Identification method and device for ventricular flapping and ventricular fibrillation, electronic equipment and storage medium

Abstract

The invention relates to a ventricular flapping and ventricular fibrillation recognition method and device, electronic equipment and a storage medium. The identification method comprises the following steps: collecting a biological signal and storing the biological signal; performing differential sorting on the biological signals; constructing a quantile map for the signals subjected to differential sorting according to a set step length; selecting a predetermined number of feature points in the quantile to calculate a target curvature of a connection line of the feature points; and whether ventricular flapping and ventricular fibrillation exist or not is recognized according to comparison of the target curvature and a set threshold value. According to the recognition method, the signal fidelity in the feature extraction process can be improved, misleading of local interference to recognition is reduced, and the accuracy of ventricular flapping and ventricular fibrillation recognition is improved.

Description

Technical Field

[0001] The present application relates to the field of signal control technology, and in particular to a method, device, electronic device and storage medium for identifying ventricular flutter and ventricular fibrillation. Background Art

[0002] Ventricular flutter and ventricular fibrillation are serious ectopic rhythms, in which the ventricles lose their effective overall contraction ability and are replaced by rapid and uncoordinated tremors of various parts of the myocardium. Ventricular flutter lasts for a very short time and generally turns into ventricular fibrillation within a few minutes. It is a fatal arrhythmia before death. Ventricular fibrillation refers to the sudden loss of overall coordination of the patient's myocardium, irregular contraction, and a clinically dangerous state in which blood cannot be pumped out. Since both are malignant and extremely serious fatal arrhythmias, the hemodynamic changes and clinical manifestations caused are almost the same. Both are cardiac arrests in terms of blood discharge function. After occurrence, they almost always cause Ass syndrome and sudden death in a very short time, so rapid and effective cardiopulmonary resuscitation and early electric defibrillation and cardioversion treatment are required.

[0003] The devices usually used for defibrillation and cardioversion therapy are AED (automatic external defibrillator) or WCD (wearable defibrillator). AED or WCD connects electrodes to the human body, collects electrocardiogram data, automatically analyzes the heart rhythm, and determines whether to perform electric shock defibrillation based on whether it is a malignant arrhythmia such as ventricular fibrillation or ventricular flutter.

[0004] How to provide an accurate identification method for ventricular fibrillation and ventricular flutter has become a hot topic in the field of defibrillation and cardioversion therapy. Summary of the invention

[0005] Embodiments of the present application provide a method, device, electronic device, and storage medium for controlling a target signal.

[0006] In a first aspect, an embodiment of the present application provides a method for identifying ventricular fibrillation and ventricular flutter, comprising:

[0007] Collect and store biological signals;

[0008] performing differential sorting on the biological signal;

[0009] Construct a quantile map for the differentially sorted signal according to the set step size;

[0010] Selecting a predetermined number of feature points in the bitmap to calculate a target curvature of a line connecting the feature points;

[0011] Whether ventricular flutter or ventricular fibrillation exists is identified based on the comparison between the target curvature and a set threshold.

[0012] The step of selecting a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points includes:

[0013] Selecting three segmentation points in a predetermined area of ​​the quantile map;

[0014] Obtaining the vertical coordinates corresponding to the three segmentation points;

[0015] Connect the three vertical coordinates and calculate the angle cosine with the dividing point located in the middle as the corner point;

[0016] The obtained angle cosine is used as the target curvature.

[0017] Optionally, the quantile map is a percentile map;

[0018] Optionally, the selection range of the segmentation points is greater than 40% and less than 99%;

[0019] As another implementation, selecting a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points includes:

[0020] Selecting a first feature point and a second feature point in the bitmap;

[0021] Calculating the slope of the line connecting the first feature point and the second feature point;

[0022] The slope is taken as the target curvature.

[0023] Optionally, the quantile map is a percentile map;

[0024] Optionally, the segmentation point selection range of the first feature point is 40%-80%, the segmentation point selection range of the second feature point is 90%-99%, and more preferably, the segmentation point selection range of the second feature point is 95%-99%.

