Audio-Visual Target Speaker Extraction on AVSEC3 (dev)
17.5SISDRSAV-GridNet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SAV-GridNet2025.09 | 17.5 | 3.23 | 95 | |
| Lip-Face2025.09 | 17.5 | 3.08 | 95 | |
| AV-CrossNet2025.09 | 17.3 | 3.23 | 95 | |
| Lip-Face2025.09 | 14.4 | 2.57 | 91 | |
| AV-CrossNet2025.09 | 14.3 | 2.75 | 92 | |
| SAV-GridNet2025.09 | 14.2 | 2.68 | 91 | |
| Mix2025.09 | -4.4 | 1.15 | 68 | |
| Mix2025.09 | -5 | 1.17 | 60 |