Biometric identification method for ECG based on wavelet transform and piecewise correction

摘要

In this paper, we proposed a novel biometric identification system which can extract more distinguishable features from the ECG signal. Based on the reference point detection, piecewise correction method was used to solve the problem of heart rate variability (HRV). Besides, we used some details of parts of the wavelet coefficient structure to reconstruct more distinguishable signal. At last, the feature dimension was reduced to 41 by PCA. 254 subjects' ECG records from the publicly available databases were selected to verify this method. The accuracies were 98.50% and 98.87% using the k-NN classifier and SVM classifier separately. The experiments demonstrate that our method is efficient in identification applications.