Boosting Diffusion Guidance via Learning Degradation-Aware Models for Blind Super Resolution

TL;DR

This paper introduces DADiff, a diffusion-based blind super-resolution method that learns degradation-aware models to improve high-fidelity image reconstruction without requiring known degradation kernels.

Contribution

DADiff integrates degradation-aware models into diffusion guidance for blind SR and proposes input perturbation and guidance scalar techniques to enhance performance.

Findings

Outperforms state-of-the-art blind SR methods

Achieves high-frequency detail with improved fidelity

Demonstrates robustness across various degradation conditions

Abstract

Recently, diffusion-based blind super-resolution (SR) methods have shown great ability to generate high-resolution images with abundant high-frequency detail, but the detail is often achieved at the expense of fidelity. Meanwhile, another line of research focusing on rectifying the reverse process of diffusion models (i.e., diffusion guidance), has demonstrated the power to generate high-fidelity results for non-blind SR. However, these methods rely on known degradation kernels, making them difficult to apply to blind SR. To address these issues, we present DADiff in this paper. DADiff incorporates degradation-aware models into the diffusion guidance framework, eliminating the need to know degradation kernels. Additionally, we propose two novel techniques: input perturbation and guidance scalar, to further improve our performance. Extensive experimental results show that our proposed method has superior performance over state-of-the-art methods on blind SR benchmarks.