WO2021169296A1 - 心电图数据处理的方法、装置、计算机设备及存储介质 - Google Patents

心电图数据处理的方法、装置、计算机设备及存储介质 Download PDF

Info

Publication number
WO2021169296A1
WO2021169296A1 PCT/CN2020/119053 CN2020119053W WO2021169296A1 WO 2021169296 A1 WO2021169296 A1 WO 2021169296A1 CN 2020119053 W CN2020119053 W CN 2020119053W WO 2021169296 A1 WO2021169296 A1 WO 2021169296A1
Authority
WO
WIPO (PCT)
Prior art keywords
wave
data information
waveform
information
processing
Prior art date
Application number
PCT/CN2020/119053
Other languages
English (en)
French (fr)
Inventor
绳立淼
李响
贾文笑
康延妮
高群群
Original Assignee
平安科技(深圳)有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技(深圳)有限公司 filed Critical 平安科技(深圳)有限公司
Publication of WO2021169296A1 publication Critical patent/WO2021169296A1/zh

Links

Images

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/24Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316Modalities, i.e. specific diagnostic methods
    • A61B5/318Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/24Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316Modalities, i.e. specific diagnostic methods
    • A61B5/318Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/346Analysis of electrocardiograms
    • A61B5/349Detecting specific parameters of the electrocardiograph cycle
    • A61B5/366Detecting abnormal QRS complex, e.g. widening
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/72Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/72Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7225Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation

