Research map / Adversarial robustness and security / Attacks on text and graphs
Visual-Friendly Concept Protection via Selective Adversarial Perturbations
Xiaoyue Mi, Fan Tang, You Wu, Juan Cao, Peng Li, Yang Liu
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 VCPro instead of Anti-DreamBooth | Stable Diffusion v1-4 personalization protection · VGGFace2 (eps 8/255) | FID (lower = less visible): 96.2 -> 27.0 | Improved |
| use VCPro instead of Anti-DreamBooth (protection strength) | Textual Inversion protection · VGGFace2 (eps 8/255) | LIQE (lower = stronger protection): 1.18 -> 2.31 | Hurt |
| add regional adversarial loss (RAL) to Anti-DreamBooth (ablation) | Anti-DreamBooth · VGGFace2 ablation | FID: 96.2 -> 68.2 | Improved |
| add Lagrangian multiplier loose solution (LMLS) to Anti-DreamBooth (ablation) | Anti-DreamBooth · VGGFace2 ablation | FID: 96.2 -> 30.1 | Improved |
| add LMLS to Anti-DreamBooth (ablation) | Anti-DreamBooth · VGGFace2 ablation, Textual Inversion | CLIP-FACE: 0.05 -> 0.17 | Hurt |
| add pull term to regional adversarial loss (ablation) | VCPro · VGGFace2 ablation, Textual Inversion | CLIP-FACE: 0.14 -> 0.03 | Improved |
| add pull term to regional adversarial loss (ablation) | VCPro · VGGFace2 ablation | FID: 17.2 -> 27.0 | Hurt |
| constrain perturbation with input mask instead of loss-guided mask (ablation) | VCPro · VGGFace2 ablation | FID: 18.6 -> 27.0 | Hurt |
| increase perturbation size from 8/255 to 16/255 (ablation) | VCPro · VGGFace2 | FID: 27.0 -> 48.7 | Hurt |
| increase perturbation size from 8/255 to 16/255 (ablation) | VCPro · VGGFace2 | LIQE: 2.19 -> 1.84 | Improved |
| use smaller mask with regional loss (ablation) | Anti-DreamBooth+RAL · VGGFace2 ablation | FID: 68.2 -> 58.4 | Improved |
Closest papers by meaning
- Perturb a Model, Not an Image: Towards Robust Privacy Protection via Anti-Personalized Diffusion Models (NeurIPS 2025 · similarity 0.90)
- Can Protective Perturbation Safeguard Personal Data from Being Exploited by Stable Diffusion? (CVPR 2024 · similarity 0.89)
- Prompt-Agnostic Adversarial Perturbation for Customized Diffusion Models (NeurIPS 2024 · similarity 0.89)
- Nearly Zero-Cost Protection Against Mimicry by Personalized Diffusion Models (CVPR 2025 · similarity 0.88)
- CAT: Contrastive Adversarial Training for Evaluating the Robustness of Protective Perturbations in Latent Diffusion Models (ICML 2025 · similarity 0.87)
- Harnessing Global-Local Collaborative Adversarial Perturbation for Anti-Customization (CVPR 2025 · similarity 0.87)
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.
