[Resource Topic] 2023/736: Private Eyes: Zero-Leakage Iris Searchable Encryption

Welcome to the resource topic for 2023/736

Private Eyes: Zero-Leakage Iris Searchable Encryption

Authors: Julie Ha, Chloe Cachet, Luke Demarest, Sohaib Ahmad, Benjamin Fuller


Biometric databases are being deployed with few cryptographic protections. Because of the nature of biometrics, privacy breaches affect users for their entire life.
This work introduces Private Eyes, the first zero-leakage biometric database. The only leakage of the system is unavoidable: 1) the log of the dataset size and 2) the fact that a query occurred. Private Eyes is built from symmetric searchable encryption. Proximity queries are the required functionality: given a noisy reading of a biometric, the goal is to retrieve all stored records that are close enough according to a distance metric.
Private Eyes combines locality sensitive-hashing or LSHs (Indyk and Motwani, STOC 1998) and encrypted maps. One searches for the disjunction of the LSHs of a noisy biometric reading. The underlying encrypted map needs to efficiently answer disjunction queries.
We focus on the iris biometric. Iris biometric data requires a large number of LSHs, approximately 1000. The most relevant prior work is in zero-leakage k-nearest-neighbor search (Boldyreva and Tang, PoPETS 2021), but that work is designed for a small number of LSHs.
Our main cryptographic tool is a zero-leakage disjunctive map designed for the setting when most clauses do not match any records. For the iris, on average at most 6% of LSHs match any stored value.
To aid in evaluation, we produce a synthetic iris generation tool to evaluate sizes beyond available iris datasets. This generation tool is a simple generative adversarial network. Accurate statistics are crucial to optimizing the cryptographic primitives so this tool may be of independent interest.
Our scheme is implemented and open-sourced. For the largest tested parameters of 5000 stored irises, search requires 26 rounds of communication and 26 minutes of single-threaded computation.

ePrint: https://eprint.iacr.org/2023/736

See all topics related to this paper.

Feel free to post resources that are related to this paper below.

Example resources include: implementations, explanation materials, talks, slides, links to previous discussions on other websites.

For more information, see the rules for Resource Topics .