Action Recognition on UCF101 VISPR
70.4Top-1 AccuracyRaw data
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Raw dataEvaluation Model=GPT-4V2026.03 | 70.4 | 64.4 | |
| OursEvaluation Model=GPT-4V2026.03 | 67.1 | 54.9 | |
| Obf-WeakBlurEvaluation Model=GPT-4V2026.03 | 63.1 | 63.5 | |
| Obf-BlackeningEvaluation Model=GPT-4V2026.03 | 62.4 | 56.4 | |
| Obf-StrongBlurEvaluation Model=GPT-4V2026.03 | 61.7 | 55.9 | |
| DownsampleEvaluation Model=GPT-4V2026.03 | 61.5 | 57.2 | |
| SelectivePrivacyEvaluation Model=GPT-4V2026.03 | 61.2 | 59.1 | |
| Surrogate model of STPrivacyEvaluation Model=GPT-4V2026.03 | 59.8 | 57.4 | |
| Surrogate model of MPPAREvaluation Model=GPT-4V2026.03 | 59.2 | 57.8 | |
| Surrogate model of SPActEvaluation Model=GPT-4V2026.03 | 59.1 | 57.9 | |
| Surrogate model of VITAEvaluation Model=GPT-4V2026.03 | 58.9 | 58.5 | |
| Surrogate model of BDQEvaluation Model=GPT-4V2026.03 | 58.7 | 58.1 | |
| Raw dataEvaluation Model=Video-LLaVA2026.03 | 55.4 | — | |
| OursEvaluation Model=Video-LLaVA2026.03 | 48.4 | — | |
| Obf-WeakBlurEvaluation Model=Video-LLaVA2026.03 | 46 | — | |
| Obf-StrongBlurEvaluation Model=Video-LLaVA2026.03 | 43.6 | — | |
| DownsampleEvaluation Model=Video-LLaVA2026.03 | 43.1 | — | |
| Obf-BlackeningEvaluation Model=Video-LLaVA2026.03 | 40.1 | — | |
| SelectivePrivacyEvaluation Model=Video-LLaVA2026.03 | 39.8 | — | |
| Surrogate model of STPrivacyEvaluation Model=Video-LLaVA2026.03 | 37.4 | — | |
| Surrogate model of MPPAREvaluation Model=Video-LLaVA2026.03 | 36.9 | — | |
| Surrogate model of SPActEvaluation Model=Video-LLaVA2026.03 | 36.8 | — | |
| Surrogate model of BDQEvaluation Model=Video-LLaVA2026.03 | 35.6 | — | |
| Surrogate model of VITAEvaluation Model=Video-LLaVA2026.03 | 34.2 | — |