Depth Estimation on TartanAir old town structured
0.256Abs RelM4Depth
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| M4DepthTest size=384 × 512, Zero-shot generalization (trained on KITTI)=true2021.05 | 0.256 | 6.759 | 7.211 | 0.37 | 80.4 | 88 | 91.8 | |
| PWCDC-NetTest size=384 × 512, Zero-shot generalization (trained on KITTI)=true2021.05 | 0.338 | 8.679 | 16.76 | 0.703 | 62.7 | 74.1 | 80 | |
| DeepV2DTest size=384 × 512, Zero-shot generalization (trained on KITTI)=true2021.05 | 0.694 | 18.777 | 7.551 | 0.498 | 49.4 | 72.2 | 83 | |
| ManyDepthTest size=320 × 1024, Zero-shot generalization (trained on KITTI)=true2021.05 | 0.759 | 20.895 | 15.604 | 0.649 | 35.1 | 58.3 | 73 | |
| Monodepth2Test size=320 × 1024, Zero-shot generalization (trained on KITTI)=true2021.05 | 0.775 | 17.973 | 15.8 | 0.727 | 32.2 | 54 | 69.2 | |
| MonodepthTest size=384 × 512, Zero-shot generalization (trained on KITTI)=true2021.05 | 0.909 | 30.616 | 19.203 | 0.782 | 26.7 | 48.9 | 65.5 |