Depth Prediction on KITTI original (Eigen split)
0.063Abs RelFusionDepth (Refined Depth + GDC)
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| FusionDepth (Refined Depth + GDC)Train=Monocular + LiDAR (M+L), Resolution=640x192, Iterative correction=true, Refinement=Iterative, Post-processing=GDC2021.09 | 0.063 | 0.364 | 3.291 | 0.139 | 94.5 | 97.8 | 98.8 | |
| FusionDepth (Initial Depth + GDC)Train=Monocular + LiDAR (M+L), Resolution=640x192, Iterative correction=true, Post-processing=GDC2021.09 | 0.067 | 0.423 | 3.42 | 0.144 | 94.1 | 97.7 | 98.8 | |
| FusionDepth (Refined Depth)Train=Monocular + LiDAR (M+L), Resolution=640x192, Refinement=Iterative2021.09 | 0.074 | 0.423 | 3.61 | 0.15 | 93.6 | 97.3 | 98.6 | |
| FusionDepth (Initial Depth)Train=Monocular + LiDAR (M+L), Resolution=640x192, Refinement=None2021.09 | 0.078 | 0.515 | 3.67 | 0.154 | 93.5 | 97.3 | 98.6 | |
| Guizilini et al.Train=Monocular + LiDAR (M+L), Resolution=640x192, Note=Only use LiDAR data in training phase, but tested on the KITTI improved dataset2021.09 | 0.082 | 0.424 | 3.73 | 0.131 | 91.7 | — | — | |
| BTSTrain=Monocular + Supervised (M+Sup), Resolution=640x1922021.09 | 0.091 | 0.555 | 4.033 | 0.174 | 90.4 | 96.7 | 98.4 | |
| DynamicDepthTest frames=2 (-1, 0), WxH=640 x 1922022.03 | 0.096 | 0.72 | 4.458 | 0.175 | 89.7 | 96.4 | 98.4 | |
| MonoPL++Train=Monocular + LiDAR (M+L), Resolution=640x192, Iterative correction=true2021.09 | 0.098 | 0.714 | 4.3 | 0.176 | 89.9 | 96.7 | 98.4 | |
| ManyDepthTest frames=2 (-1, 0), WxH=640 x 1922022.03 | 0.098 | 0.77 | 4.459 | 0.176 | 90 | 96.5 | 98.3 | |
| DornTrain=Monocular + Supervised (M+Sup), Resolution=640x1922021.09 | 0.099 | 0.593 | 3.714 | 0.161 | 89.7 | 96.6 | 98.6 | |
| GLNetTrain=Monocular (M), Resolution=640x1922021.09 | 0.099 | 0.796 | 4.743 | 0.186 | 88.4 | 95.5 | 97.9 | |
| Guizilini et al.Test frames=1, WxH=640 x 1922022.03 | 0.102 | 0.698 | 4.381 | 0.178 | 89.6 | 96.4 | 98.4 | |
| MonoDepth2Train=Monocular + Stereo (M+S), Resolution=640x1922021.09 | 0.106 | 0.818 | 4.75 | 0.196 | 87.4 | 95.7 | 97.9 | |
| Johnston et al.Test frames=1, WxH=640 x 1922022.03 | 0.106 | 0.861 | 4.699 | 0.185 | 88.9 | 96.2 | 98.2 | |
| Wang et al.Test frames=2 (-1, 0), WxH=640 x 1922022.03 | 0.106 | 0.799 | 4.662 | 0.187 | 88.9 | 96.1 | 98.2 | |
| Struct2DepthTrain=Monocular (M), Resolution=640x1922021.09 | 0.109 | 0.825 | 4.75 | 0.187 | 87.4 | 95.8 | 98.3 | |
| PackNet-SfMTrain=Monocular (M), Resolution=640x1922021.09 | 0.111 | 0.785 | 4.601 | 0.189 | 87.8 | 96 | 98.2 | |
| Packnet-SFMTest frames=1, WxH=640 x 1922022.03 | 0.111 | 0.785 | 4.601 | 0.189 | 87.8 | 96 | 98.2 | |
| Patil et al.Test frames=N, WxH=640 x 1922022.03 | 0.111 | 0.821 | 4.65 | 0.187 | 88.3 | 96.1 | 98.2 | |
| InstaDMTest frames=1, WxH=832 x 2562022.03 | 0.112 | 0.777 | 4.772 | 0.191 | 87.2 | 95.9 | 98.2 | |
| Lee et al.Test frames=1, WxH=832 x 2562022.03 | 0.114 | 0.876 | 4.715 | 0.191 | 87.2 | 95.5 | 98.1 | |
| Monodepth2Test frames=1, WxH=640 x 1922022.03 | 0.115 | 0.903 | 4.863 | 0.193 | 87.7 | 95.9 | 98.1 | |
| Videos in the wildTest frames=1, WxH=416 x 1282022.03 | 0.128 | 0.959 | 5.23 | 0.212 | 84.5 | 94.7 | 97.6 | |
| Li et al.Test frames=1, WxH=416 x 1282022.03 | 0.13 | 0.95 | 5.138 | 0.209 | 84.3 | 94.8 | 97.8 | |
| MonoDepthTrain=Stereo (S), Resolution=640x1922021.09 | 0.133 | 1.142 | 5.533 | 0.23 | 83 | 93.6 | 97 | |
| EPC++Test frames=1, WxH=832 x 2562022.03 | 0.141 | 1.029 | 5.35 | 0.216 | 81.6 | 94.1 | 97.6 | |
| Struct2depth (M)Test frames=1, WxH=416 x 1282022.03 | 0.141 | 1.026 | 5.291 | 0.215 | 81.6 | 94.5 | 97.9 | |
| Ranjan et al.Test frames=1, WxH=832 x 2562022.03 | 0.148 | 1.149 | 5.464 | 0.226 | 81.5 | 93.5 | 97.3 | |
| LEGOTrain=Monocular (M), Resolution=640x1922021.09 | 0.162 | 1.352 | 6.276 | 0.252 | 78.3 | 92.1 | 96.9 |