Peak value positioning method and device of ballistocardiogram signal and computer equipment

Abstract

The invention relates to the technical field of signal processing, in particular to a ballistocardiogram signal peak positioning method and device, computer equipment and a storage medium. Ballistocardiogram signals of a plurality of channels are sequentially subjected to filtering downsampling, motion artifact detection, signal enhancement and signal synchronization; the problem of large morphological difference among multichannel ballistocardiogram signals is solved, multichannel information fusion and signal reconstruction are carried out through an end-to-end signal mapping model, and reconstructed ballistocardiogram signals are obtained. According to the method, deep feature representation of each channel signal and extraction of time sequence feature information are realized, noise introduction and multi-channel information fusion are effectively reduced, cardiac interval detection is performed based on the reconstructed ballistocardiogram signal, the cardiac interval is used for peak recognition, and the accuracy of peak recognition is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular, to a method, device, computer device, and storage medium for peak localization of ballistocardiogram signals. Background Art

[0002] With the continuous development of computer technology and wearable devices, the technology for daily monitoring of cardiac function that is intelligent, convenient, and comfortable has become an increasingly urgent need for people. In recent years, as a non-invasive cardiac monitoring technology, ballistocardiogram signals (BCG) have become a research hotspot among many scholars at home and abroad. The BCG signal contains rich vital sign information, and the human index information reflecting cardiac function is of great significance for the early diagnosis and treatment of cardiovascular diseases.

[0003] Most of the current technical solutions adopt the cardiac cycle detection method and use the detected cardiac cycle for peak localization. Currently, whether it is a heartbeat detection method based on traditional signal processing, machine learning, or deep learning, usually only the cardiac cycle is detected from a single-channel signal. However, when facing multi-channel BCG signals (BCG signals collected simultaneously by multiple sensors), only considering cardiac cycle detection from a single channel will face two problems: (1) the channel selection problem, which requires selecting the channel with the best signal quality from multiple channels for cardiac cycle detection; (2) the problem of loss of effective information. Usually, due to the non-stationary characteristics of the BCG signal, the signal waveforms between different channels may vary at different times, and it is easy to form a situation where the effective cardiac cycle information between signals is complementary. At this time, not using the information between multiple channels will lead to a decline in the performance of the algorithm for cardiac cycle detection, making it difficult to accurately localize the peaks of the BCG signal. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide a method, device, computer device, and storage medium for peak localization of ballistocardiogram signals. The ballistocardiogram signals of several channels are successively filtered and downsampled, motion artifact detected, signal enhanced, and signal synchronized, overcoming the problem of large morphological differences between multi-channel ballistocardiogram signals. Through an end-to-end signal mapping model, multi-channel information fusion and signal reconstruction are performed to obtain a reconstructed ballistocardiogram signal, realizing the extraction of deep feature representations and temporal feature information of signals from each channel, effectively reducing noise introduction and multi-channel information fusion. Based on the reconstructed ballistocardiogram signal, cardiac cycle detection is performed, and the cardiac cycle is used for peak identification, improving the accuracy of peak identification.

[0005] In a first aspect, an embodiment of the present application provides a method for peak localization of ballistocardiogram signals, including the following steps:

[0006] Obtain the initial ballistocardiogram signals of several channels of the user to be measured and a preset signal mapping model, where the signal mapping model includes a temporal convolutional network and a bidirectional long short-term memory network;

[0007] Perform filtering downsampling and motion artifact detection on the initial ballistocardiogram signals of several channels to obtain the ballistocardiogram signals after motion artifact detection of several channels; perform signal enhancement on the ballistocardiogram signals after motion artifact detection of several channels to obtain the ballistocardiogram signals after signal enhancement of several channels;

[0008] Perform signal synchronization on the ballistocardiogram signals after signal enhancement of several channels to obtain the ballistocardiogram signals of several channels after signal synchronization;

[0009] Input the ballistocardiogram signals of several channels after signal synchronization into the temporal convolutional network for signal mapping to obtain mapped ballistocardiogram signals; input the mapped ballistocardiogram signals into the bidirectional long short-term memory network for signal reconstruction to obtain reconstructed ballistocardiogram signals;

[0010] Perform cardiac cycle detection on the reconstructed ballistocardiogram signals to obtain the cardiac cycle; according to the cardiac cycle, perform peak localization on the reconstructed ballistocardiogram signals to obtain the peak localization result.

[0011] In a second aspect, an embodiment of the present application provides a peak localization device for ballistocardiogram signals, including:

[0012] A data acquisition module, configured to obtain the initial ballistocardiogram signals of several channels of the user to be measured and a preset signal mapping model, where the signal mapping model includes a temporal convolutional network and a bidirectional long short-term memory network;

[0013] A signal preprocessing module, configured to perform filtering downsampling and motion artifact detection on the initial ballistocardiogram signals of several channels to obtain the ballistocardiogram signals after motion artifact detection of several channels; perform signal enhancement on the ballistocardiogram signals after motion artifact detection of several channels to obtain the ballistocardiogram signals after signal enhancement of several channels;

[0014] A signal synchronization module, configured to perform signal synchronization on the ballistocardiogram signals after signal enhancement of several channels to obtain the ballistocardiogram signals of several channels after signal synchronization;

[0015] A signal reconstruction module is configured to input the ballistocardiogram signals of several channels after signal synchronization into the temporal convolutional network for signal mapping to obtain mapped ballistocardiogram signals; input the mapped ballistocardiogram signals into the bidirectional long short-term memory network for signal reconstruction to obtain reconstructed ballistocardiogram signals;

[0016] A signal peak positioning module is configured to perform cardiac cycle detection on the reconstructed ballistocardiogram signal to obtain a cardiac cycle; and perform peak positioning on the reconstructed ballistocardiogram signal according to the cardiac cycle to obtain a peak positioning result.

[0017] In a third aspect, an embodiment of the present application provides a computer device, including a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for peak positioning of a ballistocardiogram signal as described in the first aspect are implemented.

[0018] In a fourth aspect, an embodiment of the present application provides a storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the method for peak positioning of a ballistocardiogram signal as described in the first aspect are implemented.

