Violence Detection on Hockey Fight
99.62Accuracy3D ConvNet
Evaluation Results
| Method | Links | |
|---|---|---|
| 3D ConvNetType=Deep-Learning Based2019.11 | 99.62 | |
| Hachiuma et al.2024.12 | 99.5 | |
| Zhang at al.2024.12 | 99.4 | |
| Islam et al.2024.12 | 99 | |
| DIFEMclassifier=Random Forest, protocol=5-fold cross validation2024.12 | 98.8 | |
| SSHAInput modality=RGB2022.02 | 98.7 | |
| DIFEMclassifier=Nearest Neighbor, protocol=5-fold cross validation2024.12 | 98.7 | |
| I3D(RGB only)Type=Deep-Learning Based2019.11 | 98.5 | |
| I3DInput modality=RGB2022.02 | 98.5 | |
| DIFEMclassifier=AdaBoost, protocol=5-fold cross validation2024.12 | 98.5 | |
| Flow Gated NetworkType=Deep-Learning Based2019.11 | 98 | |
| Cheng et al. ModelArchitecture=P3D2022.02 | 98 | |
| SSHAInput modality=RGB, Localization=Disabled2022.02 | 98 | |
| Cheng et al.2024.12 | 98 | |
| I3D(Fusion)Type=Deep-Learning Based2019.11 | 97.5 | |
| I3DInput modality=Two-stream2022.02 | 97.5 | |
| ConvLSTMType=Deep-Learning Based2019.11 | 97.1 | |
| ConvLSTM2022.02 | 97.1 | |
| DIFEMclassifier=Decision Tree, protocol=5-fold cross validation2024.12 | 97.1 | |
| FightNetType=Deep-Learning Based2019.11 | 97 | |
| SSHAInput modality=Two-stream2022.02 | 97 | |
| SPIL2024.12 | 96.8 | |
| C3DType=Deep-Learning Based2019.11 | 96.5 | |
| LHOG+LOFType=Hand-Crafted Features2019.11 | 95.1 | |
| LHOG+LOF2024.12 | 95.1 | |
| Garcia-Cobo et al.2024.12 | 94.5 | |
| I3Devaluation_source=Reported in [24]2024.12 | 93.4 | |
| MoWLD+BOWType=Hand-Crafted Features2019.11 | 91.9 | |
| MoWLD+BoW2024.12 | 91.9 | |
| HOG+HIKType=Hand-Crafted Features2019.11 | 91.7 | |
| HOG+HIK2024.12 | 91.7 | |
| TSNevaluation_source=Reported in [24]2024.12 | 91.5 | |
| MOSIFT+HIKType=Hand-Crafted Features2019.11 | 90.9 | |
| DGCNNevaluation_source=Reported in [24]2024.12 | 90.2 | |
| HOF+HIKType=Hand-Crafted Features2019.11 | 88.6 | |
| ViF + OViF2024.12 | 87.5 | |
| SSHAInput modality=Optical-flow2022.02 | 86.2 | |
| I3D(Flow only)Type=Deep-Learning Based2019.11 | 84 | |
| I3DInput modality=Optical-flow2022.02 | 84 | |
| ViFType=Hand-Crafted Features2019.11 | 82.9 | |
| ViF2024.12 | 82.9 | |
| 3D CNNevaluation_source=Reported in [24]2024.12 | 82.6 | |
| Tran et al.2024.12 | 71.39 |