A Human Identification Method Based on ECG’s Frequency Band Features

Objective: As a physiological signal in vivo,ECG(electrocardiogram) can be easily extracted from fingers.It is easy to collect but difficult to replicate.Therefore,the paper proposed a human identification method based on frequency band features of single lead ECG.Methods: Firstly,decomposition of the single-cycle ECG signal is proposed by using wavelet packet and the waveform and energy of each sub-band is extracted.It uses the sub-band energy,the ECG waveform,the sub-band waveform as classification features;Then,the DTW algorithm is introduced to determine the optimal matching distance of ECG waveforms and sub-band waveforms between test data and the template.The research extracts the energy ratio and difference of each sub-band between test data and the template at the same time.It sets suitable thresholds for above parameters for human identification.Results: The same frequency sub-band energy,ECG waveform,sub-band waveform differences are small when test data and the template are from same sample.The sub-band energy’s ratio is close to one.The differences of same frequency sub-band energy value and waveform differences are close to zero.Conversely,the parameter differences are bigger than former and it provides good classification features for ECG.Conclusions: The price of ECG identification device is low,and the ECG signal is difficult to steal.It is a relatively safe means of identification.Experimental results show that the proposed method has higher recognition rate than traditional algorithm.