[0019] In an embodiment of the present application, a method, device, computer device, and storage medium for peak positioning of a ballistocardiogram signal are provided. The ballistocardiogram signals of several channels are sequentially subjected to filtering and downsampling, motion artifact detection, signal enhancement, and signal synchronization, overcoming the problem of large morphological differences between multi-channel ballistocardiogram signals. Through an end-to-end signal mapping model, multi-channel information fusion and signal reconstruction are performed to obtain a reconstructed ballistocardiogram signal, realizing the extraction of deep feature representations and temporal feature information of signals of each channel, effectively reducing noise introduction and multi-channel information fusion. Cardiac cycle detection is performed based on the reconstructed ballistocardiogram signal, and peak recognition is performed using the cardiac cycle, improving the accuracy of peak recognition.

[0020] For better understanding and implementation, the present invention will be described in detail below with reference to the accompanying drawings. Description of the Drawings

[0021] Figure 1 It is a schematic flowchart of the method for peak positioning of a ballistocardiogram signal provided in the first embodiment of the present application;

[0022] Figure 2 It is a schematic flowchart of S2 in the method for peak positioning of a ballistocardiogram signal provided in the first embodiment of the present application;

[0023] Figure 3 It is a schematic flowchart of S3 in the method for peak positioning of a ballistocardiogram signal provided in the first embodiment of the present application;

[0024] Figure 4 It is a schematic flowchart of S4 in the method for peak positioning of a ballistocardiogram signal provided in the first embodiment of the present application;

[0025] Figure 5Schematic flowchart of S5 in the peak positioning method for ballistocardiogram signals provided in the first embodiment of the present application;

[0026] Figure 6 Schematic flowchart of S5 in the peak positioning method for ballistocardiogram signals provided in the second embodiment of the present application;

[0027] Figure 7 Schematic flowchart of S53 in the peak positioning method for ballistocardiogram signals provided in the third embodiment of the present application;

[0028] Figure 8 Schematic flowchart of S6 in the peak positioning method for ballistocardiogram signals provided in the fourth embodiment of the present application;

[0029] Figure 9 Schematic structural diagram of the peak positioning device for ballistocardiogram signals provided in the fifth embodiment of the present application;

[0030] Figure 10 Schematic structural diagram of the computer device provided in the sixth embodiment of the present application. Detailed implementation manners

[0031] Here, exemplary embodiments will be described in detail, and examples thereof are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0032] The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. The singular forms "a", "the" and "said" used in the present application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0033] It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" / "when" as used herein may be interpreted as "when" or "when" or "in response to determining".

[0034] Please refer to Figure 1 ,Figure 1 A schematic flowchart of the peak localization method for the ballistocardiogram signal provided in the first embodiment of this application. The method includes the following steps:

[0035] S1: Obtain the initial ballistocardiogram signals of several channels of the user to be measured and a preset signal mapping model.

[0036] The execution subject of the peak localization method for the ballistocardiogram signal is the localization device for the peak localization method of the ballistocardiogram signal (hereinafter referred to as the localization device). In an optional embodiment, the localization device can be a computer device, a server, or a server cluster composed of multiple computer devices.

[0037] In this embodiment, the localization device can adopt an electroencephalogram interface device to detect the initial ballistocardiogram signals of several channels of the user to be measured through several sensors.

[0038] The localization device obtains a preset signal mapping model, where the signal mapping model includes a temporal convolutional network and a bidirectional long short-term memory network.

[0039] S2: Perform filtering downsampling and motion artifact detection on the initial ballistocardiogram signals of several channels to obtain the ballistocardiogram signals after motion artifact detection of several channels; perform signal enhancement on the ballistocardiogram signals after motion artifact detection of several channels to obtain the ballistocardiogram signals after signal enhancement of several channels.

[0040] In this embodiment, the localization device performs filtering downsampling and motion artifact detection on the initial ballistocardiogram signals of several channels to obtain the ballistocardiogram signals after motion artifact detection of several channels.

[0041] Specifically, since the initial ballistocardiogram signals are mixed with noise and respiratory signals, the localization device uses a Butterworth band-pass filter with a low-pass cut-off frequency set to 1 Hz, a high-pass cut-off frequency set to 20 Hz, and an order set to 4 to filter out the respiratory signals and power frequency noise of the initial ballistocardiogram signals of each channel. And the localization device performs normalization processing on the ballistocardiogram signals of several channels after Butterworth band-pass filtering to obtain the ballistocardiogram signals of several channels after normalization processing, as follows:

[0042]

[0043] In the formula, X[n] is the value of the nth sampling point in the ballistocardiogram signal after normalization processing, and x[n] is the value of the nth sampling point in the ballistocardiogram signal after Butterworth band-pass filtering. $\mu$ is the signal mean of the ballistocardiogram signal after Butterworth band-pass filtering, and $S$ is the signal standard deviation of the ballistocardiogram signal after Butterworth band-pass filtering.

[0044] The positioning device downsamples the ballistocardiogram signals of several channels after the normalization process to 100 Hz to ensure the retention of effective information and improve the calculation speed.

[0045] Motion artifacts are caused by the body movement or muscle activity of the subject during the monitoring process. For example, turning over, swinging of the arms, and large expansion and contraction of the chest cavity caused by deep breathing may all cause the occurrence of motion artifacts. When the motion artifacts are relatively severe, it will lead to the loss of effective information in the ballistocardiogram signal. The positioning device takes 3.5 times the 90th percentile of the absolute value signal of the ballistocardiogram signal after downsampling as the threshold, and through a preset first sliding window, the first sliding window can be set to 1 s, and traverses the ballistocardiogram signal after downsampling. If the absolute value of the sampling point in the ballistocardiogram signal after downsampling is higher than the threshold, it is determined that the sampling point is a motion artifact. If one sampling point in the first sliding window is determined to be a motion artifact, all sampling points in the entire first sliding window will be determined to be motion artifacts. Remove the sampling points determined to be motion artifacts to obtain the ballistocardiogram signals of several channels after motion artifact detection, reduce the negative impact brought by noise or artifacts, improve the signal-to-noise ratio of the ballistocardiogram signal, and thus more accurately extract the characteristic information in the ballistocardiogram signal.

[0046] The positioning device performs signal enhancement on the ballistocardiogram signals of several channels after the motion artifact detection to obtain the ballistocardiogram signals of several channels after signal enhancement.

