Table 11 Generalization and robustness analysis on the basketball-action dataset.
From: Learning spatio-temporal context for basketball action pose estimation with a multi-stream network
Model | Sideline view (MAP) (%) | Baseline view (MAP) (%) | High-angle view (MAP) (%) | Overall MAP (%) |
|---|---|---|---|---|
AlphaPose | 65.1 | 61.8 | 66.8 | 64.5 |
MDPN | 66.4 | 63.2 | 68.9 | 66.2 |
Dynamic-GNN | 68.9 | 66.5 | 71.1 | 68.3 |
MSTCFL | 70.1 | 68.6 | 72.4 | 70.3 |
Ours | 76.5 | 74.2 | 77.6 | 76.1 |
