Table 1.

Main papers published in peer-reviewed journals for COVID-19 detection using CXR

Ref	Available code	Algorithms	Performance Index	Sets of images	Number of images per class
[10]	no	-VGG19 -MobileNetv2 -Inception -Xception -Inception ResNet v2	Acc=96.78% Se=98.66% Sp=96.46%	-(Cohen3,RSNA2, Radiopediaa, SIRMb)c -NIH14	224 COVID-19 / 700 bacterial pneumonia / 504 normal 224 COVID-19 / 400 bacterial pneumonia y 314 viral pneumonia / 504 normal
[12]	no	-MobileNetv2, -SqueezeNet -ResNet18 -ResNet101 -DenseNet201 -CheXNet, -Inceptionv3 -VGG19	Acc=99.7% Pr=99.7% Se=99.7% Sp=99.55%	-Cohen3,RSNA2, Radiopediaa, SIRMb	423 COVID-19 / 1485 viral pneumonia / 1579 normal
[64]	no	FrMEM, manta-ray Foraging Optimization, Knn	Acc=96.09% Pr=98.75% Acc=98.09% Pr=98.91%	Dataset 1 -Cohen3, Kaggled Dataset 2 -same set of images used in [12]	216 COVID-19 / 1675 negatives 219 COVID-19 / 1341 negatives
[13]	no	CNN-LSTM combinada	Acc=99.4% AUC=99.9 Se=99.3% Sp=99.2% F1score=98.9%	-(Cohen3, Agchunge,f, Radiopediaa, TCIAg, SIRMb) -Kaggled -NIHh	613 COVID-19 / 1525 pneumonias / 1525 normal
[65]	no	Resne50 Resnet101	Acc=97.77%	Cohen3, Kaggled	440 COVID-19 / 480 viral pneumonia / 457 bacterial pneumonia / 455 normal
[5]	no	SVM RF BPN ANFIS CNN VGGNet ResNet50 Alexnet GoogleNet Inception V3 Xception modificada	Acc=97.4% Fmeausre=96.96% Se=97.09% Sp=97.29% Kappa=97.19%	Same set of images used in [16]
[14]	no	CNN+Knn CNN+DT CNN+SVM	Acc=98.97% Se=89.39% Sp=99.75 Fscore=96.72%	Same set of images used in [12]	219 COVID-19 / 1345 viral pneumonia / 1341 normal
[15]	no	Ensemble Resnet18	Acc=88.9% Pr=83.4% Recall=85.9% F1score=84.4% Sp=96.4% Acc=88.9% Pr=83.4% Recall=85.9% F1score=84.4% Sp=96.4%	Dataset 1 [Cohen3, CoronaHacki, NLC(MC)j, JSRTk] Dataset 2 COVIDxp	180 COVID-19 / 54 bacterial pneumonia / 20 viral pneumonia / 57 tuberculosis 191 normal 180 COVID-19 / 6012 pneumonias / 8851 normal
[16]	yes	DarkCovidNet	Acc=87.02% Se=85.35% Sp=92.18% Pr=89.96% F1score=87.37	Cohen3, ChestX-ray8l	127 COVID-19 / 500 pneumonias / 500 normal
[66]	no	nCOVnet	Acc=88.09% Se=97.62% Sp=78.57%	Cohen3, Fig. 1 Actuale Kaggle4	192 COVID-19 / 5863 negatives
[17]	yes	Feature Extraction LBP, EQP, LDN, LETRIST, BSIF, LPQ, oBIFs, Inception-V3 Classifiers Knn, SVM, MLP, DT, RF	F1score=88.89%	RYDLS-20 [Cohen3, Radiopediaa, Chest X-ray14m]	180 COVID-19 / 20 MERS / 22 SARS / 20 Varicella / 24 Streptococcus / 22 Pneumocystis / 2000 normal
[33]	no	COVID-SDNet	Acc=97.37%	COVIDGR-1.0n	377 COVID-19 / 377 negatives
[59]	yes	MobileNetV2 SqueezeNet SVM	Acc=99.27%	Cohen3, Radiopediaa, Kaggle6	295 COVID-19 / 98 pneumonias / 65 normal
[67]	yes	Inception V3	Binary Acc=100% Se=99.0% Sp=100% AUC=100% Ternary Acc=85% Se=94% Sp=92.7% AUC=96% Quaternary Acc=76% Se=93% Sp=91.8% AUC=93%	Cohen3, RSNA2,Kaggled, Kermanyo	122 COVID-19 / 150 bacterial pneumonias / 150 viral pneumonias / 150 normal
[58]	no	COVIDiagnosis-Net based on SqueezeNet with Bayesian optimization	Acc=98.3% Spe=99.1% F1score=98.3% MCC=97.4%	COVIDxu	76 COVID-19, 4290 pneumonias / 1583 normal
[68]	yes	VGG-19 ResNet-50 COVID-Net	Acc=93.3% Se=91%	COVIDxu (Cohen3, Fig. 1 COVID-19j, ActualMed COVID-19k, RSNA2, COVID-19 radiography database5)	190 COVID-19, 8614 Pneumonia, 8066 normal

^a https://radiopaedia.org/articles/pneumonia

^bhttps://www.sirm.org/en/category/articles/covid-19-database/

^chttps://www.kaggle.com/andrewmvd/convid19-X-rays

^dhttps://www.kaggle.com/paultimothymooney/chest-xray-pneumonia

^ehttps://github.com/agchung/Figure1-COVID-chestxray-dataset

^fhttps://github.com/agchung/Actualmed-COVID-chestxray-dataset

^ghttps://www.cancerimagingarchive.net/

^hhttps://www.kaggle.com/nih-chest-xrays/data?select=Data_Entry_2017.csv

ⁱhttps://www.kaggle.com/praveengovi/coronahack-chest-xraydataset

^jhttp://archive.nlm.nih.gov/repos/chestImages.php

^khttp://db.jsrt.or.jp/eng.php

^lhttps://www.cc.10.nih.gov/drd/summers.html

^mhttps://nihcc.app.box.com/v/ChestXray-NIHCC

ⁿhttps://github.com/ari-dasci/OD-covidgr/releases/tag/1.0

^ohttps://doi.org/10.17632/rscbjbr9sj.3

^phttps://github.com/lindawangg/COVID-Net