CN-122019978-A - Electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation

CN122019978ACN 122019978 ACN122019978 ACN 122019978ACN-122019978-A

Abstract

The invention discloses an electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation, which solves the problems that electrocardiosignals are easily affected by myoelectric interference, power frequency interference, baseline drift and other noise in the acquisition process of the electrocardiosignals, noise interference of different degrees can be caused between different leads, so that the signal to noise ratio is reduced, characteristics are fuzzy, further, the problems of inaccurate characteristic extraction, low recognition precision, weak generalization capability and the like occur in deep learning model training and reasoning, and the like are solved by constructing a 12-lead independent visualization system and combining a single-lead/multi-lead/full-lead differential pretreatment mechanism and a median filtering layering denoising technology, so that the flexible and fine regulation and control of filtering parameters are supported, and the functions of cutting, resetting, multi-format storage, file skip and full-flow log tracing are matched, so that the accurate noise suppression and the complete retention of electrocardiosignal key characteristics are realized, and the signal input with high quality and high reliability is provided for deep learning.

Inventors

WANG HAIYAN
CHEN JIANFENG
LI LINGLING
WANG XUAN
LIANG SHAOHUA
LI WEICHAO
ZHANG PENGFEI
LI SHAOLIN
TAN JIAN

Assignees

郑州航空工业管理学院

Dates

Publication Date: 20260512
Application Date: 20260202

Claims (10)

1. The electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation is characterized by comprising a visual information display module, a lead cutting and adjusting module, a differential preprocessing module and a file operation module, wherein the visual information display module is used for visually displaying electrocardiosignals in a lead way, the lead cutting and adjusting module is used for cutting and adjusting the electrocardiosignals in the visual information display module, the differential preprocessing module is used for carrying out differential processing on different leads on the electrocardiosignals cut and adjusted by the lead cutting and adjusting module and reducing noise, and the file management module is used for switching, storing and resetting electrocardiosignal files in a recording system and using log records to process the electrocardiosignal files.
2. The electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation and control as claimed in claim 1, wherein the visual information display module comprises a Matplotlib-based 12-lead independent drawing system and an information display module; The 12-lead independent drawing system uses a sub-graph layout of 12 rows and 1 column to respectively correspond to I, II, III, aVR, aVL, aVF, V-V6 leads, allocates independent drawing areas for each lead and configures a proprietary X/Y axis coordinate system; the information display module comprises a hierarchical analysis and rendering module, integrates three types of core information of file names, patient diagnosis information and electrocardiosignals, realizes event-driven dynamic update, analyzes and displays the current file names, processing progress and patient clinical information dimension in real time, accurately extracts patient information from HEA labeling files, completes Chinese-English mapping conversion of diagnosis names, automatically calculates and displays signal total duration obtained based on sampling points and sampling rates, and is used for solving the problem that the current file names, the processing progress and the patient clinical information dimension are not consistent.
3. The electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation and control as claimed in claim 2, wherein an X axis in the X/Y axis coordinate system is unified as a time axis, and a Y axis is a signal amplitude.
4. The system for processing the electrocardiosignal based on multi-scale analysis and parameter regulation and control according to claim 2, wherein the lead clipping and adjusting module comprises a lead clipping module and a ordinate self-adaptive adjusting module; the lead clipping module accurately marks clipping time periods through a red-blue double-color semitransparent mask; The ordinate adjustment module adopts a real-time self-adaptive mechanism, calculates the maximum value and the minimum value of the signal in each lead clipping region in real time by means of a multi-segment signal extremum fusion algorithm, automatically adds 10% of margin, and immediately refreshes the ordinate range of the corresponding lead when clipping time and filtering parameters are adjusted.
5. The electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation and control as claimed in claim 2, wherein the differential preprocessing module adopts median filtering layering application and comprises a denoising layer and a baseline drift removal layer; the denoising layer generates an odd kernel size filter based on a 0-0.1 s designated window; the baseline drift removal layer is subjected to double-window cascade median filtering.
6. The electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation and control as claimed in claim 5, wherein the differential preprocessing module supports two modes of full-guide and single-guide, parameters are not interfered with each other, the full-guide is used for uniformly regulating all leads, the single-guide is used for regulating only selected leads, current lead parameter configuration is automatically stored when the leads are switched, and history parameters are loaded when the leads are switched back.
7. The electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation and control as claimed in claim 5, wherein the window 1 in the baseline drift removal layer is 0.1-0.5 s for roughly estimating a baseline, the window 2 is 0.3-1.0 s for finely correcting, and the window can be adjusted for different lead drift characteristics.
8. The electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation and control as claimed in claim 5, wherein the denoising function and the baseline drift removal function in the lead clipping and adjusting module are respectively provided with independent check boxes and can be independently switched, three processing modes of denoising only and baseline removing only are met, and the filtering switch state can be independently configured for any single lead or all leads in 12 leads.
9. The system of claim 1 wherein the file management module comprises a reset save module for saving a reset electrocardiosignal file, a file skip module for file navigation operations, and a log record module for creating a full-flow operation log record system.
10. The system of claim 9, wherein the reset preservation module has a multi-format synchronous processing function.

