Sparse Adversarial Attack to Object Detection
This work addresses the problem of creating sparse adversarial attacks for object detectors, which is an incremental step in adversarial machine learning research.
This paper introduces Sparse Adversarial Attack (SAA), a method for generating adversarial examples against object detectors using l0 norm perturbations. SAA effectively evades detection on YOLOv4 and FasterRCNN, and demonstrates strong transferability across different detectors in black-box attack scenarios.
Adversarial examples have gained tons of attention in recent years. Many adversarial attacks have been proposed to attack image classifiers, but few work shift attention to object detectors. In this paper, we propose Sparse Adversarial Attack (SAA) which enables adversaries to perform effective evasion attack on detectors with bounded \emph{l$_{0}$} norm perturbation. We select the fragile position of the image and designed evasion loss function for the task. Experiment results on YOLOv4 and FasterRCNN reveal the effectiveness of our method. In addition, our SAA shows great transferability across different detectors in the black-box attack setting. Codes are available at \emph{https://github.com/THUrssq/Tianchi04}.