System and method for detecting user fraud in automated teller machine transactions

• US 5,386,104 A
• Filed: 12/27/1993
• Issued: 01/31/1995
• Est. Priority Date: 11/08/1993
• Status: Expired due to Term

• Current Assignee: NCR Corporation
• Original Assignee: NCR Corporation
• Inventors: Sime, Iain R. F.
• Primary Examiner(s): Hajec, Donald
• Assistant Examiner(s): CHIN, ESTHER
• Application Number:
  US08/173,035
• Time in Patent Office: 400 Days
• Field of Search: 235/379, 235/380, 235/382, 235/382.5, 235/438, 235/457, 902/3, 902/5, 902/31, 902/32, 902/10, 382/3, 382/4, 382/13, 340/825.3, 340/825.31, 340/825.34, 340/825.35, 364/408, 359/2
• US Class Current: 235/379
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06Q 20/4014 Identity check for transactions
• CPC Class Codes: G06Q 20/4016 involving fraud or risk level assessment in transaction processing
• CPC Class Codes: G07C 9/257 electronically G07C9/26 takes precedence
• CPC Class Codes: G07F 19/20 Automatic teller machines [ATMs]
• CPC Class Codes: G07F 19/207 Surveillance aspects at ATMs

Abstract

A self-service automated teller system includes a fraud detection module which uses a biometric recognition technique, such as voice or fingerprint recognition, in combination with predictive technology. The biometric recognition test is normally decisive, but if this test produces a result which is close to a predetermined threshold, the predictive technology checks to determine whether the transaction is consistent with the user'"'"'s normal behavior before permitting the transaction to proceed. A suspicion count may also be included in the decision-making process, to determine whether or not the user has a previous record of suspicious transactions.

Claims

1. A system for detecting user fraud in connection with automated teller machine transactions, comprising:

means for measuring at least one biometric value for a user;

means for comparing that value to a reference biometric value for that user;

means for predicting the type of transaction to be requested by said user;

means for comparing the predicted transaction with the transaction actually requested by said user; and

means for determining whether the user is valid, based upon the biometric value comparison and the result of comparing the predicted transaction with the actual transaction if said biometric value comparison is not conclusive.

2. The system of claim 1, also including means for maintaining a suspicion count of suspicious transactions involving said user, threshold means for establishing a maximum threshold number of suspicious transactions, and comparison means for comparing the number of suspicious transactions with the threshold number and conveying the result of the comparison to said means for determining whether the user is valid.

3. The system of claim 1, in which said means for determining whether the user is valid includes means for comparing the difference between the biometric value of the user and the reference biometric value with a threshold biometric value.

4. The system of claim 2, in which said means for determining whether the user is valid comprises a user transaction authorization module having inputs which include the following:

biometric user value;

biometric reference value;

suspicion count;

type of predicted transaction; and

type of actual transaction.

5. The system of claim 4, in which said user transaction authorization module can, depending upon the various inputs, either continue with a user transaction, terminate a user transaction, or obtain further information concerning the user before making a final determination.

6. The system of claim 5, in which said user transaction authorization module can increment the suspicion count when user validity is questionable.

7. The system of claim 1, in which the means for measuring at least one biometric value comprises signature verification measuring means.

8. The system of claim 1, in which the means for measuring at least one biometric value comprises hand geometry measuring means.

9. The system of claim 1, in which the means for measuring at least one biometric value comprises finger print comparison.

10. The system of claim 1, in which the means for measuring at least one biometric value comprises palm print comparison.

11. The system of claim 1, in which the means for measuring at least one biometric value comprises voiceprint measurement.

12. The system of claim 1, in which the means for measuring at least one biometric value comprises retinal eye scanning.

13. The system of claim 1, in which the means for measuring at least one biometric value comprises iris eye scanning.

14. The system of claim 1, in which the means for measuring at least one biometric value comprises vein pattern determination.

15. The system of claim 1, in which said system includes a host computer and a plurality of automated teller machines.

16. The system of claim 15, in which said host computer includes a user reference file which maintains predicted transaction and biometric value reference files for each user.

17. The system of claim 16, in which said host computer also includes a suspicion count which maintains a record of the number of suspicious transactions for each user.

