Table 4.
Evaluation of the proposed U-Net–CNN over 10-fold tests.
| Fold | HGG | LGG | ||||
|---|---|---|---|---|---|---|
| Dice Score (%) | Sensitivity (%) | Classification Accuracy (%) | Dice Score (%) | Sensitivity (%) | Classification Accuracy (%) | |
| Fold 1 | 87.16% | 88.61% | 85.54% | 80.5% | 77.02% | 80.04% |
| Fold 2 | 89.1% | 88.92% | 88.76% | 85.3% | 79.5% | 79.96% |
| Fold 3 | 87.98% | 89.8% | 89.03% | 83.6% | 80.64% | 80.2% |
| Fold 4 | 88.74% | 89.53% | 89.2% | 82.1% | 81.2% | 81.4% |
| Fold 5 | 88.95% | 88.76% | 88.12% | 85.84% | 79.14 | 83.5% |
| Fold 6 | 88.97% | 88.81% | 89.98% | 86.23% | 80.65% | 84.6% |
| Fold 7 | 88.47% | 89.5% | 88.73% | 86.41% | 80.12% | 86.73% |
| Fold 8 | 88.62% | 90.14% | 88.91% | 85.33% | 80.79% | 84.1% |
| Fold 9 | 89.68% | 89.94% | 87.66% | 85.98% | 80.2% | 85.9% |
| Fold 10 | 89.83% | 89.97% | 90.22% | 85.96% | 80.74% | 85.3% |
| Avg. (%) | 88.8 | 89.4 | 88.6 | 84.7 | 80 | 83.1 |