In this paper, new biometric method using electrocardiogram (ECG) signal is proposed. In contrast to conventional lead Ⅱ type ECG based biometrics, to enhance the consistency of biometric data, lead Ⅲ type signals are selected. For characterizatio...
In this paper, new biometric method using electrocardiogram (ECG) signal is proposed. In contrast to conventional lead Ⅱ type ECG based biometrics, to enhance the consistency of biometric data, lead Ⅲ type signals are selected. For characterization, twenty-two different features are defined and feed-forward error back-propagation neural networks (BPNN) are applied for classification. To consider five kinds of accuracy evaluation methods, unknown two extra peoples are tested together with twelve biometrics candidates. Experimental results showed 100% of accuracy for identification of pre-defined candidates and 97.77% of overall identification accuracy.