A pulse data processing method and system for remote pulse diagnosis

Abstract

The present invention discloses a pulse data processing method and system for remote pulse diagnosis, which relates to the field of data processing technology. The method includes using a dynamic iterative sampling method to adjust the sampling frequency of a pulse detector in real time, and obtain a patient's real-time pulse signal; sending the real-time pulse signal to a remote workstation, so that the remote workstation can parse the pulse waveform curve of the patient in at least one state according to the real-time pulse signal, and input the patient's pulse waveform curve into a prediction model, output an abnormal pulse value, set an abnormal pulse threshold, and determine whether the patient's pulse waveform curve is abnormal, so as to remind the patient. The present invention can collect the patient's real-time pulse signal through dynamic iterative sampling, and perform data analysis on the patient's real-time pulse signal to determine whether the patient's pulse waveform curve is abnormal, so as to help the patient better understand his or her pulse state.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a pulse data processing method and system for remote pulse diagnosis. Background Art

[0002] With the continuous advancement of telemedicine technology and the rapid development of Internet communication technology, telemedicine services have become an important means to solve the uneven distribution of medical resources, improve medical efficiency, and improve patients' medical experience. In telemedicine, the remote cloud clinic workstation plays a vital role as a bridge connecting patients and doctors. As one of the commonly used physiological parameter monitoring devices in telemedicine, the accuracy and real-time data processing of the pulse detector are crucial to the diagnosis and treatment effect of telemedicine.

[0003] For example, a Chinese patent with publication number CN114937488A discloses a pulse data processing method, device and storage medium, the method comprising: receiving a pulse calibration data group; performing nonlinear fitting on the pulse calibration data group to calculate an intermediate calibration function of the calibration output data with respect to the calibration feedback data; calculating output reference data corresponding to the calibration output data according to the intermediate calibration function; performing linear fitting with the output reference data as explained variables and the calibration output data as explanatory variables to generate an output calibration function; receiving pressure output data, and calculating and outputting corresponding pressure calibration data according to the output calibration function and the pressure output data. The pulse data processing method provided by the invention can improve the consistency of data, keep the data type and format uniform, and improve the work efficiency of analysis and the accuracy of output results.

[0004] The above patent has the problem raised by the background technology: it relies on the algorithm on the local device to process and analyze the data, and then sends the results to the doctor through remote communication. However, due to performance limitations, data transmission delays, and insufficient data processing accuracy of the local device, the accuracy and real-time performance of the data processing are low, and the sampling frequency remains unchanged and cannot guarantee the collection of complete patient pulse signal values. To solve these problems, the present invention proposes a pulse data processing method and system for remote pulse diagnosis. Summary of the invention

[0005] Summary of the invention In view of the deficiencies in the prior art, main purpose of the present invention is to provide a pulse data processing method and system for remote pulse diagnosis, which can effectively solve the problems in the background technology. Specific technical scheme of the present invention is as follows:

[0006] A pulse data processing method for remote pulse diagnosis is used for a pulse data adjustment device, wherein the pulse data adjustment device is used to control a pulse detector, and the pulse detector is used to collect a patient's pulse signal. The pulse data processing method comprises:

[0007] The sampling frequency of the pulse detector is adjusted in real time by using a dynamic iterative sampling method, and the real-time pulse signal of the patient is obtained;

[0008] The real-time pulse signal is sent to a remote workstation, so that the remote workstation parses the pulse waveform curve of the patient in at least one state according to the real-time pulse signal, and inputs the pulse waveform curve of the patient into a prediction model, outputs an abnormal pulse value, sets an abnormal pulse threshold, and determines whether the pulse waveform curve of the patient is abnormal, so as to remind the patient.

[0009] Specifically, the dynamic iterative sampling method is used to adjust the sampling frequency of the pulse detector in real time and obtain the real-time pulse signal of the patient, including:

[0010] Divide the patient's pulse signal according to fixed duration to obtain N time intervals;

[0011] Set the sampling frequency initial value k0, sample the patient's pulse signal, obtain N pulse signals collected in N time intervals, and count the number of pulse signal peak points in each time interval, where the number of pulse signal peak points in the nth time interval is m. n , n is an integer greater than 0 and less than or equal to N;

[0012] When m n Less than or equal to m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n -q n k0, when m n Greater than m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n +q n k0, where m n-1 is the number of occurrences of the pulse signal peak point in the n-1th time interval, k n is the sampling frequency of the nth time interval, q n is the degree of change in the number of occurrences of the peak point of the pulse signal in the nth time interval;

[0013] This process is repeated to dynamically iterate the sampling frequency and obtain the complete real-time pulse signal of the patient.

[0014] Specifically, the sampling frequency initial value k0 is set to sample the patient's pulse signal, and the N pulse signals obtained by sampling in N time intervals include:

[0015] Set a sliding window, the size of the sliding window is T;

[0016] The sliding window is used to slide on the patient's real-time pulse signal to gradually cover the patient's entire real-time pulse signal.

