Automated multi-class classification of lung diseases from CXR-images using pre-trained convolutional neural networks
作者：
Highlights：
• Efficient five-class classification of CXR-images.
• Experimentation on eight pretrained networks.
• Use of a larger dataset and calculation of nine different performance metrics.
• Accuracy of 97.2% in detecting lung disorders. Analysis of misclassified images.
• Deep learning visualization techniques to locate the useful areas in decision-making.
摘要
•Efficient five-class classification of CXR-images.•Experimentation on eight pretrained networks.•Use of a larger dataset and calculation of nine different performance metrics.•Accuracy of 97.2% in detecting lung disorders. Analysis of misclassified images.•Deep learning visualization techniques to locate the useful areas in decision-making.
论文关键词：CXR-images,COVID-19,Pneumonia,Pneumothorax,Tuberculosis
论文评审过程：Received 11 June 2022, Revised 13 August 2022, Accepted 19 August 2022, Available online 27 August 2022, Version of Record 2 September 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2022.118650
