User terminal electrocardio data analysis system based on cloud serverTechnical Field
The invention relates to the technical field of data analysis, in particular to a user terminal electrocardiograph data analysis system based on a cloud server.
Background
With the popularization of health examination, more and more people begin to pay attention to management of body data, and portable body detection instruments such as portable blood pressure meters, portable blood oxygen meters, portable electrocardiographs and the like are also layered endlessly, wherein unlike portable blood pressure meters, portable blood oxygen meters for directly observing data, portable electrocardiographs generally measure complex electrocardiographic data, and cannot be understood in the absence of professional knowledge. Therefore, the conventional electrocardiograph often uses a joint analysis system to automatically give more accurate diagnosis results so as to be convenient for common users to use, the conventional analysis system is established on the basis of picture analysis of a neural network model, and the conventional analysis system determines corresponding electrocardiograph events by acquiring characteristic vectors of different electrocardiograph events and comparing the characteristic vectors with characteristic vectors extracted from electrocardiographs of users and outputs statement information of the electrocardiograph events for the users to refer to.
The problem is that the feature vector is extracted in the picture recognition process based on the neural network model and is used as the subsequent comparison, namely, the output electrocardiogram event is the item most similar to the input user electrocardiogram, a special and regular waveform can be found when an abnormal event occurs in the heart, for example, a P wave with regular and normal morphology appears in the electrocardiogram of the sinus rhythm, the general frequency is 40-150 times/min, and the P-R interval is 0.12-0.20 seconds, so that a common analysis system has great advantages in finding the special and regular waveform.
In practical electrocardiography, especially in healthy people, the electrocardiography is irregular and lacks similarity, for example, almost no two P-P gaps of a healthy heart are completely equal, so that the healthy heart is not uniform, and the heart dynamics research also supports the strict periodic non-healthy sign, so that the neural network model based on the feature vector is based on the result of a large amount of data training, but the electrocardiographic data generated by different people at different times and even at different altitude places are changed, and the relatively regular electrocardiographic data is identified as healthy electrocardiographic data, which obviously gives users an error conclusion to generate poor using experience.
In view of the above, the invention provides a cloud server-based user terminal electrocardiographic data analysis system, which can provide a secondary judgment basis for each electrocardiographic data judged to be healthy and give a more accurate analysis result to a user.
Disclosure of Invention
The invention aims to provide a cloud server-based user terminal electrocardiograph data analysis system, which solves the following technical problems:
how to solve the problem of providing a secondary judgment basis for each electrocardiographic data judged to be healthy and giving a more accurate analysis result to a user.
The aim of the invention can be achieved by the following technical scheme:
A user terminal electrocardio data analysis system based on a cloud server comprises:
The system comprises a detection module, a storage module, a diagnosis module, a calculation module and a network module, wherein the diagnosis module is used for outputting a first diagnosis result according to the electrocardiograph data of a user, and the diagnosis result comprises normal and abnormal;
The calculation module calculates a confidence coefficient according to the electrocardiograph data, outputs a second diagnosis result for electrocardiograph data input by a current user based on the confidence coefficient and user history data, and adds a reliability percentage to the first diagnosis result based on the first diagnosis result and the second diagnosis result;
the storage module is used for storing diagnosis results and user electrocardio data.
Through the technical scheme: the confidence coefficient is calculated, the second diagnosis result is output, and the first diagnosis result and the second diagnosis result are combined to further diagnose and analyze the electrocardiogram with the normal first diagnosis result, so that more reference information is given based on the historical data and the current detection result of the user, the user is helped to judge the self-electrocardiographic data more accurately, the problem that abnormal regular signals are generated but normal results are detected is avoided, and the reliability of the analysis result is improved.
As a further technical scheme of the invention: the confidence coefficient obtaining process comprises the following steps:
Acquiring a plurality of waveform data values from input user electrocardio data, and establishing a data set for each waveform data, wherein the data set establishment process comprises arranging a plurality of continuous identical waveform data values corresponding to the process from the beginning of atrial depolarization to the beginning of ventricular depolarization according to a time sequence;
The storage module is used for obtaining the average value of the same-term waveform data values of each diagnosis result under the condition that the first diagnosis result output by the diagnosis module is normal by the current user, and carrying out weighted summation on the average value of the same-term waveform data values of each diagnosis result to obtain a waveform data value reference value of a corresponding term;
and the calculation module calculates and acquires a confidence coefficient according to the data in the data set and the reference value.
