Depth Estimation on DIODE
97.4Delta-1 AccuracyMoGe
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| MoGeZero-shot=true2026.01 | 97.4 | — | — | — | — | — | — | — | — | — | |
| MoGe-2Zero-shot=true2026.01 | 97.4 | — | — | — | — | — | — | — | — | — | |
| InfiniDepth (Ours-Relative)Zero-shot=true2026.01 | 97.4 | — | — | — | — | — | — | — | — | — | |
| DepthAnythingZero-shot=true2026.01 | 97.3 | — | — | — | — | — | — | — | — | — | |
| DepthProZero-shot=true2026.01 | 97.1 | — | — | — | — | — | — | — | — | — | |
| DepthAnythingv2Zero-shot=true2026.01 | 97 | — | — | — | — | — | — | — | — | — | |
| PPDZero-shot=true2026.01 | 96.2 | — | — | — | — | — | — | — | — | — | |
| MarigoldZero-shot=true2026.01 | 94.7 | — | — | — | — | — | — | — | — | — | |
| Vision Bananadepth evaluation protocol=absolute-depth2026.07 | 91.7 | — | — | — | — | — | — | — | — | — | |
| Unified-IOTraining Samples=48K, NFEs=-, Zero-shot=true2025.12 | 90.6 | — | — | 36.9 | — | — | 1.9 | — | — | — | |
| Metric3D v2Training Samples=16M, NFEs=-, Zero-shot=true2025.12 | 89.2 | — | — | 14.7 | — | — | 1.2 | — | — | — | |
| MoGe-2Method category=Geometry-specialized2026.07 | 82.3 | — | — | — | — | — | — | — | 23 | — | |
| MTDsetting=zero-shot, seeds=metric depth estimator2026.05 | 81.6 | — | — | — | — | — | — | — | 20.7 | — | |
| FE2EMethod category=Generation-based2026.07 | 81.2 | — | — | — | — | — | — | — | 22.8 | — | |
| DepthFMTraining Samples=63K, NFEs=10 x 4, Zero-shot=true2025.12 | 80 | — | — | 22.5 | — | — | 5.2 | — | — | — | |
| GeoWizardsetting=zero-shot, seeds=metric depth estimator2026.05 | 79.2 | — | — | — | — | — | — | — | 29.7 | — | |
| Hierarchical Depth NormalizationTraining Data=~ 300K, Zero-shot transfer=true2022.10 | 78 | — | — | 24.6 | — | — | — | — | — | — | |
| HDN2024.11 | 78 | — | — | 24.6 | — | — | — | — | — | — | |
| HDNTraining Samples=300K, NFEs=-, Zero-shot=true2025.12 | 78 | — | — | 24.6 | — | — | 3.9 | — | — | — | |
| HDNsetting=zero-shot, seeds=metric depth estimator2026.05 | 78 | — | — | — | — | — | — | — | 24.6 | — | |
| D3-PredictorTraining Samples=30K, NFEs=1 x 1, Zero-shot=true2025.12 | 77.9 | — | — | 29 | — | — | 2.1 | — | — | — | |
| E2E-FTTraining Samples=74K, NFEs=1 x 1, Zero-shot=true2025.12 | 77.6 | — | — | 30.3 | — | — | 3.1 | — | — | — | |
| DepthMastersetting=zero-shot, seeds=metric depth estimator2026.05 | 77.6 | — | — | — | — | — | — | — | 21.5 | — | |
| Marigold2024.11 | 77.3 | — | — | 30.8 | — | — | — | — | — | — | |
| MarigoldTraining Samples=74K, NFEs=10 x 50, Zero-shot=true2025.12 | 77.3 | — | — | 30.8 | — | — | 4.4 | — | — | — | |
| Marigoldsetting=zero-shot, seeds=metric depth estimator2026.05 | 77.3 | — | — | — | — | — | — | — | 30.8 | — | |
| MarigoldMethod category=Generation-based2026.07 | 77.3 | — | — | — | — | — | — | — | 30.8 | — | |
| DiT-based Diffusion for Perceptual TasksBackbone=DiT-XL2024.11 | 77.2 | — | — | 31 | — | — | — | — | — | — | |
| GenperceptTraining Samples=90K, NFEs=1 x 1, Zero-shot=true2025.12 | 76.7 | — | — | 30.2 | — | — | 4 | — | — | — | |
| OursBackbone=ResNeXt101, Zero-shot=true2020.12 | 76.6 | — | — | 9.5 | — | — | — | — | — | — | |
| LeresTraining Data=~ 360K, Zero-shot transfer=true2022.10 | 76.6 | — | — | 27.1 | — | — | — | — | — | — | |
| LeReS2024.11 | 76.6 | — | — | 27.1 | — | — | — | — | — | — | |
| LeReSsetting=zero-shot, seeds=metric depth estimator2026.05 | 76.6 | — | — | — | — | — | — | — | 27.1 | — | |