Definitions

  • This application relates to the field of data processing, and in particular to a method, device, computer equipment, and storage medium for processing electrocardiogram data.
  • the existing electrocardiogram automatic diagnosis technology is to perform information labeling and feature extraction on the P wave, QRS wave and T wave of each heartbeat of the electrocardiogram.
  • the inventor realized that the accuracy of the detection result will be caused by the influence of baseline wandering noise, myoelectric noise, power frequency interference and other sudden noises. Compared with the apex of the R wave, it is greatly reduced.
  • a method for processing electrocardiogram data includes:
  • An electrocardiogram data processing device including:
  • the acquiring unit is used to acquire all heartbeat data information of the sample electrocardiogram
  • a processing unit configured to perform mean wave processing on the heartbeat data information according to a preset clustering and merging algorithm, and determine the processing result as the mean waveform of the sample electrocardiogram;
  • the labeling unit is used to label the P wave, QRS wave, T wave and ST segment information on the average waveform;
  • the output unit is used to output the labeled P wave, QRS wave, T wave and ST segment waveform information.
  • a storage medium storing at least one executable instruction, and the execution instruction causes a processor to perform the following steps:
  • the average waveform is labeled with P wave, QRS wave, T wave, and ST segment information, and the labeled P wave, QRS wave, T wave, and ST segment waveform information are output.
  • a computer device includes a processor, a memory, a communication interface, and a communication bus.
  • the processor, the memory, and the communication interface communicate with each other through the communication bus, and the memory is used to store at least one executable Instructions, the executable instructions cause the processor to perform the following steps:
  • the average waveform is labeled with P wave, QRS wave, T wave, and ST segment information, and the labeled P wave, QRS wave, T wave, and ST segment waveform information are output.
  • This application can synthesize the mean waveform through clustering, smoothing and filtering, effectively weakening the influence of baseline wandering noise, electromyographic noise, power frequency interference and sudden noise, and greatly improving the P wave, QRS wave and T wave The accuracy of labeling the start and end points in order to improve the accuracy of the ECG data processing, thereby improving the efficiency of using the data obtained from the ECG to analyze.
  • Fig. 1 shows a flow chart of a method for processing electrocardiogram data provided by an embodiment of the present application
  • Figure 2 shows a schematic structural diagram of an electrocardiogram data processing device provided by an embodiment of the present application
  • FIG. 3 shows a schematic structural diagram of another electrocardiogram data processing device provided by an embodiment of the present application.
  • FIG. 4 shows a schematic diagram of the physical structure of a computer device provided by an embodiment of the present application
  • the existing electrocardiogram automatic diagnosis technology is to perform information labeling and feature extraction on the P wave, QRS wave and T wave of each heartbeat of the electrocardiogram.
  • the accuracy of the detection result will be more than R
  • the wave apex is greatly reduced compared to that.
  • an embodiment of the present application provides a method for processing electrocardiogram data. As shown in FIG. 1, the method includes:
  • the heartbeat data information may include waveform data information in the sample electrocardiogram.
  • the waveform data information in the sample electrocardiogram can be clustered and merged by the preset clustering and merging algorithm to obtain the lead electrocardiogram
  • the average waveform can reflect the common features of most waveforms, thereby improving the accuracy of marking the start and end points of P wave, QRS wave, and T wave.
  • the specific process of performing mean wave processing on the heartbeat data information according to the preset clustering and merging algorithm may include: sequentially performing R wave detection, screening, and processing on all heartbeat data information of the obtained sample electrocardiogram. Clustering and merging, the obtained waveform can be determined as the mean waveform of the sample electrocardiogram.
  • the average waveform obtained by aggregating heartbeats can reflect the common characteristics of most waveforms.
  • the ECG signal is relatively weak, it is susceptible to interference from power frequency interference, baseline drift, etc., during acquisition.
  • the accuracy of marking the start and end points of P wave, QRS wave, and T wave, and the marked position will also fluctuate with interference, and the average waveform obtained by aggregating heartbeats can reduce interference, thereby improving the start and end of P wave, QRS wave, and T wave. Accuracy of point labeling.
  • the marked waveform information is output.
  • the embodiment of the present application provides several optional embodiments, but is not limited thereto, and the details are as follows:
  • the step 102 may specifically include: performing noise processing on the heartbeat data information, and performing R wave detection on the processed heartbeat data information to obtain R wave data information;
  • the R wave data information and a preset screening algorithm are used to screen the heartbeat data information; the screening results are clustered and merged, and the result obtained by the clustering and merging is determined as the sample electrocardiogram The mean waveform of.
  • the performing noise processing on the heartbeat data information may specifically include: performing noise processing on the sample electrocardiogram data by band-pass filtering.
  • band-pass filtering can filter baseline wandering noise, myoelectric noise, power frequency interference, and sudden noise, so as to improve the accuracy of marking the start and end points of P wave, QRS wave, and T wave.
  • a band-pass filter with a low-frequency cut-off frequency of 14 Hz and a high-frequency cut-off frequency of 20 Hz may be used for filtering during R wave detection.
  • a second-order median filter can be used for baseline drift, which can effectively suppress baseline drift after filtering, and there is no significant change in ST segment after filtering.
  • An adaptive filter can also be used to filter out 50Hz power frequency interference in the ECG signal.
  • the screening of the heartbeat data information according to the R wave data information and a preset screening algorithm may specifically include: according to the acquired RR interval data information of all heartbeats, Calculate the average RR interval data information; configure a screening interval according to the average RR interval data information, and screen heartbeat data information whose RR interval data information is within the screening interval.
  • the RR interval may be the position difference between the apex of the current R wave and the previous R wave, and the heart rate is inversely proportional to the RR interval.
  • the specific conversion formula may be:
  • the average RR interval may be the average of the RR intervals of all heartbeats, and the screening interval of the average RR interval may be configured.
  • an average RR interval of 0.8 to 1.2 times may be selected as the screening Interval.
  • the aggregation of heartbeats to obtain the mean waveform is to merge most heartbeats with the same attribute, but for occasional premature beats, especially ventricular premature beats, because the pace point is different from other heartbeats, that is, the attributes of the heartbeat are different. It may bring some differences in the shape of the heartbeat. If it participates in the superimposed aggregation, it will affect the accuracy of the aggregated heartbeat waveform. Therefore, the occasional premature beat does not participate in the superimposed aggregation.
  • the clustering and merging of the screening results, and determining the result of the clustering and merging as the mean waveform of the sample electrocardiogram may specifically include: performing the screening of the heartbeat data
  • the information configures sampling points, and performs superposition and averaging processing on all the sampling points; the processing result is determined as the average waveform of the sample electrocardiogram.
  • sampling points may be configured for the selected heartbeat data information, and the sampling points may be superimposed and averaged; the processing result may be determined as the average waveform of the sample electrocardiogram.
  • the overlap range on both sides of the R wave can be determined based on the average RR interval. For example, if the average RR interval is 1000ms, it can be determined that the apex of the R wave is the center point, and the sampling points in the range of R-500ms ⁇ R+500ms are superimposed and averaged.
  • the method may further include: performing band analysis on the average waveform to obtain the R wave notch, QRS wave shape, and QRS wave voltage change information of the average waveform.
  • the first half of the QRS wave of left bundle branch block is normal in shape and time, the second half of the time is prolonged, and the shape of the QRS wave is wide, flat or notched R wave, the ECG of the above left bundle branch block
  • the prior art pays little attention to the shape of the QRS wave.
  • the QRS wave in the right bundle branch block presents an "M" shape, and the above-mentioned shape is in The features marked and extracted by the prior art cannot be reflected either.
  • the notch, morphology, and voltage change process of QRS waves are recorded, so that the above-mentioned electrocardiogram features can play a certain role in AI prediction.
  • performing band analysis on the average waveform to obtain the R wave notch, QRS wave shape, and QRS wave voltage change information of the average waveform may specifically include: detecting the R wave according to the average waveform Determine whether there is a notch in the R wave according to the position information of the inflection point on both sides of the vertex; calculate the number of QRS wave zero-crossing points in the average waveform, and determine the shape of the QRS wave according to the number; The apex of the R wave in the average waveform is the center point, and the voltage change processes of the waveforms on both sides of the center point are recorded respectively.
  • calculating the number of QRS wave zero-crossing points in the average waveform, and determining the shape of the QRS wave according to the number may specifically include: calculating the number of QRS wave zero-crossing points in the average waveform, and according to the number , Determine the form of the QRS wave, the QRS wave can present the form of rsR', RSR' or rR'. Specifically, the number of zero-crossing points on the left and right sides of the R wave can be determined as the center. For example, if there is one zero-crossing point on the left side of the R-wave apex and three zero-crossing points on the right side of the R-wave apex, it can be judged that the QRS wave form is RSR'.
  • the method of recording the voltage change process of the QRS wave may include: taking the peak of the R wave in the average waveform as the center point, and respectively recording the voltage change process of the waveforms on both sides of the voltage change process.
  • the voltage change process may reflect the width of the second half of the QRS wave.
  • the characteristics of frustration Specifically, a point can be taken every Nms (for example, 10ms) with the apex of the R wave as the center. For example, the sampling rate is 500 Hz, the position of the R wave apex in the average waveform is point x, and the voltage of the average wave is y x .
  • the detecting position information of inflection points on both sides of the apex of the R wave according to the average waveform, and determining whether there is a notch in the R wave according to the inflection point position information may specifically include: performing a first order on the average waveform Differential processing, detecting the position information of the inflection point on both sides of the apex of the R wave; comparing the position information of the inflection point with the P wave, QRS wave, T wave, and ST segment start and end position information; judge the R wave according to the comparison result Whether there are notches.
  • the specific process of judging whether there is a notch on the R wave may include: performing a first-order difference processing on the average waveform, and detecting the inflection points on both sides of the apex of the R wave. For the voltage difference at point x+1, if the value of the first-order difference on the left and right sides of a certain point of the mean waveform is opposite, it can be judged that the point is the inflection point in the QRS wave; according to the position of the inflection point and the P wave, QRS wave, The comparison result of the start and end positions of the T wave can determine whether the R wave has a notch. For example, if there are two inflection points between the apex of the R wave and the end of the QRS wave, it means that the R wave has a notch.