As a further technical scheme of the invention: the process of calculating the confidence coefficient comprises the following steps:
By the formula:
Obtaining a confidence coefficient Tru, wherein mu1 and mu2 are preset first correlation coefficients and second correlation coefficients and are constants, g is the total number of current data sets, Nk is the number of digital elements with periodic sequences in numerical elements in a kth data set, q is the total number of the numerical elements in the kth item data set, t is the median of period lengths corresponding to all the digital elements with periodic sequences,Is the jth numerical element in the kth data set,Is the j+t number element in the kth data set, and Vk is the reference value for the kth item data set.
Through the technical scheme: the invention provides a process for obtaining the confidence coefficient, the confidence coefficient is calculated based on the periodic relation of the digital elements of the data set, the self-correlation degree among the digital elements of the data set can be reflected, the higher the correlation degree among the digital elements is, the stronger the periodicity and the correlation among the digital elements are represented, the characteristics of the disorder of the normal heart rate are not met, otherwise, the lower the confidence coefficient is, the more correct the first diagnosis result which is output as normal is indicated, the data sets of a plurality of items can be accurately evaluated in an integral way through the confidence coefficient, and the periodic state of the electrocardiographic data can be globally reflected.
As a further technical scheme of the invention: the process of outputting the second diagnostic result includes:
comparing the confidence coefficient with a preset diagnosis interval;
If the confidence coefficient exceeds the preset diagnosis interval, outputting an abnormal second diagnosis result;
if the confidence coefficient falls into the preset diagnosis interval, outputting a second diagnosis result which is normal;
If the confidence coefficient is smaller than the preset diagnosis interval, outputting a second diagnosis result which is normal.
Through the technical scheme: the invention further distinguishes the second diagnosis result which is output as normal based on the confidence coefficient, can provide more detailed auxiliary evaluation for the first diagnosis result, and helps the user to know the physical state corresponding to the electrocardiogram data more accurately.
As a further technical scheme of the invention: the process of adding a reliability percentage to the first diagnostic result based on the first diagnostic result and the second diagnostic result includes:
comparing the first diagnosis result with the second diagnosis result to obtain a basic value of the reliability percentage;
Carrying out normalization calculation through the confidence coefficient to obtain a correction value;
and calculating the basic value and the correction value to obtain a reliability percentage and adding the reliability percentage to the first diagnosis result and the second diagnosis result.
As a further technical scheme of the invention: the process of obtaining the base value of the reliability percentage comprises the following steps:
If the first diagnosis result is the same as the second diagnosis result, selecting from the interval [ a1,a2 ] from the basic value a, wherein when the confidence coefficient of the second diagnosis result is smaller than the preset diagnosis interval, the basic value a1 is taken, and when the confidence coefficient of the second diagnosis result falls into the preset diagnosis interval, the basic value a is selected from (a1,a2);
If the first diagnosis result is normal and the second diagnosis result is abnormal, the basic value is b.
As a further technical scheme of the invention: the process of obtaining the correction value through normalization calculation by the confidence coefficient comprises the following steps:
By the formula:
A correction value ΔCro is obtained, where a is the base of the logarithmic function and a > 1, y1 is the current user age, y2 is the current user gender, f is the state value conversion function, and U is the rounding function.
As a further technical scheme of the invention: the process of obtaining the reliability percentage comprises the following steps:
if the first diagnosis result is the same as the second diagnosis result, the reliability percentage is
If the first diagnosis result is normal and the second diagnosis result is abnormal, the basic value is (b-deltaCro)%.
Through the technical scheme: the reliability percentage is generated based on the current data of the user, the data highly related to the electrocardiographic data of the user can reflect the accuracy of the first diagnosis result, namely, the first diagnosis result is subjected to further subdivision evaluation, and electrocardiographic data which is partially regular but is judged to be normal heart rhythm by a diagnosis module can be found and reminded of the user in time, so that the user experience is improved, the physical health of the user is further maintained, and the output diagnosis result is clear, so that the method is obtained.
As a further technical scheme of the invention: the storage module does not store user data having a reliability percentage below 60 percent, and the calculation module alerts the user to the inaccuracy of the first diagnostic result when the reliability percentage is below 60 percent, and simultaneously suggests advice including re-measurement and hospital-going measurement.
The invention has the beneficial effects that:
(1) According to the invention, the confidence coefficient is calculated, the second diagnosis result is output, and the first diagnosis result and the second diagnosis result are combined to further diagnose and analyze the electrocardiogram with the normal first diagnosis result, so that more reference information is given based on the historical data and the current detection result of the user, the user is helped to judge the self-electrocardiographic data more accurately, the problem that abnormal regular signals are generated but normal results are detected is avoided, and the reliability of the analysis result is improved.
