Electrocardiogram analysis method and device

Abstract

The invention relates to the technical field of electrocardiogram analysis, in particular to an electrocardiogram analysis method and device.The method comprises the following steps that S1, comprehensive information collection is conducted, specifically, an electrocardiogram, heart rate variability and blood pressure of a patient are collected in real time through a comprehensive collection module, and preliminary data of the patient are obtained; s2, preliminary analysis and prediction: according to the preliminary data, a comprehensive analysis module compares the electrocardiogram data of the healthy person with the electrocardiogram of the patient, and analyzes the image data difference; the comprehensive analysis module preliminarily evaluates the heart health condition of the patient according to the image data difference; predicting a potential cardiac problem of the patient; s3, recording and learning ST-segment abnormal signals: establishing a deep learning module based on a convolutional neural network algorithm, recording and storing the electrocardiogram of the patient collected each time by using the deep learning module, and learning and judging ST-segment abnormal signals in different electrocardiograms; the system is complete in logic and comprehensive in function, and can effectively assist doctors in electrocardiogram analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of electrocardiogram analysis, and in particular to an electrocardiogram analysis method and device. Background Art

[0002] Electrocardiogram reflects the health of human heart and is an important basis for clinical diagnosis of cardiovascular diseases. With the rapid growth of the number of electrocardiograms, the demand for computer-assisted electrocardiogram analysis has become more urgent. Automatic electrocardiogram classification, as an indispensable technical means to achieve computer-assisted electrocardiogram analysis, has important medical value.

[0003] The electrocardiogram analysis method in the prior art mainly relies on the doctor's experience for interpretation, which is not only time-consuming and labor-intensive, but also has the risk of excessive subjectivity and misdiagnosis, resulting in poor results of the electrocardiogram analysis method in the prior art in actual clinical application.

[0004] In summary, how to solve the problem of excessive subjectivity and risk of misdiagnosis in the existing electrocardiogram analysis methods has become a difficult problem that needs to be solved urgently in the field. Therefore, it is necessary to propose an electrocardiogram analysis method and device that can automatically analyze the electrocardiogram and provide doctors with corresponding analysis references to reduce subjectivity and the risk of misdiagnosis. Summary of the invention

[0005] To solve the above problems, the present invention provides an electrocardiogram analysis method and device, which collects and automatically analyzes the patient's electrocardiogram through the design of a deep learning module and a comprehensive analysis module, and provides corresponding reference diagnostic information to doctors based on the analysis results, thereby reducing the doctor's subjectivity and greatly reducing the risk of misdiagnosis.

[0006] In order to achieve the above object, the technical solution of the present invention is as follows: an electrocardiogram analysis method, comprising the following steps:

[0007] S1, collection of comprehensive information: The comprehensive collection module collects the patient's clinical information, electrocardiogram, heart rate variability and blood pressure in real time to obtain the patient's preliminary data; clinical information includes the patient's age, gender, clinical symptoms and past medical history.

[0008] The comprehensive acquisition module is used to collect the patient's preliminary data and transmit the preliminary data to the comprehensive analysis module and deep learning module.

[0009] S2, preliminary analysis and prediction: the preliminary data collected by the comprehensive acquisition module is received through the comprehensive analysis module, and the normal electrocardiogram image is recorded in the comprehensive analysis module; the comprehensive analysis module compares the normal electrocardiogram image with the electrocardiogram image of the patient collected by the comprehensive acquisition module, and analyzes the abnormalities in the electrocardiogram image of the patient.

[0010] The comprehensive analysis module conducts a preliminary assessment of the patient's heart health based on the differences in the electrocardiogram images; and combined with the patient's heart rate variability and blood pressure collected by the comprehensive acquisition module, it comprehensively analyzes the correlation between the patient's heart problems and heart rate variability and blood pressure, and uses this as a basis for predicting the patient's potential heart problems.

[0011] S3, recording and learning of abnormal ST segment signals: A deep learning module is established based on the convolutional neural network algorithm. The deep learning module records and stores the electrocardiogram images of the patient acquired by the comprehensive acquisition module each time, and learns to judge the abnormal ST segment signals in different electrocardiograms; abnormal ST segment signals include abnormal ST segment morphology changes and abnormal ST segment amplitude.

