Pith Number

pith:Y7ZMGWJX

pith:2026:Y7ZMGWJXJF7EROQ2MFT5ZR4O7F

not attested not anchored not stored refs pending

Lightweight, Practical Encrypted Face Recognition with GPU Support

Bahattin Yildiz, Eduardo L. Cominetti, Gabrielle De Micheli, Geovandro Pereira, Jina Choi, Marcos A. Simplicio Jr, Syed Mahbub Hafiz, Thales B. Paiva

BSGS-Diagonal reorders rotations to cut rotation keys by 91 percent while preserving correctness in encrypted face recognition.

arxiv:2604.00546 v3 · 2026-04-01 · cs.CR

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{Y7ZMGWJXJF7EROQ2MFT5ZR4O7F}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

BSGS-Diagonal yields a 91% reduction in the number of rotation keys, translating to approximately 14 GB less memory used on the client, and reducing overall CPU peak RAM from over 30 GB in the original HyDia to under 10 GB for databases up to size 1M; runtime improved by up to 1.57x for membership verification and 1.43x for identification; GPU implementations achieve up to 9x and 21x speedups, enabling sub-second encrypted face recognition for databases up to 32K entries.

C2weakest assumption

That the BSGS-Diagonal reordering of rotations preserves both the exact correctness of the similarity scores and the semantic security of the underlying CKKS scheme while only changing performance characteristics; this is stated implicitly in the description of the algorithm but not accompanied by a formal security reduction or accuracy verification in the provided abstract.

C3one line summary

BSGS-Diagonal and fused GPU kernels reduce memory and computation for FHE-based face recognition, enabling sub-second encrypted matching on databases up to 32K entries with 91% fewer rotation keys.

Receipt and verification

First computed	2026-06-02T01:03:46.414706Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

c7f2c35937497e48ba1a6167dcc78ef95a2cdac0373488fe184f767a894b19e7

Aliases

arxiv: 2604.00546 · arxiv_version: 2604.00546v3 · doi: 10.48550/arxiv.2604.00546 · pith_short_12: Y7ZMGWJXJF7E · pith_short_16: Y7ZMGWJXJF7EROQ2 · pith_short_8: Y7ZMGWJX

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/Y7ZMGWJXJF7EROQ2MFT5ZR4O7F \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: c7f2c35937497e48ba1a6167dcc78ef95a2cdac0373488fe184f767a894b19e7

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "79f182439f5cedc1e3dbcf91e2fd8aa5c0498a0aa3f5bf5ed2b9f8b6b0340466",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CR",
    "submitted_at": "2026-04-01T06:43:36Z",
    "title_canon_sha256": "334f73e698e7d7d53f68dec7989f465d5816d4f9f96c08f233318d76a7f0b808"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2604.00546",
    "kind": "arxiv",
    "version": 3
  }
}