Feature learning method and system for ECG identity recognition

Abstract

The invention discloses a feature learning method and system for ECG identity recognition. The method comprises the steps of obtaining an electrocardiosignal of an individual; performing filtering processing on the electrocardiosignal of the individual; extracting morphological features and wavelet features of the filtered electrocardiosignal; on the basis of the morphological features and wavelet features of the filtered electrocardiosignal of the individual, building a sparse own-coding recognition neural network. The problem that when an electrocardiosignal is collected in real life, identity recognition precision is lowered due to electrocardio abnormal states such as individual body mental and emotional states (for example, arrhythmia and strenuous exercise) is can be solved.

Description

technical field [0001] The invention relates to a method and a system in the field of biological feature identification, in particular to a feature learning method and system for ECG identification. Background technique [0002] With the impact of the Internet boom and the rapid development of information technology, people pay more and more attention to information security and property security, and identification is the most important part of information security. Biometric identification generally refers to the identification of personal identity based on human physiological characteristics (such as fingerprints, iris and facial features, etc.) and behavioral characteristics (such as blinking, nodding and shaking heads, etc.). Since the physiological characteristics of human beings are unique, biometric identification is more secure and reliable than traditional identity authentication methods. Traditional identity authentication methods are more likely to be lost, forg...

AI Technical Summary

Problems solved by technology

[0006] The present invention combines the actual application scene of ECG identity recognition, and the heart rate, physical health status and emotional state of each individual are not limited during the collection process, and the combined features of morphology and wavelet are used as the initial features of the system, although the accuracy of identity recognition is improved. Accuracy, but at the same time, it also caused a sharp increase in the feature dimension and introduced too much feature redundancy, resulting in high computational complexi

Images