System and method for preventing cashier and customer fraud at retail checkout
Systems and methods for preventing fraud during retail checkout are described. A system includes: item identifier acquisition devices to acquire identifiers (e.g., barcodes) of items to be transacted; cameras or imagers to acquire images of the items to be transacted; an object recognition component to perform visual recognition of the acquired items through comparison of the acquired images of the items with a database and obtain identifiers of items represented in the database that correspond to an acquired image according to a correspondence criterion; a comparison component to compare a set of identifiers acquired through the item identifier acquisition devices with a set of identifiers obtained through the object recognition component; and an alerting component to provide an alert in case of discrepancies between the two sets.
1. A method for detecting discrepancies between first and second sets of merchandise identifiers that correspond to merchandise items to be identified in a retail checkout transaction, the method comprising:
acquiring encoded information from optical codes borne by at least some of the merchandise items to be identified in the retail checkout transaction, the encoded information corresponding to a first set of merchandise identifiers;
acquiring images of at least some of the merchandise items to be identified in the retail checkout transaction to obtain scale-invariant visual feature data based on visual appearance of the at least some of the merchandise items;
comparing the scale-invariant visual feature data with a database of stored scale-invariant visual feature data corresponding to known merchandise items to obtain a second set of merchandise identifiers;
detecting differences between the first and second sets of merchandise identifiers by comparing presence or absence of merchandise identifiers of the first set of merchandise identifiers with the presence or absence of merchandise identifiers of the second set of merchandise identifiers, in which the comparing is configured to be performed identifier by identifier or between the sets as a whole;
determining a price difference between merchandise items of the first and second sets; and
generating, in response to detecting a merchandise identifier of the second set of merchandise identifiers is absent from the first set of merchandise identifiers, transaction information indicating a user responsible for facilitating acquisition of the encoded information has avoided inclusion of a merchandise item in the retail checkout transaction.
2. The method of claim 1 , in which the transaction information includes a suspicion level based on a total number or cash value of discrepancies.
3. The method of claim 2 , further comprising:
adjusting a suspicion level in response to a discrepancy detected between the first and second sets of merchandise identifiers.
4. The method of claim 3 , in which the step of generating comprises:
generating an exception in response to the suspicion level exceeding a predetermined threshold level.
5. The method of claim 1 , in which the step of generating comprises:
immediately generating an alert in response to detecting a merchandise identifier of the second set of merchandise identifiers is absent from the first set of merchandise identifiers.
6. The method of claim 5 , further comprising:
providing the alert to store management or security personnel.
7. The method of claim 5 , further comprising:
providing the alert to the user or a customer.
8. The method of claim 1 , further comprising:
suspending the retail checkout transaction in response to detecting a merchandise identifier of the second set of merchandise identifiers is absent from the first set of merchandise identifiers.
9. The method of claim 1 , in which the step of generating comprises:
generating an alert in response to detecting a merchandise identifier of the second set of merchandise identifiers is absent from the first set of merchandise identifiers, and in response to the user indicating no further merchandise items are to be included in the retail checkout transaction.
10. The method of claim 1 , in which the optical codes are universal product codes (UPCs).
11. The method of claim 1 , in which the scale-invariant visual feature data comprises scale-invariant feature transform (SIFT) features, and in which a merchandise item is represented by a spatial relationship between multiple SIFT features.
12. The method of claim 1 , further comprising:
generating an alert in response to the price difference exceeding a predetermined threshold level.
13. The method of claim 1 , further comprising:
adding the price difference to a total of price differences associated with the user.
14. The method of claim 13 , further comprising:
generating an alert in response to the total of price differences exceeding a predetermined threshold for a predetermined period.
15. The method of claim 1 , in which the first and second sets of merchandise identifiers include universal product code (UPC) numbers.
16. The method of claim 1 , in which the optical codes are universal product codes (UPCs), the method further comprising:
generating, in response to detecting a merchandise identifier of the first set of merchandise identifiers being absent from the second set of merchandise identifiers, transaction information indicating a UPC has been substituted.
17. The method of claim 1 , further comprising:
generating a report of the transaction information for delivery to store management or security personnel at a predetermined time.
18. A method for detecting discrepancies between first and second sets of merchandise identifiers that correspond to merchandise items to be identified in a retail checkout transaction, the method comprising:
acquiring encoded information from optical codes borne by at least some of the merchandise items to be identified in the retail checkout transaction, the encoded information corresponding to a first set of merchandise identifiers;
acquiring images of at least some of the merchandise items to be identified in the retail checkout transaction to obtain scale-invariant visual feature data based on a physical property of the at least some of the merchandise items;
comparing the scale-invariant visual feature data with a database of stored scale-invariant visual feature data corresponding to known merchandise items to obtain a second set of merchandise identifiers;
detecting a merchandise identifier of the second set of merchandise identifiers as being absent from the first set of merchandise identifiers;
determining a price difference between merchandise items of the first and second sets; and
generating a report indicating a user responsible for facilitating acquisition of the encoded information has avoided inclusion of a merchandise item in the retail checkout transaction.
19. The method of claim 18 , further comprising:
providing reports periodically to store management or security personnel.