[0047] Please refer to Figure 2 , Figure 2 which is a schematic flowchart of S2 in the peak positioning method of the ballistocardiogram signal provided by the first embodiment of the present application, including steps S21 to S23, specifically as follows:

[0048] S21: Divide the ballistocardiogram signal after the motion artifact detection process into several ballistocardiogram unit signals, perform J peak positioning on the several ballistocardiogram unit signals, and obtain several J peak positioning points of the several ballistocardiogram unit signals.

[0049] In this embodiment, the positioning device divides the ballistocardiogram signal after the motion artifact detection process into several ballistocardiogram unit signals, performs J peak positioning on the several ballistocardiogram unit signals, and obtains several J peak positioning points of the several ballistocardiogram unit signals.

[0050] Specifically, the positioning device detects all peak positioning points and valley positioning points in several ballistocardiogram unit signals through a preset second sliding window, and the time window can be set to 0 - 2 s. For each peak positioning point, calculate the difference between it and the two adjacent valley positioning points, and take the largest difference as the peak-to-valley distance of this peak positioning point. According to the peak-to-valley distances of several peak positioning points, take the peak positioning point with the largest peak-to-valley distance as the first J peak positioning point of this ballistocardiogram unit signal, denoted as J0. The calculation of the maximum peak-to-valley distance can refer to the following formula:

[0051] d = max(p i - v i , p i - v i+1 )

[0052] In the formula, d is the peak-to-valley distance, p i is the value of the i-th peak positioning point, v i , v i+1 are the values of the two adjacent valley positioning points of the i-th peak positioning point, and max(·) is the maximum value function.

[0053] The positioning device estimates the next J peak positioning point, denoted as J1, in the next second sliding window [J0 + 0.5 s, J0 + 2 s]. Repeat the above process until the entire ballistocardiogram unit signal is traversed to obtain the J peak positioning sequence [J0, J1, J2, … J M-1 of several ballistocardiogram unit signals, where the J peak positioning sequence includes several J peak positioning points, and M is the number of J peak positioning points.

[0054] S22: According to the preset template length, respectively, with several J peak positioning points of several ballistocardiogram unit signals as the centers, intercept signal segments from several ballistocardiogram unit signals to obtain several ballistocardiogram signal segments corresponding to several J peak positioning points of several ballistocardiogram unit signals; perform an average calculation on several ballistocardiogram signal segments corresponding to several J peak positioning points of the same ballistocardiogram unit signal to obtain the ballistocardiogram template signal of several ballistocardiogram unit signals.

[0055] In this embodiment, the positioning device respectively intercepts signal segments from several ballistocardiogram unit signals with several J peak positioning points of several ballistocardiogram unit signals as the centers according to the preset template length to obtain several ballistocardiogram signal segments corresponding to several J peak positioning points of several ballistocardiogram unit signals.

[0056] The positioning device performs an averaging calculation on the cardiogram signal segments corresponding to several J-peak positioning points of the same cardiogram unit signal to obtain a cardiogram template signal of several cardiogram unit signals.

[0057] S23: According to several cardiogram unit signals and the cardiogram template signal of the cardiogram unit signals, using the Pearson coefficient calculation method, perform signal enhancement on several cardiogram unit signals to obtain several cardiogram unit signals after signal enhancement, and combine several cardiogram unit signals after signal enhancement in the same channel to obtain a cardiogram signal after signal enhancement.

[0058] In this embodiment, the positioning device performs signal enhancement on several cardiogram unit signals according to several cardiogram unit signals and the cardiogram template signal of the cardiogram unit signals, using the Pearson coefficient calculation method, to obtain several cardiogram unit signals after signal enhancement.

[0059] Specifically, the positioning device obtains the Pearson correlation coefficients of several sampling points in several cardiogram unit signals according to several cardiogram unit signals, the cardiogram template signal of the cardiogram unit signals, and a preset Pearson coefficient calculation algorithm, and uses them as the values of several sampling points in several cardiogram unit signals after signal enhancement, where the Pearson coefficient calculation algorithm is:

[0060]

[0061] In the formula, is the value of the i-th sampling point in the cardiogram unit signal after signal enhancement, bcg k,i [l] is the value of the l-th sampling point in the signal segment intercepted with the i-th sampling point in the cardiogram unit signal as the center and L / 2 as the radius, and L is the length of the cardiogram template signal corresponding to the cardiogram unit signal. is the signal mean value of the signal segment intercepted with the i-th sampling point in the cardiogram unit signal as the center and L / 2 as the radius, T k [l] is the value of the l-th sampling point in the cardiogram template signal corresponding to the cardiogram unit signal. is the signal mean value of the cardiogram template signal corresponding to the cardiogram unit signal.

[0062] The positioning device combines several cardiogram unit signals after signal enhancement in the same channel to obtain a cardiogram signal after signal enhancement.

[0063] S3: Synchronize the enhanced ballistocardiogram signals of several channels to obtain the ballistocardiogram signals of several channels after signal synchronization.

[0064] In this embodiment, the positioning device synchronizes the enhanced ballistocardiogram signals of several channels to obtain the ballistocardiogram signals of several channels after signal synchronization, so as to overcome the problem of large morphological differences between multi-channel ballistocardiogram signals and improve the accuracy of multi-channel information fusion.

[0065] Please refer to Figure 3 , Figure 3 , which is a schematic flowchart of S3 in the peak positioning method of the ballistocardiogram signal provided in the first embodiment of this application, including steps S31 to S32, specifically as follows:

[0066] S31: Obtain the inter-beat stability index of the enhanced ballistocardiogram signals of several channels. According to the inter-beat stability index, divide the enhanced ballistocardiogram signals of several channels into a reference channel ballistocardiogram signal and several non-reference channel ballistocardiogram signals.

[0067] In this embodiment, the positioning device obtains the inter-beat stability index of the enhanced ballistocardiogram signals of several channels. Specifically, the positioning device respectively obtains several sampling points with values greater than 0.7 in the enhanced ballistocardiogram signals of several channels. Calculate the standard deviation and mean of the distances between the sampling points according to several sampling points greater than 0.7 in the enhanced ballistocardiogram signals of the same channel to obtain the standard deviation and mean of the distances of the enhanced ballistocardiogram signals of several channels. Take the ratio of the standard deviation and mean of the distances of the enhanced ballistocardiogram signals of the same channel to obtain the inter-beat stability index of the enhanced ballistocardiogram signals of several channels.

