Research map / Federated learning and privacy / Privacy attacks and secure learning
Personalized and Invertible Face De-Identification by Disentangled Identity Information Manipulation
Jingyi Cao, Bo Liu, Yunqian Wen, Rong Xie, Li Song
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 |
|---|---|---|---|
| use ours instead of Gu et al. | face de-identification · FaceNet (CASIA) | successful protection rate: 0.704 -> 0.913 | Improved |
| use ours instead of Gu et al. | face de-identification · FaceNet (VGGFace2) | successful protection rate: 0.695 -> 0.928 | Improved |
| use ours instead of Gu et al. | face de-identification · Face Recognition library | identity distance: 0.82234 -> 0.79195 | Hurt |
| use ours instead of DeepPrivacy | face de-identification · FaceNet (VGGFace2) | successful protection rate: 0.816 -> 0.928 | Improved |
| use ours instead of DeepPrivacy | face de-identification · CelebA-HQ face region | pixel distance: 5.005 -> 0.225 | Improved |
| use ours instead of DeepPrivacy | face de-identification · CelebA-HQ | face detectability: 1.0 -> 0.9973 | No clear effect |
| use ours instead of Gu et al. | face de-identification · CelebA-HQ | face detectability: 0.8585 -> 0.9973 | Improved |
| restore with the correct password instead of a wrong one | face recovery · Face Recognition library | successful protection rate: 0.904 -> 0.035 | Improved |
| restore with the correct password instead of a wrong one | face recovery · Face Recognition library | identity distance: 0.794 -> 0.228 | Improved |
| use ours instead of Gu et al. for restoration | face recovery · Restored image quality | LPIPS: 0.186 -> 0.062 | Improved |
| use ours instead of Gu et al. for restoration | face recovery · Restored image quality | SSIM: 0.827 -> 0.902 | Improved |
| use ours instead of Gu et al. for restoration | face recovery · Restored image quality | PSNR: 27.602 -> 27.501 | Hurt |
Closest papers by meaning
- Disguise without Disruption: Utility-Preserving Face De-identification (AAAI 2024 · similarity 0.88)
- Privacy-Preserving Optics for Enhancing Protection in Face De-Identification (CVPR 2024 · similarity 0.87)
- Privacy-Preserving Face Recognition Using Trainable Feature Subtraction (CVPR 2024 · similarity 0.87)
- How Private is Private? A Comparative Study for Face De-Identification (NeurIPS 2026 · similarity 0.86)
- Asymmetric Invertible Threat: Learning Reversible Privacy Defense for Face Recognition (NeurIPS 2026 · similarity 0.85)
- FaceMe: Robust Blind Face Restoration with Personal Identification (AAAI 2025 · similarity 0.84)
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.
