Image Super-Resolution Via Iterative Refinement

TLDR

The paper introduces SR3, a method for image super‑resolution through repeated refinement. SR3 adapts denoising diffusion probabilistic models to image‑to‑image translation, using a U‑Net trained on denoising across noise levels to iteratively refine a Gaussian‑noise initialization conditioned on a low‑resolution image. SR3 achieves state‑of‑the‑art results, attaining near‑50% fool rates on 8× face super‑resolution, outperforming baselines on 4× ImageNet in human and classifier metrics, and delivering competitive FID scores in cascaded high‑resolution image generation.

Abstract

We present SR3, an approach to image Super-Resolution via Repeated Refinement. SR3 adapts denoising diffusion probabilistic models (Ho et al. 2020), (Sohl-Dickstein et al. 2015) to image-to-image translation, and performs super-resolution through a stochastic iterative denoising process. Output images are initialized with pure Gaussian noise and iteratively refined using a U-Net architecture that is trained on denoising at various noise levels, conditioned on a low-resolution input image. SR3 exhibits strong performance on super-resolution tasks at different magnification factors, on faces and natural images. We conduct human evaluation on a standard 8× face super-resolution task on CelebA-HQ for which SR3 achieves a fool rate close to 50%, suggesting photo-realistic outputs, while GAN baselines do not exceed a fool rate of 34%. We evaluate SR3 on a 4× super-resolution task on ImageNet, where SR3 outperforms baselines in human evaluation and classification accuracy of a ResNet-50 classifier trained on high-resolution images. We further show the effectiveness of SR3 in cascaded image generation, where a generative model is chained with super-resolution models to synthesize high-resolution images with competitive FID scores on the class-conditional 256×256 ImageNet generation challenge.