Electrocardiosignal interference recognition method and system

Abstract

This application discloses a method and system for judging ECG signal interference. By splitting the user's ECG signal into several identification signal segments, time-domain analysis of the ECG signal is achieved by calculating the RMSSD difference between adjacent normal cardiac cycles of the identification signal segments. Furthermore, nonlinear dynamic analysis of the ECG signal is achieved by performing EMD decomposition on the RMSSD difference between adjacent normal cardiac cycles. This method combines time-domain analysis and nonlinear dynamic analysis to identify interference in the ECG signal, solving the problem that existing interference detection methods cannot extract the dynamic characteristics of the ECG signal, resulting in poor accuracy in identifying ECG signal interference, and thus improving the accuracy of ECG signal interference identification.

Description

Technical Field

[0001] This invention relates to signal recognition, and more particularly to a method and system for judging electrocardiogram signal interference. Background Technology

[0002] Electrocardiogram (ECG) signals are an important tool for doctors to diagnose heart diseases, making accurate diagnosis and timely detection of arrhythmias crucial. However, ECG signals are easily interfered with by external factors or other bioelectrical signals. Common ECG interferences include electromyography (EMG) interference, electrode interference, and baseline drift.

[0003] Common interference detection methods are usually based on the time and frequency domains. However, relying solely on time and frequency domain signals to identify heart rate bioelectrophysiological signals often suffers from poor accuracy and real-time performance, making it difficult to accurately distinguish whether interference exists through bioelectrophysiological signals. Therefore, how to accurately determine whether interference exists in ECG signals and monitor interference in real time has become an urgent problem to be solved. Summary of the Invention

[0004] Based on this, it is necessary to propose a method and system for judging electrocardiogram interference that can improve the real-time performance and accuracy of bioelectrophysiological signal interference identification, in order to address the above problems.

[0005] This application provides a method for judging electrocardiogram signal interference, including:

[0006] The user's ECG signal is acquired and then split into several identification signal segments.

[0007] The RMSSD of adjacent normal cardiac cycles of the identified signal segment is calculated, and the obtained RMSSD is decomposed into EMD to generate the identification parameter S.

[0008] Determine whether the identification parameter S is within the set range; if so, determine that the ECG signal is subject to interference.

[0009] Furthermore, the step of acquiring the user's ECG signal and splitting the ECG signal to generate several identification signal segments specifically includes:

[0010] The ECG signal is filtered to generate a filtered electrocardiogram signal;

[0011] A signal segment v is set within the filtered ECG signal, with an initial value of 2 for v. A signal segment 1 is already set at the starting point of the filtered ECG signal, and the step size between each signal segment is a user-preset step size, and the length of each signal segment is a user-preset length.

[0012] Determine whether the v-th signal segment has reached the end point of the filtered ECG signal;

[0013] If so, all the extracted signal segments will be recorded as the identified signal segments;

[0014] If not, then after using the v-th signal segment as the identification signal segment, set v = v + 1, and return to the step of setting the v-th signal segment in the filtered ECG signal.

[0015] Furthermore, the step of calculating the RMSSD of the identified signal segment and performing EMD decomposition on the obtained RMSSD to generate the identification parameter S specifically includes:

[0016] Extract the peak of the waveform within the identified signal segment and denot it as peak R;

[0017] Two adjacent wave peaks R are denoted as the same wave peak period, and are referred to as wave peak period RR;

[0018] The number of peak periods RR within the identified signal segment is obtained and denoted as the number N;

[0019] Let RR be the period of two adjacent wave crests. i and RR i-1 and the RR i The RR i-1 The N is input into the set RMSSD calculation formula for calculation, thereby calculating the RMSSD corresponding to the identified signal segment;

[0020] The RMSSD is decomposed into EMD to generate identification parameters S.

[0021] Furthermore, the step of performing EMD decomposition on the RMSSD to generate identification parameters S specifically includes:

[0022] The RMSSD is decomposed into EMD components, which are then decomposed into a first IMF component, a second IMF component, and a third IMF component.

[0023] The sample entropy of the first IMF component, the second IMF component, and the third IMF component are calculated respectively and denoted as the first sample entropy SE1, the second sample entropy SE2, and the third sample entropy SE3.

[0024] The sample complexity of the first IMF component, the second IMF component, and the third IMF component are calculated respectively and denoted as the first complexity L1, the second complexity L2, and the third complexity L3.

[0025] The recognition parameter S is calculated based on the first sample entropy SE1, the second sample entropy SE2, the third sample entropy SE3, the first complexity L1, the second complexity L2, and the third complexity L3.

[0026] Furthermore, the step of performing EMD decomposition on the RMSSD to decompose it into a first IMF component, a second IMF component, and a third IMF component specifically includes:

[0027] Extract all maxima H contained in the RMSSD n and the minimum value Y n The initial value of n is 1, and n≥1;

[0028] The maximum value H is connected using a preset cubic spline curve. n Thus, in the preset coordinate system, a coordinate system is generated corresponding to the maximum value H. n The corresponding upper envelope is fmax(n);

[0029] Connect the minimum value Y using the preset cubic spline curve. n Thus, in the preset coordinate system, a minimum value Y is generated. n The corresponding upper envelope is fmin(n);

[0030] The fmax(n) and fmin(n) are input into a preset mean calculation formula to calculate and output the curve mean c(n);

[0031] Let z(n) = x(n) - c(n), where x(n) is the same as the value of RMSSD, and z(n) is the initial component that matches the first IMF component;

[0032] Determine the maximum value H n With the minimum value Y n Whether the sum is consistent with the number of zero-crossing points in the preset coordinate system and whether the sum between fmax(n) and fmin(n) is 0;

[0033] If so, then the initial component z(n) is recorded as the first IMF component;

[0034] The first IMF component is denoted as I(n), and r(n) = x(n) - I(n) is set as the first estimation parameter used to estimate the second IMF component.

[0035] Return to the extraction of all maxima H contained in the RMSSD. n and the minimum value Y n The steps are as follows: replace the RMSSD with the r(n) to complete the calculation of the second IMF component;

[0036] The second IMF component is denoted as E(n), and O(n) = x(n) - E(n) is set as a second calculation parameter used to calculate the third IMF component.