[0025] As another implementation manner, selecting a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points includes:

[0026] Selecting a plurality of feature points in the bitmap;

[0027] Substitute multiple feature points into Feature(i)=SECG(Ai1) / SECG(Ai0) and calculate them respectively; where Ai0 is the high percentile point and Ai1 is the low percentile point;

[0028] The calculated multiple Features (i) are brought into the pre-trained learning model for learning, and the target curvature is output.

[0029] As another implementation manner, selecting a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points includes:

[0030] The longitudinal coordinate values ​​corresponding to all the segmentation points in the quantile map are used as deep learning input data, and are brought into the deep learning model for learning, and the target curvature is output.

[0031] As a preferred embodiment, the differential sorting of the biological signals includes:

[0032] Differentiating the biological signal according to a predetermined step length;

[0033] Take the absolute value of the array data after difference and record it in the difference buffer;

[0034] Sort the data in the differential buffer from small to large and record them in the sort buffer.

[0035] As a preferred embodiment, constructing a quantile map according to a set step size for the differentially sorted signal includes:

[0036] Set the quantile step size;

[0037] The quantile sequence obtained according to the quantile step length is used as the horizontal axis;

[0038] Discretize the data of each array after taking the absolute value into a predetermined number of values, where the predetermined number is the same as the number of segmentation points of the quantile sequence;

[0039] The discretized values ​​are matched one by one with the horizontal coordinates and used as the quantile amplitude of the horizontal coordinates to form the quantile map.

[0040] In a second aspect, an embodiment of the present application provides a device for identifying ventricular flutter and ventricular fibrillation, comprising:

[0041] A storage unit, used for collecting and storing biological signals;

[0042] A differential sorting unit, used for performing differential sorting on the biological signal;

[0043] A quantile map construction unit, used for constructing a quantile map according to a set step size for the differentially sorted signal;

[0044] A calculation unit, used for selecting a predetermined number of feature points in the bitmap to calculate a target curvature of a line connecting the feature points;

[0045] The identification unit is used to identify whether ventricular flutter or ventricular fibrillation exists based on the comparison between the target curvature and a set threshold.

[0046] In a third aspect, an embodiment of the present application provides an electronic device comprising a memory, a processor, a bus, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for identifying ventricular fibrillation and ventricular flutter as described in the first aspect is implemented.

[0047] In a fourth aspect, an embodiment of the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for identifying ventricular flutter and ventricular fibrillation as described in the first aspect.

[0048] The technical solution of the present application is applied. In the technical solution of the present application, the biosignal is divided into biosignal data segments, and the data in each biosignal data segment is differentially sorted to construct a quantile map. The amplitudes of the biosignals in the quantile map are sorted in sequence, and the arrangement order is usually from low-amplitude signals to high-amplitude signals. Since the energy characteristics of large body motion interference signals are higher than the energy characteristics of biological signals, the large body motion interference signals will be arranged in sequence to the far right of the quantile map. The present application uses the target curvature (curvature of the curve) on the quantile map as an identification feature on the left side of the quantile map, or extracts approximate features that can reflect the curvature of the curve as feature recognition of ventricular fibrillation and flutter. Since the high-amplitude signal on the far right of the quantile map, that is, the large body motion interference signal, is filtered out, the recognition method in the present application improves the signal fidelity of the feature extraction process, reduces the misleading of local interference to recognition, and improves the accuracy of ventricular fibrillation and flutter recognition.

[0049] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present application.

[0051] Figure 1 A schematic diagram showing a flow chart of an embodiment of a method for identifying ventricular fibrillation and ventricular flutter of the present application is shown;

[0052] Figure 2 A percentile diagram of an embodiment of a method for identifying ventricular fibrillation and ventricular flutter according to the present application;

[0053] Figure 3 for Figure 2 A schematic diagram of the distribution of the segmentation points of the percentile map in the illustrated embodiment;

[0054] Figure 4 A schematic diagram of the structure of an embodiment of a ventricular fibrillation and ventricular flutter identification device of the present application;

[0055] Figure 5 An exemplary system architecture diagram of a method for identifying ventricular flutter and ventricular fibrillation is shown;

[0056] Figure 6A schematic structural diagram of an embodiment of an electronic device of the present application. DETAILED DESCRIPTION

[0057] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0058] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0059] In the absence of conflict, the embodiments in this application and the features in the embodiments may be combined with each other.