[0012] The deep learning module combines the clinical information collected by the comprehensive acquisition module and the potential heart problems of the patient preliminarily analyzed by the comprehensive analysis module to determine the patient's heart problems represented by different ST segment abnormal signals.

[0013] S4, analysis of abnormal ST segment signals: The deep learning module determines the isoelectric lines based on the patient's electrocardiogram, and judges the changes in ST segment morphology and measures the changes in ST segment amplitude in real time based on the isoelectric lines; the deep learning module judges the patient's possible heart problems based on the specific morphologies of different ST segments analyzed, and sends corresponding reference diagnostic information to the doctor's mobile phone.

[0014] The deep learning module updates itself in real time based on the accuracy of the reference diagnostic information; the doctor confirms or modifies the reference diagnostic information. If the reference diagnostic information is modified by the doctor, the deep learning module learns the modified content and updates the reference diagnostic information.

[0015] S5. Use of cloud computing platform: A cloud computing platform is constructed based on big data analysis technology. The cloud computing platform is used to receive ECG data input by users. The cloud computing platform analyzes and stores the input ECG data through a deep learning module. Users input abnormal ST segment signals and query corresponding ECG data through the cloud computing platform. Big data analysis technology selects corresponding ECG data for user reference based on the corresponding abnormal ST segment signals, and provides users with PDF version of ECG images and WORD version of reference diagnostic information.

[0016] Furthermore, the convolutional neural network algorithm is expressed as the following formula:

[0017] Y ij =∑m∑nX(i+m)(j+n)·Wmn+b

[0018] in:

[0019] Y ijIndicates the output of the patient's ST segment abnormality signal.

[0020] X is the input patient's ECG.

[0021] W is the convolution kernel, which is used to perform a sliding window operation on the input patient ECG.

[0022] b is the bias term, which is used to adjust the output patient ST segment abnormality signal.

[0023] m and n are the size parameters of the convolution kernel.

[0024] Furthermore, in S4, the isoelectric line is a relatively stable baseline on the electrocardiogram without obvious fluctuations.

[0025] Furthermore, in S4, the specific morphologies of the ST segment include arched upward, arched downward, horizontal, upsloping and downsloping; the abnormal morphological changes of the ST segment include ST segment elevation, depression, prolongation and shortening.

[0026] Furthermore, in S3, the deep learning module for different ST segment abnormal signals issues corresponding reference diagnostic information as follows:

[0027] ST segment elevation: Arched upward type elevation: highly suspected acute myocardial infarction; Arched downward type elevation: suspected pericarditis.

[0028] ST segment depression: horizontal depression: suspected myocardial ischemia; downward sloping depression: suspected myocardial ischemia.

[0029] ST segment prolongation or shortening: Suspected abnormality of cardiac electrophysiological activity.

[0030] Furthermore, in S3, the deep learning module sets corresponding abnormal thresholds for ST segment amplitude changes, and the abnormal thresholds are set to ST segment elevation > 0.1 mV and ST segment depression ≥ 0.05 mV.

[0031] Furthermore, in S5, a remote interaction module is provided in the cloud computing platform, and the remote interaction module is used for doctors and patients to conduct remote consultations based on the patient's condition.

[0032] The following principles and beneficial effects are achieved by adopting the above scheme:

[0033] Basic principle: Use the comprehensive acquisition module to comprehensively collect the patient's three key physiological indicators: electrocardiogram, heart rate variability and blood pressure, as preliminary data for analysis.

[0034] Through the comprehensive analysis module, the patient's electrocardiogram is compared with the electrocardiogram data of healthy people, the image data differences are analyzed, and combined with heart rate variability and blood pressure data, the patient's heart health is comprehensively evaluated to predict potential heart problems.

[0035] A deep learning module is established based on the convolutional neural network algorithm; the deep learning module analyzes the electrocardiogram in real time and, combined with clinical information, determines the specific morphology and amplitude changes of the ST segment, thereby determining the patient's heart problems.

