Table 6 Statistical analysis of various state-of-art models.
From: ResNet-32 and FastAI for diagnoses of ductal carcinoma from 2D tissue slides
| Â | Sensitivity (%) | Specificity (%) | Accuracy (%) | F1-Score (%) |
|---|---|---|---|---|
Multi-layer perceptron46 | 76.0 | 74.0 | 76.0 | – |
Random Forest28 | 93.0 | 92.6 | 93.3 | 93.0 |
SVM47 | 90.0 | 95.50 | 92.75 | – |
SVM48 | 70.00 | 88.64 | 84.22 | – |
Random Forest48 | 90.44 | 70.30 | 75.55 | – |
Naïve Bayesian48 | 70.98 | 72.50 | 72.14 | – |
CNN49 | 84.70 | – | 84.93 | 76.07 |
AlexNet50 | 84.38 | 82.35 | 87.50 | 84.85 |
V66-1651 | – | – | 79.2 | – |
VGG-1650 | 83.24 | 81.29 | 86.36 | 83.74 |
VGG-1950 | 82.39 | 80.98 | 84.66 | 82.78 |
Inception V3 + SVM51 | – | – | 83.4 | – |
Inception V328 | 80.5 | 82.0 | 79.0 | 81.0 |
Inception V3 + Bi-LSTM51 | – | – | 91.3 | – |
AlexNet46 | 93.6 | 91.7 | 92.7 | – |
AlexNet + SVM52 | 86.2 | 87.7 | 87.2 | – |
Faster RCNN (VGG)53 | 94.67 | 89.69 | 91.68 | – |
GoogLeNet54 | 91.70 | 97.66 | 91.70 | 91.92 |
ResNet-3455 | 89.37 | 81.79 | 90.66 | 84.19 |
VGG1956 | 91.16 | 97.66 | 91.16 | 91.18 |
Multiple interface learning-CNN56 | 94.43 | 77.78 | 88.81 | – |
Deep multiple instance learning-CNN57 | 94.44 | 88.89 | 93.06 | – |
Proposed model | 94.83 | 91.48 | 93.60 | 95.90 |
