Heart rate deceleration force prediction method, device and storage medium

Abstract

The present application provides a method, device and storage medium for predicting heart rate deceleration force. In this method, since the function expression of momentum update is determined according to the calculation formula of the RR interval corresponding to each preset time segment within the first set time length, when determining the second heart rate deceleration force corresponding to the preset time period according to the first heart rate deceleration force corresponding to the current preset time period and the function expression of momentum update, the first heart rate deceleration force corresponding to the preset time period can be integrated with the second heart rate deceleration force corresponding to the time period before the preset time period, so as to reflect the second heart rate deceleration force of the user at the current moment as a whole. In addition, since the PPG sensor is small in size, easy to carry, and cheap. Therefore, while facilitating daily use by users, it can conveniently and accurately predict the heart rate deceleration force over a long period of time.

Description

Technical Field

[0001] The present application relates to the field of terminal device technology, and in particular to a heart rate deceleration force prediction method, device and storage medium. Background Art

[0002] Deceleration Capacity of Rate (DC) is an indicator to measure the excitability of the vagus nerve. In clinical practice, DC is of great significance in predicting the risk of cardiovascular events such as arrhythmia and heart failure.

[0003] At present, the prediction of heart rate deceleration force needs to be determined based on electrocardiograph (ECG) signals over a longer period of time, such as 24 hours (h). However, the acquisition of ECG signals requires the use of professional equipment, such as an electrocardiograph. Since electrocardiographs are large in size, expensive, and complex to operate, they require users to have professional knowledge and therefore cannot meet the daily needs of users. Summary of the invention

[0004] In order to solve the above technical problems, the embodiments of the present application provide a heart rate deceleration force prediction method, device and storage medium, which aim to conveniently and accurately predict the heart rate deceleration force.

[0005] In the first aspect, the embodiment of the present application provides a method for predicting heart rate deceleration force. The method includes: receiving a continuous photoplethysmogram (PPG) signal collected by a first electronic device within a first set time period; after receiving each PPG signal of a preset time segment, determining the first heart rate deceleration force corresponding to the preset time segment according to the heart rate deceleration force regression model and the PPG signal of the preset time segment, the heart rate deceleration force regression model is based on physiological invariance, and is obtained by training a twin network with unshared parameters through unsupervised comparative learning; determining the second heart rate deceleration force corresponding to the preset time segment according to the function expression of momentum update and the first heart rate deceleration force corresponding to the preset time segment, the function expression of momentum update is determined according to the calculation formula of the RR interval corresponding to each preset time segment within the first set time period.

[0006] The first set duration may be a relatively long period of time, such as 24 hours.

[0007] The first electronic device is, for example, a wearable device integrated with a PPG sensor, such as a smart watch, a smart bracelet, a head-mounted device, etc.

[0008] In which, when the heart rate deceleration force monitoring function is turned on in the first electronic device and the user wears the first electronic device, the PPG sensor in the first electronic device can continuously collect PPG signals within a first set time period.

[0009] Accordingly, the first electronic device can transmit the collected PPG signal to the server in real time via a wireless network, such as a WiFi network.

[0010] The duration of the preset time segment is shorter than the first set duration, for example, 1 hour, or a longer time, or a shorter time.

[0011] Among them, the network architecture of the twin network with unshared parameters, such as Fig. 9 The network architecture of the Siamese network encoder where parameters are not shared.

[0012] Since the function expression of momentum update is determined according to the calculation formula of the RR interval corresponding to each preset time segment within the first set time length, when the second heart rate deceleration force corresponding to the preset time period is determined according to the first heart rate deceleration force corresponding to the current preset time period and the function expression of momentum update, the second heart rate deceleration force corresponding to the time period before the preset time period can be integrated with the first heart rate deceleration force corresponding to the preset time period, thereby reflecting the second heart rate deceleration force of the user at the current moment as a whole, and making the subsequent warning information made according to the second heart rate deceleration force more reasonable, reducing false warnings caused by interference from environmental factors, wearing methods, exercise status, etc.

[0013] In addition, since the PPG sensor is small in size, easy to carry, and cheap, based on the PPG signal, according to the above heart rate deceleration force prediction method, while being convenient for users to use in daily life, it can conveniently and accurately predict the heart rate deceleration force over a long period of time, thereby better monitoring the user's heart health.

