Metraipolis › archive › dc_b03b16eef45d

Mind the RefGAP: Correcting Reference Attention in Diffusion-Based Visual Editing — arXiv:2609.35708

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.35708
License (as archived)
CC0-1.0
Archived
2026-10-07T23:54:56.282Z
SHA-256 of the archived text
5597502b89bf57ef6b5de0227e225dd931b6f2647430990e67b735fbdc08e6f6
Stable URL of this page
https://metraipolis.com/documento/dc_b03b16eef45d

How to cite

Mind the RefGAP: Correcting Reference Attention in Diffusion-Based Visual Editing — arXiv:2609.35708 — arXiv (en), https://arxiv.org/abs/2609.35708 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b03b16eef45d (Metraipolis archive dc_b03b16eef45d, sha256 5597502b89bf57ef6b5de0227e225dd931b6f2647430990e67b735fbdc08e6f6)
BibTeX
@misc{dcb03b16eef45d,
  title = {Mind the RefGAP: Correcting Reference Attention in Diffusion-Based Visual Editing — arXiv:2609.35708},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.35708},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b03b16eef45d; sha256 5597502b89bf57ef6b5de0227e225dd931b6f2647430990e67b735fbdc08e6f6},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35708. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Reference-guided diffusion editors struggle to faithfully reproduce user-provided references. We identify a potential bottleneck in diffusion editors: many methods provide limited reference-attention allocation. For example, in LoomVideo, edit-region queries assign less than 1% of their attention mass to the reference. We introduce RefGAP, a training-free correction that determines logit-offset magnitudes online at each layer from the reference-attention mass measured during the forward pass. Positive offsets to reference logits strengthen reference usage by edit-region queries, while negative offsets for keep-region queries limit reference-induced changes outside the edit. Two global coefficients control the correction; they are selected once on validation data from four development diffusion editors and held fixed. Across seven diffusion-based image/video editors, RefGAP improves identity fidelity in head swapping and face swapping. RefGAP achieves a fidelity-preservation trade-off comparable to separately tuned constant edit-side biases, without per-approach strength sweeps. Additional experiments on virtual try-on and background replacement evaluate transfer beyond identity editing.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.35708 — 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 2026-10-07T23:29:27.621Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).