[0037] Return to the extraction of all maxima H contained in the RMSSD. n and the minimum value Y n The steps are as follows: replace the RMSSD with the O(n) to complete the calculation of the third IMF component.

[0038] Furthermore, the step of calculating the sample entropy of the first IMF component, the second IMF component, and the third IMF component respectively, and denoting them as the first sample entropy SE1, the second sample entropy SE2, and the third sample entropy SE3, specifically includes:

[0039] Obtain a preset length P, and split the first IMF component into several m-dimensional vector data H segments according to the preset length P;

[0040] Calculate the data H of two adjacent vectors i With H j Distance D between ij where i ≠ j;

[0041] The distance D is calculated based on the preset length P and the m-dimensional vector data H. ij The probability of similarity tolerance r

[0042] Calculate the vector data H corresponding to each segment. And calculate each of the above. The mean, denoted as mean B. m ;

[0043] After setting m = m + 1, return to the step of obtaining the preset length P and splitting the first IMF component into several m-dimensional vector data H segments according to the preset length P, thereby calculating another mean B. m+1 ;

[0044] According to the B m and the B m+1 The input is entered into the preset entropy calculation formula for calculation, thereby calculating the first sample entropy SE1 corresponding to the first IMF component;

[0045] Return to the step of obtaining the preset length P, and replace the first IMF component with the second IMF component to deduce the second sample entropy SE2 corresponding to the second IMF component;

[0046] Return to the step of obtaining the preset length P, and replace the second IMF component with the third IMF component to deduce the third sample entropy SE3 corresponding to the third IMF component.

[0047] Furthermore, the step of calculating the sample complexity of the first IMF component, the second IMF component, and the third IMF component respectively, and denoting them as the first complexity L1, the second complexity L2, and the third complexity L3, specifically includes:

[0048] The first IMF component input is coarse-grained using the average value method to generate a coarse-grained sequence group SQ, wherein the coarse-grained sequence group SQ includes the sequence S = (s(1), s(2), s(3), ..., s(m)) and the sequence Q = (s(m+1), s(m+2), ..., s(m+k)).

[0049] The last data point in the coarse-grained signal sequence group SQ is removed to obtain the coarse-grained signal sequence SQ. V ;

[0050] After setting Q = s(q + e + 1), it is determined whether Q can be copied from SQv, and the values ​​of q and e are preset values ​​set by the user.

[0051] If possible, then the coarse-grained signal sequence SQ V The next character of the sequence to be determined is concatenated into Q, and the step of coarsely calculating the first IMF component of the input using the average value method is returned.

[0052] If not, then obtain the number of cascades of Q and record it as the complexity c(n) of matching the first IMF component;

[0053] The complexity c(n) is normalized to calculate the first complexity L1 that matches the first IMF component.

[0054] Return to the step of using the average value method to perform coarse-grained calculation on the first IMF component and replace the first IMF component with the second IMF component, thereby deriving the second complexity L2 that matches the second IMF component;

[0055] Return to the step of using the average value method to perform coarse-grained calculation on the first IMF component and replace the first IMF component with the second IMF component, thereby deriving the second complexity L2 that matches the second IMF component;

[0056] Return to the step of using the average value method to perform coarse-grained calculation on the first IMF component and replace the first IMF component with the third IMF component, thereby deriving the third complexity L3 that matches the third IMF component.

[0057] Further, the step of calculating the recognition parameter S based on the first sample entropy SE1, the second sample entropy SE2, the third sample entropy SE3, the first complexity L1, the second complexity L2, and the third complexity L3 specifically includes:

[0058] Determine whether the RMSSD is greater than a preset empirical threshold T;

[0059] If so, the first sample entropy SE1, the second sample entropy SE2, and the third sample entropy SE3 are weighted and summed to generate a summation value SQ;

[0060] The summation value SQ, the first complexity L1, the second complexity L2, and the third complexity L3 are input into the preset parameter S calculation formula for deduction, thereby outputting the recognition parameter S;

[0061] The step of determining whether the RMSSD is greater than a preset empirical threshold T further includes:

[0062] If not, the first complexity L1, the second complexity L2 and the third complexity L3 are weighted and summed to generate a sum value LQ;

[0063] The summation value LQ, the first sample entropy SE1, and the second sample entropy SE2 are input into the preset parameter S calculation formula for deduction, thereby outputting the recognition parameter S.

[0064] Furthermore, the set range includes the electromyographic interference range, the electrode interference range, and the baseline drift range;

[0065] The step of determining whether the identification parameter S is within a set range, and if so, determining that the ECG signal is subject to interference, specifically includes:

[0066] Determine whether the identification parameter S is within the electromyographic interference range;

[0067] If so, then the RMSSD is determined to have electromyographic interference;

[0068] Determine whether the identification parameter S is within the electrode interference range;

[0069] If so, it is determined that the RMSSD has electrode interference;

[0070] Determine whether the identification parameter S is within the baseline drift range;

[0071] If so, then the RMSSD is determined to have baseline drift.

[0072] This application also provides a system for judging electrocardiogram signal interference, including:

[0073] The acquisition unit is used to acquire the user's ECG signal and split the ECG signal to generate several identification signal segments;

[0074] The calculation unit is used to calculate the RMSSD of adjacent normal cardiac cycles of the identified signal segment, and to perform EMD decomposition on the obtained RMSSD to generate the identification parameter S.

[0075] The judgment unit is used to determine whether the identification parameter S is within a set range. If so, it is determined that the ECG signal is subject to interference.

[0076] A computer device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps:

[0077] The user's ECG signal is acquired and then split into several identification signal segments.

[0078] The RMSSD of adjacent normal cardiac cycles of the identified signal segment is calculated, and the obtained RMSSD is decomposed into EMD to generate the identification parameter S.

[0079] Determine whether the identification parameter S is within the set range; if so, determine that the ECG signal is subject to interference.

[0080] A computer-readable medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps:

[0081] The user's ECG signal is acquired and then split into several identification signal segments.

