The Research Of Person Identification Using Electrocardiograms

With the advent of information network era, identity theft as a new words appeared in public view. Identity theft occurs when someone uses your personally identifying information, like your social security identification number, driver’s license number or credit card number, without your permission. Identity theft is a crime. Identity theft is serious. Victims spend hundreds of dollars and many days repairing damage to their good name and credit record. Some consumers victimized by identity theft may lose out on job opportunities, or be denied loans for education, housing because of negative information on their credit reports. In rare cases, they may even be arrested for crimes they did not commit. In view of this, the society needs an effective security mechanism to combat identity theft crime. The methods of traditional biologic character identification such as fingerprint, iris, DNA, voice and soon, now is also facing the security problem of imitation and falsification. Therefore, this paper presents the implementation of a new biometric identification technology-ECG signals (Electrocardiogram, ECG), as a kind of feasible methods for solving the security problem.This paper mainly studies the technology and method for identity recognition through ECG signal.On the basis of analysis of ECG signal and noise characteristics,the thesis uses butterworth digital filter on ECG signal denoising preprocessing.The autocorrelation transform method (Autocorrelation, AC) is used for feature extraction of ECG signals and the classification performance of the AC coefficient of different individuals are evaluated and compared. Finally the echo state networks (Echo State Network, ESN) is introducted for training on part of the feature vectors, then the test set was tested using the trained echo state network, and comparing with the result of using linear discriminant analysis(LDA) and K nearest neighbor algorithm for classification.The work of this paper are as follows:(1) The principles and characteristics of the ECG signal is summarized. The feasibility analysis that ECG signal can be used for identity recognition is made to lay the foundation of the application of identity recognition based on ECG signal.(2) The present research situation at home and abroad of the identification recognition based on ECG signal are analyzed and summarized in the paper.Have the deep leading meaning and draw lessons from the value to the paper.(3) A preprocess method is proposed,including ECG signal denoising and the standardization of ECG signal. First of all, butterworth digital filter is used to eliminate noise and baseline drift in the ECG signal. Then the ECG signal is resampled and the amplitude gain of ECG signal is unified.(4) This paper presents the ECG signal recognition algorithm based on ESN. The earliest ESN method used for the prediction of time series, here we introduce the ESN method for the identification of ECG signals for the first time. Finally, the ECG data in the public database and automatically collected is used to train the ESN network, and the results were analyzed.The studies of this paper provide a novel way with good capability to prevent counterfeiting,high accuracy of recognition rate, strong robustness,and which laid a solid theoretical foundation and technical support for the development of ECG identification...