Table 1.

List of related methods of COVID-19 detection from CXR images with pros/ cons.

Author	Year	Methods with advantages	Drawbacks
Wang et al. [22]	2020	Simple CNN based model in deep learning framework	Low accuracy and applied on limited datasets
Purohit et al. [23]	2020	Multi-image augmented deep learning model; synthesis images was created for low size of datasets	Average accuracy and low F1-score
Karim et al. [24]	2020	DeepCOVIDExplainer: it was a multi-class neural ensemble model using gradient-guided class activation maps.	Applied on imbalanced datasets and produced average accuracy.
Farooq et al. [25]	2020	Covid-ResNet: a residual neural network based deep learning model for COVID-19 detection.	It was a pre-trained model with fixed size of images and not applicable on real time datasets.
Khan et al. [26]	2020	CoroNet: it was a 4-class classifier Xception architecture pre-trained on ImageNet dataset and trained end-to-end on a dataset prepared.	It was applied on limited dataset and model suffer overfitting problem in large dataset.
Ozturk et al. [27]	2020	DarkNet: It was used in real time COVID detection for both binary and multi-class problem	Though it achieved good accuracy in case of binary class problem but less accuracy in multiclass problem.
Wang et al. [28]	2020	COVID-Net: It was a multi-class classification model with normal, pneumonia, and COVID-19 applied on real time datasets.	The model suffers in sensitivity and positive predictive value (PPV).
S. Albahli [29]	2020	GAN based COVID detection with synthetic data generator.	Sometimes model detect viral pneumonia as COVID and suffers in low positive predictive value (PPV).