Monocular Depth Estimation on KITTI (Eigen)
0.034Abs RelIID-RDepth
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| IID-RDepthType=Classification-Regression, Mode=Fully Trainable, Modules=InvT-IndDiffusion and AV-LFE2026.04 | 0.034 | 0.057 | 1.295 | 0.052 | 99.2 | 99.9 | 100 | — | — | 37.77 | |
| Depth AnythingEvaluation protocol=Fine-tuned, Backbone=ViT-L, Framework=ZoeDepth2024.01 | 0.046 | — | 1.896 | 0.069 | 98.2 | 99.8 | 100 | — | — | — | |
| perceptual diffusion modelMethod category=diffusion-based2023.12 | 0.047 | — | 1.929 | — | 98.2 | 99.8 | 100 | — | — | — | |
| IID-RDepthType=Classification-Regression, Mode=Compatible, Modules=InvT-IndDiffusion and AV-LFE2026.04 | 0.047 | 0.107 | 1.722 | 0.069 | 98.4 | 99.9 | 100 | — | — | 17.25 | |
| NDDepthCap=0-50m2023.09 | 0.048 | 0.107 | 1.513 | 0.071 | 98.1 | 99.8 | 100 | — | — | — | |
| GBDepthMethod category=non-diffusion-based2023.12 | 0.048 | — | 2.044 | — | 97.6 | 99.7 | 99.9 | — | — | — | |
| GEDepthEvaluation protocol=Fine-tuned2024.01 | 0.048 | — | 2.044 | 0.076 | 97.6 | 99.7 | 99.9 | — | — | — | |
| EVPType=Regression2026.04 | 0.048 | 0.136 | 2.015 | 0.073 | 98 | 99.9 | 100 | — | — | 3.17 | |
| ECoDepthType=Regression2026.04 | 0.048 | 0.139 | 2.039 | 0.074 | 97.9 | 99.7 | 100 | — | — | 2.02 | |
| URCDC-DepthCap=0-50m, Backbone=Swin-Large2023.02 | 0.049 | 0.108 | 1.528 | 0.072 | 98.1 | 99.8 | 100 | — | — | — | |
| MAMOMethod category=non-diffusion-based2023.12 | 0.049 | — | 1.984 | — | 97.7 | 99.8 | 100 | — | — | — | |
| WorDepthType=Regression2026.04 | 0.049 | — | 2.039 | 0.074 | 97.9 | 99.8 | 99.9 | — | — | 2.02 | |
| GEDepthType=Classification-Regression2026.04 | 0.049 | 0.143 | 2.054 | 0.076 | — | — | — | — | — | 1.3 | |
| URCDC-DepthCap=0-80m, Backbone=Swin-Large2023.02 | 0.05 | 0.142 | 2.032 | 0.076 | 97.7 | 99.7 | 99.9 | — | — | — | |
| NDDepthCap=0-80m2023.09 | 0.05 | 0.141 | 2.025 | 0.075 | 97.8 | 99.8 | 99.9 | — | — | — | |
| SwinV2-MIMMethod category=non-diffusion-based2023.12 | 0.05 | — | 1.966 | — | 97.7 | 99.8 | 100 | — | — | — | |
| DDPMethod category=diffusion-based2023.12 | 0.05 | — | 2.072 | — | 97.5 | 99.7 | 99.9 | — | — | — | |
| SwinV2-LEvaluation protocol=Fine-tuned2024.01 | 0.05 | — | 1.966 | 0.075 | 97.7 | 99.8 | 100 | — | — | — | |
| NDDepthEvaluation protocol=Fine-tuned2024.01 | 0.05 | — | 2.025 | 0.075 | 97.8 | 99.8 | 99.9 | — | — | — | |
| DiffusionDepth2024.12 | 0.05 | — | — | — | 97.7 | — | — | — | — | — | |
| URCDCType=Regression2026.04 | 0.05 | 0.142 | 2.032 | 0.076 | 97.7 | 99.7 | 99.9 | — | — | 2.35 | |
| NDDepthType=Regression2026.04 | 0.05 | 0.141 | 2.025 | 0.075 | 97.8 | 99.8 | 99.9 | — | — | 2.69 | |
| DDPType=Regression2026.04 | 0.05 | 0.148 | 2.072 | 0.076 | 97.5 | 99.7 | 99.9 | — | — | 0.43 | |
