Heartbeat interval estimation method, device, equipment, medium and program product

Abstract

This disclosure provides a method, apparatus, device, medium, and program product for estimating heart rate intervals, which can be applied in the field of vital sign monitoring. The heart rate interval estimation method includes: acquiring a heartbeat signal; identifying a steady-state segment signal in the heartbeat signal; extracting the time interval between adjacent heartbeats in the steady-state segment signal to obtain a steady-state heart rate interval sequence; and reconstructing the steady-state heart rate interval sequence into a complete heart rate interval sequence using a pre-trained machine learning model. This disclosure can improve the accuracy of heart rate interval estimation.

Description

Technical Field

[0001] This disclosure relates to the field of vital sign monitoring, and more specifically, to a method, apparatus, device, medium, and program product for estimating heart rate intervals. Background Technology

[0002] Long-term, continuous monitoring of vital signs is crucial for health risk assessment. Currently, various types of sensors are available for vital sign monitoring, such as contact sensors and sensing technologies based on vision, acoustics, and radio frequency (RF). However, in practical applications, many sensors are not suitable for prolonged use. An ideal monitoring method should be comfortable, interference-free, adaptable to environmental changes, and able to operate stably and continuously while protecting user privacy. In comparison, RF-based vital sign monitoring methods have significant advantages in these aspects, demonstrating broad application prospects.

[0003] Nevertheless, radio frequency (RF)-based methods still face significant challenges, the most significant being motion interference. In daily life, user actions such as posture adjustments, drinking water, and eating introduce motion artifacts, interfering with heart rate information in RF signals. Currently, there are two main solutions to motion artifacts: one attempts to reconstruct the heart rate signal from motion-affected data segments, but direct reconstruction of the heart rate signal has limited effectiveness and struggles to accurately restore the true heart rhythm; the other method simply discards the interfered data segments. While this avoids artifact interference, it results in significant data loss, causing the monitoring results to deviate from the actual physiological state. Summary of the Invention

[0004] In view of the above problems, this disclosure provides methods, apparatus, devices, media and program products for estimating heart rate intervals that can improve the accuracy of heart rate interval estimation.

[0005] According to a first aspect of this disclosure, a method for estimating heart rate intervals is provided, comprising:

[0006] Obtain heartbeat signals;

[0007] Identify steady-state segments in heartbeat signals;

[0008] The time intervals between adjacent heartbeats in the steady-state segment signal are extracted to obtain the steady-state heartbeat interval sequence;

[0009] A pre-trained machine learning model is used to reconstruct a complete heartbeat interval sequence from a steady-state heartbeat interval sequence.

[0010] According to embodiments of this disclosure, identifying steady-state segments in a heartbeat signal includes:

[0011] Obtain the frequency domain signal of the heartbeat signal;

[0012] The motion index is obtained by accumulating the frequency domain signal along the frequency dimension.

[0013] The heartbeat signal corresponding to the concentrated interval of the motion index distribution is taken as the steady-state segment signal.

[0014] According to embodiments of this disclosure, the method further includes:

[0015] The portion of the heartbeat signal that differs from the steady-state segment is taken as the motion segment signal;

[0016] Generate a mask for the signal in the motion segment.

[0017] According to embodiments of this disclosure, obtaining the frequency domain signal of the heartbeat signal includes:

[0018] A short-time Fourier transform is performed on the heartbeat signal to obtain the heartbeat Doppler spectrum, which serves as the frequency domain signal of the heartbeat signal.

[0019] According to embodiments of this disclosure, extracting the time interval between adjacent heartbeats in the steady-state segment signal includes:

[0020] The steady-state segment signal is input into a pre-trained neural network model based on a self-attention mechanism to obtain the identified heartbeat segments;

[0021] By splicing adjacent identified heartbeat segments, a combined heartbeat segment is obtained;

[0022] The heartbeat segments are translated and combined with the highest similarity as the objective.

[0023] Calculate the time deviation corresponding to the translation amount of adjacent heartbeat segments to obtain the time interval between adjacent heartbeats.

[0024] According to embodiments of this disclosure, inputting a steady-state segment signal into a pre-trained neural network model based on a self-attention mechanism includes:

[0025] The steady-state signal is divided into multiple overlapping segments of a set length; the signals of adjacent overlapping segments partially overlap.

[0026] Overlapping segments are input into a pre-trained neural network model based on a self-attention mechanism.

[0027] According to embodiments of this disclosure, translating combined heartbeat segments with the goal of maximizing similarity includes: translating adjacent combined heartbeat segments and calculating the cosine similarity between adjacent combined heartbeat segments after translation; when the cosine similarity is maximized, the translation operation is completed.

[0028] According to embodiments of this disclosure, the machine learning model is trained based on a momentum learning contrast mechanism, and the training samples of the machine learning model are obtained by randomly adding a mask to a complete intercardiac interval sequence.

[0029] Another aspect of this disclosure provides a heartbeat interval estimation device, the device comprising:

[0030] Heartbeat signal acquisition module, used to acquire heartbeat signals;

[0031] The steady-state signal recognition module is used to identify steady-state segments in heartbeat signals.

