Biometric Spoof Detection Using Neural Network Embedding Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fingerprint recognition technologies struggle to differentiate between genuine and fake biometric information, leading to potential security breaches when finely fabricated fake fingerprint patterns are used.
Innovation Solution
A method involving a neural network that extracts an embedding vector from an intermediate layer to detect spoofed biometric information, using a combination of scoring and classification techniques to determine whether the biometric information is live or spoofed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If traditional fingerprint recognition is used, then convenience and ease of access are improved, but security against spoofed biometric information deteriorates
Solution Approach 1:
The patent segments the fingerprint recognition process into multiple independent analysis components: texture analysis, frequency analysis, and spatial distribution analysis. Each component extracts specific features from different aspects of the fingerprint image, allowing the system to comprehensively evaluate both genuine and spoofed fingerprints through multiple independent verification channels
Solution Approach 2:
The patent transforms the fingerprint image into different parameter domains for analysis, including frequency domain transformation and texture parameter extraction. By analyzing the same fingerprint data through multiple parameter transformations, the system can identify characteristics that distinguish genuine fingerprints from spoofed ones without changing the underlying authentication convenience
2Productivity
If fingerprint recognition accepts fake fingerprint patterns, then authentication speed is improved, but measurement precision of biometric authenticity deteriorates
Solution Approach 1:
The patent performs preliminary analysis of fingerprint characteristics before making the final authentication decision. By pre-calculating texture features, frequency components, and spatial distribution patterns, the system prepares verification data in advance, enabling rapid spoof detection without compromising authentication speed
Solution Approach 2:
The patent replaces traditional mechanical fingerprint matching with computational analysis methods, including frequency domain transformation and texture analysis. This substitution enables the system to perform complex spoof detection calculations rapidly, maintaining high authentication speed while significantly improving measurement precision for detecting fake fingerprint patterns
Data Source
AI summary
A method with biometric information spoof detection includes extracting an embedding vector from an intermediate layer of a neural network configured to detect whether biometric information of a user is spoofed from an image including the biometric information; detecting first information regarding whether the biometric information is spoofed, based on the embedding vector; and detecting second information regarding whether the biometric information is spoofed based on whether the first information is detected, using an output vector output from an output layer of the neural network.