[0036] When the patient's electrocardiogram ST segment amplitude is abnormal, the deep learning module will also capture and analyze the abnormal ST segment amplitude. After capturing the abnormal ST segment amplitude, the deep learning module will further analyze this abnormal feature and combine it with the clinical knowledge and rules learned in the past to determine the patient's possible heart problems. For example, if the ST segment elevation exceeds the set threshold of 0.1mV, the deep learning module will issue a reference diagnosis information of "highly suspected acute myocardial infarction" to the doctor. Similarly, if the ST segment depression exceeds another set threshold of 0.05mV, the deep learning module will issue a reference diagnosis information of "highly suspected myocardial ischemia" to the doctor.

[0037] In addition, the deep learning module will also consider other relevant factors, such as the patient's age, gender, medical history, etc., to further improve the accuracy of the diagnosis.

[0038] During the analysis process, the deep learning module will also update its ECG analysis and comparison capabilities and reference diagnostic information analysis and output capabilities in real time based on the doctor's confirmation of the reference diagnostic information, thereby performing deep learning and achieving self-update.

[0039] Ultimately, when analyzing a patient's electrocardiogram, doctors can combine the comprehensive analysis module's predictive analysis to make a preliminary judgment on the patient's potential heart disease, and then combine the patient's real-time electrocardiogram ST segment abnormality data and reference diagnostic information provided by the deep learning module to make a more accurate judgment and analysis, thereby reducing the doctor's subjectivity and the risk of misdiagnosis.

[0040] During the diagnosis and treatment of patients, doctors can also upload various ECG images and data for analysis and storage through the cloud computing platform. They can also input abnormal ST segment signal data or heart problems through the cloud computing platform and search for related ECG images and data for reference. Doctors can conduct technical exchanges and remote consultations through the cloud computing platform, and patients can also transmit their own ECG images and data as well as diagnostic information to doctors through the cloud computing platform for remote consultation.

[0041] The cloud computing platform will also provide users with PDF versions of electrocardiogram images and WORD versions of reference diagnostic information, which are convenient for users to download and use to improve medical services for patients.

[0042] The above scheme has the following beneficial effects:

[0043] 1. The method of the present invention can more comprehensively and accurately analyze the patient's electrocardiogram, heart rate variability, blood pressure and other physiological indicators through the combination of the comprehensive analysis module and the deep learning module, timely discover potential heart problems, and reduce the risk of misdiagnosis and missed diagnosis. Through the automated and intelligent analysis of the comprehensive analysis module and the deep learning module, the subjective judgment of the doctor in the diagnosis process is reduced, making the diagnosis result more objective and reliable. Doctors can combine the dual judgment of the comprehensive analysis module and the deep learning module to conduct a more comprehensive evaluation and reduce the impact of personal experience differences on the diagnosis results.

[0044] 2. The comprehensive acquisition module is used to monitor the patient's electrocardiogram, heart rate variability and blood pressure in real time, which can promptly detect the patient's heart health problems. The comprehensive analysis module can then provide the patient with timely prevention and treatment suggestions, which will help the patient intervene early and improve their health.

[0045] 3. The cloud computing platform provides doctors with a convenient way to upload, store and analyze ECG images and data, promoting the sharing and collaboration of medical resources. Doctors can conduct technical exchanges and remote consultations through the platform, breaking geographical restrictions and improving the accessibility and efficiency of medical services. Patients can also remotely and conveniently obtain their own ECG images and diagnostic information through the cloud computing platform, conduct self-management and tracking, and improve health awareness. Patients can also conduct remote consultations with doctors through the platform, saving time and transportation costs and improving the medical experience.

[0046] Furthermore, an electrocardiogram analysis device includes a controller and a monitoring suit, wherein an electrocardiogram is fixedly connected to the chest of the monitoring suit, and a blood pressure monitor is fixedly connected to the wrist of the monitoring suit near the electrocardiogram. Both the electrocardiogram and the blood pressure monitor are electrically connected to the controller, and the controller is used to control the opening and closing of the electrocardiogram and the blood pressure monitor.