(2) The confidence coefficient is calculated based on the periodic relation of the digital elements of the data set, the autocorrelation among the digital elements of the data set can be reflected, the higher the correlation among the digital elements is, the stronger the periodicity and the correlation among the digital elements are, the less the confidence coefficient is in accordance with the characteristic that the normal heart rate is disordered, otherwise, the lower the confidence coefficient is, the more the first diagnosis result which is output as normal is correct, the integrity evaluation can be accurately carried out on the data sets of a plurality of items through the confidence coefficient, and the periodic state of the electrocardiographic data can be globally reflected.
(3) According to the invention, the second diagnosis result is used as the reference quantity of the auxiliary first diagnosis result and is determined according to the electrocardiographic data of the user through the process of outputting the second diagnosis result by the confidence coefficient, and the second diagnosis result which is normally output is further distinguished based on the confidence coefficient, so that more detailed auxiliary evaluation can be provided for the first diagnosis result, and the user is helped to know the physical state corresponding to the electrocardiographic data more accurately.
(4) The reliability percentage of the invention is based on the current data of the user, the data highly related to the electrocardiographic data of the user can reflect the accuracy of the first diagnosis result, namely, the first diagnosis result is normal, further subdivision evaluation is carried out, and electrocardiographic data which is partially regular but is judged to be normal rhythm by the diagnosis module can be found and reminded of the user in time, thus improving the user experience, further maintaining the physical health of the user, and the output diagnosis result is clear, thus obtaining the heart rhythm diagnosis device.
Drawings
The invention is further described below with reference to the accompanying drawings.
FIG. 1 is a schematic diagram of the overall module relationship of the present invention;
FIG. 2 is a flow chart of a confidence coefficient acquisition process of the present invention;
FIG. 3 is a process flow diagram of an additional reliability percentage of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1-3, in one embodiment, a system for analyzing electrocardiographic data of a user terminal based on a cloud server is provided, including: the system comprises a detection module, a storage module, a diagnosis module, a calculation module and a network module for remote data receiving and transmitting, wherein the detection module, the storage module, the diagnosis module, the calculation module and the network module for acquiring the user electrocardio data are used for outputting a first diagnosis result according to the user electrocardio data, the diagnosis result comprises normal and abnormal diagnosis results, the diagnosis result is acquired based on a trained neural network model, corresponding training is carried out through different types of electrocardiograph events, so that corresponding new electrocardiograph events can be output after electrocardiograph pictures are input, statement information of the corresponding electrocardiograph events is used as the first diagnosis result, the detailed process is omitted, wherein health represents a normal state, and electrocardiograph events except for healthy electrocardiograph such as sinus rhythm and the like are marked as abnormal;
the calculation module calculates a confidence coefficient according to the electrocardiograph data, outputs a second diagnosis result for electrocardiograph data input by a current user based on the confidence coefficient and user history data, and adds a reliability percentage to the first diagnosis result based on the first diagnosis result and the second diagnosis result;
The storage module is used for storing diagnosis results and user electrocardio data, and is also provided with a display for displaying the diagnosis results and the user electrocardio data.
In this embodiment, further diagnosis and analysis are performed on the electrocardiogram with the normal first diagnosis result by calculating the confidence coefficient and outputting the second diagnosis result and combining the first diagnosis result and the second diagnosis result, so that multiple reference information is given based on the historical data and the current detection result of the user, the user is helped to more accurately judge the self-electrocardiographic data, the problem that abnormal regular signals are generated but normal results are detected is avoided, and reliability of the analysis result is improved.
Referring to fig. 2, the process of obtaining the confidence coefficient includes:
Acquiring a plurality of waveform data values from input user electrocardiographic data, and establishing a data set for each waveform data, wherein the data set establishment process comprises arranging a plurality of continuous homonymous waveform data values corresponding to the process from the beginning of atrial depolarization to the beginning of ventricular depolarization according to a time sequence, for example, the data set of interval time R-R of two heartbeats can be expressed as R= { R1,r2,……,rn } and the sum of peak-peak values P-P of adjacent peaks is expressed as P= { P1,p2,……,pm } and the like;
under the condition that the first diagnosis result output by the diagnosis module is normal, the storage module obtains the average value of the same-term waveform data values of each diagnosis result, and performs weighted summation to obtain the waveform data value reference value of the corresponding term, wherein each weight value in the weighted summation process is reduced according to the increase of the difference between the time of the obtained diagnosis result and the current diagnosis time;
And the calculation module calculates and acquires a confidence coefficient according to the data in the data set and the reference value.
The process of calculating the confidence coefficient comprises the following steps:
By the formula:
Obtaining a confidence coefficient Tru, wherein mu1 and mu2 are preset first correlation coefficients and second correlation coefficients and are constants, g is the total number of current data sets, Nk is the number of digital elements with periodic sequences in numerical elements in a kth data set, q is the total number of the numerical elements in the kth item data set, t is the median of period lengths corresponding to all the digital elements with periodic sequences,Is the jth numerical element in the kth data set,Is the j+t number element in the kth data set, and Vk is the reference value of the kth item data set, where X may be the R-R value or the P-P value.
