A Fully Progressive Approach to Single-Image Super-Resolution

TL;DR

This paper introduces a progressive deep learning framework for single-image super-resolution that improves high upsampling factor results and speed, using curriculum learning and multi-scale GANs.

Contribution

The paper presents ProSR, a novel progressive architecture and training method, along with ProGanSR, a multi-scale GAN for photorealistic super-resolution at large scales.

Findings

ProSR achieves high-quality results for large upsampling factors.

ProGanSR produces photorealistic images with multi-scale GAN design.

Our model is faster and competitive in SSIM and PSNR metrics.

Abstract

Recent deep learning approaches to single image super-resolution have achieved impressive results in terms of traditional error measures and perceptual quality. However, in each case it remains challenging to achieve high quality results for large upsampling factors. To this end, we propose a method (ProSR) that is progressive both in architecture and training: the network upsamples an image in intermediate steps, while the learning process is organized from easy to hard, as is done in curriculum learning. To obtain more photorealistic results, we design a generative adversarial network (GAN), named ProGanSR, that follows the same progressive multi-scale design principle. This not only allows to scale well to high upsampling factors (e.g., 8x) but constitutes a principled multi-scale approach that increases the reconstruction quality for all upsampling factors simultaneously. In particular ProSR ranks 2nd in terms of SSIM and 4th in terms of PSNR in the NTIRE2018 SISR challenge [34]. Compared to the top-ranking team, our model is marginally lower, but runs 5 times faster.