Table 12.
Comparison between accuracies per class and over all accuracy of different models applied on MosMed dataset
| Accuracy/class model | CT-0 (%) | CT-1 (%) | CT-2 (%) | CT-3–4 (%) | Overall (%) |
|---|---|---|---|---|---|
| Spatial articulated feature (HOG + DAISY) fusion + SVM | 53.36 | 95.92 | 94.85 | 93.52 | 85.57 |
| Inception-v3 | 87.79 | 95.01 | 80.76 | 79.63 | 91.13 |
| MobileNet-v2 | 89.51 | 96.53 | 85.22 | 84.26 | 93.14 |
| ResNet-50 | 89.51 | 96.53 | 87.63 | 82.41 | 93.33 |
| Xception | 92.80 | 97.38 | 85.22 | 89.81 | 94.67 |
| The proposed modified Norm-VGG16 | 96.08 | 98.97 | 91.07 | 90.74 | 97.09 |
| The proposed COV-CAF model | 97.65 | 98.36 | 95.53 | 95.37 | 97.76 |