Challenges of Biometric Data Preservation
The challenges related with the student identity management in HEIs entails several aspects. The most pressing risks concern:
- 1. Security of biometric data, where in several cases biometric data is not secret (e.g., fingerprints can be extracted from surfaces that the user touched, faces can be easily acquired from public online sources, voices can be recorded, etc.);
- 2. Privacy of biometric data that is stored could expose sensitive information about the end-users (e.g., ethnic origin, health information, etc.); and
- 3. Revocability of biometric data, which in case they are compromised, they are nearly impossible to be revoked by the end-users.
At present one of the most commonly applied solution is deployment of the biometric templates, which are digital representations of certain features extracted from a biometric sample (e.g., the shape of a user’s hand), that do not require to store the exact raw biometric data of the user. This approach helps to avoid potential privacy issues in case of a compromised data set. This method relates to transforming the biometric template into a new domain through the integration of biometric data with externally generated randomness in a non-invertible way. Therefore, this solution protects the sensitive biometric data in a manner equivalent to a cryptographic cipher.
Additional Methods for Biometric Data Preservation
It is important to acknowledge additional non-invertible transformation methods, which have been proposed to create multiple cancelable identifiers from a biometric template, such as Cartesian, polar, and surface folding. Furthermore, the biometric cryptosystem approach, that binds a key to the biometric template, can be combined with neural networks-based transform approaches. But as these last are stored in an unprotected manner, they are vulnerable and could be prone to attacks.
The biometric encryption techniques have been employed to address privacy issues in biometrics. Given the high variability of biometric data, traditional cryptographic hashing approaches may not be suitable for this type of data. Therefore, different cryptographic tools have been applied such as homomorphic encryption.
Homomorphic encryption is an approach designed to protect biometric templates, where the encrypted biometric template is stored in the database and during verification, the matching module calculates the similarity score between the encrypted stored template and the encrypted query template. However, the feature extraction methods and template protection methods have been developed independent of each other and it cause a reciprocity between matching performance, privacy, and computational cost.
Protocol-based approaches are another methods proposed to protect the privacy of biometric data, such as secure multi party computation (SMC) protocol or zero-knowledge proof (ZKP) protocol. SMC protocols are cryptographic protocols that preserve the privacy of each participant and can be used within privacy-preserving biometric systems. ZKP protocols also represent cryptographic protocols which can be used for privacy-preservation of biometric systems. Within this approach user is obliged to prove to the verifier a certain type of knowledge without revealing any additional information.
Distributed ledger technologies (e.g., private blockchain technologies) have specific features that can address several of the existing challenges in privacy-preserving biometrics, i.e., the distributed nature addresses the problem of single-point of failure, elimination of third-parties and potential privacy leakage, monitoring and access to trustable and unmodifiable history logs. Recent blockchain-basedworks in the literature for privacy-preservation of biometrics include a protocol for storing biometric credentials in a decentralized way that utilizes decentralized identifiers and docunts implemented by W3C Verifiable Claims, and a method for protecting fingerprint templates via blockchain technology. In this method the fingerprint features are extracted, encrypted with a AES block cipher, and then uploaded to a symmetric distributed storage system.