  • the method may further include: when an electrocardiogram analysis processing request is received, according to the electrocardiogram data information carried in the request, And the pre-trained target state recognition model performs predictive analysis, and outputs the analysis result to determine the target state information.
  • the training sample data of the pre-trained target state recognition model is obtained by band analysis through mean wave processing.
  • an electrocardiogram analysis and processing request is received, and the request can carry the data information of the electrocardiogram; the predictive analysis can be performed according to the pre-trained target state recognition model, and the analysis result can be output, so that the corresponding analysis result can be corresponded to Determine target status information.
  • This application can synthesize the mean waveform through clustering, thereby smoothing and filtering, effectively weakening the influence of baseline wandering noise, electromyographic noise, power frequency interference and sudden noise, and greatly improving the P wave, QRS wave and T wave The accuracy rate of the start and end points.
  • the method may further include: acquiring sample data information of the sample electrocardiogram, the sample data information including P wave, QRS wave, T wave, ST Segment, R wave notch, QRS wave shape, QRS wave voltage change information, and RR interval data information of all heartbeats; according to the sample data information and the target state identification label corresponding to the sample electrocardiogram, the preset machine learning algorithm is trained, Determine the target state recognition model.
  • the sample data information such as P wave, QRS wave, T wave and ST segment, R wave notch, QRS wave shape, and RR interval information of all heartbeats in the QRS voltage change process of the sample electrocardiogram can be obtained, and all the data can be used.
  • the target state recognition label corresponding to the sample electrocardiogram trains a preset machine learning algorithm to determine the target state recognition model, and the target state recognition label corresponding to the sample electrocardiogram may be the difference between the recognition result of the sample electrocardiogram and the target state
  • the corresponding target state identification tag can be identified according to the corresponding target state information.
  • the recognition target state recognition model may be used to determine target state information, and the machine learning algorithm may include a deep learning algorithm, an SVM algorithm, and the like.
  • an embodiment of the present application provides an electrocardiogram data processing device. As shown in FIG. 2, the device includes: an acquisition unit 21, a processing unit 22, a labeling unit 23 and an output unit 24.
  • the acquiring unit 21 may be used to acquire all heartbeat data information of the sample electrocardiogram
  • the processing unit 22 may be configured to perform mean wave processing on the heartbeat data information according to a preset clustering and merging algorithm, and determine the processing result as the mean waveform of the sample electrocardiogram;
  • the labeling unit 23 may be used to label the average waveform with P wave, QRS wave, T wave, and ST segment information.
  • the output unit 24 can be used to output the labeled P wave, QRS wave, T wave and ST segment waveform information.
  • the processing unit 22 may include: a processing module 221, a detection module 222, a screening module 223, and a clustering and merging module 224, as shown in FIG. 3.
  • the processing module 221 may be used to perform noise processing on the heartbeat data information
  • the detection module 222 may be used to perform R wave detection on the processed heartbeat data information to obtain R wave data information;
  • the screening module 223 may be used to screen the heartbeat data information according to the R wave data information and a preset screening algorithm
  • the clustering and merging module 224 may be used for clustering and merging the screening results, and determining the result of the clustering and merging as the mean waveform of the sample electrocardiogram.
  • processing module 221 may be specifically used to perform noise processing on the sample electrocardiogram data by using band-pass filtering.
  • the screening module 223 may be specifically configured to calculate the average RR interval data information based on the RR interval data information of all the obtained heartbeats, configure the screening interval according to the average RR interval data information, and filter the RR interval.
  • Interval data information is the heartbeat data information within the screening interval.
  • the clustering and merging module 224 may be specifically used to configure sampling points for the selected heartbeat data information, perform superposition and averaging processing on all sampling points, and determine the result of the processing as the result of the sample electrocardiogram Mean waveform.
  • the device in order to obtain the R wave notch, QRS wave shape, and QRS wave voltage change information of the average waveform, the device may further include: a first analysis unit 25.
  • the first analysis unit 25 may be used to perform band analysis on the average waveform to obtain R wave notch, QRS wave shape, and QRS wave voltage change information of the average waveform.
  • the first analysis unit 25 may include: a judgment module 251, a determination module 252, and a recording module 253.
  • the judging module 251 can be used to detect the position information of the inflection point on both sides of the apex of the R wave according to the average waveform, and judge whether there is a notch in the R wave according to the information of the inflection point;
  • the determining module 252 may be used to calculate the number of zero crossing points of the QRS wave in the average waveform, and determine the shape of the QRS wave according to the number;
  • the recording module 253 may be used to record the voltage change process of the waveform on both sides of the center point with the apex of the R wave in the average waveform as the center point.
  • the judgment module 251 may be specifically used to perform first-order difference processing on the average waveform, detect the position information of the inflection point on both sides of the apex of the R wave, and compare the position information of the inflection point with the P wave, QRS wave, The position information of the start and end points of the T wave and the ST segment are compared, and according to the comparison result, it is judged whether there is a notch in the R wave.
  • the device in order to determine the target state information, may further include: a second analysis unit 26.
  • the second analysis unit 26 may be used to perform predictive analysis based on the ECG data information carried in the request and the pre-trained target state recognition model when an ECG analysis processing request is received, and output the analysis result to determine Target state information, the training sample data of the pre-trained target state recognition model is obtained by band analysis through mean wave processing.
  • the device in order to determine the target state recognition model, may further include: an acquisition unit 27 and a training unit 28.
  • the acquiring unit 26 may be used to acquire sample data information of the sample electrocardiogram, the sample data information including P wave, QRS wave, T wave, ST segment, R wave notch, QRS wave shape, QRS wave voltage change information And RR interval data information of all heartbeats;
  • the training unit 28 may be used to train a preset machine learning algorithm according to the sample data information and the target state recognition label corresponding to the sample electrocardiogram to determine the target state recognition model.
  • an embodiment of the present application also provides a storage medium.
  • the storage medium may include a high-speed RAM memory, may include a volatile memory, or may also include a non-volatile memory.
  • non-volatile memory for example, at least one disk memory, in which at least one executable instruction is stored in the storage medium, and the execution instruction causes the processor to perform the following steps: obtain all heartbeat data information of the sample electrocardiogram; The clustering and merging algorithm of the heartbeat data information is processed by the mean wave, and the processing result is determined as the mean waveform of the sample electrocardiogram; P wave, QRS wave, T wave, and ST wave are performed on the mean waveform Segment information labeling, output the labeled P wave, QRS wave, T wave and ST segment waveform information.
  • an embodiment of the present application also provides a computer device.
  • the processor 31, the communication interface 32, and the memory 33 communicate with each other through the communication bus 34.
  • the communication interface 34 is used to communicate with other devices such as network elements such as user terminals or other servers.
  • the processor 31 is configured to execute a program, and specifically can execute the relevant steps in the foregoing electrocardiogram data processing method embodiment.
  • the program may include program code, and the program code includes computer operation instructions.
  • the processor 31 may be a central processing unit CPU, or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.
  • ASIC Application Specific Integrated Circuit
  • the one or more processors included in the terminal may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.
  • the memory 33 is used to store programs.
  • the memory 33 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), for example, at least one disk memory.
  • the program can specifically be used to make the processor 31 perform the following operations: obtain all heartbeat data information of the sample electrocardiogram; perform mean wave processing on the heartbeat data information according to a preset clustering and merging algorithm, and combine the processing results Determine the mean waveform of the sample electrocardiogram; label the mean waveform with P wave, QRS wave, T wave, and ST section information, and output the marked P wave, QRS wave, T wave, and ST section waveform information.
  • the ECG data processing method provided in this application further ensures the privacy and security of all the above-mentioned data
  • all the above-mentioned data can also be stored in a node of a blockchain.
  • sample ECG and waveform information, etc. these data can be stored in the blockchain node.
  • the blockchain referred to in this application is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.
  • all the heartbeat data information of the sample electrocardiogram is obtained; according to a preset clustering and merging algorithm, the heartbeat data information is processed by the mean wave, and the processing result is determined as the sample electrocardiogram
  • the average waveform of the average waveform; P wave, QRS wave, T wave, and ST segment information are labeled on the average waveform, and the labeled P wave, QRS wave, T wave, and ST section waveform information are output.
  • This can synthesize the mean waveform through clustering, thereby smoothing the filtering, effectively weakening the influence of baseline wandering noise, electromyographic noise, power frequency interference and sudden noise, and greatly improving the P wave, QRS wave and T wave The accuracy of the start and end points of the label.
  • modules or units or components in the embodiments can be combined into one module or unit or component, and in addition, they can be divided into multiple sub-modules or sub-units or sub-components. Except that at least some of such features and/or processes or units are mutually exclusive, any combination can be used to compare all the features disclosed in this specification (including the accompanying claims, abstract and drawings) and any method or methods disclosed in this manner or All the processes or units of the equipment are combined. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.
  • the various component embodiments of the present application may be implemented by hardware, or by software modules running on one or more processors, or by a combination of them.
  • a microprocessor or a digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components in the embodiments of the present application.
  • This application can also be implemented as a device or device program (for example, a computer program and a computer program product) for executing part or all of the methods described herein.
  • Such a program for implementing the present application may be stored on a computer-readable medium, or may have the form of one or more signals.
  • Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Molecular Biology (AREA)
  • Animal Behavior & Ethology (AREA)
  • Pathology (AREA)
  • Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Cardiology (AREA)
  • Surgery (AREA)
  • Biophysics (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Signal Processing (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physiology (AREA)
  • Psychiatry (AREA)
  • Power Engineering (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

