AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

TAFIM: Targeted Adversarial Attacks against Facial Image Manipulations

Shivangi Aneja, Lev Markhasin, Matthias Nießner

ECCV 2022 · Adversarial robustness and security › Attacks on text and graphs · approach: Diffusion-based adversarial purification · 41 citations (OpenAlex)

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 changedWhereMeasuredEffect
replace per-image optimization with a conditional perturbation networkI-PGD · Self-reconstructionrun time per image: 17523.01 ms -> 10.66 msImproved
condition perturbation model on global perturbation in addition to the image (ablation)Ours Model (I) · Self-reconstruction (white target)RMSE/PSNR/LPIPS/VGG vs perturbation levelImproved
train with differentiable JPEG compression (ablation)TAFIM · Self-reconstruction under JPEG C-80protection qualityImproved
train with random compression quality instead of fixed C-80 (ablation)TAFIM · Self-reconstruction under JPEG C-30protection qualityImproved
add attention-based fusion of model-specific perturbations (ablation)Combined model · Multiple manipulations (pSp, SimSwap, StyleClip)protection qualityImproved
use L2 norm instead of L1 or L-inf in the loss (ablation)TAFIM · Self-reconstruction (white target)PSNR vs perturbationImproved
apply two successive JPEG compressions (C-30 then C-80)TAFIM · Self-reconstructionprotection qualityNo clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy