Action spotting on SoccerNet (test)
62.5Average mAPContext-aware loss function
Evaluation results
| Method | Links | |
|---|---|---|
| Context-aware loss functionFrame features=ResNet2019.12 | 62.5 | |
| Cioppa et. al2020.04 | 62.5 | |
| Multi-tower temporal convolutional networkTolerance=Averaged 5s to 60s2020.04 | 60.1 | |
| Context-aware loss functionFrame features=C3D2019.12 | 57.7 | |
| Vats et al.Frame features=ResNet2019.12 | 57.5 | |
| Context-aware loss functionFrame features=I3D2019.12 | 53.6 | |
| SoccerNet baseline 20sFrame features=ResNet2019.12 | 49.7 | |
| Giancola et. alTolerance=20s2020.04 | 49.7 | |
| SoccerNet baseline 60sFrame features=ResNet2019.12 | 40.6 | |
| Giancola et. alTolerance=60s2020.04 | 40.6 | |
| SoccerNet baseline 5sFrame features=ResNet2019.12 | 34.5 | |
| Giancola et. alTolerance=5s2020.04 | 34.5 |