Unsupervised Real-World Super-Resolution via Rectified Flow Degradation Modelling

2025-08-10

AI Summary

Unsupervised real-world super-resolution (SR) suffers from severe domain shift between synthetic and realistic degradations, limiting model generalization. To address this, we propose a corrective flow-based unsupervised degradation modeling framework. First, we design an invertible corrective flow degradation module to explicitly model continuous, reversible real-world degradation trajectories. Second, we introduce a Fourier prior-guided module that leverages phase structural priors to enhance degradation realism. Finally, we synthesize realistic low-resolution/high-resolution (LR/HR) image pairs from unpaired LR and HR images, enabling unsupervised domain adaptation and paired training. Extensive experiments on multiple real-world benchmarks demonstrate substantial improvements in both reconstruction fidelity and perceptual realism. Our method effectively bridges the domain gap between synthetic and real data, establishing a novel paradigm for unsupervised SR.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Unsupervised real-world super-resolution (SR) faces critical challenges due to the complex, unknown degradation distributions in practical scenarios. Existing methods struggle to generalize from synthetic low-resolution (LR) and high-resolution (HR) image pairs to real-world data due to a significant domain gap. In this paper, we propose an unsupervised real-world SR method based on rectified flow to effectively capture and model real-world degradation, synthesizing LR-HR training pairs with realistic degradation. Specifically, given unpaired LR and HR images, we propose a novel Rectified Flow Degradation Module (RFDM) that introduces degradation-transformed LR (DT-LR) images as intermediaries. By modeling the degradation trajectory in a continuous and invertible manner, RFDM better captures real-world degradation and enhances the realism of generated LR images. Additionally, we propose a Fourier Prior Guided Degradation Module (FGDM) that leverages structural information embedded in Fourier phase components to ensure more precise modeling of real-world degradation. Finally, the LR images are processed by both FGDM and RFDM, producing final synthetic LR images with real-world degradation. The synthetic LR images are paired with the given HR images to train the off-the-shelf SR networks. Extensive experiments on real-world datasets demonstrate that our method significantly enhances the performance of existing SR approaches in real-world scenarios.

Problem

• Model real-world degradation for unsupervised super-resolution
• Bridge domain gap between synthetic and real-world image pairs
• Enhance realism of generated low-resolution images

Innovation

• Rectified Flow Degradation Module models degradation
• Fourier Prior Guided Degradation Module enhances precision
• Synthetic LR-HR pairs train SR networks effectively