[0017] Specifically, when m n Less than or equal to m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n -q n k0, when m n Greater than m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n +q n k0, where q n The calculation method is: Among them, m n is the number of occurrences of the pulse signal peak point in the nth time interval, m n-1 is the number of occurrences of the pulse signal peak point in the n-1th time interval.

[0018] A pulse data processing method for remote pulse diagnosis is used in a remote workstation, wherein the remote workstation is connected to a pulse data adjustment device for communication, and the pulse data adjustment device is used to control a pulse detector. The pulse data processing method comprises:

[0019] Receiving a real-time pulse signal sent by the pulse data adjustment device, wherein the real-time pulse signal is obtained by the pulse detector collecting the patient's pulse signal after the pulse data adjustment device adjusts the sampling frequency of the pulse detector in real time using a dynamic iterative sampling method;

[0020] Analyzing the pulse waveform of the patient in at least one state according to the real-time pulse signal to obtain a pulse waveform curve of the patient;

[0021] The patient's pulse waveform curve is input into the prediction model, the pulse abnormality value is output, the pulse abnormality threshold is set, and it is determined whether the patient's pulse waveform curve is abnormal, so as to remind the patient.

[0022] Specifically, the step of analyzing the pulse waveform of the patient in at least one state according to the real-time pulse signal to obtain the pulse waveform curve of the patient includes:

[0023] The received real-time pulse signal is pre-processed by using wavelet transform to remove noise, and a pulse waveform curve y=f(x) is drawn in chronological order based on the pre-processed real-time pulse signal, wherein the horizontal axis of the pulse waveform curve represents the time x, and the vertical axis represents the pulse signal value y corresponding to each time;

[0024] Analyzing the pulse waveform curve of the patient to determine that the pulse waveform curve includes a pulse waveform of the patient in at least one state, wherein the patient's state types include a calm state and a motion state;

[0025] Calculate the variance of the pulse waveform curve at each moment and set the pulse waveform variance threshold U;

[0026] The time point that satisfies the conditions that the variance is greater than the pulse waveform variance threshold U, the pulse signal value is less than or equal to the pulse waveform peak value, and the first-order derivative of the pulse waveform is greater than 0 is taken as the starting point x1 of the motion state pulse waveform;

[0027] The time point that satisfies the conditions that the variance is greater than the pulse waveform variance threshold U, the pulse signal value is less than or equal to the pulse waveform peak value, and the first-order derivative of the pulse waveform is less than 0 is taken as the end point x2 of the motion state pulse waveform;

[0028] The pulse waveform curve area between the time points x1 and x2 is taken as the pulse waveform area in the motion state, and the other areas are the calm state areas, and the pulse waveform curve of the patient after analysis is obtained.

[0029] Specifically, the patient's pulse waveform curve is input into the prediction model, the abnormal pulse value is output, the abnormal pulse threshold is set, and whether the patient's pulse waveform curve is abnormal is determined to remind the patient, including:

[0030] Inputting the patient's pulse waveform curve into the prediction model to calculate the abnormal pulse value of the patient in at least one state;

[0031] The patient's pulse abnormality value is calculated by combining the patient's pulse abnormality values ​​in different states using the weight method;

[0032] An abnormal pulse threshold is set, and whether the patient's pulse waveform curve is abnormal is determined based on the patient's abnormal pulse value and the abnormal pulse threshold. When the patient's pulse waveform curve is abnormal, the patient is reminded.

[0033] Specifically, the prediction model uses the area method to calculate the area enclosed by the pulse waveform curve and the horizontal axis in at least one state of the patient, and uses the area as the abnormal pulse value of the patient in this state.

[0034] Specifically, the abnormal pulse threshold includes a maximum abnormal pulse threshold and a minimum abnormal pulse threshold. When the abnormal pulse value of the patient is higher than the maximum abnormal pulse threshold or lower than the minimum abnormal pulse threshold, the patient's pulse is regarded as abnormal.

[0035] A pulse data processing system for remote pulse feeling, used to implement a pulse data processing method for remote pulse feeling, the system comprising:

[0036] The pulse data adjustment device uses a dynamic iterative sampling method to adjust the sampling frequency of the pulse detector in real time, obtains the patient's real-time pulse signal, and sends the real-time pulse signal to a remote workstation;

