Transaction Terminal Fraud Processing

Abstract

Image analysis is performed on a user at a transaction terminal. Based on behaviors, expressions, and activities of the user, fraud or potential fraud is flagged. When fraud is flagged, the transaction terminal stops processing an active transaction on behalf of the user and alerts are sent. When potential fraud is flagged, images/video associated with the active transaction are sent for review and the active transaction may be suspended or permitted to proceed at the transaction terminal. In an embodiment, a same user conducting multiple transactions with different accounts at a same transaction terminal or multiple different transaction terminals within a configured period of time is automatically identified as fraud based on a fraud rule.

1 . (canceled) 
     
     
         2 . A method, comprising:
 identifying from a first image an operator at a first terminal performing a first transaction utilizing a first account;   identifying from a second image the operator at a second terminal performing a second transaction utilizing a second account; and   flagging the second transaction as potential fraud based on factors associated with the operator, the first transaction, and the second transaction.   
     
     
         3 . The method of  claim 2  further comprising, preventing the second transaction from completing based on the flagging. 
     
     
         4 . The method of  claim 2 , wherein flagging further includes sending the factors to a remote terminal for determining the flagging. 
     
     
         5 . The method of  claim 2 , wherein flagging further includes providing the factors to the a machine-learning algorithm as input and determining the flagging based on output from the machine-learning algorithm. 
     
     
         6 . The method of  claim 2 , wherein identifying from the first image further includes capturing, by a camera, the first image when a first card is inserted into the first terminal for the first transaction, wherein the first card is associated with the first account. 
     
     
         7 . The method of  claim 6 , wherein identifying from the second image further includes capturing, by another camera, the second image when a second card is inserted into the second terminal for the second transaction, wherein the second card is associated with the second account. 
     
     
         8 . The method of  claim 2 , wherein flagging further includes using first features of the operator from the first image and second features of the operator from the second image to confirm that the operator is a same individual that performed the first transaction at the first terminal and is attempting to perform the second transaction at the second terminal. 
     
     
         9 . The method of  claim 2 , wherein identifying from the first image further includes setting a timer, wherein identifying from the second image further includes comparing an elapsed time of the timer to a threshold, and using the elapsed time compared to the threshold as one of the factors during the flagging. 
     
     
         10 . The method of  claim 9 , wherein identifying from the first image further includes identifying a first behavior of the operator from the first image, wherein identifying from the second image further includes identifying a second behavior of the operator from the second image and using the first behavior and the second behavior as additional ones of the factors during the flagging. 
     
     
         11 . The method of  claim 10 , identifying from the first image further includes identifying first transaction details for the first transaction, wherein identifying from the second image further includes identifying second transaction details for the second transaction and using the first transaction details and the second transaction details a further ones of the factors during the flagging 
     
     
         12 . The method of  claim 11 , wherein flagging further includes generating a fraud score based on the factors and comparing the fraud score against a threshold score causing the second transaction to be halted at the second terminal. 
     
     
         13 . A method, comprising:
 evaluating images taken of an operator that performs a first transaction at a first terminal and is attempting to perform a second transaction at a second terminal within a predefined amount of time;   obtaining first transaction details for the first transaction and second transaction details associated with the second transaction; and   determining whether to prevent the second transaction from being processed on the second terminal based on the evaluating of the images, the first transaction details, and the second transaction details.   
     
     
         14 . The method of  claim 13 , wherein evaluating further includes confirming that the operator at the first terminal is a same operator present at the second terminal from the images. 
     
     
         15 . The method of  claim 14 , wherein evaluating further includes identifying a first behavior of the operator at the first terminal from the images and a second behavior of the operator at the second terminal from the images. 
     
     
         16 . The method of  claim 13 , wherein obtaining further includes determining from the first transaction details that a first account was used by the operator for the first transaction and determining from the second transaction details that a second account is attempting to be used by the operator at the second terminal. 
     
     
         17 . The method of  claim 13 , wherein determining further includes generating a fraud score for the second transaction based on factors derived from the images, the first transaction details, and the second transaction details. 
     
     
         18 . The method of  claim 17 , wherein determining further includes preventing the second transaction from completing on the second terminal when the fraud score exceeds a threshold score. 
     
     
         19 . The method of  claim 13 , wherein determining further includes providing the images, the first transaction details, and the second transaction details as input to a machine-learning algorithm and using an output from the machine-learning algorithm to determine whether to prevent the second transaction from being completed on the second terminal. 
     
     
         20 . A system, comprising:
 a first terminal;   a second terminal;   cameras;   a server that comprises a processor;   the processor executes instructions that cause the processor to perform operations, comprising:
 obtaining, from a first camera, a first image of an operator at the first terminal performing a first transaction using a first account at a first time; 
 obtaining, from a second camera, a second image of the operator at the second terminal attempting to perform a second transaction using a second account at a second time; 
 confirming from the first image and the second image that the operator is a same individual that performed the first transaction and that is attempting to perform the second transaction; 
 confirming that a difference between the first time and the second time is within a threshold period of elapsed time; 
 using first transaction details associated with the first transaction and second transaction details associated with the second transaction and generating first factors; 
 using the first image and the second image and generating second factors; 
 generating a fraud score based on the first factors and the second factors; 
 preventing the second transaction from completing on the second terminal when the fraud score exceeds a threshold score. 
   
     
     
         21 . The system of  claim 20 , wherein the first terminal and the second terminal are automated teller machines, self-service terminals, kiosks, or point-of-sale terminals.