Deep Super-Resolution Network for Single Image Super-Resolution with Realistic Degradations

TL;DR

This paper introduces a deep CNN-based super-resolution network designed to handle realistic degradations like blurring and noise, improving the quality of high-resolution images from low-resolution inputs in practical scenarios.

Contribution

The proposed network effectively models diverse real-world degradations, advancing super-resolution techniques beyond the traditional bicubic assumption.

Findings

Outperforms existing methods on synthetic datasets with realistic degradations

Produces higher quality HR images from real low-resolution images

Computationally efficient for practical applications

Abstract

Single Image Super-Resolution (SISR) aims to generate a high-resolution (HR) image of a given low-resolution (LR) image. The most of existing convolutional neural network (CNN) based SISR methods usually take an assumption that a LR image is only bicubicly down-sampled version of an HR image. However, the true degradation (i.e. the LR image is a bicubicly downsampled, blurred and noisy version of an HR image) of a LR image goes beyond the widely used bicubic assumption, which makes the SISR problem highly ill-posed nature of inverse problems. To address this issue, we propose a deep SISR network that works for blur kernels of different sizes, and different noise levels in an unified residual CNN-based denoiser network, which significantly improves a practical CNN-based super-resolver for real applications. Extensive experimental results on synthetic LR datasets and real images demonstrate that our proposed method not only can produce better results on more realistic degradation but also computational efficient to practical SISR applications.