Action Assessment on Rhythmic Gymnastics
52.8Score (Ball)ACTION-NET
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ACTION-NET2020.08 | 52.8 | 65.7 | 70.8 | 57.8 | |
| Xu et al.2020.08 | 51.5 | 62.1 | 54 | 52.2 | |
| ResNet LSTMBackbone=ResNet, Regression/Pooling type=LSTM2020.08 | 47.1 | 47.2 | 55.5 | 41 | |
| ResNet Bi-LSTMBackbone=ResNet, Regression/Pooling type=Bi-LSTM2020.08 | 46.6 | 44.4 | 54.4 | 44.7 | |
| I3D Max + SVRBackbone=I3D, Regression/Pooling type=Max + SVR2020.08 | 44.5 | 46.5 | 38.6 | 19.1 | |
| I3D Avg + SVRBackbone=I3D, Regression/Pooling type=Avg + SVR2020.08 | 40.3 | 57.9 | 61 | 45.6 | |
| C3D + SVR2020.08 | 35.7 | 55.1 | 49.5 | 51.6 | |
| I3D Bi-LSTMBackbone=I3D, Regression/Pooling type=Bi-LSTM2020.08 | 35.5 | 65.3 | 55.7 | 36.8 | |
| I3D LSTMBackbone=I3D, Regression/Pooling type=LSTM2020.08 | 32.7 | 58.3 | 54.6 | 38.1 | |
| ResNet Avg + SVRBackbone=ResNet, Regression/Pooling type=Avg + SVR2020.08 | 27.9 | 58.9 | 60.2 | 39.5 | |
| ResNet Max + SVRBackbone=ResNet, Regression/Pooling type=Max + SVR2020.08 | 16.9 | 45.2 | 37.6 | 40.4 |