一种心电图数据处理的方法、装置、存储介质及计算机设备,涉及数据处理领域,主要目的在于能够通过聚类合成均值波形,从而平滑滤波,有效弱化了基线漂移噪声、肌电噪声、工频干扰以及突发的噪声的影响,极大的提升了对P波,QRS波和T波的起止点标注的准确率。该方法包括:获取样本心电图的所有心搏数据信息(101);根据预设的聚类合并算法,对心搏数据信息进行均值波处理,并将处理结果确定为样本心电图的均值波形(102);对均值波形进行P波,QRS波、T波以及ST段的信息标注(103),输出标注后的P波、QRS波、T波以及ST段的波形信息(104)。该方法、装置、存储介质及计算机设备适用于心电图数据处理。

Description

心电图数据处理的方法、装置、计算机设备及存储介质
本申请要求于2020年2月27日提交中国专利局、申请号为CN202010123259.0,发明名称为“心电图数据处理的方法、装置、存储介质及计算机设备”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及数据处理领域,特别是涉及一种心电图数据处理的方法、装置、计算机设备及存储介质。
背景技术
随着医疗信息技术的不断发展,心电图自动诊断技术已经在临床上得到了一定的应用,虽然目前心电图自动诊断技术还无法与专业医生的诊断水平相比,但在某些方面已经得到了一定程度的认可,极大的减轻了专业医生的负担。
目前,现有的心电图自动诊断技术是对心电图的每个心搏的P波、QRS波和T波进行信息标注和特征提取。
技术问题
在对P波,QRS波和T波的起止点标注的过程中,发明人意识到由于受到基线漂移噪声、肌电噪声、工频干扰以及其他突发噪声的影响,会导致检测结果的准确度与R波顶点相比有很大降低。
技术解决方案
一种心电图数据处理的方法,包括:
获取样本心电图的所有心搏数据信息;
根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;
对所述均值波形进行P波,QRS波、T波以及ST段的信息标注;
输出标注后的P波、QRS波、T波以及ST段的波形信息。
一种心电图数据处理的装置,包括:
获取单元,用于获取样本心电图的所有心搏数据信息;
处理单元,用于根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;
标注单元,用于对所述均值波形进行P波,QRS波、T波以及ST段的信息标注;
输出单元,用于输出标注后的P波、QRS波、T波以及ST段的波形信息。
一种存储介质,所述存储介质中存储有至少一可执行指令,所述执行指令使处理器执行以下步骤:
获取样本心电图的所有心搏数据信息;根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;对所述均值波形进行P波,QRS波、T波以及ST段的信息标注,输出标注后的P波、QRS波、T波以及ST段的波形信息。
一种计算机设备,包括处理器、存储器、通信接口和通信总线所述处理器、所述存储器和所述通信接口通过所述通信总线完成相互间的通信,所述存储器用于存放至少一可执行指令,所述可执行指令使所述处理器执行以下步骤:
获取样本心电图的所有心搏数据信息;根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;对所述均值波 形进行P波,QRS波、T波以及ST段的信息标注,输出标注后的P波、QRS波、T波以及ST段的波形信息。
有益效果
本申请能够通过聚类合成均值波形,平滑滤波,有效弱化了基线漂移噪声,肌电噪声,工频干扰以及突发的噪声的影响,极大的提升了对P波,QRS波和T波的起止点标注的准确率,以便提高对心电图数据处理的准确性,从而提高利用心电图得到的数据进行分析的效率。
附图说明
通过阅读下文优选实施方式的详细描述,各种其他的优点和益处对于本领域普通技术人员将变得清楚明了。附图仅用于示出优选实施方式的目的,而并不认为是对本申请的限制。而且在整个附图中,用相同的参考符号表示相同的部件。在附图中:
图1示出了本申请实施例提供的一种心电图数据处理的方法流程图;
图2示出了本申请实施例提供的一种心电图数据处理装置的结构示意图;
图3示出了本申请实施例提供的另一种心电图数据处理装置的结构示意图;
图4示出了本申请实施例提供的一种计算机设备的实体结构示意图;
本发明的实施方式
下面将参照附图更详细地描述本公开的示例性实施例。虽然附图中显示了本公开的示例性实施例,然而应当理解,可以以各种形式实现本公开而不应被这里阐述的实施例所限制。相反,提供这些实施例是为了能够更透彻地理解本公开,并且能够将本公开的范围完整的传达给本领域的技术人员。
如背景技术所述,目前,现有的心电图自动诊断技术是对心电图的每个心搏的P波、QRS波和T波进行信息标注和特征提取。然而,在对P波,QRS波和T波的起止点标注的过程中,由于受到基线漂移噪声、肌电噪声、工频干扰以及其他突发噪声的影响,会导致检测结果的准确度与R波顶点相比有很大降低。
为了解决上述问题,本申请实施例提供了一种心电图数据处理的方法,如图1所示,所述方法包括:
101、获取样本心电图的所有心搏数据信息。
其中,所述心搏数据信息可以包括所述样本心电图中的波形数据信息。
102、根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形。
其中,为了降低波形数据信息采集时的工频干扰,基线漂移等噪声,可以通过所述预设的聚类合并算法对上述样本心电图中的波形数据信息进行聚类合并,以获取该导联心电图的均值波形,所述均值波形可以体现大多数波形的共有特征,从而提高P波,QRS波,T波起止点标注的准确性。所述根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理的具体过程可以包括:对所述获取的样本心电图的所有心搏数据信息依次进行R波检测、筛选处理和聚类合并,所得到的波形可以确定为所述样本心电图的均值波形。
需要说明的是,在现有技术中都是对全部心搏进行R波检测以及P波、QRS波、T波和ST段进行分析,并提取特征值。而本申请只需对所述合成的均值波形的P波、QRS波、T波和ST段进行分析和特征提取,可以极大地减少运算量,提高心电图数据分析和特征提取的效率。
103、对所述均值波形进行P波、QRS波、T波以及ST段的信息标注。
具体地,可以直接根据所述均值波形,以及所述P波、QRS波、T波以及ST段对应的形状特征进行识别,并根据所述识别结果对所述P波、QRS波、T波和ST段进行信息标注。需要说明的是,所述通过聚合心搏得到的均值波形可以体现大多数波形的共有特征,另外,由于心电信号比较微弱,在采集的时候容易受到工频干扰,基线漂移等干扰,会影响P波、QRS波、T波起止点标注的准确性,标注的位置也会随着干扰波动,而通过聚合心搏得到的均值波形,可以降低干扰,从而提高P波、QRS波、T波起止点标注的准确性。
104、输出标注后的P波、QRS波、T波以及ST段的波形信息。
具体地,根据上述对所述均值波形进行P波、QRS波、T波以及ST段的信息标注后,将所述标注的波形信息进行输出。
进一步的,为了更好的说明上述心电图数据处理方法的过程,作为对上述实施例的细化和扩展,本申请实施例提供了几种可选实施例,但不限于此,具体如下所示:
在本申请的一个可选实施例,所述步骤102具体可以包括:对所述心搏数据信息进行噪声处理,并对处理后的心搏数据信息进行R波检测,获取R波数据信息;根据所述R波数据信息,以及预设的筛选算法,对所述心搏数据信息进行筛选;对所述筛选结果进行聚类合并,并将所述聚类合并所得的结果确定为所述样本心电图的均值波形。其中,所述R波检测的具体过程可以包括:带通滤波:使用5-18Hz的前向滤波器对输入的心电图信号进行滤波并做相位延迟补偿;差分:对前向滤波器输出的信号进行差分处理,形成差分信号d(n);数据整理:对差分信号进行变换:输出d'(n)=输入的绝对值/G1;对经数据整理后的信号使用公式-d'(n)*d'(n)*log(d'(n)*d'(n))进行香农能量转换;平均滤波:使用滤波宽度M=55~75点(153~208ms)的前向滤波器对处理后的信号进行滤波并做相位延迟补偿;检测极大/小点:所述极大点是指取值大于位于其左右两侧的点的选定点,所述极小点是指取值小于位于其左右两侧的点的选定点;排除假R点;纠正误排除点;在近似R波位置周围±25点范围内寻找到真正R位置。
对于本申请实施例,所述对所述心搏数据信息进行噪声处理,具体可以包括:利用带通滤波对样本心电图数据进行噪声处理。
其中,通过带通滤波能够对基线漂移噪声、肌电噪声、工频干扰以及突发的噪声等进行滤波,实现提升P波、QRS波、T波的起止点标注的准确率。具体地,可以采用低频截止频率为14Hz,高频截止频率为20Hz的带通滤波在R波检测的时候进行滤波。针对基线漂移可以采用二阶中值滤波器,滤波后能有效的抑制基线漂移,且滤波后对ST段无明显变化。还可以采用自适应滤波器滤除心电信号中的50Hz工频干扰。