[0047] Beneficial effects: The device installs an electrocardiogram (ECG) on the chest of the monitoring suit to intelligently monitor the patient's heart condition, combines it with a blood pressure monitor to detect the patient's blood pressure, and then uses a comprehensive analysis module to measure the variability of the RR interval of continuous sinus beats to analyze the patient's heart rate variability, thereby achieving a comprehensive analysis of the patient's physical condition.

[0048] The design of the monitoring suit makes it convenient to detect various types of data on patients, so that patients do not need to carry or install additional equipment when taking electrocardiogram and blood pressure measurements, which greatly improves the convenience of use.

[0049] At the same time, this integrated design also makes it easier for patients to move around. Compared with fixed designs, it allows patients to have more freedom, enhances the user experience, and reduces the risk of device loss.

[0050] It also reduces the possibility of wire entanglement due to too many devices, thereby improving the accuracy of monitoring results and improving the comfort of patient use.

[0051] Furthermore, an alarm is fixedly connected to the monitoring suit.

[0052] Beneficial effects: When the deep learning module detects abnormal blood pressure, abnormal electrocardiogram or abnormal heart rate variability of the patient, the alarm will automatically sound to remind medical staff to immediately treat the patient; this design effectively reduces the adverse consequences caused by delayed treatment and improves the safety and survival rate of patients; compared with traditional manual monitoring, it can save time and effort and has better stability.

[0053] Furthermore, a plurality of ventilation holes are opened on the monitoring suit.

[0054] Beneficial effects: Several vents can help patients dissipate heat and improve comfort; prevent patients from wearing monitoring clothing for a long time and having a high body temperature, which can cause discomfort.

[0055] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 Flow chart of the electrocardiogram analysis method in an embodiment of the present invention.

[0057] Figure 2 Schematic diagram of the cloud computing platform function in the electrocardiogram analysis method in an embodiment of the present invention.

[0058] Figure 3 It is an axonometric diagram of the electrocardiogram analysis device according to an embodiment of the present invention.

[0059] The reference numerals in the drawings of the specification include: 1. monitoring suit; 2. electrocardiograph; 3. blood pressure monitor; 4. alarm; 5. ventilation hole. DETAILED DESCRIPTION

[0060] The following is further described in detail through specific implementation methods:

[0061] Embodiment 1:

[0062] As attached Figure 1-2 As shown, an electrocardiogram analysis method comprises the following steps:

[0063] S1, collection of comprehensive information: the comprehensive collection module is used to collect the patient's clinical information, electrocardiogram, heart rate variability and blood pressure in real time to obtain the patient's preliminary data; clinical information includes the patient's age, gender, clinical symptoms and past medical history; the comprehensive collection module is used to collect the patient's preliminary data and transmit the preliminary data.

[0064] S2, preliminary analysis and prediction: using the comprehensive analysis module to receive the preliminary data collected by the comprehensive acquisition module, the comprehensive analysis module has a normal electrocardiogram image recorded in it; the comprehensive analysis module compares the normal electrocardiogram image with the electrocardiogram image of the patient collected by the comprehensive acquisition module, and analyzes the abnormalities in the electrocardiogram image of the patient.

[0065] The comprehensive analysis module makes a preliminary assessment of the patient's heart health based on the differences in the electrocardiogram images; and combines the patient's heart rate variability and blood pressure collected by the comprehensive acquisition module to comprehensively analyze the correlation between the patient's heart problems and heart rate variability and blood pressure, and use this as a basis to predict the patient's potential heart problems;

[0066] S3, recording and learning of abnormal ST segment signals: A deep learning module is established based on the convolutional neural network algorithm. The deep learning module records and stores the electrocardiogram images of the patient acquired by the comprehensive acquisition module each time, and learns to judge the abnormal ST segment signals in different electrocardiograms; abnormal ST segment signals include abnormal ST segment morphology changes and abnormal ST segment amplitude.

[0067] The deep learning module combines the clinical information collected by the comprehensive acquisition module with the existing heart problems of the patient and the predicted potential heart problems of the patient preliminarily analyzed by the comprehensive analysis module to determine the heart problems of the patient represented by different ST segment abnormal signals.

