Single-lead electrocardiosignal f-wave extraction method

Abstract

The invention discloses a single-lead electrocardiosignal f-wave extraction method. The extraction method includes the following steps: establishing a two-channel timing sequence convolutional neuralnetwork model, wherein the model includes two encoding submodules, two mask estimation submodules, an information fusion module and two encoding submodules; and constructing a training data set of ECGsignals, training the two-channel timing sequence convolutional neural network model, and saving the trained model for the extraction of f-wave time domain signals in unknown mixed electrocardiosignals. QRST wave groups and f-wave are respectively coded and decoded by the provided two-channel timing sequence convolutional neural network; the mask estimation of information fusion is adopted; regularization terms are constructed; and therefore, the distribution difference of QRST component coding characteristics can be restrained, the distortion of the QRST wave groups can be reduced, the potential characteristics of the QRST wave groups and f-wave with high signal-to-noise ratio can be selected, so that the detection precision of the f-wave can be enhanced.

Description

technical field [0001] The invention relates to the field of signal processing, in particular to a single-lead electrocardiographic signal f-wave extraction method. Background technique [0002] Atrial fibrillation is the most common supraventricular arrhythmia. People with atrial fibrillation have a stroke risk four to five times higher than those who are not affected by atrial fibrillation. Among heart disease patients, those with atrial fibrillation were twice as likely to die. When atrial fibrillation occurs, the P wave is replaced by an F wave that vibrates rapidly (fibrillation) and varies in amplitude and shape. At the same time, this f wave causes irregular and frequently changing ventricular electrical activity. The physical parameters of f wave (such as amplitude and frequency spectrum) are often used in the clinical diagnosis of atrial fibrillation, which is of great significance to the study of the mechanism of atrial fibrillation electrophysiological activity...

AI Technical Summary

Problems solved by technology

Although the single-lead method is convenient for mobile ECG monitoring, it is sensitive to abnormal heart beats (such as changes in the shape of heart beats caused by premature ventricular contractions), and the detection accuracy is low

[0004] The traditional multi-lead f-wave extraction method uses the correlation between the ECG signals of each lead to suppress the QRST component; this type of method needs to affix multiple electrodes on the patient, which limits the patient's movement and is not conducive to long-term monitoring and monitoring. Dynamic detection; and the single-lead f-wave extraction method is based on the quasi-periodicity of the ECG signal, constructing a QRST template for matching filtering to achieve f-wave extraction

However, QRST waveform distortion and cardiac beat interval abnormalities seriously affect the accuracy of QRST template estimation, thereby affecting the accuracy of f-wave extraction

Images