Zero-shot Affine-invariant Depth Estimation on NYU Indoor v2
4.3AbsRelDepthAnything
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DepthAnythingTraining Data=62.6M, Method Category=Regression Methods2026.06 | 4.3 | 98.1 | 2.25 | |
| DepthAnything V2Training Data=62.6M, Method Category=Regression Methods2026.06 | 4.5 | 97.9 | 2.88 | |
| UNIGPTraining Data=59K, Method Category=Generative Methods2026.06 | 5.2 | 96.6 | 2.38 | |
| MarigoldTraining Data=74K, Method Category=Generative Methods2026.06 | 5.5 | 96.4 | 3.56 | |
| GeoWizardTraining Data=280K, Method Category=Generative Methods2026.06 | 5.6 | 96.3 | 4.81 | |
| OmnidataTraining Data=12.2M, Method Category=Regression Methods2026.06 | 7.4 | 94.5 | 6.5 | |
| UniConTraining Data=16K, Method Category=Generative Methods2026.06 | 7.9 | 93.9 | 7.5 | |
| JoDiTraining Data=290K, Method Category=Generative Methods2026.06 | 8.3 | 92 | 9.13 | |
| OneDiffusionTraining Data=75M, Method Category=Generative Methods2026.06 | 8.9 | 92 | 8.88 | |
| DPTTraining Data=1.4M, Method Category=Regression Methods2026.06 | 9.8 | 90.3 | 6.5 | |
| MiDaSTraining Data=2M, Method Category=Regression Methods2026.06 | 11.1 | 88.5 | 9.88 | |
| JointNetTraining Data=2.56M, Method Category=Generative Methods2026.06 | 13.6 | 84.1 | 10 |