Monocular Depth Estimation on SYNTHIA to Cityscapes 16 classes UDA (val)
6.62Root Mean Squared Error (RMSE)STL target
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| STL targetArchitecture=Single-Task Learning, Strategy=Oracle (Target Labels)2022.06 | 6.62 | — | — | — | — | — | — | |
| MTL targetArchitecture=Multi-Task Learning, Strategy=Oracle (Target Labels)2022.06 | 6.79 | — | — | — | — | — | — | |
| mTEB (Ours)Architecture=Multi-Task Learning with mTEB, Strategy=Unsupervised Domain Adaptation2022.06 | 11.66 | — | — | — | — | — | — | |
| STL sourceArchitecture=Single-Task Learning, Strategy=Source only2022.06 | 13.79 | — | — | — | — | — | — | |
| STL-UDAArchitecture=Single-Task Learning, Strategy=Unsupervised Domain Adaptation2022.06 | 14.26 | — | — | — | — | — | — | |
| MTL-UDAArchitecture=Multi-Task Learning, Strategy=Unsupervised Domain Adaptation2022.06 | 14.47 | — | — | — | — | — | — | |
| MTL sourceArchitecture=Multi-Task Learning, Strategy=Source only2022.06 | 14.51 | — | — | — | — | — | — | |
| CTRLUDA Setting=SYNTHIA to Cityscapes, ISL=false2021.05 | 14.8 | 0.3 | 6.3 | 0.6 | 30 | 58 | 77 | |
| DADAUDA Setting=SYNTHIA to Cityscapes, ISL=false2021.05 | 17 | 0.6 | 10.8 | 4.4 | 14 | 28 | 41 |