Video Matting on V-HIM60 Hard
1.306MADVideoMaMa
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| VideoMaMamask_type=Downsampl. 8x2026.01 | 1.306 | 2.363 | — | — | — | |
| VideoMaMamask_type=Polygon (Easy)2026.01 | 1.416 | 2.48 | — | — | — | |
| VideoMaMamask_type=Downsampl. 32x2026.01 | 1.461 | 2.849 | — | — | — | |
| VideoMaMamask_type=Polygon (Hard)2026.01 | 1.64 | 3.024 | — | — | — | |
| MaGGIemask_type=Downsampl. 8x2026.01 | 2.379 | 5.415 | — | — | — | |
| VideoMaMamask_type=Model-Generated Masks (SAM2)2026.01 | 2.435 | 4.18 | — | — | — | |
| MaGGIemask_type=Downsampl. 32x2026.01 | 2.5375 | 5.9073 | — | — | — | |
| MaGGIemask_type=Polygon (Easy)2026.01 | 2.5808 | 5.5572 | — | — | — | |
| Input Mask (Baseline)mask_type=Downsampl. 8x2026.01 | 2.744 | 11.127 | — | — | — | |
| MaGGIemask_type=Polygon (Hard)2026.01 | 3.1567 | 6.4121 | — | — | — | |
| MaGGIemask_type=Model-Generated Masks (SAM2)2026.01 | 3.1567 | 6.4121 | — | — | — | |
| Input Mask (Baseline)mask_type=Polygon (Easy)2026.01 | 3.636 | 16 | — | — | — | |
| Input Mask (Baseline - SAM2)mask_type=Model-Generated Masks (SAM2)2026.01 | 4.666 | 21.96 | — | — | — | |
| Input Mask (Baseline)mask_type=Downsampl. 32x2026.01 | 5.132 | 33.664 | — | — | — | |
| Input Mask (Baseline)mask_type=Polygon (Hard)2026.01 | 6.771 | 43.228 | — | — | — | |
| MaGGIeInput type=Frame-by-frame binary mask, Retrained on I-HIM50K and V-HIM50K=true2024.04 | 21.23 | 7.08 | 6.89 | 29.9 | 42.98 | |
| MGM-TCVOMInput type=Frame-by-frame binary mask, Retrained on I-HIM50K and V-HIM50K=true2024.04 | 22.16 | 7.91 | 7.27 | 31 | 47.82 | |
| MGM*-TCVOMInput type=Frame-by-frame binary mask, Retrained on I-HIM50K and V-HIM50K=true, Variant=stacked guidance masks2024.04 | 22.59 | 7.86 | 7.32 | 32.75 | 37.83 | |
| SparseMatInput type=Frame-by-frame binary mask, Retrained on I-HIM50K and V-HIM50K=true2024.04 | 24.83 | 8.47 | 8.19 | 36.92 | 55.98 | |
| InstMattInput type=Frame-by-frame binary mask, Retrained on I-HIM50K and V-HIM50K=true2024.04 | 27.24 | 7.88 | 8.02 | 31.89 | 47.19 | |
| FTP-VMInput type=First-frame trimap, Retrained on I-HIM50K and V-HIM50K=false2024.04 | 46.77 | 14.4 | 15.82 | 45.04 | 76.48 | |
| FTP-VMInput type=First-frame trimap, Retrained on I-HIM50K and V-HIM50K=true2024.04 | 48.11 | 14.87 | 16.12 | 45.29 | 78.66 | |
| MGMmask_type=Downsampl. 8x2026.01 | 63.8374 | 16.5609 | — | — | — | |
| MGMmask_type=Polygon (Easy)2026.01 | 64.5457 | 17.3675 | — | — | — | |
| MGMmask_type=Model-Generated Masks (SAM2)2026.01 | 66.5221 | 18.2103 | — | — | — | |
| MGMmask_type=Polygon (Hard)2026.01 | 75.5501 | 20.5516 | — | — | — | |
| MGMmask_type=Downsampl. 32x2026.01 | 79.2273 | 22.2024 | — | — | — | |
| OTVMInput type=First-frame trimap, Retrained on I-HIM50K and V-HIM50K=true2024.04 | 140.96 | 17.6 | 47.84 | 59.66 | 298.46 | |
| OTVMInput type=First-frame trimap, Retrained on I-HIM50K and V-HIM50K=false2024.04 | 412.41 | 29.97 | 146.11 | 90.15 | 764.36 |