Biometric Authentication Feature Selection via Signal Validity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing personal recognition techniques face challenges in maintaining high accuracy under varying illumination and environmental conditions, and require subjects to be positioned unnaturally for authentication, leading to increased error rates and user discomfort.
Innovation Solution
An information processing apparatus and method that evaluates the validity of feature detection signals, selects suitable features, and uses environmental information, including image and voice data, to perform robust personal recognition while keeping the subject in a natural state, using a combination of image and voice sensing, hierarchical neural networks for feature detection and validation, and secondary authentication for accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple biometric features are combined for verification, then authentication accuracy is improved, but error occurrence rate increases when feature detection signals have low quality
Solution Approach 1:
The system dynamically adjusts the weightings of different biometric features based on their detection quality scores. Features with higher quality scores receive greater weight in the final authentication decision, while low-quality features are downweighted or discarded. This parameter adjustment allows the system to maintain high authentication accuracy while minimizing errors from poor-quality detections.
Solution Approach 2:
The system incorporates feedback loops that continuously monitor the quality of feature detection signals and adjust the authentication process accordingly. Quality scores are calculated based on signal characteristics, and this feedback information is used to modulate the verification process in real-time, ensuring that only reliable features contribute to authentication decisions.
2Measurement precision
If subjects are positioned just in front of camera or made to utter voice according to instruction, then authentication can be performed, but subject is made conscious of being authenticated and natural state is lost
Solution Approach 1:
The system performs authentication by detecting features that naturally occur in the environment without requiring active participation from the subject. The authentication process serves itself by automatically detecting and evaluating biometric features as they naturally appear, eliminating the need for subjects to pose or utter phrases according to instructions.
Solution Approach 2:
The system introduces environmental information as an intermediary element that bridges the gap between natural subject behavior and authentication requirements. By detecting features in the natural environment rather than requiring directed subject action, the intermediary approach allows authentication to occur without making the subject conscious of being authenticated.
3Adaptability or versatility
If feature detection is performed under various viewing conditions, then recognition should be robust, but recognition accuracy decreases when illumination and environmental conditions are unfavorable
Solution Approach 1:
The system dynamically adjusts detection parameters and feature selection based on environmental conditions. When illumination or environmental conditions are unfavorable, the system changes which features are detected and how they are weighted, allowing the recognition system to adapt to varying conditions while maintaining accuracy.
Solution Approach 2:
The authentication system transitions from a static feature-set approach to a dynamic feature-selection approach. The system continuously adapts its detection strategy based on real-time environmental assessment, selecting and weighting features according to current conditions rather than relying on a fixed set of features regardless of environment.
Data Source
AI summary
A plurality of kinds of feature amounts are collected from image information and voice information on a person existing in a space, valid values of the collected feature amounts are calculated, feature amounts to be used for personal recognition are determined in the collected feature amounts on the basis of the calculated valid values, and personal recognition is performed by using the determined feature amounts.