Unobtrusive User Identity Verification via Multi-Model Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mobile device user verification methods are burdensome, requiring explicit user input and relying on a single verification model, which can fail in various contexts, such as low light or noisy environments, limiting continuous and accurate verification.
Innovation Solution
A computing device creates and updates multiple verification models unobtrusively using enrollment data collected during normal usage, allowing for transparent and continuous user verification by combining data streams from different models, such as speech and vision-based models, to improve accuracy and reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single verification model is used, then the verification process is simple, but the verification accuracy decreases in certain contexts such as low light or noisy environments
Solution Approach 1:
The patent combines multiple verification models (vision-based face verification, speech-based speaker verification, and text-dependent speaker verification) into a unified verification system. The system integrates results from different models to make verification decisions, thereby improving reliability in contexts where a single model may fail due to environmental factors such as low light or noisy environments.
Solution Approach 2:
The verification system is designed to perform multiple verification functions using different models. The system can switch between or combine vision-based, speech-based, and text-dependent verification methods depending on the context, making the verification process universally applicable across various environmental conditions and use cases.
2Ease of operation
If explicit security metadata input is required, then verification security is maintained, but user convenience deteriorates due to repeated manual input
Solution Approach 1:
The system automatically collects verification data without requiring explicit user input. It passively captures vision-based data from the camera, speech-based data from the microphone, and text-dependent data from keyboard input, thereby improving ease of operation while maintaining verification security through continuous background verification.
Solution Approach 2:
The verification process operates continuously in the background rather than requiring discrete manual authentication events. The system continuously monitors and verifies user identity through multiple data streams, enabling seamless verification that maintains security without interrupting user workflow.
3Reliability
If continuous background verification is implemented, then security monitoring is improved, but user burden increases due to constant verification requirements
Solution Approach 1:
The system performs continuous verification automatically without user intervention. It collects verification data from ongoing user interactions with the device, such as speech patterns, facial recognition, and typing behavior, thereby enabling continuous security monitoring that is transparent to the user and does not increase operational burden.
Solution Approach 2:
The system collects and processes verification data in advance during normal device usage. By continuously gathering vision, speech, and text data during regular interactions, the system prepares verification results beforehand, enabling rapid authentication decisions without requiring users to explicitly initiate verification processes.
Data Source
AI summary
Techniques for unobtrusively verifying the identity of a user of a computing device are provided. In one embodiment, the computing device can establish one or more verification models for verifying the user's identity, where at least a subset of the one or more verification models is based on enrollment data that is collected in an unobtrusive manner from the user. The computing device can then verify the user's identity using the one or more verification models.