Head turning recognition on Clinical seizure videos
81.1AccuracyQwen
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| QwenModel Type=VLM, Signal Enhancement=None2026.05 | 81.1 | 57.1 | 22.2 | 32 | |
| InternModel Type=VLM, Signal Enhancement=None2026.05 | 80 | 0 | 0 | 0 | |
| Crop+InternModel=Intern, Signal Enhancement=Crop2026.05 | 80 | 0 | 0 | 0 | |
| Crop+QwenModel=Qwen, Signal Enhancement=Crop2026.05 | 76.7 | 36.4 | 22.2 | 27.6 | |
| CNNModel Type=Traditional2026.05 | 66.7 | 37.4 | 41.7 | 32.5 | |
| ViViTModel Type=Traditional2026.05 | 61.1 | 27 | 48.6 | 31.7 |