[0014] According to the first aspect, training a heart rate deceleration force regression model includes: obtaining sample PPG signals and sample electrocardiogram (ECG) signals, wherein the sample PPG signals and the sample ECG signals are PPG signals and corresponding ECG signals collected by the same user in the same time period; dividing the sample PPG signals and the corresponding sample ECG signals into multiple small batch data sets; constructing positive sample pairs and negative sample pairs according to the sample PPG signals and sample ECG signals in the same small batch data set; based on physiological invariance, using an unsupervised comparative learning method and a twin network with non-shared parameters, comparatively learning positive sample pairs and negative sample pairs, and extracting physiological features that meet physiological invariance in the sample PPG signals and the sample ECG signals; defining a loss function used in the training process according to the physiological characteristics; constructing a heart rate deceleration force regression model according to the twin network with non-shared parameters, the loss function and the heart rate deceleration force algorithm.

[0015] Therefore, by combining the part of the PPG signal that meets the requirements and the heart rate deceleration force algorithm part into one model, a heart rate deceleration force regression model can be obtained.

[0016] For specific implementation details of this aspect, please refer to the content of the heart rate deceleration force regression model training stage in the following embodiment, which will not be repeated here.

[0017] According to the first aspect, or any implementation manner of the first aspect above, each mini-batch data set includes a first PPG signal, a second PPG signal, a first ECG signal corresponding to the first PPG signal, and a second ECG signal corresponding to the second PPG signal; constructing positive sample pairs and negative sample pairs according to the sample PPG signals and the sample ECG signals in the same mini-batch data set, including: constructing positive sample pairs according to the first PPG signal and the first ECG signal, the second PPG signal and the second ECG signal in the same mini-batch data set, respectively; constructing negative sample pairs according to the first PPG signal and the second PPG signal, the first PPG signal and the second ECG signal, and the second PPG signal and the first ECG signal in the same mini-batch data set.

[0018] For example, from the perspective of distance, the distance between the physiological features of the positive sample pair is relatively close, that is, the same physiological characteristics are maintained.

[0019] For example, from the perspective of distance, the distance between the physiological features of the negative sample pair is relatively far, that is, they belong to the changing physiological characteristics.

[0020] The relationship between positive sample pairs and negative sample pairs can be as follows Fig.10 shown.

[0021] According to the first aspect, or any implementation of the first aspect above, the second heart rate deceleration force corresponding to the preset time segment is determined according to the function expression of momentum update and the first heart rate deceleration force corresponding to the preset time segment, including: determining the second heart rate deceleration force corresponding to the preset time segment according to the second heart rate deceleration force corresponding to the first time segment and the first heart rate deceleration force corresponding to the preset time segment, the first time segment is a preset time segment before the preset time segment, and the duration of the first time segment is the same as the duration of the preset time segment.

[0022] Among them, the preset time segment, such as the one mentioned in the following embodiment time.

[0023] The first time segment, for example, is described in the following embodiment time.

[0024] Therefore, based on the above momentum update function expression, we can get The heart rate deceleration force at the moment, i.e., the second heart rate deceleration force in the following embodiment .because The second heart rate deceleration force of the moment, combined with The second heart rate deceleration force before the moment, so it can more accurately reflect the user's overall The second heart rate deceleration force at all times makes the warning information more reasonable and reduces false warnings caused by interference from environmental factors, wearing methods, exercise status, etc.

[0025] For specific implementation details of this aspect, please refer to the content of the heart rate deceleration force output strategy optimization stage in the following embodiment, which will not be repeated here.

[0026] According to the first aspect, or any implementation of the first aspect above, the method also includes: storing the second heart rate deceleration force corresponding to each preset time segment within a first set time length; after reaching the first set time length, sending all second heart rate deceleration forces stored within the first set time length to a second electronic device, the second electronic device being used to generate a warning message based on the received second heart rate deceleration force, and displaying the warning message and the received second heart rate deceleration force.

[0027] The second electronic device is, for example, a mobile phone, a tablet computer or other electronic device.

[0028] For example, in some implementations, an application or a dedicated heart rate monitoring application may be installed in the second electronic device. In this way, after the second electronic device establishes a communication connection with the first electronic device, such as a Bluetooth connection, the user can configure a wireless network for the first electronic device through the second electronic device, so that the second electronic device can transmit the collected PPG signal to the server through the wireless network and interact with the second electronic device.