| SenseNova-Visiondepth evaluation protocol=affine-invariant2026.07 | 76.4 | — | — | — | — | — | — | — | — | — | |
| SenseNova-VisionMethod category=Generation-based2026.07 | 76.4 | — | — | — | — | — | — | — | 20.6 | — | |
| DepthAnythingTraining Samples=63.5K, NFEs=-, Zero-shot=true2025.12 | 75.9 | — | — | 26.1 | — | — | 2.2 | — | — | — | |
| DepthAnythingV2setting=zero-shot, seeds=metric depth estimator2026.05 | 75.9 | — | — | — | — | — | — | — | 26 | — | |
| DepthAnythingMethod category=Geometry-specialized2026.07 | 75.9 | — | — | — | — | — | — | — | 26 | — | |
| Ours + LeReSBackbone=ResNet50, Zero-shot=true2022.02 | 75.8 | — | — | 27.4 | — | — | — | — | — | — | |
| DPT2024.11 | 75.8 | — | — | 18.2 | — | — | — | — | — | — | |
| DPT-largeTraining Samples=1.4M, NFEs=-, Zero-shot=true2025.12 | 75.8 | — | — | 18.2 | — | — | 4.2 | — | — | — | |
| DepthAnything v2Training Samples=62.6M, NFEs=-, Zero-shot=true2025.12 | 75.8 | — | — | 32.1 | — | — | 2.8 | — | — | — | |
| GeoWizardTraining Samples=280K, NFEs=10 x 50, Zero-shot=true2025.12 | 75.3 | — | — | 32.8 | — | — | 5.6 | — | — | — | |
| Depth Prosetting=zero-shot, seeds=metric depth estimator2026.05 | 75.2 | — | — | — | — | — | — | — | 32.1 | — | |
| Lotus-2Method category=Generation-based2026.07 | 75.2 | — | — | — | — | — | — | — | 22.1 | — | |
| OursBackbone=ResNet50, Zero-shot=true2020.12 | 75.1 | — | — | 9.6 | — | — | — | — | — | — | |
| LeReSBackbone=ResNet50, Zero-shot=true2022.02 | 75.1 | — | — | 28.7 | — | — | — | — | — | — | |
| Omnidata2024.11 | 74.2 | — | — | 33.9 | — | — | — | — | — | — | |
| OmnidataTraining Samples=12.2M, NFEs=-, Zero-shot=true2025.12 | 74.2 | — | — | 33.9 | — | — | 5.1 | — | — | — | |
| Omnidatasetting=zero-shot, seeds=metric depth estimator2026.05 | 74.2 | — | — | — | — | — | — | — | 33.9 | — | |
| DeceptionTraining Samples=500K, NFEs=1 x 28, Zero-shot=true2025.12 | 74.1 | — | — | 24.3 | — | — | 3.6 | — | — | — | |
| Lotussetting=zero-shot, seeds=metric depth estimator2026.05 | 73.8 | — | — | — | — | — | — | — | 22.8 | — | |
| Single-task baseline2025.06 | 73.72 | 87.72 | 93.11 | — | 3.9632 | 0.3163 | — | — | 32.56 | 3.9233 | |
| StableMTL2025.06 | 73.72 | 87.45 | 92.89 | — | 4.0073 | 0.3192 | — | — | 33.03 | 3.9516 | |
| StableMTL-SModel type=single-stream2025.06 | 73.42 | 87.17 | 92.62 | — | 4.0281 | 0.3227 | — | — | 33.36 | 3.9504 | |
| 4M-XLTraining Samples=759M, NFEs=-, Zero-shot=true2025.12 | 73.4 | — | — | 33.1 | — | — | 1.7 | — | — | — | |
| DepthAnything V2Method category=Geometry-specialized2026.07 | 73.4 | — | — | — | — | — | — | — | 26.5 | — | |
| DPTsetting=zero-shot, seeds=metric depth estimator2026.05 | 73 | — | — | — | — | — | — | — | 26.9 | — | |
| DICEPTIONMethod category=Generation-based2026.07 | 72.2 | — | — | — | — | — | — | — | 28.9 | — | |
| MiDaSBackbone=ResNeXt101, Zero-shot=true2020.12 | 71.5 | — | — | 11.1 | — | — | — | — | — | — | |
| MiDaSTraining Data=~ 2M, Zero-shot transfer=true2022.10 | 71.5 | — | — | 33.2 | — | — | — | — | — | — | |
| MiDaSBackbone=ResNeXt101, Zero-shot=true2022.02 | 71.5 | — | — | 33.2 | — | — | — | — | — | — | |
| MiDaS2024.11 | 71.5 | — | — | 33.2 | — | — | — | — | — | — | |
| MiDaSTraining Samples=2M, NFEs=-, Zero-shot=true2025.12 | 71.5 | — | — | 33.2 | — | — | 6.9 | — | — | — | |
| MiDaSsetting=zero-shot, seeds=metric depth estimator2026.05 | 71.3 | — | — | — | — | — | — | — | 26.6 | — | |
