Table 1 Quantitative comparisons on the KITTI and NYUv2 datasets. Bold type indicates the best performance and underline indicates the second best performance.
From: RGB-conditioned frequency domain refinement for sparse-to-dense depth completion
Method | Param. [M] | KITTI | NYUv2 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
RMSE [mm]Â \(\downarrow\) | MAE [mm]Â \(\downarrow\) | iRMSE [1/km]Â \(\downarrow\) | iMAE [1/km]Â \(\downarrow\) | RMSE [m]Â \(\downarrow\) | REL\(\downarrow\) | \(\varvec{\delta _{1.25}}\uparrow\) | \(\varvec{\delta _{1.25^2}}\uparrow\) | \(\varvec{\delta _{1.25^3}}\uparrow\) | ||
DeepLiDAR20 | 53.4 | 887.00 | 215.38 | 2.51 | 1.10 | 0.115 | 0.022 | 99.3 | 99.9 | 100.0 |
DySPN38 | 26 | 878.50 | 228.60 | 2.50 | 1.00 | 0.090 | 0.012 | 99.6 | 99.9 | 100.0 |
SpAgNet47 | 51 | 844.79 | 218.39 | 2.39 | 0.91 | 0.114 | 0.015 | 99.3 | 99.9 | 100.0 |
NLSPN48 | 25.8 | 771.80 | 197.30 | 2.00 | 0.80 | 0.092 | 0.012 | 99.6 | 99.9 | 100.0 |
GuideNet15 | 73.5 | 777.78 | 221.59 | 2.39 | 1.00 | 0.101 | 0.015 | 99.5 | 99.9 | 100.0 |
CFormer49 | 83.5 | 741.44 | 194.99 | 2.03 | 0.85 | 0.091 | 0.012 | 99.6 | 99.9 | 100.0 |
LRRU45 | 21 | 729.50 | 188.80 | 1.90 | 0.80 | 0.091 | 0.011 | 99.6 | 99.9 | 100.0 |
OGNI-DC50 | 83.4 | 747.64 | 182.29 | 1.81 | 0.79 | 0.087 | 0.011 | 99.6 | 99.9 | 100.0 |
Ours | 36.3 | 721.46 | 184.53 | 1.86 | 0.79 | 0.085 | 0.011 | 99.6 | 99.9 | 100.0 |
