Table 3 The results of our method are compared with state-of-the-art methods in terms of ADD(-S) on the Occlusion Linemod dataset. Objects in bold names are symmetric.
From: Enhancing object pose estimation for RGB images in cluttered scenes
| Â | EfficientPose24 | Zebrapose21 | Ullah et al.26 | Deepfusion14 | 6D-diff23 | RDPN6D53 | OURS |
|---|---|---|---|---|---|---|---|
Ape | 59.39 | 57.9 | 56.57 | 64.1 | 60.6 | 64.6 | 66.46 |
Can | 93.27 | 95.0 | 91.12 | 96.1 | 97.9 | 97.0 | 94.73 |
Cat | 79.78 | 60.6 | 68.58 | 52.2 | 63.2 | 54.8 | 65.41 |
Driller | 97.77 | 94.8 | 95.64 | 95.8 | 96.6 | 93.1 | 97.09 |
Duck | 72.71 | 64.5 | 65.31 | 72.3 | 67.2 | 68.8 | 78.85 |
Eggbox | 96.18 | 70.9 | 93.46 | 75.3 | 73.5 | 78.1 | 95.07 |
Glue | 90.80 | 88.7 | 85.15 | 79.3 | 92.0 | 83.5 | 90.28 |
Holepuncher | 81.95 | 83.0 | 76.53 | 86.8 | 85.5 | 96.1 | 87.18 |
Average | 83.98 | 76.9 | 79.04 | 77.7 | 79.6 | 79.5 | 84.38 |
