Encrypted Vector Similarity Computations Using Partially Homomorphic Encryption: Applications and Performance Analysis
Sefik Serengil, Alper Ozpinar · 2025
Read Encrypted Vector Similarity Computations Using Partially Homomorphic Encryption: Applications and Performance Analysis by Sefik Serengil, Alper Ozpinar (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-05850.txt.
Description
Sefik Serengil and Alper Ozpinar test a lighter form of encryption for comparing face data in a 2025 paper. Full encryption that allows any math is slow, so they ask whether a partial form can do the one job they need. That job is to check how alike two face vectors are, without exposing the stored vectors.
Vectors are scaled in advance, so a simple product sum can stand in for the usual angle based score. Tests use a public face photo set and three face models, in a two tower setup with stored data on one side and a local device on the other. The partial method proved faster and lighter than the full method, and it suits tight memory limits.
Catalog record
| CREATOR | Sefik Serengil, Alper Ozpinar |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-05850