CN108577830B - User-oriented physical sign information dynamic monitoring method and system - Google Patents

Abstract

Translated fromChinese

本发明实施例涉及一种面向用户的体征信息动态监护方法和动态监护设备，方法包括：动态监护设备接收用户输入或者服务器下发的监测基准数据；对被测对象进行监护数据采集得到被测对象的体征监护数据，其中至少包括心电图数据；对心电图数据进行波群特征识别得到心电图数据的特征信号，根据特征信号对心电图数据进行心搏分类，结合心电图基本规律参考数据得到心搏分类信息，并生成心电图事件数据；根据心电图事件数据实时确定心电图事件信息，并确定心电图事件信息是否为心电异常事件信息；当为心电异常事件信息时生成并通过动态监护设备输出报警信息；并且基于不同的工作模式进行报警信息的上报或本地存储。

The embodiments of the present invention relate to a user-oriented dynamic monitoring method and dynamic monitoring device for physical sign information. The method includes: the dynamic monitoring device receives monitoring benchmark data input by a user or sent by a server; collecting monitoring data on a measured object to obtain the measured object The monitoring data of the signs, which includes at least the ECG data; the characteristic signals of the ECG data are obtained by the complex feature identification of the ECG data, the heartbeat classification is performed on the ECG data according to the characteristic signals, and the heartbeat classification information is obtained in combination with the reference data of the basic rules of the ECG. Generate ECG event data; determine ECG event information in real time according to ECG event data, and determine whether the ECG event information is ECG abnormal event information; generate and output alarm information through dynamic monitoring equipment when it is ECG abnormal event information; In the working mode, the alarm information is reported or stored locally.

Description

User-oriented physical sign information dynamic monitoring method and system

Technical Field

The invention relates to the technical field of data processing, in particular to a user-oriented physical sign information dynamic monitoring method and a user-oriented physical sign information dynamic monitoring system.

Background

A multi-parameter bedside monitor is a common clinical medical device. The monitoring device is characterized by being provided with a plurality of groups of sensors and being capable of monitoring vital sign indexes such as electrocardio, blood pressure, blood oxygen, pulse, respiration, body temperature and the like at the same time. In ward monitoring, the multi-parameter monitor can become an important reference of doctors for patient monitoring, so that doctors can find problems of patients in time and deal with the problems in time, and the life safety of the patients is ensured. Clinical applications of monitors can be found in: during and after surgery, trauma care, coronary heart disease, critically ill patients, neonates, premature infants, hyperbaric oxygen chambers, delivery rooms, and the like.

However, there are many cases where there is a need for physical sign monitoring in non-hospitalized people, such as patients who are oriented to home rest, people who are weak and need real-time physical sign monitoring, and social security and welfare agency support such as nursing homes/nursing homes. Currently, no corresponding equipment is available in the industry to meet the use requirement of the equipment.

Therefore, how to perform effective physical sign monitoring for non-hospitalized people and provide more effective medical guarantee service for users based on physical sign monitoring are the difficult problems and challenges to be solved by the invention.

Disclosure of Invention

The invention aims to provide a user-oriented dynamic monitoring method and a dynamic monitoring system for physical sign information, which realize data interaction through a wired or wireless communication technology, can timely find abnormality and generate alarm through completely and quickly automatically analyzing and recording electrocardiogram data, can automatically send alarm information to dynamic monitoring equipment when the processing is finished in a server, and can automatically report or manually trigger and report the alarm information to the server through local output when the processing is finished in the dynamic monitoring equipment. Further, the server can perform information distribution processing including terminal devices distributed to medical institutions or to associated users of monitored persons based on the alarm information. The dynamic monitoring device and the server also support the recording and storage of monitored data and can be configured in different modes for recording all data and only heterogeneous data.

In order to achieve the above object, a first aspect of the embodiments of the present invention provides a user-oriented method for dynamically monitoring physical sign information, including:

the dynamic monitoring equipment receives monitoring reference data input by a user or issued by a server; the monitoring reference data comprises information of a measured object and information of an electrocardio abnormal event;

the dynamic monitoring equipment acquires monitoring data of a measured object to obtain physical sign monitoring data of the measured object; the physical sign monitoring data at least comprises electrocardiogram data;

performing wave group feature identification on the electrocardiogram data to obtain feature signals of the electrocardiogram data, performing heart beat classification on the electrocardiogram data according to the feature signals, obtaining heart beat classification information by combining with electrocardiogram basic rule reference data, and generating electrocardiogram event data; the electrocardiographic event data comprises device ID information of the dynamic monitoring device;

determining corresponding electrocardiogram event information in real time according to the electrocardiogram event data, and determining whether the electrocardiogram event information is the electrocardiogram abnormal event information;

when the information is the abnormal event information of the electrocardio, generating and outputting alarm information through the dynamic monitoring equipment; the alarm information comprises the equipment ID information, the information of the measured object and the information of the electrocardio abnormal event;

determining a working mode of the dynamic monitoring equipment; the working mode comprises a local processing mode and a background processing mode;

when the electrocardiogram event data is in a local processing mode, the dynamic monitoring equipment records and stores the electrocardiogram event data and the alarm information;

when the monitoring device is in a background processing mode, the server determines the user ID of the responsible user corresponding to the detected object information according to the alarm information or receives an alarm signal triggered by the dynamic monitoring device based on the alarm information, generates notification information and sends the notification information to the user equipment of the responsible user and/or the user equipment of a preset association mechanism; wherein the notification information includes the alarm information and a user ID of the responsible user.

Preferably, before the dynamic monitoring device receives the user input or the monitoring reference data issued by the server, the method further includes:

and determining the measured object information of the measured object.

Preferably, the method further comprises:

the server receives feedback information sent by user equipment of the responsible user and sends the feedback information to the dynamic monitoring equipment according to the equipment ID information;

the dynamic monitoring device outputs the feedback information.

Preferably, the step of generating notification information and sending the notification information to the user equipment of the responsible user and/or the user equipment of the preset association mechanism specifically includes:

the server determines the position information of the dynamic monitoring equipment according to the information of the measured object and/or the ID information of the equipment;

generating the notification information according to the position information and the alarm information;

and sending the notification information to the user equipment of the responsible user and/or the user equipment of a preset association mechanism according to the user ID of the responsible user.

Preferably, the method further comprises:

the dynamic monitoring equipment sends electrocardiogram data in a preset time period to the server according to preset time information;

and the server generates monitoring report data of the measured object in a preset time period according to the electrocardiogram data.

Preferably, the method further comprises:

the dynamic monitoring equipment receives an event recording signal triggered by a user, monitors and generates alarm event recording information;

and sending the alarm event record information to the user equipment of the responsible user and/or the user equipment of a preset association mechanism.

Preferably, the vital signs monitoring data further comprises: one or more of pulse data, blood pressure data, respiration data, blood oxygen saturation data, and body temperature data.

Further preferably, the method further comprises:

and the dynamic monitoring equipment determines that one or more of the pulse data, the blood pressure data, the respiration data, the blood oxygen saturation data and the body temperature data has abnormal data exceeding a corresponding set threshold, generates sign monitoring abnormal event information by combining the electrocardiogram event data and sends the sign monitoring abnormal event information to the server.

Preferably, the performing the wave group feature identification on the electrocardiogram data to obtain the feature signal of the electrocardiogram data, performing heart beat classification on the electrocardiogram data according to the feature signal, obtaining heart beat classification information by combining with the electrocardiogram basic rule reference data, and generating the electrocardiogram event data specifically includes:

resampling the data format of the electrocardiogram data to convert the data format of the electrocardiogram data into a preset standard data format, and performing first filtering processing on the converted electrocardiogram data with the preset standard data format;

performing heart beat detection processing on the electrocardiogram data after the first filtering processing, and identifying a plurality of heart beat data included in the electrocardiogram data, wherein each heart beat data corresponds to one heart beat period and comprises corresponding P wave, QRS wave group, amplitude of T wave and start-stop time data;

determining a detection confidence for each heartbeat from the heartbeat data;

performing interference identification on the heartbeat data according to an interference identification two-classification model to obtain whether interference noise exists in the heartbeat data and a probability value for judging the interference noise;

determining validity of the heartbeat data according to the detection confidence, and generating heartbeat time series data based on the interference recognition result and the time rule combination according to the lead parameters of the heartbeat data and the heartbeat data determined to be valid; generating heartbeat analysis data from the heartbeat time series data;

extracting and analyzing the characteristics of amplitude and time characterization data of the heart beat analysis data according to a heart beat classification model to obtain primary classification information of the heart beat analysis data;

inputting heartbeat analysis data of a specific heartbeat in the primary classification information result into an ST segment and T wave change model for identification, and determining ST segment and T wave evaluation information;

according to the heart beat time sequence data, performing P wave and T wave feature detection on the heart beat analysis data, and determining detailed feature information of the P wave and the T wave in each heart beat, wherein the detailed feature information comprises data of amplitude, direction, form and start-stop time;

carrying out secondary classification processing on the heart beat analysis data under the primary classification information according to the electrocardiogram basic rule reference data, the detailed characteristic information of the P wave and the T wave and the ST segment and T wave evaluation information to obtain heart beat classification information;

and analyzing and matching the heart beat classification information to generate the electrocardiogram event data.

