Audio-driven talking face generation on VoxCeleb2
8.841ScVASA-1
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VASA-1Training data scale=Full data2024.04 | 8.841 | 6.312 | 0.468 | 0.304 | |
| VASA-1Training data scale=10% data2024.04 | 8.818 | 6.298 | 0.457 | 0.229 | |
| Real video2024.04 | 7.64 | 7.189 | 0.588 | 0.505 | |
| Audio2Head2024.04 | 6.172 | 8.47 | 0.246 | 0.26 | |
| SadTalker2024.04 | 5.843 | 8.813 | 0.441 | 0.275 | |
| MakeItTalk2024.04 | 4.176 | 15.513 | -0.051 | 0.21 |