performance limitation from the sole use of homomorphic encryption by eliminating multiplications from the two well-known dissimilarity metrics (cosine similarity and Euclidian distance) into summations and simple lookup tables.…”

Improved Multiplication-Free Biometric Recognition Under Encryption

“…In this paper, we extend the work conducted in [1] by improving the integration of the MFBR schemes with HE, as illustrated in Table I, where we denote by MFBRv1 [1] our initial version, and by MFBRv2 our improved one proposed in the present paper. Our solution, as in [1], is built upon the HELR framework [9] but applied to the IP and SED measures that do not require training.…”

“…In this paper, we extend the work conducted in [1] by improving the integration of the MFBR schemes with HE, as illustrated in Table I, where we denote by MFBRv1 [1] our initial version, and by MFBRv2 our improved one proposed in the present paper. Our solution, as in [1], is built upon the HELR framework [9] but applied to the IP and SED measures that do not require training. Assuming normalized feature vectors extracted from a well-trained DNN, we determine the probability density function (PDF) and cumulative distribution function (CDF) corresponding to the projection of a point on the unit d-ball upon which we generate the lookup tables (that we call MFIP and MFSED) in an equiprobable manner to reinforce their security.…”

“…In [5], the author considers only the cleartext decision mode, which has the downside of exposing the final score. The knowledge of the final score can lead to the reconstruction of the original biometric template, as shown by [12]- [15]. In contrast to the encryption decision mode, the comparison with the threshold performed inside the encrypted domain reveals only one bit of information (match / no match).…”

“…the influence of their integration with HE, achieving a performance better than the baseline. • We improve the integration of MFBRs with HE presented in [1], rendering the encrypted reference template more compact, restricting it to one ciphertext instead of d ciphertexts.…”