Skip to main content
. 2020 Nov 10;24(12):3595–3605. doi: 10.1109/JBHI.2020.3037127

TABLE V. Results of COVID-19 Prediction Using Retrained COVIDNet-CXR A, Retrained COVID-CAPS, ResNet-50 With and Without Segmentation, FuCiTNet and COVID-SDNet. All Four Levels of Severity in the Positive Class are Taken Into Account.

Class N P Accuracy
Metric Specificity Precision F1 Sensitivity Precision F1
COVIDNet-CXR 88.82Inline graphic0.90 3.36Inline graphic6.15 73.31Inline graphic3.79 46.82Inline graphic17.59 81.65Inline graphic6.02 56.94Inline graphic15.05 67.82Inline graphic6.11
COVID-CAPS 65.74Inline graphic9.93 65.62Inline graphic3.98 65.15Inline graphic5.02 64.93Inline graphic9.71 66.07Inline graphic4.49 64.87Inline graphic4.92 65.34Inline graphic3.26
Without seg. 79.87Inline graphic8.91 71.91Inline graphic3.12 75.40Inline graphic4.91 68.63Inline graphic6.08 78.75Inline graphic6.31 72.689Inline graphic3.45 74.25Inline graphic3.61
With seg. 78.41Inline graphic7.09 73.36Inline graphic4.66 75.46Inline graphic2.97 70.80Inline graphic8.26 77.17Inline graphic4.79 73.40Inline graphic4.01 74.60Inline graphic2.93
FuCiTNet 80.79Inline graphic6.98 72.00Inline graphic4.48 75.84Inline graphic3.18 67.90Inline graphic8.58 78.48Inline graphic4.99 72.35Inline graphic4.76 74.35Inline graphic3.34
COVID-SDNet
79.76Inline graphic6.19 74.74Inline graphic3.89 76.94Inline graphic2.82 72.59Inline graphic6.77 78.67Inline graphic4.70 75.71Inline graphic3.35 76.18Inline graphic2.70