Defense against Adversarial Cloud Attack on Remote Sensing Salient Object Detection

TL;DR

This paper introduces a novel adversarial cloud attack and a DefenseNet pre-processing method to protect remote sensing salient object detection models from adversarial perturbations, especially cloud-like camouflage, ensuring robustness in both white-box and black-box scenarios.

Contribution

It proposes a new adversarial cloud attack method and a learnable DefenseNet to defend against such attacks without modifying existing deep learning models.

Findings

DefenseNet effectively defends against adversarial cloud attacks.

The method maintains SOD performance under attack in both white-box and black-box settings.

Experimental results show promising robustness on the EORSSD dataset.

Abstract

Detecting the salient objects in a remote sensing image has wide applications for the interdisciplinary research. Many existing deep learning methods have been proposed for Salient Object Detection (SOD) in remote sensing images and get remarkable results. However, the recent adversarial attack examples, generated by changing a few pixel values on the original remote sensing image, could result in a collapse for the well-trained deep learning based SOD model. Different with existing methods adding perturbation to original images, we propose to jointly tune adversarial exposure and additive perturbation for attack and constrain image close to cloudy image as Adversarial Cloud. Cloud is natural and common in remote sensing images, however, camouflaging cloud based adversarial attack and defense for remote sensing images are not well studied before. Furthermore, we design DefenseNet as a learn-able pre-processing to the adversarial cloudy images so as to preserve the performance of the deep learning based remote sensing SOD model, without tuning the already deployed deep SOD model. By considering both regular and generalized adversarial examples, the proposed DefenseNet can defend the proposed Adversarial Cloud in white-box setting and other attack methods in black-box setting. Experimental results on a synthesized benchmark from the public remote sensing SOD dataset (EORSSD) show the promising defense against adversarial cloud attacks.