Table 2.
Performance results for the testing dataset (10 patient studies) obtained as the average of the results computed from each Net trained in the five folds, and using the region growing algorithm.
| Testing Dataset (10 Patients’ Studies) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| E-Net | U-Net | Region Growing | |||||||
| Mean | ±std | ±CI (95%) | Mean | ±std | ±CI (95%) | Mean | ±std | ±CI (95%) | |
| Sensitivity | 93.56% | 3.41% | 0.95% | 92.40% | 4.20% | 2.60% | 98.58% | 11.18% | 2.57% |
| PPV | 98.44% | 0.95% | 0.26% | 99.00% | 0.82% | 0.51% | 77.71% | 9.47% | 2.21% |
| DSC | 95.90% | 1.56% | 0.43% | 95.61% | 1.82% | 1.13% | 86.91% | 8.94% | 2.03% |
| VOE | 7.84% | 2.91% | 0.81% | 8.36% | 3.40% | 2.11% | 23.15% | 11.55% | 2.67% |
| VD | −4.89% | 3.82% | 1.06% | −6.71% | 4.43% | 2.75% | 26.85% | 14.42% | 3.33% |
| ASSD | 2.31 | 0.70 | 0.43 | 2.26 | 0.60 | 0.37 | 9.72 | 4.51 | 1.07 |