[0068] The positioning device takes the enhanced ballistocardiogram signal of the channel with the smallest inter-beat stability index as the reference channel ballistocardiogram signal according to the inter-beat stability index, and the others as non-reference channel ballistocardiogram signals, and divides the enhanced ballistocardiogram signals of several channels into a reference channel ballistocardiogram signal and several non-reference channel ballistocardiogram signals.

[0069] S32: According to a preset time window, obtain the Euclidean distances between several sampling points of the benchmark channel ballistocardiogram signal and several sampling points of several non-benchmark channel ballistocardiogram signals under the current time window. According to the Euclidean distances, obtain the relative delay between the benchmark channel ballistocardiogram signal and the several non-benchmark channel ballistocardiogram signals. According to the relative delay, synchronize the signals of the several non-benchmark channel ballistocardiogram signals to obtain the synchronized several non-benchmark channel ballistocardiogram signals.

[0070] In this embodiment, the positioning device obtains the Euclidean distances between several sampling points of the benchmark channel ballistocardiogram signal and several sampling points of several non-benchmark channel ballistocardiogram signals under the current time window according to a preset time window, such as the sampling point range [-45, 45]. According to the Euclidean distances, obtain the relative delay between the benchmark channel ballistocardiogram signal and the several non-benchmark channel ballistocardiogram signals. According to the relative delay, synchronize the signals of the several non-benchmark channel ballistocardiogram signals to obtain the synchronized several non-benchmark channel ballistocardiogram signals.

[0071] S4: Input the synchronized ballistocardiogram signals of the several channels into the temporal convolutional network for signal mapping to obtain the mapped ballistocardiogram signals; input the mapped ballistocardiogram signals into the bidirectional long short-term memory network for signal reconstruction to obtain the reconstructed ballistocardiogram signals.

[0072] In this embodiment, the positioning device inputs the synchronized ballistocardiogram signals of the several channels into the temporal convolutional network for signal mapping to obtain the mapped ballistocardiogram signals. Fusing the effective information between multiple channels can improve the accuracy of cardiac cycle detection to a certain extent.

[0073] Specifically, the temporal convolutional network adopts a TCN (Temporal Convolutional Network) network, and the temporal convolutional network is composed of several stacked residual connection modules (Temporal Block). In order to enable the network to learn richer deep feature representations and long-term temporal convolutional features, the number of channels of each residual connection module (initially, the channels_out of the residual connection module = 8, that is, twice the number of channels of the ballistocardiogram signal) and the dilation factor (initially set d = 4) are doubled in turn. The residual connection module includes a temporal convolutional sub-module, a convolutional block attention sub-module, and a convolutional layer.

[0074] Please refer to Figure 4 , Figure 4It is a schematic flowchart of S4 in the peak positioning method of the ballistocardiogram signal provided by the first embodiment of the present application, including steps S41 to S42, which are specifically as follows:

[0075] S41: Take the ballistocardiogram signals of several channels after signal synchronization as the input data of the first temporal convolutional module. According to the input data and the temporal convolutional sub-module, successively process through a causal dilated convolutional layer, weight normalization, a non-linear activation function, and a dropout layer, and repeat several times to obtain a first convolutional map; perform attention extraction according to the input signal and the convolutional attention sub-module to obtain an attention extraction map, perform convolutional processing on the attention extraction map through the convolutional layer to obtain a second convolutional map, and splice the first convolutional map and the second convolutional map to obtain a convolutional splicing map output by the first temporal convolutional module.

[0076] In this embodiment, the positioning device takes the ballistocardiogram signals of several channels after signal synchronization as the input data of the first temporal convolutional module. According to the input data and the temporal convolutional sub-module, successively process through a causal dilated convolutional layer, weight normalization, a non-linear activation function, and a dropout layer, and repeat several times to obtain a first convolutional map.

[0077] The positioning device performs attention extraction according to the input signal and the convolutional attention sub-module to obtain an attention extraction map, performs convolutional processing on the attention extraction map through the convolutional layer to obtain a second convolutional map, and splices the first convolutional map and the second convolutional map to obtain a convolutional splicing map output by the first temporal convolutional module. An attention mechanism is introduced on the residual branch to enhance the input features, enabling the model to adaptively focus on key information while weakening the influence of interference features.

[0078] S42: Take the convolutional splicing map output by the first temporal convolutional module as the input data of the next temporal convolutional module, repeat the data processing, and obtain the convolutional splicing map output by the last temporal convolutional module as the mapped ballistocardiogram signal.

[0079] In this embodiment, the positioning device takes the convolutional splicing map output by the first temporal convolutional module as the input data of the next temporal convolutional module, repeats the data processing, and obtains the convolutional splicing map output by the last temporal convolutional module as the mapped ballistocardiogram signal.

[0080] The positioning device inputs the mapped ballistocardiogram signal into the bidirectional long short-term memory network for signal reconstruction to obtain a reconstructed ballistocardiogram signal, where the bidirectional long short-term memory network is a BiLSTM (Bi-directional Long Short-Term Memory) network. The bidirectional long short-term memory network consists of a forward LSTM layer and a backward LSTM layer and is a special type of recurrent neural network that can selectively retain important feature information when processing long time series. Through the bidirectional mechanism, BiLSTM can consider the context information in the sequence, thereby enhancing the ability to represent temporal features.

[0081] S5: Perform cardiac cycle detection on the reconstructed ballistocardiogram signal to obtain the cardiac cycle; based on the cardiac cycle, perform peak localization on the reconstructed ballistocardiogram signal to obtain the peak localization result.

[0082] In this embodiment, the positioning device performs cardiac cycle detection on the reconstructed ballistocardiogram signal to obtain the cardiac cycle; based on the cardiac cycle, perform peak localization on the reconstructed ballistocardiogram signal to obtain the peak localization result.

[0083] The ballistocardiogram signals of several channels are successively subjected to filtering and downsampling, motion artifact detection, signal enhancement, and signal synchronization, overcoming the problem of large morphological differences between multi-channel ballistocardiogram signals. Through an end-to-end signal mapping model, multi-channel information fusion and signal reconstruction are performed to obtain a reconstructed ballistocardiogram signal, realizing the deep feature representation of signals in each channel and the extraction of temporal feature information, effectively reducing noise introduction and multi-channel information fusion. Cardiac cycle detection is performed based on the reconstructed ballistocardiogram signal, and peak recognition is performed using the cardiac cycle, improving the accuracy of peak recognition.

