| CPC G06Q 20/1085 (2013.01) [G06F 21/83 (2013.01); G06V 10/761 (2022.01); G06V 20/52 (2022.01); G06V 40/10 (2022.01); G06V 40/28 (2022.01); G07F 19/2055 (2013.01); G08B 13/19613 (2013.01)] | 20 Claims |

|
1. A method comprising:
receiving a plurality of images from at least one camera;
initiating, by a behavior analysis machine-learning platform trained using images of normal and abnormal behavior at kiosks, an abnormal behavior determination process including:
identifying, by at least one processor of the behavior analysis machine-learning platform, at least one image of the plurality of images, which includes one or more hands of a person that extend towards a keypad or a card reader on an electronic kiosk to perform an action;
determining, by the at least one processor of the behavior analysis machine-learning platform, that the one or more hands are engaged in abnormal behavior associated with an unauthorized device being installed at the electronic kiosk,
wherein the abnormal behavior is determined based on a comparison of at least one area occupied by the one or more hands of the person, that extend towards the keypad or the card reader on the electronic kiosk, in the identified at least one image to a previously identified abnormal area for at least one hand in an image illustrating abnormal behavior; and
in response to determining the abnormal behavior, preventing at least one transaction from occurring at the electronic kiosk.
|