NYU Depth
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Depth Estimation | NYU Depth v2 | RMSE0.1232 | 209 | |
| Depth Completion | NYU-depth-v2 official (test) | RMSE0.085 | 200 | |
| Semantic Segmentation | NYU Depth V2 (test) | mIoU61.2 | 183 | |
| Depth Prediction | NYU-depth v2 (test) | Accuracy (δ < 1.25)99.5 | 113 | |
| Monocular Depth Estimation | NYU-Depth v2 (official) | Abs Rel0.059 | 75 | |
| Monocular Depth Estimation | NYU Depth Eigen v2 (test) | A.Rel0.072 | 49 | |
| Depth Completion | NYU-Depth v2 | RMSE0.089 | 43 | |
| Edge accuracy | NYU-Depth V2 (test) | Precision79.4 | 30 | |
| Semantic Segmentation | NYU-depth v2 | mIoU58.8 | 28 | |
| Monocular Depth Estimation | NYU-Depth v2 (val) | A.Rel3.9 | 27 | |
| Semantic Segmentation | NYU Depth V2 | mIoU68.1 | 27 | |
| Monocular Depth Estimation | NYU Depth V2 | Recall0.641 | 18 | |
| Object Detection | NYU Depth v2 (test) | mAP55.4 | 17 | |
| Monocular Depth Estimation | NYU-Depth V2 (test) | AbsRel4.3 | 13 | |
| Indoor Depth Completion | NYU-Depth Setting C: T* -> T V2 (test) | RMSE0.09 | 11 | |
| Vanishing Point Detection | NYU Depth (test) | AUC @10°71.62 | 11 | |
| Monocular Depth Estimation | NYU Depth v2 (Underrepresented classes) | Delta 197.5 | 10 | |
| Geometric Correspondence | NYU-Depth V2 | RGB-to-D Correspondence Score33.5 | 10 | |
| Indoor Scene Recognition | NYU Depth v2 (test) | Mean Class Accuracy70.4 | 9 | |
| Indoor Depth Completion | NYU-Depth Setting A: R -> T V2 (test) | RMSE0.12 | 9 | |
| Single Image Dehazing | NYU-Depth D-HAZY (full) | PSNR15.54 | 9 | |
| Indoor Depth Completion | NYU-Depth Setting B: R* -> T V2 (test) | RMSE0.242 | 8 | |
| Depth Completion | NYU-Depth Setting C: T* ⇒ T V2 (test) | CD47.39 | 7 | |
| Depth Completion | NYU-Depth Setting B: R* ⇒ T V2 (test) | CD249.05 | 7 | |
| Depth Completion | NYU-Depth Setting A: R ⇒ T V2 (test) | CD33.15 | 7 |
Showing 25 of 44 rows