2020/07/06 by Mohammad Saeed Rad, Rad, Mohammad Saeed, Thomas Yu +9
Computer Science · Engineering · #Advanced Image Processing Techniques #Advanced Vision and Imaging #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #FOS: Electrical engineering #Image Processing Techniques and Applications #Image and Video Processing (eess.IV) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2007.03053
openalex publication_date 2020/07/06 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
Super-resolution (SR) has traditionally been based on pairs of\nhigh-resolution images (HR) and their low-resolution (LR) counterparts obtained\nartificially with bicubic downsampling. However, in real-world SR, there is a\nlarge variety of realistic image degradations and analytically modeling these\nrealistic degradations can prove quite difficult. In this work, we propose to\nhandle real-world SR by splitting this ill-posed problem into two comparatively\nmore well-posed steps. First, we train a network to transform real LR images to\nthe space of bicubically downsampled images in a supervised manner, by using\nboth real LR/HR pairs and synthetic pairs. Second, we take a generic SR network\ntrained on bicubically downsampled images to super-resolve the transformed LR\nimage. The first step of the pipeline addresses the problem by registering the\nlarge variety of degraded images to a common, well understood space of images.\nThe second step then leverages the already impressive performance of SR on\nbicubically downsampled images, sidestepping the issues of end-to-end training\non datasets with many different image degradations. We demonstrate the\neffectiveness of our proposed method by comparing it to recent methods in\nreal-world SR and show that our proposed approach outperforms the\nstate-of-the-art works in terms of both qualitative and quantitative results,\nas well as results of an extensive user study conducted on several real image\ndatasets.\n