[0037] The remote workstation receives the real-time pulse signal sent by the pulse data adjustment device, analyzes the pulse waveform curve of the patient in at least one state according to the real-time pulse signal, inputs the pulse waveform curve of the patient into the prediction model, outputs the abnormal pulse value, sets the abnormal pulse threshold, determines whether the pulse waveform curve of the patient is abnormal, and reminds the patient.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] The present invention uses a pulse data adjustment device to adjust the sampling frequency of the pulse detector in real time by using a dynamic iterative sampling method, collects the patient's real-time pulse signal, and sends it to a remote workstation; the remote workstation parses the pulse waveform curve of the patient in at least one state according to the received real-time pulse signal, and inputs the patient's pulse waveform curve into a prediction model, outputs a pulse abnormality value, sets a pulse abnormality threshold, determines whether the patient's pulse waveform curve is abnormal, and reminds the patient when the patient's pulse waveform curve is abnormal. The present invention can collect the patient's complete real-time pulse signal by adjusting the sampling frequency in real time, control the sampling frequency within a reasonable range, reduce space and resource waste, and improve data processing efficiency; parse the pulse waveform area of ​​the patient in at least one state, analyze the patient's pulse state according to the patient's pulse waveform curve, determine whether the patient's pulse waveform curve is abnormal, and remind the patient in time when the patient's pulse waveform curve is abnormal, so that the patient can find problems through self-detection and control his own pulse condition in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A work flow chart of a pulse data processing method for remote pulse diagnosis of the present invention;

[0041] Figure 2 A workflow diagram of a dynamic iteration sampling method for a pulse data processing method for remote pulse diagnosis of the present invention;

[0042] Figure 3 A schematic diagram of a sliding window technology application of a pulse data processing method for remote pulse diagnosis of the present invention;

[0043] Figure 4 A work flow chart of a remote workstation of a pulse data processing method for remote pulse feeling of the present invention;

[0044] Figure 5 The present invention is a module schematic diagram of a pulse data processing system for remote pulse diagnosis. DETAILED DESCRIPTION

[0045] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0046] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0047] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0048] Example 1

[0049] This embodiment provides a pulse data processing method for remote pulse diagnosis, the specific scheme is as follows: Figure 1 As shown, a pulse data processing method for remote pulse diagnosis, the pulse data processing method comprising:

[0050] S101, the pulse data adjustment device uses a dynamic iterative sampling method to adjust the sampling frequency of the pulse detector in real time and obtain the patient's real-time pulse signal;

[0051] S102. The pulse data adjustment device sends the real-time pulse signal to a remote workstation, so that the remote workstation parses the pulse waveform curve of the patient in at least one state according to the real-time pulse signal, and inputs the pulse waveform curve of the patient into a prediction model, sets an abnormal pulse threshold, determines whether the pulse waveform curve of the patient is abnormal, and reminds the patient.

[0052] In this embodiment, the pulse data adjustment device refers to a device that controls the sampling frequency of the pulse detector and has data transceiver functions. Among them, the pulse detector in this embodiment is portable and can be carried by the patient. The patient collects the pulse signal by wearing the pulse detector. Due to the uncertainty of the patient's state, the pulse signal frequency in different states is also different. When the pulse signal frequency is accelerated, the sampling frequency remains unchanged, which may lead to the loss of the collected pulse signal. When the pulse signal frequency slows down, the sampling frequency remains unchanged, which may cause problems such as waste of resources. Therefore, the pulse data adjustment device adopts a dynamic iterative sampling method to adjust the sampling frequency of the pulse signal in real time. Real time means that within a shorter preset time interval, for example, 1 / 60 second is a real-time frame, and the steps processed within the same 1 / 60 second are all real-time, ensuring that the patient's complete real-time pulse signal value can be collected, and improving the sampling efficiency, which helps to control resource consumption.

[0053] The pulse data adjustment device sends the collected real-time pulse signal to a remote workstation, which processes and analyzes the received real-time pulse signal. The patient's state includes a calm state and a motion state. The pulse waveform curve of the patient in at least one state is obtained by analysis, and the pulse waveform curve of the patient is input into a prediction model. The prediction model can be a differential equation, a difference equation, a difference fitting model, etc. Specifically, the prediction model outputs the patient's pulse abnormality value using an integral method, and determines whether the patient's pulse waveform curve is abnormal based on a set pulse abnormality threshold. When the patient's pulse waveform curve is abnormal, the patient is reminded.

[0054] The present invention uses a pulse data adjustment device to adjust the sampling frequency of the pulse detector in real time by using a dynamic iterative sampling method, collects the patient's real-time pulse signal, and sends it to a remote workstation; the remote workstation parses the pulse waveform curve of the patient in at least one state according to the received real-time pulse signal, and inputs the patient's pulse waveform curve into a prediction model, outputs a pulse abnormality value, sets a pulse abnormality threshold, determines whether the patient's pulse waveform curve is abnormal, and reminds the patient when the patient's pulse waveform curve is abnormal. The present invention can collect the patient's complete real-time pulse signal by adjusting the sampling frequency in real time, control the sampling frequency within a reasonable range, reduce space and resource waste, and improve data processing efficiency; parse the pulse waveform area of ​​the patient in at least one state, analyze the patient's pulse state according to the patient's pulse waveform curve, determine whether the patient's pulse waveform curve is abnormal, and remind the patient in time when the patient's pulse waveform curve is abnormal, so that the patient can find problems through self-detection and control his own pulse condition in real time.

