Cardioelectric characteristic extracting process based on evolutive wavelet wiener deconvolution

Abstract

The invention relates to a feature extraction method of electrocardiogram data, comprising the following steps: 1. preprocessing of electrocardiographic data; sub-band, and then apply the evolutionary Wiener deconvolution method to the sub-band signal to extract the feature point position of the QRS complex. ③ Feature extraction of P wave and T wave: According to the feature point position information of the QRS complex obtained in the first step, the time period where the QRS complex is located is replaced by the baseline. Then, wavelet transform is used to extract the sub-bands of P wave and T wave from the ECG signal after removing the QRS complex, and finally the evolutionary Wiener deconvolution method is applied to the obtained sub-bands until the accurate P wave, T wave are extracted. T wave feature point location. By adopting the present invention, insignificant key position points in the original ECG data can be highlighted and transformed into obvious peak points or valley points, thus providing a powerful basis for feature detection.

Description

technical field [0001] The invention relates to a feature extraction method of electrocardiographic data, which belongs to the technical field of biomedical engineering. Background technique [0002] Feature extraction of ECG signals is one of the most powerful means for heart disease detection. There are several algorithms commonly used for feature extraction of ECG signals: length transform and energy transform, hidden Markov model, artificial neural network and wavelet transform. However, the above methods have certain defects: the length transformation and energy transformation cannot accurately detect the characteristic information of the abnormal QRS complex; The method also lacks effectiveness in the detection of abnormal ECG waveforms; the wavelet transform method is superior to the above-mentioned methods in the detection of pathological ECG signals, but the detection of unsteady changes in ECG waveforms, especially in the case of low signal-to-noise ratio, is diff...

AI Technical Summary

Problems solved by technology

[0004] The gist of the present invention is to propose an efficient and accurate feature extraction method for ECG data in order to overcome the problem of remote ECG data feature extraction in the remote ECG monitoring where multiple patients need to be monitored online at the same time and to detect heart disease in a timely and accurate manner. Deficiencies in cardiac disease detection in electrical monitoring

Images