Table 5 Model performance table (statistical measures and information about the models).
From: CAD-PsorNet: deep transfer learning for computer-assisted diagnosis of skin psoriasis
Transfer Learning Models | Hyper-parameters | Precision (%) | Specificity (%) | Sensitivity (%) | AUC (%) | Accuracy (%) |
|---|---|---|---|---|---|---|
VGG16 | Optimizer: Adam, Learning rate:0.08%, Loss function- Binary cross-entropy | 78.22 | 84.73 | 78.59 | 91.63 | 89.43 |
VGG19 | Optimizer: Adam, learning rate-0.08%, Loss function- Binary cross-entropy | 66.20 | 85.85 | 75.13 | 92.57 | 90.67 |
MobileNetV1 | Optimizer: Adam, learning rate-0.08%, Loss function- Binary cross-entropy | 96.87 | 89.37 | 94.84 | 98.24 | 97.24 |
ResNet50 | Optimizer: Adam, learning rate-0.08%, Loss function- Binary cross-entropy | 30.78 | 65.36 | 83.66 | 88.69 | 86.96 |
Fine-Tuned MobileNetV1 based methodology | Optimizer: Adam and AdaGrad, Learning rate:0.08%, Loss function- Binary cross-entropy | 94.15 | 96.50 | 93.65 | 99.32 | 98.02 |
95.51 | 96.42 | 94.25 | 99.43 | 99.13 |
