Research map / Video / Action recognition
Anonymizing Egocentric Videos
Daksh Thapar, Aditya Nigam, Chetan Arora
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 proposed homography perturbation to egocentric videos | Th epic [42] wearer recognition · EPIC-Kitchens 'Mix' | CRR: 66.3 -> 19.2 | Improved |
| add proposed perturbation to egocentric videos | Furnari [10] activity recognition · EPIC-Kitchens 'Mix' | accuracy: 89.6 -> 81.8 | Hurt |
| fine-tune activity recognizer on perturbed videos | Furnari [10] activity recognition · EPIC-Kitchens 'Mix' | accuracy: 81.8 -> 87.4 | Improved |
| fine-tune wearer recognizer on perturbed videos | Th epic [42] wearer recognition · EPIC-Kitchens 'Cut' | CRR: 15.4 -> 17.8 | No clear effect |
| add proposed perturbation (open set) | Th fpsi [41] wearer recognition · FPSI Subject-1 | CRR: 85.5 -> 13.6 | Improved |
| add proposed perturbation (closed set) | Hoshen [15] wearer recognition · FPSI (closed set) | CRR: 76.0 -> 08.7 | Improved |
| fine-tune Hoshen on perturbed videos | Hoshen [15] wearer recognition · FPSI (closed set) | CRR: 08.7 -> 10.4 | No clear effect |
| add proposed perturbation (closed set) | Poleg [31] activity recognition · FPSI (closed-set protocol) | accuracy: 89 -> 82 | Hurt |
| replace optical-flow gait features with camera rotation (EgoIDNet) for open-set recognition | wearer recognition · EPIC-Kitchens (open set) | CRR: 55.06 -> 59.67 | Improved |
| replace [41] with EgoIDNet in closed set | wearer recognition · DB-01 (closed set) | CA: 99.2 -> 99.1 | No clear effect |
| replace [41] with EgoIDNet in open set | wearer recognition · DB-03 (open set) | CRR: 80.56 -> 83.46 | Improved |
| use random rotation noise instead of crafted perturbation (ablation) | anonymization of EPIC-Kitchens 'cut' videos · EPIC-Kitchens 'cut' | wearer vs activity accuracy | Hurt |
Closest papers by meaning
- Is Sharing of Egocentric Video Giving Away Your Biometric Signature? (ECCV 2020 · similarity 0.94)
- An Egocentric Look at Video Photographer Identity (CVPR 2016 · similarity 0.89)
- EgoPrivacy: What Your First-Person Camera Says About You? (ICML 2025 · similarity 0.86)
- Ego-Only: Egocentric Action Detection without Exocentric Transferring (ICCV 2023 · similarity 0.86)
- EgoSim: An Egocentric Multi-view Simulator and Real Dataset for Body-worn Cameras during Motion and Activity (NeurIPS 2024 · similarity 0.85)
- Estimating Ego-Body Pose from Doubly Sparse Egocentric Video Data (NeurIPS 2024 · 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.
