VFHQ
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Blind Face Video Restoration | VFHQ (test) | PSNR29.563 | 14 | |
| Talking Head Generation | VFHQ (test) | FID13.67 | 10 | |
| Cross-identity Reenactment | VFHQ (test) | CSIM0.6757 | 8 | |
| Self-reenactment | VFHQ (test) | PSNR23.24 | 8 | |
| 3D Portrait Animation (Cross Reenactment) | VFHQ 1.0 (test) | CSIM66.3 | 7 | |
| 3D Portrait Animation (Self Reenactment) | VFHQ 1.0 (test) | PSNR23.47 | 7 | |
| Lip-Syncing | VFHQ | FID9.12 | 7 | |
| Talking Head Synthesis | VFHQ (first 100 frames) | FID26.22 | 6 | |
| Image Reconstruction | VFHQ | rFID3.42 | 6 | |
| Brightness de-flickering | VFHQ 1.0 (test) | FVD14.53 | 5 | |
| Inpainting | VFHQ (test) | PSNR29.119 | 4 | |
| Colorization | VFHQ (test) | PSNR23.079 | 4 |
Showing 12 of 12 rows