Speech-driven 3D talking head generation on EmbodiedHead speaking (test)
5.7LVEEmbodiedHead (Stage 1)
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| EmbodiedHead (Stage 1)Training Stage=Stage 1, Rendering engine=GAGAvatar [7]2026.04 | 5.7 | 1.66 | 2.98 | 47 | 2.64 | 17.53 | 60 | 0.199 | |
| EmbodiedHead (Stage 2)Training Stage=Stage 2, Rendering engine=GAGAvatar [7]2026.04 | 5.71 | 1.58 | 3.04 | 46 | 2.64 | 18.28 | 61.7 | 0.184 | |
| ARTalkRendering engine=GAGAvatar [7]2026.04 | 6.86 | 2.42 | 3.21 | 46 | 2.39 | 16.81 | 57.9 | 0.221 | |
| DiffPoseTalkRendering engine=GAGAvatar [7]2026.04 | 10.3 | 2.69 | 4.24 | 44 | 2.01 | 16.09 | 56.5 | 0.268 |