Monocular Depth Estimation on NYU v2
99.5Delta 1 AccDepth Pro
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Depth Prowith Ours=true2026.05 | 99.5 | — | — | — | — | — | — | 0.012 | — | |
| DepthAnythingV2with Ours=true2026.05 | 99.5 | — | — | — | — | — | — | 0.013 | — | |
| LeReSwith Ours=true2026.05 | 99.4 | — | — | — | — | — | — | 0.014 | — | |
| DPTwith Ours=true2026.05 | 99.4 | — | — | — | — | — | — | 0.016 | — | |
| GeoWizardwith Ours=true2026.05 | 99.4 | — | — | — | — | — | — | 0.014 | — | |
| Metric3Dv2with Ours=true2026.05 | 99.4 | — | — | — | — | — | — | 0.014 | — | |
| Lotuswith Ours=true2026.05 | 99.3 | — | — | — | — | — | — | 0.015 | — | |
| DepthMasterwith Ours=true2026.05 | 99.3 | — | — | — | — | — | — | 0.016 | — | |
| UniDepthV2with Ours=true2026.05 | 99.3 | — | — | — | — | — | — | 0.017 | — | |
| Marigoldwith Ours=true2026.05 | 99.1 | — | — | — | — | — | — | 0.017 | — | |
| Modality ForcingMethod Type=I2D from Joint Model2026.06 | 98.9 | — | — | — | — | — | — | 2.52 | — | |
| DepthLabSynthetic Training Samples=74k2024.12 | 98.8 | — | — | — | — | — | — | 2.5 | — | |
| HD-VGGT2026.03 | 98.8 | — | — | — | — | — | — | 0.035 | — | |
| pi^32026.03 | 98.6 | — | — | — | — | — | — | 0.038 | — | |
| UniDepth V2Method Type=Discriminative depth estimation2026.06 | 98.6 | — | — | — | — | — | — | 2.96 | — | |
| MoGe-2Method Type=Discriminative depth estimation2026.06 | 98.6 | — | — | — | — | — | — | 2.89 | — | |
| Cross-Context DistillationZero-shot=true, Teacher Model=MiDaS v3.12025.02 | 98.5 | — | — | — | 0.046 | — | — | — | — | |
| MoGeFine-tuning=false, Evaluation Protocol=Scale-invariant relative depth2025.11 | 98.5 | — | — | — | 3.02 | — | — | — | — | |
| MoGe 2Type=Discriminative, Evaluation Protocol=Zero-shot2026.01 | 98.4 | — | — | — | — | — | — | 3.1 | — | |
| VGGTFine-tuning=false, Evaluation Protocol=Scale-invariant relative depth2025.11 | 98.3 | — | — | — | 3.14 | — | — | — | — | |
| VGGT+OursFine-tuning=true, Evaluation Protocol=Scale-invariant relative depth2025.11 | 98.3 | — | — | — | 3.1 | — | — | — | — | |
| PPDType=Generative, Evaluation Protocol=Zero-shot2026.01 | 98.2 | — | — | — | — | — | — | 3.3 | — | |
| DINOv3ViT=H+/16, #Params=841M, Zero-shot=true, DPT head=true2025.12 | 98.2 | — | — | — | 0.044 | — | — | — | — | |
| Depth ProMethod Type=Discriminative depth estimation2026.06 | 98.2 | — | — | — | — | — | — | 3.67 | — | |
| Cross-Context DistillationZero-shot=true, Teacher Model=DepthAnythingv2-Large2025.02 | 98.1 | — | — | — | 0.043 | — | — | — | — | |
| Depth AnythingTraining samples=63.5M, Feed-forward=true, Diffusion=false, Zero-shot=true2024.07 | 98 | — | — | — | 4.3 | — | — | — | — | |
| BetterDepthTraining samples=74K, Feed-forward=true, Diffusion=true, Zero-shot=true2024.07 | 98 | — | — | — | 4.2 | — | — | — | — | |
| BetterDepthGT aligned=true2024.11 | 98 | — | — | — | 0.04 | — | — | — | — | |
| UniDepthGT aligned=false2024.11 | 98 | — | — | — | 0.05 | — | — | — | — | |
| MiDaS v3.1Zero-shot=true2025.02 | 98 | — | — | — | — | — | — | — | — | |
| PixioViT=H/16, #Params=631M, Zero-shot=true, DPT head=true2025.12 | 98 | — | — | — | 0.041 | — | — | — | — | |
| DepthAnythingV2Real Training Samples=62M, Synthetic Training Samples=595K2024.12 | 98 | — | — | — | — | — | — | 4.4 | — | |
| DepthAnythingV2with Ours=false2026.05 | 98 | — | — | — | — | — | — | 0.043 | — | |