[0055] In this embodiment, if Figure 2The pulse data adjustment device uses a dynamic iterative sampling method to adjust the sampling frequency of the pulse detector in real time and obtain the patient's real-time pulse signal, including:

[0056] S201, dividing the patient's pulse signal according to fixed time lengths to obtain N time intervals;

[0057] S202, set the sampling frequency initial value k0, sample the patient's pulse signal, obtain N pulse signals collected in N time intervals, and count the number of pulse signal peak points in each time interval, where the number of pulse signal peak points in the nth time interval is m. n , n is an integer greater than 0 and less than or equal to N;

[0058] S203, when m n Less than or equal to m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n -q n k0, when m n Greater than m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n +q n k0, where m n-1 is the number of occurrences of the pulse signal peak point in the n-1th time interval, k n is the sampling frequency of the nth time interval, q n is the degree of change in the number of occurrences of the peak point of the pulse signal in the nth time interval;

[0059] S204. Repeat S202-S203, dynamically iterate the sampling frequency, collect the complete patient's real-time pulse signal, avoid signal loss, and help control resource consumption.

[0060] The pulse signal data adjustment device divides the sampling duration according to a set fixed duration, obtaining a total of N time intervals. Set the initial sampling frequency value k0. The sampling frequency k1 of the first time interval is k0, and the sampling frequency k2 of the second time interval is k1. Sample the pulse signals in the first two time intervals and count the number of occurrences m1 and m2 of the peak points of the pulse signals in the first two time intervals. When m1≥m2, at the same sampling frequency, the number of occurrences of the peak points of the pulse signal decreases, indicating that the patient's pulse shows a slowing trend. If a relatively high sampling frequency is still maintained, there will be a problem of resource waste. Therefore, reduce the sampling frequency k3 of the next time interval to k3 = k2 - q2k0, which helps to control resource consumption. When m1<m2, at the same sampling frequency, the number of occurrences of the peak points of the pulse signal increases, indicating that the patient's pulse shows an accelerating trend. If a relatively low sampling frequency is still maintained, it may lead to missing signals in the collected signals. Therefore, increase the sampling frequency k3 of the next time interval to k3 = k2 + q2k0 to avoid signal loss.

[0061] Therefore, repeat this process. When m n-1 ≥m n , the sampling frequency of the (n + 1)-th time interval is k n+1 =k n -q n k0. When m n-1 <m n , the sampling frequency of the (n + 1)-th time interval is k n+1 =k n +q n k0, where n is an integer greater than 0 and less than or equal to N. Stop the iteration until the sampling frequency of the N-th time interval is calculated. This dynamic iterative sampling method realizes the dynamic iteration of the sampling frequency, ensures the collection of the patient's complete real-time pulse signal, avoids signal loss, and helps to control resource consumption.

[0062] For example Figure 3 , in this embodiment, set the initial sampling frequency k0, sample the patient's pulse signal, and correspondingly obtain N pulse signals collected in N time intervals, including:

[0063] S301. Set a sliding window, and the size of the sliding window is T.

[0064] S302. Slide the sliding window on the patient's real-time pulse signal to gradually cover the entire real-time pulse signal of the patient.

[0065] In Figure 3In this example, taking two sliding windows of size T as an example, in the first sliding window, there are more sampling points, 12 in total. Among them, x2 and x4 are peak points. Three points can determine a peak point. According to the three points x1, x2, and x3, x2 can be determined as the peak point. According to the three points x3, x4, and x5, x4 can be determined as the peak point. Then, the other points collected in the first sliding window are redundant points that are not helpful for determining the peak point. Collecting too many redundant points is a waste of resources. When the time period is long enough, by adjusting the sampling frequency, the collected pulse signal can be like the pulse signal in the second sliding window, and all the collected points can play a role in determining the pulse peak point, there are no redundant points, and there is no signal loss. At this time, the sampling frequency is highly consistent with the change of the pulse signal.

[0066] In this embodiment, when m n is less than or equal to m n-1 the sampling frequency of the (n + 1)-th time interval is k n+1 = k n - q n k0. When m n is greater than m n-1 the sampling frequency of the (n + 1)-th time interval is k n+1 = k n + q n k0. In the formula, the calculation method of q n is as follows: where m n is the number of occurrences of the peak point of the pulse signal in the n-th time interval, and m n-1 is the number of occurrences of the peak point of the pulse signal in the (n - 1)-th time interval.

[0067] For example, set the initial sampling frequency k0 = 4Hz. The pulse of a normal adult is 60 to 100 times per minute. The sampling frequency k1 of the first time interval is k0, and the sampling frequency k2 of the second time interval is k1. When the sampling frequency is 4Hz, it corresponds to 240 samples per minute.

