EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multi-Source Training-Free Controllable Style Transfer via Diffusion Models

Cuihong Yu, Cheng Han, Chao Zhang

Symmetry · 13.02.2025 · 17 (2) · 290

DOI: 10.3390/sym17020290 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Symmetry, online erschienen am 13.02.2025. Verfasst von Cuihong Yu und 2 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Natural Science Foundation of Jilin Province, National Natural Science Foundation of China, Science and Technology Development Plan Project of Jilin Province.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Diffusion models, as representative models in the field of artificial intelligence, have made significant progress in text-to-image synthesis. However, studies of style transfer using diffusion models typically require a large amount of text to describe semantic content or specific painting attributes, and the style and layout of semantic content in synthesized images are frequently uncertain. To accomplish high-quality fixed content style transfer, this paper adopts text-free guidance and proposes a multi-source, training-free and controllable style transfer method by using single image or video as content input and single or multiple style images as style guidance. To be specific, the proposed method firstly fuses the inversion noise of a content image with that of a single or multiple style images as the initial noise of stylized image sampling process. Then, the proposed method extracts the self-attention mechanism’s query, key, and value vectors from the DDIM inversion process of content and style images and injects them into the stylized image sampling process to improve the color, texture and semantics of stylized images. By setting the hyperparameters involved in the proposed method, the style transfer effect of symmetric style proportion and asymmetric style distribution can be achieved. By comparing with state-of-the-art baselines, the proposed method demonstrates high fidelity and excellent stylized performance, and can be applied to numerous image or video style transfer tasks.

Förderung

  • Natural Science Foundation of Jilin Province
  • National Natural Science Foundation of China
  • Science and Technology Development Plan Project of Jilin Province

Zitieren

APA

Yu, C., Han, C., & Zhang, C. (2025). Multi-Source Training-Free Controllable Style Transfer via Diffusion Models. Symmetry, 17(2), 290. https://doi.org/10.3390/sym17020290

Vancouver

Yu C, Han C, Zhang C. Multi-Source Training-Free Controllable Style Transfer via Diffusion Models. Symmetry. 2025;17(2):290. doi:10.3390/sym17020290

BibTeX
@article{Yu2025,
  author = {Yu, Cuihong and Han, Cheng and Zhang, Chao},
  title = {Multi-Source Training-Free Controllable Style Transfer via Diffusion Models},
  journal = {Symmetry},
  year = {2025},
  volume = {17},
  number = {2},
  pages = {290},
  doi = {10.3390/sym17020290}
}
Kostenlos starten →