[0084] Please refer to Figure 5 , Figure 5 which is a schematic flowchart of S5 in the peak localization method for ballistocardiogram signals provided in the first embodiment of this application, including steps S51 to S52, specifically as follows:

[0085] S51: Use the Hilbert transform method to perform signal conversion on the reconstructed ballistocardiogram signal to obtain a Hilbert transform signal; calculate an envelope signal based on the reconstructed ballistocardiogram signal and the Hilbert transform signal.

[0086] In this embodiment, the positioning device uses the Hilbert transform method to perform signal conversion on the reconstructed ballistocardiogram signal to obtain a Hilbert transform signal.

[0087] Specifically, the positioning device performs a fast Fourier transform on the reconstructed ballistocardiogram signal to obtain an initial spectrum. The DC component and the Nyquist frequency of the initial spectrum are kept unchanged (when N is even), the negative frequency part is set to 0, and the positive frequency part is multiplied by 2 to compensate for the energy, and then the adjusted spectrum is obtained. An inverse fast Fourier transform is performed on the adjusted spectrum to obtain a Hilbert transform signal. Among them, the adjusted spectrum is:

[0088]

[0089] In the formula, is the value of the k-th sampling point in the adjusted spectrum.

[0090] The positioning device calculates an envelope signal according to the reconstructed ballistocardiogram signal, the Hilbert transform signal, and a preset envelope signal calculation algorithm, and obtains an envelope signal. Among them, the envelope signal calculation algorithm is:

[0091]

[0092] In the formula, H[n] is the value of the n-th sampling point in the envelope signal, z[n] is the value of the n-th sampling point in the reconstructed ballistocardiogram signal, is the value of the n-th sampling point in the Hilbert transform signal, and j is the imaginary unit.

[0093] S52: Perform a moving average filter on the envelope signal to obtain an envelope signal after the moving average filter; count the number of peaks of the envelope signal after the moving average filter, and calculate the cardiac interbeat interval according to the number of peaks of the envelope signal after the moving average filter and the signal length, and obtain the cardiac interbeat interval.

[0094] In this embodiment, the positioning device performs a moving average filter on the envelope signal according to a preset filtering algorithm to obtain an envelope signal after the moving average filter. Among them, the filtering algorithm is:

[0095]

[0096] In the formula, y[n] is the value of the n-th sampling point in the envelope signal after the moving average filter, W is a preset third sliding window, and H[n - w] is the value of the (n - w)-th sampling point in the envelope signal.

[0097] In this embodiment, the positioning device counts the number of peaks of the envelope signal after the moving average filter, and calculates the cardiac interbeat interval according to the number of peaks of the envelope signal after the moving average filter, the signal length, and a cardiac interbeat interval calculation algorithm, and obtains the cardiac interbeat interval. Among them, the cardiac interbeat interval calculation algorithm is:

[0098]

[0099] Wherein, interval is the cardiac cycle, M is the number of peaks of the envelope signal after moving average filtering, and P num is the signal length of the envelope signal after moving average filtering.

[0100] Please refer to Figure 6 , Figure 6 which is a schematic flowchart of S5 in the peak positioning method of the ballistocardiogram signal provided in the second embodiment of the present application, and further includes steps S53 to S54, specifically as follows:

[0101] S53: Obtain the initial peak positioning point of the reconstructed ballistocardiogram signal, and construct a cardiac cycle window interval according to the initial peak positioning point and the cardiac cycle; traverse the reconstructed ballistocardiogram signal according to the cardiac cycle window interval to obtain several peak positioning points of the reconstructed ballistocardiogram signal.

[0102] In this embodiment, the positioning device obtains the initial peak positioning point of the reconstructed ballistocardiogram signal. Specifically, the positioning device searches for a maximum peak within the interval of [0, 2s] of the reconstructed ballistocardiogram signal as the initial peak position, which serves as the initial peak positioning point.

[0103] The positioning device constructs a cardiac cycle window interval according to the initial peak positioning point and the cardiac cycle, wherein the cardiac cycle window interval is:

[0104]

[0105] Wherein, P0 is the initial peak positioning point.

[0106] The positioning device traverses the reconstructed ballistocardiogram signal according to the cardiac cycle window interval to obtain several peak positioning points of the reconstructed ballistocardiogram signal.

[0107] S54: Combine the initial peak positioning point and several peak positioning points to obtain a peak positioning result.

[0108] In this embodiment, the positioning device combines the initial peak positioning point and several peak positioning points to obtain a peak positioning result.

[0109] Please refer to Figure 7 , Figure 7 which is a schematic flowchart of S53 in the peak positioning method of the ballistocardiogram signal provided in the third embodiment of the present application, and further includes steps S531 to S532, specifically as follows:

[0110] S531: During the traversal of the reconstructed cardiogram signal within the cardiac cycle window interval, detect the number of peak positioning points obtained at the current moment. If the number of peak positioning points meets a preset number threshold, update the cardiac cycle according to a preset cardiac cycle update algorithm to obtain an updated cardiac cycle.

[0111] The cardiac cycle update algorithm is as follows:

[0112] interval′=mean(PPI i-c ,PPI i-x-1 ,...,PPI i )

[0113] In the formula, interval′ is the updated cardiac cycle, mean(·) is the mean function, c is the number threshold, PPI i is the interval between the i-th peak positioning point and the previous peak positioning point, and PPI i-c is the interval between the (i - c)-th peak positioning point and the previous peak positioning point.

[0114] In this embodiment, during the traversal of the reconstructed cardiogram signal within the cardiac cycle window interval by the positioning device, detect the number of peak positioning points obtained at the current moment. If the number of peak positioning points meets a preset number threshold, update the cardiac cycle according to a preset cardiac cycle update algorithm to obtain an updated cardiac cycle.

[0115] S532: Re - count the number of obtained peak positioning points, update the cardiac cycle window interval according to the initial peak positioning point and the updated cardiac cycle, and continue to traverse the reconstructed cardiogram signal with the updated cardiac cycle window interval until the last peak positioning point is obtained, thereby obtaining several peak positioning points of the reconstructed cardiogram signal.

