ATM Fraud Detection Using Conditional Data Collection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting bank transfer fraud at ATMs, such as those described in Patent Literature 1, collect personal user information even in cases where no incident occurs, leading to a significant psychological burden on users.
Innovation Solution
An information processing apparatus and system that determine whether a user at an ATM is talking on a phone using first data like images, voice data, or radio waves, and only acquire second data suggesting the user's state if a phone call is detected, thereby reducing unnecessary data collection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If personal information is collected by camera and microphone even when no incident occurs, then detection capability is improved, but psychological burden on users increases
Solution Approach 1:
The system performs preliminary detection using only first data (captured image, voice data, or radio wave data) to determine whether a phone call is taking place before deciding to collect second data (user state information). This preliminary action avoids unnecessary collection of personal information when no incident is detected, thereby reducing psychological burden while maintaining detection capability.
Solution Approach 2:
The data collection process is segmented into two stages: first data collection (always performed) and second data collection (conditionally performed only when phone call is detected). This segmentation allows the system to balance detection reliability with user privacy protection by collecting detailed personal information only when necessary.
2Measurement precision
If second data is collected in all cases, then incident detection accuracy is improved, but data processing load increases
Solution Approach 1:
The system applies partial action by collecting second data (detailed user state information) only in specific cases where a phone call is detected, rather than in all cases. This reduces the overall data processing load while maintaining sufficient incident detection accuracy for high-risk situations.
Data Source
AI summary
An information processing apparatus (10) includes a call determination unit (14) that determines whether a user who has visited an ATM is talking on a phone based on first data that is any one of a captured image, voice data, and received radio wave data related to a target area based on the ATM, an operation control unit (15) that starts acquisition of second data that suggests a state of the user and has a data type different from that of the first data in a case where it is determined that the user is talking on a phone, and a case determination control unit (16) that acquires a determination result of presence or absence of an incident using the first data and the second data.