[0060] In order to make the technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0061] Electric defibrillation therapy needs to be based on the measurement of the distribution and complexity of physiological signals. At present, sample entropy is used in evaluating the complexity of physiological time series (EEG, sEMG, etc.) and diagnosing pathological conditions. Sample entropy (SampleEntropy, SampEn) measures the complexity of time series by measuring the probability of generating new patterns in the signal. The greater the probability of generating new patterns, the greater the complexity of the sequence. In other words, the lower the value of sample entropy, the higher the self-similarity of the sequence. The larger the value of sample entropy, the more complex the sample sequence.

[0062] When processing data of biological signals such as heart rate or blood pressure, by calculating the sample entropy of each segment of the entire time period, a sequence of sample entropy trend graphs can be generated, which helps to observe the complexity changes of biological signals such as heart rate or blood pressure. For example, irregular high-frequency fluctuations may lead to large sample entropy, while large increases and decreases in heart rate may cause sample entropy to decrease.

[0063] When using sample entropy to process data to identify ventricular flutter and ventricular fibrillation, the usual approach is to divide the entire analysis data segment into multiple segments according to the duration, and perform some energy feature calculations on the multiple segments; due to differences in the scale of energy features, many methods also standardize the energy features, such as selecting the maximum value in multiple segments to standardize the data, or selecting the maximum value in multiple segments as the basis for screening other data. The standardized value will be used as a hyperparameter for data learning.

[0064] When patients have clinical symptoms such as ventricular flutter and ventricular fibrillation, they may experience convulsions and other body movements due to discomfort, which may cause occasional large body movement interference in the collected physiological signals. When using the sample entropy method to analyze the signal, two hyperparameter settings are required, one is m and the other is r; m is a step size hyperparameter used in the calculation. r is a threshold, which is generally selected as 0.1 to 0.25 times the signal standard deviation. When calculating the sample entropy, a certain transformation is performed based on the signal, and then the transformed value is compared with the threshold, and then the next step is processed. When the energy characteristics of the large body movement interference signal are higher than the energy characteristics of the cardiac signal, the fluctuation characteristics of the large body movement interference signal will affect the standard deviation of the signal. Moreover, the characteristics of the standard deviation calculation method, even if the value is very small, but the amplitude is large, will have a great impact on the result, and then will also have a great impact on the threshold r. Therefore, in the method of using sample entropy to identify ventricular flutter and ventricular fibrillation, the standard deviation characteristics of the large body movement interference signal will interfere with the threshold calculation, resulting in subsequent calculation errors, and thus affecting the accurate identification of ventricular flutter and ventricular fibrillation.

[0065] Figure 1 FIG. 2 shows a flow chart of an embodiment of the method for identifying ventricular fibrillation and ventricular flutter of the present application. Figure 1 As shown, the method for identifying ventricular fibrillation and ventricular flutter in this embodiment may include the following steps:

[0066] Step 101, collect biological signals and store them.

[0067] Specifically, the biological signal includes but is not limited to an electrocardiogram signal, a blood pressure signal, etc. The biological signal is collected in real time and stored in a raw data buffer OECG.

[0068] The real-time data collection can be performed by collecting data segments. Specifically, at the beginning of real-time data collection, a counter can be used to count, and when the collected data reaches a predetermined number, the collection of the data segment is stopped, and the collected data segment is stored in the data buffer and prepared for the next step of processing, and the collection unit for collecting biological signals starts counting again to start the collection of the next data segment.

[0069] Furthermore, data preprocessing can be performed on the data segments stored in the original data buffer OECG, and the preprocessing includes but is not limited to filtering. As one of the options, when the biological signal is an electrocardiogram signal, the filtering process can use a 5-20Hz Butterworth filter to digitally filter the data. The data after data preprocessing can remove problems such as power frequency interference, electromyographic interference and baseline drift, making the biological signal purer and facilitating subsequent signal analysis and comparison. The pre-accepted data segments can be stored in the preprocessing buffer MECG.

[0070] Step 102: perform differential sorting on the biological signals.

[0071] The biological signal data segments stored in the original data buffer OECG or the original data buffer OECG by step 101 are differentiated according to a predetermined step length, and then the absolute value of the array data after differentiation is taken and recorded in the differential buffer DECG, and the data in the differential buffer DECG is sorted from small to large and recorded in the arrangement buffer SECG.