| BetterDepth-400Training samples=400, Feed-forward=true, Diffusion=true, Zero-shot=true2024.07 | 97.9 | — | — | — | 4.6 | — | — | — | — | |
| BetterDepth-2KTraining samples=2K, Feed-forward=true, Diffusion=true, Zero-shot=true2024.07 | 97.9 | — | — | — | 4.4 | — | — | — | — | |
| DepthAnything v2Zero-shot=true2025.02 | 97.9 | — | — | — | 0.045 | — | — | — | — | |
| CUT3RFine-tuning=false, Evaluation Protocol=Scale-invariant relative depth2025.11 | 97.9 | — | — | — | 3.73 | — | — | — | — | |
| CUT3R+OursFine-tuning=true, Evaluation Protocol=Scale-invariant relative depth2025.11 | 97.9 | — | — | — | 3.68 | — | — | — | — | |
| DINOv2ViT=g/14, #Params=1137M, Zero-shot=true, DPT head=true2025.12 | 97.9 | — | — | — | 0.046 | — | — | — | — | |
| DA-v2Method Type=Discriminative depth estimation2026.06 | 97.9 | — | — | — | — | — | — | 4.16 | — | |
| DUSt3R+OursFine-tuning=true, Evaluation Protocol=Scale-invariant relative depth2025.11 | 97.8 | — | — | — | 3.68 | — | — | — | — | |
| Depth ProType=Discriminative, Evaluation Protocol=Zero-shot2026.01 | 97.8 | — | — | — | — | — | — | 4 | — | |
| DUSt3RFine-tuning=false, Evaluation Protocol=Scale-invariant relative depth2025.11 | 97.7 | — | — | — | 3.83 | — | — | — | — | |
| PPDMethod Type=Generative depth estimation2026.06 | 97.7 | — | — | — | — | — | — | 3.82 | — | |
| Depth Anything v2Type=Discriminative, Evaluation Protocol=Zero-shot2026.01 | 97.6 | — | — | — | — | — | — | 4.1 | — | |
| DINOv2ViT=L/14, #Params=304M, Zero-shot=true, DPT head=true2025.12 | 97.6 | — | — | — | 0.05 | — | — | — | — | |
| DepthAnythingReal Training Samples=63.5M2024.12 | 97.6 | — | — | — | — | — | — | 4.4 | — | |
| Depth Prowith Ours=false2026.05 | 97.5 | — | — | — | — | — | — | 0.044 | — | |
| Metric3Dv2with Ours=false2026.05 | 97.3 | — | — | — | — | — | — | 0.067 | — | |
| ZoeDepthMethod Type=Discriminative depth estimation2026.06 | 97.3 | — | — | — | — | — | — | 4.76 | — | |
| LotusMethod Type=Generative depth estimation2026.06 | 97.3 | — | — | — | — | — | — | 4.31 | — | |
| DepthMasterwith Ours=false2026.05 | 97.2 | — | — | — | — | — | — | 0.05 | — | |
| LotusGT aligned=false2024.11 | 97 | — | — | — | 0.05 | — | — | — | — | |
| Metric3Dv2GT aligned=false2024.11 | 97 | — | — | — | 0.07 | — | — | — | — | |
| SharpDepthGT aligned=false2024.11 | 97 | — | — | — | 0.06 | — | — | — | — | |
| MAEViT=H/14, #Params=631M, Zero-shot=true, DPT head=true2025.12 | 97 | — | — | — | 0.054 | — | — | — | — | |
| UniDepthV2with Ours=false2026.05 | 97 | — | — | — | — | — | — | 0.062 | — | |
| GenPerceptZero-shot=true2025.02 | 96.9 | — | — | — | 0.058 | — | — | — | — | |
| VGGT2026.03 | 96.9 | — | — | — | — | — | — | 0.062 | — | |
| JointDiTMethod Type=I2D from Joint Model2026.06 | 96.9 | — | — | — | — | — | — | 5.12 | — | |
| LotusType=Generative, Evaluation Protocol=Zero-shot2026.01 | 96.8 | — | — | — | — | — | — | 5.4 | — | |
| WorldMirror2026.03 | 96.8 | — | — | — | — | — | — | 0.063 | — | |
| MarigoldMethod Type=Generative depth estimation2026.06 | 96.8 | — | — | — | — | — | — | 4.88 | — | |
| MoGe-2Fine-tuning=false, Evaluation Protocol=Metric depth2025.11 | 96.7 | — | — | — | 6.92 | — | — | — | — | |
| Lotuswith Ours=false2026.05 | 96.7 | — | — | — | — | — | — | 0.053 | — | |
| MASt3RMethod Type=Discriminative depth estimation2026.06 | 96.7 | — | — | — | — | — | — | 4.67 | — | |
