Electrocardiographic signal feature extraction and its instrument development based on continuous wavelet transform / 生物医学工程学杂志

ABSTRACT

This paper introduces a virtual recording and analysis instrumentation system for recording and identifying electrocardiographic (ECG) signals. The system developed is aimed at constructing a PC-based virtual instrumentation which enables to record. investigate and measure the ECG signal of 12 leads simultaneously and perfectly in order to improve the measuring precision of ECG. Based on the proper feature in time domain of Mexican hat wavelet expressed by positioning and analysis precision for QRS complex, the instrumentation system uses continuous wavelet transform(CWT) and uses the Mexican hat as the wavelet base to measure precisely the characteristic information and generate the precise characteristic parameters of ECG. The analysis of measured ECG signals in hospital demonstrated that even in the condition with serious noise interference, the method presented is still easily to describe the characteristics of ECG on line precisely which makes the instrumentation system valuable in practical application.