[0032] The steady-state heartbeat interval sequence acquisition module is used to extract the time interval between adjacent heartbeats in the steady-state segment signal to obtain the steady-state heartbeat interval sequence; and

[0033] The reconstruction module is used to reconstruct a complete heartbeat interval sequence from a steady-state heartbeat interval sequence using a pre-trained machine learning model.

[0034] Another aspect of this disclosure provides an electronic device, including: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the method described above.

[0035] Another aspect of this disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the processor to perform the method described above.

[0036] Another aspect of this disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the method described above.

[0037] The above-described one or more embodiments have the following beneficial effects: This disclosure provides a robust, non-contact, sustainable, and real-world applicable long-term cardiac activity monitoring solution. By identifying heartbeat information in steady-state intervals, extracting its inter-beat interval (IBI) features, and using this information to accurately reconstruct the missing IBI during motion interference, accurate estimation of the heartbeat interval is achieved, effectively avoiding the influence of motion artifacts on heartbeat information in non-steady-state states. Compared with existing signal processing or deep learning methods based on peak detection and filtering, this disclosure exhibits higher accuracy and robustness in both steady-state and motion (non-steady-state) intervals. Furthermore, this disclosure demonstrates high stability and generalization ability under different individual, environmental, time span, and radar deployment conditions, showcasing strong application potential in long-term, non-contact cardiac activity monitoring tasks and promoting the practical application of radio frequency sensing technology in health monitoring, smart healthcare, and other fields. Attached Figure Description

[0038] The foregoing contents, as well as other objects, features, and advantages of this disclosure, will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0039] Figure 1 A flowchart illustrating a heartbeat interval estimation method according to an embodiment of the present disclosure is shown schematically.

[0040] Figure 2 This schematically illustrates the overall framework of a heartbeat interval estimation method according to an embodiment of the present disclosure;

[0041] Figure 3 A flowchart illustrating the identification of steady-state segment signals according to embodiments of the present disclosure is shown schematically.

[0042] Figure 4 The schematic diagram illustrates a process for identifying steady-state segment signals according to an embodiment of the present disclosure;

[0043] Figure 5 A flowchart illustrating the process of extracting the time interval between adjacent heartbeats in a steady-state segment signal according to an embodiment of the present disclosure is shown.

[0044] Figure 6 The illustration shows a schematic diagram of the process for extracting the time interval between adjacent heartbeats in a steady-state segment signal according to an embodiment of the present disclosure;

[0045] Figure 7 The diagram illustrates a training process of a machine learning model according to an embodiment of the present disclosure.

[0046] Figure 8 The diagram illustrates the effect of the heart rate interval estimation method according to an embodiment of the present disclosure on the reconstruction of tachycardia and bradycardia signals, wherein (a) corresponds to the tachycardia signal and (b) corresponds to the bradycardia signal;

[0047] Figure 9 The illustration shows the effect of the heartbeat interval estimation method according to the embodiments of the present disclosure in reconstructing the IBI sequence in a steady-state scenario, where (a) corresponds to the true value of the IBI sequence and (b) corresponds to the estimated value of the IBI sequence.

[0048] Figure 10 The diagram illustrates the effect of the heart rate interval estimation method according to the embodiments of the present disclosure on reconstructing the IBI sequence in the presence of motion; wherein (a) corresponds to the true value of the IBI sequence, (b) corresponds to the estimated value of the IBI sequence in the steady state interval, and (c) corresponds to the reconstructed value of the complete IBI sequence;

[0049] Figure 11 A schematic block diagram of a heartbeat interval estimation device according to an embodiment of the present disclosure is shown.

[0050] Figure 12 A block diagram schematically illustrates an electronic device suitable for implementing a heartbeat interval estimation method according to an embodiment of the present disclosure. Detailed Implementation

[0051] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0052] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0053] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0054] When expressions such as "at least one of A, B, and C, etc." are used, they should generally be interpreted in accordance with the meaning commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0055] Figure 1 A flowchart illustrating a heartbeat interval estimation method according to an embodiment of the present disclosure is shown schematically.

[0056] like Figure 1 As shown, this embodiment includes operations S110~S140:

[0057] Operate S110 to obtain heartbeat signals.

[0058] Operate S120 to identify the steady-state segment of the heartbeat signal.

[0059] Operation S130 extracts the time interval between adjacent heartbeats in the steady-state segment signal to obtain the steady-state heartbeat interval sequence.

[0060] Operation S140 uses a pre-trained machine learning model to reconstruct the steady-state intercardia sequence into a complete intercardia sequence.

[0061] In operation S110, exemplarily, the heartbeat signal Obtained through preprocessing of radar echo signals. See also Figure 2 The radar system transmits signals to the human body and receives echo signals. The received echo signals are then preprocessed to obtain a heartbeat signal that can be used to analyze the heartbeat.

