Action Recognition on NTU RGB+D V2 (Cross Subject)
76.1AccuracyGCA-LSTM
Evaluation Results
| Method | Links | |
|---|---|---|
| GCA-LSTMTraining method=stepwise training2017.07 | 76.1 | |
| Visualization CNN2017.07 | 76 | |
| SkeletonNet2017.07 | 75.9 | |
| GCA-LSTMTraining method=direct training2017.07 | 74.3 | |
| JTM CNN2017.07 | 73.4 | |
| STA Model2017.07 | 73.4 | |
| ST-LSTMvariant=Global (2)2017.07 | 70.7 | |
| ST-LSTMvariant=Global (1)2017.07 | 70.5 | |
| ST-LSTM2017.07 | 69.2 | |
| Part-aware LSTM2017.07 | 62.9 | |
| Deep LSTM2017.07 | 60.7 | |
| Dynamic Skeletons2017.07 | 60.2 | |
| HBRNN2017.07 | 59.1 | |
| Deep RNN2017.07 | 56.3 | |
| Lie Group2017.07 | 50.1 | |
| Skeletal Quads2017.07 | 38.6 |