Abstract: An automated checkout system aims to detect and disable certain functionalities of an auto-checkout client application if a user exhibits anomalous behavior. The system collects user interaction data to describe the user's interactions with the client device through the application. It then uses a behavior scoring model on the data to generate an anomalous behavior score, indicating the likelihood of anomalous behavior. For instance, if a user scans an item far from its usual location, the score may suggest atypical behavior. The system employs the behavior score to identify functionalities to disable by comparing it to a set of threshold values. Each threshold corresponds to specific auto-checkout application functionalities to be disabled if the behavior score reaches the thresholds. For example, if a threshold is related to identifying items by image capture, the system may disable this feature if the score exceeds the threshold value.

Abstract: A device may capture an image of a first machine-readable code, send a request to a server to obtain access to a user application at the device. The device may capture an image of an item offered by a retailer, detect information about the item included in the image of the item, and add the item to a virtual shopping cart. The device may receive payment information for purchasing the item in the virtual shopping cart, and generate a second machine-readable code comprising purchase information of the item. The device may present the second machine-readable code to a verification device for verifying the purchase of the item.