Face Reenactment on VoxCeleb2 (test)
24.92FIDCainGAN
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CainGANK (source frames)=8, Training epochs=202020.04 | 24.92 | 0.77 | 0.91 | — | |
| Ours-FTNumber of frames (T)=32, Evaluation protocol=Fine-tuning2019.05 | 30.6 | 0.72 | 0.45 | 0.33 | |
| CainGANK (source frames)=1, Training epochs=302020.04 | 35 | 0.69 | 0.85 | — | |
| Ours-FFNumber of frames (T)=32, Evaluation protocol=Feed-forward2019.05 | 40.4 | 0.65 | 0.48 | 0.38 | |
| Ours-FFNumber of frames (T)=8, Evaluation protocol=Feed-forward2019.05 | 42.2 | 0.64 | 0.47 | 0.4 | |
| Ours-FTNumber of frames (T)=8, Evaluation protocol=Fine-tuning2019.05 | 42.2 | 0.68 | 0.42 | 0.39 | |
| FSHM-FF-fullK (source frames)=8, Training Scope=Entire dataset, Training epochs=302020.04 | 42.2 | 0.64 | — | — | |
| FSHM-FT-fullK (source frames)=8, Training Scope=Entire dataset, Training epochs=302020.04 | 42.2 | 0.68 | — | — | |
| Ours-FFNumber of frames (T)=1, Evaluation protocol=Feed-forward2019.05 | 46.1 | 0.61 | 0.42 | 0.43 | |
| FSHM-FF-fullK (source frames)=1, Training Scope=Entire dataset, Training epochs=302020.04 | 46.61 | 0.61 | — | — | |
| Ours-FTNumber of frames (T)=1, Evaluation protocol=Fine-tuning2019.05 | 48.5 | 0.64 | 0.35 | 0.46 | |
| FSHM-FT-fullK (source frames)=1, Training Scope=Entire dataset, Training epochs=302020.04 | 48.5 | 0.64 | — | — | |
| pix2pixHDK (source frames)=8, Training epochs=302020.04 | 71.89 | 0.66 | 0.81 | — | |
| pix2pixHDK (source frames)=1, Training epochs=302020.04 | 72.26 | 0.66 | 0.8 | — | |
| FSHMK (source frames)=8, Training epochs=302020.04 | 83.13 | 0.65 | 0.73 | — | |
| FSHMK (source frames)=1, Training epochs=302020.04 | 93.17 | 0.64 | 0.72 | — |