pith:MF2PILEK
Towards Generalized Image Manipulation Localization via Score-based Model
DiffIML approximates the score function of mask distributions to iteratively recover coherent manipulation masks from noise, improving generalization over discriminative methods.
arxiv:2605.16879 v1 · 2026-05-16 · cs.CV
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{MF2PILEKDCT4U3FWAVS5CC2FG2}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
DiffIML approximates the score function of mask distributions to iteratively recover coherent masks from noise, circumventing the brittleness of discriminative models and yielding consistent generalization improvements on diverse unseen datasets across eight non-generative and three generative benchmarks.
The assumption that the learned score function, combined with edge supervision and error prior in a lightweight latent-space diffusion process, will reliably produce coherent masks without overfitting to training artifacts or requiring extensive per-dataset tuning.
DiffIML applies score-based generative modeling to image manipulation localization, recovering coherent masks iteratively from noise to improve generalization on unseen manipulation types.
References
Formal links
Receipt and verification
| First computed | 2026-05-20T00:03:27.915983Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
6174f42c8a18a7ca6cb60565d10b4536b545bc34706b02b6dea4ec375db2b397
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/MF2PILEKDCT4U3FWAVS5CC2FG2 \
| 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: 6174f42c8a18a7ca6cb60565d10b4536b545bc34706b02b6dea4ec375db2b397
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "c2a634d64184f07067a94ce47daa550e9b2729552775d3c9e6362b9a8cf3a06f",
"cross_cats_sorted": [],
"license": "http://creativecommons.org/licenses/by-nc-nd/4.0/",
"primary_cat": "cs.CV",
"submitted_at": "2026-05-16T08:39:54Z",
"title_canon_sha256": "b2ff810814c5769dcfc475c020bc82816a9f4b49f4ac4a2fd565c581bb248da4"
},
"schema_version": "1.0",
"source": {
"id": "2605.16879",
"kind": "arxiv",
"version": 1
}
}