[0062] For example, the radar equipment employs Frequency Modulated Continuous Wave Radar (FMCW) and combines it with Multi-Input Multiple Output (MIMO) array technology. A 12-channel virtual antenna array is constructed based on a two-dimensional MIMO array, and high-resolution spatial sensing is achieved by combining it with three-dimensional beamforming. The formulaic expression of three-dimensional beamforming is as follows:

[0063] ;

[0064] in, The output of three-dimensional beamforming represents the radar system's view of a specific location. The spatial response intensity, where r represents the distance. Indicates pitch angle, Indicates the direction angle. It is the distance guide vector. The Kronecker product of the transmit and receive steering vectors is used for spatial pointing enhancement. This is the intermediate frequency signal matrix.

[0065] Then along The last two dimensions are summed to obtain a distance profile. Then, the constant false alarm rate (CFAR) is used to detect the distance to the human target within the distance profile. Statistical analysis of the pitch and azimuth angles is then performed based on the autocorrelation function. The autocorrelation function helps identify periodic components in the signal, facilitating heartbeat signal analysis. The pitch and azimuth angles with the strongest periodicity are selected as the optimal observation angles for subsequent heartbeat signal extraction. A three-dimensional receptive field matching the human chest cavity structure is constructed centered on the determined optimal observation angle. Phase sequences are extracted from different angles to form a time series matrix. Finally, the time series matrix is ​​divided into segments according to a fixed time window. The signal angle with the smallest variance in each time segment is selected, which helps remove noise and irrelevant motion effects, resulting in a purer heartbeat signal. The selected signals are spliced ​​along the time dimension to form the final heartbeat signal used for analysis. .

[0066] In operation S120, the obtained heartbeat signal is divided into steady-state and motion intervals, and the steady-state segment signal and motion segment signal are separated. The steady-state heartbeat interval data is obtained based on the steady-state segment signal, avoiding the influence of interference signals in the motion segment on the heartbeat interval calculation.

[0067] For example, the steady-state and motion regions of the heartbeat signal can be divided by performing a frequency domain transformation on the signal and combining frequency domain analysis and statistical modeling, based on the spectral energy distribution and histogram strategy. Alternatively, the division of steady-state and motion regions can also be achieved using bandpass filtering, wavelet transform, and other methods. Figure 3 A flowchart illustrating the identification of steady-state segment signals according to embodiments of the present disclosure is shown schematically. Figure 4 This schematically illustrates a processing diagram for identifying steady-state segment signals according to an embodiment of the present disclosure. See also... Figure 3 , Figure 4 In this embodiment of the disclosure, identifying the steady-state segment signal in the heartbeat signal includes:

[0068] In operation S310, the frequency domain signal of the heartbeat signal is acquired. Exemplarily, in this embodiment of the disclosure, the frequency domain signal of the heartbeat signal is acquired by performing a short-time Fourier transform (STFT) on the heartbeat signal, thereby obtaining a heartbeat Doppler spectrum. The short-time Fourier transform can provide resolution in both time and frequency, making it more suitable for signal processing of heartbeat signals containing periodic components. Furthermore, in some other embodiments, wavelet transform or other methods can also be used to acquire the frequency domain signal.

[0069] In operating the S320, the motion index is obtained by accumulating the frequency domain signal along the frequency dimension. The motion index indicates the intensity of an individual's movement; a higher motion index indicates more intense movement. The motion index can clearly distinguish between the steady-state interval and the motion interval, allowing for better extraction of information from the steady-state interval and thus better reconstruction of the IBI information of the motion interval. Specifically, the motion index is obtained by accumulating the heart rate Doppler spectrum along the frequency dimension. The formula is expressed as:

[0070] ;

[0071] ;

[0072] in, This represents the short-time Fourier transform spectrum of the heartbeat signal. n in the text represents The frequency index in the table, where N represents the total number of frequency points in the short-time Fourier transform. This represents the short-time Fourier transform.

[0073] The advantage of using the motion index to distinguish between steady-state and motion regions is that slight body movements, posture adjustments, and even changes in breathing patterns can cause variations in the spectral characteristics of a Doppler spectrum. By accumulating along the frequency dimension, this method can sensitively capture these subtle perturbations that might not be effectively identified by other time-domain statistical methods or energy thresholding methods. This makes the "motion index" superior to other methods in distinguishing between "absolute stillness" and "slight perturbations."

[0074] In operation S330, the signal corresponding to the concentrated interval of the motion index distribution is taken as the steady-state segment signal. The motion index is used to measure the intensity of motion disturbance in a time segment. By analyzing the distribution of the motion index, the most stable heartbeat frequency band is determined. In this embodiment, a motion index histogram is obtained by statistically analyzing the motion index. The motion index histogram is a frequency statistics of a specific interval of the motion index, which can be used to more accurately distinguish between the steady-state interval and the motion interval. The interval where the main peak of the motion index histogram is located is taken as the most stable heartbeat frequency band, and the corresponding heartbeat signal is taken as the steady-state segment signal.

[0075] In some embodiments, after completing the steady-state segment signal identification, the method further includes:

[0076] In operation S340, the portion of the heartbeat signal that differs from the steady-state segment signal is taken as the motion segment signal, thereby dividing the heartbeat signal into steady-state segment signal and motion segment signal according to the motion index histogram.

