Table 9.
Comparing the accuracies of the proposed algorithm and the existing models.
| Models | Accuracy (%) | Architecture | Imaging Type | Number of Images(Covid-19, Non-Covid-19) |
|---|---|---|---|---|
| Jaiswal et al. [27] | 96.25 | DenseNet-201 | CT | (1262, 1230) |
| Hall et al. [28] | 89.2 | ResNet-50 | CXR | (135, 320) |
| Sethy and Behr [29] | 95.38 | ResNet-50 + SVM | CXR | (25, 25) |
| Hemdan et al. [30] | 90.0 | COVIDX-Net | CXR | (25, 25) |
| Xu et al. [32] | 86.7 | ResNet + Location Attention | CT | (219, 224) |
| Apostolopoulos and Mpesiana. [33] | 96.78 | MobileNet V2 | CXR | (224, 700) |
| Hu et al. [37] | 91.21 | ShuffleNet V2 | CT | (521, 397) |
| Zheng et al. [53] | 90.8 | DeCoVNet | CT | (313, 229) |
| Ozturk et al. [54] | 98.08 | DarkCovidNet | CXR | (125, 500) |
| Wang et al. [55] | 82.9 | M−Inception | CT | (195, 258) |
| Our method | 98.50 | DenseNet-201 | CXR | (360, 4200) |