Self-Cascaded Diffusion Models for Arbitrary-Scale Image Super-Resolution

TL;DR

This paper introduces CasArbi, a self-cascaded diffusion model that progressively enhances image resolution for arbitrary-scale super-resolution, outperforming previous methods in quality and flexibility.

Contribution

The paper proposes a novel self-cascaded diffusion framework with coordinate-guided residual diffusion for continuous and flexible arbitrary-scale image super-resolution.

Findings

Outperforms prior methods in perceptual quality

Achieves better distortion metrics across benchmarks

Enables efficient diffusion sampling for continuous scales

Abstract

Arbitrary-scale image super-resolution aims to upsample images to any desired resolution, offering greater flexibility than traditional fixed-scale super-resolution. Recent approaches in this domain utilize regression-based or generative models, but many of them are a single-stage upsampling process, which may be challenging to learn across a wide, continuous distribution of scaling factors. Progressive upsampling strategies have shown promise in mitigating this issue, yet their integration with diffusion models for flexible upscaling remains underexplored. Here, we present CasArbi, a novel self-cascaded diffusion framework for arbitrary-scale image super-resolution. CasArbi meets the varying scaling demands by breaking them down into smaller sequential factors and progressively enhancing the image resolution at each step with seamless transitions for arbitrary scales. Our novel coordinate-guided residual diffusion model allows for the learning of continuous image representations while enabling efficient diffusion sampling. Extensive experiments demonstrate that our CasArbi outperforms prior arts in both perceptual and distortion performance metrics across diverse arbitrary-scale super-resolution benchmarks.