[0116] In this embodiment, the positioning device re - counts the number of obtained peak positioning points, updates the cardiac cycle window interval according to the initial peak positioning point and the updated cardiac cycle, and continues to traverse the reconstructed cardiogram signal with the updated cardiac cycle window interval until the last peak positioning point is obtained, thereby obtaining several peak positioning points of the reconstructed cardiogram signal.

[0117] In an alternative embodiment, it further includes step S6: Train the signal mapping model. Please refer to Figure 8 , Figure 8 which is the flowchart of S6 in the peak positioning method of the cardiogram signal provided in the fourth embodiment of this application, including steps S61 - S63, specifically as follows:

[0118] S61: Obtain the ballistocardiogram signals of several channels after signal synchronization of several sample users and the labeled ballistocardiogram signals.

[0119] In this embodiment, the positioning device obtains the ballistocardiogram signals of several channels after signal synchronization of several sample users and the labeled ballistocardiogram signals.

[0120] S62: Input the ballistocardiogram signals of several channels after signal synchronization of several sample users into the signal mapping model for convolution processing and signal reconstruction to obtain the reconstructed ballistocardiogram signals of several sample users.

[0121] In this embodiment, the positioning device inputs the ballistocardiogram signals of several channels after signal synchronization of several sample users into the signal mapping model for convolution processing and signal reconstruction to obtain the reconstructed ballistocardiogram signals of several sample users.

[0122] S63: Obtain a loss value according to the reconstructed ballistocardiogram signals, labeled ballistocardiogram signals of several sample users, and the mean squared error loss function, and train the signal mapping model according to the loss value.

[0123] In this embodiment, the positioning device obtains a loss value according to the reconstructed ballistocardiogram signals, labeled ballistocardiogram signals of several sample users, and the mean squared error loss function, and trains the signal mapping model according to the loss value, where the mean squared error loss function is:

[0124]

[0125] In the formula, is the loss value, B is the number of sample users, y i is the reconstructed ballistocardiogram signal of the i-th sample user, is the labeled ballistocardiogram signal of the i-th sample user.

[0126] By calculating the mean square value of the differences between the predicted values and the true values, the prediction error of the model can be effectively measured. The Adam optimization algorithm is used to update the parameters in the signal mapping model. The initial learning rate η is set to 0.001, the maximum number of iterations is set to 64 times, the batch size for each iteration is 128, and the model with the minimum validation loss during the iteration process is retained during the model training process.

[0127] Please refer to Figure 9 , Figure 9 which is a schematic structural diagram of a ballistocardiogram signal peak positioning device provided by an embodiment of the present application. This device can implement all or part of the ballistocardiogram signal peak positioning device through software, hardware, or a combination of both. This device 9 includes:

[0128] A data acquisition module 91, configured to acquire initial ballistocardiogram signals of several channels of a user to be measured and a preset signal mapping model, wherein the signal mapping model includes a temporal convolutional network and a bidirectional long short-term memory network;

[0129] A signal preprocessing module 92, configured to perform filtering, downsampling, and motion artifact detection on the initial ballistocardiogram signals of several channels to obtain ballistocardiogram signals after motion artifact detection of several channels; perform signal enhancement on the ballistocardiogram signals after motion artifact detection of several channels to obtain ballistocardiogram signals after signal enhancement of several channels;

[0130] A signal synchronization module 93, configured to synchronize the ballistocardiogram signals after signal enhancement of several channels to obtain ballistocardiogram signals of several channels after signal synchronization;

[0131] A signal reconstruction module 94, configured to input the ballistocardiogram signals of several channels after signal synchronization into the temporal convolutional network for signal mapping to obtain mapped ballistocardiogram signals; input the mapped ballistocardiogram signals into the bidirectional long short-term memory network for signal reconstruction to obtain reconstructed ballistocardiogram signals;

[0132] A signal peak positioning module 95, configured to detect the cardiac interbeat interval of the reconstructed ballistocardiogram signals to obtain the cardiac interbeat interval; perform peak positioning on the reconstructed ballistocardiogram signals according to the cardiac interbeat interval to obtain a peak positioning result.

[0133] In an embodiment of the present application, through a data acquisition module, initial ballistocardiogram signals of a plurality of channels of a user to be measured and a preset signal mapping model are acquired, wherein the signal mapping model includes a temporal convolutional network and a bidirectional long short-term memory network; through a signal preprocessing module, the initial ballistocardiogram signals of the plurality of channels are filtered, downsampled, and motion artifact detection is performed to obtain ballistocardiogram signals after motion artifact detection of the plurality of channels; signal enhancement is performed on the ballistocardiogram signals after motion artifact detection of the plurality of channels to obtain ballistocardiogram signals after signal enhancement of the plurality of channels; through a signal synchronization module, signal synchronization is performed on the ballistocardiogram signals after signal enhancement of the plurality of channels to obtain ballistocardiogram signals of the plurality of channels after signal synchronization; through a signal reconstruction module, the ballistocardiogram signals of the plurality of channels after signal synchronization are input into the temporal convolutional network for signal mapping to obtain mapped ballistocardiogram signals; the mapped ballistocardiogram signals are input into the bidirectional long short-term memory network for signal reconstruction to obtain reconstructed ballistocardiogram signals; through a signal peak positioning module, cardiac interbeat interval detection is performed on the reconstructed ballistocardiogram signals to obtain a cardiac interbeat interval; according to the cardiac interbeat interval, peak positioning is performed on the reconstructed ballistocardiogram signals to obtain a peak positioning result. By sequentially performing filtering, downsampling, motion artifact detection, signal enhancement, and signal synchronization on the ballistocardiogram signals of the plurality of channels, the problem of large morphological differences between multi-channel ballistocardiogram signals is overcome, and multi-channel information fusion and signal reconstruction are performed through an end-to-end signal mapping model to obtain reconstructed ballistocardiogram signals, realizing deep feature representation of signals of each channel and extraction of temporal feature information, effectively reducing noise introduction and multi-channel information fusion, performing cardiac interbeat interval detection based on the reconstructed ballistocardiogram signals, and using the cardiac interbeat interval for peak identification, thereby improving the accuracy of peak identification.

