US 12,493,882 B2
Computer-vision based detection of scams at physical point of sale terminals
Ethan Sommer, Minneapolis, MN (US)
Assigned to Target Brands, Inc., Minneapolis, MN (US)
Filed by TARGET BRANDS, INC., Minneapolis, MN (US)
Filed on Apr. 20, 2023, as Appl. No. 18/303,634.
Claims priority of provisional application 63/390,727, filed on Jul. 20, 2022.
Prior Publication US 2024/0029069 A1, Jan. 25, 2024
Int. Cl. G06Q 20/00 (2012.01); G06Q 20/20 (2012.01); G06Q 20/40 (2012.01); G06V 10/70 (2022.01); G06V 20/52 (2022.01); G06V 40/20 (2022.01); G07G 1/12 (2006.01)
CPC G06Q 20/4016 (2013.01) [G06Q 20/204 (2013.01); G06V 10/70 (2022.01); G06V 20/52 (2022.01); G06V 40/20 (2022.01); G07G 1/12 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system for detecting a scam attempt in a physical retail environment, the system comprising:
at least one camera positioned within a physical retail environment, the at least one camera configured to capture image data of a checkout area in the physical retail environment;
a cash register having a cash register drawer, wherein the cash register generates event data during a checkout process; and
an edge computing device in communication with the at least one camera and the cash register, the edge computing device being configured to:
access, from a data storage device, scam detection criteria that identifies combinations of visual features and event features that correspond to one or more checkout scams;
receive, from the at least one camera, the image data of the checkout area;
receive, from the cash register, the event data that is associated with the checkout process in the checkout area, wherein the event data includes a cash register drawer open event;
detect, based on processing the image data, physical movement of a form of payment in or around the checkout area;
correlate timestamps for the detected physical movement of the form of payment with the event data to generate correlated event data, wherein the correlating is based on the timestamps for the detected physical movement of the form of payment being within a threshold amount of time of the cash register drawer open event;
generate a confidence value based on the correlated event data, wherein the confidence value indicates a likelihood of occurrence of the one or more checkout scams;
determine whether the confidence value satisfies the scam detection criteria;
generate, based on a determination that the confidence value satisfies the scam detection criteria, an indication of the one or more checkout scams; and
return the indication of the one or more checkout scams being performed at the cash register and in the checkout area.