对于本申请实施例,所述根据所述R波数据信息,以及预设的筛选算法,对所述心搏数据信息进行筛选,具体可以包括:根据获取的所有心搏的RR间期数据信息,计算平均RR间期数据信息;根据所述平均RR间期数据信息,配置筛选区间,筛选RR间期数据信息处于所述筛选区间内的心搏数据信息。
其中,所述RR间期可以为当前R波与前一个R波顶点所在的位置差值,心率与RR间期成反比关系,具体换算公式可以为:
Figure PCTCN2020119053-appb-000001
对于本申请实施例,平均RR间期可以为所有心搏的RR间期的平均值,配置所述平均RR间期的筛选区间,例如,可以选取0.8倍~1.2倍的平均RR间期作为筛选区间。
需要说明的是,聚合心搏以获取均值波形是为了合并多数属性相同的心搏,但对于偶发早搏,尤其是室性早搏,由于起搏点与其他心搏不同,即心搏的属性不同,可能会带来心搏的形状上的一些差异,如果参与叠加聚合,会对聚合心搏波形的准确性产生影响,因 此,所述偶发早搏不参与叠加聚合。
对于本申请实施例,所述对所述筛选结果进行聚类合并,并将所述聚类合并所得的结果确定为所述样本心电图的均值波形,具体可以包括:对所述筛选的心搏数据信息配置采样点,并对所有采样点进行叠加求均值处理;将处理所得结果确定为所述样本心电图的均值波形。
其中,可以对所述筛选的心搏数据信息配置采样点,并对所述采样点进行叠加求均值处理;将处理所得结果确定为所述样本心电图的均值波形。需要说明的是,R波两侧的叠加范围可以根据平均RR间期确定。如:平均RR间期为1000ms,则可以确定以R波顶点为中心点,对R-500ms~R+500ms范围内的采样点进行叠加求均值处理。
在本申请的一个可选实施例,为了得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息,所述方法还可以包括:对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息。
需要说明的是,在特征标注和提取方面,现有技术通常只关注QRS波的Q波,R波和S波的波电压以及QRS波时间等信息,而往往忽略了对于Q波,R波和S波每个波的时间,以及左右心室除极过程中的变化的信息的收集和分析。例如:左束支传导阻滞的QRS波前半部分形态和时间正常,后半部分时间延长以及QRS波的形态呈宽大、平顶或有切迹的R波,上述左束支传导阻滞的心电图特征在通过现有技术标注和提取的特征中无法得到体现;另外,现有技术对QRS波的形态关注不足,如右束支传导阻滞中的QRS波呈现“M”形态,而上述形态在通过现有技术标注和提取的特征中也无法得到体现。而本申请对QRS波的切迹、形态的判断以及电压变化过程的记录,使得上述心电图特征得以在AI预测中起到一定作用。
对于本申请实施例,所述对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息,具体可以包括:根据所述均值波形,检测R波顶点两侧的拐点位置信息,并根据所述拐点位置信息,判断R波是否存在切迹;计算所述均值波形中QRS波过零点的数目,并根据所述数目,确定QRS波的形态;以所述均值波形中R波顶点为中心点,分别记录所述中心点两侧波形的电压变化过程。
其中,计算所述均值波形中QRS波过零点的数目,并根据所述数目,确定QRS波的形态的过程具体可以包括:计算所述均值波形中QRS波过零点的数目,并根据所述数目,确定QRS波的形态,所述QRS波可以呈现rsR’、RSR’或rR’形态。具体地,可以以R波顶点为中心,分别判断其左右两侧过零点的数目。例如,R波顶点左侧存在一个过零点,R波顶点右侧存在三个过零点,则可以判断QRS波形态为RSR’。
所述QRS波电压变化过程的记录方式可以包括:以所述均值波形中R波顶点为中心点,分别记录其两侧波形的电压变化过程,所述电压变化过程可以体现QRS波后半部分宽大顿挫的特点。具体地,可以以R波顶点为中心,每隔Nms(如10ms)取一个点。例如,采样率为500Hz,均值波形中R波顶点的位置为点x,均值波的电压为y x,以均值波形的R波顶点为中心,向其左右两侧每隔10ms记录电压值为y x+5、y x-5、y x+10、y x-10……,横坐标的取值范围可以为x-0.1s~x+0.1s。
进一步地,所述根据所述均值波形,检测R波顶点两侧的拐点位置信息,并根据所述拐点位置信息,判断R波是否存在切迹,具体可以包括:对所述均值波形进行一阶差分处理,检测R波顶点两侧的拐点位置信息;将所述拐点位置信息与所述P波、QRS波、T波、ST段起止点位置信息进行对比;根据所述对比结果,判断R波是否存在切迹。
其中,所述判断R波是否存在切迹的具体过程可以包括:对所述均值波形进行一阶差分处理,检测R波顶点两侧的拐点,所述一阶差分就是对均值波形求x点与x+1点的电压差,如果均值波形的某点的左右两侧一阶差分的值正负相反,则可以判断该点为QRS波中的拐点;根据拐点的位置与P波、QRS波、T波起止点位置的对比结果,可以判断R 波是否有切迹。例如,若R波顶点与QRS波终点中间存在两个拐点,则说明该R波有切迹。
在本申请的一个可选实施例,为了利用数据处理后的心电图数据确定目标状态信息,所述方法还可以包括:当接收到心电图分析处理请求时,根据所述请求中携带的心电图数据信息,以及预先训练的目标状态识别模型进行预测分析,并输出分析结果,以确定目标状态信息,所述预先训练的目标状态识别模型的训练样本数据是通过均值波处理进行波段分析得到的。
具体地,接收心电图分析处理请求,所述请求中可以携带有所述心电图的数据信息;可以根据预先训练的目标状态识别模型进行预测分析,并输出分析结果,从而可以根据所述分析结果,对应确定目标状态信息。
本申请可以通过聚类合成均值波形,从而平滑滤波,有效地弱化基线漂移噪声、肌电噪声、工频干扰以及突发的噪声的影响,极大的提升对P波,QRS波和T波的起止点标注的准确率。
在本申请的一个可选实施例,为了确定目标状态识别模型,所述方法还可以包括:获取所述样本心电图的样本数据信息,所述样本数据信息包括P波、QRS波、T波、ST段,R波切迹、QRS波形态、QRS波电压变化信息以及全部心搏的RR间期数据信息;根据所述样本数据信息、样本心电图对应的目标状态识别标签对预设机器学习算法进行训练,确定目标状态识别模型。
具体地,可以获取所述样本心电图的P波、QRS波、T波和ST段,R波切迹、QRS波形态、QRS电压变化过程全部心搏的RR间期信息等样本数据信息,并利用所述样本心电图对应的目标状态识别标签对预设机器学习算法进行训练,以确定所述目标状态识别模型,所述样本心电图对应的目标状态识别标签可以为样本心电图的识别结果与所述目标状态的对应关系,例如:对所述样本数据信息进行特征提取后,可以根据对应的目标状态识别标签,识别对应的目标状态信息,如所述目标状态信息可以为心脏状态信息,具体可以包括起自心房或心室的异常搏动以及房室、束支传导阻滞、心率不齐等。所述识别目标状态识别模型可以用于确定目标状态信息,所述机器学习算法可以包括深度学习算法、SVM算法等。
进一步地,作为图1的具体实现,本申请实施例提供了一种心电图数据处理装置,如图2所示,所述装置包括:获取单元21、处理单元22、标注单元23和输出单元24。
所述获取单元21,可以用于获取样本心电图的所有心搏数据信息;
所述处理单元22,可以用于根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;
所述标注单元23,可以用于对所述均值波形进行P波,QRS波、T波以及ST段的信息标注。
所述输出单元24,可以用于输出标注后的P波、QRS波、T波以及ST段的波形信息。
所述处理单元22,可以包括:处理模块221、检测模块222、筛选模块223和聚类合并模块224,如图3所示。
处理模块221,可以用于对所述心搏数据信息进行噪声处理;
检测模块222,可以用于对处理后的心搏数据信息进行R波检测,获取R波数据信息;
筛选模块223,可以用于根据所述R波数据信息,以及预设的筛选算法,对所述心搏数据信息进行筛选;
聚类合并模块224,可以用于对所述筛选结果进行聚类合并,并将所述聚类合并所得的结果确定为所述样本心电图的均值波形。
进一步地,所述处理模块221,具体可以用于利用带通滤波对样本心电图数据进行噪声处理。
进一步地,所述筛选模块223,具体可以用于根据获取的所有心搏的RR间期数据信息,计算平均RR间期数据信息,根据所述平均RR间期数据信息,配置筛选区间,筛选RR间期数据信息处于所述筛选区间内的心搏数据信息。
进一步地,所述聚类合并模块224,具体可以用于对所述筛选的心搏数据信息配置采样点,并对所有采样点进行叠加求均值处理,将处理所得结果确定为所述样本心电图的均值波形。
对于本申请实施例,为了得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息,所述装置还可以包括:第一分析单元25。