[0072] Specifically, in one embodiment, the raw data buffer OECG or the biological signal data segment of the raw data buffer OECG is differentiated with a step length of m (including but not limited to 4 milliseconds). The collection of biological signals can be considered as time series data. The differential model is not specifically limited in this application. Any model that can eliminate the fluctuation data in the raw data buffer OECG or the raw data buffer OECG and make the data more stable can be used to implement the differential processing in this application.

[0073] Step 502: construct a quantile map for the differentially sorted signal according to a set step size.

[0074] Specifically, in this step, the quantile step size can be set first. The quantile sequence obtained according to the quantile step size is used as the horizontal coordinate; the data of each array after taking the absolute value is discretized into a predetermined number of values, which is the same as the number of segmentation points of the quantile sequence; the discretized values ​​are matched one by one with the horizontal coordinate and used as the quantile amplitude of the horizontal coordinate to form a quantile map.

[0075] The following uses the example of selecting 1% as the quantile step size. If the quantile step size is 1%, the quantile sequence is 1%, 2%, 3%, ..... 100%. The number of segmentation points in the quantile sequence is 100.

[0076] The quantile sequence obtained according to the quantile step length is used as the horizontal coordinate Xqq=1%, 2%, ... 100%.

[0077] Discretize the data in the arrangement buffer SECG into 100 values, and use the 100 values ​​as the one-to-one correspondence with the horizontal axis and as the quantile amplitude of the horizontal axis, recorded as the vertical axis, Yqq = SECG (Xqq), Xqq as the horizontal axis, Yqq as the vertical axis, and the horizontal and vertical axes form a quantile map, see Figure 2 .

[0078] It should be noted that the quantile step size of 1% is only exemplary. Values ​​such as 1 / 20, 1 / 50, 1 / 200, etc. can all be applied to the production of quantile maps. To ensure the acquisition density of the quantile amplitude, the quantile step size should be selected to be less than 1 / 20 as much as possible.

[0079] Step 503 : Select a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points.

[0080] In one of the implementation schemes for obtaining the target curvature, specifically, three segmentation points are selected in a predetermined area of ​​the quantile map; the vertical coordinates corresponding to the three segmentation points are obtained; the three vertical coordinates are connected, and the angle cosine is calculated with the segmentation point located in the middle position as the corner point; and the obtained angle cosine is used as the target curvature.

[0081] by Figure 2 Taking the percentile graph shown as an example, the calculation of the target curvature is described in detail.

[0082] Select the quantile amplitudes corresponding to the three segmentation points with abscissas of 80%, 60%, and 40% in the quantile map, that is, the characteristic points A, B, and C corresponding to the three ordinates, see Figure 3 . Calculate the angle cosine value with 60% as the corner point as the target curvature. Specifically, the steps are as follows:

[0083] A calculates the coordinates: the horizontal coordinate of the quantile Ax, substitutes it into the sort buffer SECG(Ax), and obtains the vertical coordinate, thereby obtaining the coordinates of the three feature points (X80, Y80), (X60, Y60), (X40, Y40).

[0084] B calculates the side length:

[0085] Hypotenuse a = ((X80-X60)^2+(Y80-Y60)^2)^.05;

[0086] Hypotenuse b = ((X40-X60)^2+(Y40-Y60)^2)^.05;

[0087] Opposite side c = ((X80-X40)^2+(Y80-Y40)^2)^.05.

[0088] C calculates the cosine of an angle:

[0089] Cosa=(a^2+b^2-C^2) / (2*a*b).

[0090] The obtained cosine value of the angle is used as the target curvature.

[0091] Step 105: Identify whether ventricular flutter or ventricular fibrillation exists based on the comparison between the target curvature and the set threshold.

[0092] Set the angle Thr corresponding to the corner point to 150 degrees as the threshold.

[0093] If Cosa>cos(2*pi*150 / 360) holds, it is identified as ventricular fibrillation or ventricular flutter, otherwise it is identified as negative.

[0094] , where pi is π, π=3.1415926.

[0095] It should be noted that the selection of Thr=150 degrees as the threshold value is only exemplary, and Thr may also be any angle value greater than 150 degrees and less than 180 degrees.

