Biometric Data Classification Using Neural Network Similarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing biometric authentication systems, particularly face recognition, face challenges in improving performance due to dependency on the composition of training and performance test databases, with unclear requirements for characteristic information to be stored, leading to inconsistent results.
Innovation Solution
A biometric authentication data classification device and method using an artificial neural network model to extract and calculate overall similarity between candidate data and performance test data, determining the appropriate database for new data based on similarity thresholds and distribution ratios, specifically for age, race, and gender characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If deep learning models are applied to face image authentication, then recognition performance can be improved, but performance becomes dependent on the composition of training and test databases
Solution Approach 1:
The patent changes the parameter of database composition by systematically varying the racial and gender distribution in training and test databases. This allows identification of optimal parameter settings (database compositions) that improve authentication performance while maintaining consistency across different scenarios.
Solution Approach 2:
The patent implements feedback mechanisms by evaluating authentication performance across multiple database compositions and using this feedback to optimize the training database. The system continuously adjusts database composition based on performance metrics to achieve both high reliability and adaptability.
2Measurement precision
If the training database is improved to enhance recognition performance, then authentication accuracy increases, but it becomes difficult to determine which characteristic information should be stored
Solution Approach 1:
The patent segments the database composition determination process into distinct components: racial characteristic analysis, gender characteristic analysis, and performance evaluation. This segmentation makes it manageable to identify which specific characteristic information should be stored by analyzing each component separately.
Solution Approach 2:
The patent performs preliminary analysis of database composition requirements before actual training. By pre-determining the optimal racial and gender distribution and characteristic information needs, it simplifies the subsequent training process and avoids the complexity of ad-hoc database composition decisions.
3Ease of manufacture
If 2D face image authentication is used, then the system is simple to implement, but authentication performance decreases depending on face detection direction
Solution Approach 1:
The patent transitions from 2D face image authentication to 3D face image authentication by introducing depth information as an additional dimension. This dimensional change enables the system to maintain simplicity while significantly improving authentication performance across different face detection directions through stereoscopic vision.
Data Source
AI summary
A method for adding biometric authentication training data into databases performed by a biometric authentication data classification device includes: extracting first biometric characteristic information from at least one candidate biometric training data for biometric authentication using an artificial neural network model; calculating an overall similarity between the first biometric characteristic information and second biometric characteristic information extracted from a performance test database of which a biometric authentication performance is lower than a threshold level, the performance test database being selected among performance test databases for the biometric authentication; and adding the at least one candidate biometric training data into one of the biometric authentication training database and the performance test database based on the calculated overall similarity.