[0068] If after sampling, it is statistically found that the number of occurrences of the peak point of the pulse signal in the first time interval m1 = 80 times, and the number of occurrences of the peak point of the pulse signal in the second time interval m2 = 70 times. Without changing the sampling frequency, m2 < m1 indicates that the pulse signal of the patient shows a slow trend. If the sampling frequency remains unchanged in the third time interval, it will cause a waste of resources. Therefore, in the present invention, the sampling frequency of the third time interval is calculated to be approximately 3.73Hz, and the sampling frequency decreases.

[0069] If after sampling, the number of occurrences of the pulse signal peak point in the first time interval is m1=60 times, and the number of occurrences of the pulse signal peak point in the second time interval is m2=70 times, and the sampling frequency remains unchanged, m2>m1 indicates that the patient's pulse signal is accelerating. If the sampling frequency remains unchanged in the third time interval, the collected pulse signal may be missing. Therefore, the present invention sets the sampling frequency of the third time interval to The calculated k3 is about 4.31Hz, and the sampling frequency increases; if after sampling, m2=m1, indicating that the patient's pulse signal is in a stable state, then the calculated sampling frequency k3=k2, and the sampling frequency remains unchanged, ensuring that the patient's complete real-time pulse signal is collected.

[0070] Calculate q based on the difference in the number of peak points in the previous two time intervals n When the difference in the number of peak points in adjacent time intervals is large, it reflects that the patient's pulse signal fluctuation amplitude is large. At this time, the calculated q n is also large, corresponding to the calculated k n+1 The adjustment amplitude is also large; when the difference in the number of peak points in adjacent time intervals is small, it reflects that the fluctuation amplitude of the patient's pulse signal is small. At this time, the calculated q n is also small, corresponding to the calculated k n+1 The adjustment amplitude is also small. This adjustment method can dynamically adjust the sampling frequency according to the fluctuation amplitude of the patient's pulse signal, and dynamically adjust the degree of change of the sampling frequency to make it more consistent with the fluctuation of the patient's pulse signal.

[0071] Further, the remote workstation parses the pulse waveform curve of the patient in at least one state according to the real-time pulse signal, inputs the pulse waveform curve of the patient into the prediction model, outputs the abnormal pulse value, sets the abnormal pulse threshold, and determines whether the pulse waveform curve of the patient is abnormal, so as to remind the patient, including:

[0072] S401, receiving the real-time pulse signal sent by the pulse data adjustment device;

[0073] S402, analyzing the pulse waveform of the patient in at least one state according to the real-time pulse signal to obtain a pulse waveform curve of the patient;

[0074] S403, inputting the patient's pulse waveform curve into the prediction model, outputting the pulse abnormality value, setting the pulse abnormality threshold, determining whether the patient's pulse waveform curve is abnormal, and reminding the patient.

[0075] The remote workstation can receive the real-time pulse signal sent by the pulse data adjustment device and analyze the real-time pulse signal of the patient. The patient may be in a calm state or a state of exercise. According to the pulse waveform characteristics of the patient in different states, the pulse waveform areas of the patient in different states can be analyzed. When the patient has pulse waveform areas in multiple states, the multi-state pulse waveform of the patient can be combined for analysis, and the patient's pulse abnormality value can be calculated using a prediction model. According to the set pulse abnormality threshold, it is determined whether the patient's pulse waveform curve is abnormal. When the patient's pulse waveform curve is abnormal, the patient is reminded.

[0076] In this embodiment, the step of analyzing the pulse waveform of the patient in at least one state according to the real-time pulse signal to obtain the pulse waveform curve of the patient includes:

[0077] S501, performing denoising preprocessing on the received real-time pulse signal using wavelet transform, and drawing a pulse waveform curve y=f(x) in chronological order based on the preprocessed real-time pulse signal, wherein the horizontal axis of the pulse waveform curve represents the time x, and the vertical axis represents the pulse signal value y corresponding to each time;

[0078] S502, analyzing the pulse waveform curve of the patient, and determining that the pulse waveform curve includes a pulse waveform of the patient in at least one state, wherein the patient's state types include a calm state and a motion state;

[0079] S503, finding the variance of the pulse waveform curve at each moment, and setting the pulse waveform variance threshold value U;

[0080] S504, taking the time point that satisfies the conditions that the variance is greater than the pulse waveform variance threshold U, the pulse signal value is less than or equal to the pulse waveform peak value, and the first-order derivative of the pulse waveform is greater than 0 as the starting point x1 of the motion state pulse waveform;

[0081] S505, taking the time point that satisfies the conditions that the variance is greater than the pulse waveform variance threshold U, the pulse signal value is less than or equal to the pulse waveform peak value, and the first-order derivative of the pulse waveform is less than 0 as the end point x2 of the motion state pulse waveform;

[0082] S506: The pulse waveform curve region between the time points x1 and x2 is regarded as the pulse waveform region in the motion state, and the other regions are the calm state regions, thereby obtaining the pulse waveform curve of the patient after analysis.

