Table 1 Evaluation and comparison on the Linemod dataset in terms of the ADD(-S) metric. Symmetric objects are marked with * ).
From: MagicCubePose, A more comprehensive 6D pose estimation network
Method | YOLO6D | Pix2Pose | PVNet | DPOD | HybridPose | EfficientPose(\(\varphi\)=0) | RNNPose31 | Our |
|---|---|---|---|---|---|---|---|---|
Ape | 21.62 | 58.1 | 43.62 | 53.28 | 63.1 | 87.71 | 88.19 | 87.71 |
Benchvise | 81.80 | 91.0 | 99.90 | 95.34 | 99.9 | 99.71 | 100 | 100 |
Cam | 36.57 | 60.9 | 86.86 | 90.36 | 90.4 | 97.94 | 98.04 | 98.24 |
Can | 68.80 | 84.4 | 95.47 | 94.10 | 98.5 | 98.52 | 99.31 | 99.02 |
Cat | 41.82 | 65.0 | 79.34 | 60.38 | 89.4 | 98.00 | 96.41 | 98.20 |
Driller | 63.51 | 76.3 | 96.43 | 97.72 | 98.5 | 99.90 | 99.70 | 99.80 |
Duck | 27.23 | 43,8 | 52.58 | 66.01 | 65.0 | 90.99 | 89.30 | 90.70 |
Eggbox * | 69.58 | 96.8 | 99.15 | 99.72 | 100 | 100 | 99.53 | 100 |
Glue * | 80.02 | 79.4 | 95.66 | 93.83 | 98.8 | 100 | 99.71 | 99.90 |
Holepuncher | 42.63 | 74.8 | 81.92 | 65.83 | 89.7 | 95.15 | 97.43 | 95.62 |
Iron | 74.97 | 83.4 | 98.88 | 99.80 | 100 | 99.69 | 100 | 99.69 |
Lamp | 71.11 | 82.0 | 99.33 | 88.11 | 99.5 | 100 | 99.81 | 100 |
phone | 47.74 | 45.0 | 92.41 | 74.24 | 94.9 | 97.98 | 98.39 | 98.56 |
Average | 55.95 | 72.4 | 86.27 | 82.98 | 91.3 | 97.35 | 97.37 | 97.50 |
