KXNet: A Model-Driven Deep Neural Network for Blind Super-Resolution

TL;DR

KXNet is a model-driven deep neural network for blind super-resolution that explicitly incorporates the physical relationship between blur kernels and high-resolution images, leading to improved accuracy and generality.

Contribution

The paper introduces KXNet, a novel blind SISR method that embeds the physical generation mechanism into a deep network via iterative algorithm unfolding.

Findings

Outperforms current state-of-the-art blind SISR methods.

Demonstrates superior accuracy on synthetic and real data.

Shows strong generality across different datasets.

Abstract

Although current deep learning-based methods have gained promising performance in the blind single image super-resolution (SISR) task, most of them mainly focus on heuristically constructing diverse network architectures and put less emphasis on the explicit embedding of the physical generation mechanism between blur kernels and high-resolution (HR) images. To alleviate this issue, we propose a model-driven deep neural network, called KXNet, for blind SISR. Specifically, to solve the classical SISR model, we propose a simple-yet-effective iterative algorithm. Then by unfolding the involved iterative steps into the corresponding network module, we naturally construct the KXNet. The main specificity of the proposed KXNet is that the entire learning process is fully and explicitly integrated with the inherent physical mechanism underlying this SISR task. Thus, the learned blur kernel has clear physical patterns and the mutually iterative process between blur kernel and HR image can soundly guide the KXNet to be evolved in the right direction. Extensive experiments on synthetic and real data finely demonstrate the superior accuracy and generality of our method beyond the current representative state-of-the-art blind SISR methods. Code is available at: https://github.com/jiahong-fu/KXNet.