Denoising Diffusion Restoration Models

TL;DR

Denoising Diffusion Restoration Models (DDRM) provide an efficient, unsupervised approach for solving various linear inverse image restoration problems by leveraging pre-trained diffusion models, outperforming existing methods in quality and speed.

Contribution

Introduces DDRM, a novel unsupervised, diffusion-based method for linear inverse problems that is versatile, fast, and does not require problem-specific training.

Findings

DDRM outperforms current unsupervised methods in reconstruction and perceptual quality.

DDRM is approximately 5 times faster than the nearest competitor.

DDRM generalizes well to natural images outside the training distribution.

Abstract

Many interesting tasks in image restoration can be cast as linear inverse problems. A recent family of approaches for solving these problems uses stochastic algorithms that sample from the posterior distribution of natural images given the measurements. However, efficient solutions often require problem-specific supervised training to model the posterior, whereas unsupervised methods that are not problem-specific typically rely on inefficient iterative methods. This work addresses these issues by introducing Denoising Diffusion Restoration Models (DDRM), an efficient, unsupervised posterior sampling method. Motivated by variational inference, DDRM takes advantage of a pre-trained denoising diffusion generative model for solving any linear inverse problem. We demonstrate DDRM's versatility on several image datasets for super-resolution, deblurring, inpainting, and colorization under various amounts of measurement noise. DDRM outperforms the current leading unsupervised methods on the diverse ImageNet dataset in reconstruction quality, perceptual quality, and runtime, being 5x faster than the nearest competitor. DDRM also generalizes well for natural images out of the distribution of the observed ImageNet training set.