Eye-Tracking Liveness Detection for Mobile Anti-Spoofing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing security systems face challenges in distinguishing between live individuals and photographic images, with methods like 3D scanning being costly and time-consuming, and manual identification relying on human memory, making them susceptible to spoofing.
Innovation Solution
A computer-implemented method using eye gaze detection on a mobile device with a front-facing camera to determine spoofing attempts by comparing user eye gaze images with a registration threshold, employing random sequences of eye gaze icons and machine learning to verify live presence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3D scanning is used for biometric authentication, then liveness detection accuracy is improved, but device cost and processing time increase significantly
Solution Approach 1:
The patent uses 2D photographs as simplified copies of 3D biometric data. Instead of capturing complex 3D facial geometry with expensive scanners, the system captures standard 2D images that can be stored and processed easily. Multiple 2D images from different angles are combined to create a comprehensive biometric profile, achieving adequate liveness detection without the cost and complexity of true 3D scanning hardware.
Solution Approach 2:
The patent replaces mechanical 3D scanning systems with a combination of 2D imaging and software-based analysis. The eye gaze detection algorithm processes 2D images to infer three-dimensional spatial relationships and movement patterns, substituting physical 3D capture hardware with computational methods that achieve similar authentication goals at lower cost.
2Device complexity
If manual identification by security guards is used, then implementation cost is reduced, but reliability and anti-spoofing capability deteriorate
Solution Approach 1:
The system enables automated self-service authentication where the user's eye movements and gaze patterns are automatically captured and analyzed without human intervention. The algorithm independently evaluates whether the captured images represent a live person or a photograph, eliminating the need for security guards while maintaining or improving reliability through consistent algorithmic decision-making.
Solution Approach 2:
The system incorporates feedback mechanisms where the authentication algorithm continuously analyzes eye gaze patterns and provides real-time evaluation of liveness. The system can detect subtle indicators of spoofing attempts and adjust its decision-making based on multiple observed parameters, creating a reliable automated system that surpasses human capability in detecting photographic fraud.
3Productivity
If photographic identification is used, then processing speed is improved, but susceptibility to spoofing increases
Solution Approach 1:
The patent employs periodic eye movement stimulation where the system presents moving or changing visual stimuli and observes the user's periodic eye tracking responses. Live humans naturally follow moving targets with their eyes in predictable patterns, while photographs remain static. This periodic action creates a dynamic verification process that maintains fast processing speeds while effectively distinguishing live users from photographic spoofs.
Solution Approach 2:
The system transitions from static photographic comparison to dynamic eye movement analysis. By capturing and analyzing the dynamic behavior of eye gaze over time, the system adds a temporal dimension to authentication. The algorithm evaluates how eyes naturally move, focus, and track objects, creating a living authentication process that is inherently resistant to static photographic spoofing while maintaining rapid processing.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided herein is are computer-implemented methods, computer-implemented systems, and non-transitory computer-readable storage media for registering a user's eye gaze profile via a mobile device having a screen and a front facing camera. Also provided herein is are computer-implemented methods, computer-implemented systems, and non-transitory computer-readable storage media for determining a spoofing attempt by a user during a request to access a resource via a mobile device.