Biometric Authentication Using Single Sensor Multi-Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing biometric authentication systems face challenges in extracting non-overlapping biometric features from multiple body sites, such as the face and finger, without reducing authentication accuracy, especially when users wear masks or have skin issues, and in maintaining accuracy despite changes in position or posture.
Innovation Solution
A biometric authentication system that uses a single general-purpose sensor, like a visible light camera, to guide and image multiple non-overlapping body sites to an optimum presentation position, extracting biometric features with low pattern correlation for stable and accurate authentication, employing a combination of image processing and machine learning techniques to normalize posture and extract feature amounts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensors are used to capture different body sites, then authentication accuracy is improved, but device complexity increases
Solution Approach 1:
The patent employs a single visible light camera to perform multiple authentication functions by capturing images of different body sites (face, finger, hand vein) and extracting various biometric features (facial features, fingerprint features, vein pattern features) from the same sensor, thereby avoiding the need for multiple specialized sensors while maintaining authentication accuracy
Solution Approach 2:
The system changes the imaging parameters (illumination conditions, image processing methods) to extract different biometric features from the same captured image, enabling multiple authentication modes (face recognition, fingerprint recognition, vein pattern recognition) using a single sensor with adjusted parameters
2Device complexity
If a single sensor is used to capture multiple body sites, then device complexity is reduced, but authentication accuracy deteriorates
Solution Approach 1:
The patent segments the captured image to identify and extract different body sites (face region, finger region, hand region) and their corresponding biometric features independently, allowing accurate extraction of multiple non-overlapping features from a single composite image
Solution Approach 2:
The system uses image processing algorithms and feature extraction techniques as intermediaries to separate and identify different biometric features from the single captured image, enabling accurate authentication without requiring multiple physical sensors
3Measurement precision
If biometric features are extracted from overlapping body sites, then authentication accuracy is improved, but usability deteriorates due to position and posture changes
Solution Approach 1:
The patent dynamically adjusts the selection of biometric features based on the user's presentation posture and which body sites are visible in the captured image, allowing the system to flexibly switch between different authentication modes (face, finger, hand) depending on user convenience rather than requiring a fixed posture
4Measurement precision
If multiple biometric features are extracted from the same body site, then authentication accuracy is improved, but the system becomes vulnerable to position and posture changes
Solution Approach 1:
The patent transitions from extracting multiple features from the same body site (one-dimensional approach) to extracting features from multiple different body sites (multi-dimensional approach), thereby maintaining authentication accuracy while reducing vulnerability to position and posture changes by having backup authentication options
Data Source
AI summary
A biometric authentication system, including an image input unit configured to obtain an image by imaging a living body, a storage unit configured to store registration information relating to a plurality of biological features obtained from a biological region of an image of each person, and an authentication processing unit configured to process the biological region of the image obtained by the image input unit to execute biometric authentication based on the registration information, wherein the plurality of biological features obtained from the biological region of the each person are a plurality of biological features having a low pattern correlation with one another, and wherein the authentication processing unit is configured to combine the plurality of biological features having a low pattern correlation with one another, which are obtained by processing the image, to execute the biometric authentication.