18. A self-service system comprising:

a user interface unit having user identification means;

input means operable by a user for initiating a selected self-service transaction;

biometric means for producing biometric output data relating to a user;

storage means for storing biometric reference data relating to a plurality of authorized users of the system, said storage means being arranged to store data relating to previous transactions initiated by said authorized users;

prediction means coupled to said identification means and arranged to predict the type of transaction to be requested by a given user on the basis of the stored data relating to previous transactions; and

transaction authorization means coupled to said biometric means and to said prediction means and arranged to make a determination as to whether a transaction initiated by a user is to be permitted to proceed to completion, said determination being made on the basis of a comparison between said biometric output data and said biometric reference data for the relevant user, and, if said biometric comparison is not conclusive on the basis of a comparison between the transaction predicted for, and the transaction actually requested by, the relevant user.

19. The system of claim 18, in which said transaction authorization means is arranged to allow a transaction to proceed if said biometric output data conclusively matches said biometric reference data, regardless of the comparison between the predicted and requested transaction, but that if said biometric output data does not conclusively match said biometric reference data but does match said reference data within a predetermined limit of discrepancy then the comparison between the predicted and requested transaction is used in determining whether the transaction is permitted to proceed to completion.

20. The system of claim 19, in which a first value is derived representative of the difference between said biometric output data and said biometric reference data, said output data being considered to be a conclusive match with said reference data if said first value is less than a predetermined threshold value.

21. The system of claim 19, in which, if said biometric output data does not conclusively match said biometric reference data but does match said reference data within said predetermined limit, and said requested transaction is not consistent with said predicted transaction, said transaction authorization means is arranged to obtain additional information concerning the relevant user before a final determination is made as to whether the transaction is permitted to proceed to completion.

22. The system of claim 21, in which, in the course of transaction, a further use is made of said biometric means provide said additional information.

23. The system of claim 21, in which, in the course of a transaction, the user enters a personal identification number by means of said input means to provide additional information.

24. The system of claim 19, also including further storage means arranged to store a suspicion count for each of said authorized users, each suspicion count being arranged to be incremented by one each time said biometric output data for the relevant user is not a conclusive match with said biometric reference data for that user, but being reset to a start value each time said output data is a conclusive match with said reference data, said transaction authorization means being arranged to terminate a transaction if said suspicion count reaches a predetermined threshold value.

25. The system of claim 18, in which said user identification means comprises a card reader for reading from a card account information relating to a user.

26. The system of claim 18, in which the stored data relating to previous transactions is updated each time a transaction is initiated by one of said authorized users.

27. The system of claim 18, also including a host computer and a plurality of automated teller machines, each including a user interface unit, prediction means and authorization means as called for in claim 18.

28. A method for detecting user fraud in connection with automated teller machine transactions, comprising the following steps:

(a) measuring at least one biometric value for a user;

(b) comparing that value to a reference biometric value for that user;

(c) predicting the type of transaction to be requested by said user;

(d) comparing the predicted transaction with the transaction actually requested by said user; and

(e) determining whether the user is valid, based upon the biometric value comparison and the result of comparing the predicted transaction with the actual transaction if said biometric value comparison is not conclusive.

29. The method of claim 28, also including the following additional steps:

(f) maintaining a suspicion count of suspicious transactions involving said user;

(g) establishing a maximum threshold number of suspicious transactions;

(h) comparing the number of suspicious transactions with the threshold number; and

(i) utilizing said comparison in the determination of whether the user is valid.

30. The method of claim 28, in which the step of determining whether the user is valid includes comparing the difference between the biometric value of the user and the reference biometric value with a threshold biometric value.

31. The method of claim 30, also including the step obtaining more information concerning the user when the difference between the biometric value of the user and the reference biometric value is close to the threshold biometric value and the transaction requested by the user is different than the predicted transaction.

32. The method of claim 31, in which the step of obtaining additional information includes obtaining a PIN number from the user.

33. The method of claim 29, also including the step of incrementing the suspicion count whenever it is determined in step (e) that the user is not valid.

34. The method of claim 33, also including the step of decrementing the suspicion count to zero whenever the user transaction is determined to be non-suspicious.