Neural Network Authentication Using PPG Blood Vessel Pulse Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing biometric authentication systems can be fooled by forgeries and are affected by environmental obstacles, which compromises their accuracy in recognizing unique personal characteristics.
Innovation Solution
An authentication system utilizing a photoplethysmogram (PPG) sensor and a neural network architecture, specifically a double convolutional neural network (CNN), to generate a result value indicating whether the identification of a subject passes an authentication test by comparing sensed PPG signals with reference signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional biometric authentication methods are used, then the system is easy to implement, but the accuracy is compromised due to forgeries and environmental obstacles
Solution Approach 1:
The patent replaces traditional mechanical/optical biometric recognition systems with a photoplethysmogram-based physiological signal analysis system. Instead of using cameras or scanners that are susceptible to environmental interference and forgeries, the system uses PPG sensors to detect blood vessel pulse waves, which provide more reliable and difficult-to-fake physiological data for authentication.
Solution Approach 2:
The patent changes the authentication parameter from external physical characteristics (fingerprint, face) to internal physiological parameters (blood vessel pulse wave characteristics). This parameter change makes the authentication more resistant to environmental obstacles and forgeries, as internal physiological signals are harder to replicate and less affected by external conditions.
2Measurement precision
If feature detection and statistical analysis are used for PPG signal processing, then the system provides interpretability, but the accuracy decreases due to information loss
Solution Approach 1:
The patent extracts only the essential authentication result from the complex PPG signal analysis, using a neural network to directly output whether the signals match without displaying intermediate feature extractions or statistical analyses. This extraction approach maintains high accuracy while reducing the complexity of the processing pipeline by removing unnecessary intermediate steps.
Solution Approach 2:
The patent replaces traditional mechanical signal processing methods (feature detection, statistical analysis) with a neural network-based automated processing system. The neural network automatically learns and extracts relevant features from raw PPG signals without manual feature engineering, thereby preserving more information and achieving higher authentication precision.
3Reliability
If a double convolutional neural network is used, then the authentication accuracy increases, but the computational complexity increases
Solution Approach 1:
The patent divides the authentication process into two separate sub-networks within the double convolutional neural network, each processing one PPG signal independently. This segmentation allows the system to process signals in parallel, reducing the overall computational energy consumption while maintaining high authentication reliability through the comparison of results from both sub-networks.
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
The system achieves high accuracy in authenticating subjects, with approximately 94.3% accuracy in determining a valid identification and 97.8% accuracy in rejecting invalid identifications, enhancing security by directly using raw PPG signals without feature detection or statistical analysis.
Implementation Method 1
a photoplethysmogram (PPG) sensor configured to sense pulses of a blood vessel of the subject to generate a sensed PPG signal of the subject
Data Source
AI summary
An authentication system for authenticating an identification of a subject is provided. The authentication system includes a photoplethysmogram (PPG) sensor, a storage device, and a processor. The PPG sensor is configured to sense pulses of a blood vessel of the subject to generate a sensed PPG signal of the subject. The storage device stores an authentication model. The processor is configured to load the authentication model from the storage device and input the sensed PPG signal and a reference PPG signal into the authentication model to generate a result value which indicates whether the identification of the subject passes an authentication test.