A wavelet method for biometric identification using wearable ECG sensors

This paper reports a new signal processing method for biometric identification purposes that utilizes signals collected from a cost-effective wearable electro-cardiogram (ECG) sensor. In this method, raw ECG signals were first prepared in the time domain and then decomposed into a structure of coefficients using wavelet algorithms. This coefficient structure was further extended to a coefficient matrix, whose principle components were used as discriminant quantities to identify individual subjects. In our research, 47 datasets collected from 20 subjects were used to test the method. Results demonstrate that this identification method is effective.