Speaker Separation on LRS2 (test)
12.71SDRVoiceFormer (Ours A+V)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| VoiceFormer (Ours A+V)Video conditioning=true, Text conditioning=false2025.01 | 12.71 | 92 | 2.06 | |
| Lee et al.Video conditioning=true, Text conditioning=false2025.01 | 10.01 | 88 | 0.94 |