Monocular Depth Estimation on DDAD outdoor (test)
99Delta < 1.25^3 AccuracyLinear Fit
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Linear FitBackbone=DPT-Hybrid, Scaling=Linear Fit, Train=DDAD (oracle)2026.01 | 99 | 0.163 | 10.342 | |
| OursBackbone=DPT-Hybrid, Scaling=Ours, Train=NYUv2, KITTI2026.01 | 98.4 | 0.167 | 12.847 | |
| Linear FitBackbone=DepthAnything-S, Scaling=Linear Fit, Train=DDAD (oracle)2026.01 | 98.3 | 0.182 | 18.423 | |
| RSABackbone=DPT-Hybrid, Scaling=RSA, Train=NYUv2, KITTI2026.01 | 98.1 | 0.171 | 13.539 | |
| OursBackbone=DepthAnything-S, Scaling=Ours, Train=NYUv2, KITTI2026.01 | 97.9 | 0.195 | 19.054 | |
| RSABackbone=DepthAnything-S, Scaling=RSA, Train=NYUv2, KITTI2026.01 | 97.6 | 0.207 | 19.715 | |
| ImageBackbone=DPT-Hybrid, Scaling=Image, Train=NYUv2, KITTI2026.01 | 97.5 | 0.179 | 14.468 | |
| GlobalBackbone=DPT-Hybrid, Scaling=Global, Train=NYUv2, KITTI2026.01 | 96.9 | 0.183 | 15.967 | |
| ImageBackbone=DepthAnything-S, Scaling=Image, Train=NYUv2, KITTI2026.01 | 96.8 | 0.217 | 20.834 | |
| GlobalBackbone=DepthAnything-S, Scaling=Global, Train=NYUv2, KITTI2026.01 | 96.3 | 0.221 | 21.345 |