New black-box attack method ColorFD targets remote sensing object detectors

Researchers have developed ColorFD, a novel black-box physical adversarial attack method designed for remote sensing object detectors. This method utilizes multiple pure-color patches, optimizing their positions and colors through Differential Evolution. ColorFD demonstrates effectiveness against detectors like YOLOv3u, YOLOv5u, and Faster R-CNN, outperforming existing black-box patch attacks and remaining competitive with white-box methods. Physical-world experiments confirm the transferability of these optimized patches from digital simulations to real-world imaging conditions. AI

IMPACT This research could lead to more robust object detection systems in remote sensing by highlighting vulnerabilities to physical adversarial attacks.

RANK_REASON The cluster contains a research paper detailing a new method for adversarial attacks on computer vision models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• ColorFD
• DagsHub
• differential evolution
• Faster R-CNN
• Hugging Face
• YOLOv3u
• YOLOv5u

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →