Method, device, medium and equipment for detecting premature beat based on PPG waveform region features

Abstract

The application discloses a method, device, medium and equipment for detecting premature beat based on PPG waveform region characteristics, and the flow of the method is as follows: preprocessing the collected original PPG signal; obtaining the average heartbeat time interval and the PPG pulse segment by using an improved adaptive low-complexity peak and valley detection algorithm; establishing three PPG waveform templates based on the waveform region characteristics of the PPG pulse segment and the average heartbeat time interval; dividing the preprocessed original PPG signal to obtain the PPG signal segment to be classified; classifying the PPG signal segment to be classified based on the multivariate PPG signal quality index and the template to obtain the PPG signal segment classification result; obtaining the number of premature beat occurrences and the total number of heartbeats contained in the preprocessed original PPG signal according to the classification result; and obtaining the premature beat percentage of the total heartbeats by dividing the number of premature beat occurrences by the total number of heartbeats. The application can detect the premature beat condition of the subject by starting from the PPG waveform region characteristics, and the algorithm is effective and feasible, and has low complexity and wide applicability.

Description

Technical Field

[0001] This invention belongs to the field of biomedical optoelectronic signal processing, specifically a method, device, medium, and equipment for detecting premature beats based on PPG waveform region features. Background Technology

[0002] Premature beats (PVCs) are a heart condition, commonly including atrial and ventricular PVCs. Patients experience discomfort when PVCs occur. Detecting PVCs and monitoring their percentage of total heartbeats is crucial for determining the need for medication or other interventions. Traditional algorithms for detecting PVCs mostly use electrocardiogram (ECG) data analysis because ECGs have distinct waveforms that accurately distinguish abnormal heart rhythms. However, ECG detection requires electrodes to be placed on the body surface, which can cause discomfort or allergic reactions. Therefore, photoplethysmography (PPG) is now widely used for PVC detection. PPG is a non-invasive method that uses photoelectric methods to detect changes in blood volume by detecting reflected or transmitted light from tissues. PPG data comes from a sensor, typically attached to the fingertip, avoiding the discomfort of ECG electrodes on the body surface, making it particularly suitable for long-term monitoring. This is extremely useful for applications requiring long-term patient monitoring (such as sleep monitoring) or continuous monitoring (such as fitness trackers).

[0003] Many existing PPG methods for detecting premature beats rely on machine learning approaches that train models using large amounts of prior data or on fixed prior thresholds to identify premature beats. This increases algorithm complexity, reduces the speed of premature beat detection, limits the range of embedded platforms to which it can be applied, and reduces the applicability of the detection method to different populations and different PPG acquisition methods. Therefore, it is particularly important to propose a premature beat detection algorithm that does not require a large amount of prior data and prior thresholds, thus making it widely applicable to different populations and different PPG acquisition methods, and applicable to a variety of embedded platforms. Summary of the Invention

[0004] Based on this, and in response to the above problems, this invention proposes a method, apparatus, medium, and device for detecting premature beats based on PPG waveform region features, in order to at least solve the problems of high algorithm complexity, limited applicability, and limited application on embedded platforms in premature beat detection methods.

[0005] In a first aspect, the present invention provides a method for detecting premature beats based on PPG waveform region features, the method comprising:

[0006] S1. Preprocess the acquired raw PPG signal;

[0007] S2. The preprocessed raw PPG signal is further processed using an improved adaptive low-complexity peak and valley detection algorithm to obtain the average heartbeat time interval and PPG pulse segments.

[0008] S3. Based on the waveform region characteristics of the PPG pulse segment and the average heartbeat time interval, three PPG waveform templates are established; the preprocessed original PPG signal is divided based on the PPG pulse segment to obtain PPG signal segments to be classified; the PPG signal segments to be classified are classified based on multivariate PPG signal quality indicators and the templates to obtain PPG signal segment classification results; heartbeat information is obtained according to the PPG signal segment classification results, the heartbeat information including the number of premature beats and the total number of heartbeats;

[0009] S4. Divide the number of premature beats by the total number of heartbeats to obtain the percentage of premature beats in the total heartbeats.

