Title: SRZoo: An integrated repository for super-resolution using deep learning

Abstract: Deep learning-based image processing algorithms, including image super-resolution methods, have been proposed with significant improvement in performance in recent years. However, their implementations and evaluations are dispersed in terms of various deep learning frameworks and various evaluation criteria. In this paper, we propose an integrated repository for the super-resolution tasks, named SRZoo, to provide state-of-the-art super-resolution models in a single place. Our repository offers not only converted versions of existing pre-trained models, but also documentation and toolkits for converting other models. In addition, SRZoo provides platform-agnostic image reconstruction tools to obtain super-resolved images and evaluate the performance in place. It also brings the opportunity of extension to advanced image-based researches and other image processing models. The software, documentation, and pre-trained models are publicly available on GitHub.