[0077] In operation S350, a mask is generated for the moving segment signal. Based on the steady-state and moving segment signals divided in operation S240, the steady-state segment signal is marked as 1, and the moving segment signal is marked as 0, generating a mask for the moving segment signal. This ensures that subsequent processing only analyzes the steady-state signal segment, avoiding the influence of the moving segment signal. The mask generation process can be represented as follows:

[0078] ;

[0079] in, Represents a mask sequence. , Indicates the threshold boundary of the histogram. Indicates the threshold boundary on the histogram. The boundary index generator represents the peak of the histogram. This indicates the operation of generating a histogram.

[0080] After obtaining the steady-state segment signal through operation S130, it is necessary to further estimate the corresponding inter-beat interval (IBI) sequence of the steady-state segment signal. The IBI estimation process requires identifying each heartbeat segment in the steady-state segment signal and calculating the time interval between adjacent heartbeats, that is, the time difference between corresponding points in adjacent heartbeat segments.

[0081] Figure 5 A flowchart illustrating the process of extracting the time interval between adjacent heartbeats in a steady-state segment signal according to an embodiment of the present disclosure is shown. Figure 6 This illustration schematically depicts a process for extracting the time interval between adjacent heartbeats in a steady-state segment signal according to an embodiment of the present disclosure. See also... Figure 5 , Figure 6 This disclosure embodiment identifies heartbeat segments using a neural network model and calculates the time interval between two adjacent heartbeats using a correlation-based pattern matching method. In this disclosure embodiment, extracting the time interval between adjacent heartbeats in the steady-state signal includes:

[0082] In operation S510, the steady-state segment signal is input into a pre-trained neural network model based on a self-attention mechanism to obtain the identified heartbeat segments. The self-attention mechanism can capture global dependencies across the entire signal sequence, thus maintaining the ability to recognize heartbeat signals with slight changes or distortions, improving the algorithm's robustness to different users and physiological states. For example, in this embodiment, the neural network model uses a Transformer model based on a multi-head self-attention mechanism. In this mechanism, different attention heads can focus on different components of the signal, automatically separating interference from valid information during feature extraction, thus improving discrimination accuracy. In some embodiments, other models based on self-attention mechanisms, such as BERT, can also be used for heartbeat segment recognition.

[0083] In some embodiments, the steady-state segment signal is further divided into multiple overlapping segments of a predetermined length, and the resulting overlapping segments are input into a pre-trained neural network model based on a self-attention mechanism for heartbeat segment recognition. The signals between adjacent overlapping segments partially overlap, so that the neural network can fully extract features for recognition, and also facilitates subsequent splicing to obtain combined heartbeat segments.

[0084] The following explanation uses the Transformer model as an example. For the obtained steady-state signal... First, it is divided into overlapping segments according to a segment length of 400ms and a step size of 100ms. That is, there is a 300ms signal overlap between adjacent overlapping segments, where This indicates the number of overlapping segments. The segments are mapped to embedding vectors via a multilayer perceptron network, and learnable positional codes of the same dimension and size as the embedding vectors are added. These codes are then input as tokens to the Transformer encoder. The Transformer encoder employs a multi-head self-attention mechanism, which can capture global dependencies across the entire signal sequence. It retains the ability to recognize slightly changed or distorted heartbeat signals, improving the algorithm's robustness to different users and physiological states. Furthermore, different attention heads can focus on different components of the signal, a mechanism that automatically separates interference from valid information during feature extraction, improving discrimination accuracy.

[0085] In operation S520, adjacent identified heartbeat segments are spliced ​​together to obtain a set of combined heartbeat segments. ,in The number of combined heartbeat segments is indicated. In the embodiments of this disclosure, the length of each combined heartbeat segment is fixed at 600ms, and the maximum similarity offset is calculated for each combined heartbeat segment signal.

[0086] In operation S530, the combined heartbeat segments are translated with the goal of maximizing similarity. Time alignment and interval calculation are performed on the identified heartbeat segments. Specifically, adjacent detected heartbeat segments are concatenated into a new sequence set (combined heartbeat segments). Then, the maximum similarity offset is calculated for each combined heartbeat segment signal. The cosine similarity between adjacent combined heartbeat segments after translation is calculated. When the cosine similarity is maximized, the translation operation is completed, thereby maximizing the cosine similarity between the segments.

[0087] In operation S540, the time deviation corresponding to the translation amount of adjacent combined heartbeat segments is calculated to obtain the time interval between adjacent heartbeats. For ease of calculation, the time difference between the center points of adjacent combined heartbeat segments is used to calculate the time interval between adjacent heartbeats, for example. After calculating the offset corresponding to the maximum similarity index of the signal, the time difference between the center points of each adjacent heartbeat segment is finally calculated to obtain the time interval between two adjacent heartbeats. This achieves steady-state IBI extraction without peak values ​​or specific templates, significantly improving the accuracy and reliability of non-contact heart rate measurement, and is especially suitable for monitoring environments with frequent changes in physiological state and frequent exercise in real life.

[0088] In operation S140, missing intercardiac interval (IBI) sequences are reconstructed under the influence of motion interference to address the problem of discontinuities in steady-state IBI data caused by motion artifacts. The core idea is to learn the potential change patterns from the effective IBI sequences obtained from the steady-state interval, guide the model to learn the dynamic distribution law of IBI under the condition of mask information through a contrastive learning mechanism, and finally reconstruct the complete IBI sequence.