[0082] The RMSSD of adjacent normal cardiac cycles of the identified signal segment is calculated, and the obtained RMSSD is decomposed into EMD to generate the identification parameter S.

[0083] Determine whether the identification parameter S is within the set range; if so, determine that the ECG signal is subject to interference.

[0084] The aforementioned ECG signal interference judgment method and system, by splitting the user's ECG signal to generate several identification signal segments, realizes time-domain analysis of the ECG signal by calculating the RMSSD difference between adjacent normal cardiac cycles of the identification signal segments, and realizes nonlinear dynamic analysis of the ECG signal by performing EMD decomposition on the RMSSD difference between adjacent normal cardiac cycles. This achieves the function of identifying ECG signal interference by combining time-domain analysis and nonlinear dynamic analysis, solving the problem that existing interference detection methods cannot extract the dynamic characteristics of ECG signals, resulting in poor accuracy in identifying ECG signal interference, and thus improving the accuracy of ECG signal interference identification. Attached Figure Description

[0085] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0086] in:

[0087] Figure 1 This is a flowchart illustrating a method for determining ECG signal interference in one embodiment;

[0088] Figure 2 This is a schematic diagram of the structure of an electrocardiogram signal interference detection system in one embodiment;

[0089] Figure 3 This is a structural block diagram of a computer device in one embodiment. Detailed Implementation

[0090] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0091] refer to Figure 1 This application provides a method for judging electrocardiogram signal interference, including:

[0092] S1. Acquire the user's ECG signal and split the ECG signal to generate several identification signal segments;

[0093] As described in step S1 above, the background system acquires the user's ECG signal. It can be understood that the ECG signal is the commonly used electrocardiogram signal, which is acquired by an external electrocardiogram acquisition device and sent to the background system. Then, the background system splits the ECG signal to generate several identification signal segments. In this embodiment, the background system splits the ECG signal with a window length of 750 points and a sliding step size of 100 points to generate several identification signal segments. This application does not limit the specific number of segments to be split into for the identification signal segments.

[0094] S2. Calculate the RMSSD of adjacent normal cardiac cycles of the identified signal segment, and perform EMD decomposition on the obtained RMSSD to generate the identification parameter S.

[0095] As described in step S2 above, after obtaining the identified signal segment, the background system uses a preset mathematical morphology method to calculate the RMSSD of adjacent normal cardiac cycles corresponding to the identified signal segment. It can be understood that the background system calculates the identified signal segments obtained from each segment breakdown to generate the RMSSD that matches each identified signal segment. Then, the background system performs EMD (empirical mode decomposition) decomposition on each RMSSD to generate the identification parameter S.

[0096] S3. Determine whether the identification parameter S is within the set range. If so, determine that the ECG signal is interfered with.

[0097] As described in step S3 above, the background system obtains a preset range and determines whether the identification parameter S is within the preset range. If so, it determines that the ECG signal has interference, thereby realizing the identification of interference in the ECG by combining time-domain analysis of the ECG (i.e., the step of calculating the RMSSD difference between adjacent normal cardiac cycles of the identified signal segment) and nonlinear dynamic analysis (i.e., the step of EMD decomposition).

[0098] This embodiment, through the above method, achieves time-domain analysis of the ECG signal by splitting the user's ECG signal into several identification signal segments and calculating the RMSSD difference between adjacent normal cardiac cycles of the identification signal segments. It also achieves nonlinear dynamic analysis of the ECG signal by performing EMD decomposition on the RMSSD difference between adjacent normal cardiac cycles. This combines time-domain analysis and nonlinear dynamic analysis to identify interference in the ECG signal, solving the problem that existing interference detection methods cannot extract the dynamic characteristics of the ECG signal, resulting in poor accuracy in identifying interference in the ECG signal, and thus improving the accuracy of interference identification in the ECG signal.

[0099] In one embodiment, step S1 specifically includes:

[0100] S11. Filter the ECG signal to generate a filtered ECG signal. Set a v-th signal segment within the filtered ECG signal, with an initial value of 2 for v. A first signal segment is already set at the starting point of the filtered ECG signal, and the step size between each signal segment is a user-preset step size, and the length of each signal segment is a user-preset length. Determine whether the v-th signal segment reaches the end point of the filtered ECG signal. If yes, record all extracted signal segments as the identified signal segments. If not, after setting the v-th signal segment as the identified signal segment, set v = v + 1, and return to the step of setting the v-th signal segment within the filtered ECG signal.

[0101] As described in the above embodiments, the background system uses a 4th-order IIR Butterworth bandpass filter to filter the ECG signal. The sampling frequency of the filter is 500Hz and the cutoff frequency is 0.5Hz-40Hz, thereby realizing the function of removing 50Hz power frequency interference from the ECG signal and finally outputting the filtered ECG signal.

[0102] Furthermore, during the process of splitting the filtered ECG signal, the backend system first sets a first signal segment with a step size of 750 points at the starting point of the filtered ECG signal. Then, the backend system generates a second signal segment (v) 100 steps after the first signal segment. It can be understood that the initial value of v is 2, meaning the backend system generates a second signal segment with a step size of 750 points 100 steps after the first signal segment. The backend system then determines whether the second signal segment has been reached. If the endpoint of the filtered ECG signal is reached, the background system records the first signal segment and the second signal segment as the identified signal segment. If not, the background system generates a third (v+1) interception point 100 step units after the second signal segment, and intercepts the third (v+1) signal segment with a step size of 750 points based on the third interception point. This process continues until the generated v signal segment reaches the endpoint of the filtered ECG signal, and then all the obtained signal segments are recorded as the identified signal segments.

[0103] Furthermore, step S2 specifically includes:

[0104] S21. Extract the peaks of the waveform within the identified signal segment and denote them as peaks R. Record two adjacent peaks R as the same peak period and denote them as peak periods RR. Obtain the number of peak periods RR within the identified signal segment and denote it as the number N. Record two adjacent peak periods RR as RR... i and RR i-1 and the RR i The RR i-1 The N is input into the set RMSSD calculation formula for calculation, thereby calculating the RMSSD corresponding to the identification signal segment, and the RMSSD is decomposed into EMD to generate the identification parameter S.

