YCB
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Lossless compression | YCB-Slide | Bits per Byte1.073 | 17 | |
| Single-reference novel object pose estimation | YCB-V | mAP83.1 | 15 | |
| Video Reconstruction | YCB-Sim (test) | PSNR27.71 | 15 | |
| Stable Placement Prediction | YCB | Rotational Error (deg)2.5 | 10 | |
| 6-DoF Grasping | YCB (held-out objects) | Success Rate88.2 | 10 | |
| 6D Object Pose Tracking | YCB Synthetic tomato_fast | RMSE (Translation, ep)0.89 | 9 | |
| 6D Object Pose Tracking | YCB Synthetic cracker_fast | Translation RMSE (ep)0.9 | 9 | |
| 6D Object Pose Tracking | YCB Synthetic potted_fast | Translation RMSE (ep)1.04 | 9 | |
| 6D Object Pose Tracking | YCB Synthetic tomato_regular | Translation RMSE (ep)0.48 | 9 | |
| 6D Object Pose Tracking | YCB Synthetic cracker_regular | Translation RMSE (ep)0.4 | 9 | |
| 6D Object Pose Tracking | YCB Synthetic potted_regular | Translation RMSE0.56 | 9 | |
| 6D Object Pose Tracking | YCB Synthetic mustard_regular | Translation RMSE0.77 | 9 | |
| 6D Object Pose Refinement | YCB-V | Avg. Success82.6 | 9 | |
| Tactile Registration | YCB Reg | Rotational Error (deg)0.94 | 8 | |
| Dexterous Grasping | YCB (Unseen Hands) | LEAP Score93 | 7 | |
| Dexterous Grasping | YCB (Seen Hands) | Allegro Score92 | 7 | |
| Hand-only Grasp Data Generation | YCB 043-phillips-screwdriver | GWS Epsilon0.0027 | 7 | |
| Hand-only Grasp Data Generation | YCB 005-tomato-soup-can | GWS Epsilon0.0194 | 7 | |
| Hand-only Grasp Data Generation | YCB 004-sugar-box | GWS Epsilon0.0092 | 7 | |
| Object Identification | YCB-Slide | Top-1 Accuracy96.9 | 6 | |
| Lossy compression performance (video-based settings) | YCB-Slide (test) | BD-Rate-65.5 | 6 | |
| 6D Object Pose Estimation | YCB-V | Master Chef Can Error3.6 | 5 | |
| 6D Pose Estimation | YCB-V | AR VSD77.2 | 5 | |
| 6D Pose Estimation | YCB-V | AR49.86 | 5 | |
| Hand-only Grasp Data Generation | YCB 062-dice | Validation Success Rate0 | 3 |
Showing 25 of 37 rows