Video-driven Talking Head Generation (Self-Reenactment) on HDTF
18.12FIDIM-Portrait
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| IM-PortraitView Point=Random within ±5°2025.04 | 18.12 | — | — | — | — | |
| MGGTalk2025.04 | 18.95 | 32.4 | 0.786 | 0.102 | 0.129 | |
| Ours 2D + DepthView Point=Random within ±5°2025.04 | 19.72 | — | — | — | — | |
| Face-V2VView Point=Random within ±5°2025.04 | 22.27 | — | — | — | — | |
| EmoPortraitView Point=Random within ±5°2025.04 | 27.71 | — | — | — | — | |
| Portrait4d-v2View Point=Random within ±5°2025.04 | 27.83 | — | — | — | — | |
| DaGAN2025.04 | 33.23 | 30.94 | 0.729 | 0.106 | 0.172 | |
| Real3DPortrait2025.04 | 33.26 | 31.62 | 0.702 | 0.145 | 0.163 | |
| OTAvatar2025.04 | 36.47 | 30.67 | 0.654 | 0.133 | 0.226 | |
| Portrait4D-v22025.04 | 36.57 | 30.12 | 0.79 | 0.111 | 0.158 | |
| Styleheat2025.04 | 61.92 | 30.23 | 0.664 | 0.209 | 0.245 | |
| ROME2025.04 | 76.44 | 30.99 | 0.775 | 0.135 | 0.142 |