Deep Equilibrium Diffusion Restoration with Parallel Sampling

TL;DR

This paper introduces DeqIR, a novel diffusion-based image restoration method that models the entire sampling process as a fixed point system, enabling parallel sampling and faster, more efficient high-quality image recovery.

Contribution

It proposes a deep equilibrium fixed point formulation for diffusion IR models, allowing parallel sampling and gradient computation without additional training.

Findings

Effective parallel sampling for image restoration

Improved image quality through initialization optimization

Demonstrated superior performance on benchmark IR tasks

Abstract

Diffusion model-based image restoration (IR) aims to use diffusion models to recover high-quality (HQ) images from degraded images, achieving promising performance. Due to the inherent property of diffusion models, most existing methods need long serial sampling chains to restore HQ images step-by-step, resulting in expensive sampling time and high computation costs. Moreover, such long sampling chains hinder understanding the relationship between inputs and restoration results since it is hard to compute the gradients in the whole chains. In this work, we aim to rethink the diffusion model-based IR models through a different perspective, i.e., a deep equilibrium (DEQ) fixed point system, called DeqIR. Specifically, we derive an analytical solution by modeling the entire sampling chain in these IR models as a joint multivariate fixed point system. Based on the analytical solution, we can conduct parallel sampling and restore HQ images without training. Furthermore, we compute fast gradients via DEQ inversion and found that initialization optimization can boost image quality and control the generation direction. Extensive experiments on benchmarks demonstrate the effectiveness of our method on typical IR tasks and real-world settings.