SEAL changes 5% of face pixels to block diffusion identity cloning
On CelebA-HQ and VGGFace2, protected images stay above 41 dB PSNR while identity similarity falls to at most 0.15.
Stories per day · 30 days
Relateddiffusion-models1encrypted-communication1face-privacy1identity-protection1info-nce1leakage-protection1semantic-communication1vae1
6 stories, newest first · one card per story
Stories tagged #adversarial-perturbation
Key-conditioned perturbation reduces eavesdropper leakage in semantic communication
InfoNCE-guided training suppresses eavesdropper BLEU score to 0.07 while LAPI adds 5-9% leakage reduction.
LETHE: Defending against Unauthorized Image-to-Video Generation via Value Diversity Collapse
LETHE protects photos from unauthorized image-to-video generation by collapsing reference value diversity.
Mind the Spike: Mechanisms and Brittleness of Visual Massive Activations in Large Vision-Language Models
Visual activation spikes in vision-language models are brittle and can be created or removed with tiny image perturbations.
When Agents See Differently: Exposing UI Desynchronization Threats in Mobile Agents
Researchers expose 'human-agent UI desynchronization' attacks where repackaged APKs invisibly mislead mobile AI agents into attacker-chosen actions.
RAPID: A Real-Time Defense Against Unauthorized Model Distillation for Text-to-Image Services
RAPID embeds defensive perturbations in a T2I model's shared VAE decoder to block unauthorized black-box distillation in real time.