[0010] Optionally, the preprocessing of the acquired raw PPG signal specifically includes:

[0011] Acquire raw PPG signals;

[0012] The PPG signal is filtered using a Butterworth bandpass filter, and the filtered PPG signal is then normalized in amplitude.

[0013] Optionally, the improved adaptive low-complexity peak and valley detection algorithm is used to further process the preprocessed raw PPG signal to obtain the average heartbeat time interval and PPG pulse segments, specifically including:

[0014] An improved adaptive low-complexity peak and valley detection algorithm is used to locate the peak and valley values ​​of the preprocessed raw PPG signal. The average heartbeat time interval is obtained through the peak location, and the PPG pulse segments are obtained by segmenting the preprocessed raw PPG signal through the valley location.

[0015] Optionally, the process involves establishing three PPG waveform templates based on the waveform region features of the PPG pulse segments and the average heartbeat time interval; dividing the preprocessed raw PPG signal based on the PPG pulse segments to obtain PPG signal segments to be classified; classifying the templates and the PPG signal segments to be classified based on multivariate PPG signal quality indicators to obtain PPG signal segment classification results; and obtaining heartbeat information based on the PPG signal segment classification results, wherein the heartbeat information includes the number of premature beats and the total number of heartbeats, specifically including:

[0016] Three PPG waveform templates are established based on the waveform region characteristics of the PPG pulse segment and the average heartbeat time interval;

[0017] Based on the PPG pulse segments, the preprocessed original PPG signal is divided using two division methods: independent segments and combinations of adjacent segments, to obtain PPG signal segments to be classified.

[0018] The template and the PPG signal segment to be classified are assessed for similarity based on a multivariate PPG signal quality index to obtain the assessment result; the PPG signal segment to be classified is then classified based on the assessment result to obtain the PPG signal segment classification result.

[0019] Heartbeat information is obtained based on the PPG signal segment classification results, and the heartbeat information includes the number of premature beats and the total number of heartbeats.

[0020] Optionally, the waveform region features of the PPG pulse segment include feature values; the establishment of three PPG waveform templates based on the waveform region features of the PPG pulse segment and the average heartbeat time interval specifically includes:

[0021] Three PPG waveform templates are established based on the aforementioned characteristic values ​​and the average heartbeat time interval; wherein, the formula for calculating the characteristic values ​​is:

[0022]

[0023] in,

[0024]

[0025] Where K is the characteristic value, y(x) is the amplitude of the PPG pulse segment, x is the pulse time, and y min y is the minimum amplitude of a PPG pulse segment. max y represents the maximum amplitude of a PPG pulse segment. m X is the average amplitude of the PPG pulse segment, and X is the duration of the pulse segment.

[0026] Optionally, the three PPG waveform templates are: a PPG waveform template corresponding to a sinus rhythm; a PPG waveform template consisting of a first type premature heartbeat and a sinus rhythm preceding the premature heartbeat; and a PPG waveform template consisting of a second type premature heartbeat and a sinus rhythm preceding the premature heartbeat.

[0027] Optionally, the first type of premature heartbeat is a premature heartbeat with no obvious pulse fluctuation in the PPG waveform; the second type of premature heartbeat is a premature heartbeat with obvious pulse fluctuation in the PPG waveform.

[0028] In a second aspect, the present invention provides a device for detecting premature beats based on PPG waveform region features, the device comprising:

[0029] The preprocessing module is used to preprocess the acquired raw PPG signal;

[0030] The first processing module is used to further process the preprocessed raw PPG signal using an improved adaptive low-complexity peak and valley detection algorithm to obtain the average heartbeat time interval and PPG pulse segments.

[0031] The signal segment classification module is used to establish three PPG waveform templates based on the waveform region characteristics of the PPG pulse segments and the average heartbeat time interval; to divide the preprocessed original PPG signal based on the PPG pulse segments to obtain PPG signal segments to be classified; to classify the PPG signal segments to be classified based on multivariate PPG signal quality indicators and the templates to obtain PPG signal segment classification results; and to obtain heartbeat information based on the PPG signal segment classification results, wherein the heartbeat information includes the number of premature beats and the total number of heartbeats.

[0032] The final data acquisition module is used to divide the number of premature beats by the total number of heartbeats to obtain the percentage of premature beats in the total heartbeats.

[0033] In a third aspect, the present invention provides a computer-readable storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, the processor performs the steps of any one of the methods described in the first aspect above.

[0034] In a fourth aspect, the present invention provides a computer device including a memory and a processor, characterized in that the memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of any one of the methods described in the first aspect above.

[0035] Based on the above, the present invention has the following advantages or beneficial effects:

[0036] This invention proposes a method, apparatus, medium, and device for detecting premature beats based on PPG waveform region features. This method detects premature beats by classifying heartbeat types based on PPG waveform region features. The percentage of premature beats in the total heartbeats can be calculated based on the classification and detection results. Compared with traditional premature beat detection methods, its advantages are low algorithm complexity and the use of only a very small number of prior thresholds, i.e., low requirements for prior conditions of specific signals. It can be widely applied to different populations and different PPG acquisition methods, and can be widely used in various embedded development platforms. Attached Figure Description

[0037] 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.

[0038] in:

[0039] Figure 1 This is a flowchart of a method for detecting premature beats based on PPG waveform region features proposed in this invention;

[0040] Figure 2 The image shows the detection results of the improved adaptive low-complexity peak and valley detection algorithm of this invention on an example PPG signal.

[0041] Figure 3 The waveform morphology is labeled for sinus rhythm and the two types of premature heartbeats mentioned in this invention.

[0042] Figure 4 These are three PPG waveform templates established for example signals in this invention;

[0043] Figure 5 This is a schematic diagram of the PPG signal segment classification results obtained on the example signal in this invention;

[0044] Figure 6 This is a schematic diagram of the device for detecting premature beats based on PPG waveform region features in this invention.

[0045] Figure 7 This is a schematic diagram of the internal structure of a device provided by the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0047] The present invention will be further described in detail below with reference to the accompanying drawings. The present invention provides a method for detecting premature beats based on PPG waveform region features. Using this method, the number of segments containing premature heartbeats in the acquired PPG signal can be counted, and then the percentage of premature beats in the total heartbeats of the subject during the acquisition time can be calculated.

[0048] PPG signal, or photoplethysmography signal, can be acquired in two modes: transmission and reflection. In transmission mode, light transmitted through the medium is detected by a photodetector opposite the light source. In reflection mode, the photodetector detects light backscattered or reflected from tissue, bone, or blood vessels.

[0049] Premature beats (PVCs) are a heart condition, with the most common types being atrial premature beats (PACs) and ventricular premature beats (PVCs). PACs refer to atrial premature beats, while PVCs refer to ventricular premature beats. PAC1 / PVC1 is a type of premature beat with no obvious pulse fluctuations in the PPG waveform, while PAC2 / PVC2 is a type of premature beat with obvious pulse fluctuations in the PPG waveform.

[0050] Many existing premature beat detection algorithms rely on machine learning for classification and detection. However, machine learning suffers from weak model interpretability and requires a large amount of prior data for training. Publicly available PPG datasets containing accurately labeled premature beats are extremely rare. Using manually labeled PAV / PVC datasets as training data can lead to inaccurate classification results due to labeling errors, and the applicability to embedded platforms is limited. Furthermore, existing technologies have proposed feature extraction algorithms based on continuous difference root mean square and Shannon entropy for premature beat identification, but these rely heavily on fixed thresholds, which reduces the robustness and applicability of the algorithms.

[0051] Based on the above problems, this invention proposes a method, apparatus, medium, and device for detecting premature beats based on PPG waveform region features. This addresses at least the issues of high algorithm complexity, limited application to embedded platforms, and limited applicability of traditional premature beat detection methods. Compared with traditional premature beat detection methods, this invention has the advantages of low algorithm complexity and the use of only a very small number of prior thresholds, i.e., low requirements for prior conditions of specific signals. It can be widely applied to different populations and different PPG acquisition methods, and can be widely used in various embedded development platforms.

[0052] Please see Figure 1 This is a flowchart of a method for detecting premature beats based on PPG waveform region features proposed in this invention. Figure 1The present invention comprises four steps: S1, preprocessing the acquired raw PPG signal; S2, further processing the preprocessed raw PPG signal using an improved adaptive low-complexity peak and valley detection algorithm to obtain the average heartbeat interval and PPG pulse segments; S3, establishing three PPG waveform templates based on the waveform region features of the PPG pulse segments and the average heartbeat interval; dividing the preprocessed raw PPG signal based on the PPG pulse segments to obtain PPG signal segments to be classified; classifying the PPG signal segments to be classified based on multivariate PPG signal quality indicators and the three PPG waveform templates to obtain PPG signal segment classification results; obtaining heartbeat information based on the PPG signal segment classification results, including the number of premature beats and the total number of heartbeats; S4, dividing the number of premature beats by the total number of heartbeats to obtain the percentage of premature beats in the total heartbeats. It should be noted that, for the sake of clarity in explaining the actual content of this invention, the four steps are broken down into eight smaller steps for detailed and clear description. Specifically, S1 includes step a; S2 includes steps b and c; S3 includes steps d, e, f, and g; and S4 includes step h.

[0053] The specific algorithm flow in step a is as follows: The acquired PPG signal is preprocessed, which refers to removing noise and baseline drift from the original PPG signal and normalizing the signal amplitude. Specifically, to remove high-frequency noise and baseline drift, a third-order Butterworth bandpass filter with cutoff frequencies of 0.5Hz and 4Hz is used to filter the signal and remove unwanted frequency components. Then, the filtered PPG signal is normalized, and the normalization formula is given by (1):

[0054]

[0055] Where x represents the data to be normalized, y represents the normalization result, min(x) represents the minimum value of x, and max(x) represents the maximum value of x.

[0056] Step b includes the following specific algorithm flow: The preprocessed original PPG signal is further processed using an improved adaptive low-complexity peak and valley detection algorithm. This further processing refers to locating peaks and valleys in the preprocessed original PPG signal using the improved adaptive low-complexity peak and valley detection algorithm. The advantages of the improved adaptive low-complexity peak and valley detection algorithm proposed in this invention are:

[0057] (1) The algorithm does not rely on prior thresholds obtained from prior knowledge of a specific type of signal to be measured to set thresholds. Setting thresholds based on prior thresholds will reduce the robustness of the algorithm.

[0058] (2) Under the premise of meeting the adaptive requirements, the algorithm has low complexity and can be applied to a variety of embedded platforms.

[0059] The improved adaptive low-complexity peak and valley detection algorithm proposed in this invention includes the following steps: First, perform a Fourier transform (FFT) on the preprocessed original PPG signal to obtain the dominant frequency of the preprocessed original PPG signal, thus obtaining a rough estimate of the heartbeat interval duration. For PPG signals with a duration six times the rough estimate of the heartbeat interval duration, perform local minimum location to obtain the predicted average heartbeat interval duration. For the preprocessed original PPG signal, divide it into windows according to a duration ten times the predicted average heartbeat interval duration, ensuring that adjacent windows overlap by a window length of three times the predicted average heartbeat interval duration, thus obtaining... For each window segment of the signal, local maxima and minima are located to preserve signal characteristics while reducing computational complexity. Then, local maxima (M_max) and local minima (M_min) matrices are initialized, with a size of L×N, where N is the length of the window segment and L is half the length of the window segment rounded up and subtracted by 1. Linear detrending processing is performed on the window segments, and the local maxima and minima matrices are filled in. A moving window method is used to determine the local maxima and minima of each window segment, where the moving window length (W) is... k ) is: {W k =2k|k=1,2,…,L}, where k is the row coordinate of the matrix, i is the column coordinate of the matrix, and m_max is the element of the partial local maximum matrix. k,i Based on formula (2), the remaining local maximum matrix elements are assigned a value of 1; some local minimum matrix elements (m_min) are determined. k,i Based on formula (3), the remaining local minimum matrix elements are assigned a value of 1.

[0060]

[0061]

[0062] Fill the local maximum matrix according to formula (4), and fill the local minimum matrix according to formula (5);

[0063]

[0064]

[0065] Next, within each segment, the present invention performs row-by-row summation on the local maximum and local minimum matrices to determine the optimal scale, where the scale with the smallest row sum is considered the optimal scale. All local maximum and local minimum elements with row coordinates greater than the optimal scale are removed, resulting in the remaining local maximum and local minimum matrices. Finally, the standard deviation of each column in the remaining local maximum and local minimum matrices is calculated, and the positions where the standard deviation of the remaining local maximum and local minimum matrices is equal to zero are located, yielding preliminary peak and valley location results for the window signal segment. These preliminary peak and valley location results from all window signal segments divided based on the preprocessed original PPG signal are combined. The x-coordinate of each valley and peak position is reduced by 1, and within a small range on either side (e.g., within one-tenth of the predicted average heartbeat interval), local maxima and maxima are found to correct the located peak and valley values. The found local maxima and maxima represent the peak and valley location results of the preprocessed original PPG signal. This method performs excellently in processing PPG signals, exhibiting adaptive and low-complexity characteristics. Please see Figure 2 This is a graph showing the detection results of the improved adaptive low-complexity peak and valley detection algorithm of this invention on an example PPG signal. Peak points are marked with stars, and valley points are marked with circles. According to... Figure 2 The results show the location of peaks and valleys in an example PPG signal using the improved adaptive low-complexity peak and valley detection algorithm.

[0066] The specific algorithm flow in step c is as follows: The average heartbeat time interval and PPG pulse segments are obtained from the positioning results in step b. The average heartbeat time interval refers to the approximate time interval between two heartbeats of the subject during the acquisition of the original PPG signal, obtained from the peak positioning results. The PPG pulse segment refers to the segment in the preprocessed original PPG signal that corresponds to a significant pulse fluctuation. The average heartbeat time interval is obtained by calculating the average time interval between every two peak positioning points on the preprocessed PPG signal. The preprocessed original PPG signal is segmented by valley positioning to obtain PPG pulse segments.

[0067] Step d includes the following specific algorithm flow: Based on the waveform region features of PPG pulse segments and the average heartbeat time interval, three PPG waveform templates are established. The waveform region features of PPG pulse segments include the waveform region feature quantity K value. The three PPG waveform templates are as follows:

[0068] (1) PPG fragment template corresponding to a sinus rhythm;

[0069] (2) A PPG waveform template consisting of a type I premature heartbeat and the sinus rhythm preceding the premature heartbeat;

[0070] (3) The PPG waveform template is formed by a type II premature heartbeat and the sinus rhythm preceding the premature heartbeat.

[0071] The first type of premature beats is characterized by a PPG waveform without significant pulse fluctuations; the second type is characterized by a PPG waveform with significant pulse fluctuations. Please refer to [link / reference]. Figure 3 The waveform morphology is labeled for sinus rhythm and the two types of premature beats mentioned in this invention, based on... Figure 3 You can see the waveforms corresponding to the three heartbeat types involved in template creation. Figure 3 NORM indicates sinus rhythm, PAC1 / PVC1 indicates type I premature beats: premature beats with no obvious pulse fluctuations in the PPG waveform, and PAC2 / PVC2 indicates type II premature beats: premature beats with obvious pulse fluctuations in the PPG waveform.

[0072] The waveform region characteristics of PPG pulse segments and the average heartbeat time interval are used as reference data for template selection. The formula for calculating the waveform region characteristic value K is as follows:

[0073]

[0074] in,

[0075]

[0076] Where K is the characteristic value, y(x) is the amplitude of the PPG pulse segment, x is the pulse time, and y min y is the minimum amplitude of a PPG pulse segment. max y represents the maximum amplitude of a PPG pulse segment. m X is the average amplitude of the PPG pulse segment, and X is the duration of the pulse segment.