[0105] As described in the above embodiment, the background system extracts the peaks of each waveform within the identified signal segment and records them as peaks R. Then, the background system records two adjacent peaks R as the same peak period and records it as peak period RR. For example, if there are peaks R1, R2, R3, and R4, then two adjacent peaks RR are (R1, R2), (R2, R3), and (R3, R4). The background system then obtains the number of peak periods RR within the identified signal segment and records it as the number N. Finally, the background system records two adjacent peak periods RR as RR. i and RR i-1 and the RR i The RR i-1The N is input into the set RMSSD calculation formula for calculation. In this embodiment, the RMSSD calculation formula is specifically expressed as follows:

[0106]

[0107] The background system then uses the aforementioned mathematical morphology algorithm to calculate the specific value of the RMSSD that matches the identified signal segment.

[0108] In one embodiment, the step of performing EMD decomposition on the RMSSD to generate identification parameters S specifically includes:

[0109] S22. The RMSSD is decomposed into EMD components, namely a first IMF component, a second IMF component, and a third IMF component. The sample entropy of the first IMF component, the second IMF component, and the third IMF component is calculated and denoted as the first sample entropy SE1, the second sample entropy SE2, and the third sample entropy SE3, respectively. The sample complexity of the first IMF component, the second IMF component, and the third IMF component is calculated and denoted as the first complexity L1, the second complexity L2, and the third complexity L3, respectively. The recognition parameter S is calculated based on the first sample entropy SE1, the second sample entropy SE2, the third sample entropy SE3, the first complexity L1, the second complexity L2, and the third complexity L3.

[0110] As described in the above embodiment, the backend system performs EMD decomposition on the RMSSD, thereby decomposing it into the first IMF component, the second IMF component, and the third IMF component. It can be understood that the IMF is the normal output result after EMD decomposition, that is, the Intrinsic Mode Functions (IMF). After calculating each IMF component, the backend system calculates the sample entropy and sample complexity of the first IMF component, the second IMF component, and the third IMF component, and denotes them as the first complexity L1, the second complexity L2, the third complexity L3, the first complexity L1, the second complexity L2, and the third complexity L3, respectively. Then, the backend system calculates the recognition parameter S based on the first complexity L1, the second complexity L2, the third complexity L3, the first complexity L1, the second complexity L2, and the third complexity L3.

[0111] In one embodiment, the step of performing EMD decomposition on the RMSSD to decompose it into a first IMF component, a second IMF component, and a third IMF component specifically includes:

[0112] S23. Extract all maxima H contained in the RMSSD. n and the minimum value Y n The initial value of n is 1, and n≥1;

[0113] The maximum value H is connected using a preset cubic spline curve. n Thus, in the preset coordinate system, a coordinate system is generated corresponding to the maximum value H. n The corresponding upper envelope fmax(n) is connected to the minimum value Y using the preset cubic spline curve. n Thus, in the preset coordinate system, a minimum value Y is generated. n The corresponding upper envelope fmin(n) is calculated by inputting fmax(n) and fmin(n) into a preset mean calculation formula, thereby outputting the curve mean c(n). Let z(n) = x(n) - c(n), where x(n) is consistent with the value of RMSSD, and z(n) is the initial component matching the first IMF component. The maximum value H is then determined. n With the minimum value Y n If the sum of the values ​​is consistent with the number of zero-crossing points in the preset coordinate system, and if the sum between fmax(n) and fmin(n) is 0, then the initial component z(n) is denoted as the first IMF component, the first IMF component is denoted as I(n), and r(n) = x(n) - I(n) is set, where r(n) is the first estimation parameter used to estimate the second IMF component. Then, all maxima H contained in the RMSSD are extracted. n and the minimum value Y n The steps are as follows: The RMSSD is replaced with r(n) to complete the estimation of the second IMF component. The second IMF component is denoted as E(n), and O(n) = x(n) - E(n) is set as the second estimation parameter used to estimate the third IMF component. The process returns the extraction of all maxima H contained in the RMSSD. n and the minimum value Y n The steps are as follows: replace the RMSSD with the O(n) to complete the calculation of the third IMF component.

[0114] As described in the above embodiments, the background system extracts all maximum values ​​H contained in the RMSSD. n and the minimum value Y n The initial value of n is 1, and n≥1. It is understandable that the maximum value H... n and the minimum value Y n There was more than one; subsequently, the backend system used preset cubic spline curves to connect the maxima H respectively. nand the minimum value Y n Thus, in the preset coordinate system, a coordinate system is generated corresponding to the maximum value H. n The corresponding upper envelope fmax(n) and the minimum value Y n The corresponding lower envelope fmin(n) is calculated by inputting fmax(n) and fmin(n) into a preset mean calculation formula, thereby outputting the curve mean c(n). In this embodiment, the preset mean calculation formula is:

[0115]

[0116] The backend system then records the RMSSD as the signal value x(n) and determines whether the signal value x(n) and the curve mean c(n) satisfy z(n) = x(n) - c(n). It can be understood that z(n) is the initial component. If satisfied, the backend system continues to determine the maximum value H. n With the minimum value Y n Whether the sum is consistent with the number of zero-crossing points in the preset coordinate system, it can be understood that the RMSSD includes more than just the maximum value H. n and the minimum value Y n If other values ​​also fall within the preset coordinate system, then the background system simultaneously determines the maximum value H. n With the minimum value Y n If the sum of the values ​​is equal to or differs by 1 from the number of zero-crossing points in the preset coordinate system, and if the sum of fmax(n) and fmin(n) is 0, then the background system will record the initial component z(n) as the first IMF component.

[0117] The backend system then denotes the first IMF component as I(n) and sets r(n) = x(n) - I(n), where r(n) is the first estimation parameter used to estimate the second IMF component. The backend system then returns the extracted maxima H contained in the RMSSD. n and the minimum value Y n The step simultaneously replaces the RMSSD with the r(n), thereby completing the calculation of the second IMF component;

[0118] Similarly, the backend system denotes the second IMF component as E(n) and sets O(n) = x(n) - E(n), where O(n) is the second estimation parameter used to estimate the third IMF component. Then, the backend system returns the extraction of all maxima H contained in the RMSSD. n and the minimum value Y nThe steps are as follows: replace the RMSSD with the O(n) to complete the calculation of the third IMF component.