[0134] Please refer to Figure 10 , Figure 10 which is a schematic structural diagram of a computer device provided by an embodiment of the present application. The computer device 10 includes: a processor 101, a memory 102, and a computer program 103 stored on the memory 102 and executable on the processor 101; the computer device may store multiple instructions, and the instructions are suitable for being loaded and executed by the processor 101 to perform the method steps of the above Figures 1 to 8 shown embodiment. The specific execution process may refer to the specific description of the Figures 1 to 8 shown embodiment and will not be elaborated here.

[0135] Among them, the processor 101 may include one or more processing cores. The processor 101 uses various interfaces and lines to connect various parts within the server. By running or executing instructions, programs, code sets, or instruction sets stored in the memory 102, and by calling the data in the memory 102, it executes various functions of the peak positioning device 9 for the ballistocardiogram signal and processes data. Optionally, the processor 101 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), or programmable logic array (PLA). The processor 101 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, application programs, etc.; the GPU is responsible for rendering and drawing the content to be displayed on the touch display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 101 and may be implemented separately by a single chip.

[0136] Among them, the memory 102 may include a random access memory (RAM), and may also include a read-only memory (ROM). Optionally, the memory 102 includes a non-transitory computer-readable storage medium. The memory 102 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 102 may include a program storage area and a data storage area. Among them, the program storage area may store instructions for implementing the operating system, instructions for at least one function (such as touch instructions, etc.), instructions for implementing the above-mentioned method embodiments, etc.; the data storage area may store the data involved in the above-mentioned method embodiments. Optionally, the memory 102 may also be at least one storage device located far from the aforementioned processor 101.

[0137] The embodiment of the present application also provides a storage medium, which can store multiple instructions, and the instructions are suitable for being loaded and executed by a processor to perform the above Figures 1 to 8 method steps of the embodiment, and the specific execution process can be referred to Figures 1 to 8 the specific description of the embodiment, and details are not described herein again.

[0138] Those skilled in the art can clearly understand that, for the convenience and conciseness of description, only the above division of each functional unit and module is used as an example. In practical applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiments can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of this application. The specific working processes of the units and modules in the above system can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated here.

[0139] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0140] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0141] In the embodiments provided by the present invention, it should be understood that the disclosed device / terminal device and method can be implemented in other ways. For example, the device / terminal device embodiments described above are only illustrative. For example, the division of the module or unit is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of the device or unit can be in an electrical, mechanical or other form.

[0142] The unit described as a separate component may or may not be physically separated, and the component displayed as a unit may or may not be a physical unit, that is, it may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0143] In addition, in each embodiment of the present invention, each functional unit can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0144] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above-mentioned embodiment methods of the present invention, it can also be completed by a computer program instructing relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc.

[0145] The present invention is not limited to the above-mentioned embodiments. If various modifications or deformations of the present invention do not depart from the spirit and scope of the present invention, and if these modifications and deformations are within the scope of the claims of the present invention and equivalent technical scope, then the present invention also intends to include these modifications and deformations.

Claims

1. A method for peak localization of ballistocardiogram signals, characterized in that, Including the following steps: Obtain the initial ballistocardiogram signals of several channels of the user to be measured and a preset signal mapping model, where the signal mapping model includes a temporal convolutional network and a bidirectional long short-term memory network; Perform filtering, downsampling, and motion artifact detection on the initial ballistocardiogram signals of several channels to obtain the ballistocardiogram signals after motion artifact detection for several channels; perform signal enhancement on the ballistocardiogram signals after motion artifact detection for several channels to obtain the ballistocardiogram signals after signal enhancement for several channels; Perform signal synchronization on the ballistocardiogram signals after signal enhancement for several channels to obtain the ballistocardiogram signals of several channels after signal synchronization; Input the ballistocardiogram signals of several channels after signal synchronization into the temporal convolutional network for signal mapping to obtain mapped ballistocardiogram signals; input the mapped ballistocardiogram signals into the bidirectional long short-term memory network for signal reconstruction to obtain reconstructed ballistocardiogram signals; Perform cardiac interbeat interval detection on the reconstructed ballistocardiogram signals to obtain the cardiac interbeat interval; according to the cardiac interbeat interval, perform peak localization on the reconstructed ballistocardiogram signals to obtain the peak localization result.

2. The method for peak localization of ballistocardiogram signals according to claim 1, wherein The step of performing signal enhancement on the ballistocardiogram signals after motion artifact detection for several channels to obtain the ballistocardiogram signals after signal enhancement for several channels includes the steps: Divide the ballistocardiogram signals after motion artifact detection processing into several ballistocardiogram unit signals, perform J peak localization on the several ballistocardiogram unit signals to obtain several J peak localization points of the several ballistocardiogram unit signals; According to the preset template length, respectively use the several J peak localization points of the several ballistocardiogram unit signals as the centers, and intercept signal segments from the several ballistocardiogram unit signals to obtain several ballistocardiogram signal segments corresponding to the several J peak localization points of the several ballistocardiogram unit signals; perform average calculation on the several ballistocardiogram signal segments corresponding to the several J peak localization points of the same ballistocardiogram unit signal to obtain the ballistocardiogram template signals of the several ballistocardiogram unit signals; According to the several ballistocardiogram unit signals and the ballistocardiogram template signals of the ballistocardiogram unit signals, use the Pearson coefficient calculation method to perform signal enhancement on the several ballistocardiogram unit signals to obtain the several ballistocardiogram unit signals after signal enhancement, and combine the several ballistocardiogram unit signals after signal enhancement of the same channel to obtain the ballistocardiogram signals after signal enhancement.

3. The method for peak localization of ballistocardiogram signals according to claim 1, characterized in that The step of performing signal synchronization on the ballistocardiogram signals after signal enhancement for several channels to obtain the ballistocardiogram signals of several channels after signal synchronization includes the steps: Obtain the interbeat stability index of the ballistocardiogram signals after signal enhancement for several channels, and according to the interbeat stability index, divide the ballistocardiogram signals after signal enhancement for several channels into reference channel ballistocardiogram signals and several non-reference channel ballistocardiogram signals; According to a preset time window, obtain the Euclidean distances between a number of sampling points of the fiducial channel ballistocardiogram signal and a number of sampling points of a number of non-fiducial channel ballistocardiogram signals under the current time window. According to the Euclidean distances, obtain the relative delay between the fiducial channel ballistocardiogram signal and the number of non-fiducial channel ballistocardiogram signals. According to the relative delay, perform signal synchronization on the ballistocardiogram signals of the number of non-fiducial channels to obtain the ballistocardiogram signals of the number of non-fiducial channels after signal synchronization.