The user-oriented physical sign information dynamic monitoring method provided by the embodiment of the invention adopts data preprocessing, heart beat characteristic detection, interference signal detection and heart beat classification and lead combination based on a deep learning method, heart beat audit, electrocardiogram event and parameter analysis and calculation, and finally automatic electrocardiogram detection analysis of a complete and rapid process for automatically outputting electrocardiogram event result data, and can output alarm based on the electrocardiogram detection analysis result, or generate alarm by combining blood pressure, blood oxygen, pulse, respiration and body temperature data, and perform information distribution processing based on the alarm information, wherein the information distribution processing comprises terminal equipment distributed to a medical institution or to an associated user of a monitored person, so that the monitored person can obtain effective and timely medical rescue service. The user-oriented dynamic monitoring method for the physical sign information is used for effectively monitoring physical signs for non-hospitalized people and providing more effective medical guarantee service for users based on physical sign monitoring.

A second aspect of an embodiment of the present invention provides a dynamic monitoring system, which includes one or more dynamic monitoring devices described in the first aspect above and a server.

Drawings

Fig. 1 is a flowchart of a user-oriented method for dynamically monitoring physical sign information according to an embodiment of the present invention;

FIG. 2 is a flow chart of a method for processing ECG data according to an embodiment of the present invention;

FIG. 3 is a diagram of an interference recognition two-class model according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of a heart beat classification model provided by an embodiment of the invention;

FIG. 5 is a schematic diagram of an ST-band and T-wave alteration model provided by an embodiment of the present invention;

fig. 6 is a schematic structural diagram of a dynamic monitoring system according to an embodiment of the present invention.

Detailed Description

The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.

The dynamic monitoring system provided by the invention is used for carrying out multi-parameter physical sign monitoring on non-hospitalized patients. Different from clinical monitoring, the real-time performance of data transmission, storage, abnormal reporting and response mechanisms related to the dynamic monitoring system are essentially different from the clinical monitoring.

In addition, through research on the existing monitor, the fact that the electrocardiograph monitoring is different from other parameters in vital sign indexes of electrocardiograph, blood pressure, blood oxygen, pulse, respiration, body temperature and the like is found, effective information in the electrocardiograph signals obtained through the sensor can be extracted through a series of complex algorithm calculation, and compared with other signals, the processing process is complex and difficult, and the link is easy to cause detection errors.

The electrocardiosignal is a weak current reflected by the electrical activity of the myocardial cells on the body surface and is recorded by the body surface electrode and the amplifying and recording system. In the recording process, other non-cardiogenic electric signals, such as myoelectric signal interference caused by skeletal muscle activity, and the like, can be recorded. Therefore, it is considered that effective interference identification and elimination are required for cardiac electrical signals to effectively reduce false alarms caused by interference signals.

In addition, the electrocardiosignals represent the myocardial electrical activity process, so the electrocardiosignals can be used for detecting the heart rate and also represent a large amount of information of the heart state. When the heart state is in a problem, the electrocardiosignals are changed correspondingly. In the course of research on the existing multi-parameter monitoring devices in the industry, we have found that the existing monitoring devices can only perform very limited analysis and alarm on the electrocardiosignals. In addition to effective interference identification and elimination for cardiac electrical signals to reduce false positives due to interfering signals, we believe that improvements can be made from:

firstly, accurate identification of P-waves and T-waves is required in heart beat feature extraction, which can avoid multiple detections and missed detections of heart beat detection, such as multiple detections of some special electrocardiogram signals, for example, high and large T-waves of patients with slower heart rhythm, or signals with large T-waves.

Secondly, the classification of heart beats is divided more carefully, and can not only stay in three classifications of sinus, supraventricular and ventricular, thereby meeting the complex and comprehensive analysis requirements of the clinician.

Third, accurate identification of atrial flutter and ST-T changes can help provide assistance in myocardial ischemia analysis for ST segment and T wave changes.

Fourth, accurate identification of cardiac beats and cardiac electrical events.

In the invention, aiming at the points, through the analysis and calculation of the electrocardiogram data, particularly by introducing an Artificial Intelligence (AI) technology, arrhythmia analysis, long intermittent asystole, flutter and flutter, conduction block, premature beat and escape, bradycardia, tachycardia, ST segment change detection and analysis and classification of electrocardiogram events are carried out on the acquired digital signals, so as to achieve the purpose of generating accurate alarm signals, thereby effectively monitoring the vital signs of patients.

There are data showing that more than 90% of heart disease episodes occur outside of the medical institution, and therefore it is very necessary to record and monitor the heart condition in daily situations for people with a history of heart disease.

Based on the discovery, the invention provides a user-oriented physical sign information dynamic monitoring method, which can be applied to user-oriented monitoring outside a medical institution. Can be implemented in various dynamic monitoring devices, including wearable devices. The user-oriented physical sign information dynamic monitoring method of the present invention is described in detail below with reference to the flowchart of the user-oriented physical sign information dynamic monitoring method shown in fig. 1. In the dynamic monitoring of the physical sign information of the present invention, the monitoring of the cardiac electrical signal is the most important.

As shown in fig. 1, the user-oriented physical sign information dynamic monitoring method of the present invention mainly includes the following steps:

step 110, the dynamic monitoring device receives monitoring reference data input by a user or issued by a server;

specifically, the dynamic monitoring device may be a multi-parameter monitoring device including a single-lead or multi-lead electrocardiographic monitoring function, and each dynamic monitoring device has a unique device ID. When the dynamic monitoring device is assigned to a user to be monitored for use, corresponding monitoring reference data can be configured in the dynamic monitoring device according to the condition of the user. Therefore, it is necessary to first determine the object information of the object to be measured.

For the electrocardiographic monitoring, the monitoring reference data can be understood as reference data or information indicating whether the monitored electrocardiographic signal of the user is normal or not and whether an alarm needs to be generated or not, and for different users, the setting of the monitoring reference data can be different, and the monitoring reference data can be specifically obtained by a mode of configuring and inputting on the dynamic monitoring device or a mode of configuring and issuing to the dynamic monitoring device through a server according to user information. In this embodiment, the monitoring reference data may include information of the object to be measured and set electrocardiographic abnormal event information. The abnormal event information of the electrocardio comprises information of various abnormal events of the electrocardio which need to generate abnormal alarm of the electrocardio, when the dynamic monitoring equipment carries out a series of processing such as acquisition and analysis on the electrocardiogram data to obtain the abnormal event of the electrocardio indicated by the electrocardiogram data, whether the abnormal event of the electrocardio is generated can be determined by determining whether the abnormal event of the electrocardio is the event specified in the abnormal event information of the electrocardio.

For other vital sign parameters, such as pulse data, blood pressure data, respiration data, blood oxygen saturation data, and body temperature data, the monitoring reference data may be set corresponding parameter thresholds. The parameter threshold of each parameter can have different sets of parameter thresholds, and can be selected according to the actual situation of the monitored person. Preferably, in the present invention, before monitoring, monitoring reference data may be determined in advance according to the object to be measured, and then a corresponding set threshold may be determined according to the monitoring reference data.

In this example, the monitoring reference data at least includes information of the object to be measured and information of an abnormal event of electrocardiography.

Step 120, the dynamic monitoring device collects monitoring data of the object to be monitored to obtain physical sign monitoring data of the object to be monitored;

specifically, the dynamic monitoring device is provided with an electrode, a probe, a cuff and other sign signal acquisition devices which are in contact with the measured object, and the sign signal acquisition devices acquire the sign signals of the measured object and digitally process the sign signals to obtain sign monitoring data. The physical sign monitoring data at least comprises electrocardiogram data, and also can comprise the pulse data, the blood pressure data, the respiration data, the blood oxygen saturation data, the body temperature data and the like. The physical sign monitoring data has time attribute information, and each data point has corresponding data acquisition time which is the time attribute information. When data acquisition is carried out, the data acquisition time is also recorded and stored as the time attribute information of the physical sign monitoring data.

For better understanding of the intent and implementation of the present invention, the following brief introduction describes the acquisition method and principle of various physical sign monitoring data:

electrocardiogram data: the electrocardiosignal acquisition and recording instrument for non-invasive electrocardiographic examination acquires and records signals generated by the electrophysiological activity of heart cells in a single-lead or multi-lead mode.

Pulse data: the pulse is a phenomenon in which arterial blood vessels periodically pulsate with the relaxation of the heart, and the pulse includes changes in various physical quantities such as intravascular pressure, volume, displacement, and wall tension. The sensor is composed of a light source and a photoelectric transducer, and can be clamped on the fingertip or the auricle of a measured person. The light source is selected to have a wavelength selective for oxygenated hemoglobin in arterial blood, such as by using a light emitting diode with a spectrum of 700 and 900 nm. The light beam is transmitted through peripheral blood vessel of human body, when the arterial blood congestion volume is changed, the light transmittance of said light beam is changed, and the light transmitted or reflected by tissue can be received by photoelectric transducer, converted into electric signal, and amplified and outputted by amplifier, so that the change of arterial blood vessel volume can be reflected. The pulse is a signal that periodically changes with the pulsation of the heart, the arterial blood vessel volume also periodically changes, and the signal change period of the photoelectric transducer is the pulse rate, i.e., pulse data.

