Liveness Detection via Device Motion and Illumination Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current liveness detection methods in biometric authentication are inadequate in distinguishing live users from spoofing attempts, particularly with high-definition video playback, leading to unreliable authentication results.
Innovation Solution
A computing device equipped with a gyroscope, accelerometer, and machine learning algorithms captures face biometric data as video, generating signals from angular changes and illumination variations to determine user liveness by moving from a first to a second position, and uses similarity scores and classification models to verify identity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional liveness detection methods are used, then the authentication process is simple, but the detection accuracy is insufficient against high-definition video playback spoofing
Solution Approach 1:
The system transitions from static image-based liveness detection to dynamic video-based detection with multiple captured frames. The method analyzes temporal changes in biometric data across sequential frames, including facial movements, head rotations, and illumination changes, making the detection process dynamic and adaptable to spoofing attempts.
Solution Approach 2:
The patent introduces angular position as an additional dimension for analysis. By tracking the angular position of the computing device and correlating it with changes in biometric image data, the system creates a multi-dimensional verification approach that goes beyond traditional two-dimensional image analysis.
2Reliability
If multiple parameters and signals are analyzed for liveness detection, then the reliability of authentication improves, but the computational complexity increases
Solution Approach 1:
The system segments the liveness detection process into distinct analytical components: capturing biometric video data, generating parameters from individual frames, creating signals from parameter changes, and calculating similarity scores. This segmentation allows each component to be processed independently and efficiently.
Solution Approach 2:
The system employs feedback mechanisms by comparing signals generated during authentication with signals stored in a database. Similarity scores are calculated based on this feedback comparison, and the system uses this information to determine whether to grant or deny access, creating a closed-loop verification process.
Data Source
Figure 1
Figure 2~4
Figure 5~7
AI summary
A method for verifying the identity of a user is provided that includes generating, by a computing device, a parameter for each processed frame in a video of biometric data captured from a user. The parameter results from movement of the computing device during capture of the biometric data. Moreover, the method includes generating a signal for the parameter and calculating a confidence score based on the generated signal and a classification model specific to the user. The classification model is generated from other signals generated for the parameter. Furthermore, the method includes verifying the identity of the user as true when the confidence score is at least equal to a threshold score.