[0119] In one embodiment, the step of calculating the sample entropy of the first IMF component, the second IMF component, and the third IMF component respectively, and denoting them as the first sample entropy SE1, the second sample entropy SE2, and the third sample entropy SE3, specifically includes:

[0120] S24. Obtain a preset length P, and according to the preset length P, split the first IMF component into several m-dimensional vector data H segments, and calculate the values ​​of two adjacent vector data segments H. i With H j Distance D between ij Where i ≠ j, the distance D is calculated based on the preset length P and the m-dimensional vector data H. ij The probability of similarity tolerance r Calculate the vector data H corresponding to each segment. And calculate each of the above. The mean of () is denoted as mean B. m After setting m = m + 1, return to the step of obtaining the preset length P and splitting the first IMF component into several m-dimensional vector data H segments according to the preset length P, thereby calculating another mean B. m+1 According to the aforementioned B m and the B m+1 The input is entered into a preset entropy calculation formula to calculate the first sample entropy SE1 corresponding to the first IMF component. Then, the process returns to the step of obtaining the preset length P, and the first IMF component is replaced with the second IMF component to calculate the second sample entropy SE2 corresponding to the second IMF component. Then, the process returns to the step of obtaining the preset length P, and the second IMF component is replaced with the third IMF component to calculate the third sample entropy SE3 corresponding to the third IMF component.

[0121] As described in the above embodiments, the background system obtains a preset length P and splits the first IMF component into several m-dimensional vector data H segments according to the preset length P, specifically as follows:

[0122] H(i)={h(i),h(i+1),h(i+2)....h(i+m-1)}

[0123] The backend system then calculates the data H of two adjacent vectors. i With H j Distance D between ij Where i≠j, specifically:

[0124] D ij =max(|x(i+k)-x(j+k)|)

[0125] Here, k is the index for traversing the i-th or j-th data segment, which is a user-preset value, and the initial values ​​of i and j are 1;

[0126] The backend system then calculates the distance D based on the preset length P and the m-dimensional vector data H. ij The probability of similarity tolerance r Specifically, this manifests as follows:

[0127]

[0128] When the background system calculates the corresponding vector data H based on the above steps, Then, continue to calculate each of the aforementioned... The mean of, and denoted as mean B. m Specifically, it manifests as follows:

[0129]

[0130] And when the background system completes the processing of the mean B m After the calculation, let m = m + 1, then return to the step of obtaining the preset length P and splitting the first IMF component into several m-dimensional vector data H segments according to the preset length P, thereby deriving another mean B. m+1 And according to the B m and the B m+1 The input is fed into a preset entropy calculation formula for calculation, thereby outputting the first sample entropy SE1 that matches the first IMF component. The preset entropy calculation formula is specifically expressed as follows:

[0131]

[0132] It is understood that in this embodiment, P, m, and r in the above formulas need to be determined based on the actual sampling frequency and the update time of the BIS index. In this embodiment, P = 1280, m = 2, and r = 0.15 * the standard deviation of the vector data H in this segment.

[0133] When the background system calculates the first sample entropy SE1 matching the first IMF component, the background system returns to the step of obtaining the preset length P in this embodiment, and replaces the first IMF component with the second IMF component, thereby calculating the second sample entropy SE2 corresponding to the second IMF component. Then, the background system returns to the step of obtaining the preset length P in this embodiment again, and replaces the second IMF component with the third IMF component, thereby calculating the third sample entropy SE3 corresponding to the third IMF component, thus completing the calculation of the first sample entropy SE1, the second sample entropy SE2, and the third sample entropy SE3.

[0134] In one embodiment, the step of calculating the sample complexity of the first IMF component, the second IMF component, and the third IMF component respectively, and denoting them as the first complexity L1, the second complexity L2, and the third complexity L3, specifically includes:

[0135] S25. The first IMF component input is coarse-grained using the averaging method to generate a coarse-grained sequence group SQ, wherein the coarse-grained sequence group SQ includes the sequence S = (s(1), s(2), s(3), ..., s(m)) and the sequence Q = (s(m+1), s(m+2), ..., s(m+k)); the last data in the coarse-grained signal sequence group SQ is removed to obtain the coarse-grained signal sequence SQ. V After setting Q = s(q + e + 1), it is determined whether Q can be copied from SQv. The values ​​of q and e are preset values ​​set by the user. If so, the coarse-grained signal sequence SQv is... V The next character of the sequence to be determined is concatenated into Q, and the process returns to the step of coarsely calculating the first IMF component using the averaging method. If this is not possible, the number of concatenations of Q is obtained and recorded as the complexity c(n) matching the first IMF component. The complexity c(n) is normalized to deduce the first complexity L1 matching the first IMF component. The process returns to the step of coarsely calculating the first IMF component using the averaging method and replaces the first IMF component with the second IMF component to deduce the second complexity L2 matching the second IMF component. The process returns to the step of coarsely calculating the first IMF component using the averaging method and replaces the first IMF component with the third IMF component to deduce the third complexity L3 matching the third IMF component.

[0136] As described in the above embodiments, the background system uses the averaging method to simplify the input first IMF component into a signal sequence composed of 0s and 1s, that is, to perform coarse-grained calculation. The specific coarse-grained calculation formula is as follows:

[0137]

[0138] After the above coarse-grained calculation is completed, a coarse-grained signal sequence SQ is generated. Specifically, the coarse-grained signal sequence SQ is: S = (s(1), s(2), s(3), ..., s(m)), Q = (s(m+1), s(m+2), ..., s(m+k)). It can be understood that if k ≥ 1, the background system removes the last data point from the coarse-grained signal sequence group SQ, thus obtaining the coarse-grained signal sequence SQ. V The background system then causes the coarse-grained signal sequence SQ to... V After Q = s(q + e + 1), it is determined whether Q can be copied from SQv. It is understood that the values ​​of q and e are preset values ​​set by the user.