[0068] S4, analysis of abnormal ST segment signals: The deep learning module determines the isoelectric lines based on the patient's electrocardiogram, and judges the changes in ST segment morphology and measures the changes in ST segment amplitude in real time based on the isoelectric lines; the deep learning module judges the patient's possible heart problems based on the specific morphologies of different ST segments analyzed, and sends corresponding reference diagnostic information to the doctor's mobile phone.

[0069] The specific morphology of the ST segment includes arched upward, arched downward, horizontal, upsloping and downsloping; the abnormal morphological changes of the ST segment include ST segment elevation, depression, extension and shortening. The threshold of abnormal ST segment amplitude is set as ST segment elevation> 0.1mV and ST segment depression ≥ 0.05mV.

[0070] For different ST segment abnormal signals, the deep learning module will issue corresponding reference diagnostic information as follows:

[0071] ST segment elevation: Arched upward type elevation: highly suspected acute myocardial infarction; Arched downward type elevation: suspected pericarditis.

[0072] ST segment depression: horizontal depression: suspected myocardial ischemia; downward sloping depression: suspected myocardial ischemia.

[0073] ST segment prolongation or shortening: Suspected abnormality of cardiac electrophysiological activity.

[0074] The deep learning module updates itself in real time based on the accuracy of the reference diagnostic information; the doctor confirms or modifies the reference diagnostic information. If the reference diagnostic information is modified by the doctor, the deep learning module learns the modified content and updates the reference diagnostic information.

[0075] S5. Use of cloud computing platform: A cloud computing platform is constructed based on big data analysis technology. The cloud computing platform is used to receive ECG data input by users. The cloud computing platform analyzes and stores the input ECG data through a deep learning module. Users input abnormal ST segment signals and query corresponding ECG data through the cloud computing platform. Big data analysis technology selects corresponding ECG data for user reference based on the corresponding abnormal ST segment signals, and provides users with PDF version of ECG images and WORD version of reference diagnostic information.

[0076] The convolutional neural network algorithm is expressed as the following formula:

[0077] Y ij =∑m∑nX(i+m)(j+n)·Wmn+b

[0078] Where: Y ij represents the output abnormal ST segment signal of the patient; X is the input electrocardiogram of the patient; W is the convolution kernel, which is used to perform a sliding window operation on the input electrocardiogram of the patient; b is the bias term, which is used to adjust the output abnormal ST segment signal of the patient; m and n are the size parameters of the convolution kernel, which determine the size of the convolution kernel.

[0079] The specific implementation process is as follows: Use the comprehensive acquisition module to comprehensively collect the patient's electrocardiogram, heart rate variability and blood pressure three key physiological indicators as preliminary data for analysis.

[0080] Through the comprehensive analysis module, the patient's electrocardiogram is compared with the electrocardiogram data of healthy people, the image data differences are analyzed, and combined with heart rate variability and blood pressure data, the patient's heart health is comprehensively evaluated to predict potential heart problems.

[0081] For example, if the patient's heart rate variability is reduced, and the absolute values ​​of the peak and trough waveforms of the electrocardiogram image are smaller than normal values, the comprehensive analysis module will determine that the patient's cardiac autonomic nervous regulation function is impaired, and preliminarily analyze the patient's heart may have potential myocardial ischemia, heart failure, diabetic cardiac autonomic neuropathy or arrhythmia and other diseases, and provide patients with corresponding prevention and examination suggestions, and provide doctors with corresponding diagnosis and treatment suggestions.

[0082] A deep learning module is established based on the convolutional neural network algorithm; the deep learning module analyzes the electrocardiogram in real time and, combined with clinical information, determines the specific morphology and amplitude changes of the ST segment, thereby determining the patient's heart problems.

[0083] For example, the deep learning module analyzes the patient's electrocardiogram through a neural network algorithm and finds that the ST segment in the patient's electrocardiogram is elevated in a hunched-up manner and the patient suddenly develops symptoms of myocardial infarction. It will then determine that the patient has acute myocardial infarction. At the same time, it will send a reference diagnosis message to the doctor: "Acute myocardial infarction is highly suspected." and remind the doctor to treat the patient immediately.

