Metraipolis › archive › dc_aea8dabcfebd

FlowTranslator: Training-Free In-Image Machine Translation with Flow-Based Editing — arXiv:2611.04303

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.04303
License (as archived): CC0-1.0
Archived: 2026-11-04T18:36:36.070Z
SHA-256 of the archived text: f0714c69baa15ba985e848b1e5362f150c2c5a3c08fdd54aed24e1e065a98fad
Stable URL of this page: https://metraipolis.com/documento/dc_aea8dabcfebd

How to cite

FlowTranslator: Training-Free In-Image Machine Translation with Flow-Based Editing — arXiv:2611.04303 — arXiv (en), https://arxiv.org/abs/2611.04303 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_aea8dabcfebd (Metraipolis archive dc_aea8dabcfebd, sha256 f0714c69baa15ba985e848b1e5362f150c2c5a3c08fdd54aed24e1e065a98fad)

BibTeX

@misc{dcaea8dabcfebd,
  title = {FlowTranslator: Training-Free In-Image Machine Translation with Flow-Based Editing — arXiv:2611.04303},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04303},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_aea8dabcfebd; sha256 f0714c69baa15ba985e848b1e5362f150c2c5a3c08fdd54aed24e1e065a98fad},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04303. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In-image machine translation (IIMT) replaces text in an image with its translation while preserving the surrounding visual context. Existing task-specific models achieve convincing performance but require massive paired training images before and after translation. As advanced text-to-image (T2I) models have reached a practically usable level of multilingual text generation, we argue that this capability can be transferred to IIMT and present FlowTranslator. It is a training-free IIMT framework that first employs an OCR model and an LLM to generate the translation prompt, which is injected into the T2I-based diffusion model. To optimize the accurate and natural generation of translated text, we introduce two trajectory controls: Region Coupling Release (RCR) partially decorrelates target-branch noise inside the text region early in the trajectory, while Coupling Expansion (CE) amplifies the emerging edit displacement mid-trajectory. Under the EasyOCR read-back protocol, FlowTranslator achieves average PRIM BLEU and COMET scores of 20.6 and 57.2, respectively, substantially outperforming the strong training-based baseline UniTranslator. On the more challenging and realistic whole-image ScenePair photographs translated into Chinese, it reaches 40.1\% exact-match accuracy and surpasses existing editing systems, some of which only address the text-rendering stage rather than the complete IIMT task. These results establish flow-based editing as a strong, fully training-free baseline for IIMT.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.04303 — 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 2026-11-04T18:32:37.434Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).