Liveness Detection via Randomized Physiological Stimuli
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current liveness detection methods in network security, particularly biometric systems, are vulnerable to spoofing attacks as they rely on machine learning techniques that are resource-intensive and can be fooled by fake biometric samples, lacking robustness in distinguishing real from pre-captured or synthesized data.
Innovation Solution
A liveness detection system that uses randomized visual or audible outputs to elicit physiological responses from human features like the eyes or mouth, comparing these reactions with expected responses to determine if an entity is living, employing techniques such as pupillary response analysis, eye tracking, and lip reading without requiring machine learning for pattern recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning techniques are used for liveness detection, then the system can analyze biometric data patterns, but the processing resources required become excessively high and the system becomes vulnerable to spoofing attacks
Solution Approach 1:
The patent extracts and isolates specific physiological response characteristics (pupillary response, blink patterns, eye movements) from complex biometric data, focusing only on the most reliable liveness indicators rather than processing entire biometric datasets through resource-intensive machine learning models
Solution Approach 2:
Instead of using complex models to detect liveness, the patent inverts the approach by using simple, well-understood physiological responses that are inherently difficult to spoof. The system leverages the fact that living beings naturally exhibit specific physiological reactions to stimuli, rather than trying to prove liveness through complex pattern recognition
2Reliability
If machine learning techniques are used for liveness detection, then the system can recognize biometric patterns, but the device complexity and implementation cost increase significantly
Solution Approach 1:
The system uses the entity's own physiological responses to itself (how its eyes naturally react to light, how its pupils constrict and dilate) rather than requiring external complex analysis systems. The physiological response serves as its own verification mechanism
Solution Approach 2:
The patent changes the detection parameters from complex biometric feature sets to specific, measurable physiological response parameters such as pupillary constriction rate, blink frequency, and eye movement patterns, which can be captured and analyzed with simpler systems
3Object-affected harmful factors
If pre-captured or synthesized biometric samples are presented, then unauthorized access can be gained, but physiological response analysis can distinguish these from real living beings
Solution Approach 1:
The system presents a stimulus (such as a light source or visual target) before capturing the response, establishing a cause-effect relationship that pre-captured samples cannot replicate. The timing and sequence of stimulus-response are built into the verification process in advance
Solution Approach 2:
The system creates a closed-loop feedback mechanism where a stimulus is presented and the resulting physiological response is immediately captured and analyzed. This real-time feedback loop ensures that only living beings that can process and respond to the stimulus can be verified
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
This approach effectively differentiates between living beings and non-living entities by analyzing predictable physiological reactions, enhancing the security of biometric systems and reducing the risk of spoofing attacks without the need for resource-intensive machine learning models.
Implementation Method 1
a camera to record a moving image of the entity over the interval of time
Data Source
Figure 1
Figure 2A
Figure 2B~2C
AI summary
A computer-implemented liveness detection method comprising implementing, by a liveness detection system, the following steps: selecting at random a first set of one or more parameters of a first liveness test; transmitting, to a user device available to an entity, the first parameter set, thereby causing the user device to perform the first liveness test according to the first parameter set; receiving from the user device results of the first liveness test; receiving results of a second liveness test pertaining to the entity; detecting whether a timeout condition has occurred, the timeout condition caused by an unacceptable delay in receiving the results relative to a timing of the transmitting step; and if the timeout condition occurs, refusing the entity access to a remote computer system, otherwise determining whether the entity is a living being using the results of the liveness tests.