Normalized Electrocardiogram Generation for Personal Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional personal identification methods using facial images, fingerprints, or iris images are vulnerable to security issues and difficult to implement in real-time due to data processing requirements, while electrocardiogram-based identification faces challenges in differentiating abnormal signals that affect accuracy.
Innovation Solution
A method for generating a normalized electrocardiogram by extracting and connecting single-cycle electrocardiogram signals with a similarity degree equal to or higher than a critical similarity degree, filtered to remove noise and adjust amplitudes, to improve personal identification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional personal identification methods (facial image, fingerprint, iris image) are used, then personal authentication can be performed, but security vulnerabilities arise because these data can be reproduced or duplicated by absent or deceased persons
Solution Approach 1:
The patent changes the identification parameter from static biological data (facial image, fingerprint) to dynamic physiological signals (electrocardiogram). The ECG signal's temporal and amplitude characteristics provide a more secure identifier that cannot be easily reproduced or duplicated, resolving the security vulnerability of conventional methods.
2Reliability
If conventional personal identification methods are used, then authentication can be performed, but real-time identification is difficult due to large data processing requirements
Solution Approach 1:
The patent segments the continuous ECG signal into individual cardiac cycles (P-Q-R-S-T waves) for analysis. This segmentation reduces the data processing burden compared to analyzing complete facial images or fingerprint scans, enabling faster real-time authentication while maintaining identification capability.
3Reliability
If electrocardiogram data is used for personal identification, then security is improved because ECG cannot be reproduced by others, but identification accuracy deteriorates when abnormal heart signals are present
Solution Approach 1:
The patent extracts and removes abnormal segments from the ECG signal, isolating only the normal P-Q-R-S-T wave cycles suitable for identification. By extracting clean signal portions and excluding abnormal beats, the system maintains high identification accuracy even when the overall ECG contains irregularities.
Solution Approach 2:
The patent performs preliminary filtering and normalization of the ECG signal before identification processing. By pre-processing the signal to remove noise and standardize amplitude, the system prepares clean data that improves subsequent identification accuracy while preserving the security benefits of using ECG data.
4Measurement precision
If multiple single-cycle electrocardiograms are extracted and connected to generate normalized ECG, then identification accuracy is dramatically increased, but data processing complexity increases
Solution Approach 1:
The patent merges multiple extracted single-cycle ECG segments into a normalized composite signal. By combining multiple valid cardiac cycles and averaging their characteristics, the system enhances identification accuracy through more robust data while managing processing complexity through systematic integration.
Data Source
AI summary
Disclosed are a method for generating an electrocardiogram for personal identification and a method for identifying a person using the electrocardiogram. The electrocardiogram generation method generates a normalized electrocardiogram by extracting single-cycle electrocardiogram signals meaningful for personal identification from an electrocardiogram of a person and by connecting the extracted single-cycle electrocardiogram signals arranged in temporal order. Therefore, the electrocardiogram generation method dramatically increases identification accuracy in personal identification.