4. The method for peak localization of ballistocardiogram signals according to claim 1, characterized in that The temporal convolutional network includes a number of stacked residual connection modules, and the residual connection module includes a temporal convolution sub-module, a convolutional block attention sub-module, and a convolutional layer; The step of inputting the ballistocardiogram signals of the number of channels after signal synchronization into the temporal convolutional network for signal mapping to obtain mapped ballistocardiogram signals includes: Taking the ballistocardiogram signals of the number of channels after signal synchronization as the input data of the first temporal convolution module. According to the input data and the temporal convolution sub-module, sequentially perform processing through a causal dilated convolutional layer, weight normalization, a non-linear activation function, and a dropout layer, and repeat a number of times to obtain a first convolutional map. Perform attention extraction according to the input signal and the convolutional attention sub-module to obtain an attention extraction map. Perform convolutional processing on the attention extraction map through the convolutional layer to obtain a second convolutional map. Perform splicing processing on the first convolutional map and the second convolutional map to obtain a convolutional splicing map output by the first temporal convolution module; Taking the convolutional splicing map output by the first temporal convolution module as the input data of the next temporal convolution module, repeat the data processing to obtain a convolutional splicing map output by the last temporal convolution module as the mapped ballistocardiogram signal.

5. The peak positioning method of the ballistocardiogram signal according to claim 1, characterized in that The step of performing cardiac interbeat interval detection on the reconstructed ballistocardiogram signal to obtain a cardiac interbeat interval includes: Adopt a Hilbert transform method to perform signal conversion on the reconstructed ballistocardiogram signal to obtain a Hilbert transform signal; perform envelope signal calculation according to the reconstructed ballistocardiogram signal and the Hilbert transform signal to obtain an envelope signal; Perform moving average filtering on the envelope signal to obtain an envelope signal after moving average filtering; count the number of peaks of the envelope signal after moving average filtering, and perform cardiac interbeat interval calculation according to the number of peaks of the envelope signal after moving average filtering and the signal length to obtain a cardiac interbeat interval.

6. The method for peak localization of the ballistocardiogram signal according to claim 1, wherein The step of performing peak localization on the reconstructed ballistocardiogram signal according to the cardiac interbeat interval to obtain a peak localization result includes: Obtain an initial peak localization point of the reconstructed ballistocardiogram signal, and construct a cardiac interbeat interval window interval according to the initial peak localization point and the cardiac interbeat interval; traverse the reconstructed ballistocardiogram signal according to the cardiac interbeat interval window interval to obtain a number of peak localization points of the reconstructed ballistocardiogram signal; Combine the initial peak localization point and the number of peak localization points to obtain a peak localization result.

7. The method for peak localization of the ballistocardiogram signal according to claim 6, wherein Traverse the reconstructed cardiogram signal according to the cardiac cycle window interval to obtain several peak positioning points of the reconstructed cardiogram signal During the traversal of the reconstructed cardiogram signal in the cardiac cycle window interval, detect the number of peak positioning points obtained at the current moment. If the number of peak positioning points meets a preset number threshold, update the cardiac cycle according to a preset cardiac cycle update algorithm to obtain an updated cardiac cycle. The cardiac cycle update algorithm is as follows: interval′ = mean(PPI i-c , PPI i-x-1 ,..., PPI i ) where interval′ is the updated cardiac cycle interval, mean(·) is the mean function, c is the number threshold, and PPI i is the interval between the i-th peak location point and the previous peak location point, and PPI i-c is the interval between the (i - c)-th peak location point and the previous peak location point; Recount the number of obtained peak positioning points, update the cardiac cycle window interval according to the initial peak positioning points and the updated cardiac cycle, and continue to traverse the reconstructed cardiogram signal with the updated cardiac cycle window interval until the last peak positioning point is obtained, and obtain several peak positioning points of the reconstructed cardiogram signal.

8. The method for peak localization of the ballistocardiogram signal according to claim 4, characterized in that, It further includes the step of training the signal mapping model; The training of the signal mapping model includes the steps of: Obtain the cardiogram signals of several channels and the labeled cardiogram signals after signal synchronization of several sample users; Input the cardiogram signals of several channels after signal synchronization of several sample users into the signal mapping model for convolution processing and signal reconstruction to obtain the reconstructed cardiogram signals of several sample users; According to the reconstructed cardiogram signals, labeled cardiogram signals of several sample users and the mean square error loss function, obtain a loss value, and train the signal mapping model according to the loss value.

9. A peak positioning device for ballistocardiogram signals, characterized in that, It includes: A data acquisition module for obtaining the initial cardiogram signals of several channels of a user to be measured and a preset signal mapping model, where the signal mapping model includes a temporal convolutional network and a bidirectional long short-term memory network; A signal preprocessing module for filtering, downsampling and motion artifact detection of the initial cardiogram signals of several channels to obtain the cardiogram signals after motion artifact detection of several channels; perform signal enhancement on the cardiogram signals after motion artifact detection of several channels to obtain the cardiogram signals after signal enhancement of several channels; A signal synchronization module for synchronizing the cardiogram signals of several channels after signal enhancement to obtain the cardiogram signals of several channels after signal synchronization; A signal reconstruction module for inputting the cardiogram signals of several channels after signal synchronization into the temporal convolutional network for signal mapping to obtain a mapped cardiogram signal; inputting the mapped cardiogram signal into the bidirectional long short-term memory network for signal reconstruction to obtain a reconstructed cardiogram signal; A signal peak positioning module for detecting the cardiac cycle of the reconstructed cardiogram signal to obtain a cardiac cycle; performing peak positioning on the reconstructed cardiogram signal according to the cardiac cycle to obtain a peak positioning result.

10. A computer device, characterized in that, Comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, the processor, when executing the computer program, implements the steps of the method for peak positioning of the ballistocardiogram signal according to any one of claims 1 to 8.

Images