Stereo Depth Estimation on SQUID zero-shot
0.0705Relative Error (Rel)StereoAdapter-2
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| StereoAdapter-2Training Set=UW-StereoDepth-80K2026.02 | 0.0705 | 0.6396 | 1.7481 | 0.1285 | 94.25 | 97.65 | 98.62 | |
| StereoAdapterTraining Set=UW-StereoDepth-40K2026.02 | 0.0806 | 0.7082 | 1.8843 | 0.1469 | 94.13 | 97.48 | 98.52 | |
| RAFT-StereoTraining Set=Scene Flow2026.02 | 0.0831 | 0.6946 | 1.9625 | 0.1441 | 92.35 | 96.34 | 98.35 | |
| IGEV-StereoTraining Set=5 Datasets* + TartanAir2026.02 | 0.0932 | 1.4685 | 2.4741 | 0.1523 | 93.46 | 97.12 | 98.2 | |
| Stereo AnywhereTraining Set=Scene Flow2026.02 | 0.0952 | 1.1017 | 2.4317 | 0.1586 | 91.79 | 96.05 | 97.63 | |
| Selective IGEVTraining Set=5 Datasets* + TartanAir2026.02 | 0.096 | 0.9617 | 1.9268 | 0.1665 | 91.71 | 95.55 | 97.2 | |
| FoundationStereoTraining Set=FoundationStereo dataset2026.02 | 0.1095 | 0.7012 | 2.251 | 0.1584 | 89.95 | 94.33 | 95.01 | |
| GwcNetTraining Set=Scene Flow2026.02 | 0.2294 | 1.2275 | 3.0003 | 0.3799 | 74.23 | 85.17 | 90.05 | |
| PSMNetTraining Set=Scene Flow2026.02 | 0.5182 | 7.1404 | 4.9186 | 0.5902 | 71.39 | 79.99 | 83.11 | |
| LEAStereoTraining Set=Scene Flow2026.02 | 0.5574 | 3.9434 | 5.4659 | 0.4335 | 65.12 | 80.42 | 88.69 | |
| HSMNetTraining Set=Scene Flow2026.02 | 0.9772 | 7.2766 | 8.2301 | 4.0887 | 0 | 0 | 0 | |
| TiO-DepthTraining Set=KITTI20122026.02 | 1.3154 | 11.6828 | 7.093 | 0.8121 | 17.53 | 33.46 | 51.33 | |
| ACVNetTraining Set=Scene Flow2026.02 | 1.603 | 65.6518 | 10.3828 | 0.7293 | 70.19 | 79.25 | 83.21 | |
| CREStereoTraining Set=ETH3D2026.02 | 2.5746 | 9.8789 | 8.4526 | 5.1297 | 48.9 | 57.32 | 70.01 | |
| GMStereoTraining Set=5 Datasets* + TartanAir2026.02 | 3.3442 | 140.3211 | 18.7829 | 1.0219 | 53 | 60.76 | 65.78 | |
| AANetTraining Set=Scene Flow2026.02 | 7.4801 | 314.1577 | 34.7612 | 1.8994 | 6.02 | 10.87 | 15.7 |