Skip to main content

Table 8 Mean ± standard deviation of the results obtained in the test stage for the classification of chest X-ray images between Normal VS COVID-19 after 5 independent repetitions

From: Does imbalance in chest X-ray datasets produce biased deep learning approaches for COVID-19 screening?

Exp.

Class

Precision

Recall

F1-Score

<40

Normal

0.9453 ± 0.0514

0.9749 ± 0.0349

0.9588 ± 0.0277

 

COVID-19

0.9742 ± 0.0355

0.9438 ± 0.0588

0.9573 ± 0.0276

40-50

Normal

0.9673 ± 0.0288

0.9814 ± 0.0141

0.9742 ± 0.0193

 

COVID-19

0.9816 ± 0.0125

0.9693 ± 0.0250

0.9753 ± 0.0156

50-60

Normal

0.9911 ± 0.0057

0.9846 ± 0.0080

0.9878 ± 0.0053

 

COVID-19

0.9844 ± 0.0066

0.9907 ± 0.0058

0.9875 ± 0.0043

60-70

Normal

0.9925 ± 0.0055

0.9826 ± 0.0062

0.9875 ± 0.0053

 

COVID-19

0.9828 ± 0.0064

0.9926 ± 0.0055

0.9877 ± 0.0054

70-80

Normal

0.9780 ± 0.0171

0.9846 ± 0.0086

0.9812 ± 0.0102

 

COVID-19

0.9832 ± 0.0111

0.9774 ± 0.0173

0.9802 ± 0.0115

≥ 80

Normal

0.9373 ± 0.0691

0.8912 ± 0.0849

0.9125 ± 0.0690

 

COVID-19

0.8859 ± 0.0805

0.9255 ± 0.0800

0.9043 ± 0.0734