Through the technical scheme: in this embodiment, a process of obtaining a confidence coefficient is provided, the confidence coefficient is calculated based on a periodic relationship of digital elements of a data set, so that the self-correlation degree between the digital elements of the data set can be reflected, the higher the correlation degree between the digital elements is, the stronger the periodicity and the correlation between the digital elements are represented, the characteristics of the disorder of normal heart rate are not met, otherwise, the lower the confidence coefficient is, the more correct the first diagnosis result output as normal is indicated, the integrity evaluation can be accurately carried out on the data sets of a plurality of items through the confidence coefficient, and the periodic state of electrocardiographic data can be globally reflected.
The process of outputting the second diagnostic result includes: comparing the confidence coefficient with a preset diagnosis interval; if the confidence coefficient exceeds the preset diagnosis interval, outputting an abnormal second diagnosis result; if the confidence coefficient falls into the preset diagnosis interval, outputting a second diagnosis result which is normal; if the confidence coefficient is smaller than the preset diagnosis interval, outputting a second diagnosis result which is normal.
Through the technical scheme: in this embodiment, a process of outputting a second diagnosis result through a confidence coefficient is provided, the second diagnosis result is used as a reference for assisting the first diagnosis result, and is determined according to the electrocardiographic data of the user.
Referring to fig. 3, the process of adding a reliability percentage to the first diagnostic result based on the first diagnostic result and the second diagnostic result includes:
comparing the first diagnosis result with the second diagnosis result to obtain a basic value of the reliability percentage;
Carrying out normalization calculation through the confidence coefficient to obtain a correction value;
and calculating the basic value and the correction value to obtain a reliability percentage and adding the reliability percentage to the first diagnosis result and the second diagnosis result.
The process of obtaining the base value of the reliability percentage comprises the following steps:
If the first diagnosis result is the same as the second diagnosis result, selecting from the interval [ a1,a2 ] from the basic value a, wherein when the confidence coefficient of the second diagnosis result is smaller than the preset diagnosis interval, the basic value a1 is taken, and when the confidence coefficient of the second diagnosis result falls into the preset diagnosis interval, the basic value a is selected from (a1,a2), and it is to be noted that the basic value and the confidence coefficient are positively correlated, in this embodiment, linear correlation is adopted as a selection rule selected from (a1,a2), and obviously other types of correlation functions can be selected;
if the first diagnosis result is normal and the second diagnosis result is abnormal, taking a basic value as b, wherein a1,a2 and b are constants selected according to the historical data of the current user, and the selected values are different for different users.
In this embodiment, the process of obtaining the correction value by performing normalization calculation through the confidence coefficient includes:
By the formula:
The correction value ΔCro is obtained, where a is the base of the logarithmic function and a > 1, y1 is the current user age, y2 is the current user gender, f is the state value conversion function, preferably the look-up table function, and U is the rounding function, in this embodiment rounding down.
The process for obtaining the reliability percentage comprises the following steps:
if the first diagnosis result is the same as the second diagnosis result, the reliability percentage is
If the first diagnosis result is normal and the second diagnosis result is abnormal, the basic value is (b-deltaCro)%.
Through the technical scheme: in the embodiment, the acquisition process of the reliability percentage is provided, specifically, the basic value of the reliability percentage is firstly acquired, then the correction value is acquired through normalization calculation by the confidence coefficient, and finally the reliability percentage is acquired through calculation of the basic value and the correction value, the reliability percentage is generated based on the current data of the user and is highly related to the electrocardiographic data of the user, the accuracy of the first diagnosis result can be reflected, namely, the condition that the first diagnosis result is normal is further subdivided and evaluated, the electrocardiographic data of which part is extremely regular but is judged to be normal rhythm by the diagnosis module can be found, the user is timely found and reminded, the user experience is improved, the physical health of the user is further maintained, and the output diagnosis result is clear, so that the heart rhythm diagnosis device is obtained.
Wherein the storage module does not store user data having a reliability percentage below 60 percent and the calculation module alerts the user that the first diagnostic result is inaccurate when the reliability percentage is below 60 percent while providing advice including re-measurement and hospital-going measurements, it should be understood that a user having a reliability percentage above 60 percent still cannot be obtained for a number of repeated measurements if a heart related medical history is present or the first diagnostic result is abnormal should be advised to go to a hospital measurement.
The foregoing describes one embodiment of the present invention in detail, but the description is only a preferred embodiment of the present invention and should not be construed as limiting the scope of the invention. All equivalent changes and modifications within the scope of the present invention are intended to be covered by the present invention.