Information Code Fraud Detection via Visual Class Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for detecting fraud in information codes, such as QR codes, are inadequate as they cannot reliably determine if an information code has been replaced, especially when fraudulent information is encoded along with detection information, making it difficult for users to visually recognize the original information.
Innovation Solution
An information processing system that compares class information obtained from original and displayed visual information with class information generated from the information code, using image classification and position information to verify the authenticity of the displayed information code, thereby preventing fraudulent processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If information for detection is encoded into the information code, then fraud detection capability is improved, but fraudulent information can also be encoded alongside it making detection impossible
Solution Approach 1:
The patent divides the verification process into two independent parts: (1) decoding the information code to obtain encoded information, and (2) obtaining verification information from visual information that should be displayed together with the code. By segmenting the verification data into separate visual elements rather than embedding all verification data within the code itself, the system prevents fraudsters from encoding both fraudulent content and valid verification data in the same code.
Solution Approach 2:
The patent introduces visual information as an intermediary element that serves as the verification medium. Instead of relying solely on data within the information code, the system uses separate visual information (such as images or text displayed near the code) as a mediator to verify authenticity. This intermediary approach allows the system to maintain verification capability while preventing fraudulent encoding.
2Productivity
If information code is configured with geometric figures, then encoding capability is improved, but visual recognizability of original information deteriorates
Solution Approach 1:
The patent uses visual information as an intermediary that bridges the gap between the geometric information code and human understanding. While the information code itself remains as geometric figures for efficient encoding, the accompanying visual information serves as a mediator that presents the original information in a visually recognizable format, allowing users to verify content without decoding the geometric code.
Solution Approach 2:
The patent segments the information presentation into two distinct components: the geometric information code for machine reading and encoding, and separate visual information for human visual recognition. This segmentation allows each component to optimize its respective function without compromising the other.
3Object-affected harmful factors
If information code is replaced with another code, then fraud occurs, but visual information comparison becomes difficult without class classification
Solution Approach 1:
The patent transforms the verification task from direct visual comparison of complex images to comparison of simplified class parameters. By converting visual information into class classifications (such as object categories or semantic labels), the system changes the parameter space from pixel-level complexity to concept-level simplicity, making fraud detection more accessible while managing system complexity through standardized classification schemes.
Data Source
AI summary
To easily determine whether an information code has been replaced, an information processing device obtains, based on original information obtained by photographing an optically readable information code and decoding the same, first class information indicating a class into which an object visually represented by first visual information which should be displayed together with an information code is classified. The information processing device obtains second class information indicating a class into which an object visually represented by second visual information photographed together with the information code because of display together with the information code is classified. The information processing device compares the first class information with the second class information. The information processing device controls, based on a result of the comparison, execution of predetermined processing using the original information.