Table 14 Performance analysis of existing and proposed models on MIAS dataset using different optimizers (before data preprocessing).
CNN Classifier | Optimizer | Accuracy (%) | Sensitivity (%) | Specificity (%) | Precision (%) | AUC (%) | F1-score (%) | Cohen’s Kappa (κ) |
|---|---|---|---|---|---|---|---|---|
VGG-16 | Adam | 75.00 | 65.50 | 85.00 | 70.50 | 77 | 67.00 | 0.50 |
RMSProp | 73.50 | 64.00 | 83.50 | 69.00 | 76 | 65.00 | 0.48 | |
SGD | 72.00 | 63.00 | 82.00 | 68.00 | 75 | 64.00 | 0.46 | |
VGG-19 | Adam | 78.00 | 69.00 | 87.00 | 73.00 | 80 | 70.50 | 0.53 |
RMSProp | 76.50 | 67.00 | 85.50 | 71.50 | 78 | 69.00 | 0.51 | |
SGD | 75.00 | 66.00 | 84.50 | 70.50 | 77 | 68.00 | 0.50 | |
DenseNet169 | Adam | 76.50 | 67.50 | 86.50 | 72.00 | 79 | 70.00 | 0.54 |
RMSProp | 75.00 | 66.00 | 85.00 | 71.00 | 78 | 69.00 | 0.52 | |
SGD | 73.50 | 64.50 | 83.50 | 70.00 | 77 | 68.00 | 0.50 | |
ResNet-50 | Adam | 79.00 | 70.00 | 88.00 | 74.00 | 81 | 71.00 | 0.55 |
RMSProp | 77.50 | 68.50 | 87.50 | 73.50 | 79 | 70.50 | 0.53 | |
SGD | 76.00 | 67.00 | 86.50 | 72.00 | 78 | 69.00 | 0.51 | |
DenseNet201 | Adam | 80.00 | 72.00 | 89.00 | 76.00 | 83 | 74.00 | 0.57 |
RMSProp | 78.00 | 70.00 | 88.00 | 74.50 | 81 | 72.00 | 0.55 | |
SGD | 77.00 | 69.00 | 87.50 | 73.00 | 79 | 71.00 | 0.53 | |
Proposed hybrid model | Adam | 91.00 | 83.00 | 92.50 | 88.00 | 94 | 86.50 | 0.72 |
RMSProp | 89.00 | 80.50 | 91.00 | 87.50 | 93 | 85.00 | 0.70 | |
SGD | 87.00 | 78.50 | 89.00 | 85.00 | 92 | 83.00 | 0.68 |
