Research map / Label-efficient and robust learning / OOD and anomaly detection
Secure On-Device Video OOD Detection Without Backpropagation
Shawn Li, Peilin Cai, Yuxiao Zhou, Zhiyu Ni, Renjie Liang, You Qin, Yi Nian, Zhengzhong Tu, Xiyang Hu, Yue Zhao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| mask the least important 50% of feature channels (dynamic Shapley-based selection) before encryption (ablation) | SecDOOD · Near-OOD UCF101 | FPR95: 11.55 -> 7.48 | Improved |
| mask 50% of channels (dynamic) instead of 75% (ablation) | SecDOOD · Far-OOD HMDB as ID, Kinetics OOD | FPR95: 43.79 -> 15.39 | Improved |
| select masked channels by importance (dynamic) instead of randomly (ablation) | SecDOOD · Far-OOD Kinetics ID, HAC OOD | AUROC: 75.38 -> 84.98 | Improved |
| select masked channels by importance instead of randomly (ablation) | SecDOOD · Near-OOD UCF101 | FPR95: 12.62 -> 7.48 | Improved |
| mask 50% of channels vs no masking (ablation) | SecDOOD · Near-OOD HMDB/EPIC/Kinetics | FPR95: 37.69 -> 37.03 (HMDB) | No clear effect |
| use HyperNetwork-generated classifier parameters trained in the cloud instead of randomly initialized classifier | SecDOOD · Far-OOD Kinetics-600 as ID (avg) | AUROC: 62.55 -> 82.15 | Improved |
| use SecDOOD without on-device training instead of on-device-trained post-hoc methods | Near-OOD detection · UCF101 50/51 | FPR95: 12.52 -> 7.48 | Improved |
| SecDOOD instead of best on-device-trained method on HMDB-as-ID far-OOD average | Far-OOD detection · HMDB as ID (avg) | FPR95: 16.73 -> 22.23 | Hurt |
| offload training to the cloud, no on-device training | SecDOOD · device FLOPs | FLOPs: 3x lower than full on-device training; 16% above baseline inference | Improved |
| encrypt only 25% or 50% of channels instead of 100% (ablation) | SecDOOD · encryption time per sample | encryption/decryption time | Improved |
Closest papers by meaning
- Spurious-Aware Prototype Refinement for Reliable Out-of-Distribution Detection (NeurIPS 2025 · similarity 0.84)
- Image Background Serves as Good Proxy for Out-of-distribution Data (ICLR 2024 · similarity 0.83)
- MultiOOD: Scaling Out-of-Distribution Detection for Multiple Modalities (NeurIPS 2024 · similarity 0.83)
- MetaOOD: Automatic Selection of OOD Detection Models (ICLR 2025 · similarity 0.83)
- Can We Ignore Labels in Out of Distribution Detection? (ICLR 2025 · similarity 0.82)
- OODD: Test-time Out-of-Distribution Detection with Dynamic Dictionary (CVPR 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
