Research map / Video / Action recognition
Temporal Unlearnable Examples: Preventing Personal Video Data from Unauthorized Exploitation by Object Tracking
Qiangqiang Wu, Yi Yu, Chenqi Kong, Ziquan Liu, Jia Wan, Haoliang Li, Alex C. Kot, Antoni B. Chan
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 |
|---|---|---|---|
| add temporal contrastive learning (TCL) (ablation) | TUE generator, SiamFC · GOT-10k AO | AO: 16.1 -> 12.1 | Improved |
| add temporal contrastive learning (TCL) (ablation) | TUE generator, OSTrack · GOT-10k AO | AO: 41.6 -> 18.0 | Improved |
| add temporal contrastive learning (TCL) (ablation) | TUE generator, SiamFC · OTB AUC | AUC: 17.6 -> 11.4 | Improved |
| add target state condition to generator (ablation) | TUE generator, SiamFC · GOT-10k AO | AO: 22.5 -> 16.1 | Improved |
| add target state condition to generator (ablation) | TUE generator, SiamFC · OTB AUC | AUC: 19.7 -> 17.6 | Improved |
| add context-region noise to EM baseline (ablation) | EM baseline, SiamFC · GOT-10k AO | AO: 27.0 -> 21.4 | Improved |
| replace EM+Context with learned generator (ablation) | SiamFC · GOT-10k AO | AO: 21.4 -> 16.1 | Improved |
| train generator with only 25% of GOT-10k videos (ablation) | TUE generator · GOT-10k | AO | No clear effect |
| generator versus EM+Context optimisation for training cost (ablation) | TUE generator · training cost | training time: 33 Hours -> 7 Hours | Improved |
| apply TUE instead of EM when transferring to SeqTrack-256 | SeqTrack-256 · GOT-10k AO | AO: 8.1 -> 2.1 | Improved |
| apply TUE instead of EM when transferring to HIPTrack | HIPTrack · GOT-10k AO | AO: 63.0 -> 43.2 | Improved |
| apply TUE instead of EM on OSTrack | OSTrack-256 · GOT-10k AO | AO: 26.3 -> 18.0 | Improved |
Closest papers by meaning
- Few-Shot Backdoor Attacks on Visual Object Tracking (ICLR 2022 · similarity 0.86)
- Learning To Track Objects From Unlabeled Videos (ICCV 2021 · similarity 0.86)
- Towards Privacy-Preserving Visual Recognition via Adversarial Training: A Pilot Study (ECCV 2018 · similarity 0.85)
- Privacy-Preserving Deep Action Recognition: An Adversarial Learning Framework and A New Dataset (TPAMI 2020 · similarity 0.85)
- Privacy Beyond Pixels: Latent Anonymization for Privacy-Preserving Video Understanding (ICLR 2026 · similarity 0.85)
- Video Diffusion Models Excel at Tracking Similar-Looking Objects Without Supervision (NeurIPS 2025 · similarity 0.83)
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.
