Unsupervised Learning for Real-World Super-Resolution

TL;DR

This paper introduces an unsupervised super-resolution method that learns to invert bicubic downsampling, enabling training on unpaired real-world images and improving generalization to natural images.

Contribution

It proposes a novel unsupervised approach that restores natural image characteristics, allowing super-resolution training without paired data and better real-world applicability.

Findings

Outperforms bicubic-based methods on real-world images

Generates realistic high-resolution images from unpaired data

Shows strong generalization to natural image distributions

Abstract

Most current super-resolution methods rely on low and high resolution image pairs to train a network in a fully supervised manner. However, such image pairs are not available in real-world applications. Instead of directly addressing this problem, most works employ the popular bicubic downsampling strategy to artificially generate a corresponding low resolution image. Unfortunately, this strategy introduces significant artifacts, removing natural sensor noise and other real-world characteristics. Super-resolution networks trained on such bicubic images therefore struggle to generalize to natural images. In this work, we propose an unsupervised approach for image super-resolution. Given only unpaired data, we learn to invert the effects of bicubic downsampling in order to restore the natural image characteristics present in the data. This allows us to generate realistic image pairs, faithfully reflecting the distribution of real-world images. Our super-resolution network can therefore be trained with direct pixel-wise supervision in the high resolution domain, while robustly generalizing to real input. We demonstrate the effectiveness of our approach in quantitative and qualitative experiments.