williamqf-ai / ucdir

A Unified Conditional Framework for Diffusion-based Image Restoration

A Unified Conditional Framework for Diffusion-based Image Restoration

Yi Zhang¹, Xiaoyu Shi¹, Dasong Li¹, Xiaogang Wang¹, Jian Wang², Hongsheng Li^1
1The Chinese University of Hong Kong, ²Snap Research

Project

---

Abstract: Diffusion Probabilistic Models (DPMs) have recently shown remarkable performance in image generation tasks, which are capable of generating highly realistic images. When adopting DPMs for image restoration tasks, the crucial aspect lies in how to integrate the conditional information to guide the DPMs to generate accurate and natural output, which has been largely overlooked in existing works. In this paper, we present a unified conditional framework based on diffusion models for image restoration. We leverage a lightweight UNet to predict initial guidance and the diffusion model to learn the residual of the guidance. By carefully designing the basic module and integration module for the diffusion model block, we integrate the guidance and other auxiliary conditional information into every block of the diffusion model to achieve spatially-adaptive generation conditioning. To handle high-resolution images, we propose a simple yet effective inter-step patch-splitting strategy to produce arbitrary-resolution images without grid artifacts. We evaluate our conditional framework on three challenging tasks: extreme low-light denoising, deblurring, and JPEG restoration, demonstrating its significant improvements in perceptual quality and the generalization to restoration tasks.

---

Network Architecture

Training

 Coming soon.

Evaluation

 Coming soon.

Citation

If you use KBNet, please consider citing:

@article{zhang2023UCDIR,
  author    = {Zhang, Yi and Shi, Xiaoyu and Li, Dasong and Wang, Xiaogang and Wang, Jian and Li, Hongsheng},
  title     = {A Unified Conditional Framework for Diffusion-based Image Restoration},
  journal   = {arXiv preprint arXiv:2305.20049},
  year      = {2023},
}

Contact

Should you have any question, please contact [email protected]

Acknowledgment: BasicSR, S3.