Unsupervised Image Super-Resolution using Cycle-in-Cycle Generative Adversarial Networks
This addresses the problem of super-resolution without paired data for applications in image processing, though it is incremental as it builds on existing GAN-based image-to-image translation methods.
The paper tackles unsupervised single image super-resolution where low-/high-resolution pairs and down-sampling processes are unavailable, and inputs are degraded by noise and blur, achieving results comparable to state-of-the-art supervised models on NTIRE2018 datasets.
We consider the single image super-resolution problem in a more general case that the low-/high-resolution pairs and the down-sampling process are unavailable. Different from traditional super-resolution formulation, the low-resolution input is further degraded by noises and blurring. This complicated setting makes supervised learning and accurate kernel estimation impossible. To solve this problem, we resort to unsupervised learning without paired data, inspired by the recent successful image-to-image translation applications. With generative adversarial networks (GAN) as the basic component, we propose a Cycle-in-Cycle network structure to tackle the problem within three steps. First, the noisy and blurry input is mapped to a noise-free low-resolution space. Then the intermediate image is up-sampled with a pre-trained deep model. Finally, we fine-tune the two modules in an end-to-end manner to get the high-resolution output. Experiments on NTIRE2018 datasets demonstrate that the proposed unsupervised method achieves comparable results as the state-of-the-art supervised models.