EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Differential Diffusion: Giving Each Pixel Its Strength

Eran Levin, Ohad Fried

Computer Graphics Forum · 2025-05 · 44 (2)

DOI: 10.1111/cgf.70040 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Computer Graphics Forum, online erschienen am 14.04.2025. Verfasst von Eran Levin und 1 weiteren Autor:innen. Die Arbeit wurde bisher 13-mal zitiert (Crossref) und nennt 66 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Israel Science Foundation, Tel Aviv University.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Diffusion models have revolutionized image generation and editing, producing state‐of‐the‐art results in conditioned and unconditioned image synthesis. While current techniques enable user control over the degree of change in an image edit, the controllability is limited to global changes over an entire edited region. This paper introduces a novel framework that enables customization of the amount of change per pixel or per image region . Our framework can be integrated into any existing diffusion model, enhancing it with this capability. Such granular control opens up a diverse array of new editing capabilities, such as control of the extent to which individual objects are modified, or the ability to introduce gradual spatial changes. Furthermore, we showcase the framework's effectiveness in soft‐inpainting—the completion of portions of an image while subtly adjusting the surrounding areas to ensure seamless integration. Additionally, we introduce a new tool for exploring the effects of different change quantities. Our framework operates solely during inference, requiring no model training or fine‐tuning. We demonstrate our method with the current open state‐of‐the‐art models, and validate it via both quantitative and qualitative comparisons, and a user study. Our code is published and integrated into several platforms.

Förderung

  • Israel Science Foundation
  • Tel Aviv University

Zitieren

APA

Levin, E., & Fried, O. (2025). Differential Diffusion: Giving Each Pixel Its Strength. Computer Graphics Forum, 44(2). https://doi.org/10.1111/cgf.70040

Vancouver

Levin E, Fried O. Differential Diffusion: Giving Each Pixel Its Strength. Computer Graphics Forum. 2025;44(2). doi:10.1111/cgf.70040

BibTeX
@article{Levin2025,
  author = {Levin, Eran and Fried, Ohad},
  title = {Differential Diffusion: Giving Each Pixel Its Strength},
  journal = {Computer Graphics Forum},
  year = {2025},
  volume = {44},
  number = {2},
  doi = {10.1111/cgf.70040}
}
Kostenlos starten →