Research map / Adversarial robustness and security / Attacks on text and graphs
ID-Guard: A Universal Framework for Combating Facial Manipulation via Breaking Identification
Zuomin Qu, Wei Lu, Xiangyang Luo, Qian Wang, Xiaochun Cao
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 mask-constrained loss to IDM (ablation) | ID-Guard · CelebAMask-HQ | DSR: 0.866 -> 0.999 | Improved |
| add identity consistency loss to IDM (ablation) | ID-Guard · CelebAMask-HQ | ID sim.: 0.148 -> 0.031 | Improved |
| add feature confusion loss to IDM (ablation) | ID-Guard · CelebAMask-HQ | DSR: 0.973 -> 0.999 | Improved |
| add all three IDM losses (ablation) | ID-Guard · CelebAMask-HQ | DSR: 0.358 -> 0.999 | Improved |
| add identity consistency loss to IDM (ablation) | ID-Guard · CelebAMask-HQ | L2face: 0.431 -> 0.425 | Hurt |
| replace equivalent weights with dynamic MGDA weighting (ablation) | ID-Guard · 5 manipulation models | average DSR: 0.766 -> 0.999 | Improved |
| replace equivalent weights with KPI-based weighting (ablation) | ID-Guard · HiSD | DSR: 0.420 -> 0.998 | Improved |
| use KPI-based instead of MGDA-based weighting for training cost (ablation) | ID-Guard · training time | s/iter: 1.50 -> 1.24 | Improved |
| add gradient prior perturbation over random noise (ablation) | ID-Guard · defense performance | average improvement: 31.2% | Improved |
| use Resnet generator instead of Transformer (ablation) | ID-Guard · generator size | params: 179.3M -> 7.9M | Improved |
| add compression-resistant strategy (ablation) | ID-Guard · training cost | GFLOPs/sample: 1670.22 -> 1751.39 | Hurt |
| adversarial training with ID-Guard on StarGAN | StarGAN · PGD attack | ID sim.: 0.004 -> 0.712 | Improved |
Closest papers by meaning
- Adversarial Robust Safeguard for Evading Deep Facial Manipulation (AAAI 2024 · similarity 0.89)
- Guard Me If You Know Me: Protecting Specific Face-Identity from Deepfakes (NeurIPS 2025 · similarity 0.87)
- IDProtector: An Adversarial Noise Encoder to Protect Against ID-Preserving Image Generation (CVPR 2025 · similarity 0.86)
- Edit Away and My Face Will not Stay: Personal Biometric Defense against Malicious Generative Editing (CVPR 2025 · similarity 0.86)
- FSFM: A Generalizable Face Security Foundation Model via Self-Supervised Facial Representation Learning (CVPR 2025 · similarity 0.85)
- DiffAM: Diffusion-based Adversarial Makeup Transfer for Facial Privacy Protection (CVPR 2024 · similarity 0.85)
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.