Description

Electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation Technical Field The invention belongs to the technical field of biological signal processing and medical signal analysis, and relates to an electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation. Background Cardiovascular disease, the leading cause of death and disability worldwide, has become a major public health problem that severely threatens human health. Along with the acceleration of the aging process of population and the change of life style, the incidence rate and the death rate of cardiovascular diseases such as hypertension, coronary heart disease, arrhythmia and the like continuously rise, and the data according to the national health statistics annual survey 2023 show that the death rate of the cardiovascular diseases of urban and rural residents in partial areas of China is the first in 2021, so that the early and timely intervention and the accurate treatment of lesions are particularly critical. Electrocardiogram (ECG) is a core routine for arrhythmia detection by virtue of real-time, noninvasive, low cost. The system can intuitively reflect various pathophysiological states of heart such as rhythm, conduction, myocardial ischemia, myocardial hypertrophy and the like by recording bioelectric changes generated when the heart is excited by a pacing point, an atrium and a ventricle in each cardiac cycle, and is one of the most basic and widely applied examination means in clinical cardiovascular disease diagnosis. In recent years, the deep learning technology has great potential in the field of arrhythmia intelligent detection by virtue of strong characteristic automatic extraction and pattern recognition capability, and provides a new path for solving pain points such as low efficiency, strong subjectivity, high missed diagnosis error rate and the like of clinical manual analysis of ECG. The arrhythmia detection model based on deep learning can realize rapid identification and classification of abnormal rhythms such as atrial fibrillation, ventricular premature beat and myocardial infarction by a large number of labeled ECG data training, and is hopeful to become an important auxiliary tool for screening basic medical institutions and large-scale crowds. However, the detection precision and generalization capability of the deep learning model are highly dependent on the quality of input ECG signals, the ECG signals are easily affected by myoelectric interference, power frequency interference, baseline drift and other noise in the current clinic, and the signal characteristics of different leads and different individuals have obvious differences, so that the signal-to-noise ratio of the original signals is reduced, key pathological waveforms are fuzzy, the limitations of the existing preprocessing technology further restrict the model performance, a static scale is adopted in the visualization link, waveform compression or truncation is easily caused by signal amplitude variation, details cannot be completely reserved, a 'unification' fixed filtering mode is adopted in preprocessing, the differential characteristics of the signals are difficult to adapt, the parameter regulation and the data safety are difficult to be compatible, the functions of real-time preview and processing contrast are lacked, and the problems all lead to inaccurate feature extraction and insufficient abnormal detection rate of the deep learning model, so that the reliability and the clinical conversion effect of intelligent diagnosis are seriously affected. The system provides a comprehensive solution integrating dynamic self-adaptive visualization, differential filtering and safe and efficient file management, realizes deep fusion of signal preprocessing and interactive design, realizes ordinate self-adaptive adjustment by combining a multi-section extremum fusion algorithm, ensures complete presentation of signal details, and meanwhile, a multi-scale analysis system constructs an independent lead parameter storage space, supports single-lead/multi-lead/full-lead differential filtering, adopts a layered processing architecture, dynamically generates a filter through a noise reduction window, performs double-window cascade correction of a baseline drift layer, realizes fine regulation and control of one-lead one parameter, provides high-quality and high-reliability signal input for a deep learning model, further improves the accuracy and generalization capability of arrhythmia intelligent detection, provides powerful support for accurate diagnosis and research of cardiovascular diseases, and has obvious clinical and scientific research application values. Disclosure of Invention Aiming at the problems, the invention provides an electrocardiosignal interactive processing system based on multi-scale analysis and parameter regulation, which well solves the problems in th