[0089] Specifically, this embodiment uses a machine learning model to learn the dynamic distribution pattern of the IBI sequence, uses an encoder-decoder structure to reconstruct the IBI sequence, and introduces a momentum learning comparison mechanism. By adding a momentum encoder to work in conjunction with the original main encoder, the stability of features is ensured through momentum updates.

[0090] Figure 7 The illustration shows a schematic diagram of the training process of a machine learning model according to an embodiment of the present disclosure, wherein the pre-trained machine learning model is trained based on a momentum learning contrast mechanism, and the training samples of the machine learning model are obtained by randomly adding a mask to a complete heartbeat interval sequence.

[0091] First, data from different signal modalities, such as ECG (Electrocardiogram), SCG (Seismocardiogram), and PPG (Photoplethysmogram), are uniformly processed into a standard IBI time series format. For each modal signal, the heartbeat interval (i.e., the time difference between two adjacent heartbeats) is calculated by detecting the R wave or other equivalent feature points, thereby constructing a cross-modal consistent IBI training sample and obtaining an IBI sequence library for model training and testing.

[0092] To enable the model to process mask data, this embodiment introduces a dynamic mask generation function. For IBI sequences Perform masking, where R is the real number field and L is the length of the IBI sequence. For any complete IBI sequence... Randomly generate two versions of mask views The generation process can be represented as:

[0093] ;

[0094] The mask employs a geometric distribution control mechanism to simulate the temporal structure of occlusion in a real-world scene, specifically controlling the average segment length and mask ratio for each mask. The average segment length varies between 10 and 40, and the mask ratio varies between 0.2 and 0.8. Furthermore, the mask is generated randomly during the training phase, while the mask information for the testing phase is obtained based on operations S310~S350 described earlier; the two sets are, in principle, disjoint.

[0095] To enhance the model's adaptability to different mask structures and improve its robust modeling ability for IBI temporal features, a momentum contrastive learning mechanism is introduced. This mechanism constructs two sets of encoder networks with identical structures but different parameter update methods: encoder... The parameters are updated via backpropagation, while the momentum encoder... It does not participate in backpropagation; its parameters are inherited from the main encoder via a moving average, with the following update rules:

[0096] ;

[0097] in, and These are the parameters for the encoder and the momentum encoder, respectively. The momentum coefficient. The encoder outputs the latent vector. , representing the temporal representation of the same IBI sequence under different masking conditions. The goal is to maximize the similarity between different masked views of the same original IBI in the embedding space, while maintaining distinction from masked views of other sequences.

[0098] Loss function for a sample with index i in the dataset Using the standard InfoNCE contrast loss form:

[0099] ;

[0100] in, Represents the normalized vector and Dot product similarity between them; It is a temperature parameter; For one of the masked views of the i-th sample After encoder The obtained latent vectors correspond to the positive samples as follows: , representing another masked view of the i-th sample. After momentum encoder The resulting latent vector; Represents the i-th potential vector The feature corresponding to the j-th negative sample; K represents the number of negative samples provided in each query.

[0101] Given that contrastive learning performs better when a large number of negative samples are introduced, this embodiment introduces a feature queue much larger than the batch size, and its size is controlled as an independent hyperparameter. During the IBI reconstruction stage, the momentum encoder is discarded. Only the main encoder is retained. The main encoder and decoder Together, they form a complete encoder-decoder architecture, and the reconstruction loss of the IBI sequence It can be represented as:

[0102] ;

[0103] in, This indicates the reconstructed IBI sequence result. The size of the dataset.

[0104] Finally, the loss function The definition is as follows:

[0105] ;

[0106] in, It is a hyperparameter used to balance reconstruction error and contrastive learning loss. For the dataset size, This means calculating and summing the contrastive loss for each potential vector.

[0107] After training the machine learning model, its reconstruction performance on IBI sequences was verified. The IBI reconstruction effect was measured by comparing the IBI estimation results with motion artifact signals to the ground truth results.

[0108] For example, the hardware platform in this embodiment consists of a millimeter-wave radar module and a data transmission module. The millimeter-wave radar module operates in the frequency range of 60–64 GHz, with a maximum bandwidth of 4 GHz, and is equipped with three transmitting antennas and four receiving antennas. To achieve efficient data transmission, an ESP32-based module is designed as an intermediate communication unit. This module is responsible for receiving data from the RF front-end module and transmitting it to the main control computer. In addition, to obtain a high-precision electrocardiogram (ECG) as a truth reference, an ECG acquisition board is synchronously connected to the system.

[0109] In the IBI estimation process, Adam was used for optimization, with a batch size of 1024 and a learning rate of 0.001.

[0110] In the IBI mask reconstruction process, the encoder contains five convolutional layers with feature channels of 64, 128, 256, 512, and 1024 respectively. Each layer consists of two sets (1D convolution plus ReLU activation function plus batch normalization layer). The decoder structure is symmetrical to the encoder, but uses deconvolution instead of ordinary convolution. Adam is also used for optimization, with a batch size of 256 and a learning rate of 0.03.

