Non-invasive intracardiac hyperexcitability point positioning method based on stack type auto-encoder

Abstract

The invention discloses a non-invasive intracardiac hyperexcitability point positioning method based on a stack type auto-encoder. The method mainly includes a training stage and a prediction stage. At the training stage, acquired QRS data serves as input, three-dimensional coordinates of corresponding mapping points of the QRS data serve as labels to train the auto-encoder, input data is preprocessed through the denoising auto-encoder, in the process of learning high-dimensional information, an upper layer of all variable factors is utilized, the result is more precise, and finally the stack type auto-encoder is established. At the prediction stage, an input data set x in a test set is input in a trained stack type model, and the final coordinate prediction result (x-y-z) is obtained on the output layer of the model. Position prediction of VT hyperexcitability points is achieved on the aspect of data drive, and the problems of time and labor waste of clinical catheter ablation is effectively solved.

Description

technical field [0001] The invention belongs to the technical field of cardiac electrophysiological analysis, and in particular relates to a method for locating non-invasive intracardiac abnormal excitation points based on a stacked autoencoder. Background technique [0002] Electrocardiography (Electrocardiography) has become a very important method for non-invasively revealing cardiac electrophysiological activities by recording the potential changes caused by the depolarization and repolarization process during the cardiac beating cycle on the body surface. If the heart is regarded as the source of the electric field, then the corresponding electric field is distributed in the space from the heart to the body surface, and the potentials at various positions can be recorded on the body surface. [0003] Ventricular tachycardia (VT) refers to the rapid arrhythmia that occurs in the bundle branches, myocardial conduction fibers, and ventricular myocardium below the His bundl...

AI Technical Summary

Problems solved by technology

[0004] At present, the clinical diagnosis of ventricular tachycardia is mainly based on 12-lead electrocardiogram. However, this method can only make a preliminary diagnosis of VT, and cannot provide more detailed information such as the specific location of tachycardia.

On the other hand, in the ablation of VT, the surgeon also uses invasive means to directly measure the electrophysiological activity of the target location of the heart to locate the abnormal activation point of VT. However, the efficiency of invasive pacing mapping method is not high. Time-consuming and labor-intensive, but also has certain risks

Images