Detection of Asynchronous movement on Seizure Video Dataset
70AccuracyViViT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ViViTModel Architecture=ViViT, Signal Enhancement (Pose)=false2026.05 | 70 | 66.5 | 71 | 67.4 | |
| CNNModel Architecture=CNN, Signal Enhancement (Pose)=false2026.05 | 67.8 | 57.9 | 86.1 | 69 | |
| InternModel Architecture=Intern, Signal Enhancement (Pose)=false2026.05 | 65.6 | 65.6 | 51.2 | 57.5 | |
| QwenModel Architecture=Qwen, Signal Enhancement (Pose)=false2026.05 | 62.2 | 62.1 | 43.9 | 51.4 | |
| Pose+QwenModel Architecture=Qwen, Signal Enhancement (Pose)=true2026.05 | 53.1 | 43.3 | 38.2 | 40.6 | |
| Pose+InternModel Architecture=Intern, Signal Enhancement (Pose)=true2026.05 | 52.4 | 30.8 | 11.4 | 16.7 |