Table 7 Statistical analysis performance comparison with recent architectures.
Model | Accuracy (mean ± SD) | Sensitivity (mean ± SD) | Specificity (mean ± SD) | F1 score (mean ± SD) | Precision (mean ± SD) | Recall (sensitivity) | AUC-ROC (mean ± SD) |
|---|---|---|---|---|---|---|---|
VGG 16 | 90.24 ± 1.10 | 87.36 ± 1.15 | 89.28 ± 1.10 | 85.71 ± 1.40 | 84.20 ± 1.25 | 87.36 ± 1.15 | 90.10 ± 1.12 |
VGG 19 | 91.20 ± 0.98 | 91.20 ± 1.08 | 88.32 ± 1.26 | 86.63 ± 1.15 | 84.92 ± 1.10 | 91.20 ± 1.08 | 91.30 ± 1.05 |
Google Net | 91.20 ± 0.95 | 88.32 ± 1.10 | 89.28 ± 1.08 | 87.55 ± 1.20 | 86.85 ± 1.15 | 88.32 ± 1.10 | 90.85 ± 1.10 |
ResNet 50 | 90.24 ± 1.25 | 91.20 ± 0.92 | 89.28 ± 0.95 | 86.63 ± 1.00 | 83.75 ± 1.00 | 91.20 ± 0.92 | 91.00 ± 0.97 |
ResNet 18 | 89.28 ± 1.05 | 87.36 ± 1.14 | 90.24 ± 1.00 | 90.24 ± 0.95 | 89.50 ± 0.85 | 87.36 ± 1.14 | 90.50 ± 0.93 |
Inception V2 | 89.28 ± 1.20 | 91.20 ± 1.02 | 87.36 ± 1.18 | 91.20 ± 1.05 | 91.10 ± 0.92 | 91.20 ± 1.02 | 91.60 ± 0.95 |
BiLSTM-CNN | 95.91 ± 0.67 | 96.82 ± 0.55 | 94.55 ± 0.72 | 93.34 ± 0.61 | 90.23 ± 0.58 | 96.82 ± 0.55 | 97.12 ± 0.45 |
DenseNet121 | 92.84 ± 0.88 | 93.22 ± 0.85 | 91.45 ± 0.92 | 91.34 ± 0.95 | 93.22 ± 0.85 | 92.27 ± 0.79 | 95.20 ± 0.61 |
EfficientNetB0 | 93.51 ± 0.81 | 94.10 ± 0.72 | 92.76 ± 0.88 | 92.83 ± 0.78 | 94.10 ± 0.72 | 93.45 ± 0.66 | 96.14 ± 0.54 |
