Video Question Answering on MedVideoCap zero-shot
90.6Accuracy (%)EgoBlur
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EgoBlurprotocol=zero-shot, evaluation model=Qwen3-VL-8B, category=Classical2026.06 | 90.6 | 0.13 | |
| Original (upper bound)protocol=zero-shot, evaluation model=Qwen3-VL-8B2026.06 | 90.47 | — | |
| DeepPrivacy2 (face)protocol=zero-shot, evaluation model=Qwen3-VL-8B, category=Face-only, generative2026.06 | 89.8 | -0.67 | |
| FAMSprotocol=zero-shot, evaluation model=Qwen3-VL-8B, category=Face-only, generative2026.06 | 89.6 | -0.87 | |
| FADMprotocol=zero-shot, evaluation model=Qwen3-VL-8B, category=Body-only, generative2026.06 | 86.3 | -4.17 | |
| Gaussian blurprotocol=zero-shot, evaluation model=Qwen3-VL-8B, category=Classical2026.06 | 85.27 | -5.2 | |
| StructHumanprotocol=zero-shot, evaluation model=Qwen3-VL-8B, conditioning=Wan2.1-VACE2026.06 | 85.2 | -5.27 | |
| Pixelationprotocol=zero-shot, evaluation model=Qwen3-VL-8B, category=Classical2026.06 | 85.13 | -5.34 | |
| StructAllprotocol=zero-shot, evaluation model=Qwen3-VL-8B, conditioning=Wan2.1-VACE2026.06 | 84.67 | -5.8 | |
| DeepPrivacy2 (body)protocol=zero-shot, evaluation model=Qwen3-VL-8B, category=Body-only, generative2026.06 | 78.07 | -12.4 | |
| Blackoutprotocol=zero-shot, evaluation model=Qwen3-VL-8B, category=Classical2026.06 | 74.8 | -15.67 |