FideDiff model offers efficient, high-fidelity image deblurring using diffusion

Researchers have developed FideDiff, a novel single-step diffusion model designed for high-fidelity image motion deblurring. This model reformulates deblurring as a diffusion-like process, where a consistency model aligns progressively blurred images to a single clean image. By integrating Kernel ControlNet for blur kernel estimation and adaptive timestep prediction, FideDiff achieves superior performance on full-reference metrics, outperforming previous diffusion-based methods and matching other state-of-the-art models. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Introduces a more efficient diffusion model for image restoration, potentially improving real-world applications.

RANK_REASON This is a research paper detailing a new model for image deblurring. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• FideDiff
• arXiv
• Xiaoyang Liu
• Kernel ControlNet

• paper
• other