Metraipolis › archive › dc_4f4863ad2ae2

Can Pixels Alone Reveal Image Origin? Minimax Limits and Learnable Interfaces for Passive Provenance — arXiv:2609.30997

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.30997
License (as archived)
CC0-1.0
Archived
2026-10-08T01:39:39.360Z
SHA-256 of the archived text
9885ce82a39353445b8cf1fd6983e336cdc4d0b06ff9df326a1b132c58854289
Stable URL of this page
https://metraipolis.com/documento/dc_4f4863ad2ae2

How to cite

Can Pixels Alone Reveal Image Origin? Minimax Limits and Learnable Interfaces for Passive Provenance — arXiv:2609.30997 — arXiv (en), https://arxiv.org/abs/2609.30997 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4f4863ad2ae2 (Metraipolis archive dc_4f4863ad2ae2, sha256 9885ce82a39353445b8cf1fd6983e336cdc4d0b06ff9df326a1b132c58854289)
BibTeX
@misc{dc4f4863ad2ae2,
  title = {Can Pixels Alone Reveal Image Origin? Minimax Limits and Learnable Interfaces for Passive Provenance — arXiv:2609.30997},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.30997},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4f4863ad2ae2; sha256 9885ce82a39353445b8cf1fd6983e336cdc4d0b06ff9df326a1b132c58854289},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30997. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Passive image provenance asks whether pixels alone can reveal where an image came from: a human, an aggregate AI class, or a particular generator. This becomes a robustness problem once a source image can be edited before the verifier sees it. We study the problem as source--target verification under adversarial distribution shift. Our first result gives the exact best-case limit for any image-only verifier: the largest robust target-acceptance gap equals the minimum total-variation distance between the target distribution and the set of attacked source distributions. This quantity depends on the source, target, and edit class, not on the verifier architecture. Our second result explains why deployed public verifiers can fail before this statistical limit is reached. If the verifier can be emulated on the attack region to error $\varepsilon$, then a surrogate black-box attack reaches target acceptance within $2\varepsilon$ plus optimization error of the white-box optimum; score-revealing logistic and softmax heads over public features are identifiable, and approximate score access gives stable recovery bounds. A finite-state experiment checks the minimax identity where both sides are computable. On same-prompt real/diffusion benchmarks, the evaluated public CLIP verifiers fail under targeted pixel attacks, while a ResNet-18 victim exhibits partial fake-to-real transfer. Binary feedback with abstention reduces measured attack success, but positive empirical gap upper bounds do not establish robustness. These results motivate separate evaluation of the source--target statistical ceiling and the information released by a deployed verifier.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-28 in cs.AI).
Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.30997 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T01:12:14.522Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).