Skip to content
NewA calmer, faster ZeroHour

Is it Possible to Generate Irreversible PolyProtected Templates from Face Embeddings using System-Specific Keys?

arXiv cs.CRby Vedrana Krivokuća Hahn ()1 min read1 · 1 read
AI summary · grok-4.7Summaries by Vani AI

Study shows PolyProtect can irreversibly protect face embeddings with shared system-specific keys, enabling protected-domain de-duplication.

The authors test whether PolyProtect remains irreversible when the same system-specific C and E parameters protect every subject's face embeddings, instead of per-subject keys. Shared keys would allow identity de-duplication in the protected domain, which subject-specific keys prevent because of unlinkability. Experiments, reproducible from open-source code, identify three systematic ways to choose such keys: reuse of pre-selected subject keys, a prior selection algorithm on random vectors, and sampling from an approximated good C/E distribution. The authors conclude template protection is not degraded.

Key facts
  • System-specific C and E keys are shared across all subjects.
  • Subject-specific keys block de-duplication because of unlinkability.
  • Three systematic selection methods still yield irreversible templates.
  • Experiments are reproducible with the authors' open-source code.
Analyze this story with Vani AIImpact, who is affected and what defenders should do, in a new chat

From the original article

1 min read at arxiv.org

This work aims to answer the question of whether it is possible to generate irreversible protected templates when the PolyProtect biometric template protection method is applied to face embeddings using system-specific keys (i. e. , the same C and E parameters, which define the transform, are applied to all subjects' face embeddings), instead of the traditional subject-specific keys (i.

arXiv cs.CR

By Vedrana Krivokuća Hahn · arxiv.org

Read the full article at arXiv cs.CR

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.