Face pulling recognition on Clinical seizure videos
61.1AccuracyQwen
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| QwenModel Type=VLM, Signal Enhancement=None2026.05 | 61.1 | 31.2 | 17.2 | 22.2 | |
| Crop+InternModel=Intern, Signal Enhancement=Crop2026.05 | 48.9 | 37.3 | 86.2 | 52.1 | |
| Crop+QwenModel=Qwen, Signal Enhancement=Crop2026.05 | 47.8 | 29.5 | 44.8 | 35.6 | |
| InternModel Type=VLM, Signal Enhancement=None2026.05 | 41.1 | 23.9 | 37.9 | 29.3 | |
| CNNModel Type=Traditional2026.05 | 33.3 | 32 | 92.6 | 46.3 | |
| ViViTModel Type=Traditional2026.05 | 31.1 | 30.9 | 92.6 | 45.3 |