| DiffusionDepthType=Regression2026.04 | 0.05 | 0.141 | 2.016 | 0.074 | 97.7 | 99.8 | 99.9 | — | — | 3.12 | |
| iDiscType=Classification-Regression2026.04 | 0.05 | 0.145 | 2.067 | 0.077 | 97.7 | 99.7 | 99.9 | — | — | 0.67 | |
| IID-RDepthType=Classification-Regression2026.04 | 0.05 | 0.14 | 1.996 | 0.075 | 97.9 | 99.8 | 100 | — | — | 4.09 | |
| PixelFormerMethod category=non-diffusion-based2023.12 | 0.051 | — | 2.081 | — | 97.6 | 99.7 | 99.9 | — | — | — | |
| PixelFormerType=Classification-Regression2026.04 | 0.051 | 0.149 | 2.081 | 0.077 | 97.6 | 99.7 | 99.9 | — | — | 0 | |
| DepthFormerCap=0-80m, Backbone=Swin-Large and ResNet-502023.02 | 0.052 | 0.158 | 2.143 | 0.079 | 97.5 | 99.7 | 99.9 | — | — | — | |
| NeWCRFsCap=0-80m, Backbone=Swin-Large2023.02 | 0.052 | 0.155 | 2.129 | 0.079 | 97.4 | 99.7 | 99.9 | — | — | — | |
| NeWCRFsCap=0-80m2023.09 | 0.052 | 0.155 | 2.129 | 0.079 | 97.4 | 99.7 | 99.9 | — | — | — | |
| DepthFormerBackbone=Swin-L+R-50-C12022.03 | 0.052 | 0.158 | 2.143 | 0.079 | 97.5 | 99.7 | 99.9 | — | — | — | |
| DepthFormerMethod category=non-diffusion-based2023.12 | 0.052 | — | 2.143 | — | 97.5 | 99.7 | 99.9 | — | — | — | |
| NeWCRFsEvaluation protocol=Fine-tuned2024.01 | 0.052 | — | 2.129 | 0.079 | 97.4 | 99.7 | 99.9 | — | — | — | |
| DepthFormer2024.12 | 0.052 | — | — | — | 97.5 | — | — | — | — | — | |
| BinsFormer2024.12 | 0.052 | — | — | — | 97.4 | — | — | — | — | — | |
| NeWCRFsType=Regression2026.04 | 0.052 | 0.155 | 2.129 | 0.079 | 97.4 | 99.7 | 99.9 | — | — | -2.31 | |
| DepthformerType=Regression2026.04 | 0.052 | 0.158 | 2.143 | 0.079 | 97.5 | 99.7 | 99.8 | — | — | -2.98 | |
| CrossFlow2024.12 | 0.053 | — | — | — | 97.3 | — | — | — | — | — | |
| BinsFormerType=Classification-Regression2026.04 | 0.053 | 0.156 | 2.141 | 0.08 | 97.4 | 99.7 | 99.9 | — | — | -2.88 | |
| ZoeDepthEvaluation protocol=Fine-tuned, Reproduced=true2024.01 | 0.054 | — | 2.281 | 0.082 | 97.1 | 99.6 | 99.9 | — | — | — | |
| P3DepthCap=0-50m2023.02 | 0.055 | 0.13 | 1.651 | 0.081 | 97.4 | 99.7 | 99.9 | — | — | — | |
| P3DepthCap=0-50m2023.09 | 0.055 | 0.13 | 1.651 | 0.081 | 97.4 | 99.7 | 99.9 | — | — | — | |
| BTSTraining Input/Auxiliar Info=SI/GT Depth, Cap=0 − 50 m2020.10 | 0.056 | 0.169 | 1.925 | 0.087 | 96.4 | 99.4 | 99.9 | — | — | — | |
| PWACap=0-50m2023.02 | 0.057 | 0.161 | 1.872 | 0.087 | 96.5 | 99.5 | 99.9 | — | — | — | |
| PWACap=0-50m2023.09 | 0.057 | 0.161 | 1.872 | 0.087 | 96.5 | 99.5 | 99.9 | — | — | — | |
| ZoeDepthMethod category=non-diffusion-based2023.12 | 0.057 | — | — | — | — | — | — | — | — | — | |
| Adabins2022.07 | 0.058 | — | 2.36 | — | 96.4 | 99.5 | 99.9 | — | — | — | |
| Depthformer2022.07 | 0.058 | — | 2.285 | — | 96.7 | 99.6 | 99.9 | — | — | — | |
| AdabinsCap=0-80m2023.02 | 0.058 | 0.19 | 2.36 | 0.088 | 96.4 | 99.5 | 99.9 | — | — | — | |
