Real-World Denoising via Diffusion Model

Understand

Real-world image denoising is an extremely important image processing problem, which aims to recover clean images from noisy images captured in natural environments.

• In recent years, diffusion models have achieved very promising results in the field of image generation, outperforming previous generation models.
• However, it has not been widely used in the field of image denoising because it is difficult to control the appropriate position of the added noise.
• Inspired by diffusion models, this paper proposes a novel general denoising diffusion model that can be used for real-world image denoising.