[0139] At this point, when Q can be copied from SQv, the background system will use the coarse-grained signal sequence SQ. V The next character of the sequence to be determined is concatenated into Q, and the step of coarsely calculating the first IMF component of the input using the average value method is returned, thereby realizing the function of concatenating multiple characters in Q;

[0140] Furthermore, when Q cannot be copied from SQv, the backend system obtains the number of characters concatenated with Q in SQv and records it as the complexity c(n) matching the first IMF component. Then, the backend system normalizes the complexity c(n) to calculate the first complexity L1 matching the first IMF component. The specific normalization formula is as follows:

[0141]

[0142] Then the L1 output by the normalization formula is the first complexity L1 of the first IMF component matching.

[0143] The backend system then returns to the step of coarsely calculating the first IMF component using the average value method, and replaces the first IMF component with the second IMF component, thereby calculating the second complexity L2 that matches the second IMF component.

[0144] The background system then returns to the step of coarsely calculating the first IMF component using the average value method and replaces the first IMF component with the third IMF component, thereby deriving the third complexity L3 that matches the third IMF component. Thus, this embodiment completes the calculation steps corresponding to the first complexity L1, the second complexity L2, and the third complexity L3.

[0145] In one embodiment, the step of calculating the recognition parameter S based on the first sample entropy SE1, the second sample entropy SE2, the third sample entropy SE3, the first complexity L1, the second complexity L2, and the third complexity L3 specifically includes:

[0146] S26. Determine whether the RMSSD is greater than the preset empirical threshold T. If so, perform a weighted summation of the first sample entropy SE1, the second sample entropy SE2, and the third sample entropy sE3 to generate a summation value SQ. Input the summation value SQ, the first complexity L1, the second complexity L2, and the third complexity L3 into the preset parameter S calculation formula for calculation, and then output the recognition parameter S.

[0147] The step of determining whether the RMSSD is greater than the preset empirical threshold T further includes, if not, weighted summation of the first complexity L1, the second complexity L2 and the third complexity L3 to generate a summation value LQ, and inputting the summation value LQ, the first sample entropy SE1 and the second sample entropy SE2 into the preset parameter S calculation formula for calculation, thereby outputting the recognition parameter S.

[0148] As described in the above embodiments, the background system obtains the preset experience threshold T and determines whether the RMSSD is greater than the preset experience threshold T. If so, the first sample entropy SE1, the second sample entropy SE2, and the third sample entropy SE3 are weighted and summed to generate a summation value SQ. The specific formula for the weighted summation value SQ is as follows:

[0149]

[0150] Wherein, the a i It is a preset weighting coefficient, and the a i satisfy The background system then records the weighted summation output as the summation value SQ. The background system then inputs the summation value SQ, the first complexity L1, the second complexity L2, and the third complexity L3 into the preset parameter S calculation formula for calculation, thereby outputting the recognition parameter S.

[0151] Furthermore, when the backend system determines that the RMSSD is less than the preset empirical threshold T, the backend system performs a weighted summation of the first complexity L1, the second complexity L2, and the third complexity L3 to generate a summation value LQ. The specific formula for the weighted summation of the summation value LQ is as follows:

[0152]

[0153] Wherein, the b i It is a preset weighting coefficient, and the b i satisfy Then the background system records the weighted summation output as the summation value LQ. The background system then inputs the summation value LQ, the first sample entropy SE1, and the second sample entropy SE2 into the preset parameter S calculation formula for calculation, thereby outputting the recognition parameter S.

[0154] It is understandable that the formula for calculating the preset parameter S is specifically embodied as follows:

[0155]

[0156] Furthermore, the preset empirical threshold T can be adaptively updated based on the RMSSD after a set time period. For example, the RMSSD can be recalculated every 1 second. This achieves adaptive updating of the preset empirical threshold T. The calculation method of the data segment is specifically as follows:

[0157]

[0158] The backend system can then adaptively update the preset experience threshold T using the above formula, thereby achieving the function of adjusting the RMSSD threshold in real time.

[0159] In one embodiment, the set range includes the electromyography interference range, the electrode interference range, and the baseline drift range;

[0160] Step S3 specifically includes:

[0161] S31. Determine whether the identification parameter S is within the electromyographic interference range. If yes, determine that the RMSSD has electromyographic interference. Determine whether the identification parameter S is within the electrode interference range. If yes, determine that the RMSSD has electrode interference. Determine whether the identification parameter S is within the baseline drift range. If yes, determine that the RMSSD has baseline drift.

[0162] As described in the above embodiments, in this embodiment, the electromyographic interference range is (S>7), the electrode interference range is (7>S>4), and the baseline drift threshold is (4>S>2). The background system determines whether the identification parameter S is greater than 7. If so, it determines that the RMSSD has electromyographic interference. In addition, the background system also determines whether the identification parameter S is less than 7 and greater than 4. If so, it determines that the RMSSD has electrode interference (such as electrode detachment or tearing). Furthermore, the background system also determines whether the identification parameter S is less than 4 and greater than 2. If so, it determines that the RMSSD has baseline offset interference, thereby realizing the function of judging the interference present in the RMSSD based on the value of the identification parameter S.

[0163] In addition, before the background system determines whether the identification parameter S meets the above three ranges, the background system first determines whether the signal amplitude of the identification signal segment is greater than 0.2mV. If not, the background system determines that the identification signal segment has a pause. At this time, the user's identification signal segment has lost its judgment significance, and the background system will no longer judge the identification signal segment.

[0164] Similarly, when the background system determines that the identification parameter S is less than the set value 2, the background system determines that there is no interference in the current identification signal segment, and continues to monitor the remaining identification signal segments within the ECG signal.

[0165] refer to Figure 2 This application also provides a system for judging electrocardiogram signal interference, including:

[0166] Acquisition unit 1 is used to acquire the user's ECG signal and split the ECG signal to generate several identification signal segments;

[0167] Calculation unit 2 is used to calculate the RMSSD of adjacent normal cardiac cycles of the identified signal segment, and to perform EMD decomposition on the obtained RMSSD to generate identification parameter S;

[0168] Judgment unit 3 is used to determine whether the identification parameter S is within a set range. If so, it is determined that the ECG signal is subject to interference.

