Traffic Sign Classifiers Under Physical World Realistic Sticker Occlusions: A Cross Analysis Study

Abstract

Recent adversarial attacks with real world applications are capable of deceiving deep neural networks (DNN), which often appear as printed stickers applied to objects in physical world. Though achieving high success rate in lab tests and limited field tests, such attacks have not been tested on multiple DNN architectures with a standard setup to unveil the common robustness and weakness points of both the DNNs and the attacks. Furthermore, realistic looking stickers applied by normal people as acts of vandalism are not studied to discover their potential risks as well the risk of optimizing the location of such realistic stickers to achieve the maximum performance drop. In this paper, (a) we study the case of realistic looking sticker application effects on traffic sign detectors performance; (b) we use traffic sign image classification as our use case and train and attack 11 of the modern architectures for our analysis; (c) by considering different factors like brightness, blurriness and contrast of the train images in our sticker application procedure, we show that simple image processing techniques can help realistic looking stickers fit into their background to mimic real world tests; (d) by performing structured synthetic and real-world evaluations, we study the difference of various traffic sign classes in terms of their crucial distinctive features among the tested DNNs.

Cybersecurity of Autonomous Vehicles: A Systematic Literature Review of Adversarial Attacks and Defense Models

“…Intentional threats occur when bad actors maliciously take advantage of the shortcomings and weaknesses of AI techniques with the goal of interfering with the AI system and impairing safety-critical operations. Such attacks include painting the road to confuse drivers or placing stickers on a stop sign to obscure the view [182], [183], [184], [185], [186], and [187]. These modifications may cause the AI system to misclassify objects, which may therefore cause the AV to act in a potentially hazardous manner.…”

Explainable and Trustworthy Traffic Sign Detection for Safe Autonomous Driving: An Inductive Logic Programming Approach

“…One recent work proposed three sticker application methods, namely RSA, SSA and MCSA, that can deceive the traffic sign recognition DNNs with realistic-looking stickers [4]. Another attack included painting the road, which targeted deep neural network models for end-to-end autonomous driving control [5].…”