Blood pressure data: the highest pressure reached during systole, called the systolic pressure, pushes blood into the aorta and maintains systemic circulation. The lowest pressure reached when the heart expands is called the diastolic pressure, which allows blood to flow back into the right atrium. The integral of the blood pressure waveform over one week divided by the cardiac cycle T is called the mean pressure. The measurement of blood pressure data can be realized by various methods, and can be specifically divided into invasive measurement and non-invasive measurement. In the multi-parameter monitor, two non-invasive measurement methods, namely a Korotkoff sound method and a vibration measurement method, are preferably adopted. The korotkoff sound method is used for measuring the blood pressure by detecting korotkoff sound (pulse sound) under the cuff, and the korotkoff sound non-invasive blood pressure monitoring system comprises a cuff inflating system, the cuff, a korotkoff sound sensor, an audio amplification and automatic gain adjustment circuit, an A/D converter, a microprocessor, a display part and the like. The vibration measurement method is to detect the oscillation wave of the gas in the cuff, the oscillation wave is originated from the pulsation of the blood vessel wall, and the blood pressure data including the systolic Pressure (PS), the diastolic Pressure (PD) and the average Pressure (PM) can be measured by measuring the relevant points of the oscillation wave. The method for obtaining the pulse vibration wave by the vibration measurement method can obtain the pulse vibration wave by measurement by means of a microphone and a pressure sensor, and then the blood pressure data can be obtained. For some special application scenarios, the blood pressure data can also be obtained by invasive measurement. For example, for monitoring some patients in Intensive Care Unit (ICU) wards, real-time blood pressure data acquisition may be accomplished by directly cannulating the artery with the other end of the cannulation connected to a sterile fluid-filled pressure sensing system. Advantages of this invasive monitoring method include: the blood pressure can be displayed in real time, and continuous blood pressure change waveforms can be displayed; accurate readings can be made in the hypotensive state; the comfort of the patient recorded for a long time is improved, and the wound caused by long-term inflation and deflation in non-invasive measurement is avoided; more information can be extracted, including the ability to estimate the volume of the blood vessel from the morphology of the blood pressure waveform.

Respiratory data: respiratory measurements are an important part of lung kinetic energy examination. The monitor measures the respiratory frequency (times/minute) by measuring the respiratory wave, and then obtains the respiratory data. The respiratory frequency can be measured by directly measuring the temperature change of the respiratory airflow through a thermistor, and the change is converted into a voltage signal through a bridge circuit; the respiratory rate can also be measured by impedance method, because the muscles of chest wall are alternatively relaxed, the thorax is alternatively deformed, and the electrical impedance of the organism tissue is also alternatively changed. The variation of the respiratory impedance value can be measured by various methods such as a bridge method, a modulation method, a constant voltage source method, a constant current source method and the like. In the monitor, the respiration impedance electrode can be used together with the electrocardio-electrode, and the respiration impedance change and the respiration frequency can be detected simultaneously when the electrocardio-signal is detected.

Blood oxygen saturation data: blood oxygen saturation is an important parameter that measures the ability of human blood to carry oxygen. The blood oxygen saturation can be measured by adopting a transmission method (or a reflection method) dual-wavelength (red light R and infrared light IR) photoelectric detection technology, detecting the ratio of alternating components caused by the light absorption of arterial blood by red light and infrared light and the stable component (direct current) value of the light absorption caused by non-pulsating tissues (epidermis, muscle, venous blood and the like), and obtaining the blood oxygen saturation value SpO2, namely blood oxygen saturation data, by calculation. Because the pulse rule of the photoelectric signal is consistent with the pulse rule of the heart, the pulse data can be simultaneously determined according to the period of the detected signal.

Body temperature data: body temperature is an important indicator for understanding the state of life. The body temperature is measured by using a thermistor with a negative temperature coefficient as a temperature sensor and a bridge as a detection circuit. In specific application, an integrated temperature measuring circuit can be used for measuring to obtain body temperature data. More than two temperature measuring circuits can be used for measuring the temperature difference of two different parts to correct the measured value. Body surface probes and body cavity probes can also be used to monitor body surface and temperature in the cavity respectively. In some special applications, in order to avoid cross infection, the questioning data may be monitored by using an infrared non-contact temperature measurement technology. In the monitor, the temperature measurement precision is set to be 0.1 ℃ so as to have faster temperature measurement response.

In the invention, the physical sign monitoring data of the measured object can be acquired by the method.

As mentioned above, the monitoring of the electrocardiogram is more complex compared with the monitoring of vital sign indexes such as other blood pressure, blood oxygen, pulse, respiration, body temperature and the like, so that the processing method different from other physical sign monitoring data is adopted for the electrocardiogram data in the invention, and particularly, the electrocardiogram automatic analysis method based on artificial intelligence self-learning is adopted for carrying out the identification, processing and abnormal judgment on the electrocardiogram data.

The following flow is also an explanation of a method for dynamically monitoring vital sign information, mainly taking the processing procedure of electrocardiogram data as an example. The other physical sign data can be used as reference information and combined with the processing result of the electrocardiogram data to judge the physical sign state of the monitored person.

Step 130, performing wave group feature identification on the electrocardiogram data to obtain feature signals of the electrocardiogram data, performing heart beat classification on the electrocardiogram data according to the feature signals, obtaining heart beat classification information by combining with electrocardiogram basic rule reference data, and generating electrocardiogram event data;

specifically, in this example, the processing of the electrocardiogram data may be performed in the dynamic monitoring device or in the server. The dynamic monitoring equipment can be connected with the server in a wired mode or a wireless network for data transmission. Preferably, the data transmission is realized by adopting a wireless network when the real-time data transmission is carried out. The wireless network includes but is not limited to wireless local area network (WIFI) based on IEEE 802.11b standard, bluetooth, 3G/4G/5G mobile communication network, internet of things, and the like.

After the dynamic monitoring device monitors and obtains the electrocardiogram data, the electrocardiogram data can be stored locally in the dynamic monitoring device, or the data can be transmitted to a server for storage, and the transmitted data is loaded with the device ID information of the dynamic monitoring device. Therefore, the information of the monitored person can be correspondingly obtained.

The electrocardiogram data processing process of the invention adopts an artificial intelligence self-learning based electrocardiogram automatic analysis method, and is realized based on an artificial intelligence Convolutional Neural Network (CNN) model. The CNN model is a supervised learning method in deep learning, namely a multi-layer network (hidden layer) connecting structure simulating a neural network, an input signal sequentially passes through each hidden layer, a series of complex mathematical treatments (Convolution, Pooling Pooling, Regularization, overfitting prevention, Dropout temporary discarding, Activation and generally using a Relu Activation function) are carried out in the hidden layer, some characteristics of an object to be recognized are automatically abstracted layer by layer, then the characteristics are taken as input and transmitted to a high-level hidden layer for calculation until a Full connecting layer (Full Connection) of the last layers reconstructs the whole signal, and a Softmax function is used for logical (logistic) regression to achieve multi-target classification.

The CNN belongs to a supervised learning method in artificial intelligence, in a training stage, an input signal is processed by a plurality of hidden layers to reach a final full-connection layer, an error exists between a classification result obtained by softmax logistic regression and a known classification result (label), and a core idea of deep learning is to continuously minimize the error through a large number of sample iterations so as to calculate and obtain parameters for connecting neurons of all hidden layers. This process generally requires constructing a special cost function (cost function), and rapidly and effectively minimizing all connection parameters in a neural network structure with very complex overall depth (the number of hidden layers) and breadth (the dimension of a feature) by using a nonlinear optimization gradient descent algorithm and an error back propagation algorithm (BP).

The deep learning inputs the data to be recognized into the training model, and the data passes through the first hidden layer, the second hidden layer and the third hidden layer, and finally the recognition result is output.

In the invention, the wave group characteristic recognition, the interference recognition, the heart beat classification and the like of the electrocardiogram data are all output results based on the artificial intelligent self-learning training model, and the analysis speed and the accuracy degree are high.

Whether the processing of the electrocardiogram data is performed in the ambulatory monitoring device or in the server, the detailed processing can be performed according to the process shown in fig. 3, and according to the following steps,

step 131, converting the data format of the electrocardiogram data into a preset standard data format through resampling, and performing first filtering processing on the converted electrocardiogram data with the preset standard data format;

specifically, the format of the electrocardiogram data is adapted to be read, different reading implementations are realized for different devices, and after reading, the baseline needs to be adjusted and the data is converted into millivolt data according to the gain. And converting the data into sampling frequency which can be processed by the whole process through data resampling. Then, high-frequency and low-frequency noise interference and baseline drift are removed through filtering, and the accuracy of artificial intelligence analysis is improved. And storing the processed electrocardiogram data in a preset standard data format.

The method solves the problem that different leads are used, the sampling frequency and the transmission data format are different, and removes the noise interference and the baseline drift of high frequency and low frequency through digital signal filtering.

The digital signal filtering can adopt a high-pass filter, a low-pass filter and median filtering respectively, so that power frequency interference, myoelectricity interference and baseline drift interference are eliminated, and the influence on subsequent analysis is avoided.

More specifically, a low-pass and high-pass Butterworth filter can be adopted for zero phase shift filtering so as to remove baseline drift and high-frequency interference and retain effective electrocardiosignals; the median filtering may replace the amplitude of the window center sequence with a median of the voltage amplitudes of the data points within a sliding window of a preset duration. The low frequency baseline drift can be removed.

