Watermark Radioactivity Detection on VCTK
100AccdetAudioMarkNet
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| AudioMarkNetMadv (Generation Model)=YourTTS, FT (Fine-tuning Strategy)=Full2026.06 | 100 | 100 | 100 | 100 | 100 | 1 | |
| LambdaMarkMadv (Generation Model)=SemanticVocoder, FT (Fine-tuning Strategy)=Full2026.06 | 100 | 97.47 | 100 | 100 | 100 | 1 | |
| LambdaMarkMadv (Generation Model)=SemanticVocoder, FT (Fine-tuning Strategy)=LoRA2026.06 | 100 | 94.16 | 100 | 100 | 100 | 1 | |
| LambdaMarkMadv (Generation Model)=YourTTS, FT (Fine-tuning Strategy)=Full2026.06 | 99 | 92.06 | 100 | 98 | 98.99 | 0.99 | |
| LambdaMarkMadv (Generation Model)=AudioLDM2, FT (Fine-tuning Strategy)=Full2026.06 | 98 | 81.09 | 100 | 96 | 97.96 | 1 | |
| LambdaMarkMadv (Generation Model)=AudioLDM2, FT (Fine-tuning Strategy)=LoRA2026.06 | 91.5 | 70.31 | 100 | 83 | 90.71 | 0.997 | |
| AudioMarkNetMadv (Generation Model)=SemanticVocoder, FT (Fine-tuning Strategy)=Full2026.06 | 50 | 52.69 | 0 | 0 | 0 | 0.5 | |
| AudioMarkNetMadv (Generation Model)=SemanticVocoder, FT (Fine-tuning Strategy)=LoRA2026.06 | 50 | 55.31 | 0 | 0 | 0 | 0.5 | |
| AudioMarkNetMadv (Generation Model)=AudioLDM2, FT (Fine-tuning Strategy)=Full2026.06 | 50 | 67.88 | 0 | 0 | 0 | 0.5 | |
| AudioMarkNetMadv (Generation Model)=AudioLDM2, FT (Fine-tuning Strategy)=LoRA2026.06 | 50 | 50.06 | 0 | 0 | 0 | 0.5 |