AI Paper Atlas

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

Universal Image Immunization against Diffusion-based Image Editing via Semantic Injection

Chanhui Lee, Donggyu Choi, Seunghyun Shin, Hae-Gon Jeon, Jeany Son

ECCV 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: Vision-language model robustness · 6 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
add real training data to target injection loss (ablation)Universal immunization UAP · Generated edit set, SD1.5 white-boxLPIPS: 0.557 -> 0.585Improved
add source semantic suppression loss (ablation)Universal immunization UAP · Generated edit set, SD1.5 white-boxLPIPS: 0.585 -> 0.606Improved
add source semantic suppression loss (ablation)Universal immunization UAP · Generated edit set, SD1.5 white-boxPSNR: 14.41 -> 14.19Improved
add source semantic suppression loss (ablation)Universal immunization UAP · InstructPix2Pix black-boxLPIPS: 0.514 -> 0.527Improved
add real data to injection loss (ablation)Universal immunization UAP · InstructPix2Pix black-boxLPIPS: 0.464 -> 0.514Improved
use semantic injection UAP instead of universalized Map baselineUniversal immunization · Generated edit set, SD1.5 white-boxLPIPS: 0.465 -> 0.606Improved
use semantic injection UAP instead of universalized Encoder baselineUniversal immunization · Generated edit set, SD1.4 black-boxPSNR: 18.08 -> 14.23Improved
universal UAP versus image-specific SAOurs (universal) · Generated edit set, SD1.5 white-boxLPIPS: 0.490 -> 0.606Improved
universal UAP versus image-specific EA on Feat. Sim.Ours (universal) · Generated edit set, SD1.5 white-boxFeat. Sim. (CLIP): 0.677 -> 0.673Hurt
universal UAP versus FastProtect optimization-freeOurs (universal) · ImageNet-Edit, SD1.5 white-boxLPIPS: 0.439 -> 0.609Improved
apply JPEG/diffusion purification to immunized imagesOurs (universal) · Conditional DiffPureLPIPS: 0.606 -> 0.568Mixed
vary the target concept (Ronaldo, Tiger, Sunflower, Peacock, Mandala) (ablation)Universal immunization · Generated edit set, SD1.5 white-boxLPIPS: 0.602 avg, 0.015 stdNo 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