| BTSCap=0-50m2023.02 | 0.058 | 0.183 | 1.995 | 0.09 | 96.2 | 99.4 | 99.9 | — | — | — | |
| AdabinsCap=0-80m2023.09 | 0.058 | 0.19 | 2.36 | 0.088 | 96.4 | 99.5 | 99.9 | — | — | — | |
| BTSCap=0-50m2023.09 | 0.058 | 0.183 | 1.995 | 0.09 | 96.2 | 99.4 | 99.9 | — | — | — | |
| AdaBinsBackbone=Mini-ViT+E-B52022.03 | 0.058 | 0.19 | 2.36 | 0.088 | 96.4 | 99.5 | 99.9 | — | — | — | |
| AdaBinsEvaluation protocol=Fine-tuned2024.01 | 0.058 | — | 2.36 | 0.088 | 96.4 | 99.5 | 99.9 | — | — | — | |
| AdaBins2024.12 | 0.058 | — | — | — | 96.4 | — | — | — | — | — | |
| AdabinType=Classification-Regression2026.04 | 0.058 | 0.19 | 2.36 | 0.088 | 96.4 | 99.5 | 99.9 | — | — | -13.41 | |
| BTSTraining Input/Auxiliar Info=SI/GT Depth, Cap=0 − 80 m2020.10 | 0.059 | 0.245 | 2.756 | 0.096 | 95.6 | 99.3 | 99.8 | — | — | — | |
| BTS2022.07 | 0.059 | — | 2.756 | — | 95.6 | 99.3 | 99.8 | — | — | — | |
| BTSBackbone=DenseNet-1612022.03 | 0.059 | 0.245 | 2.756 | 0.096 | 95.6 | 99.3 | 99.8 | — | — | — | |
| PWACap=0-80m2023.02 | 0.06 | 0.221 | 2.604 | 0.093 | 95.8 | 99.4 | 99.9 | — | — | — | |
| PWACap=0-80m2023.09 | 0.06 | 0.221 | 2.604 | 0.093 | 95.8 | 99.4 | 99.9 | — | — | — | |
| BTSSup=M, Data=K2021.03 | 0.061 | 0.261 | 2.834 | 0.099 | 95.4 | 99.2 | 99.8 | — | — | — | |
| BTSCap=0-80m2023.02 | 0.061 | 0.261 | 2.834 | 0.099 | 95.4 | 99.2 | 99.8 | — | — | — | |
| TransDepthCap=0-50m2023.02 | 0.061 | 0.185 | 1.992 | 0.091 | 96.3 | 99.5 | 99.9 | — | — | — | |
| BTSCap=0-80m2023.09 | 0.061 | 0.261 | 2.834 | 0.099 | 95.4 | 99.2 | 99.8 | — | — | — | |
| DPT-Hybrid2022.07 | 0.062 | — | 2.573 | — | 95.9 | 99.5 | 99.9 | — | — | — | |
| DPT-HybridBackbone=ViT-B+R-50-C1C22022.03 | 0.062 | — | 2.573 | 0.092 | 95.9 | 99.5 | 99.9 | — | — | — | |
| DPTEvaluation protocol=Fine-tuned2024.01 | 0.062 | — | 2.573 | 0.092 | 95.9 | 99.5 | 99.9 | — | — | — | |
| PGA-NetSup=D, Data=K2021.03 | 0.063 | 0.267 | 2.634 | 0.101 | 95.2 | 99.2 | 99.8 | — | — | — | |
| FusionDepth (Refined Depth + GDC)Train Mode=Monocular + Sparse LiDAR, Iterative Correction (†)=true, Iterative Refinement Section=true, Resolution=640x1922021.09 | 0.063 | 0.364 | 3.291 | 0.139 | 94.5 | 97.8 | 98.8 | — | — | — | |
| PGA-NetBackbone=ResNet-502022.03 | 0.063 | 0.267 | 2.634 | 0.101 | 95.2 | 99.2 | 99.8 | — | — | — | |
| Flex-X-Large (Ours)Params=32M, GFLOPs=25, Res.=640×192, Align=LS, GT Type=Imp-GT2026.07 | 0.063 | 0.328 | 3.188 | 0.102 | 95.2 | 99.1 | 99.8 | — | — | — | |
| TransDepth w/ ViTSup=M, Data=K2021.03 | 0.064 | 0.258 | 2.761 | 0.099 | 95.5 | 99.3 | 99.9 | — | — | — | |
| TransDepth (Full)Sup=M, Data=K2021.03 | 0.064 | 0.252 | 2.755 | 0.098 | 95.6 | 99.4 | 99.9 | — | — | — | |
| TransDepthCap=0-80m2023.02 | 0.064 | 0.252 | 2.755 | 0.098 | 95.6 | 99.4 | 99.9 | — | — | — | |
| TransDepthCap=0-80m2023.09 | 0.064 | 0.252 | 2.755 | 0.098 | 95.6 | 99.4 | 99.9 | — | — | — | |