[0111] For the determination of cardiac arrhythmias in the presence of motion artifacts. Figure 8 The illustration shows the effect of the heart rate interval estimation method according to the embodiments of the present disclosure on the reconstruction of tachycardia and bradycardia signals, wherein (a) corresponds to the tachycardia signal and (b) corresponds to the bradycardia signal, and the data source is a public dataset.

[0112] First, the motion-affected IBI sequence is masked (as shown by the blue scatter dots in the figure, representing the "IBI masked sequence"). Then, the pre-trained machine learning model obtained in operation S140 of this embodiment, which employs a one-dimensional encoder-decoder structure and combines a contrastive learning mechanism, is used to reconstruct the masked segment. The reconstruction process is completed through the following steps:

[0113] (1) The original IBI sequence is time-masked to form training samples with missing regions;

[0114] (2) The encoder extracts the latent feature representation of the IBI mask sequence;

[0115] (3) The decoder reconstructs the original IBI curve and generates a complete estimation result;

[0116] (4) Calculate the average IBI value of the reconstructed sequence to determine the heart rhythm type.

[0117] See Figure 8 In (a), the average IBI of the original mask sequence is 603ms, which was incorrectly identified as "normal"; while the average IBI recovered by the mask reconstruction method in this embodiment is 589ms, which is closer to the true value (as shown by the black dashed line), and was successfully identified as "tachycardia". Similarly, see Figure 8 In (b), the average value of the original mask sequence is 998ms, which may lead to a misjudgment as "normal"; while the average IBI after reconstruction is 1001ms, which accurately reflects the true physiological state of "bradycardia".

[0118] As can be seen, this embodiment can not only effectively restore the overall fluctuation trend of the IBI curve, but also assist in the realization of heart rhythm classification and abnormality identification in medical applications, and has broad practical value in remote health monitoring and intelligent medical devices.

[0119] IBI estimation performance for steady-state scenarios such as typing Figure 9 The illustration shows the effect of the heartbeat interval estimation method according to embodiments of the present disclosure in reconstructing IBI sequences in a steady-state scenario. Figure 9 The paper presents a comparison between the heartbeat signal collected by millimeter-wave radar in a static state of a subject and the IBI sequence estimated by the method provided in this embodiment, and the true value. (a) is the true value of the IBI sequence, which is obtained by calculating the RR interval from the synchronously acquired electrocardiogram and is used as an evaluation reference standard. (b) is the IBI estimation result of the heartbeat interval estimation method of this embodiment on the heartbeat data of the same time period. The comparison shows that the IBI estimated value, i.e., the red line in (b), is highly consistent with the true value of IBI, i.e., the black dotted line in (a), in terms of trend change, oscillation amplitude and overall rhythm. This verifies the estimation accuracy and temporal continuity of the method of this embodiment in the steady state interval.

[0120] The estimation process includes the following steps:

[0121] (1) Signal preprocessing: First, the original radar signal is converted into a heartbeat signal;

[0122] (2) Heartbeat pattern recognition: The preprocessed heartbeat signal is divided into several overlapping segments and embedded into the feature space. Through the multi-head self-attention mechanism, the model automatically identifies the effective segments containing heartbeat activity.

[0123] (3) Heartbeat pattern matching and alignment: For segments identified as containing heartbeats, time alignment and similarity matching methods are used to derive the final IBI sequence by calculating the time difference between the centers of adjacent segments.

[0124] The method described in this embodiment differs from traditional peak detection-based approaches and is particularly suitable for scenarios where stable waveform markers are absent in radar signals. It exhibits good generalization ability and application value. Experiments show that the IBI estimation error of this method can be controlled within 10 ms in the steady-state region, which is sufficient to meet clinical and health monitoring needs.

[0125] The effectiveness of IBI estimation for large-scale body movements such as posture adjustment. Figure 10 The diagram illustrates the effect of the heart rate interval estimation method according to the embodiments of the present disclosure on reconstructing the IBI sequence in the presence of motion. In this diagram, (a) corresponds to the true value of the IBI sequence, which is the subject's actual IBI sequence, derived from the RR interval extracted from the electrocardiogram, and used as the evaluation benchmark. In (b), the red curve corresponds to the estimated value of the IBI sequence in the steady-state interval. In (c), the red curve corresponds to the reconstructed value of the complete IBI sequence, which is the complete IBI sequence obtained after masking and reconstructing the motion-affected area.

[0126] The estimation process includes the following complete steps: (1) signal preprocessing; (2) motion detection to generate a mask; (3) IBI estimation; (4) IBI mask reconstruction.

[0127] It is evident that even when a large number of IBI fragments are missing due to strenuous exercise such as adjusting posture, the method provided in this embodiment can still effectively reconstruct the missing heart rhythm, ensuring the continuity and accuracy of the IBI time sequence. This provides a stable and reliable data foundation for subsequent heart rhythm analysis, HRV calculation, and health monitoring. It has good robustness, wide adaptability, and practical value, and is suitable for heart rate variability assessment tasks in real-world scenarios such as remote health monitoring, home medical care, and sleep monitoring.

[0128] Based on the aforementioned heartbeat interval estimation method, this disclosure also provides a heartbeat interval estimation device. The following will be combined with... Figure 11 The device is described in detail.

