Biometric Learning Data Generation for Spoofing Attack Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing threat of spoofing attacks in fingerprint recognition systems requires a method to generate data with high similarity and matching to effectively differentiate between genuine and forged biometric data, as existing technologies struggle to reliably distinguish between them.
Innovation Solution
A learning data generation method that selects and matches first and second data sets, each with a similarity greater than or equal to a threshold, to generate learning data, which is then used by a counterfeit detection system to determine whether input biometric data is normal or forged through an AI model, improving the system's ability to prevent unauthorized access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If biometric authentication is implemented to improve security, then user identification reliability is improved, but the system becomes vulnerable to spoofing attacks
Solution Approach 1:
The patent applies preliminary anti-action by proactively generating counterfeit biometric data that mimics genuine biometric characteristics before actual authentication occurs. This synthetic counterfeit data is used to train the authentication system to recognize and reject spoofing attempts, thereby preventing attacks before they can succeed.
Solution Approach 2:
The patent implements preliminary action by pre-generating and storing synthetic counterfeit biometric data that represents various spoofing methods. This preparation allows the system to be pre-trained on counterfeit examples, enabling it to quickly identify and reject actual spoofing attempts during authentication without requiring real-time analysis of attack methods.
2Ease of operation
If traditional fingerprint recognition is used to simplify authentication, then ease of operation is improved, but measurement precision between genuine and forged data deteriorates
Solution Approach 1:
The patent applies copying by creating synthetic copies of genuine biometric data that incorporate counterfeit characteristics. These copied and modified biometric templates are used to train the system to recognize the subtle differences between genuine and forged data, thereby improving detection accuracy without changing the simple authentication interface.
Solution Approach 2:
The patent implements parameter changes by modifying biometric data parameters to create synthetic counterfeit examples that vary in authenticity characteristics. By training the system on these parameter-varied examples, the system learns to detect subtle differences between genuine and forged data while maintaining the simplicity of the authentication process for users.
Data Source
AI summary
Provided is a learning data generation method including selecting first data and second data, each of the first data and the second data having a similarity to registered biometric data that is greater than or equal to a first threshold, determining a matching degree between the first data and the second data, and generating learning data based on matching the first data to the second data in response to a determination that the matching degree is greater than or equal to a second threshold.