Zoomed In, Diffused Out: Towards Local Degradation-Aware Multi-Diffusion for Extreme Image Super-Resolution

TL;DR

This paper introduces a novel multi-diffusion approach with local degradation-aware prompts that enables large-scale, high-resolution image super-resolution using pre-trained Text-to-Image diffusion models without additional training.

Contribution

The paper presents a new multi-diffusion method combined with local degradation-aware prompts to achieve ultra-high-resolution image super-resolution beyond the original training limits.

Findings

Enables 2K, 4K, and 8K image generation without retraining.

Ensures global coherence across large images.

Improves local detail reconstruction in super-resolution tasks.

Abstract

Large-scale, pre-trained Text-to-Image (T2I) diffusion models have gained significant popularity in image generation tasks and have shown unexpected potential in image Super-Resolution (SR). However, most existing T2I diffusion models are trained with a resolution limit of 512x512, making scaling beyond this resolution an unresolved but necessary challenge for image SR. In this work, we introduce a novel approach that, for the first time, enables these models to generate 2K, 4K, and even 8K images without any additional training. Our method leverages MultiDiffusion, which distributes the generation across multiple diffusion paths to ensure global coherence at larger scales, and local degradation-aware prompt extraction, which guides the T2I model to reconstruct fine local structures according to its low-resolution input. These innovations unlock higher resolutions, allowing T2I diffusion models to be applied to image SR tasks without limitation on resolution.