Liveness Authentication via Infrared Waveband Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current face recognition systems are vulnerable to photo and flash attacks, where attackers use images or videos of a person to deceive the authentication system, and existing improvements struggle to differentiate between real and synthetic human face actions.
Innovation Solution
A liveness authentication method that filters incident light into different wavebands, determining whether the light is 'cold' or 'warm' by comparing the intensity of visible and invisible light components, such as infrared or ultraviolet, to distinguish between genuine human presence and synthetic displays like liquid crystal display screens.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If face recognition system uses only visible light comparison, then authentication is simple and fast, but it is vulnerable to photo attacks
Solution Approach 1:
The patent segments the light spectrum into multiple wavebands (visible light, infrared, ultraviolet) and detects each separately using dedicated sensors. This segmentation allows the system to analyze the spectral composition of incident light and distinguish between warm light (natural light reflected from human face) and cold light (light from display screens), thereby preventing photo attacks while maintaining system reliability.
Solution Approach 2:
The patent extends the detection dimension from only visible light intensity comparison to multi-dimensional spectral analysis by incorporating infrared and ultraviolet waveband detection. This dimensional expansion enables the system to identify the thermal characteristics and spectral signature of light sources, making it possible to differentiate between genuine human faces and photographic counterfeits.
2Reliability
If the system requires designated human face actions, then anti-photo attack capability is improved, but it becomes vulnerable to flash attacks using videos
Solution Approach 1:
The patent replaces the mechanical approach of requiring physical human actions with an optical/physical property-based detection method. Instead of analyzing dynamic facial movements, the system uses optical sensors to detect the physical characteristics of incident light (thermal radiation in infrared waveband and spectral composition), thereby identifying whether the light originates from a human face or a display screen showing video content.
Solution Approach 2:
The patent changes the detection parameter from visible light intensity and facial action analysis to multi-waveband light characteristics, particularly emphasizing infrared radiation detection. By monitoring the infrared component of incident light, the system can identify the thermal emission signature unique to human faces, which distinguishes genuine presence from video playback on cold display screens.
3Measurement precision
If the system detects invisible light wavebands, then it can distinguish warm light from cold light, but the device complexity increases
Solution Approach 1:
The patent segments the light spectrum into multiple wavebands (visible light, infrared, ultraviolet) and detects each separately using dedicated sensors. This segmentation allows the system to analyze the spectral composition of incident light and distinguish between warm light (natural light reflected from human face) and cold light (light from display screens), thereby preventing photo attacks while maintaining system reliability.
Solution Approach 2:
The patent employs a camera module that serves multiple functions: it captures visible light images for face recognition and simultaneously detects infrared and ultraviolet wavebands for liveness authentication. This multi-functionality reduces overall system complexity by integrating multiple detection capabilities into a single device rather than requiring separate dedicated sensors for each waveband.
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 detects and prevents attacks from synthetic displays by identifying the presence of invisible light components in ambient light, enhancing the security of face recognition systems against photo and flash attacks.
Implementation Method 1
a filtering device configured to filter out light of a first predetermined waveband from incident light incident to a first area
Implementation Method 2
an optical-electrical converting device configured to convert the filtered light from the filtering device into a first electrical signal
Data Source
AI summary
Provided are a liveness authentication method and apparatus. The method comprises: filtering out light of a first predetermined waveband from incident light incident to a first area; detecting an intensity of the filtered light as a first light intensity; and determining that a liveness authentication is failed in the case that the first light intensity satisfies a predetermined condition. The liveness authentication is performed through determining whether the incident light is cold light. In the case that the incident light is light emitted from a liquid crystal display screen, it is determined that the incident light is the cold light, and it is determined that the liveness authentication is failed.