Infrared Image Super-Resolution via Heterogeneous Convolutional WGAN

TL;DR

This paper introduces HetSRWGAN, a lightweight deep learning framework utilizing heterogeneous convolution and adversarial training to enhance infrared image resolution, addressing the unique challenges of IR images with fewer patterns.

Contribution

It proposes a novel heterogeneous kernel-based residual block and a gradient-based loss function for IR image super-resolution, improving stability and performance.

Findings

Achieves better qualitative and quantitative results

Provides more stable training process

Outperforms existing methods in IR super-resolution

Abstract

Image super-resolution is important in many fields, such as surveillance and remote sensing. However, infrared (IR) images normally have low resolution since the optical equipment is relatively expensive. Recently, deep learning methods have dominated image super-resolution and achieved remarkable performance on visible images; however, IR images have received less attention. IR images have fewer patterns, and hence, it is difficult for deep neural networks (DNNs) to learn diverse features from IR images. In this paper, we present a framework that employs heterogeneous convolution and adversarial training, namely, heterogeneous kernel-based super-resolution Wasserstein GAN (HetSRWGAN), for IR image super-resolution. The HetSRWGAN algorithm is a lightweight GAN architecture that applies a plug-and-play heterogeneous kernel-based residual block. Moreover, a novel loss function that employs image gradients is adopted, which can be applied to an arbitrary model. The proposed HetSRWGAN achieves consistently better performance in both qualitative and quantitative evaluations. According to the experimental results, the whole training process is more stable.