--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].

The traditional retail industry is facing challenges as the rapid development and continuous improvement of AI tools and techniques ushers in the era of New Retail. Many once-successful brick and mortar shops are at risk of disappearing altogether if they fail to adapt to the changing marketplace. As the concept of "New Retail" spreads globally, many Fortune Global 500 retail companies are implementing AI technologies such as computer vision and NLP in unmanned stores and warehouses or virtual stores, and are accelerating supply chain upgrades to better position themselves for the future of shopping. Global Retail Industry Market Size The steady growth of the global economy and per capita disposable income have triggered a surge in the retail industry over the past few years. The global retail industry market was valued at nearly US $23.5 trillion in 2017, a 5.3 percent year-on increase.

The traditional retail industry is facing challenges as the rapid development and continuous improvement of AI tools and techniques ushers in the era of New Retail. Many once-successful brick and mortar shops are at risk of disappearing altogether if they fail to adapt to the changing marketplace. As the concept of "New Retail" spreads globally, many Fortune Global 500 retail companies are implementing AI technologies such as computer vision and NLP in unmanned stores and warehouses or virtual stores, and are accelerating supply chain upgrades to better position themselves for the future of shopping. Global Retail Industry Market Size The steady growth of the global economy and per capita disposable income have triggered a surge in the retail industry over the past few years. The global retail industry market was valued at nearly US $23.5 trillion in 2017, a 5.3 percent year-on increase.

Unmanned store concept adoption has increased in retail, as a growing number of convenience stores are to embrace the concept by the end of the year, Singapore retail companies Octobox, OMO Store and Pick & Go recently announced at the Singapore Retail Industry Conference and Exhibition, writes The Strait Times. Deployments such as palm scanners will allow shoppers to enter the kiosk, RFID will keep track of items purchased and AI will monitor movement inside the kiosk. These features as well as smart cameras and smart shelves will be used in data analysis to improve shoppers' journey and experience. The retail industry in the area is struggling with a lack of employees and severe competition from online retailers, but also suffers from "downside risks and the many uncertainties around the world," explained Trade and Industry Minister Chan Chun Sing at the event. With help and support in developing the concept from government agency Enterprise Singapore, Singapore retail companies hope to overcome the difficult times through innovation.

Customers at this automated JD.com store in Tianjin can pay via QR code on mobile phones and through facial recognition technology, both without sales assistance. First there was pilot-free aircraft in China, better known as drones. This year the country began setting up a tract of the South China Sea to test captain-less ships. And now there are stores without staff. This phenomenon, while not unique to China, has taken off quickly there this year with lots of growing room left.

Out of curiosity, Teresa Chen recently went to an unmanned convenience store, EasyGo, near the China Plaza shopping mall in Guangzhou, China. She used her WeChat app to scan a QR code to get into the store, where she found imported snacks, candy and soft drinks. Every product had a price tag powered by radio-frequency identification, a technology that can automatically track and identify tags attached to objects. After she selected a box of crackers, Chen stood in the payment area and the bill automatically appeared in WeChat. Once Chen paid using her WeChat Wallet (a mobile payment service developed by Tencent), the door opened to let her leave the store. "The whole experience was seamless, as I didn't even need to scan the price tag to pay," said Chen, a product manager for a nutrition supplement company in Guangzhou.