POS Fraud Recognition Using Customer and Screen Image Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-service POS terminals in retail stores face challenges in detecting and preventing fraudulent activities due to reduced oversight and interaction with employees, necessitating additional technologies for fraud detection.
Innovation Solution
A fraud behavior recognition device and method that utilizes cameras and processor units to recognize customer behaviors and operations on self-service POS terminals, detecting fraudulent actions by analyzing image data and monitoring screen transitions, and notifying attendants of suspicious activities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If self-service POS terminals are introduced to reduce labor costs and prevent disease transmission, then operational efficiency and hygiene are improved, but fraud detection capability deteriorates due to reduced employee oversight
Solution Approach 1:
The patent introduces an image recognition system as an intermediary between the customer and the POS terminal. The system captures images of the terminal screen and customer operations, processes them through AI algorithms, and generates fraud risk assessments. This intermediary automated monitoring mechanism compensates for the absence of human employee oversight while maintaining self-service operational efficiency.
Solution Approach 2:
The patent replaces the mechanical system of human employee monitoring with an automated image recognition and analysis system. The system uses cameras to capture visual data, processes images through neural networks and AI algorithms, and automatically identifies fraudulent behaviors without requiring human intervention, thus substituting physical human oversight with computational analysis.
2Measurement precision
If automated image recognition is used to detect fraud, then fraud detection accuracy is improved, but system complexity increases
Solution Approach 1:
The patent designs the image recognition system to perform multiple functions simultaneously: capturing terminal screen content, recognizing customer operations, analyzing behavior patterns, and generating fraud risk assessments. By consolidating these diverse functions into a single integrated system, the patent achieves high fraud detection accuracy without proportionally increasing overall system complexity.
Solution Approach 2:
The image recognition system operates autonomously without requiring continuous human intervention. It automatically captures images, processes them through AI algorithms, identifies fraudulent behaviors, and generates reports. This self-service capability allows the system to maintain high detection accuracy while minimizing the operational complexity of human management and intervention.
Data Source
AI summary
A detection device for detecting a fraud behavior of a customer at a point-of-sale (POS) terminal in a store, including a first interface circuit configured to receive a first image of the customer from a camera, a second interface circuit configured to receive a second image that is displayed on the POS terminal, and a processor configured to acquire one or more first images via the first interface circuit and determine one or more behaviors of the customer based on the acquired first images, acquire one or more second images via the second interface circuit and determine, based on the acquired second images, one or more operations that have been made by the customer on the POS terminal, and determine that one of the behaviors is fraudulent based on one or more of the operations that have been made by the customer before said one of the behaviors.