Table 2.
Ablation study on ILN-GNet for PD Detection using Dataset 1.
| Metrics | ILN-GNet | Model with conventional LinkNet | Model with conventional LinkNet and Ghostnet | Model with conventional PHOG | Model with conventional WF | Model without Feature Extraction | Model without MDSCM | Model without WAP-BN |
|---|---|---|---|---|---|---|---|---|
| Accuracy | 96.40% | 93.10% | 95.20% | 90.30% | 86.30% | 92.30% | 89.30% | 90.80% |
| Sensitivity | 95.70% | 92.20% | 95.70% | 89.60% | 84.30% | 93.90% | 89.10% | 91.50% |
| Specificity | 97.00% | 94.00% | 94.70% | 91.00% | 88.00% | 91.00% | 89.50% | 90.20% |
| Precision | 96.50% | 93.00% | 94.00% | 89.60% | 85.80% | 90.00% | 87.90% | 89.00% |
| F-measure | 96.10% | 92.60% | 94.80% | 89.60% | 85.10% | 91.90% | 88.50% | 90.20% |
| MCC | 92.70% | 86.20% | 90.30% | 80.50% | 72.40% | 84.70% | 78.60% | 81.60% |
| NPV | 96.30% | 93.30% | 96.20% | 91.00% | 86.70% | 94.50% | 90.60% | 92.60% |
| FPR | 3.00% | 6.00% | 5.30% | 9.00% | 12.00% | 9.00% | 10.50% | 9.80% |
| FNR | 4.30% | 7.80% | 4.30% | 10.40% | 15.70% | 6.10% | 10.90% | 8.50% |
| FDR | 3.50% | 7.00% | 6.00% | 10.40% | 14.20% | 10.00% | 12.10% | 11.00% |