Abstract: There are provided systems and methods relating to checking and confirming merchandise purchased at shopping facilities. In one form, the system includes: a shopping cart containing merchandise items to be purchased; a sales transaction database; a point-of-sales system that creates transaction records identified by transaction identifiers; and a mobile device used by an employee that includes a sensor to scan a paper or digital receipt to obtain the transaction identifier and a camera to capture images of the items in the shopping cart. The system also includes a control circuit that receives the transaction identifier, accesses the database using the identifier to determine the purchased items; analyzes the images of the merchandise items in the shopping cart and creates a computer vision receipt listing detected items; compares the purchased items with the detected items; and takes an action if there is a discrepancy.

Abstract: An improved approach for returning aesthetically relevant search results is disclosed. A training set of images (e.g., designer-created images) is used to train a detection engine that detects items in the images as features. A classification engine is configured to analyze the features and generate classification indices for the features. A user can select an item, and the classification index for the feature corresponding to the item is retrieved. The classification index is used to identify result images, which can be returned ranked according user action data and other parameters, such as style.

Abstract: A method of processing a search query includes, for each search context of a plurality of search contexts, for each scoring primitive of a plurality of scoring primitives, and for a set of previously executed search queries that are consistent with the search context, determining a correlation between the scoring primitive and actual user selections of results of the previously executed search queries by a plurality of users. For each search context, machine learning is performed on the correlations to identify a predicted performance function comprising a weighted subset of the scoring primitives that meet predefined predictive quality criteria. Executing a user submitted search query includes associating the user submitted search query with a respective search context, and ordering at least a portion of the search results in accordance with the predicted performance function for the search context for the user submitted search query.