Multi-Spectral Illumination for Biometric Liveness Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Biometric facial recognition systems are vulnerable to spoofing by counterfeit versions of authorized users' facial features, such as photographs or masks, which can deceive the system into authorizing unauthorized access.
Innovation Solution
The use of controlled illumination with specific spectral content, including ultraviolet, visible, near-infrared, and short-wave infrared bands, to distinguish between real human faces and counterfeit images by measuring pixel values and comparing them to enrollment image data, determining liveness and preventing unauthorized access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If biometric facial recognition systems use standard imaging to identify authorized users, then authentication speed and ease of operation are improved, but the system becomes vulnerable to spoofing by counterfeit images
Solution Approach 1:
The system changes the spectral parameter of illumination by using multiple wavelength bands (ultraviolet, visible, near-infrared, and short-wave infrared) to illuminate the face. Different materials (real skin vs. counterfeit materials) reflect these wavelengths differently, creating distinctive spectral signatures that enable reliable differentiation between live faces and spoofing attempts while maintaining authentication speed.
Solution Approach 2:
The patent introduces spectral analysis as an intermediary layer between image capture and authentication decision. The system captures images across multiple spectral bands, analyzes the reflectance characteristics as an intermediate step, and only then makes the authentication decision. This intermediary spectral verification prevents spoofing without significantly slowing down the overall authentication process.
2Measurement precision
If the system captures images across multiple spectral bands to detect counterfeits, then measurement precision for distinguishing real faces from counterfeit images is improved, but device complexity increases
Solution Approach 1:
The imaging device is designed to capture images across multiple spectral bands (ultraviolet, visible, near-infrared, and short-wave infrared) using a single multi-functional sensor system. This universal device performs both standard facial recognition and spectral-based liveness detection simultaneously, avoiding the need for separate specialized hardware for each function and thereby reducing overall system complexity.
Solution Approach 2:
The patent merges the liveness detection function with the standard facial recognition imaging process. By combining multiple spectral band captures into a single integrated imaging workflow, the system achieves high measurement precision for counterfeit detection without proportionally increasing device complexity. The same imaging hardware serves dual purposes: authentication and anti-spoofing.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Effectively prevents spoofing by accurately differentiating between real human faces and counterfeit images, enhancing the security of biometric authentication systems by ensuring only live human faces can gain access.
Implementation Method 1
controlled illumination with specific spectral content, including ultraviolet, visible, near-infrared, and short-wave infrared bands
Implementation Method 2
acquire a first image of the human face using a camera and controlled illumination
Data Source
AI summary
A computer, including a processor and a memory, the memory including instructions to be executed by the processor to acquire a first image from a first camera by illuminating a first object with a first light and determine an object status as one of a real object or a counterfeit object by comparing a first measure of pixel values corresponding to the first object to a threshold.