[0083] The pulse signal is denoised using wavelet transform. The denoising process of the pulse signal can be divided into three steps: wavelet decomposition, selecting a wavelet and determining a wavelet decomposition level M, and then performing M-layer wavelet decomposition calculations on the signal; threshold quantization of high-frequency coefficients of wavelet decomposition, selecting a hard threshold method for quantization of each layer of high-frequency coefficients from the 1st layer to the Mth layer, so as to better retain local features such as signal edges; wavelet reconstruction of the signal, performing wavelet reconstruction of the signal based on the low-frequency coefficients of the Mth layer of wavelet decomposition and the high-frequency coefficients of the 1st layer to the Mth layer after quantization, and finally obtaining the pulse signal after noise removal, and drawing the pulse waveform curve y=f(x) in chronological order based on the preprocessed pulse signal.

[0084] According to the variance calculation formula of continuous signal: In the formula, is the variance of the pulse signal, y is the pulse signal value corresponding to the pulse waveform curve, μ is the mean of the pulse signal, E{} represents the expected value in the calculation brackets, calculates the variance corresponding to each moment of the pulse waveform curve, reflects the degree of waveform change at that moment, and sets the pulse waveform variance threshold U.

[0085] When the patient's state changes from a calm state to a moving state, the pulse signal will show a sudden rising trend, so the variance of the pulse waveform is greater than the variance threshold U. When it is in the rising state, the first-order derivative of the pulse waveform is greater than 0, and the pulse signal value is less than or equal to the peak value of the pulse waveform. Therefore, the moment when the patient's state changes from a calm state to a moving state can be determined, and this point is taken as the starting point x1 of the moving state pulse waveform;

[0086] When the patient's state recovers from the stimulated state to the calm state, the pulse signal will show a rapid downward trend, so the variance of the pulse waveform is greater than the variance threshold U. When in a declining state, the first-order derivative of the pulse waveform is less than 0, and the pulse signal value is less than or equal to the peak value of the pulse waveform. Therefore, the moment when the patient's state changes from the motion state to the calm state can be determined. This point is taken as the end point x2 of the motion state pulse waveform, and the pulse waveform curve area between the time points x1 and x2 is taken as the motion state pulse waveform area, and the other areas are the calm state areas. In this way, the pulse waveform curves of the patient in different states are analyzed, and combined with the analysis of the pulse waveform curves of the patient in different states, the analysis results are more accurate.

[0087] In this embodiment, the patient's pulse waveform curve is input into the prediction model, the pulse abnormality value is output, the pulse abnormality threshold is set, and whether the patient's pulse waveform curve is abnormal is determined to remind the patient, including:

[0088] S601: Input the pulse waveform curve of the patient into the prediction model to calculate the abnormal pulse value of the patient in at least one state;

[0089] S602 uses a weight method to combine the abnormal pulse values ​​of the patient in different states to calculate the abnormal pulse value of the patient;

[0090] S603 sets an abnormal pulse threshold, and determines whether the patient's pulse waveform curve is abnormal according to the patient's abnormal pulse value and the abnormal pulse threshold. When the patient's pulse waveform curve is abnormal, the patient is reminded.

[0091] When the patient has multiple states, the patient's pulse waveform curve is input into the prediction model to calculate the abnormal pulse value P1 of the patient's calm state and the abnormal pulse value P2 of the patient's motion state. The weight method is used to combine the abnormal pulse values ​​of the patient's calm state and motion state to calculate the abnormal pulse value of the patient:

[0092]

[0093] Wherein, P is the abnormal pulse value of the patient, x1 is the starting point of the pulse waveform in the motion state, x2 is the end point of the pulse waveform in the motion state, and t is the pulse detection time of the patient.

[0094] By comparing the calculated abnormal pulse value of the patient with the set abnormal pulse threshold, it is determined whether the patient's pulse waveform curve is abnormal. When the patient's pulse waveform curve is abnormal, the patient is reminded so that the patient can control his or her pulse status in real time.

[0095] In this embodiment, the prediction model uses the area method to calculate the area enclosed by the pulse waveform curve and the horizontal axis of the patient in at least one state, and uses the area as the abnormal pulse value of the patient in this state.

[0096] The prediction model uses the area method to obtain the abnormal pulse value of the patient in different states. When the patient's pulse waveform curve has different states, the abnormal pulse value of the patient in the calm state is:

[0097]

[0098] The abnormal pulse values ​​of patients in motion are:

[0099]

[0100] The abnormal pulse values ​​of the patient in different states are calculated respectively, and the abnormal pulse values ​​of the patient in different states are weighted to obtain the abnormal pulse value of the patient, which combines the different states of the patient to make the calculated result more accurate.

