.png){kind=icon}
Return
BoostCount: Diffusion-based position-sensitive adversarial purification for crowd counting
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2025.129041.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Develop a position-sensitive adversarial purification framework for crowd counting. • Formulate adversarial purification into image-to-image conversion for target task. • Firstly introduce the conditional diffusion into the purification for crowd counting. • Design a position-sensitive adversarial attack module for crowd counting. • Conduct extensive experiments to validate the defense effect of the developed method.
Keywords:
adversarial purification
crowd counting
conditional diffusion
position-sensitive
image-to-image conversion
Journal
IF:
7.5
Papers:
3.0W
Citations:
10.2W
Organization
Cited Papers
No cited papers available