Table 3.
Model | Sensitivity | Specificity | Accuracy | AUC |
---|---|---|---|---|
Baseline | 74.6% | 80.4% | 76.4% | 0.854 |
Baseline + Augmentation | 80.2% | 81.4% | 78.8% | 0.884 |
VGG16 | 75.3% | 82.6% | 76.5% | 0.862 |
Baseline + Nucleus U-net | 75.4% | 86.6% | 77.7% | 0.887 |
Baseline + Mitosis U-net | 74.2% | 86.9% | 79.9% | 0.882 |
Baseline + Epithelium U-net | 80.0% | 82.3% | 79.3% | 0.895 |
Baseline + Tubule U-net | 76.1% | 86.9% | 76.1% | 0.870 |
ConcatNet (+all U-nets) | 82.0% | 87.8% | 84.1% | 0.924 |