Table 5.
Comparison of the proposed method with previous state-of-the-art studies.
| Previous Study | Method | Accuracy (%) |
|---|---|---|
| Wang et al. [68] | DL | 92.30 |
| Gupta et al. [69] | InstaCovNet-19 | 99.08 |
| Afshar et al. [70] | Capsule Network | 95.70 |
| Islam et al. [30] | LSTM & CNN | 99.40 |
| Chowdhury et al. [26] | TL | 97.94 |
| Sethy et al. [28] | ResNet50 - SVM | 95.40 |
| Farooq et al. [71] | TL | 96.20 |
| Das et al. [72] | Xception | 97.40 |
| Ucar et al. [73] | Bayes - SqueezeNet | 98.30 |
| Ismael et al. [74] | DL & SVM | 94.70 |
| Apostolopoulos et al. [75] | TL | 93.48 |
| Hemdan et al. [76] | VGG19 | 90.00 |
| Xu et al. [77] | ResNet - Location Attention | 86.70 |
| Brunese et al. [78] | TL | 97.00 |
| Ozturk et al. [79] | DarkCovidNet | 87.02 |
| Narin et al. [25] | TL | 98.00 |
| Rahimzadeh et al. [80] | Xception & ResNet50V2 | 91.40 |
| Asif et al. [24] | TL | 96.00 |
| Nour et al. [27] | DL & SVM | 98.97 |
| Khan et al. [23] | TL | 95.00 |
| Wang et al. [81]. | ResNet - Feature Pyramid Network | 94.00 |
| Proposed Method | DL & SVM | 99.80 |