US 12,468,990 B2
Auditing purchasing system
Susan W. Brosnan, Raleigh, NC (US); Jessica Snead, Cary, NC (US); Patricia S. Hogan, Raleigh, NC (US); and Daniel R. Goins, Wake Forest, NC (US)
Assigned to Toshiba Global Commerce Solutions Holdings Corporation, Tokyo (JP)
Filed by Toshiba Global Commerce Solutions Holdings Corporation, Tokyo (JP)
Filed on Mar. 18, 2022, as Appl. No. 17/698,922.
Prior Publication US 2023/0297905 A1, Sep. 21, 2023
Int. Cl. G06Q 10/0631 (2023.01); G06Q 20/20 (2012.01); G06V 10/764 (2022.01)
CPC G06Q 10/06311 (2013.01) [G06Q 20/208 (2013.01); G06V 10/764 (2022.01)] 15 Claims
OG exemplary drawing
 
1. A method comprising:
receiving an item for purchase;
determining a first identity of the item for purchase;
adding the item for purchase to a transaction;
receiving, from a first camera located on a self-service kiosk, a first image of the item for purchase captured from the first camera;
determining a second identity of the item for purchase from the first image using a trained machine learning (ML) model; and
responsive to determining the first and second identities of the item for purchase do not match, executing an audit of the transaction at an audit station, wherein executing the audit of the transaction comprises:
generating a set of instructions on the audit station, wherein the set of instructions instruct a user to position the item for purchase in a first position relative to a field of view of a second camera located on the audit station, wherein the second camera is a higher resolution than the first camera;
causing the generated set of instructions to display on a user interface in the field of view of the user;
receiving, from the first camera, a second image of the item for purchase;
determining a third identity of the item for purchase from the second image using a second trained ML model; and
determining that the first identity matches the third identity.