Astep 132 of performing heartbeat detection processing on the electrocardiogram data after the first filtering processing to identify a plurality of heartbeat data included in the electrocardiogram data;

specifically, each heartbeat data corresponds to a heartbeat cycle, including the amplitude and start-stop time data of the corresponding P-wave, QRS complex, T-wave. The heart beat detection in the step is composed of two processes, namely a signal processing process, wherein a characteristic frequency band of a QRS wave group is extracted from the electrocardiogram data after the first filtering processing; secondly, the occurrence time of the QRS complex is determined by setting a reasonable threshold. In an electrocardiogram, the P-wave, QRS complex, T-wave component and noise component are generally included. The frequency range of the QRS complex is typically between 5 and 20Hz, and the QRS complex signal can be extracted by a band-pass filter in this range. However, the frequency bands of P wave and T wave and the frequency bands of noise partially overlap with the frequency bands of QRS complex, so that the signals except QRS complex cannot be completely removed by the signal processing method. It is therefore necessary to extract the QRS complex position from the signal envelope by setting a reasonable threshold. The specific detection process is a process based on peak detection. And (4) performing threshold judgment on each peak value sequence in the signal, and entering a QRS complex judgment process when the threshold value is exceeded, so as to perform detection of more features, such as RR intervals, forms and the like.

Multi-parameter monitors tend to record for long periods of time during which the amplitude and frequency of the heartbeat signal change from moment to moment, and this characteristic is more pronounced in disease states. When the threshold is set, the threshold needs to be dynamically adjusted according to the change situation of the data characteristics in the time domain. In order to improve the detection accuracy and the positive rate, the QRS complex detection is mostly carried out in a mode of combining a double-amplitude threshold value with a time threshold value, a high threshold value has a higher positive rate, a low threshold value has a higher sensitivity rate, and when an RR interval exceeds a certain time threshold value, the detection is carried out by using the low threshold value, so that the condition of missed detection is reduced. On the other hand, since the low threshold is easily affected by T waves and myoelectric noise and is likely to cause multiple detections, detection is preferentially performed using the high threshold.

The heartbeat data for the different leads have lead parameters to characterize which lead heartbeat data the heartbeat data is. Therefore, the electrocardiogram data can be obtained, and simultaneously, the lead information can be determined according to the transmission source, and the information is used as the lead parameters of the heart beat data.

Step 133, determining a detection confidence of each heartbeat according to the heartbeat data;

specifically, the confidence calculation module can provide an estimated value of the confidence of QRS complex detection according to the amplitude of QRS complex and the amplitude proportion of the noise signal in RR interval during the process of heart beat detection.

134, performing interference identification on the heartbeat data according to the interference identification two-classification model to obtain whether interference noise exists in the heartbeat data and a probability value for judging the interference noise;

the multi-parameter monitor is susceptible to various influences and generates interference phenomena in the long-time recording process, so that the acquired heartbeat data is invalid or inaccurate, the condition of a testee cannot be correctly reflected, and the diagnosis difficulty and workload of a doctor are increased; and the interference data is also a main factor causing the intelligent analysis tool to not work effectively. Therefore, it is important to minimize the interference of external signals.

The method is based on an end-to-end binary identification model taking a deep learning algorithm as a core, has the characteristics of high precision and strong generalization performance, can effectively solve the problem of disturbance generated by main interference sources such as electrode plate falling, motion interference, electrostatic interference and the like, and solves the problem of poor identification effect caused by various and irregular interference data changes in the traditional algorithm.

The method can be specifically realized by the following steps:

step A, performing interference identification on the heartbeat data by using an interference identification binary classification model;

step B, identifying data segments of which the heart beat interval is greater than or equal to a preset interval judgment threshold in the heart beat data;

step C, carrying out signal abnormality judgment on the data segment of which the cardiac interval is greater than or equal to the preset interval judgment threshold value, and determining whether the data segment is an abnormal signal;

the identification of the abnormal signal mainly includes whether the electrode plate falls off or not, low voltage and the like.

Step D, if the abnormal signal is not detected, determining an initial data point and an end data point of sliding sampling in the data segment according to a set time value by a preset time width, starting sliding sampling on the data segment from the initial data point until the end data point, and obtaining a plurality of sampling data segments;

and E, carrying out interference identification on each sampling data segment.

The above steps A to E are explained as a specific example. Cutting and sampling heartbeat data of each lead by a set first data volume, and then respectively inputting the heartbeat data into an interference recognition two-classification model for classification to obtain an interference recognition result and a probability value of a corresponding result; judging whether the heartbeat data of the interval of the heartbeat is more than or equal to 2 seconds or not, wherein the heartbeat data is signal overflow, low voltage and electrode falling; if the condition is not the above, the sampling is performed by the first data amount, starting from the left heart beat and continuously sliding to the right by the first data amount without overlapping.

The method has the advantages that the first data volume heartbeat data which can be any lead is input, then the interference identification binary model is adopted for classification, whether the classification result is interference or not is directly output, the obtained result is fast, the accuracy is high, the stability is good, and more effective and high-quality data can be provided for subsequent analysis.

Because the interference data is often caused by the action of external disturbance factors, mainly the conditions of electrode plate falling, low voltage, electrostatic interference, motion interference and the like exist, not only are the interference data generated by different disturbance sources different, but also the interference data generated by the same disturbance source are various; meanwhile, considering that although the interference data is wide in diversity and large in difference with normal data, the diversity is guaranteed as much as possible when the interference training data are collected, meanwhile, the diversity of the interference data is increased as much as possible by adopting sliding sampling of a moving window, so that a model is more robust to the interference data, and even if the future interference data is different from any previous interference, the similarity of the future interference data and the interference data is larger than that of the normal data, so that the capacity of the model for identifying the interference data is enhanced.

The interference recognition two-class model used in this step can be as shown in fig. 3, where the network uses 2 layers of convolution layers first, the size of the convolution kernel is 1 × 5, and a maximum pooling is added after each layer. The number of convolution kernels starts at 128 and doubles for each maximum pooling layer. The convolutional layer is followed by two fully connected layers and one softmax classifier. Because the classification number of the model is 2, softmax has two output units which sequentially correspond to corresponding classes, and cross entropy is adopted as a loss function.

For training of this model, we used precisely labeled data segments from approximately 400 million patients, 30 million patients. Annotations fall into two categories: a normal electrocardiogram signal or a significantly disturbed electrocardiogram signal segment. The method comprises the steps of labeling the segments through a customized developed tool, and then storing interference segment information in a custom standard data format.

In the training process, two GPU servers are used for dozens of times of round-robin training. In one specific example, the sampling rate is 200Hz, the data length is a segment D [300] of 300 ECG voltage values (millivolts), and the input data is: InputData (i, j), where i is the ith lead and j is the jth segment D of lead i. The training is started after all input data are scattered randomly, so that the convergence of the training process is ensured, and meanwhile, too many samples are collected from the electrocardiogram data of the same patient under control, the generalization capability of the model is improved, and the accuracy in a real scene is improved. After the training is converged, 100 ten thousand independent test data are used for testing, and the accuracy can reach 99.3%. Additional specific test data are shown in table 1 below.

	Interference	Is normal
			Sensitivity (Sensitivity)	99.14％	99.32％
Positive prediction rate (Positive predictivity)	96.44％	99.84％

TABLE 1

Step 135, determining validity of the heartbeat data according to the detection confidence, merging and generating heartbeat time series data based on the interference recognition result and the time rule according to the lead parameters of the heartbeat data and the heartbeat data, and generating heartbeat analysis data according to the heartbeat time series data;

specifically, because the complexity of the electrocardiogram signal and each lead may be affected by different degrees of interference, there are multiple detection and missed detection conditions when detecting the heart beat by means of a single lead, and the time characterization data of the heart beat result detected by different leads are not aligned, so it is necessary to combine the heart beat data of all leads according to the interference identification result and the time rule to generate a complete heart beat time sequence data, unify the time characterization data of all lead heart beat data. Wherein the time characterizing data is used to represent time information of each data point on a time axis of the electrocardiogram data signal. According to the unified heartbeat time sequence data, in the subsequent analysis and calculation, preset threshold values can be used for cutting the heartbeat data of each lead, so that the heartbeat analysis data of each lead required by specific analysis can be generated.

The cardiac activity data for each of the above leads needs to be validated according to the detection confidence obtained instep 133 before being combined.

Specifically, the lead heart beat merging module performs the heart beat data merging process as follows: acquiring time characterization data combinations of different lead heart beat data according to refractory periods of electrocardiogram basic rule reference data, discarding heart beat data with larger deviation, voting the time characterization data combinations to generate combined heart beat positions, adding the combined heart beat positions into a combined heart beat time sequence, moving to the next group of heart beat data to be processed, and circularly executing until all heart beat data are combined.

Wherein the electrocardiographic activity refractory period may preferably be between 200 milliseconds and 280 milliseconds. The time characterization data combination of the acquired different lead heart beat data should satisfy the following conditions: temporal characterization data of heart beat data each lead in the combination contains at most one temporal characterization data of heart beat data. When voting is carried out on the time characterization data combination of the heartbeat data, the percentage of the lead number of the detected heartbeat data in the effective lead number is used for determining; if the position of the lead corresponding to the time characterization data of the heartbeat data is a low voltage section, an interference section and the electrode is fallen off, the lead is considered as an invalid lead for the heartbeat data. In calculating the specific location of the merged heart beat, the average value of the time characterization data of the heart beat data can be used. During the merging process, the method sets a refractory period to avoid error merging.

In this step, one unified heart beat time series data is output by the merge operation. The step can simultaneously reduce the multi-detection rate and the missing detection rate of the heart beat, and effectively improve the sensitivity and the positive prediction rate of heart beat detection.

