Table 3.
Deep learning networks for COVID-19 identification.
| Ref. | Deep learning model | Modality | Total samples | Evaluation metrics | ||
|---|---|---|---|---|---|---|
| Normal | Pneumonia | COVID-19 | ||||
| [80] | COVID-Net | X-ray | 8,066 | 5,521 | 183 | Accuracy 94.3%, precision 90.9%, recall 96.8% |
| [83] | EfficientNet | X-ray | 8,066 | 5,521 | 183 | Accuracy 93.9%, precision 100%, recall 96.8% |
| [84] | NASNet | X-ray | 533 | 515 | 108 | Accuracy 95%, precision 95%, recall 90% |
| [85] | GAN and VGG16 | X-ray | 721 | 0 | 403 | Accuracy 95%, recall 90% |
| [86] | DenseNet103, ResNet18 | X-ray | 191 | 20 | 180 | Accuracy 88.9%, precision 83.4%, recall 85.9% |
| [87] | ResNet101, ResNet152 | X-ray | 8851 | 9576 | 140 | Accuracy 96.1% |
| [88] | DenseNet and graph attention network | X-ray | 10192 | 7399 | 399 | Accuracy 94.1%, precision 94.47%, recall 91.9% |
| [89] | VGG19 | X-ray | 3181 | 0 | 2049 | Accuracy 98.36% |
| [90] | ResNet34, HRNet | X-ray | 400 | 0 | 400 | Accuracy 99.99%, precision 100%, recall 99.9% |
| [91] | VGG16 | CT | 49800 | 23652 | 80800 | Accuracy93.57%, precision 89.40%, recall 94% |
| [92] | Deep long short-term memory network | CT | 547 | 631 | 612 | Accuracy 97.93%, recall 98.18% |
| [93] | VGG19 | Ultrasound | 235 | 277 | 399 | Accuracy 100% |