Biometric Template Evolution for Aging Identity Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine learning models for biometric authentication, such as facial recognition, struggle to account for changes in physical characteristics over time, leading to authentication failures due to aging, and require extensive resources to store numerous samples.
Innovation Solution
A system that trains a first machine learning model on a dataset of user images, generates image feature templates associated with positive authentications, and uses a second model to predict these templates, allowing for authentication based on comparisons between current and predicted feature templates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If numerous samples of physical characteristics are taken and stored over time, then authentication accuracy for aging individuals is improved, but technical resource requirements increase
Solution Approach 1:
The patent extracts only the essential feature representations from biometric samples rather than storing complete raw images or extensive sample sets. The system extracts salient features that capture aging patterns while discarding redundant information, achieving accurate authentication with minimal storage requirements.
Solution Approach 2:
The system transforms biometric data from raw image format into compressed feature vector representations. This parameter transformation reduces storage requirements while preserving the essential characteristics needed for authentication, including aging-related changes.
2Speed
If a machine learning model is trained on initial biometric data, then authentication speed is improved, but the model cannot account for changes in physical characteristics over time
Solution Approach 1:
The patent implements dynamic template updating where the reference biometric template is not static but evolves over time. The system periodically updates the reference template to reflect current physical characteristics, allowing the model to adapt to aging while maintaining fast authentication through the preserved feature-based comparison mechanism.
Solution Approach 2:
The system incorporates feedback loops where authentication results and new biometric samples are used to continuously refine and update the reference template. This feedback mechanism enables the model to adapt to physical changes over time while maintaining efficient authentication performance.
3Adaptability or versatility
If the reference template is frequently updated to reflect aging, then adaptability to physical changes is improved, but system complexity and processing overhead increase
Solution Approach 1:
The patent implements periodic template updates rather than continuous updates. The reference template is updated at predetermined intervals or under specific conditions, reducing processing overhead while maintaining adaptability to aging. This periodic approach balances the need for adaptability with system simplicity and efficiency.
4Measurement precision
If extensive biometric data is stored to improve authentication reliability, then measurement precision is improved, but loss of time for data management increases
Solution Approach 1:
The system extracts only the essential feature representations from biometric samples rather than storing complete raw images or extensive sample sets. This extraction process maintains measurement precision by preserving critical characteristics while dramatically reducing data management time and computational overhead.
Data Source
AI summary
Provided are systems for authenticating an individual using image feature templates that include at least one processor to train a first machine learning model based on a training dataset of a plurality of images of a user, generate a plurality of image feature templates using the first machine learning model, wherein each image feature template of the plurality of image feature templates is associated with a positive authentication of the identity of the user during a time interval, generate a second machine learning model based on the plurality of image feature templates, generate a predicted image feature template using the second machine learning model, determine whether to authenticate the identity of the user based on an input image of the user, and perform an action based on determining whether to authenticate the identity of the user. Methods and computer program products are also provided.