| TransDepthBackbone=ViT-B+ResNet-502022.03 | 0.064 | 0.252 | 2.755 | 0.098 | 95.6 | 99.4 | 99.9 | — | — | — | |
| DepthGenMethod category=diffusion-based2023.12 | 0.064 | — | 2.985 | — | 95.3 | 99.1 | 99.8 | — | — | — | |
| TransDepth2024.12 | 0.064 | — | — | — | 95.6 | — | — | — | — | — | |
| TransDepth w/ AGDSup=M, Data=K2021.03 | 0.065 | 0.261 | 2.766 | 0.101 | 95.3 | 99.3 | 99.8 | — | — | — | |
| FusionDepth (Initial Depth + GDC)Train Mode=Monocular + Sparse LiDAR, Iterative Correction (†)=true, Iterative Refinement Section=true, Resolution=640x1922021.09 | 0.067 | 0.423 | 3.42 | 0.144 | 94.1 | 97.7 | 98.8 | — | — | — | |
| FAL-NetSup=S, Data=K+CS2021.03 | 0.068 | 0.276 | 2.906 | 0.106 | 94.4 | 99.1 | 99.8 | — | — | — | |
| Flex-X-Large (Ours)Params=32M, GFLOPs=25, Res.=640×192, Align=Med, GT Type=Imp-GT2026.07 | 0.068 | 0.32 | 3.136 | 0.106 | 94.9 | 99.1 | 99.8 | — | — | — | |
| DA2 (ViT-L)Params=335M, GFLOPs=1028, Res.=1246×378, Align=LS, GT Type=Imp-GT2026.07 | 0.07 | 0.376 | 3.178 | 0.104 | 95.6 | 99.2 | 99.8 | — | — | — | |
| DA2 (ViT-L)Params=335M, GFLOPs=1947, Res.=1722×518, Align=LS, GT Type=Imp-GT2026.07 | 0.07 | 0.347 | 3.083 | 0.104 | 95.6 | 99.3 | 99.8 | — | — | — | |
| DSNTraining Input/Auxiliar Info=SI/GT and SN, Cap=0 − 50 m2020.10 | 0.071 | 0.267 | 2.351 | 0.111 | 94.4 | 98.9 | 99.7 | — | — | — | |
| PackNetSup=V, Data=K+CS2021.03 | 0.071 | 0.359 | 3.153 | 0.109 | 94.4 | 99 | 99.7 | — | — | — | |
| P3DepthCap=0-80m2023.02 | 0.071 | 0.27 | 2.842 | 0.103 | 95.3 | 99.3 | 99.8 | — | — | — | |
| Fu et al.Cap=0-50m2023.02 | 0.071 | 0.268 | 2.271 | 0.116 | 93.6 | 98.5 | 99.5 | — | — | — | |
| P3DepthCap=0-80m2023.09 | 0.071 | 0.27 | 2.842 | 0.103 | 95.3 | 99.3 | 99.8 | — | — | — | |
| P3DepthEvaluation protocol=Fine-tuned2024.01 | 0.071 | — | 2.842 | 0.103 | 95.3 | 99.3 | 99.8 | — | — | — | |
| P3DepthType=Regression2026.04 | 0.071 | 0.27 | 2.842 | 0.103 | 95.3 | 99.3 | 99.8 | — | — | -36.56 | |
| P3DepthType=Regression, Note=Second listing in table2026.04 | 0.071 | 0.27 | 2.842 | 0.103 | 95.3 | 99.3 | 99.8 | — | — | -36.57 | |
| DORN et al.Supervision=Depth (Supervised)2020.01 | 0.072 | 0.307 | 2.727 | 0.12 | 93.2 | 98.4 | 99.4 | — | — | — | |
| DORNSupervision=Depth (D)2020.03 | 0.072 | 0.307 | 2.727 | 0.12 | 93.2 | 98.4 | 99.4 | — | — | — | |
| VNLTraining Input/Auxiliar Info=SI/GT Depth and VN, Cap=0 − 80 m2020.10 | 0.072 | — | 3.258 | 0.117 | 93.8 | 99 | 99.8 | — | — | — | |
| DORNTraining Input/Auxiliar Info=SI/GT Depth, Cap=0 − 80 m2020.10 | 0.072 | 0.307 | 2.727 | 0.12 | 93.2 | 98.4 | 99.4 | — | — | — | |
| DORNSup=D, Data=K2021.03 | 0.072 | 0.307 | 2.727 | 0.12 | 93.2 | 98.4 | 99.4 | — | — | — | |
| Yin et al.Sup=M, Data=K2021.03 | 0.072 | — | 3.258 | 0.117 | 93.8 | 99 | 99.8 | — | — | — |