Skeleton-based Action Recognition on SBU Interaction Dataset (5-fold cross validation)
93.3AccuracyST-LSTM
Evaluation results
| Method | Links | |
|---|---|---|
| ST-LSTMTraversal=Tree, Trust Gate=true2016.07 | 93.3 | |
| ST-LSTM (Tree) + Trust GateFeature=Geometric, Structure=Tree, Trust Gate=true2017.06 | 93.3 | |
| Co-occurrence LSTM2016.07 | 90.4 | |
| Co-occurrence LSTMFeature=Geometric2017.06 | 90.4 | |
| ST-LSTMTraversal=Tree2016.07 | 88.6 | |
| ST-LSTM (Tree)Feature=Geometric, Structure=Tree2017.06 | 88.6 | |
| Ji et al.2016.07 | 86.9 | |
| Ji et al.Feature=Geometric2017.06 | 86.9 | |
| Deep LSTMReported by=[51]2016.07 | 86 | |
| Deep LSTMFeature=Geometric2017.06 | 86 | |
| ST-LSTMTraversal=Joint Chain2016.07 | 84.7 | |
| ST-LSTM (Joint Chain)Feature=Geometric, Structure=Joint Chain2017.06 | 84.7 | |
| CHARM2016.07 | 83.9 | |
| CHARMFeature=Geometric2017.06 | 83.9 | |
| HBRNNReported by=[51]2016.07 | 80.4 | |
| Hierarchical RNNFeature=Geometric2017.06 | 80.4 | |
| Yun et al.2016.07 | 80.3 | |
| Yun et al.Feature=Geometric2017.06 | 80.3 |