| MegaDepthTraining Data=~ 150K, Zero-shot transfer=true2022.10 | 71.2 | — | — | 39.1 | — | — | — | — | — | — | |
| DMPTraining Samples=10K, NFEs=1 x 5, Zero-shot=true2025.12 | 70.6 | — | — | 36.1 | — | — | 7.7 | — | — | — | |
| OneDiffusionTraining Samples=500K, NFEs=-, Zero-shot=true2025.12 | 66.1 | — | — | 39.9 | — | — | 2.4 | — | — | — | |
| ChenBackbone=ResNet50, Zero-shot=true2020.12 | 66 | — | — | 16.5 | — | — | — | — | — | — | |
| Chen et al.Training Data=~ 795K, Zero-shot transfer=true2022.10 | 66 | — | — | 37.9 | — | — | — | — | — | — | |
| WSVDBackbone=ResNet50, Zero-shot=true2020.12 | 63.8 | — | — | 18.9 | — | — | — | — | — | — | |
| WSVDTraining Data=~ 1.5M, Zero-shot transfer=true2022.10 | 63.8 | — | — | 35.8 | — | — | — | — | — | — | |
| DiverseDepthBackbone=ResNeXt50, Zero-shot=true2020.12 | 63.1 | — | — | 10.8 | — | — | — | — | — | — | |
| DiverseDepthTraining Data=~ 300K, Zero-shot transfer=true2022.10 | 63.1 | — | — | 37.6 | — | — | — | — | — | — | |
| DiverseDepthBackbone=ResNeXt50, Zero-shot=true2022.02 | 63.1 | — | — | 37.6 | — | — | — | — | — | — | |
| DiverseDepth2024.11 | 63.1 | — | — | 37.6 | — | — | — | — | — | — | |
| DiverseDepthTraining Samples=320K, NFEs=-, Zero-shot=true2025.12 | 63.1 | — | — | 37.6 | — | — | 7 | — | — | — | |
| DiverseDepthsetting=zero-shot, seeds=metric depth estimator2026.05 | 63.1 | — | — | — | — | — | — | — | 37.6 | — | |
| XianBackbone=ResNet50, Zero-shot=true2020.12 | 61.8 | — | — | 17.4 | — | — | — | — | — | — | |
| Xian et al.Training Data=~ 30K, Zero-shot transfer=true2022.10 | 61.8 | — | — | 42.5 | — | — | — | — | — | — | |
| Xian et al.Backbone=ResNet50, Zero-shot=true2022.02 | 61.8 | — | — | 42.5 | — | — | — | — | — | — | |
| MegaDepthBackbone=Hourglass, Zero-shot=true2020.12 | 61.5 | — | — | 19 | — | — | — | — | — | — | |
| MegaDepthBackbone=Hourglass, Zero-shot=true2022.02 | 61.5 | — | — | 39.1 | — | — | — | — | — | — | |
| DiffusionMTL2025.06 | 56.93 | 78.91 | 88.56 | — | 4.9657 | 0.4106 | — | — | 45.19 | 4.6659 | |
| OASISBackbone=ResNet50, Zero-shot=true2020.12 | 53.4 | — | — | 19.8 | — | — | — | — | — | — | |
| OASISTraining Data=~ 140K, Zero-shot transfer=true2022.10 | 53.4 | — | — | 48.4 | — | — | — | — | — | — | |
| OASISBackbone=ResNet50, Zero-shot=true2022.02 | 53.4 | — | — | 48.4 | — | — | — | — | — | — | |
| DiffusionMTL*2025.06 | 49.57 | 73.62 | 84.55 | — | 7.6032 | 0.5166 | — | — | 58.17 | 7.2017 | |
| JTR2025.06 | 42.26 | 67.66 | 81.59 | — | 8.4119 | 0.5482 | — | — | 66.39 | 8.1448 | |
| JTR*2025.06 | 40.31 | 64.24 | 78.1 | — | 8.9962 | 0.5852 | — | — | 73.14 | 9.2842 | |
| AI PlaygroundTraining Scenario=Brown / Night / High2020.07 | 35.63 | 59.48 | 79.45 | 76.59 | 3.6897 | 0.2148 | — | — | — | — | |
| AI PlaygroundTraining Scenario=Brown / Night / Low2020.07 | 31.63 | 56.47 | 73.45 | 77.43 | 3.7898 | 0.2149 | — | — | — | — | |
| CUT3R2026.03 | — | — | — | 16 | — | — | — | 78.4 | — | — | |
| DAGE2026.03 | — | — | — | 8.7 | — | — | — | 92.5 | — | — | |
| DepthPro2026.03 | — | — | — | 31 | — | — | — | 58.8 | — | — | |
| DiffusionMTLFull Setting Adaptation=No2025.06 | — | — | — | — | — | — | — | — | 45.19 | — | |
| DiffusionMTLFull Setting Adaptation=Yes2025.06 | — | — | — | — | — | — | — | — | 58.17 | — | |
| GeoCrafter2026.03 | — | — | — | 9.1 | — | — | — | 93.4 | — | — | |
| JTRFull Setting Adaptation=No2025.06 | — | — | — | — | — | — | — | — | 66.39 | — | |
| JTRFull Setting Adaptation=Yes2025.06 | — | — | — | — | — | — | — | — | 73.14 | — |