Black-box DNN Watermarking on CIFAR-100
75.87CDAClean Model
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Clean ModelBackbone=ResNet182026.02 | 75.87 | — | — | |
| ADIBackbone=ResNet182026.02 | 75.7 | 100 | 0.0436 | |
| AFSBackbone=ResNet182026.02 | 75.62 | 100 | 0.2155 | |
| SEW-PreBackbone=ResNet18, Enhancement=Pre-specificity2026.02 | 75.51 | 100 | 0.3372 | |
| ISSBABackbone=ResNet182026.02 | 75.35 | 100 | 0.0363 | |
| SEW-PostBackbone=ResNet18, Enhancement=Post-specificity2026.02 | 75.34 | 100 | 0.0342 | |
| EWBackbone=ResNet182026.02 | 75.22 | 100 | 0.0929 | |
| WESBackbone=ResNet182026.02 | 75.11 | 100 | 0.0358 | |
| ContentBackbone=ResNet182026.02 | 74.32 | 100 | 0.3184 | |
| EWEBackbone=ResNet182026.02 | 74.19 | 100 | 0.3282 | |
| NoiseBackbone=ResNet182026.02 | 73.85 | 100 | 0.3717 | |
| UnrelatedBackbone=ResNet182026.02 | 73.85 | 100 | 0.1567 | |
| MWBackbone=ResNet182026.02 | 70.36 | 100 | 0.1044 |