Robustness Analysis against Adversarial Patch Attacks in Fully Unmanned Stores

--The advent of convenient and efficient fully unmanned stores equipped with artificial intelligence-based automated checkout systems marks a new era in retail. However, these systems have inherent artificial intelligence security vulnerabilities, which are exploited via adversarial patch attacks, particularly in physical environments. This study demonstrated that adversarial patches can severely disrupt object detection models used in unmanned stores, leading to issues such as theft, inventory discrepancies, and interference. We investigated three types of adversarial patch attacks--Hiding, Creating, and Altering attacks--and highlighted their effectiveness. We also introduce the novel color histogram similarity loss function by leveraging attacker knowledge of the color information of a target class object. Besides the traditional confusion-matrix-based attack success rate, we introduce a new bounding-boxes-based metric to analyze the practical impact of these attacks. Starting with attacks on object detection models trained on snack and fruit datasets in a digital environment, we evaluated the effectiveness of adversarial patches in a physical testbed that mimicked a real unmanned store with RGB cameras and realistic conditions. Furthermore, we assessed the robustness of these attacks in black-box scenarios, demonstrating that shadow attacks can enhance success rates of attacks even without direct access to model parameters. Our study underscores the necessity for robust defense strategies to protect unmanned stores from adversarial threats. Highlighting the limitations of the current defense mechanisms in real-time detection systems and discussing various proactive measures, we provide insights into improving the robustness of object detection models and fortifying unmanned retail environments against these attacks. NMANNED stores operate without sellers or employees, utilizing artificial intelligence-(AI-)based automation technologies. These stores provide consumers with a convenient shopping experience and effectively reduce operational costs, thus playing a crucial role in shaping the future of automated retail[1], [2]. This work was partially supported by the Institute of Information and Communications Technology Planning and Evaluation (IITP) grant through the Korean Government [Ministry of Science and ICT (MSIT)] (Robust AI and Distributed Attack Detection for Edge AI Security) under Grant 2021-0-00511 and by the IITP grant funded by the Korea government (MSIT) (No. H. Na, W . Lee, S. Roh and S. Park are with Department of Software Convergence, Soongsil University, Seoul 07027, South Korea, Email: rn-rud7932@soongsil.ac.kr. D. Choi is with Department of Software, Soongsil University, Seoul 07027, South Korea, Email: sunchoi@ssu.ac.kr. Enhanced detection capabilities are achieved using RGB and depth cameras, supplemented by sensors like RFID, LiDAR, and weight sensors [3], [4].