Engagement estimation on PAFE
78Weighted F1PriorNet
Evaluation results
| Method | Links | |
|---|---|---|
| PriorNetPrinciple=Prior-guided2026.05 | 78 | |
| TCCT-Net + VLMPrinciple=VLM-assisted filtering2026.05 | 74 | |
| TCCT-NetPrinciple=Hybrid temporal2026.05 | 72 | |
| DNNPrinciple=Neural features, Window=20 sec2026.05 | 68 | |
| XGBoostPrinciple=Boosted features, Window=20 sec2026.05 | 65 | |
| XGBoostPrinciple=Boosted features, Window=10 sec2026.05 | 64 | |
| DNNPrinciple=Neural features, Window=10 sec2026.05 | 64 | |
| XGBoostPrinciple=Boosted features, Window=5 sec2026.05 | 62 | |
| DNNPrinciple=Neural features, Window=5 sec2026.05 | 62 | |
| SVMPrinciple=Classical features, Window=10 sec2026.05 | 61 | |
| SVMPrinciple=Classical features, Window=20 sec2026.05 | 60 | |
| SVMPrinciple=Classical features, Window=5 sec2026.05 | 57 |