Landmark Detection on Human public femoral landmarks dataset (test)
1.44SGTDGCNN
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DGCNN2026.02 | 1.44 | 2.11 | 3.36 | 10.01 | 2.87 | 8.74 | 2.88 | 2.89 | 13.81 | 1.88 | 5 | |
| LmPT-v2Backbone=PTv22026.02 | 1.58 | 1.57 | 2.98 | 3.05 | 2.33 | 3.67 | 2.4 | 2.9 | 3.31 | 1.62 | 2.54 | |
| LmPT-v3Backbone=PTv32026.02 | 3.41 | 3.86 | 4.39 | 2.69 | 3.68 | 3.55 | 2.89 | 4.86 | 5.96 | 3.38 | 3.87 | |
| AME (Average manual error)2026.02 | 4.1 | 2.87 | 4.91 | 2.8 | 2.66 | 3.63 | 3.91 | 2.5 | 5.14 | 3.08 | 3.56 | |
| A&A2026.02 | 6.94 | 6.32 | 29.32 | 1.76 | 3.37 | 3.88 | 4.1 | 1.37 | 5.37 | 2.71 | 6.51 | |
| MME (Maximum manual error)2026.02 | 18.77 | 8.5 | 14.64 | 10.58 | 10.06 | 11.2 | 13.99 | 9.05 | 13.24 | 9.86 | 11.99 |