[0084] When the patient's electrocardiogram ST segment amplitude is abnormal, the deep learning module will also capture and analyze the abnormal ST segment amplitude. After capturing the abnormal ST segment amplitude, the deep learning module will further analyze this abnormal feature and combine it with the clinical knowledge and rules learned in the past to determine the patient's possible heart problems. For example, if the ST segment elevation exceeds the set threshold of 0.1mV, the deep learning module will issue a reference diagnosis information of "highly suspected acute myocardial infarction" to the doctor. Similarly, if the ST segment depression exceeds another set threshold of 0.05mV, the deep learning module will issue a reference diagnosis information of "highly suspected myocardial ischemia" to the doctor.

[0085] In addition, the deep learning module will also consider other relevant factors, such as the patient's age, gender, medical history, etc., to further improve the accuracy of the diagnosis.

[0086] During the analysis process, the deep learning module will also update its ECG analysis and comparison capabilities and reference diagnostic information analysis and output capabilities in real time based on the doctor's confirmation of the reference diagnostic information, thereby performing deep learning and achieving self-update.

[0087] Ultimately, when analyzing a patient's electrocardiogram, doctors can combine the comprehensive analysis module's predictive analysis to make a preliminary judgment on the patient's potential heart disease, and then combine the patient's real-time electrocardiogram ST segment abnormality data and reference diagnostic information provided by the deep learning module to make a more accurate judgment and analysis, thereby reducing the doctor's subjectivity and the risk of misdiagnosis.

[0088] During the diagnosis and treatment of patients, doctors can also upload various ECG images and data for analysis and storage through the cloud computing platform. They can also input abnormal ST segment signal data or heart problems through the cloud computing platform and search for related ECG images and data for reference. Doctors can conduct technical exchanges and remote consultations through the cloud computing platform, and patients can also transmit their own ECG images and data as well as diagnostic information to doctors through the cloud computing platform for remote consultation.

[0089] The cloud computing platform will also provide users with PDF versions of electrocardiogram images and WORD versions of reference diagnostic information, which are convenient for users to download and use to improve medical services for patients.

[0090] The method of the present invention can more comprehensively and accurately analyze the patient's electrocardiogram, heart rate variability, blood pressure and other physiological indicators through the combination of the comprehensive analysis module and the deep learning module, timely discover potential heart problems, and reduce the risk of misdiagnosis and missed diagnosis. Through the automated and intelligent analysis of the comprehensive analysis module and the deep learning module, the subjective judgment of the doctor in the diagnosis process is reduced, making the diagnosis result more objective and reliable. The doctor can combine the dual judgment of the comprehensive analysis module and the deep learning module to make a more comprehensive assessment and reduce the impact of personal experience differences on the diagnosis results.

[0091] The method of the present invention utilizes a comprehensive acquisition module to monitor the patient's electrocardiogram, heart rate variability and blood pressure in real time, which can promptly detect the patient's heart health problems, and then provide the patient with timely prevention and treatment suggestions through a comprehensive analysis module, which helps the patient to intervene early and improve their health status.

[0092] The design of the cloud computing platform provides doctors with a convenient way to upload, store and analyze ECG images and data, promoting the sharing and collaboration of medical resources. Doctors can conduct technical exchanges and remote consultations through the platform, breaking geographical restrictions and improving the accessibility and efficiency of medical services. Patients can also remotely and conveniently obtain their own ECG images and diagnostic information through the cloud computing platform, conduct self-management and tracking, and improve health awareness. Patients can also conduct remote consultations with doctors through the platform, saving time and transportation costs and improving the medical experience.

[0093] Embodiment 2:

[0094] As attached Figure 1-3 As shown, different from the above-mentioned embodiments, an electrocardiogram analysis device includes a controller and a monitoring suit 1, wherein an electrocardiogram 2 is fixedly bonded to the chest of the monitoring suit 1, and a blood pressure monitor 3 is fixedly bonded to the wrist of the monitoring suit 1 near the electrocardiogram 2. The electrocardiogram 2 and the blood pressure monitor 3 are both electrically connected to the controller, and the controller is used to control the opening and closing of the electrocardiogram 2 and the blood pressure monitor 3.