[0029] The user interface of the warning information and the second heart rate deceleration force displayed in the second electronic device may be as shown in interface 10c.

[0030] Therefore, by storing the second heart rate deceleration force corresponding to each preset time segment within the first set time length, the server can batch-send all the second heart rate deceleration forces to the second electronic device after the first set time length is reached.

[0031] According to the first aspect, or any implementation of the first aspect above, the method also includes: for each preset time segment, after determining the second heart rate deceleration force corresponding to the preset time segment, sending the second heart rate deceleration force corresponding to the preset time segment to a second electronic device, the second electronic device is used to generate a warning message based on the received second heart rate deceleration force, and display the warning message and the received second heart rate deceleration force.

[0032] Therefore, when the server fails to obtain the second heart rate deceleration force of a preset time segment, it promptly sends the second heart rate deceleration force of the preset time segment to the second electronic device, so that the user can promptly check the second heart rate deceleration force corresponding to the current preset time period, and then adjust the work and rest schedule in time and go to see a doctor when there is an abnormality.

[0033] According to the first aspect, or any implementation of the first aspect above, the first electronic device is communicatively connected to the second electronic device, and the second electronic device is further configured to send the warning information and the corresponding second heart rate deceleration force to the first electronic device.

[0034] Therefore, the user can check the second heart rate deceleration force and the corresponding warning information in time through the first electronic device worn without unlocking the second electronic device.

[0035] According to the first aspect, or any implementation of the first aspect above, the method also includes: storing the second heart rate deceleration force corresponding to each preset time segment within a first set time length; after reaching the first set time length, sending all second heart rate deceleration forces stored within the first set time length to a first electronic device, the first electronic device being used to generate a warning message based on the received second heart rate deceleration force, and displaying the warning message and the received second heart rate deceleration force.

[0036] Therefore, in the absence of a second electronic device, the first electronic device can also timely display the second heart rate deceleration force and the corresponding warning information, so that the user can timely check the second heart rate deceleration force corresponding to the current preset time period, and then adjust the work and rest schedule in time and go to see a doctor when there is an abnormality.

[0037] According to the first aspect, or any implementation of the first aspect above, the method also includes: for each preset time segment, after determining the second heart rate deceleration force corresponding to the preset time segment, sending the second heart rate deceleration force corresponding to the preset time segment to a first electronic device, the first electronic device being used to generate a warning message based on the received second heart rate deceleration force, and displaying the warning message and the received second heart rate deceleration force.

[0038] Therefore, in the absence of a second electronic device, the first electronic device can also timely display the second heart rate deceleration force and the corresponding warning information, so that the user can timely check the second heart rate deceleration force corresponding to the current preset time period, and then adjust the work and rest schedule in time and go to see a doctor when there is an abnormality.

[0039] According to the first aspect, or any implementation of the first aspect above, the first electronic device is a wearable device integrating a PPG sensor.

[0040] For example, smart watches, smart bracelets, head-mounted devices, etc. that integrate PPG sensors.

[0041] Therefore, users can monitor their heart health by using wearable devices, which is convenient for daily use.

[0042] In a second aspect, an embodiment of the present application provides a terminal device. The terminal device includes: a memory and a processor, the memory and the processor are coupled; the memory stores program instructions, and when the program instructions are executed by the processor, the terminal device executes instructions of the method in the first aspect or any possible implementation of the first aspect.

[0043] Exemplarily, in some implementations, the terminal device may include a server device and a client device.

[0044] Among them, the server device can be understood as a server.

[0045] In some implementations, the server may be a cloud server.

[0046] In other implementations, the server may be a physical server.

[0047] The client device may include a first electronic device, which may be a wearable device, such as a smart watch or smart bracelet integrated with a PPG sensor.

[0048] The client device may also include a second electronic device, which may be a mobile phone, a tablet computer, or other device.

[0049] In a third aspect, an embodiment of the present application provides a heart rate deceleration force prediction system. The system includes: a server, a first electronic device, and a second electronic device; wherein the server, the first electronic device, and the second electronic device are used to execute instructions of the method in the first aspect or any possible implementation of the first aspect.

