A Generative Victim Model for Segmentation

TL;DR

This paper introduces a novel generative victim model for segmentation that enables effective adversarial attack generation without relying on specialized segmentation models, shifting focus from robustness to image generation.

Contribution

It proposes a new generative victim model for segmentation, providing a different approach to adversarial attacks that does not depend on task-specific models.

Findings

Effective adversarial attacks with good transferability.

The approach diverges from traditional white-box and black-box attack methods.

The model does not require task-specific segmentation models.

Abstract

We find that the well-trained victim models (VMs), against which the attacks are generated, serve as fundamental prerequisites for adversarial attacks, i.e. a segmentation VM is needed to generate attacks for segmentation. In this context, the victim model is assumed to be robust to achieve effective adversarial perturbation generation. Instead of focusing on improving the robustness of the task-specific victim models, we shift our attention to image generation. From an image generation perspective, we derive a novel VM for segmentation, aiming to generate adversarial perturbations for segmentation tasks without requiring models explicitly designed for image segmentation. Our approach to adversarial attack generation diverges from conventional white-box or black-box attacks, offering a fresh outlook on adversarial attack strategies. Experiments show that our attack method is able to generate effective adversarial attacks with good transferability.