[0077] The K value is related to the PPG waveform of the subject. The PPG waveform and area will change under different physiological and pathological conditions. With the change of physiological factors such as peripheral vascular resistance and the degree of vascular wall hardening, the K value will also change accordingly. It can be considered that the corresponding K value of the PPG pulse segment of the same subject at the same time should be kept within a certain range. This invention proposes to use the K value as the basis for selecting the reference data of the third template. If a premature beat of type PAC2 / PVC2 occurs after a normal heartbeat, the amplitude of the PPG pulse segment corresponding to the normal heartbeat will not be able to drop to the trough baseline normally. Therefore, the K value of this PPG pulse segment calculated by formula (6) will exceed the reasonable range of the K value of the patient's PPG pulse segment. Therefore, this invention proposes to screen out PPG pulse segments with unreasonable K values ​​by calculating the K value of each PPG pulse segment. The segment should be initially classified as the PPG pulse segment of the most recent sinus rhythm before the occurrence of PAC2 / PVC2. The PPG pulse segment corresponding to this heartbeat and the next heartbeat that occurs immediately afterward constitutes the reference data for the third template to be established in this invention. Since the duration of the PPG pulse segment formed by PAC1 / PVC1 and its preceding sinus rhythm is approximately 1.5 to 2 times the average heartbeat interval, after excluding the reference data corresponding to the third template, the remaining PPG pulse segments are selected as reference data based on either the first or second template, according to the PPG pulse duration. This invention uses 1.6 times the average heartbeat interval as a threshold; when the PPG pulse segment duration exceeds this threshold, it is classified as reference data based on the second template, and vice versa. Please refer to [link to relevant documentation]. Figure 4 These are the three PPG waveform templates established for the example signals in this invention. Figure 4 As can be seen, the three PPG waveform templates established based on waveform region characteristics and average heartbeat time interval of an example PPG signal have obvious differences in shape and can well represent the three heartbeat types or combinations of heartbeat types.

[0078] The specific algorithm flow included in step e is as follows: Based on PPG pulse segments, the preprocessed original PPG signal is divided into two partitioning methods: independent segments and adjacent segment combinations, to obtain PPG signal segments to be classified. The independent segment partitioning method means that a single PPG pulse segment is regarded as a PPG signal segment to be classified. The adjacent segment combination partitioning method means that each PPG pulse segment, together with its preceding and following PPG pulse segments, forms a PPG signal segment to be classified.

[0079] Step f includes the following specific algorithm flow: Classifying the PPG signal segments to be classified based on multivariate PPG signal quality indices and templates to obtain PPG signal segment classification results. Specifically: Similarity assessment is performed on the three types of templates and the PPG signal segments to be classified using multivariate PPG signal quality indices (SQI), and assessment results are obtained; the PPG signal segments to be classified are then classified according to the assessment results to obtain PPG signal segment classification results. Multivariate PPG signal quality indices refer to a voting combination of direct matching quality indices, linear resampling quality indices, dynamic time warping quality indices, and clipping detection quality indices. Multivariate PPG signal quality indices are used as similarity assessment indicators when matching the three established PPG waveform templates with the PPG signal segments to be classified. That is, using the established templates as a reference, the multivariate signal quality index (SQI) of the PPG segment is calculated (SQI values ​​are between 0 and 1). The higher the SQI value of the PPG segment, the higher the matching degree between the PPG segment and the corresponding template. When matching the PPG signal segment to be classified with the three PPG waveform templates, the amplitude of the PPG signal segment to be classified is normalized using formula (1). When the PPG signal segment to be classified is matched with the first template and the second template, the PPG signal segment to be classified divided by the independent segment division method is used to assess the similarity with the template; when the PPG signal segment to be classified is matched with the third template, the PPG signal segment to be classified divided by the adjacent segment combination segment division method is used to assess the similarity with the template. When the multivariate signal quality index is greater than 0.75, the PPG signal segment to be classified is initially considered to be the heartbeat type or heartbeat type combination contained in the corresponding template. Combined with the characteristics that the pulse amplitude corresponding to the PAC2 / PVC2 premature beat is lower than the pulse amplitude of its previous sinus rhythm and the time interval between the peak point of the previous sinus rhythm and the peak point of the previous heartbeat of the sinus rhythm is less than the average heartbeat time interval, the PPG signal segment to be classified is finally classified into one of the heartbeat types or heartbeat type combinations contained in the three PPG waveform templates.

