COVID-19 detection on Chest X-ray images: A comparison of CNN architectures and ensembles

作者：

Highlights：

• This paper compares 21 CNN architectures for COVID-19 detection in CXR images.

• DenseNet169 achieved an accuracy of 98.15% and an F1 score of 98.12%.

• An ensemble of DenseNet169 instances increased the F1 score to 99.24%.

• The training was repeated five times for each model to get more reliable results.

摘要

•This paper compares 21 CNN architectures for COVID-19 detection in CXR images.•DenseNet169 achieved an accuracy of 98.15% and an F1 score of 98.12%.•An ensemble of DenseNet169 instances increased the F1 score to 99.24%.•The training was repeated five times for each model to get more reliable results.

论文关键词：Convolutional neural networks,Transfer learning,Chest X-ray images

论文评审过程：Received 25 November 2021, Revised 6 May 2022, Accepted 7 May 2022, Available online 21 May 2022, Version of Record 26 May 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117549