所述第一分析单元25,可以用于对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息。
所述第一分析单元25,可以包括:判断模块251、确定模块252和记录模块253。
判断模块251,可以用于根据所述均值波形,检测R波顶点两侧的拐点位置信息,并根据所述拐点位置信息,判断R波是否存在切迹;
确定模块252,可以用于计算所述均值波形中QRS波过零点的数目,并根据所述数目,确定QRS波的形态;
记录模块253,可以用于以所述均值波形中R波顶点为中心点,分别记录所述中心点两侧波形的电压变化过程。
进一步地,所述判断模块251,具体可以用于对所述均值波形进行一阶差分处理,检测R波顶点两侧的拐点位置信息,将所述拐点位置信息与所述P波、QRS波、T波、ST段起止点位置信息进行对比,根据所述对比结果,判断R波是否存在切迹。
对于本申请实施例,为了确定目标状态信息,所述装置还可以包括:第二分析单元26。
所述第二分析单元26,可以用于当接收到心电图分析处理请求时,根据所述请求中携带的心电图数据信息,以及预先训练的目标状态识别模型进行预测分析,并输出分析结果,以确定目标状态信息,所述预先训练的目标状态识别模型的训练样本数据是通过均值波处理进行波段分析得到的。
对于本申请实施例,为了确定目标状态识别模型,所述装置还可以包括:获取单元27和训练单元28。
所述获取单元26,可以用于获取所述样本心电图的样本数据信息,所述样本数据信息包括P波、QRS波、T波、ST段,R波切迹、QRS波形态、QRS波电压变化信息以及全部心搏的RR间期数据信息;
所述训练单元28,可以用于根据所述样本数据信息、样本心电图对应的目标状态识别标签对预设机器学习算法进行训练,确定目标状态识别模型。
基于上述如图1所示方法,相应的,本申请实施例还提供了一种存储介质,所述存储介质可能包含高速RAM存储器,可能包含易失性存储器,也可能还包括非易失性存储器(non-volatile memory),例如至少一个磁盘存储器,所述存储介质中存储有至少一可执行指令,所述执行指令使处理器执行以下步骤:获取样本心电图的所有心搏数据信息;根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;对所述均值波形进行P波,QRS波、T波以及ST段的信息标注,输出标注后的P波、QRS波、T波以及ST段的波形信息。
基于上述如图1所示方法和如图2所示装置的实施例,本申请实施例还提供了一种计算机设备,如图4所示,处理器(processor)31、通信接口(Communications Interface)32、存储器(memory)33、以及通信总线34。其中:处理器31、通信接口32、以及存储器33通过通信总线34完成相互间的通信。通信接口34,用于与其它设备比如用户端或其它服务器等的网元通信。处理器31,用于执行程序,具体可以执行上述心电图数据处理的方法实施例中的相关步骤。具体地,程序可以包括程序代码,该程序代码包括计算机操作指令。处 理器31可能是中央处理器CPU,或者是特定集成电路ASIC(Application Specific Integrated Circuit),或者是被配置成实施本申请实施例的一个或多个集成电路。
终端包括的一个或多个处理器,可以是同一类型的处理器,如一个或多个CPU;也可以是不同类型的处理器,如一个或多个CPU以及一个或多个ASIC。存储器33,用于存放程序。存储器33可能包含高速RAM存储器,也可能还包括非易失性存储器(non-volatile memory),例如至少一个磁盘存储器。程序具体可以用于使得处理器31执行以下操作:获取样本心电图的所有心搏数据信息;根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;对所述均值波形进行P波,QRS波、T波以及ST段的信息标注,输出标注后的P波、QRS波、T波以及ST段的波形信息。
在另一实施例中,本申请所提供的心电图数据处理的方法,为进一步保证上述所有出现的数据的私密和安全性,上述所有数据还可以存储于一区块链的节点中。例如样本心电图及波形信息等等,这些数据均可存储在区块链节点中。
需要说明的是,本申请所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。
通过本申请的技术方案,获取样本心电图的所有心搏数据信息;根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;对所述均值波形进行P波,QRS波、T波以及ST段的信息标注,输出标注后的P波、QRS波、T波以及ST段的波形信息。从而能够能够通过聚类合成均值波形,从而平滑滤波,有效弱化了基线漂移噪声,肌电噪声,工频干扰以及突发的噪声的影响,极大的提升了对P波,QRS波和T波的起止点标注的准确率。
在上述实施例中,对各个实施例的描述都各有侧重,某个实施例中没有详述的部分,可以参见其他实施例的相关描述。
可以理解的是,上述方法及装置中的相关特征可以相互参考。另外,上述实施例中的“第一”、“第二”等是用于区分各实施例,而并不代表各实施例的优劣。
所属领域的技术人员可以清楚地了解到,为描述的方便和简洁,上述描述的系统,装置和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
在此提供的算法和显示不与任何特定计算机、虚拟系统或者其它设备固有相关。各种通用系统也可以与基于在此的示教一起使用。根据上面的描述,构造这类系统所要求的结构是显而易见的。此外,本申请也不针对任何特定编程语言。应当明白,可以利用各种编程语言实现在此描述的本申请的内容,并且上面对特定语言所做的描述是为了披露本申请的最佳实施方式。
在此处所提供的说明书中,说明了大量具体细节。然而,能够理解,本申请的实施例可以在没有这些具体细节的情况下实践。在一些实例中,并未详细示出公知的方法、结构和技术,以便不模糊对本说明书的理解。
类似地,应当理解,为了精简本公开并帮助理解各个发明方面中的一个或多个,在上面对本申请的示例性实施例的描述中,本申请的各个特征有时被一起分组到单个实施例、图、或者对其的描述中。然而,并不应将该公开的方法解释成反映如下意图:即所要求保护的本申请要求比在每个权利要求中所明确记载的特征更多的特征。更确切地说,如下面的权利要求书所反映的那样,发明方面在于少于前面公开的单个实施例的所有特征。因此,遵循具体实施方式的权利要求书由此明确地并入该具体实施方式,其中每个权利要求本身都作为本申请的单独实施例。
本领域那些技术人员可以理解,可以对实施例中的设备中的模块进行自适应性地改变并且把它们设置在与该实施例不同的一个或多个设备中。可以把实施例中的模块或单元或组件组合成一个模块或单元或组件,以及此外可以把它们分成多个子模块或子单元或子组 件。除了这样的特征和/或过程或者单元中的至少一些是相互排斥之外,可以采用任何组合对本说明书(包括伴随的权利要求、摘要和附图)中公开的所有特征以及如此公开的任何方法或者设备的所有过程或单元进行组合。除非另外明确陈述,本说明书(包括伴随的权利要求、摘要和附图)中公开的每个特征可以由提供相同、等同或相似目的的替代特征来代替。
此外,本领域的技术人员能够理解,尽管在此所述的一些实施例包括其它实施例中所包括的某些特征而不是其它特征,但是不同实施例的特征的组合意味着处于本申请的范围之内并且形成不同的实施例。例如,在下面的权利要求书中,所要求保护的实施例的任意之一都可以以任意的组合方式来使用。
本申请的各个部件实施例可以以硬件实现,或者以在一个或者多个处理器上运行的软件模块实现,或者以它们的组合实现。本领域的技术人员应当理解,可以在实践中使用微处理器或者数字信号处理器(DSP)来实现根据本申请实施例中的一些或者全部部件的一些或者全部功能。本申请还可以实现为用于执行这里所描述的方法的一部分或者全部的设备或者装置程序(例如,计算机程序和计算机程序产品)。这样的实现本申请的程序可以存储在计算机可读介质上,或者可以具有一个或者多个信号的形式。这样的信号可以从因特网网站上下载得到,或者在载体信号上提供,或者以任何其他形式提供。
应该注意的是上述实施例对本申请进行说明而不是对本申请进行限制,并且本领域技术人员在不脱离所附权利要求的范围的情况下可设计出替换实施例。在权利要求中,不应将位于括号之间的任何参考符号构造成对权利要求的限制。单词“包含”不排除存在未列在权利要求中的元件或步骤。位于元件之前的单词“一”或“一个”不排除存在多个这样的元件。本申请可以借助于包括有若干不同元件的硬件以及借助于适当编程的计算机来实现。在列举了若干装置的单元权利要求中,这些装置中的若干个可以是通过同一个硬件项来具体体现。单词第一、第二、以及第三等的使用不表示任何顺序。可将这些单词解释为名称。

