Real-world dataset
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Part-level Pose Estimation | Real-world dataset | mAP (5° 5cm)37.38 | 24 | |
| Articulated Object Joint Estimation | Real world dataset Seen categories 1.0 (seen_objects) | Axis Orientation Error (°)13.88 | 12 | |
| Question Answering | Real-World Dataset A | Scenario 1 Score88.9 | 9 | |
| Image Deraining | Real-world dataset (test) | PSNR49 | 8 | |
| 3D tomographic reconstruction | Real-World Dataset (FIPS) (36-view) | PSNR37.48 | 7 | |
| 3D tomographic reconstruction | Real-World Dataset FIPS (24-view) | PSNR36.399 | 7 | |
| Novel View Synthesis | Real-world dataset Subsampled 5x Speed | PSNR (Full Image)28.666 | 6 | |
| Novel View Synthesis | Real-world dataset Original | PSNR (Full Image)29.776 | 6 | |
| Articulated Object Joint Estimation | Real world dataset Unseen categories 1.0 | Axis Orientation Error (°)14.18 | 6 | |
| Initial state generalization | Real-world dataset Rabbit | PSNR30.35 | 5 | |
| Initial state generalization | Real-world dataset Peanut | PSNR33.01 | 5 | |
| Initial state generalization | Real-world dataset Mandarin | PSNR32.83 | 5 | |
| Initial state generalization | Real-world dataset Chick1 | PSNR32.05 | 5 | |
| Novel View Synthesis | Real-world Dataset (test) | CLIP-S Score98.5 | 5 | |
| Tactile Recognition | Augmented Real-World Dataset | Accuracy88.89 | 5 | |
| 4D scene generation | Real-world Dataset | PSNR22.53 | 4 | |
| Physical Simulation Realism | Real-world dataset (PhysDreamer, Instant-NGP, NeRFStudio, Mip-NeRF 360, etc.) (test) | Physical Realism3.44 | 4 | |
| 3D Human Pose Estimation | Real-world dataset (test) | PA-MPJPE (Upper Body)34 | 4 | |
| Egocentric Pose Estimation | Real-world dataset (test) | Ecam0.475 | 4 |
Showing 19 of 19 rows