Metraipolis › archive › dc_682fdcb9df08

In-Token Learning for High-Fidelity Image Restoration via Diffusion Transformers — arXiv:2609.33523

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.33523
License (as archived): CC0-1.0
Archived: 2026-10-07T23:51:40.052Z
SHA-256 of the archived text: a849dcef56820f940fa4ff003b6a0c4ac49c4c650d5b52dbcbb60198ef28dc4c
Stable URL of this page: https://metraipolis.com/documento/dc_682fdcb9df08

How to cite

In-Token Learning for High-Fidelity Image Restoration via Diffusion Transformers — arXiv:2609.33523 — arXiv (en), https://arxiv.org/abs/2609.33523 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_682fdcb9df08 (Metraipolis archive dc_682fdcb9df08, sha256 a849dcef56820f940fa4ff003b6a0c4ac49c4c650d5b52dbcbb60198ef28dc4c)

BibTeX

@misc{dc682fdcb9df08,
  title = {In-Token Learning for High-Fidelity Image Restoration via Diffusion Transformers — arXiv:2609.33523},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33523},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_682fdcb9df08; sha256 a849dcef56820f940fa4ff003b6a0c4ac49c4c650d5b52dbcbb60198ef28dc4c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33523. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We present In-Token Learning, an image restoration framework that adapts a pretrained diffusion transformer using conditional rectified flow matching. Clean targets paired with degraded inputs supervise transport from Gaussian noise to restored images. Spatially aligned degraded-image tokens are fused with evolving latent tokens along the channel dimension, preserving the image-token count at a given resolution. Direct Low-Quality Guidance (DLG) combines frozen degraded-image embeddings with a fixed task prompt through the native conditioning pathway, without a trainable ControlNet-style branch or image captioning. We evaluate super-resolution and denoising on DIV2K, LSDIR, FFHQ, RealLQ250, and RealPhoto60, and automatic colorization on DIV2K and LSDIR. The tasks use separately trained checkpoints under the same framework. Results show competitive fidelity and perceptual quality under the evaluated protocols, with weaker generalization on RealLQ250. We report full-image QHD ($2560{\times}1440$) inference and a tiled $12$K restoration demonstration of Along the River During the Qingming Festival. Attention cost still increases with resolution. This technical report preserves the early broader study underlying Fill2SR, which subsequently developed the real-world super-resolution direction.

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/2609.33523 — 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 2026-10-07T23:29:27.223Z 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).