Self-Checkout POS Registration with Weight Tolerance and Image Check
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing self-service POS terminals face errors in product registration even when purchasers correctly register commodities due to discrepancies between measured weights and expected reference weights, and are vulnerable to fraudulent activities like swapping commodities.
Innovation Solution
A POS terminal equipped with a weight sensor, camera, and image processing capabilities that verify the registered commodity's weight and visual appearance against stored data to ensure accurate registration, minimizing errors and detecting fraud.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If weight-based fraud detection is implemented using a scale, then fraudulent activities can be detected, but registration errors occur when measured weight does not exactly match expected reference weight
Solution Approach 1:
The patent combines multiple detection methods (weight measurement, image recognition, and commodity information matching) into a unified fraud detection system. The processor integrates data from the weight sensor, camera, and commodity database to comprehensively verify whether the placed commodity matches the registered commodity, thereby maintaining high fraud detection accuracy while reducing false positives from weight variations alone.
Solution Approach 2:
The system introduces an intermediary verification mechanism that compares the measured weight against a reference weight range derived from commodity information stored in memory. Rather than requiring exact weight matching, the system uses the reference weight as a mediator to establish an acceptable tolerance range, allowing for natural weight variations while still detecting fraudulent substitutions.
2Measurement precision
If strict weight matching is enforced for registration, then measurement accuracy improves, but registration smoothness deteriorates due to frequent errors
Solution Approach 1:
The patent changes the parameter from exact weight matching to weight range matching. By storing reference weight information in the commodity database and comparing measured weights against this reference within an acceptable tolerance, the system maintains measurement precision while accommodating natural weight variations, thereby ensuring smooth registration operations without frequent error interruptions.
3Device complexity
If only weight measurement is used for verification, then device complexity is reduced, but reliability of fraud detection deteriorates
Solution Approach 1:
The patent merges multiple verification modalities (weight sensor data, camera image capture, commodity information from database, and visual appearance comparison) into a unified fraud detection process. The processor synthesizes information from all these sources to reliably determine whether the placed commodity matches the registered commodity, achieving high fraud detection reliability without excessive system complexity.
Solution Approach 2:
The system implements a multi-functional verification mechanism where the same processor handles weight measurement analysis, image recognition, commodity information retrieval, and fraud determination. This universal approach allows a single integrated system to perform multiple verification functions, maintaining reliability while avoiding the need for separate specialized devices for each function.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables smooth and accurate product registration by reducing errors and detecting fraudulent activities, ensuring reliable transaction processing.
Implementation Method 1
a weight sensor disposed in or under the first table
Implementation Method 2
a camera positioned to image the first table
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A point-of-sale terminal includes a table, a weight sensor, a camera, a memory storing commodity information each indicating a commodity and a weight range and a preset image thereof, and a processor configured to: determine whether an operation for registering a first commodity is input, upon determining that the operation is input, search the memory for first commodity information corresponding to the first commodity, determine whether a weight measured by the sensor increases, upon determining that the weight increases, determine whether an increase in the weight is within a first range indicated by the first commodity information, and upon determining that the increase is not within the range, determine whether an image captured by the camera shows the first commodity based on a first preset image indicated by the first commodity information, and upon determining that the image shows the first commodity, register the first commodity.