Cross-Sensor Biometric Authentication Using a Local ML Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing biometric authentication systems are tied to specific sensors, limiting flexibility and requiring users to use the same sensor for both capturing reference and inference biometric data, which can lead to compatibility issues with different devices and manufacturers.
Innovation Solution
Decouple the training phase of a machine learning model from the inference phase by training the model using biometric data from a first sensor and authenticating using biometric data from a potentially different second sensor, allowing authentication to be independent of the sensor used.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If biometric authentication systems use the same sensor for both capturing reference and inference biometric data, then authentication reliability is improved, but device adaptability deteriorates
Solution Approach 1:
The patent segments the authentication system into two independent components: a training phase that creates a sensor-agnostic biometric model, and an inference phase that uses different sensors to capture biometric data. This segmentation allows the system to train on data from one sensor while making authentication decisions using data from potentially different sensors, thereby resolving the contradiction between maintaining high authentication reliability and achieving device adaptability.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between the sensor data and the authentication decision. This model is trained to recognize biometric patterns in a sensor-agnostic manner, acting as a mediator that can process and interpret biometric data from various sensor types and manufacturers. The intermediary model translates sensor-specific data into a universal biometric representation, enabling reliable authentication across different devices.
2Reliability
If additional tokens or smart cards are used for authentication, then security is improved, but device complexity deteriorates
Solution Approach 1:
The patent enables the communication device itself to perform authentication functions by storing and executing the trained machine learning model locally. The device uses its own sensors to capture biometric data and processes this data through the embedded model to make authentication decisions. This self-service approach eliminates the need for external authentication tokens or smart cards, as the device becomes self-sufficient in performing secure authentication, thereby reducing device complexity while maintaining or improving security.
Data Source
AI summary
Methods and devices for supporting authentication of a user to a service provided by a second communication device, wherein a first communication device sends to the second communication device a request for the user to access the service, wherein the request comprises credentials of the user. The first communication device receives inference biometric data of the user from the second communication device; determines whether the user can be authenticated using a machine learning, ML, model trained for classifying biometric data of the user and the received inference biometric data as input; and in response thereto, sends to the second communication device a message indicative of a confirmation or a rejection of the authentication of the user.