[0101] In this embodiment, the abnormal pulse threshold includes the maximum abnormal pulse threshold and the minimum abnormal pulse threshold. When the patient's abnormal pulse value is higher than the maximum abnormal pulse threshold or lower than the minimum abnormal pulse threshold, the patient's pulse is regarded as abnormal.

[0102] Whether the patient's abnormal pulse value is too high or too low, it is in an abnormal state. The setting of the abnormal pulse threshold includes the minimum abnormal pulse threshold Z1 and the maximum abnormal pulse threshold Z2. When the abnormal pulse value P < Z1 or P > Z2, it is determined that the patient's pulse is in an abnormal state, and the patient's pulse is regarded as abnormal. When the pulse waveform curve of the patient is abnormal, the patient is reminded.

[0103] Embodiment 2

[0104] As Figure 5 , this embodiment provides a pulse condition data processing system for remote pulse taking, which is used to implement the method for processing pulse condition data for remote pulse taking described in Embodiment 1. The system includes:

[0105] A pulse condition data adjustment device, which uses a dynamic iterative sampling method to adjust the sampling frequency of the pulse detector in real time, obtains the patient's real-time pulse signal, and sends the real-time pulse signal to the remote workstation;

[0106] A remote workstation, which receives the real-time pulse signal sent by the pulse condition data adjustment device, analyzes at least one state of the patient's pulse waveform curve according to the real-time pulse signal, inputs the patient's pulse waveform curve into the prediction model, outputs the abnormal pulse value, sets the abnormal pulse threshold, and determines whether the patient's pulse waveform curve is abnormal to remind the patient.

[0107] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements fall within the scope of the present invention claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A pulse data processing method for remote pulse diagnosis, characterized in that: Used in a pulse data adjustment device, the pulse data adjustment device is used to control a pulse detector, the pulse detector is used to collect a patient's pulse signal, and the pulse data processing method comprises: The sampling frequency of the pulse detector is adjusted in real time by using a dynamic iterative sampling method, and the real-time pulse signal of the patient is obtained; The real-time pulse signal is sent to a remote workstation, so that the remote workstation analyzes the pulse waveform curve of the patient in at least one state according to the real-time pulse signal, inputs the pulse waveform curve of the patient into a prediction model, outputs an abnormal pulse value, sets an abnormal pulse threshold, determines whether the pulse waveform curve of the patient is abnormal, and reminds the patient; The method of using the dynamic iterative sampling method to adjust the sampling frequency of the pulse detector in real time and obtain the real-time pulse signal of the patient includes: Divide the patient's pulse signal according to fixed duration to obtain N time intervals; Set the sampling frequency initial value k0, sample the patient's pulse signal, obtain N pulse signals collected in N time intervals, and count the number of pulse signal peak points in each time interval, where the number of pulse signal peak points in the nth time interval is m. n , n is an integer greater than 0 and less than or equal to N; When m n Less than or equal to m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n -q n k0, when m n Greater than m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n +q n k0, where m n-1 is the number of occurrences of the pulse signal peak point in the n-1th time interval, k n is the sampling frequency of the nth time interval, q n is the degree of change in the number of occurrences of the peak point of the pulse signal in the nth time interval; Repeat this process, dynamically iterate the sampling frequency, and acquire the complete patient's real-time pulse signal; The initial value k0 of the sampling frequency is set, and the pulse signal of the patient is sampled, and N pulse signals corresponding to the N time intervals are obtained, including: Set a sliding window, the size of the sliding window is T; Sliding the sliding window on the patient's real-time pulse signal to gradually cover the entire real-time pulse signal of the patient; When m n Less than or equal to m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n -q n k0, when m n Greater than m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n +q n k0, where q n The calculation method is: Among them, m n is the number of occurrences of the pulse signal peak point in the nth time interval, m n-1 is the number of occurrences of the pulse signal peak point in the n-1th time interval.