[0169] The aforementioned units are used to execute the ECG signal interference judgment system, and will not be described in detail here.

[0170] Figure 3 An internal structural diagram of a computer device in one embodiment is shown. This computer device may specifically be a server, including but not limited to high-performance computers and high-performance computer clusters. Figure 3As shown, the computer device includes a processor, memory, and network interface connected via a system bus. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and may also store a computer program. When executed by the processor, this computer program enables the processor to implement the employee status determination method. The internal memory may also store a computer program, which, when executed by the processor, enables the processor to implement the electrocardiogram signal interference determination method.

[0171] In one embodiment, the customer behavior recognition method provided by the present invention can be implemented as a computer program, which can be implemented in the form of, for example... Figure 3 The system runs on the computer device shown. The computer device's memory can store the various program templates that make up the automatic mail sorting and aggregation device. For example: acquisition unit 1, calculation unit 2, and judgment unit 3.

[0172] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, performs the following steps:

[0173] The user's ECG signal is acquired and split into several identification signal segments. The RMSSD of adjacent normal cardiac cycles of the identification signal segments is calculated, and the obtained RMSSD is decomposed into EMD to generate identification parameters S. It is determined whether the identification parameters S are within a set range. If so, it is determined that the ECG signal is interfered with.

[0174] As can be seen from the above embodiments, the greatest beneficial effect of the present invention lies in the fact that, by splitting the user's ECG signal into several identification signal segments, time-domain analysis of the ECG signal is achieved by calculating the RMSSD difference between adjacent normal cardiac cycles of the identification signal segments, and nonlinear dynamic analysis of the ECG signal is achieved by performing EMD decomposition on the RMSSD difference between adjacent normal cardiac cycles. This achieves the function of identifying interference in the ECG signal by combining time-domain analysis and nonlinear dynamic analysis, solving the problem that existing interference detection methods in the prior art cannot extract the dynamic characteristics of the ECG signal, resulting in poor accuracy in identifying interference in the ECG signal, and thus improving the accuracy of identifying interference in the ECG signal.

[0175] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), and double data rate RAM.

[0176] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0177] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for judging interference in electrocardiogram signals, characterized in that, include: The user's ECG signal is acquired and then split into several identification signal segments. The RMSSD of adjacent normal cardiac cycles of the identified signal segment is calculated, and the obtained RMSSD is decomposed into EMD to generate the identification parameter S. Determine whether the identification parameter S is within the set range; if so, determine that the ECG signal is subject to interference. The step of performing EMD decomposition on the RMSSD to generate identification parameters S specifically includes: The RMSSD is decomposed into EMD components, which are then decomposed into a first IMF component, a second IMF component, and a third IMF component. The sample entropy of the first IMF component, the second IMF component, and the third IMF component are calculated respectively, and denoted as the first sample entropy. Second sample entropy and the third sample entropy ; The sample complexity of the first IMF component, the second IMF component, and the third IMF component are calculated separately and denoted as the first complexity. Second complexity and the third complexity ; According to the first sample entropy The second sample entropy The third sample entropy First complexity Second complexity and the third complexity Calculate the identification parameter S; Wherein, according to the first sample entropy The second sample entropy The third sample entropy First complexity Second complexity and the third complexity The steps for calculating the identification parameter S specifically include: Determine whether the RMSSD is greater than a preset empirical threshold T; If so, then the entropy of the first sample... The second sample entropy and the third sample entropy Perform a weighted summation to generate the sum value. ; The summation value First complexity Second complexity and the third complexity The input is used to calculate the preset parameter S in the formula, thereby outputting the identification parameter S; If not, then reduce the first complexity. The second complexity and the third complexity Perform a weighted summation to generate the sum value. ; The summation value The first sample entropy and the second sample entropy The input is used to calculate the preset parameter S in the formula, and then the identification parameter S is output.

2. The method for judging ECG signal interference as described in claim 1, characterized in that, The step of acquiring the user's ECG signal and splitting the ECG signal to generate several identification signal segments specifically includes: The ECG signal is filtered to generate a filtered electrocardiogram signal; The first step is set within the filtered electrocardiogram signal. signal segment, The initial value is 2, wherein a first signal segment has been set at the starting point of the filtered ECG signal, and the step size between each signal segment is a user-preset step size, and the length of each signal segment is a user-set preset length. Determine the first Whether the signal segment reaches the end point of the filtered ECG signal; If so, all the extracted signal segments will be recorded as the identified signal segments; If not, then after taking the v-th signal segment as the identification signal segment, set v=v+1, and return to the execution of setting the v-th signal segment within the filtered ECG signal. The steps for signal segments.

3. The method for judging ECG signal interference as described in claim 1, characterized in that, The step of calculating the RMSSD of the identified signal segment and performing EMD decomposition on the obtained RMSSD to generate the identification parameter S specifically includes: Extract the peak of the waveform within the identified signal segment and denot it as peak R; Two adjacent wave peaks R are denoted as the same wave peak period, and are referred to as wave peak period RR; The number of peak periods RR within the identified signal segment is obtained and denoted as the number N; Let the two adjacent peak periods RR be denoted as as well as and the The above The N is input into the set RMSSD calculation formula for calculation, thereby calculating the RMSSD corresponding to the identified signal segment; The RMSSD is decomposed into EMD to generate identification parameters S.

