Alert button

Unrestricted Adversarial Attacks for Semantic Segmentation

Add code
Bookmark button
Alert button
Oct 09, 2019
Guangyu Shen, Chengzhi Mao, Junfeng Yang, Baishakhi Ray

Figure 1 for Unrestricted Adversarial Attacks for Semantic Segmentation
Figure 2 for Unrestricted Adversarial Attacks for Semantic Segmentation
Figure 3 for Unrestricted Adversarial Attacks for Semantic Segmentation
Figure 4 for Unrestricted Adversarial Attacks for Semantic Segmentation

Share this with someone who'll enjoy it:

Twitter IconFacebook IconLinkedin IconWhatsapp IconMessenger Icon

Semantic segmentation is one of the most impactful applications of machine learning; however, their robustness under adversarial attack is not well studied. In this paper, we focus on generating unrestricted adversarial examples for semantic segmentation models. We demonstrate a simple yet effective method to generate unrestricted adversarial examples using conditional generative adversarial networks (CGAN) without any hand-crafted metric. The na\"ive implementation of CGAN, however, yields inferior image quality and low attack success rate. Instead, we leverage the SPADE (Spatially-adaptive denormalization) structure with an additional loss item, which is able to generate effective adversarial attacks in a single step. We validate our approach on the well studied Cityscapes and ADE20K datasets, and demonstrate that our synthetic adversarial examples are not only realistic, but also improve the attack success rate by up to 41.0\% compared with the state of the art adversarial attack methods including PGD attack.

View paper onarxiv iconopen_review iconOpenReview

Share this with someone who'll enjoy it:

Twitter IconFacebook IconLinkedin IconWhatsapp IconMessenger Icon