2. A pulse data processing method for remote pulse diagnosis, characterized in that: Used for a remote workstation, the remote workstation is connected to the pulse data adjustment device for communication, the pulse data adjustment device is used to control the pulse detector, the pulse data processing method comprises: Receiving a real-time pulse signal sent by the pulse data adjustment device, wherein the real-time pulse signal is obtained by the pulse detector collecting the patient's pulse signal after the pulse data adjustment device adjusts the sampling frequency of the pulse detector in real time using a dynamic iterative sampling method; Analyzing the pulse waveform of the patient in at least one state according to the real-time pulse signal to obtain a pulse waveform curve of the patient; Inputting the patient's pulse waveform curve into the prediction model, outputting the pulse abnormality value, setting the pulse abnormality threshold, determining whether the patient's pulse waveform curve is abnormal, and reminding the patient; The process of adjusting the sampling frequency of the pulse detector in real time by using the dynamic iterative sampling method includes: Divide the patient's pulse signal according to fixed duration to obtain N time intervals; Set the sampling frequency initial value k0, sample the patient's pulse signal, obtain N pulse signals collected in N time intervals, and count the number of pulse signal peak points in each time interval, where the number of pulse signal peak points in the nth time interval is m. n , n is an integer greater than 0 and less than or equal to N; When m n Less than or equal to m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n -q n k0, when m n Greater than m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n +q n k0, where m n-1 is the number of occurrences of the pulse signal peak point in the n-1th time interval, k n is the sampling frequency of the nth time interval, q n is the degree of change in the number of occurrences of the peak point of the pulse signal in the nth time interval; Repeat this process, dynamically iterate the sampling frequency, and acquire the complete patient's real-time pulse signal; The initial value k0 of the sampling frequency is set, and the pulse signal of the patient is sampled, and N pulse signals corresponding to the N time intervals are obtained, including: Set a sliding window, the size of the sliding window is T; Sliding the sliding window on the patient's real-time pulse signal to gradually cover the entire real-time pulse signal of the patient; When m n Less than or equal to m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n -q n k0, when m n Greater than m n-1 When the sampling frequency of the n+1th time interval is k n+1 =k n +q n k0, where q n The calculation method is: Among them, m n is the number of occurrences of the pulse signal peak point in the nth time interval, m n-1 is the number of occurrences of the pulse signal peak point in the n-1th time interval.

3. A pulse data processing method for remote pulse diagnosis according to claim 2, characterized in that: The step of analyzing the pulse waveform of the patient in at least one state according to the real-time pulse signal to obtain the pulse waveform curve of the patient includes: The received real-time pulse signal is pre-processed by using wavelet transform to remove noise, and a pulse waveform curve y=f(x) is drawn in chronological order based on the pre-processed real-time pulse signal, wherein the horizontal axis of the pulse waveform curve represents the time x, and the vertical axis represents the pulse signal value y corresponding to each time; Analyzing the pulse waveform curve of the patient to determine that the pulse waveform curve includes a pulse waveform of the patient in at least one state, wherein the patient's state types include a calm state and a motion state; Calculate the variance of the pulse waveform curve at each moment and set the pulse waveform variance threshold U; The time point that satisfies the conditions that the variance is greater than the pulse waveform variance threshold U, the pulse signal value is less than or equal to the pulse waveform peak value, and the first-order derivative of the pulse waveform is greater than 0 is taken as the starting point x1 of the motion state pulse waveform; The time point that satisfies the conditions that the variance is greater than the pulse waveform variance threshold U, the pulse signal value is less than or equal to the pulse waveform peak value, and the first-order derivative of the pulse waveform is less than 0 is taken as the end point x2 of the motion state pulse waveform; The pulse waveform curve area between the time points x1 and x2 is taken as the pulse waveform area in the motion state, and the other areas are the calm state areas, and the pulse waveform curve of the patient after analysis is obtained.

4. A pulse data processing method for remote pulse diagnosis according to claim 3, characterized in that: The step of inputting the patient's pulse waveform curve into the prediction model, outputting the pulse abnormality value, setting the pulse abnormality threshold, and determining whether the patient's pulse waveform curve is abnormal so as to remind the patient includes: Inputting the patient's pulse waveform curve into the prediction model to calculate the abnormal pulse value of the patient in at least one state; The patient's pulse abnormality value is calculated by combining the patient's pulse abnormality values ​​in different states using the weight method; An abnormal pulse threshold is set, and whether the patient's pulse waveform curve is abnormal is determined based on the patient's abnormal pulse value and the abnormal pulse threshold. When the patient's pulse waveform curve is abnormal, the patient is reminded.

5. A pulse condition data processing method for remote pulse diagnosis according to claim 4, characterized in that: The prediction model uses the area method to calculate the area enclosed by the pulse waveform curve and the horizontal axis in at least one state of the patient, and uses the area as the abnormal pulse value of the patient in this state.

6. A pulse condition data processing method for remote pulse diagnosis according to claim 5, characterized in that: The abnormal pulse threshold includes a maximum abnormal pulse threshold and a minimum abnormal pulse threshold. When the abnormal pulse value of the patient is higher than the maximum abnormal pulse threshold or lower than the minimum abnormal pulse threshold, the patient's pulse is regarded as abnormal.

7. A pulse data processing system for remote pulse diagnosis, used to implement a pulse data processing method for remote pulse diagnosis as claimed in any one of claims 1 to 6, characterized in that: The system comprises: The pulse data adjustment device uses a dynamic iterative sampling method to adjust the sampling frequency of the pulse detector in real time, obtains the patient's real-time pulse signal, and sends the real-time pulse signal to a remote workstation; The remote workstation receives the real-time pulse signal sent by the pulse data adjustment device, analyzes the pulse waveform curve of the patient in at least one state according to the real-time pulse signal, inputs the pulse waveform curve of the patient into the prediction model, outputs the abnormal pulse value, sets the abnormal pulse threshold, determines whether the pulse waveform curve of the patient is abnormal, and reminds the patient.

Images