Step 136, extracting and analyzing the characteristics of amplitude and time characterization data of the heartbeat analysis data according to the heartbeat classification model to obtain primary classification information of the heartbeat analysis data;

the difference of different electrocardio monitoring devices in the aspects of signal measurement, signal acquisition or lead data output and the like exists, so that a simple single-lead classification method or a multi-lead classification method can be used according to specific conditions. The multi-lead classification method comprises a lead voting decision classification method and a lead synchronous association classification method. The lead voting decision classification method is a voting decision method which carries out lead independent classification based on the heart beat analysis data of each lead and then fuses result voting to determine a classification result; the lead synchronous correlation classification method adopts a method of carrying out synchronous correlation analysis on heartbeat analysis data of each lead. The single-lead classification method is to directly use a corresponding lead model to classify the heartbeat analysis data of the single-lead equipment without a voting decision process. The following is a description of the above classification methods.

The single-lead classification method comprises the following steps:

and cutting the single-lead heart beat data according to the heart beat time sequence data to generate single-lead heart beat analysis data, inputting the single-lead heart beat analysis data into a trained heart beat classification model corresponding to the lead, and extracting and analyzing the characteristics of amplitude and time characterization data to obtain single-lead primary classification information.

The lead voting decision classification method may specifically include:

firstly, cutting heartbeat data of each lead according to the heartbeat time sequence data to generate heartbeat analysis data of each lead;

secondly, extracting and analyzing the characteristics of amplitude and time characterization data of the heart beat analysis data of each lead according to the heart beat classification model corresponding to each lead obtained by training to obtain the classification information of each lead;

and thirdly, performing classification voting decision calculation according to the classification information of each lead and the lead weight value reference coefficient to obtain the primary classification information. Specifically, the lead weight value reference coefficient is a voting weight coefficient of each lead for different heart beat classifications obtained based on the Bayesian statistical analysis of the electrocardiographic data.

The lead synchronous association classification method may specifically include:

cutting the heartbeat data of each lead according to the heartbeat time sequence data so as to generate heartbeat analysis data of each lead; and then, carrying out feature extraction and analysis on synchronous amplitude and time characterization data of the heartbeat analysis data of each lead according to the trained multi-lead synchronous association classification model to obtain primary classification information of the heartbeat analysis data.

The synchronous association classification method of heart beat data includes inputting all lead data of dynamic electrocardiogram equipment, intercepting data points with the same position and certain length on each lead according to heart beat positions of unified heart beat analysis data, synchronously transmitting the data points to a trained artificial intelligent deep learning model for calculation and analysis, and outputting accurate heart beat classification of heart beat characteristics in which all lead electrocardiogram signal characteristics and heart beat characteristics associated with each other in time are comprehensively considered at each heart beat position point.

The method fully considers that the different lead data of the electrocardiogram actually measures the information flow transmitted by the electrocardio signal in different electrocardio axis vector directions, comprehensively analyzes the multidimensional digital characteristics transmitted by the electrocardiogram signal in time and space, greatly improves the defect that the traditional method is easy to obtain wrong classification by only depending on single lead independent analysis and then summarizing the result and performing some statistical voting modes, and greatly improves the accuracy of heart beat classification.

The heartbeat classification model adopted in this step may be as shown in fig. 4, and specifically may be an end-to-end multi-label classification model inspired by models such as a convolutional neural network AlexNet, VGG16, inclusion and the like based on artificial intelligence deep learning. Specifically, the network of the model is a 7-layer convolutional network, each convolution being followed by an activation function. The first layer is two different sized convolutional layers, followed by six convolutional layers. The convolution kernels of the seven-layer convolution are 96,256,256,384,384,384,256, respectively. The scale of the convolution kernel for the other layers is 5, except that the first layer has two scales of 5 and 11, respectively. The third, fifth, sixth and seventh convolution layers are followed by a pooling layer. Finally followed by two fully connected layers.

In the heart beat classification model in the step, 1700 ten thousand data samples of a training set containing 30 ten thousand patients are adopted for training. The samples are generated by accurately marking data according to the requirements of dynamic electrocardiogram analysis and diagnosis, and the marking is mainly aimed at common arrhythmia, conduction block and ST segment and T wave change, and can meet the requirements of model training of different application scenes. Specifically, the marked information is stored in a preset standard data format. In the preprocessing of training data, in order to increase the generalization capability of the model, a small sliding is performed on the classification with a small sample size to amplify the data, specifically, the data is moved for 2 times according to a certain step length (for example, 10 to 50 data points) on the basis of each heartbeat, so that 2 times of data can be increased, and the identification accuracy of the classification samples with a small data size is improved. The generalization capability is also improved through the verification of actual results.

In an actual training process, two GPU servers are used for dozens of times of round-robin training, 500 thousands of independent test data are used for testing after the training is converged, and the accuracy can reach 91.92%.

The length of the truncation of the training data may be 1 second to 10 seconds. For example, the sampling rate is 200Hz, 2.5s is taken as the sampling length, the length of the obtained data is a segment D [500] of 500 electrocardiogram voltage values (millivolts), and the input data is: InputData (i, j), where i is the ith lead and j is the jth segment D of lead i. The training is started after all input data are scattered randomly, so that the convergence of the training process is ensured, and meanwhile, too many samples are collected from the electrocardiogram data of the same patient under control, the generalization capability of the model is improved, and the accuracy in a real scene is improved. During training, segment data D corresponding to all leads are synchronously input, and lead data of a plurality of space dimensions (different electrocardiogram axial vectors) at each time position are synchronously learned according to a multi-channel analysis method of image analysis, so that a classification result which is more accurate than that of a conventional algorithm is obtained.

Step 137, inputting the heartbeat analysis data of the specific heartbeat in the primary classification information result into an ST segment and T wave change model for identification, and determining the ST segment and T wave evaluation information;

the ST segment and T wave evaluation information is specifically lead position information of ST segment and T wave change corresponding to heartbeat analysis data. Since clinical diagnosis requires localization to specific leads for ST-segment and T-wave changes.

Here, the specific heartbeat data of the one-time classification information refers to heartbeat analysis data including sinus heartbeat (N) and other heartbeat types that may include ST change.

The ST-segment and T-wave change lead positioning module inputs specific heartbeat data of primary classification information into an artificial intelligence deep learning training model for identifying ST-segment and T-wave changes in sequence according to each lead, calculation analysis is carried out, and the output result shows whether the characteristics of the lead segments conform to the conclusion of ST-segment and T-wave changes or not, so that the information of the specific leads, namely the ST-segment and T-wave evaluation information, of the ST-segment and T-wave changes can be determined. The specific method can be as follows: and inputting the analysis data of each lead heart beat with the result of the sinus heart beat in the primary classification information into an ST segment and T wave change model, and identifying and judging the analysis data of the sinus heart beat one by one to determine whether the analysis data of the sinus heart beat has ST segment and T wave change characteristics and generated specific lead position information and determine the evaluation information of the ST segment and the T wave.

The ST segment and T wave change model used in this step may be, as shown in fig. 5, specifically, an end-to-end classification model inspired by models such as an artificial intelligence deep learning-based convolutional neural network AlexNet and VGG 16. Specifically, the model is a 7-layer network, and the model comprises 7 convolutions, 5 pooling and 2 full connections. Convolution kernels used in convolution are all 1x5, and the number of filters in each layer of convolution is different. The number of the 1 st layer convolution filters is 96; the convolution of the 2 nd layer and the convolution of the 3 rd layer are used together, and the number of the filters is 256; the 4 th layer convolution and the 5 th layer convolution are used together, and the number of the filters is 384; the number of the 6 th layer convolution filters is 384; the number of the 7 th layer convolution filters is 256; the 1 st, 3 rd, 5 th, 6 th and 7 th convolution layers are then pooled. Two full joins follow, and finally the results are classified into two categories using a Softmax classifier. In order to increase the nonlinearity of the model and extract the characteristics of the data with higher dimensionality, two convolution modes are adopted.

Because the proportion of the heart beats with the ST segment changes and the T wave changes in all heart beats is low, in order to take the diversity of training data and the balance of various types of data volumes into consideration, the proportion of the selected training data without the ST segment changes and the T wave changes and with the ST segment changes and the T wave changes is about 2:1, the good generalization capability of the model in the classification process is ensured, and the tendency that the training data accounts for more classes does not appear. Because the heart beat has various forms and the forms expressed by different individuals are different, in order to better estimate the distribution of each classification by a model, the characteristics can be effectively extracted, and training samples are collected from individuals of different ages, weights, sexes and living areas; in addition, because the electrocardiogram data of a single individual in the same time period are often highly similar, in order to avoid over-learning, when the data of the single individual is acquired, a small number of samples in different time periods are randomly selected from all the data; finally, due to the fact that the heart beat forms of the patients have the characteristics of large difference among individuals and high similarity among the individuals, when training and testing sets are divided, different patients are divided into different data sets, the situation that data of the same individual are simultaneously shown in the training sets and the testing sets is avoided, therefore, the testing result of the obtained model is closest to a real application scene, and the reliability and universality of the model are guaranteed.

Step 138, according to the heart beat time sequence data, performing P wave and T wave feature detection on the heart beat analysis data, and determining detailed feature information of the P wave and the T wave in each heart beat;

specifically, the detailed feature information includes data of amplitude, direction, form and start-stop time; in the analysis of the heart beat signal, the characteristics of P wave, T wave and QRS wave are also important basis in the electrocardiogram analysis.

In a P wave and T wave feature detection module, various features in the P wave, the T wave and the QRS wave complex are extracted by calculating the positions of the splitting points in the QRS wave complex and the positions of the splitting points of the P wave and the T wave. The method can be realized by QRS complex segmentation point detection, single-lead PT detection algorithm and multi-lead PT detection algorithm respectively.

