Efficient Burst Super-Resolution with One-step Diffusion

TL;DR

This paper introduces an efficient burst super-resolution method using a diffusion model that significantly reduces runtime while maintaining high-quality, sharp super-resolved images from low-resolution burst inputs.

Contribution

It presents a novel diffusion-based burst super-resolution approach with a stochastic sampler and one-step diffusion, achieving high efficiency and quality.

Findings

Reduces diffusion model runtime to 1.6% of baseline

Maintains high perceptual and distortion quality

Uses knowledge distillation for efficiency

Abstract

While burst Low-Resolution (LR) images are useful for improving their Super Resolution (SR) image compared to a single LR image, prior burst SR methods are trained in a deterministic manner, which produces a blurry SR image. Since such blurry images are perceptually degraded, we aim to reconstruct sharp and high-fidelity SR images by a diffusion model. Our method improves the efficiency of the diffusion model with a stochastic sampler with a high-order ODE as well as one-step diffusion using knowledge distillation. Our experimental results demonstrate that our method can reduce the runtime to 1.6 % of its baseline while maintaining the SR quality measured based on image distortion and perceptual quality.