Claims (20)

  1. 一种心电图数据处理的方法,其中,包括:
    获取样本心电图的所有心搏数据信息;
    根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;
    对所述均值波形进行P波、QRS波、T波以及ST段的信息标注;
    输出标注后的P波、QRS波、T波以及ST段的波形信息。
  2. 根据权利要求1所述的方法,其中,所述根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形,包括:
    对所述心搏数据信息进行噪声处理,并对处理后的心搏数据信息进行R波检测,获取R波数据信息;
    根据所述R波数据信息,以及预设的筛选算法,对所述心搏数据信息进行筛选;
    对所述筛选结果进行聚类合并,并将所述聚类合并所得的结果确定为所述样本心电图的均值波形。
  3. 根据权利要求2所述的方法,其中,所述对所述心搏数据信息进行噪声处理,包括:
    利用带通滤波对样本心电图数据进行噪声处理;
    所述根据所述R波数据信息,以及预设的筛选算法,对所述心搏数据信息进行筛选,包括:
    根据获取的所有心搏的RR间期数据信息,计算平均RR间期数据信息;
    根据所述平均RR间期数据信息,配置筛选区间,筛选RR间期数据信息处于所述筛选区间内的心搏数据信息;
    所述对所述筛选结果进行聚类合并,并将所述聚类合并所得的结果确定为所述样本心电图的均值波形,包括:
    对所述筛选的心搏数据信息配置采样点,并对所有采样点进行叠加求均值处理;
    将处理所得结果确定为所述样本心电图的均值波形。
  4. 根据权利要求1所述的方法,其中,所述对所述均值波形进行P波,QRS波、T波以及ST段的信息标注之后,所述方法还包括:
    对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息。
  5. 根据权利要求4所述的方法,其中,所述对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息,包括:
    根据所述均值波形,检测R波顶点两侧的拐点位置信息,并根据所述拐点位置信息,判断R波是否存在切迹;
    计算所述均值波形中QRS波过零点的数目,并根据所述数目,确定QRS波的形态;
    以所述均值波形中R波顶点为中心点,分别记录所述中心点两侧波形的电压变化过程。
  6. 根据权利要求5所述的方法,其中,所述根据所述均值波形,检测R波顶点两侧的拐点位置信息,并根据所述拐点位置信息,判断R波是否存在切迹,包括:
    对所述均值波形进行一阶差分处理,检测R波顶点两侧的拐点位置信息;
    将所述拐点位置信息与所述P波、QRS波、T波、ST段起止点位置信息进行对比;
    根据所述对比结果,判断R波是否存在切迹。
  7. 根据权利要求4所述的方法,其中,所述对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息之后,所述方法还包括:
    当接收到心电图分析处理请求时,根据所述请求中携带的心电图数据信息,以及预先 训练的目标状态识别模型进行预测分析,并输出分析结果,以确定目标状态信息,所述预先训练的目标状态识别模型的训练样本数据是通过均值波处理进行波段分析得到的。
  8. 一种心电图数据处理装置,其中,包括:
    获取单元,用于获取样本心电图的所有心搏数据信息;
    处理单元,用于根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;
    标注单元,用于对所述均值波形进行P波,QRS波、T波以及ST段的信息标注;
    输出单元,用于输出标注后的P波、QRS波、T波以及ST段的波形信息。
  9. 一种存储介质,其上存储有计算机程序,所述存储介质中存储有至少一可执行指令,所述执行指令使处理器执行如下步骤:
    获取样本心电图的所有心搏数据信息;
    根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;
    对所述均值波形进行P波、QRS波、T波以及ST段的信息标注;
    输出标注后的P波、QRS波、T波以及ST段的波形信息。
  10. 根据权利要求9所述的存储介质,其中,所述根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形,包括:
    对所述心搏数据信息进行噪声处理,并对处理后的心搏数据信息进行R波检测,获取R波数据信息;
    根据所述R波数据信息,以及预设的筛选算法,对所述心搏数据信息进行筛选;
    对所述筛选结果进行聚类合并,并将所述聚类合并所得的结果确定为所述样本心电图的均值波形。
  11. 根据权利要求10所述的存储介质,其中,所述对所述心搏数据信息进行噪声处理,包括:
    利用带通滤波对样本心电图数据进行噪声处理;
    所述根据所述R波数据信息,以及预设的筛选算法,对所述心搏数据信息进行筛选,包括:
    根据获取的所有心搏的RR间期数据信息,计算平均RR间期数据信息;
    根据所述平均RR间期数据信息,配置筛选区间,筛选RR间期数据信息处于所述筛选区间内的心搏数据信息;
    所述对所述筛选结果进行聚类合并,并将所述聚类合并所得的结果确定为所述样本心电图的均值波形,包括:
    对所述筛选的心搏数据信息配置采样点,并对所有采样点进行叠加求均值处理;
    将处理所得结果确定为所述样本心电图的均值波形。
  12. 根据权利要求9所述的存储介质,其中,所述对所述均值波形进行P波,QRS波、T波以及ST段的信息标注之后,所述执行指令使处理器还执行如下步骤:
    对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息。
  13. 根据权利要求12所述的存储介质,其中,所述对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息,包括:
    根据所述均值波形,检测R波顶点两侧的拐点位置信息,并根据所述拐点位置信息,判断R波是否存在切迹;
    计算所述均值波形中QRS波过零点的数目,并根据所述数目,确定QRS波的形态;
    以所述均值波形中R波顶点为中心点,分别记录所述中心点两侧波形的电压变化过程。
  14. 根据权利要求13所述的存储介质,其中,所述根据所述均值波形,检测R波顶点两侧的拐点位置信息,并根据所述拐点位置信息,判断R波是否存在切迹,包括:
    对所述均值波形进行一阶差分处理,检测R波顶点两侧的拐点位置信息;
    将所述拐点位置信息与所述P波、QRS波、T波、ST段起止点位置信息进行对比;
    根据所述对比结果,判断R波是否存在切迹。
  15. 根据权利要求12所述的存储介质,其中,所述对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息之后,所述执行指令使处理器还执行如下步骤:
    当接收到心电图分析处理请求时,根据所述请求中携带的心电图数据信息,以及预先训练的目标状态识别模型进行预测分析,并输出分析结果,以确定目标状态信息,所述预先训练的目标状态识别模型的训练样本数据是通过均值波处理进行波段分析得到的。
  16. 一种计算机设备,包括处理器、存储器、通信接口和通信总线所述处理器、所述存储器和所述通信接口通过所述通信总线完成相互间的通信,所述存储器用于存放至少一可执行指令,所述可执行指令使所述处理器执行如下步骤:
    获取样本心电图的所有心搏数据信息;
    根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形;
    对所述均值波形进行P波、QRS波、T波以及ST段的信息标注;
    输出标注后的P波、QRS波、T波以及ST段的波形信息。
  17. 根据权利要求16所述的计算机设备,其中,所述根据预设的聚类合并算法,对所述心搏数据信息进行均值波处理,并将所述处理结果确定为所述样本心电图的均值波形,包括:
    对所述心搏数据信息进行噪声处理,并对处理后的心搏数据信息进行R波检测,获取R波数据信息;
    根据所述R波数据信息,以及预设的筛选算法,对所述心搏数据信息进行筛选;
    对所述筛选结果进行聚类合并,并将所述聚类合并所得的结果确定为所述样本心电图的均值波形。
  18. 根据权利要求17所述的计算机设备,其中,所述对所述心搏数据信息进行噪声处理,包括:
    利用带通滤波对样本心电图数据进行噪声处理;
    所述根据所述R波数据信息,以及预设的筛选算法,对所述心搏数据信息进行筛选,包括:
    根据获取的所有心搏的RR间期数据信息,计算平均RR间期数据信息;
    根据所述平均RR间期数据信息,配置筛选区间,筛选RR间期数据信息处于所述筛选区间内的心搏数据信息;
    所述对所述筛选结果进行聚类合并,并将所述聚类合并所得的结果确定为所述样本心电图的均值波形,包括:
    对所述筛选的心搏数据信息配置采样点,并对所有采样点进行叠加求均值处理;
    将处理所得结果确定为所述样本心电图的均值波形。
  19. 根据权利要求16所述的计算机设备,其中,所述对所述均值波形进行P波,QRS波、T波以及ST段的信息标注之后,所述可执行指令使所述处理器还执行如下步骤:
    对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息。
  20. 根据权利要求19所述的计算机设备,其中,所述对所述均值波形进行波段分析,得到所述均值波形的R波切迹、QRS波形态、QRS波电压变化信息,包括:
    根据所述均值波形,检测R波顶点两侧的拐点位置信息,并根据所述拐点位置信息,判断R波是否存在切迹;
    计算所述均值波形中QRS波过零点的数目,并根据所述数目,确定QRS波的形态;
    以所述均值波形中R波顶点为中心点,分别记录所述中心点两侧波形的电压变化过程。
PCT/CN2020/119053 2020-02-27 2020-09-29 心电图数据处理的方法、装置、计算机设备及存储介质 WO2021169296A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010123259.0A CN111419212A (zh) 2020-02-27 2020-02-27 心电图数据处理的方法、装置、存储介质及计算机设备
CN202010123259.0 2020-02-27

Publications (1)