[0095] The specific implementation process is as follows: the patient puts on the monitoring suit 1, installs the electrocardiograph 2 on the chest, and aligns the blood pressure monitor 3 on the wrist; after putting on the suit, the electrocardiograph 2 will monitor the patient's heart condition in real time, the blood pressure monitor 3 will monitor the patient's blood pressure in real time, and the comprehensive analysis module will measure the variability of the RR interval of continuous sinus beats to analyze the patient's heart rate variability; the comprehensive analysis module will jointly analyze the patient's electrocardiogram, heart rate variability and blood pressure, so as to achieve a comprehensive analysis of the patient's physical condition.

[0096] Through the design of the monitoring suit 1, it is possible to conveniently detect various types of data on patients, so that patients do not need to carry or install additional equipment when taking electrocardiogram and blood pressure measurements, which greatly improves the convenience of use.

[0097] At the same time, this integrated design also makes it easier for patients to move around. Compared with fixed designs, it allows patients to have more freedom, enhances the user experience, and reduces the risk of device loss.

[0098] It also reduces the possibility of wire entanglement due to too many devices, thereby improving the accuracy of monitoring results and improving the comfort of patient use.

[0099] Embodiment 3:

[0100] As attached Figure 3 As shown, what is different from the above-mentioned embodiment is that an alarm 4 is fixedly bonded to the monitoring suit 1 , and the controller is used to sound an alarm on the alarm 4 .

[0101] The specific implementation process is as follows: when the deep learning module detects abnormal blood pressure, abnormal electrocardiogram or abnormal heart rate variability of the patient, the controller will control the alarm 4 to alarm immediately, reminding medical staff to treat the patient immediately; such a design effectively reduces the adverse consequences caused by delayed treatment and improves the safety and survival rate of patients; compared with traditional manual monitoring, it can save time and effort and has better stability.

[0102] Embodiment 4:

[0103] As attached Figure 3 As shown, different from the above-mentioned embodiment, a plurality of ventilation holes 5 are opened on the monitoring suit 1 .

[0104] The specific implementation process is as follows: when the patient wears the monitoring suit 1 for a long time, it is easy to cause a high body temperature, which may cause physical discomfort; a number of ventilation holes 5 make the monitoring suit 1 have better breathability, thereby helping the patient to dissipate heat and improve the patient's comfort.

[0105] Obviously, the above embodiments are merely examples for the purpose of clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the invention.

Claims

1. An electrocardiogram analysis method, characterized in that: The following steps are involved: S1, Collection of comprehensive information: The comprehensive collection module collects the patient's clinical information, electrocardiogram, heart rate variability and blood pressure in real time to obtain the patient's preliminary data; clinical information includes the patient's age, gender, clinical symptoms and past medical history; The comprehensive acquisition module is used to collect preliminary data of patients and transmit the preliminary data to the comprehensive analysis module and the deep learning module; S2, preliminary analysis and prediction: receiving preliminary data collected by the comprehensive acquisition module through the comprehensive analysis module, and recording a normal electrocardiogram image in the comprehensive analysis module; the comprehensive analysis module compares the normal electrocardiogram image with the electrocardiogram image of the patient collected by the comprehensive acquisition module, and analyzes the abnormalities in the electrocardiogram image of the patient; The comprehensive analysis module provides a preliminary assessment of the patient’s heart health based on the differences in the ECG images; Combined with the heart rate variability and blood pressure of the patient collected by the comprehensive acquisition module, the correlation between the patient's heart problem and heart rate variability and blood pressure is comprehensively analyzed, and the patient's potential heart problem is predicted based on this; S3, recording and learning of abnormal ST segment signals: A deep learning module is established based on the convolutional neural network algorithm. The deep learning module records and stores the electrocardiogram images of the patient collected by the comprehensive acquisition module each time, and learns to judge the abnormal ST segment signals in different electrocardiograms; abnormal ST segment signals include abnormal ST segment morphology changes and abnormal ST segment amplitude; The deep learning module combines the clinical information collected by the comprehensive acquisition module and the potential heart problems of the patient preliminarily analyzed by the comprehensive analysis module to determine the heart problems of the patient represented by different ST segment abnormal signals; S4, analysis of abnormal ST segment signals: the deep learning module determines the isoelectric line based on the patient's electrocardiogram, and judges the ST segment morphology changes and measures the ST segment amplitude changes in real time based on the isoelectric line; the deep learning module judges the patient's possible heart problems based on the specific morphology of the different ST segments analyzed, and sends the corresponding reference diagnosis information to the doctor's mobile phone; The deep learning module updates itself in real time according to the accuracy of the reference diagnostic information; the doctor confirms or modifies the reference diagnostic information. If the reference diagnostic information is modified by the doctor, the deep learning module learns the modified content and updates the reference diagnostic information; S5. Use of cloud computing platform: A cloud computing platform is constructed based on big data analysis technology. The cloud computing platform is used to receive ECG data input by users. The cloud computing platform analyzes and stores the input ECG data through a deep learning module. Users input abnormal ST segment signals and query corresponding ECG data through the cloud computing platform. Big data analysis technology selects corresponding ECG data for user reference based on the corresponding abnormal ST segment signals, and provides users with PDF version of ECG images and WORD version of reference diagnostic information.