[0096] In this embodiment, the selection range of the three segmentation points is within the range of 40% to 99%. The original biological signal has signals with different amplitudes and large fluctuation signals. Through differential processing, a part of the noise signal can be filtered out. After the data is absolute valued and sorted, the biological signals in each biological signal data segment are arranged in order. Figure 2 , the amplitude of the biological signal increases from left to right, and the signal with the largest amplitude is located on the far right of the quantile graph. When the biological signal is an ECG signal, and the patient has a large body movement due to ventricular flutter or ventricular fibrillation, the amplitude of the ECG signal will be larger than the normal ECG amplitude, corresponding to Figure 2 , the amplitude of the signal is often located at the far right of the quantile diagram. According to the characteristics of the distribution diagram of the interference situation analyzed experimentally, this type of signal is usually between 99% and 100%. In this embodiment, the segmentation point is selected in the range of 40% to 99%. This value range can effectively exclude signals with larger body movements, thereby making the identification of ventricular fibrillation and flutter more accurate.

[0097] It should be noted that the method for obtaining the target curvature in the present application is not limited to the method in step 104, and other various methods may also be used to obtain the target curvature.

[0098] In another embodiment of obtaining the target curvature, the method for obtaining the target curvature includes:

[0099] Select the first feature point and the second feature point in the quantile map; calculate the slope of the line connecting the first feature point and the second feature point; and use the slope as the target curvature.

[0100] In one feasible solution, the quantile map may be a percentile map, see Figure 2 When the percentile map is used, the segmentation point selection range of the first feature point is 40%-80%, and the segmentation point selection range of the second feature point is 90%-99%. More preferably, the segmentation point selection range of the second feature point is 95%-99%. In this embodiment, the slope is used to characterize the rate of change of the fluctuation size difference. Since the biosignals in each biosignal data segment are arranged in sequence, see Figure 2The amplitude of the biological signal increases from left to right. The signal with the largest amplitude is located on the far right of the quantile diagram. In the range of 40%-80%, it belongs to the stable segment of the signal, and 90%-99% belongs to the fluctuating segment. According to the characteristics of the distribution diagram of the interference situation analyzed experimentally, the signal generated when the patient has a large body movement due to ventricular flutter or ventricular fibrillation is usually between 99% and 100%. Therefore, when the second characteristic point uses a value in the range of 95%-99%, the abnormal signal can be excluded. At the same time, the high amplitude of the ECG signal is usually concentrated in the range of 95%-99%, so it can accurately identify whether ventricular fibrillation or ventricular flutter occurs.

[0101] Specifically, Figure 2 Take the quantile map shown in as an example, select two percentiles and the quantile values ​​corresponding to the two percentiles. Among them, the first percentage point A0 is the 95% percentile, and A1 is the 65% percentile; substitute the formula Feature = SECG (A0) / SECG (A1), and the calculated result Feature is used as the target curvature.

[0102] The target curvature is compared with a set threshold. When the calculated result of the Feature is greater than the threshold, it is identified as ventricular fibrillation or ventricular flutter, otherwise it is identified as negative.

[0103] In yet another embodiment of obtaining the target curvature, the method for obtaining the target curvature includes:

[0104] Selecting a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points includes:

[0105] Select multiple feature points in the quantile map; substitute multiple feature points into

[0106] Feature(i)=SECG(Ai1) / SECG(Ai0) are calculated respectively; Ai0 is the high percentile point and Ai1 is the low percentile point; the calculated multiple Feature(i) are brought into the pre-trained learning model for learning, and the target curvature is output.

[0107] In this embodiment, the principle for selecting Ai0 and Ai1 is that the percentile of Ai0 only needs to be greater than the percentile of Ai1.

[0108] In some implementations, multiple Features (i) may be brought into a pre-trained machine, including but not limited to a SVM (Support Vector Machine), an adaboost model, and other models for learning.

[0109] In yet another embodiment of obtaining the target curvature, the method for obtaining the target curvature includes:

[0110] The longitudinal coordinate values ​​corresponding to all segmentation points in the quantile map are used as deep learning input data, and are brought into the deep learning model for learning, and the target curvature is output.

