Liveness Detection via Displacement Instructions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing liveness detection methods for network-based biometric authentication are inconvenient and inaccurate, making it difficult to distinguish between genuine user image data and fraudulent biometric data, such as photographs or replayed images, during remote transactions.
Innovation Solution
A method involving displacement instructions is used to capture and analyze facial image data by translating, scaling, and/or rotating the image on a screen to determine if the image is taken by a live person, by comparing the calculated translation distances against predefined instructions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing liveness detection methods (motion structure, eye blinks, pattern illumination) are used, then liveness detection capability is provided, but convenience deteriorates and accuracy remains insufficient
Solution Approach 1:
The system performs preliminary actions by displaying a challenge image before capturing the user's response image. The challenge image is processed to generate displacement instructions that define expected spatial relationships. This preliminary setup enables accurate liveness detection without requiring complex real-time analysis during the actual capture moment, thus improving both accuracy and user convenience.
Solution Approach 2:
The system implements feedback by comparing the captured image's feature points against the displacement instructions generated from the challenge image. The calculated translation distance is compared with the expected translation distance, providing immediate feedback on whether the captured image represents a live user. This feedback mechanism enhances detection accuracy while maintaining simple user interaction.
2Reliability
If displacement instruction method is implemented, then liveness detection accuracy is improved, but device complexity increases
Solution Approach 1:
The system uses copying by creating a virtual representation of the challenge image and its expected displacement through computational geometry. Instead of complex hardware modifications, the patent copies the challenge image's spatial relationships into displacement instructions that can be processed algorithmically. This approach achieves high accuracy through software-based virtual modeling rather than physical system complexity.
Solution Approach 2:
The patent replaces mechanical or physical liveness detection mechanisms with computational and algorithmic approaches. Instead of using complex physical sensors or mechanical verification systems, the invention substitutes these with image processing algorithms that calculate translation distances and compare them against predefined displacement instructions, thereby reducing hardware complexity while maintaining or improving detection accuracy.
Data Source
AI summary
A method for capturing genuine user image data is provided that includes the steps of creating a displacement instruction, and displaying, by an electronic device, facial image data of a user in accordance with the displacement instruction. Moreover, the method includes the steps of positioning the displayed user facial image data to be located within a screen of the electronic device, capturing facial image data of the user, and calculating a translation distance of the captured user facial image data. Furthermore, the method includes the steps of comparing the calculated translation distance against a translation distance calculated from information in the displacement instruction to determine whether the calculated translation distance is in accordance with the displacement instruction. In response to determining the calculated translation distance is in accordance with the displacement instruction, determining the captured user facial image data is genuine and thus taken of a live person.