.png){kind=icon}
.png){kind=icon}
Return
FTUAttack: Feature truncation unrestricted attack based on stable diffusion model
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2025.129604.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Propose a transferable Feature Truncation Unrestricted Attack based on diffusion model. • Propose a Global Deep Feature Extractor module to boost attack transferability. • Propose a Critical Latent Feature Extractor module to further enhance attack effect. • Propose a Multi-Loss Fusion strategy to balance the transferability and imperceptibility.
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W