Shadows can be Dangerous: Stealthy and Effective Physical-world Adversarial Attack by Natural Phenomenon
This addresses the need for more covert adversarial attacks in real-world machine learning deployments, particularly for security-critical applications like autonomous driving, though it is incremental by building on existing optical attack methods.
The paper tackles the problem of creating stealthy physical-world adversarial attacks by using natural shadows instead of artificial patterns, achieving success rates of 98.23% and 90.47% on traffic sign recognition datasets and over 95% effectiveness in real-world moving camera scenarios.
Estimating the risk level of adversarial examples is essential for safely deploying machine learning models in the real world. One popular approach for physical-world attacks is to adopt the "sticker-pasting" strategy, which however suffers from some limitations, including difficulties in access to the target or printing by valid colors. A new type of non-invasive attacks emerged recently, which attempt to cast perturbation onto the target by optics based tools, such as laser beam and projector. However, the added optical patterns are artificial but not natural. Thus, they are still conspicuous and attention-grabbed, and can be easily noticed by humans. In this paper, we study a new type of optical adversarial examples, in which the perturbations are generated by a very common natural phenomenon, shadow, to achieve naturalistic and stealthy physical-world adversarial attack under the black-box setting. We extensively evaluate the effectiveness of this new attack on both simulated and real-world environments. Experimental results on traffic sign recognition demonstrate that our algorithm can generate adversarial examples effectively, reaching 98.23% and 90.47% success rates on LISA and GTSRB test sets respectively, while continuously misleading a moving camera over 95% of the time in real-world scenarios. We also offer discussions about the limitations and the defense mechanism of this attack.