4. The method for judging ECG signal interference as described in claim 3, characterized in that, The step of performing EMD decomposition on the RMSSD to decompose it into a first IMF component, a second IMF component, and a third IMF component specifically includes: Extract all maxima contained in the RMSSD and minimum value The initial value of n is 1, and n≥1; The maximum value is connected using a preset cubic spline curve. Thus, a maximum value is generated in the preset coordinate system. The corresponding upper envelope is fmax(n); Connect the minimum value using the preset cubic spline curve. Thus, a minimum value is generated in the preset coordinate system. The corresponding upper envelope is fmin(n); The fmax(n) and fmin(n) are input into a preset mean calculation formula to calculate and output the curve mean c(n). Let z(n) = x(n) – c(n), where x(n) is consistent with the value of RMSSD, and z(n) is the initial component that matches the first IMF component; Determine the maximum value With the minimum value Whether the sum is consistent with the number of zero-crossing points in the preset coordinate system and whether the sum between fmax(n) and fmin(n) is 0; If so, then the initial component z(n) is recorded as the first IMF component; The first IMF component is denoted as I(n), and r(n) = x(n) - I(n) is set as the first estimation parameter used to estimate the second IMF component. Return to the extraction of all maxima contained in the RMSSD. and minimum value The steps are as follows: replace the RMSSD with the r(n) to complete the calculation of the second IMF component; The second IMF component is denoted as E(n), and O(n) = x(n) - E(n) is set as a second calculation parameter used to calculate the third IMF component. Return to the extraction of all maxima contained in the RMSSD. and minimum value The steps are as follows: replace the RMSSD with the O(n) to complete the calculation of the third IMF component.

5. The method for judging ECG signal interference as described in claim 3, characterized in that, The sample entropy of the first IMF component, the second IMF component, and the third IMF component is calculated respectively and denoted as the first sample entropy. Second sample entropy and the third sample entropy The steps specifically include: Obtain a preset length P, and split the first IMF component into several m-dimensional vector data H segments according to the preset length P; Calculate two adjacent vector data and Distance between where i ≠ j; The distance is calculated based on the preset length P and the m-dimensional vector data H. The probability of similarity tolerance r ; Calculate the vector data H corresponding to each segment. and calculate each of the above. The mean, denoted as mean. ; After setting m=m+1, return to the step of obtaining the preset length P and splitting the first IMF component into several m-dimensional vector data H segments according to the preset length P, thereby calculating another mean. ; According to the above and the The input is fed into a preset entropy calculation formula for calculation, thereby calculating the first sample entropy corresponding to the first IMF component. ; Returning to the step of obtaining the preset length P, the first IMF component is replaced with the second IMF component, thereby calculating the second sample entropy corresponding to the second IMF component. ; Returning to the step of obtaining the preset length P, the second IMF component is replaced with the third IMF component, thereby calculating the third sample entropy corresponding to the third IMF component. .

6. The method for judging ECG signal interference as described in claim 3, characterized in that, The sample complexity of calculating the first IMF component, the second IMF component, and the third IMF component respectively is denoted as the first complexity. Second complexity and the third complexity The steps specifically include: The first IMF component input is coarsely calculated using the averaging method, thereby generating a coarse-grained sequence set. The coarse-grained sequence group This includes the sequence S = (s(1), s(2), s(3), ..., s(m)) and the sequence Q = (s(m+1), s(m+2), ..., s(m+k)). The last data in the coarse-grained sequence group SQ is removed to obtain the coarse-grained signal sequence. ; After setting Q = s(q+e+1), it is determined whether Q can be copied from SQv, and the values ​​of q and e are preset values ​​set by the user. If possible, then the coarse-grained signal sequence The next character of the sequence to be determined is concatenated into Q, and the step of coarsely calculating the first IMF component of the input using the average value method is returned. If not, then obtain the number of cascades of Q and record it as the complexity c(n) of matching the first IMF component; The complexity c(n) is normalized to deduce the first complexity that matches the first IMF component. ; Returning to the step of coarsely calculating the first IMF component using the averaging method, and replacing the first IMF component with the second IMF component, the second complexity matching the second IMF component is calculated. ; Returning to the step of coarsely calculating the first IMF component using the averaging method, and replacing the first IMF component with the second IMF component, the second complexity matching the second IMF component is calculated. ; Returning to the step of coarsely calculating the first IMF component using the average method, and replacing the first IMF component with the third IMF component, the third complexity matching the third IMF component is calculated. .

7. The method for judging ECG signal interference as described in claim 1, characterized in that, The set range includes the electromyography interference range, the electrode interference range, and the baseline drift range; The step of determining whether the identification parameter S is within a set range, and if so, determining that the ECG signal is subject to interference, specifically includes: Determine whether the identification parameter S is within the electromyographic interference range; If so, then the RMSSD is determined to have electromyographic interference; Determine whether the identification parameter S is within the electrode interference range; If so, it is determined that the RMSSD has electrode interference; Determine whether the identification parameter S is within the baseline drift range; If so, then the RMSSD is determined to have baseline drift.

8. A system for judging electrocardiogram signal interference, characterized in that, include: The acquisition unit is used to acquire the user's ECG signal and split the ECG signal to generate several identification signal segments; The calculation unit is used to calculate the RMSSD of adjacent normal cardiac cycles of the identified signal segment, and to perform EMD decomposition on the obtained RMSSD to generate the identification parameter S. The judgment unit is used to determine whether the identification parameter S is within a set range; if so, it is determined that the ECG signal is subject to interference. The computing unit is further configured to perform EMD decomposition on the RMSSD, thereby decomposing it into a first IMF component, a second IMF component, and a third IMF component; calculate the sample entropy of the first IMF component, the second IMF component, and the third IMF component respectively, and record it as the first sample entropy. Second sample entropy and the third sample entropy Calculate the sample complexity of the first IMF component, the second IMF component, and the third IMF component respectively, and denote it as the first complexity. Second complexity and the third complexity According to the entropy of the first sample The second sample entropy The third sample entropy First complexity Second complexity and the third complexity Calculate the identification parameter S; The calculation unit is further configured to determine whether the RMSSD is greater than a preset empirical threshold T; if so, the first sample entropy is... The second sample entropy and the third sample entropy Perform a weighted summation to generate the sum value. The summation value First complexity Second complexity and the third complexity The input is used to calculate the preset parameter S in the formula, thereby outputting the identification parameter S; If not, then reduce the first complexity. The second complexity and the third complexity Perform a weighted summation to generate the sum value. The summation value The first sample entropy and the second sample entropy The input is used to calculate the preset parameter S in the formula, and then the identification parameter S is output.

Images