[0111] Specifically, the quantile step size is still 1% for example. If the quantile step size is 1%, the quantile sequence is 1%, 2%, 3%, .....100%. The number of segmentation points of the quantile sequence is 100.

[0112] The quantile sequence obtained according to the quantile step is used as the horizontal coordinate Xqq=1%, 2%...100%.

[0113] Discretize the data in the arrangement buffer SECG into 100 values, and use the 100 values ​​as the corresponding values ​​of the horizontal axis and the quantile amplitude of the horizontal axis, recorded as the vertical axis, Yqq = SECG (Xqq), Xqq as the horizontal axis, Yqq as the vertical axis, see Figure 2 .

[0114] Use Yqq as the original data and input it into the trained deep learning models such as CNN (convolutional neural network) and DNN (deep neural network), and output the calculated result of the target curvature.

[0115] It can be seen from the above technical solution that this patent constructs a quantile map by dividing the biological signal into biological signal data segments, performing differential sorting on the data in each biological signal data segment, and then sorting the amplitude of the biological signal in the quantile map in sequence, usually in the order of gradually increasing from low-amplitude signals to high-amplitude signals. Since the energy characteristics of the large body motion interference signal are higher than the energy characteristics of the biological signal, the large body motion interference signal will be arranged in order to the far right of the quantile map. The present application is located on the left side of the quantile map based on the target curvature (curvature of the curve) on the quantile map as an identification feature, or extracts approximate features that can reflect the curvature of the curve as a feature identification of ventricular fibrillation and ventricular flutter. Since the high-amplitude signal located on the far right of the quantile map, that is, the large body motion interference signal, is filtered out, the identification method in the present application improves the signal fidelity of the feature extraction process, reduces the misleading of local interference to the identification, and improves the accuracy of ventricular flutter and ventricular fibrillation identification.

[0116] In addition, since the ventricular fibrillation and ventricular flutter identification method in the present application processes the biological signal data segment by segment, compared with the sample entropy which requires the selection of multiple segments and the selection of the maximum value among the multiple segments for data screening, the method shown in the present application has the advantages of saving computing time and improving recognition efficiency.

[0117] CUDB (CU Ventricular Tachyarrhythmia Database) was used as the database for evaluation, and the sample entropy features and the quantile map features in this application were used as algorithm inputs respectively. It was verified that the classification accuracy using the quantile map features was nearly 10% higher than that using the sample entropy features.

[0118] For the calculation of accuracy, confusion matrix can be used for calculation. Confusion matrix is ​​a visualization tool, which is mainly used to compare classification results and actual measured values. The accuracy of classification results can be displayed in a confusion matrix.

[0119] The structure of the confusion matrix is ​​generally shown in Table 1 below.

[0120] Table 1:

[0121]

[0122] Among them, the confusion matrix includes four symbols TP (True Positives), FP (False Positives), TN (True Negatives), FN (False Negatives), and their corresponding meanings are [True] samples are positive, the predicted result is positive, [False Positive] samples are negative, the predicted result is positive, [True Negative] samples are negative, the predicted result is negative, and [False Negative] samples are positive, and the predicted result is negative.

[0123] Precision is the most commonly used classification performance indicator. It can be used to represent the accuracy of the model, that is, the number of correct model recognitions / total number of samples. Generally speaking, the higher the accuracy of the model, the better the model effect. The formula for calculating the precision using the confusion matrix is:

[0124] accuracy=(TP+TN) / (TP+FN+FP+TN)

[0125] Taking the total number of samples of 745 as an example, the sample entropy features and the quantile map features described in this application are used as algorithm inputs to calculate the accuracy, and the corresponding confusion matrices are shown in Table 2 and Table 3. Among them, Table 2 is the confusion matrix of the quantile map features, and Table 3 is the confusion matrix of the sample entropy features.

[0126]

[0127] Substituting into the above calculation formula of accuracy, the results are 0.87 and 0.79 respectively, and the accuracy is nearly 10%. It can be seen that the recognition method of the bitmap feature used in this application has a better recognition effect.

[0128] Further references Figure 4As an implementation of the methods shown in the above figures, the present application provides an embodiment of a ventricular fibrillation and ventricular flutter identification device. Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0129] like Figure 4 As shown, the target signal control device 400 of this embodiment includes: a storage unit 401, a differential sorting unit 402, a bitmap construction unit 403, a calculation unit 404 and an identification unit 405.