[0080] The specific algorithm flow in step g is as follows: Obtain heartbeat information based on the PPG signal segment classification results. The heartbeat information includes the number of premature beats and the total number of heartbeats. Specifically, the number of PPG signal segments to be classified into the second template and the third template is counted as the number of premature beats. The total number of heartbeats is the sum of the number of PPG signal segments to be classified into the first template and twice the number of segments classified into the second and third templates.

[0081] The specific algorithm flow in step h is as follows: The percentage of premature beats (PVCs) in the total heart rate is obtained by dividing the number of PVCs by the total heart rate. This is calculated by dividing the number of PVCs obtained in step g by the total heart rate obtained in step g. The percentage of PVCs in the total heart rate is an important indicator for medical monitoring of premature beats in patients. (See also...) Figure 5 This is a schematic diagram illustrating the classification results of PPG signal segments obtained from the example signal in this invention. Circles mark the starting points of PPG signal segments corresponding to the first template, stars mark the starting points of PPG signal segments corresponding to the second template, and triangles mark the starting points of PPG signal segments corresponding to the third template. According to... Figure 5 The classification results show that premature beats accounted for 34.6% of the total heartbeats during the data collection period for this subject.

[0082] Please see Figure 6 This is a schematic diagram of the device for detecting premature beats based on PPG waveform region features in this invention. The device includes:

[0083] The preprocessing module 601 is used to preprocess the acquired raw PPG signal;

[0084] The first processing module 602 is used to further process the preprocessed raw PPG signal using an improved adaptive low-complexity peak and valley detection algorithm to obtain the average heartbeat time interval and PPG pulse segments.

[0085] The signal segment classification module 603 is used to establish three PPG waveform templates based on the waveform region characteristics and average heartbeat time interval of PPG pulse segments; to divide the preprocessed original PPG signal based on the PPG pulse segments to obtain PPG signal segments to be classified; to classify the PPG signal segments to be classified based on multivariate PPG signal quality indicators and templates to obtain PPG signal segment classification results; and to obtain heartbeat information based on the PPG signal segment classification results, which includes the number of premature beats and the total number of heartbeats.

[0086] The final data acquisition module 604 is used to obtain the percentage of premature beats in the total number of heartbeats by dividing the number of premature beats by the total number of heartbeats.

[0087] In this embodiment, the relevant content of the preprocessing module 601, the first processing module 602, the signal segment classification module 603, and the final data acquisition module 604 can be found in the following references. Figures 1 to 5 The contents of the illustrated embodiments will not be repeated here.

[0088] In this application embodiment, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the processor performs any one of the methods described in the above method embodiments.

[0089] In one embodiment, an apparatus is provided, including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform any of the methods described in the above method embodiments.

[0090] Figure 7 An internal structural diagram of a device in one embodiment is shown. This computer device can specifically be a terminal, a server, or a gateway. Figure 7 As 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 causes the processor to perform the steps in the above-described method embodiments. The internal memory may also store a computer program, which, when executed by the processor, causes the processor to perform the steps in the above-described method embodiments. Those skilled in the art will understand that... Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0091] Those skilled in the art will understand that all or part of the processes in 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 described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application 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), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0092] 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.

[0093] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for detecting premature beats based on PPG waveform region features, characterized in that, The method includes: S1. Preprocess the acquired raw PPG signal; S2. The preprocessed raw PPG signal is further processed using an improved adaptive low-complexity peak and valley detection algorithm to obtain the average heartbeat time interval and PPG pulse segments. S3. Based on the waveform region characteristics of the PPG pulse segment and the average heartbeat time interval, three PPG waveform templates are established; the preprocessed original PPG signal is divided based on the PPG pulse segment to obtain PPG signal segments to be classified; the PPG signal segments to be classified are classified based on multivariate PPG signal quality indicators and the templates to obtain PPG signal segment classification results; heartbeat information is obtained according to the PPG signal segment classification results, the heartbeat information including the number of premature beats and the total number of heartbeats; S4. Divide the number of premature beats by the total number of heartbeats to obtain the percentage of premature beats in the total heartbeats.

2. The method according to claim 1, characterized in that, The preprocessing of the acquired raw PPG signal specifically includes: Acquire raw PPG signals; The PPG signal is filtered using a Butterworth bandpass filter, and the filtered PPG signal is then normalized in amplitude.

