Table 7 Comparison experiments under data imbalance.
From: A framework for hardware trojan detection based on contrastive learning
| Â | Method | TPR % | TNR % | FPR % | FNR % | Accuracy % | Precision % | F1 Score |
|---|---|---|---|---|---|---|---|---|
T100-2100 | DCNN-ResNet | 22.00 | 82.00 | 0.18 | 0.78 | 52.00 | 55.00 | 0.31 |
Our Proposal | 42.00 | 98.00 | 0.02 | 0.58 | 70.00 | 95.45 | 0.58 | |
T2100-100 | DCNN-ResNet | 70.00 | 30.00 | 0.70 | 0.30 | 50.00 | 50.00 | 0.58 |
Our Proposal | 56.00 | 78.00 | 0.22 | 0.44 | 67.00 | 71.79 | 0.63 |
