Momentum feature comparison network based on generative adversarial network for single image super-resolution

作者：

Highlights：

• A feature comparison network model based on momentum update is proposed to solve the single image super-resolution problem.

• Using generative adversarial network algorithms to improve the quality of reconstructed details for image super-resolution.

• In the process of network training, only the input low-resolution images are used, which is a completely unsupervised image super-resolution algorithm.

摘要

•A feature comparison network model based on momentum update is proposed to solve the single image super-resolution problem.•Using generative adversarial network algorithms to improve the quality of reconstructed details for image super-resolution.•In the process of network training, only the input low-resolution images are used, which is a completely unsupervised image super-resolution algorithm.

论文关键词：Unsupervised learning,Momentum feature comparison,Generative adversarial networks,Super resolution

论文评审过程：Received 14 October 2021, Revised 22 March 2022, Accepted 18 April 2022, Available online 29 April 2022, Version of Record 13 May 2022.

论文官网地址：https://doi.org/10.1016/j.image.2022.116726