Presents VVA-Bench for vision-centric prompt attacks on I2V generation showing high ASR on SOTA models and VPA-Guard defense reducing ASR by 44.2% on average.
arXiv preprint arXiv:2312.02216 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
Training-free motion conditioning for latent video diffusion by direct injection of low-frequency phase from a reference video into the diffusion noise.
citing papers explorer
-
VPA-Guard: Defending and Benchmarking Image-to-Video Generation Against Visual Prompt Attacks
Presents VVA-Bench for vision-centric prompt attacks on I2V generation showing high ASR on SOTA models and VPA-Guard defense reducing ASR by 44.2% on average.
-
{\Phi}-Noise: Training-Free Temporal Video Conditioning via Phase-Based Noise Manipulation
Training-free motion conditioning for latent video diffusion by direct injection of low-frequency phase from a reference video into the diffusion noise.