Liveness Detection via Eye Movement Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current liveness detection methods are inadequate for distinguishing between a living human and artificial representations, as they often require specific user cooperation and rely on hardware like infrared or depth cameras, which can be circumvented by static images or simple attacks, and their accuracy and robustness are insufficient for high-security applications.
Innovation Solution
A liveness detection method and system that utilizes eye movement analysis by tracking pupil position changes in response to a target on a screen, determining whether the correlation or variance of these movements meets predetermined thresholds to verify the presence of a living being, thereby reducing user interaction and enhancing system accuracy and usability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional liveness detection techniques are used that depend on specific hardware devices (such as infrared camera, depth camera), then the system can detect liveness, but the device complexity increases and cost rises
Solution Approach 1:
The patent replaces complex hardware-based liveness detection systems (infrared cameras, depth cameras) with a software-based eye movement analysis system that uses standard video acquisition devices. The solution substitutes mechanical/optical complex systems with algorithmic processing of ordinary video data, analyzing pupil position changes and eye movement patterns to determine liveness without requiring specialized hardware
Solution Approach 2:
The patent enables standard video cameras to perform both regular video capture and liveness detection functions. By analyzing eye movement patterns in ordinary video data, the system makes a single device serve multiple purposes, eliminating the need for separate specialized hardware while maintaining detection capability
2Reliability
If cooperative-style liveness detection systems are used that require users to make specific actions or stay still, then liveness can be verified, but user experience deteriorates and detection efficiency decreases
Solution Approach 1:
The patent implements a self-service liveness detection mechanism where the system automatically analyzes eye movement patterns without requiring user cooperation or specific actions. The analysis occurs passively during normal video capture, detecting liveness through unconscious eye movements, blinking, and pupil position changes that occur naturally during regular interaction with the device
Solution Approach 2:
The patent performs liveness detection continuously during the video capture process itself, rather than requiring a separate verification step. By analyzing eye movement patterns throughout the normal interaction period, the system completes liveness verification as part of the primary function, eliminating additional user burden
3Ease of operation
If simple image border detection methods are used to determine liveness, then the detection process is simple, but accuracy and robustness are insufficient for high-security applications
Solution Approach 1:
The patent transitions from simple geometric analysis (image borders) to analyzing dynamic physiological parameters (eye movement patterns, pupil position changes, blinking frequency). By measuring these biological parameters that change naturally in living subjects, the system achieves high accuracy while maintaining implementation simplicity through software-based analysis
Data Source
AI summary
The application provides a liveness detection method capable of implementing liveness detection on a human body, and a liveness detection system that employs the liveness detection method. The liveness detection method comprises: obtaining video data acquired via a video acquisition module; determining, based on the video data, a feature signal of an object to be detected; judging whether the feature signal meets a first predetermined condition, and if the feature signal meets the first predetermined condition, identifying that the object to be detected is a living body, wherein the feature signal is indicative of eye movement of the object to be detected.