| GeoWizardTraining samples=280K, Feed-forward=false, Diffusion=true, Zero-shot=true2024.07 | 96.6 | — | — | — | 5.2 | — | — | — | — | |
| GeoWizardType=Generative, Evaluation Protocol=Zero-shot2026.01 | 96.6 | — | — | — | — | — | — | 5.2 | — | |
| GenPerceptType=Generative, Evaluation Protocol=Zero-shot2026.01 | 96.6 | — | — | — | — | — | — | 5.2 | — | |
| GeoWizardwith Ours=false2026.05 | 96.6 | — | — | — | — | — | — | 0.052 | — | |
| GeoWizardMethod Type=Generative depth estimation2026.06 | 96.6 | — | — | — | — | — | — | 4.89 | — | |
| GeoWizardSynthetic Training Samples=278K2024.12 | 96.5 | — | — | — | — | — | — | 5.2 | — | |
| VPD2023.10 | 96.4 | 0.254 | 99.5 | 99.9 | 0.069 | 0.03 | — | — | — | |
| MeSa2023.10 | 96.4 | 0.238 | 99.5 | 99.9 | 0.066 | 0.029 | — | — | — | |
| MarigoldTraining samples=74K, Feed-forward=false, Diffusion=true, Zero-shot=true2024.07 | 96.4 | — | — | — | 5.5 | — | — | — | — | |
| MarigoldType=Generative, Evaluation Protocol=Zero-shot2026.01 | 96.4 | — | — | — | — | — | — | 5.5 | — | |
| MarigoldSynthetic Training Samples=74k2024.12 | 96.4 | — | — | — | — | — | — | 5.6 | — | |
| Marigoldwith Ours=false2026.05 | 96.4 | — | — | — | — | — | — | 0.055 | — | |
| DepthFMType=Generative, Evaluation Protocol=Zero-shot2026.01 | 96.3 | — | — | — | — | — | — | 5.5 | — | |
| MarigoldZero-shot=true2025.02 | 96.1 | — | — | — | 0.055 | — | — | — | — | |
| MarigoldGT aligned=true2024.11 | 96 | — | — | — | 0.05 | — | — | — | — | |
| DepthFMTraining samples=63K, Feed-forward=false, Diffusion=true, Zero-shot=true2024.07 | 95.6 | — | — | — | 6.5 | — | — | — | — | |
| DepthFMSynthetic Training Samples=63K2024.12 | 95.6 | — | — | — | — | — | — | 6.5 | — | |
| ZoeDepth2023.10 | 95.5 | 0.27 | 99.5 | 99.9 | 0.075 | 0.032 | — | — | — | |
| PixioViT=H/16, #Params=631M, Head type=DPT Head2025.12 | 95.5 | 0.268 | — | — | — | — | — | — | — | |
| AiT2023.10 | 95.4 | 0.275 | 99.4 | 99.9 | 0.076 | 0.033 | — | — | — | |
| AiT-PArchitecture=SwinV2-L2023.06 | 95.3 | 0.279 | — | — | 0.076 | — | — | — | — | |
| ZoeDepthGT aligned=false2024.11 | 95 | — | — | — | 0.08 | — | — | — | — | |
| Xie et al.2023.10 | 94.9 | 0.287 | 99.4 | 99.9 | 0.083 | 0.035 | — | — | — | |
| MIMArchitecture=SwinV2-L2023.06 | 94.9 | 0.287 | — | — | 0.083 | — | — | — | — | |
| MiDaSwith Ours=true2026.05 | 94.9 | — | — | — | — | — | — | 0.048 | — | |
| HDNTraining samples=300K, Feed-forward=true, Diffusion=false, Zero-shot=true2024.07 | 94.8 | — | — | — | 6.9 | — | — | — | — | |
| HDNZero-shot=true2025.02 | 94.8 | — | — | — | 0.069 | — | — | — | — | |
| HDNType=Discriminative, Evaluation Protocol=Zero-shot2026.01 | 94.8 | — | — | — | — | — | — | 6.9 | — | |
| DINOv2 ViT-B/14readout=DPT2026.02 | 94.8 | 0.297 | — | — | — | — | — | — | — | |
| Omnivorous ViT-B/14readout=DPT2026.02 | 94.8 | 0.297 | — | — | — | — | — | — | — | |
| DDVMArchitecture=Efficient U-Net, Unsupervised Pretraining=true, Auxiliary Supervised Depth Data=true, samples=42023.06 | 94.6 | 0.315 | — | — | 0.074 | — | — | — | — | |
| HDNReal Training Samples=300K2024.12 | 94.6 | — | — | — | — | — | — | 7.2 | — | |
| OmnidataTraining samples=12.2M, Feed-forward=true, Diffusion=false, Zero-shot=true2024.07 | 94.5 | — | — | — | 7.4 | — | — | — | — | |
| OmnidataZero-shot=true2025.02 | 94.5 | — | — | — | 0.074 | — | — | — | — |