Multimodal Image-to-Image Translation (RGB+Normal to Shade) on Taskonomy
0.79MAETokenFusion (S)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TokenFusion (S)Architecture=Transformer-based, Scale=Small, Resolution=256x2562022.04 | 0.79 | 1.33 | |
| TokenFusion (Ti)Architecture=Transformer-based, Scale=Tiny, Resolution=256x2562022.04 | 0.95 | 1.54 | |
| CENArchitecture=CNN-based, Resolution=256x2562022.04 | 1.1 | 1.72 | |
| Self-Att.Architecture=CNN-based, Resolution=256x2562022.04 | 1.18 | 1.76 | |
| ConcatArchitecture=CNN-based, Resolution=256x2562022.04 | 1.28 | 2.02 | |
| Concat (S)Architecture=Transformer-based, Scale=Small, Resolution=256x2562022.04 | 1.3 | 2.07 | |
| Concat (Ti)Architecture=Transformer-based, Scale=Tiny, Resolution=256x2562022.04 | 1.33 | 2.2 | |
| Align.Architecture=CNN-based, Resolution=256x2562022.04 | 1.41 | 2.21 |