Table 2 Classification accuracy performance of deep features based on leave-one-out strategy (%)
From: Rapid identification of COVID-19 severity in CT scans through classification of deep features
Backbone | Discriminant | Linear SVM | Cubic SVM | KNN | Boosted trees |
|---|---|---|---|---|---|
Inception-V3 | 78.88 | 86.15 | 92.47 | 93.69 | 85.32 |
ResNet-50 | 80.52 | 89.03 | 93.02 | 92.73 | 87.11 |
ResNet-101 | 78.74 | 90.53 | 93.69 | 93.96 | 89.44 |
DenseNet-201 | 65.95 | 90.53 | 95.34 | 94.24 | 89.57 |