[0130] The storage unit 401 is used to collect and store biological signals;

[0131] A differential sorting unit 402, used for performing differential sorting on the biological signal;

[0132] A quantile map construction unit 403, configured to construct a quantile map according to a set step size for the differentially sorted signal;

[0133] A calculation unit 404 is used to select a predetermined number of feature points in the quantile map to calculate a target curvature of a line connecting the feature points;

[0134] The identification unit 405 is used to identify whether ventricular flutter or ventricular fibrillation exists based on the comparison between the target curvature and the set threshold.

[0135] Figure 5 An exemplary system architecture 500 is shown to which the ventricular flutter and ventricular fibrillation identification method of the present application can be applied.

[0136] like Figure 1 As shown, the system architecture 500 may include a biological signal collection device 501 and a terminal device 502. The biological signal collection device 501 and the terminal device 502 are connected via a medium such as a network or a cable to achieve data transmission.

[0137] The biosignal acquisition device 501 is used to simultaneously acquire biosignals of the target object. It is understandable that the biosignal acquisition device 501 is used to acquire different types of biosignals, for example, the biosignal acquisition device 501 is used to acquire electrocardiogram signals, and can also be used to acquire blood pressure signals or venous signals, etc.

[0138] The biological signal collection device 501 may send the collected signal to the terminal device 502 in real time. After receiving the above signal, the terminal device 502 may process the above signal to identify the target signal in the above biological signal.

[0139] The terminal device 502 can be hardware or software. When the terminal device 502 is hardware, it can be various electronic devices, including but not limited to tablet computers, laptop computers, and desktop computers. When the terminal device 502 is software, it can be installed in the electronic devices listed above. It can be implemented as multiple software or software modules (for example, to provide distributed services), or it can be implemented as a single software or software module. No specific limitation is made here.

[0140] In other applications, the above system architecture may further include a database 503. The database 503 may be used to store the signals collected by the biological signal collection device 501, so that the terminal device 502 can use the above signals for learning at an appropriate time.

[0141] It should be noted that the signal processing method provided in the embodiment of the present application is generally executed by the terminal device 502. Accordingly, the signal processing device is generally disposed in the terminal device 502.

[0142] It should be understood that Figure 1 The number of biological signal collection devices, terminal devices and databases in the embodiment is only for illustration. According to the implementation requirements, any number of biological signal collection devices, terminal devices and databases may be provided.

[0143] In addition, in the technical solution of the present application, an electronic device is also proposed.

[0144] Figure 6 A schematic structural diagram of an electronic device provided in one embodiment of the present application is shown.

[0145] like Figure 6 As shown, the electronic device may include a processor 601, a memory 602, a bus 603, and a computer program stored in the memory 602 and executable on the processor 601, wherein the processor 601 and the memory 602 communicate with each other via the bus 603. When the processor 601 executes the computer program, the steps of the above method are implemented, for example, including: collecting and storing biological signals; performing differential sorting on the biological signals; constructing a quantile map for the differential sorted signals according to a set step size; selecting a predetermined number of feature points in the quantile map to calculate the target curvature of the line connecting the feature points; and identifying whether ventricular flutter or ventricular fibrillation exists based on the comparison between the target curvature and a set threshold.

[0146] In addition, in one embodiment of the present application, a non-transitory computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above method are implemented, for example, including: acquiring a first biological signal and a second biological signal of the target object in real time; determining whether a first preset condition is met based on the first biological signal; determining whether the second preset condition is met based on the second biological signal; in response to determining that both the first preset condition and the second preset condition are met, determining a predicted trigger moment based on the first biological signal and / or the second biological signal; adjusting the predicted trigger moment based on the pre-obtained target time deviation to obtain a target trigger moment; and triggering and controlling the target signal based on the target trigger moment.

