Audio-Visual Speech Separation on VoxCeleb2 Scenario 2 (1 speaker + FMA music noise)
7.49SI-SNRiDolphin
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Dolphin2025.09 | 7.49 | 8.52 | 1.85 | |
| AV-Mossformer22025.09 | 6.68 | 7.53 | 1.76 | |
| IIANet2025.09 | 2.04 | 5.92 | 1.64 |
Loading the SOTA2 catalog…