PreventSelf-Checkout Losses

Item switch is a common self-checkout fraud scenario in which a shopper intentionally selects a lower-priced product from the picklist instead of the actual item being purchased. For example, avocados may be entered as cucumbers, or a bottle of whiskey scanned as onions.

This type of fraud is especially common with unpackaged goods such as fruit, vegetables, and bakery items. Because the weight of the selected product can closely match the actual item, traditional weight-based security systems are often unable to detect it.

Otter Guard uses AI-powered visual recognition to identify every scanned item and compare it with the corresponding product image stored in the database. If a mismatch is detected, the system can immediately notify a checkout attendant or prompt the shopper to correct the selection.

This is another common self-checkout fraud scenario, where the barcode of a lower-priced item is intentionally placed on a more expensive product. Because high-value products are often targeted, the resulting losses can be significant.

Otter Guard visually verifies every scanned item by comparing the product captured at checkout with the corresponding image in the product database. Even if the barcode is scanned successfully, the system can detect visual mismatches and immediately notify store employees of the discrepancy.

Scan avoidance occurs when a shopper imitates the scanning motion without actually scanning the product or adding it to the transaction. In many cases, only a few low-cost items are scanned, while more expensive products are passed through the checkout area without being registered.

This type of fraud is difficult for checkout attendants to detect and is especially common at self-checkouts without security scales.

Otter Guard identifies scan imitation behavior by analyzing customer actions and verifying which items are transferred from the checkout counter to the bagging area. This enables the system to detect products that have not been properly scanned during the transaction.

In this scenario, Otter Guard counts the items placed at the self-checkout within a predefined and calibrated monitoring zone, then compares this number with the items scanned during the transaction.

Similar to scan avoidance, the application detects situations where not all products are scanned. Some items may be placed directly into bags without scanning, or intentionally left in the checkout area to be collected after payment.

This type of fraud is especially common at self-checkouts without security scales, where traditional loss prevention systems offer limited protection.

Otter Guard detects items left unscanned in the shopping basket after payment is completed. If any products remain in the basket, a store attendant is immediately notified.

Self-checkout monitoring should not end once scanning and payment are finished. In this fraud scenario, customers may intentionally leave unpaid items in the basket and only collect or bag them after paying for selected products.

This is another example where traditional security-scaled based loss prevention is not enough.

This application verifies that every self-checkout transaction is completed with a successful payment.

Using video analysis, the system monitors customer activity at the self-checkout area. If a shopper leaves the checkout before completing payment and remains absent for longer than a predefined period, the application immediately alerts store staff in real time.

To support quick identification and response, the alert includes two customer images captured from different camera angles — a top-down view and a side view.