Single Image Super-resolution via Dense Blended Attention Generative Adversarial Network for Clinical Diagnosis

TL;DR

This paper discusses the limitations of DenseNet in single image super-resolution for medical images, highlighting the need for efficient training and broader application beyond medical imaging.

Contribution

The authors propose an alternative approach to DenseNet-based super-resolution that reduces GPU memory usage and training time, enabling broader application.

Findings

Reduced GPU memory consumption compared to DenseNet

Faster training times for super-resolution models

Potential for application to general images

Abstract

During training phase, more connections (e.g. channel concatenation in last layer of DenseNet) means more occupied GPU memory and lower GPU utilization, requiring more training time. The increase of training time is also not conducive to launch application of SR algorithms. This's why we abandoned DenseNet as basic network. Futhermore, we abandoned this paper due to its limitation only applied on medical images. Please view our lastest work applied on general images at arXiv:1911.03464.