And (3) QRS wave group segmentation point detection: and searching the R point, the R 'point, the S point and the S' point of the QRS complex in the single lead according to the QRS complex section power maximum point and the starting and stopping points provided by the QRS complex detection algorithm. And when the multi-lead data exists, calculating the median of each segmentation point as the position of the final segmentation point.

Single lead P wave and T wave detection algorithm: the P wave and the T wave have low amplitude and smooth signals relative to the QRS complex, are easily submerged in low-frequency noise, and are difficult points in detection. According to the QRS complex detection result, after the influence of the QRS complex on the low-frequency band is eliminated, a low-pass filter is used for carrying out third filtering on the signal, so that the relative amplitude of the PT wave is increased. Then, a T wave is searched between two QRS wave groups by a peak detection method. Because the T wave is a complex generated by ventricular repolarization, there is a clear time-locked relationship between the T wave and the QRS complex. Taking the detected QRS complex as a reference, taking a midpoint (e.g. limited to a range between 400ms and 600ms after the first QRS complex) from each QRS complex to the next QRS complex as a T wave detection end point, and selecting the largest peak in the interval as a T wave. And selecting the peak with the maximum amplitude from the rest peak values as the P wave. And simultaneously, determining the direction and morphological characteristics of the P wave and the T wave according to the peak value and position data of the P wave and the T wave. Preferably, the cut-off frequency of the low-pass filtering is set between 10-30 Hz.

The multi-lead P wave and T wave detection algorithm comprises the following steps: in the case of multi-lead, P, T wave detection can be performed by a tracing algorithm because the generation time of each wave in the heart beat is the same, the spatial distribution is different, and the temporal spatial distribution of noise is different. The QRS complex removal process is first performed on the signal and the signal is third filtered using a low pass filter to remove interference. Then, each independent component in the original waveform is calculated through an independent component analysis algorithm. And in each separated independent component, selecting corresponding components as P wave and T wave signals according to the distribution characteristics of the peak values and the QRS complex positions, and simultaneously determining the direction and morphological characteristics of the P wave and the T wave.

Step 139, performing secondary classification processing on the heartbeat analysis data under primary classification information according to electrocardiogram basic rule reference data, detailed characteristic information of P waves and T waves and ST segment and T wave evaluation information to obtain heartbeat classification information; and analyzing and matching the heart beat classification information to generate electrocardiogram event data.

Specifically, the reference data of the electrocardiogram basic rule is generated by following the basic rule description of the myocardial cell electrophysiological activity and the electrocardiogram clinical diagnosis in an authoritative electrocardiogram textbook, such as the minimum time interval between two heartbeats, the minimum interval between a P wave and an R wave and the like, and is used for subdividing primary classification information after heart beat classification; mainly based on the inter-beat RR intervals and the medical significance of different heart beat signals on each lead; the heart beat auditing module divides the ventricular heart beat into more detailed heart beat classifications according to the classification recognition of the electrocardiogram basic rule reference data combined with certain continuous multiple heart beat analysis data and the detailed characteristic information of P waves and T waves, and comprises the following steps: ventricular premature beat (V), Ventricular Escape (VE), accelerated ventricular premature beat (VT), subdividing supraventricular heartbeat-like into supraventricular premature beat (S), atrial escape (SE), Junctional Escape (JE), and atrial accelerated premature beat (AT), etc.

In addition, through the secondary classification processing, the error classification identification of the reference data which does not accord with the basic rule of the electrocardiogram and occurs in the primary classification can be corrected. And carrying out pattern matching on the subdivided heart beat classifications according to the electrocardiogram basic rule reference data, finding out classification identification which does not conform to the electrocardiogram basic rule reference data, and correcting the classification to be reasonable classification according to the RR intervals and the front and back classification identifications.

Specifically, through the secondary classification process, a plurality of heartbeat classifications can be output, such as: normal sinus beats (N), complete right bundle branch block (N _ CRB), complete left bundle branch block (N _ CLB), intraventricular block (N _ VB), first degree atrioventricular conduction block (N _ B1), priming (N _ PS), ventricular premature beats (V), Ventricular Escape (VE), accelerated ventricular premature beats (VT), supraventricular premature beats (S), atrial escape (SE), Junctional Escape (JE), accelerated atrial premature beats (AT), Atrial Flutter (AF), artifact (a) and the like.

Through the steps, the calculation of the basic heart rate parameters can be completed. Wherein the underlying calculated heart rate parameters include: RR interval, heart rate, QT time, QTc time and other parameters.

Subsequently, based on the result of the classification of the heartbeat cycle, pattern matching is performed with reference to the data according to the basic rules of the electrocardiogram, and the following typical electrocardiogram events, including but not limited to, the classification corresponding to the electrocardiogram event data, can be obtained:

supraventricular premature beat

Supraventricular premature beat pairing

Bigeminal ventricular premature beat

Triple rule for supraventricular premature beat

Atrial escape

Atrial escape heart rhythm

Junctional escape

Junctional escape heart rhythm

Non-paroxysmal supraventricular tachycardia

Fastest supraventricular tachycardia

Longest supraventricular tachycardia

Supraventricular tachycardia

Short ventricular tachycardia

Atrial flutter-fibrillation

Ventricular premature beat

Ventricular premature beat pair

Bigeminal ventricular premature beat

Triple rule for ventricular premature beat

Ventricular escape

Ventricular escape heart rhythm

Accelerated ventricular autonomic cardiac rhythm

Fastest ventricular tachycardia

Maximal ventricular tachycardia

Ventricular tachycardia

Short ventricular tachycardia

Second degree I type sinoatrial block

Second degree type II sinoatrial block

First degree atrioventricular block

Second degree I type atrioventricular block

Second degree II type atrioventricular block

Second degree II type (2: 1) atrioventricular block

High atrioventricular block

Complete left bundle branch block

Complete right bundle branch block

Indoor block

Pre-excitation syndrome

ST-segment and T-wave changes

Maximum RR interval

And generating electrocardiogram event data by using the heart beat analysis data according to the heart beat classification information and the electrocardiogram basic rule reference data. The cardiac event data includes device ID information for the ambulatory monitoring device.

Step 140, determining corresponding electrocardiogram event information in real time according to the electrocardiogram event data, and determining whether the electrocardiogram event information is electrocardiogram abnormal event information;

specifically, after obtaining the electrocardiographic event data, the corresponding relationship between the electrocardiographic event data obtained by artificial intelligence learning and the electrocardiographic time information can be used to correspondingly obtain corresponding electrocardiographic event information, for example, the electrocardiographic event information corresponding to the electrocardiographic event data is a sinus heart beat event, a ventricular premature beat event, or the like. Only part of the abnormal events are the abnormal events of the electrocardio which need to generate an alarm.

The data processing process is real-time, so that the dynamic monitoring equipment can continuously generate electrocardiogram event information. In practical application, the output interval of the electrocardiogram event information can be reasonably set, so that the data calculation amount is reduced, and the condition of missing detection is avoided.

After obtaining the electrocardiographic event information, matching the electrocardiographic event information with the electrocardiographic abnormal event information recorded in the dynamic monitoring equipment, if the electrocardiographic abnormal event information is the electrocardiographic abnormal event information, executing thestep 150, otherwise, continuing to execute thestep 120, and continuing to acquire monitoring data of the measured object.

In a preferred scheme, no matter whether an abnormal event of the electrocardiogram is monitored, a monitoring report for a monitored user can be generated according to electrocardiogram data and electrocardiogram event data and preset rules.

For example, the data may be collected at preset time intervals, for example, every 24 hours, to generate monitoring report data of the measured object in corresponding time periods.

Step 150, generating and outputting alarm information through dynamic monitoring equipment;

specifically, when it is determined that the electrocardiographic event information is abnormal electrocardiographic event information, if the process of analyzing and processing electrocardiographic data is executed in the dynamic monitoring device, the dynamic monitoring device directly outputs corresponding alarm information generated according to the abnormal electrocardiographic event information.

If the process of analyzing and processing the electrocardiogram data is executed in the server, the server generates corresponding alarm information according to the electrocardiogram abnormal event information, and sends the alarm information to the dynamic monitoring equipment in real time according to the equipment ID, and the alarm information is output through the dynamic monitoring equipment. Of course, the abnormal event information of the electrocardiogram can also be synchronously sent, so that more information can be output through the dynamic monitoring equipment.

The output of alarm information lies in that the suggestion is monitored its monitoring data of person and is taken place unusually for monitored person in time knows self situation, also can in time know by other personnel near monitored person simultaneously and monitored person and taken place the monitoring data unusually, thereby realizes real-time abnormal alarm.

More preferably, the alarm information includes: and the equipment ID information, the information of the object to be tested and the information of the electrocardio abnormal event form complete information output.

When other physical sign monitoring data are considered at the same time, when abnormal data exceeding a corresponding set threshold exist in one or more of pulse data, blood pressure data, respiration data, blood oxygen saturation data and body temperature data, the abnormal event information of physical sign monitoring is generated by combining electrocardiogram event data, and the output of alarm information is determined.

Step 160, acquiring the working mode of the dynamic monitoring equipment;

in a specific application scenario, the dynamic monitoring device of the present invention can be configured to operate in different modes, i.e. provide different services according to different operating modes. The operating modes may include at least a local processing mode and a background processing mode.

In the local processing mode, the abnormal event information and the related data are only stored locally and are not reported to the background processing.

And in a background processing mode, the abnormal event information and related data are reported to the server and are dispatched and processed, so that the abnormal condition of the tested object can be timely and effectively treated, and better medical assistance service can be obtained.