[0129] Figure 11 A schematic block diagram of a heartbeat interval estimation device according to an embodiment of the present disclosure is shown.

[0130] like Figure 11 As shown, the heartbeat interval estimation device 1100 of this embodiment includes a heartbeat signal acquisition module 1110, a steady-state signal recognition module 1120, a steady-state heartbeat interval sequence acquisition module 1130, and a reconstruction module 1140.

[0131] The heartbeat signal acquisition module 1110 is used to acquire heartbeat signals. In one embodiment, the heartbeat signal acquisition module 1110 can be used to perform the operation S110 described above, which will not be repeated here.

[0132] The steady-state signal recognition module 1120 is used to identify steady-state segments in the heartbeat signal. In one embodiment, the steady-state signal recognition module 1120 can be used to perform the operation S120 described above, which will not be repeated here.

[0133] The steady-state heartbeat interval sequence acquisition module 1130 is used to extract the time interval between adjacent heartbeats in the steady-state segment signal to obtain the steady-state heartbeat interval sequence. In one embodiment, the steady-state heartbeat interval sequence acquisition module 1130 can be used to perform the operation S130 described above, which will not be repeated here.

[0134] The reconstruction module 1140 is used to reconstruct a complete heartbeat interval sequence from a steady-state heartbeat interval sequence using a pre-trained machine learning model. In one embodiment, the reconstruction module 1140 can be used to perform the operation S140 described above, which will not be repeated here.

[0135] In some embodiments, the steady-state signal recognition module 1120 is specifically used to: acquire the frequency domain signal of the heartbeat signal; accumulate the frequency domain signal in the frequency dimension to obtain a motion index histogram; and take the signal corresponding to the interval where the main peak of the motion index histogram is located as the steady-state segment signal.

[0136] In some embodiments, the steady-state signal identification module 1120 is further configured to: take the signal corresponding to the interval where the non-main peak is located in the motion index histogram as the motion segment signal; and generate a mask for the motion segment signal.

[0137] In some embodiments, the steady-state heartbeat interval sequence acquisition module 1130 is specifically used to: input the steady-state segment signal into a pre-trained neural network model based on a self-attention mechanism to obtain the identified heartbeat segments; splice adjacent identified heartbeat segments to obtain combined heartbeat segments; shift the combined heartbeat segments with the goal of maximizing similarity; calculate the time deviation corresponding to the shift amount of adjacent combined heartbeat segments to obtain the time interval between adjacent heartbeats.

[0138] In some embodiments, the steady-state cardiac interval sequence acquisition module 1130 is specifically used to: translate adjacent combined heartbeat segments and calculate the cosine similarity between adjacent combined heartbeat segments after translation; when the cosine similarity is the maximum, the translation operation is completed.

[0139] In some embodiments, in the reconstruction module 1140, the machine learning model is trained based on a momentum learning contrast mechanism, and the training samples of the machine learning model are obtained by randomly adding a mask to the complete intercardia sequence.

[0140] For any parts not mentioned in the apparatus section, please refer to the various embodiments of the above method for understanding. That is, the apparatus section includes modules for performing each step of any of the method embodiments described above. Furthermore, the implementation methods, technical problems solved, functions achieved, and technical effects of each module / unit / subunit in the apparatus section embodiments are the same as or similar to the implementation methods, technical problems solved, functions achieved, and technical effects of the corresponding steps in the method section embodiments, and will not be repeated here.

[0141] According to embodiments of this disclosure, any multiple modules among the heartbeat signal acquisition module 1110, steady-state signal recognition module 1120, steady-state heartbeat interval sequence acquisition module 1130, and reconstruction module 1140 can be combined into one module, or any one of these modules can be split into multiple modules. Alternatively, at least some of the functions of one or more of these modules can be combined with at least some of the functions of other modules and implemented in one module.

[0142] According to embodiments of this disclosure, at least one of the heartbeat signal acquisition module 1110, steady-state signal recognition module 1120, steady-state heartbeat interval sequence acquisition module 1130, and reconstruction module 1140 can be at least partially implemented as a hardware circuit, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable method of integrating or packaging the circuit, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three methods. Alternatively, at least one of the heartbeat signal acquisition module 1110, steady-state signal recognition module 1120, steady-state heartbeat interval sequence acquisition module 1130, and reconstruction module 1140 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is run.

[0143] Figure 12 A block diagram schematically illustrates an electronic device suitable for implementing a heartbeat interval estimation method according to an embodiment of the present disclosure.

[0144] like Figure 12As shown, an electronic device 1200 according to an embodiment of the present disclosure includes a processor 1201, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1202 or a program loaded from a storage portion 1208 into a random access memory (RAM) 1203. The processor 1201 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 1201 may also include onboard memory for caching purposes. The processor 1201 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0145] RAM 1203 stores various programs and data required for the operation of electronic device 1200. Processor 1201, ROM 1202, and RAM 1203 are interconnected via bus 1204. Processor 1201 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 1202 and / or RAM 1203. It should be noted that the programs may also be stored in one or more memories other than ROM 1202 and RAM 1203. Processor 1201 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in said one or more memories.

