Metraipolis › archive › dc_153e7d2c0875

ImIR: Image-Instruction Tuning for All-in-One Image Restoration — arXiv:2609.25267

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.25267
License (as archived)
CC0-1.0
Archived
2026-10-08T16:39:03.004Z
SHA-256 of the archived text
a6f729916eabc6c19d67132854585d2f2d4990757e3156c23d428765518dc839
Stable URL of this page
https://metraipolis.com/documento/dc_153e7d2c0875

How to cite

ImIR: Image-Instruction Tuning for All-in-One Image Restoration — arXiv:2609.25267 — arXiv (en), https://arxiv.org/abs/2609.25267 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_153e7d2c0875 (Metraipolis archive dc_153e7d2c0875, sha256 a6f729916eabc6c19d67132854585d2f2d4990757e3156c23d428765518dc839)
BibTeX
@misc{dc153e7d2c0875,
  title = {ImIR: Image-Instruction Tuning for All-in-One Image Restoration — arXiv:2609.25267},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.25267},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_153e7d2c0875; sha256 a6f729916eabc6c19d67132854585d2f2d4990757e3156c23d428765518dc839},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.25267. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Degradations vary widely across images, so a practical restoration system has to handle many degradation types with one model. A recent and effective recipe adapts a large pretrained image-editing model to restoration using a small low-rank adapter with a text prompt. We replace that prompt with an instruction derived from the degraded image itself. The image reaches the editor through two paths: its structure comes from the model's VAE, and its semantic instruction comes from a lightweight token mapper that shifts the degraded image's vision-language embedding toward the embedding a clean image would produce. Because the instruction is a continuous vector, scaling it yields a family of valid restorations for tasks whose target is not unique, such as low-light enhancement. We adapt one Qwen-Image-Edit model to six tasks with a single adapter trained in about three hours on one GPU. The image instruction outperforms text conditioning under a matched comparison, and it supports task agnostic restoration without a degradation label, which the text variant does not.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-23 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.25267 — 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 2026-10-08T15:56:18.078Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).