Publication Number Publication Date
WO2021169296A1 true WO2021169296A1 (zh) 2021-09-02

Family

ID=71547239

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/119053 WO2021169296A1 (zh) 2020-02-27 2020-09-29 心电图数据处理的方法、装置、计算机设备及存储介质

Country Status (2)

Country Link
CN (1) CN111419212A (zh)
WO (1) WO2021169296A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114073534A (zh) * 2021-11-25 2022-02-22 成都泰盟软件有限公司 一种心功能分析算法
CN115486855A (zh) * 2022-09-15 2022-12-20 浙江好络维医疗技术有限公司 一种基于qrs波群不定次循环叠加的心电图心搏分类方法
CN115778402A (zh) * 2022-12-02 2023-03-14 深圳华清心仪医疗电子有限公司 一种动态心电信号的伪差识别方法及系统

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111419212A (zh) * 2020-02-27 2020-07-17 平安科技(深圳)有限公司 心电图数据处理的方法、装置、存储介质及计算机设备
CN111959132B (zh) * 2020-08-17 2021-11-16 厦门纳龙科技有限公司 一种心电图胸导减半报告打印方法、终端设备及存储介质
CN111956205B (zh) * 2020-10-21 2021-01-08 平安科技(深圳)有限公司 一种心电图数据处理设备、方法、装置及存储介质
CN112799556B (zh) * 2021-01-26 2022-12-30 浙江吉利控股集团有限公司 一种车辆数据的显示控制方法及波形显示装置
CN113712569B (zh) * 2021-11-01 2022-02-08 毕胜普生物科技有限公司 高频qrs波群数据分析方法及装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104586383A (zh) * 2015-01-12 2015-05-06 北京世纪今科医疗器械有限公司 一种心电波形分类方法及装置
CN106108889A (zh) * 2016-07-20 2016-11-16 杨平 基于深度学习算法的心电图分类方法
CN108024717A (zh) * 2015-10-29 2018-05-11 美多尼克监测公司 确定室性期前收缩(pvc)类型和负荷的系统和方法
US20190167143A1 (en) * 2015-10-27 2019-06-06 Cardiologs Technologies Sas Electrocardiogram processing system for delineation and classification
CN110367969A (zh) * 2019-07-05 2019-10-25 复旦大学 一种改进的心电信号快速聚类分析方法
CN111419212A (zh) * 2020-02-27 2020-07-17 平安科技(深圳)有限公司 心电图数据处理的方法、装置、存储介质及计算机设备

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1084044A (zh) * 1992-09-11 1994-03-23 中日友好医院 三维高频心电信号分析系统及方法
CN100589754C (zh) * 2006-07-14 2010-02-17 深圳迈瑞生物医疗电子股份有限公司 心电图计算机辅助诊断系统
CN1915166A (zh) * 2006-08-15 2007-02-21 武汉大学 一种基于平板运动心电图的心脏恢复性质检测的方法
CN101268938A (zh) * 2008-04-30 2008-09-24 董军 用于心电图识别与分类的方法和设备
CN101766484B (zh) * 2010-01-18 2011-09-07 董军 用于心电图识别与分类的方法和设备
CN102188240A (zh) * 2010-03-05 2011-09-21 华东师范大学 一种心电图数据采样方法及其装置
US20130041276A1 (en) * 2011-07-25 2013-02-14 Edan Instruments, Inc. METHOD AND SYSTEM FOR AUTOMATICALLY DETECTING AND ANALYZING PEDIATRIC ECGs
CN102397067B (zh) * 2011-12-07 2014-12-03 康泰医学系统(秦皇岛)股份有限公司 一种运动负荷心电波形叠加分析方法及其系统
CN103083013B (zh) * 2013-01-18 2015-05-13 哈尔滨工业大学深圳研究生院 基于形态学与小波变换的心电信号qrs复波检测方法
CN104398257A (zh) * 2014-12-08 2015-03-11 蚌埠医学院 一种心电波形周期性统计分析方法
CN105640545B (zh) * 2015-12-31 2018-12-14 深圳先进技术研究院 一种胎儿心电信号提取方法及装置
CN110226919B (zh) * 2019-06-26 2022-05-03 广州视源电子科技股份有限公司 心电信号类型检测方法、装置、计算机设备及存储介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104586383A (zh) * 2015-01-12 2015-05-06 北京世纪今科医疗器械有限公司 一种心电波形分类方法及装置
US20190167143A1 (en) * 2015-10-27 2019-06-06 Cardiologs Technologies Sas Electrocardiogram processing system for delineation and classification
CN108024717A (zh) * 2015-10-29 2018-05-11 美多尼克监测公司 确定室性期前收缩(pvc)类型和负荷的系统和方法
CN106108889A (zh) * 2016-07-20 2016-11-16 杨平 基于深度学习算法的心电图分类方法
CN110367969A (zh) * 2019-07-05 2019-10-25 复旦大学 一种改进的心电信号快速聚类分析方法
CN111419212A (zh) * 2020-02-27 2020-07-17 平安科技(深圳)有限公司 心电图数据处理的方法、装置、存储介质及计算机设备

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114073534A (zh) * 2021-11-25 2022-02-22 成都泰盟软件有限公司 一种心功能分析算法
CN114073534B (zh) * 2021-11-25 2023-05-26 成都泰盟软件有限公司 一种心功能分析算法
CN115486855A (zh) * 2022-09-15 2022-12-20 浙江好络维医疗技术有限公司 一种基于qrs波群不定次循环叠加的心电图心搏分类方法
CN115486855B (zh) * 2022-09-15 2024-05-03 浙江好络维医疗技术有限公司 一种基于qrs波群不定次循环叠加的心电图心搏分类方法
CN115778402A (zh) * 2022-12-02 2023-03-14 深圳华清心仪医疗电子有限公司 一种动态心电信号的伪差识别方法及系统

Also Published As

Publication number Publication date
CN111419212A (zh) 2020-07-17

Similar Documents

Publication Publication Date Title
WO2021169296A1 (zh) 心电图数据处理的方法、装置、计算机设备及存储介质
Kathirvel et al. An efficient R-peak detection based on new nonlinear transformation and first-order Gaussian differentiator
WO2021143401A1 (zh) 一种用于心电信号中p波和t波的检测方法和装置
EP2030565B1 (en) Systems for safe and remote outpatient ECG monitoring
Krasteva et al. QRS template matching for recognition of ventricular ectopic beats
EP3731748A1 (en) Detecting atrial fibrillation using short single-lead ecg recordings
US10856756B2 (en) Time-frequency analysis of electrocardiograms
CN101065058A (zh) 使用部分状态空间重构监视生理活动
CN111358459A (zh) 一种心律失常识别方法、装置、设备和存储介质
CN110432895B (zh) 训练数据处理、心电波形检测方法及电子设备
Velic et al. Computer aided ECG analysis—State of the art and upcoming challenges
US12059261B2 (en) Apparatus and method for electrocardiogram (ECG) signal analysis and heart block detection
WO2015171804A1 (en) Detecting artifacts in a signal
CN111839498A (zh) 心电图分析装置及其心电图分析方法
Chouhan et al. Delineation of QRS-complex, P and T-wave in 12-lead ECG
CN110826631A (zh) 心率失常疾病的分类方法、装置、存储介质和处理器
CN111839494A (zh) 一种心率监测方法及系统
Khaustov et al. An open-source standard T-wave alternans detector for benchmarking
CN114580477B (zh) 一种基于多时间序列融合的可穿戴动态呼吸率估测系统
CN113288169B (zh) 一种心电波形信号的波形识别方法、装置及设备
Gliner et al. Identification of features for machine learning analysis for automatic arrhythmogenic event classification
CN112438735B (zh) 一种心电图p波检测方法、分析装置以及存储介质
Filatova et al. Method of automatic determination of the heart’s electrical axis in cardiological decision support systems
US9192315B2 (en) Patient signal analysis and characterization based on late potentials
Villa et al. Variational mode decomposition features for heartbeat classification

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 23/01/2023)

122 Ep: pct application non-entry in european phase

Ref document number: 20922344

Country of ref document: EP

Kind code of ref document: A1