2. The electrocardiogram analysis method according to claim 1, characterized in that: In S3, the convolutional neural network algorithm is expressed as the following formula: Y ij =∑m∑nX(i+m)(j+n)·Wmn+b in: Y ij Indicates the output of abnormal ST segment signal of the patient; X is the input patient's ECG; W is the convolution kernel, which is used to perform a sliding window operation on the input patient ECG; b is a bias term, which is used to adjust the output patient ST segment abnormality signal; m and n are the size parameters of the convolution kernel.

3. The electrocardiogram analysis method according to claim 2, characterized in that: In S4, the isoelectric line is a relatively stable baseline on the electrocardiogram without obvious fluctuations.

4. The electrocardiogram analysis method according to claim 3, characterized in that: In S4, the specific morphologies of the ST segment include arched upward, arched downward, horizontal, upsloping and downsloping; abnormal changes in ST segment morphology include ST segment elevation, depression, prolongation and shortening.

5. The electrocardiogram analysis method according to claim 4, characterized in that: In S3, for different ST segment abnormal signals, the deep learning module issues the corresponding reference diagnostic information as follows: ST segment elevation: Arched upward elevation: highly suspected acute myocardial infarction; Arched downward elevation: suspected pericarditis; ST segment depression: horizontal depression: suspected myocardial ischemia; downsloping depression: suspected myocardial ischemia; ST segment prolongation or shortening: Suspected abnormality of cardiac electrophysiological activity.

6. The electrocardiogram analysis method according to claim 5, characterized in that: In S3, the deep learning module sets corresponding abnormal thresholds for ST segment amplitude changes, and the abnormal thresholds are set as ST segment elevation > 0.1mV and ST segment depression ≥ 0.05mV.

7. The electrocardiogram analysis method according to claim 6, characterized in that: In S5, a remote interaction module is provided in the cloud computing platform, and the remote interaction module is used for doctors and patients to conduct remote consultations based on the patient's condition.

8. An electrocardiogram analysis device, according to the electrocardiogram analysis method according to any one of claims 1 to 7, characterized in that: The monitoring suit (1) comprises a controller and a monitoring suit (1), wherein the chest of the monitoring suit (1) is fixedly connected to an electrocardiogram (2), and the wrist of the monitoring suit (1) close to the electrocardiogram (2) is fixedly connected to a blood pressure monitor (3), and both the electrocardiogram (2) and the blood pressure monitor (3) are electrically connected to the controller, and the controller is used to control the opening and closing of the electrocardiogram (2) and the blood pressure monitor (3).

9. The electrocardiogram analysis device according to claim 8, characterized in that: An alarm (4) is fixedly connected to the monitoring suit (1).

10. The electrocardiogram analysis device according to claim 9, characterized in that: The monitoring suit (1) is provided with a plurality of ventilation holes (5).

Images