[0147] In summary, in the technical solution of the present application, the biosignal is divided into biosignal data segments, and the data in each biosignal data segment is differentially sorted to construct a quantile map. The amplitudes of the biosignals in the quantile map are sorted in sequence, and the arrangement order is usually from low-amplitude signals to high-amplitude signals. Since the energy characteristics of large body motion interference signals are higher than the energy characteristics of biological signals, the large body motion interference signals will be arranged in order to the rightmost side of the quantile map. The present application uses the target curvature (curvature of the curve) on the quantile map as an identification feature on the left side of the quantile map, or extracts approximate features that can reflect the curvature of the curve as feature recognition of ventricular fibrillation and flutter. Since the high-amplitude signal on the far right of the quantile map, that is, the large body motion interference signal, is filtered out, the recognition method in the present application improves the signal fidelity of the feature extraction process, reduces the misleading of local interference to the recognition, and improves the accuracy of ventricular fibrillation and flutter recognition.

[0148] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for identifying ventricular flutter and ventricular fibrillation, comprising: Collect and store biological signals; performing differential sorting on the biological signal; Construct a quantile map for the differentially sorted signal according to the set step size; Selecting a predetermined number of feature points in the bitmap to calculate a target curvature of a line connecting the feature points; Whether ventricular flutter or ventricular fibrillation exists is identified based on the comparison between the target curvature and a set threshold.

2. The method according to claim 1, wherein: The selecting a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points comprises: Selecting three segmentation points in a predetermined area of ​​the quantile map; Obtaining the vertical coordinates corresponding to the three segmentation points; Connect the three vertical coordinates and calculate the angle cosine with the dividing point located in the middle as the corner point; The obtained angle cosine is used as the target curvature. Optionally, the quantile map is a percentile map; Optionally, the selection range of the segmentation points is greater than 40% and less than 99%.

3. The method according to claim 1, wherein: The selecting a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points comprises: Selecting a first feature point and a second feature point in the bitmap; Calculating the slope of a line connecting the first feature point and the second feature point; The slope is taken as the target curvature. Optionally, the quantile map is a percentile map; Optionally, the segmentation point selection range of the first feature point is 40%-80%, the segmentation point selection range of the second feature point is 90%-99%, and more preferably, the segmentation point selection range of the second feature point is 95%-99%.

4. The method according to claim 3, wherein: The selecting a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points comprises: Selecting a plurality of feature points in the bitmap; Substitute multiple feature points into Feature(i)=SECG(Ai1) / SECG(Ai0) and calculate them respectively; where Ai0 is the high percentile point and Ai1 is the low percentile point; The calculated multiple Features (i) are brought into the pre-trained learning model for learning, and the target curvature is output.

5. The method according to claim 1, wherein: The selecting a predetermined number of feature points in the bitmap to calculate the target curvature of the line connecting the feature points comprises: The longitudinal coordinate values ​​corresponding to all the segmentation points in the quantile map are used as deep learning input data, and are brought into the deep learning model for learning, and the target curvature is output.

6. The method according to any one of claims 1 to 5, wherein: The differential sorting of the biological signals comprises: Differentiating the biological signal according to a predetermined step length; Take the absolute value of the array data after difference and record it in the difference buffer; Sort the data in the differential buffer from small to large and record them in the sort buffer.

7. The method according to claim 6, wherein: The step of constructing a quantile map for the differentially sorted signal according to a set step size includes: Set the quantile step size; The quantile sequence obtained according to the quantile step length is used as the horizontal axis; Discretize the data of each array after taking the absolute value into a predetermined number of values, where the predetermined number is the same as the number of segmentation points of the quantile sequence; The discretized values ​​are matched one by one with the horizontal coordinates and used as the quantile amplitude of the horizontal coordinates to form the quantile map.

8. A device for identifying ventricular flutter and ventricular fibrillation, comprising: A storage unit, used for collecting and storing biological signals; A differential sorting unit, used for performing differential sorting on the biological signal; A quantile map construction unit, used for constructing a quantile map according to a set step size for the differentially sorted signal; A calculation unit, used for selecting a predetermined number of feature points in the bitmap to calculate a target curvature of a line connecting the feature points; The identification unit is used to identify whether ventricular flutter or ventricular fibrillation exists based on the comparison between the target curvature and a set threshold.

9. An electronic device comprising a memory, a processor, a bus, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for identifying ventricular flutter and ventricular fibrillation according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for identifying ventricular flutter and ventricular fibrillation as claimed in any one of claims 1 to 7 is implemented.

Images