[0146] According to embodiments of this disclosure, the electronic device 1200 may further include an input / output (I / O) interface 1205, which is also connected to a bus 1204. The electronic device 1200 may also include one or more of the following components connected to the input / output (I / O) interface 1205: an input section 1206 including a keyboard, mouse, etc.; an output section 1207 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1208 including a hard disk, etc.; and a communication section 1209 including a network interface card such as a LAN card, modem, etc. The communication section 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to the input / output (I / O) interface 1205 as needed. A removable medium 1211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 1210 as needed so that computer programs read from it can be installed into the storage section 1208 as needed.

[0147] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.

[0148] According to embodiments of this disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this disclosure, the computer-readable storage medium may include ROM 1202 and / or RAM 1203 and / or one or more memories other than ROM 1202 and RAM 1203 described above.

[0149] The embodiments of the present disclosure also include a computer program product, which includes a computer program containing program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the method provided by the embodiments of the present disclosure.

[0150] When the computer program is executed by the processor 1201, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0151] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 1209, and / or installed from the removable medium 1211. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0152] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 1209, and / or installed from the removable medium 1211. When the computer program is executed by the processor 1201, it performs the functions defined in the system of this disclosure embodiment. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0153] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computer programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C" or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect via the Internet).

[0154] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0155] Those skilled in the art will appreciate that the features described in the various embodiments of the present disclosure may be combined and / or coupled in various ways, even if such combinations or couplings are not explicitly described in the present disclosure. In particular, the features described in the various embodiments of the present disclosure may be combined and / or coupled in various ways without departing from the spirit and teachings of the present disclosure. All such combinations and / or couplings fall within the scope of the present disclosure.

[0156] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. The scope of this disclosure is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A method for estimating heart rate intervals, characterized in that, The method includes: Obtain heartbeat signals; Identify the steady-state segment of the heartbeat signal; Extract the time interval between each adjacent heartbeat in the steady-state segment signal to obtain the steady-state heartbeat interval sequence; The steady-state intercardia sequence is reconstructed into a complete intercardia sequence using a pre-trained machine learning model. The identification of the steady-state segment signal in the heartbeat signal includes: The heartbeat signal is subjected to a short-time Fourier transform to obtain the heartbeat Doppler spectrum, which is used as the frequency domain signal of the heartbeat signal. The motion index is obtained by accumulating the frequency domain signal in the frequency dimension; the frequency of the motion index in a specific interval is statistically analyzed to obtain the motion index histogram. The signal corresponding to the interval where the main peak is located in the motion index histogram is taken as the steady-state segment signal.

2. The method according to claim 1, characterized in that, Also includes: The signal corresponding to the interval where the non-main peak is located in the motion index histogram is taken as the motion segment signal; A mask is generated for the signal in the motion segment.

3. The method according to claim 1, characterized in that, The extraction of the time interval between adjacent heartbeats in the steady-state segment signal includes: The steady-state segment signal is input into a pre-trained neural network model based on a self-attention mechanism to obtain the identified heartbeat segments; By splicing adjacent identified heartbeat segments, a combined heartbeat segment is obtained; The combined heartbeat segments are shifted with the goal of maximizing similarity. Calculate the time deviation corresponding to the translation amount of adjacent heartbeat segments to obtain the time interval between adjacent heartbeats.

4. The method according to claim 3, characterized in that, The step of inputting the steady-state segment signal into a pre-trained neural network model based on a self-attention mechanism includes: The steady-state signal segment is divided into multiple overlapping segments of a predetermined length; the signals of adjacent overlapping segments partially overlap. The overlapping segments are input into a pre-trained neural network model based on a self-attention mechanism.

5. The method according to claim 3, characterized in that, The step of translating the combined heartbeat segments with the goal of maximizing similarity includes: translating adjacent combined heartbeat segments and calculating the cosine similarity between adjacent combined heartbeat segments after translation; when the cosine similarity is maximized, the translation operation is completed.

6. The method according to claim 1, characterized in that, The machine learning model is trained based on a momentum learning contrast mechanism, and the training samples of the machine learning model are obtained by randomly adding a mask to a complete heartbeat interval sequence.

7. A heartbeat interval estimation device, characterized in that, The device includes: Heartbeat signal acquisition module, used to acquire heartbeat signals; A steady-state signal identification module is used to identify steady-state segment signals in the heartbeat signal; wherein, identifying steady-state segment signals in the heartbeat signal includes: performing a short-time Fourier transform on the heartbeat signal to obtain a heartbeat Doppler spectrum, which is used as the frequency domain signal of the heartbeat signal; accumulating the frequency domain signal in the frequency dimension to obtain a motion index; statistically analyzing the frequencies of a specific interval of the motion index to obtain a motion index histogram; and using the signal corresponding to the interval where the main peak of the motion index histogram is located as the steady-state segment signal; A steady-state cardiac interval sequence acquisition module is used to extract the time interval between adjacent heartbeats in the steady-state segment signal to obtain a steady-state cardiac interval sequence; and The reconstruction module is used to reconstruct the steady-state intercardia sequence into a complete intercardia sequence using a pre-trained machine learning model.

8. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images