The configuration may be done locally at the ambulatory monitoring device or by a background operation when the ambulatory monitoring device is assigned for use by the user. Of course, different operation modes, i.e. different services, are selected, and the charging mode may be different accordingly.

Step 170, when the local processing mode is adopted, the dynamic monitoring equipment records and stores electrocardiogram event data and alarm information;

step 180, when the monitoring device is in the background processing mode, the server determines the user ID of the responsible user corresponding to the information of the object to be detected according to the alarm information or according to the received alarm signal triggered by the dynamic monitoring device based on the alarm information, generates notification information and sends the notification information to the user equipment of the responsible user and/or the user equipment of a preset association mechanism;

specifically, if the dynamic monitoring device generates the alarm information, the dynamic monitoring device may send the alarm information to the server, or the dynamic monitoring device sends an alarm signal to the server to inform the server that the alarm information is generated, and the alarm signal carries the device ID information and the information of the object to be measured. At this time, the server determines information of a responsible user or an association mechanism associated with the measured object according to the measured object information carried in the alarm information or the alarm signal.

If the alarm information is generated in the server, the server can determine information of a responsible user or a related mechanism related to the measured object according to the measured object information in the alarm information.

Here, information on responsible users or preset association facilities associated with the monitored user is stored in advance in the server. The responsible user can be a guardian, a relative, a family doctor and the like of the monitored user, and the user equipment of the responsible user can be a smart phone, a tablet personal computer or other equipment at least having information receiving and displaying functions, which are installed and operated by the people and have corresponding applications; the association mechanism may be a medical mechanism designated by the monitored user, and similarly, the user device of the association structure may also be a smart phone, a tablet computer, or other device having at least information receiving and displaying functions, which is installed and operated with the corresponding application.

And the server determines the position information of the dynamic monitoring equipment according to the equipment ID information, so as to determine the position of the monitored object.

The server generates notification information according to the alarm information or the alarm signal, wherein the notification information at least carries the information of the monitored object and the position information of the position of the monitored object determined according to the equipment ID, so that a receiver, namely a responsible user, can at least know the object with the abnormal alarm, and can quickly contact the monitored object. In a specific application, the information of the object to be tested may include a mobile phone number of the object to be tested, etc. to facilitate contacting the object to be tested and provide help to the object to be tested.

In a preferred scheme, the notification information further comprises the information of the abnormal event of the electrocardiogram, so that a receiver of the notification information can perform pre-judgment according to the information of the abnormal time of the electrocardiogram, the abnormal condition can be known within the first time, and corresponding medical assistance measures can be taken according to the severity of the abnormal condition. The method comprises the steps of sending feedback information to the dynamic monitoring equipment, and prompting a monitored person to take abnormal and temporary coping measures, such as taking medicines, sitting still, seeking medical advice quickly, waiting for the medical institution to get on home, and the like.

Specifically, afterstep 160, the server receives feedback information sent by the user equipment of the responsible user, and sends the feedback information to the dynamic monitoring device according to the device ID information; feedback information is output by the dynamic monitoring device.

The above description details the implementation process of the physical sign information dynamic monitoring method according to the embodiment of the present invention, and for convenience of understanding, a specific example of practical application is described as an example below.

In practical application, the service mode can be further subdivided on the basis of the local processing mode and the background processing mode, so that more customized monitoring service can be provided for users on the basis of the same monitoring equipment.

In a specific example, the following three service modes may be set:

a full service mode, an emergency service mode, and a local service mode.

In a complete service mode, the dynamic monitoring equipment uploads the monitored physical sign monitoring data to the server in real time, the physical sign monitoring data are distributed to the terminal equipment of the association mechanism by the server in real time, 24-hour all-weather monitoring service is achieved, and the physical sign monitoring data are collected once according to a preset time interval, such as every 24 hours, so that monitoring report data of the detected object in a corresponding time period are generated.

When abnormal alarm information is generated, different processing modes can be adopted according to the severity of the alarm information.

For example, if an alarm of an abnormal electrocardio event with a common severity occurs, a remote guidance service can be performed in a mode of information feedback; if a more severe alarm is occurring, a medical aid assignment is made via the server, an on-premises medical service is provided by the associated medical facility, or the monitored person is scheduled for medical attention by the associated medical facility.

In the emergency service mode, the dynamic monitoring device reports the alarm information to the server only when the alarm information is generated, and the same server can adopt different processing modes according to the severity of the alarm information. When no abnormal alarm occurs, the dynamic monitoring equipment only monitors and displays data locally.

In the local service mode, the dynamic monitoring device only carries out local monitoring and only outputs the alarm information locally when generating the alarm information. The monitored person can report to the server in a mode of manually triggering and sending an alarm signal after obtaining the alarm information.

In addition, the dynamic monitoring device can also provide the function of recording and reporting the active event when the user feels uncomfortable, so as to better provide medical assistance service and monitoring for the user.

The dynamic monitoring equipment can be provided with a switch triggered by one key, and after a user triggers the switch, the monitoring equipment is started to monitor user input and generate alarm event record information. Listening devices may include, but are not limited to: a microphone, a camera, a touch screen, a virtual keyboard, etc. The user can describe the symptom of the subjective abnormality by means of video, voice, text input and the like.

Furthermore, the interception of the electrocardiogram data, the pulse data, the blood pressure data, the respiration data, the blood oxygen saturation data and the body temperature data before and after the user triggers the switch can be started simultaneously, and the information input by the user and the information can be generated into alarm event recording information.

And then, sending the alarm event record information to a background server, and distributing the alarm event record information to responsible users and/or user equipment of a preset correlation mechanism so as to dispose in time.

Fig. 6 is a schematic structural diagram of a dynamic monitoring system according to an embodiment of the present invention, where the monitoring system includes one or more dynamic monitoring devices and a server. The server and the dynamic monitoring device respectively comprise: a processor and a memory. The memory may be connected to the processor by a bus. The memory may be a non-volatile memory such as a hard disk drive and a flash memory, in which a software program and a device driver are stored. The software program is capable of performing various functions of the above-described methods provided by embodiments of the present invention; the device drivers may be network and interface drivers. The processor is used for executing a software program, and the software program can realize the method provided by the embodiment of the invention when being executed.

The user-oriented physical sign information dynamic monitoring method and the user-oriented physical sign information dynamic monitoring system provided by the embodiment of the invention adopt data preprocessing, heart beat characteristic detection, interference signal detection, heart beat classification and lead combination based on a deep learning method, heart beat audit, electrocardiogram event and parameter analysis and calculation, finally, automatic electrocardiogram detection and analysis of a complete and rapid process for automatically outputting electrocardiogram event result data, can output alarm based on the electrocardiogram detection and analysis result, or generate alarm by combining blood pressure, blood oxygen, pulse, respiration and body temperature data, and perform information distribution processing based on alarm information based on response processing of the alarm, and comprise terminal equipment which is distributed to a medical institution or a related user of a monitored person, so that the monitored person can obtain effective and timely medical rescue service. The user-oriented dynamic monitoring method and the user-oriented dynamic monitoring system for the physical sign information are used for effectively monitoring physical signs for non-hospitalized people and providing more effective medical guarantee service for the user based on the physical sign monitoring.

Those of skill would further appreciate that the various illustrative components and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied in hardware, a software module executed by a processor, or a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (9)

