Table 4.
Performance comparison of different models for detecting COVID-19 on the target dataset.
| Model/Method | Evaluation metrics | |||
|---|---|---|---|---|
| Accuracy | Recall | Specificity | F1 | |
| AlexNet | 74.5 | 70.4 | 79.0 | 75.0 |
| DenseNet-121 | 88.9 | 88.8 | 88.9 | 88.2 |
| DenseNet-169 | 91.2 | 93.3 | 88.9 | 90.8 |
| DenseNet-201 | 91.7 | 88.6 | 94.1 | 91.9 |
| GoogleNet | 78.9 | 75.9 | 82.3 | 79.0 |
| Inception-ResNet-v2 | 86.3 | 88.1 | 84.2 | 87.0 |
| Inception-v3 | 89.4 | 90.0 | 88.9 | 88.8 |
| MobileNet-v2 | 87.2 | 93.2 | 77.6 | 89.0 |
| NasNet-large | 85.2 | 79.3 | 91.9 | 84.0 |
| NasNet-Mobile | 83.4 | 84.8 | 81.9 | 85.0 |
| ResNet-101 | 89.7 | 82.2 | 89.2 | 89.0 |
| ResNet-18 | 90.1 | 89.4 | 90.9 | 91.0 |
| ResNet-50 | 90.8 | 90.0 | 91.0 | 90.1 |
| ResNeXt-101 | 90.9 | 93.1 | 88.9 | 90.6 |
| ResNeXt-50 | 90.6 | 93.4 | 88.2 | 90.3 |
| ShuffleNet | 86.1 | 83.5 | 89.0 | 86.0 |
| SqueezeNet | 78.5 | 86.5 | 63.8 | 82.0 |
| VGG-16 | 78.5 | 74.6 | 82.8 | 76.0 |
| VGG-19 | 83.2 | 90.7 | 74.7 | 85.0 |
| Xception | 85.6 | 88.3 | 80.6 | 87.7 |
|
| ||||
| Contrastive learning [35] | 78.6 | 78.0 | 77.0 | 78.8 |
| Decision function [72] | 88.3 | 87.0 | 87.9 | 86.7 |
| DenseNet-121 + SVM [4] | 85.9 | 84.9 | 86.8 | 86.2 |
| DenseNet-169-based [11] | 83.0 | 84.8 | 85.5 | 81.0 |
| DenseNet-169-based [76] | 87.7 | 85.6 | 86.9 | 87.8 |
| ResNet-101-based [71] | 80.3 | 85.7 | 86.0 | 81.8 |
|
| ||||
| ADA-COVID (without training) | 92.5 | 93.5 | 94.2 | 93.0 |
| ADA-COVID (with training) | 95.8 | 94.9 | 96.0 | 95.2 |