Retail Checkout Item Verification Using Image Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional point-of-sale systems in retail settings face challenges in verifying items accurately, leading to transaction errors and theft, particularly in self-checkout systems, due to issues like scan avoidance and label switching, which are not effectively addressed by existing security measures.
Innovation Solution
A data reading system that captures images of items and compares them to stored reference data using a Siamese neural network and optical character recognition, eliminating the need for item recognition and minimizing computationally intensive processes, thereby quickly and efficiently verifying transactions and detecting potential discrepancies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If security personnel monitor transactions and review receipts to prevent theft, then theft detection capability improves, but operational cost and time consumption increase
Solution Approach 1:
The system enables self-service verification by automatically capturing images of items and comparing them to reference data without requiring manual intervention from security personnel or store clerks. The automated image comparison system performs verification independently, eliminating the need for human monitoring while maintaining reliable theft detection.
Solution Approach 2:
The patent replaces manual mechanical verification processes with an automated optical and computational system. Instead of human eyes and hands physically checking items and receipts, the system uses image capture devices and computer processing to automatically verify items, substituting mechanical human labor with automated technological processes.
2Measurement precision
If weight-based verification is used to detect label switching, then detection accuracy improves for items with different weights, but it fails to detect theft for items with same or similar weights
Solution Approach 1:
The system changes the verification parameter from weight measurement to visual image comparison. Instead of relying on weight differences to detect label switching, the system captures images of items and compares visual characteristics such as label appearance, packaging features, and product morphology. This parameter change enables detection of label switching for items with identical or similar weights.
Solution Approach 2:
The patent introduces an intermediary comparison mechanism using reference data and image processing algorithms. The system captures images of items being scanned, compares them against stored reference images of legitimate products, and identifies discrepancies. This intermediary visual comparison process serves as a mediator between the scanned item and the expected product, detecting fraud regardless of weight similarities.
3Reliability
If conventional image capture and comparison methods are used, then item verification is performed, but the process is computationally intensive and time-consuming
Solution Approach 1:
The system extracts only the essential visual features and characteristics from captured item images for comparison, rather than processing the entire image data. By identifying and comparing specific key features such as label patterns, product shapes, and packaging characteristics, the system reduces computational complexity while maintaining verification accuracy. This feature extraction approach speeds up the verification process.
Solution Approach 2:
The patent implements preliminary action by pre-storing reference images and characteristic data of legitimate products in a database before actual verification occurs. During verification, the system quickly compares captured item images against these pre-prepared references using optimized algorithms, eliminating the need for complex real-time analysis from scratch and significantly improving verification speed.
Data Source
AI summary
The disclosure relates to a data reading system operable for obtaining and decoding data from items processed by a customer in a retail transaction, where the data reading system compares an image of the item obtained during the retail transaction to reference image data of the item to verify the identity of the item and ensure that optical code information for the item has not been altered prior to processing. If a discrepancy is identified during the data reading process, the data reading system generates an exception identifying the potential issue for further review.