3. The method according to claim 1, characterized in that, The improved adaptive low-complexity peak and valley detection algorithm is used to further process the preprocessed raw PPG signal to obtain the average heartbeat time interval and PPG pulse segments, specifically including: An improved adaptive low-complexity peak and valley detection algorithm is used to locate the peak and valley values ​​of the preprocessed raw PPG signal. The average heartbeat time interval is obtained through the peak location, and the PPG pulse segments are obtained by segmenting the preprocessed raw PPG signal through the valley location.

4. The method according to claim 1, characterized in that, The method establishes three PPG waveform templates based on the waveform region features of the PPG pulse segment and the average heartbeat time interval; the method divides the preprocessed original PPG signal based on the PPG pulse segment to obtain the PPG signal segment to be classified. The PPG signal segments to be classified are classified based on the multivariate PPG signal quality index and the template to obtain the PPG signal segment classification results. Heartbeat information is obtained based on the PPG signal segment classification results. This heartbeat information includes the number of premature beats and the total number of heartbeats, specifically: Three PPG waveform templates are established based on the waveform region characteristics of the PPG pulse segment and the average heartbeat time interval; Based on the PPG pulse segments, the preprocessed original PPG signal is divided using two division methods: independent segments and combinations of adjacent segments, to obtain PPG signal segments to be classified. The template and the PPG signal segment to be classified are assessed for similarity based on a multivariate PPG signal quality index to obtain the assessment result; the PPG signal segment to be classified is then classified based on the assessment result to obtain the PPG signal segment classification result. Heartbeat information is obtained based on the PPG signal segment classification results, and the heartbeat information includes the number of premature beats and the total number of heartbeats.

5. The method according to claim 1, characterized in that, The waveform region characteristics of the PPG pulse segment include characteristic values; The establishment of three PPG waveform templates based on the waveform region characteristics of the PPG pulse segment and the average heartbeat time interval specifically includes: Three PPG waveform templates are established based on the aforementioned characteristic values ​​and the average heartbeat time interval; wherein, the formula for calculating the characteristic values ​​is: in, wherein K is the characteristic value, y(x) is the PPG pulse segment amplitude, x is the pulse time, y min is the minimum amplitude of the PPG pulse segment, y max is the maximum amplitude of the PPG pulse segment, y m is the average amplitude of the PPG pulse segment, x is the duration of the pulse segment.

6. The method according to claim 5, characterized in that, The three PPG waveform templates are: a PPG waveform template corresponding to a sinus rhythm; a PPG waveform template consisting of a first type premature heartbeat and the sinus rhythm preceding the premature heartbeat; and a PPG waveform template consisting of a second type premature heartbeat and the sinus rhythm preceding the premature heartbeat.

7. The method according to claim 6, characterized in that, The first type of premature beat is a premature beat with no obvious pulse fluctuation in the PPG waveform; the second type of premature beat is a premature beat with obvious pulse fluctuation in the PPG waveform.

8. A device for detecting premature beats based on PPG waveform region features, characterized in that, The device includes: The preprocessing module is used to preprocess the acquired raw PPG signal; The first processing module is used to further process the preprocessed raw PPG signal using an improved adaptive low-complexity peak and valley detection algorithm to obtain the average heartbeat time interval and PPG pulse segments. The signal segment classification module is used to establish three PPG waveform templates based on the waveform region characteristics of the PPG pulse segments and the average heartbeat time interval; to divide the preprocessed original PPG signal based on the PPG pulse segments to obtain PPG signal segments to be classified; to classify the PPG signal segments to be classified based on multivariate PPG signal quality indicators and the templates to obtain PPG signal segment classification results; and to obtain heartbeat information based on the PPG signal segment classification results, wherein the heartbeat information includes the number of premature beats and the total number of heartbeats. The final data acquisition module is used to divide the number of premature beats by the total number of heartbeats to obtain the percentage of premature beats in the total heartbeats.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it causes the processor to perform the steps of the method as described in any one of claims 1 to 7.

10. A computer device, comprising a memory and a processor, characterized in that, The memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of the method as described in any one of claims 1 to 7.

Images