Translated fromChinese

1.一种面向用户的体征信息动态监护系统，其特征在于，所述系统包括：1. a user-oriented dynamic monitoring system for sign information, characterized in that the system comprises:动态监护设备接收用户输入或者服务器下发的监测基准数据；所述监测基准数据包括被测对象信息和心电异常事件信息；The dynamic monitoring device receives the monitoring benchmark data input by the user or issued by the server; the monitoring benchmark data includes the measured object information and the abnormal ECG event information;所述动态监护设备对被测对象进行监护数据采集，得到所述被测对象的体征监护数据；所述体征监护数据至少包括心电图数据；The dynamic monitoring device collects monitoring data on the measured object, and obtains the monitoring data of physical signs of the measured object; the monitoring data of physical signs includes at least electrocardiogram data;对所述心电图数据进行波群特征识别，并根据心搏数据确定的每个心搏的检测置信度，得到所述心电图数据的特征信号，根据所述特征信号对所述心电图数据进行心搏分类，结合心电图基本规律参考数据得到心搏分类信息，并生成心电图事件数据；所述心电图事件数据包括所述动态监护设备的设备ID信息；Perform wave group feature identification on the electrocardiogram data, and obtain a characteristic signal of the electrocardiogram data according to the detection confidence of each heartbeat determined by the heartbeat data, and classify the electrocardiogram data according to the characteristic signal. , obtain the heartbeat classification information in combination with the reference data of the basic rules of the ECG, and generate the ECG event data; the ECG event data includes the device ID information of the dynamic monitoring device;根据所述心电图事件数据实时确定对应的心电图事件信息，并确定所述心电图事件信息是否为所述心电异常事件信息；Determine the corresponding ECG event information in real time according to the ECG event data, and determine whether the ECG event information is the ECG abnormal event information;当为所述心电异常事件信息时，生成并通过所述动态监护设备输出报警信息；所述报警信息包括所述设备ID信息、所述被测对象信息、所述心电异常事件信息；When it is the abnormal ECG event information, generate and output alarm information through the dynamic monitoring device; the alarm information includes the device ID information, the measured object information, and the ECG abnormal event information;确定所述动态监护设备的工作模式；所述工作模式包括本地处理模式和后台处理模式；Determine the working mode of the dynamic monitoring device; the working mode includes a local processing mode and a background processing mode;当为本地处理模式时，所述动态监护设备对所述心电图事件数据和所述报警信息进行记录存储；When in the local processing mode, the dynamic monitoring device records and stores the ECG event data and the alarm information;当为后台处理模式时，所述服务器根据所述报警信息，或者根据接收到所述动态监护设备基于所述报警信息触发的报警信号，确定所述被测对象信息对应的责任用户的用户ID，并生成通知信息发送给所述责任用户的用户设备和/或预设的关联机构的用户设备；其中，所述通知信息包括所述报警信息和所述责任用户的用户ID；When in the background processing mode, the server determines the user ID of the responsible user corresponding to the measured object information according to the alarm information, or according to receiving an alarm signal triggered by the dynamic monitoring device based on the alarm information, and generate notification information and send it to the user equipment of the responsible user and/or the user equipment of a preset associated organization; wherein, the notification information includes the alarm information and the user ID of the responsible user;其中，所述对所述心电图数据进行波群特征识别，并根据心搏数据确定的每个心搏的检测置信度，得到所述心电图数据的特征信号，根据所述特征信号对所述心电图数据进行心搏分类，结合心电图基本规律参考数据得到心搏分类信息，并生成心电图事件数据具体包括：Wherein, the electrocardiogram data is characterized by wave group feature identification, and the characteristic signal of the electrocardiogram data is obtained according to the detection confidence of each heartbeat determined by the heartbeat data, and the electrocardiogram data is analyzed according to the characteristic signal. Perform heartbeat classification, obtain heartbeat classification information based on the reference data of the basic rules of the ECG, and generate the ECG event data specifically including:将所述心电图数据的数据格式经过重采样转换为预设标准数据格式，并对转换后的预设标准数据格式的心电图数据进行第一滤波处理；converting the data format of the electrocardiogram data into a preset standard data format through resampling, and performing a first filtering process on the converted electrocardiogram data in the preset standard data format;对所述第一滤波处理后的心电图数据进行心搏检测处理，识别所述心电图数据包括的多个心搏数据，每个所述心搏数据对应一个心搏周期，包括相应的P波、QRS波群、T波的幅值和起止时间数据；Perform heartbeat detection processing on the ECG data after the first filtering process, identify a plurality of heartbeat data included in the ECG data, each of the heartbeat data corresponds to a heartbeat cycle, including corresponding P waves, QRS Amplitude and start and end time data of complex, T wave;根据所述心搏数据确定每个心搏的检测置信度；determining the detection confidence of each heartbeat according to the heartbeat data;根据干扰识别二分类模型对所述心搏数据进行干扰识别，得到心搏数据是否存在干扰噪音，以及用于判断干扰噪音的一个概率值；Performing interference identification on the heartbeat data according to the two-class interference identification model, to obtain whether there is interference noise in the heartbeat data, and a probability value for judging the interference noise;根据所述检测置信度确定心搏数据的有效性，并且，根据确定有效的心搏数据的导联参数和心搏数据，基于所述干扰识别的结果和时间规则合并生成心搏时间序列数据；根据所述心搏时间序列数据生成心搏分析数据；Determine the validity of the heartbeat data according to the detection confidence, and, according to the lead parameters and the heartbeat data of the determined valid heartbeat data, merge and generate heartbeat time series data based on the result of the interference identification and the time rule; generating heartbeat analysis data according to the heartbeat time series data;根据心搏分类模型对所述心搏分析数据进行幅值和时间表征数据的特征提取和分析，得到所述心搏分析数据的一次分类信息；According to the heartbeat classification model, the feature extraction and analysis of the amplitude and time representation data are performed on the heartbeat analysis data to obtain primary classification information of the heartbeat analysis data;对所述一次分类信息结果中的特定心搏的心搏分析数据输入到ST段和T波改变模型进行识别，确定ST段和T波评价信息；Inputting the heartbeat analysis data of the specific heartbeat in the first classification information result into the ST segment and T wave change model for identification, and determining the ST segment and T wave evaluation information;根据所述心搏时间序列数据，对所述心搏分析数据进行P波和T波特征检测，确定每个心搏中P波和T波的详细特征信息，详细特征信息包括幅值、方向、形态和起止时间的数据；According to the heartbeat time series data, the P wave and T wave characteristic detection is performed on the heartbeat analysis data, and the detailed characteristic information of the P wave and T wave in each heartbeat is determined, and the detailed characteristic information includes the amplitude, direction, Data on shape and start and end times;对所述心搏分析数据在所述一次分类信息下根据所述心电图基本规律参考数据、所述P波和T波的详细特征信息以及所述ST段和T波评价信息进行二次分类处理，得到心搏分类信息；Perform secondary classification processing on the heartbeat analysis data under the primary classification information according to the ECG basic law reference data, the detailed feature information of the P wave and T wave, and the ST segment and T wave evaluation information, Get heartbeat classification information;对所述心搏分类信息进行分析匹配，生成所述心电图事件数据。The heartbeat classification information is analyzed and matched to generate the electrocardiogram event data.2.根据权利要求1所述的面向用户的体征信息动态监护系统，其特征在于，在所述动态监护设备接收用户输入或者服务器下发的监测基准数据之前，所述系统还包括：2. The dynamic monitoring system for user-oriented vital information according to claim 1, wherein before the dynamic monitoring device receives the monitoring reference data input by the user or issued by the server, the system further comprises:确定所述被测对象的被测对象信息。Determine the measured object information of the measured object.3.根据权利要求1所述的面向用户的体征信息动态监护系统，其特征在于，所述系统还包括：3. The user-oriented dynamic monitoring system for vital signs information according to claim 1, wherein the system further comprises:所述服务器接收所述责任用户的用户设备发送的反馈信息，并根据所述设备ID信息将所述反馈信息发送给所述动态监护设备；The server receives the feedback information sent by the user equipment of the responsible user, and sends the feedback information to the dynamic monitoring device according to the device ID information;所述动态监护设备输出所述反馈信息。The dynamic monitoring device outputs the feedback information.4.根据权利要求1所述的面向用户的体征信息动态监护系统，其特征在于，所述生成通知信息发送给所述责任用户的用户设备和/或预设的关联机构的用户设备具体包括：4. The user-oriented dynamic monitoring system for physical sign information according to claim 1, wherein the generating notification information and sending the user equipment to the responsible user and/or the user equipment of a preset associated organization specifically include:所述服务器根据所述被测对象信息和/或设备ID信息，确定所述动态监护设备的位置信息；The server determines the location information of the dynamic monitoring device according to the measured object information and/or device ID information;根据所述位置信息和所述报警信息，生成所述通知信息；generating the notification information according to the location information and the alarm information;根据所述责任用户的用户ID将所述通知信息发送给所述责任用户的用户设备和/或预设的关联机构的用户设备。The notification information is sent to the user equipment of the responsible user and/or the user equipment of a preset associated organization according to the user ID of the responsible user.5.根据权利要求1所述的面向用户的体征信息动态监护系统，其特征在于，所述系统还包括：5. The user-oriented dynamic monitoring system for vital signs information according to claim 1, wherein the system further comprises:所述动态监护设备按照预设时间信息，将预设时段内的心电图数据发送给所述服务器；The dynamic monitoring device sends the ECG data within the preset time period to the server according to the preset time information;所述服务器根据所述心电图数据生成预设时段内所述被测对象的监测报告数据。The server generates monitoring report data of the measured object within a preset time period according to the electrocardiogram data.6.根据权利要求1所述的面向用户的体征信息动态监护系统，其特征在于，所述系统还包括：6. The user-oriented dynamic monitoring system for vital signs information according to claim 1, wherein the system further comprises:所述动态监护设备接收用户触发的事件记录信号，监听并生成报警事件记录信息；The dynamic monitoring device receives the event record signal triggered by the user, monitors and generates alarm event record information;将所述报警事件记录信息发送给所述责任用户的用户设备和/或预设的关联机构的用户设备。Sending the alarm event record information to the user equipment of the responsible user and/or the user equipment of a preset association organization.7.根据权利要求1所述的面向用户的体征信息动态监护系统，其特征在于，所述体征监护数据还包括：脉搏数据、呼吸数据、血氧饱和度数据和体温数据中的一种或多种。7. The user-oriented dynamic monitoring system for physical sign information according to claim 1, wherein the physical sign monitoring data further comprises: one or more of pulse data, respiration data, blood oxygen saturation data and body temperature data kind.8.根据权利要求7所述的面向用户的体征信息动态监护系统，其特征在于，所述系统还包括：8. The user-oriented dynamic monitoring system for vital signs information according to claim 7, wherein the system further comprises:所述动态监护设备确定所述脉搏数据、呼吸数据、血氧饱和度数据和体温数据中的一个或多个存在超出相应的设定阈值的异常数据，并结合所述心电图事件数据生成体征监测异常事件信息，并发送到所述服务器。The dynamic monitoring device determines that one or more of the pulse data, respiration data, blood oxygen saturation data and body temperature data have abnormal data that exceeds a corresponding set threshold, and generates signs to monitor abnormality in combination with the ECG event data event information and sent to the server.9.根据权利要求1-8任一所述的面向用户的体征信息动态监护系统，其特征在于，所述动态监护系统包括一个或多个动态监护设备和服务器。9 . The user-oriented dynamic monitoring system for vital information according to claim 1 , wherein the dynamic monitoring system comprises one or more dynamic monitoring devices and servers. 10 .

Images