[0050] In a fourth aspect, an embodiment of the present application provides a computer-readable medium for storing a computer program, wherein the computer program includes instructions for executing the method in the first aspect or any possible implementation of the first aspect.

[0051] In a fifth aspect, an embodiment of the present application provides a computer program comprising instructions for executing the method in the first aspect or any possible implementation of the first aspect.

[0052] In a sixth aspect, an embodiment of the present application provides a chip, the chip comprising a processing circuit and a transceiver pin, wherein the transceiver pin and the processing circuit communicate with each other through an internal connection path, and the processing circuit executes the method in the first aspect or any possible implementation of the first aspect to control the receiving pin to receive a signal, and to control the sending pin to send a signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 Schematic diagram showing the relationship between the waveform of the PPG signal and the waveforms of the cardiac contraction, diastole, and ECG signal;

[0054] Figure 2 The figure is a schematic diagram showing the appearance and wearing of a wearable device with an integrated PPG sensor;

[0055] Figure 3 A schematic diagram of a communication system for exemplarily implementing the heart rate deceleration force prediction method provided in an embodiment of the present application;

[0056] Figure 4 is a schematic diagram of a hardware structure of a wearable device shown as an example;

[0057] Figure 5 The figure is a schematic diagram of a hardware structure of an electronic device shown as an example;

[0058] Figure 6 A schematic diagram of a software structure of an electronic device is shown as an example;

[0059] Figure 7 is a schematic diagram for describing physiological invariance as an example;

[0060] Figure 8 Schematic diagram of a conventional twin network encoder shown as an example;

[0061] Fig. 9 A schematic diagram of a network architecture of a twin network encoder with non-shared parameters is shown as an example;

[0062] Fig.10 is a schematic diagram of an exemplary heart rate deceleration force regression model training phase;

[0063] Fig.11 is a schematic diagram showing an exemplary heart rate deceleration force prediction stage;

[0064] Fig.12 Schematic diagram showing the relationship between a PPG signal input into a heart rate deceleration force regression model, a first heart rate deceleration force, and a second heart rate deceleration force in a heart rate deceleration force prediction stage;

[0065] Fig.13 The figure is a schematic diagram of a user interface for displaying warning information. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0067] The term "and / or" in this article is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0068] The terms "first" and "second" in the description and claims of the embodiments of the present application are used to distinguish different objects rather than to describe a specific order of objects. For example, a first target object and a second target object are used to distinguish different target objects rather than to describe a specific order of target objects.

[0069] In the description of the embodiments of the present application, words such as "exemplary", "for example", or "optionally" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary", "for example", or "optionally" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "for example", or "optionally" is intended to present related concepts in a specific way.

[0070] In the description of the embodiments of the present application, the names of various controls, areas, options, and entrances are only used for illustration and are not intended to limit the embodiments of the present application. That is, in actual use, they may also be described as other names.

[0071] In the description of the embodiments of the present application, unless otherwise specified, the meaning of "multiple" refers to two or more than two. For example, multiple processing units refer to two or more processing units; multiple systems refer to two or more systems.

[0072] In the description of the embodiments of the present application, unless otherwise specified, only one user interface can be displayed at the same time. For example, when interface 10a is displayed, interface 10b will not be displayed on the current screen. Conversely, when interface 10b is displayed, interface 10a will not be displayed on the current screen. However, the elements included in different interfaces can be the same.

[0073] In the description of the embodiments of the present application, unless otherwise specified, the dotted lines in the drawings are only for illustration, that is, they are not shown in actual use.

[0074] Cardiovascular diseases such as arrhythmia and heart failure are serious diseases that have a great impact on people's health. At present, in order to reduce the risk of cardiovascular events, the heart rate deceleration force (DC) can be used to measure the excitability of the vagus nerve, and then provide users with early warning information suitable for the current situation.

[0075] For example, when the DC value is a low-risk value (normal value), such as greater than or equal to 4.5ms (milliseconds), it indicates that the user's vagus nerve has increased the heart rate deceleration force. In this case, the user's body is not abnormal, and no warning information may be issued to the user, or the user may be prompted that the DC value is normal.

[0076] For example, when the DC value is a medium-risk value, such as between 2.5ms and 4.5ms (excluding the critical value), it indicates that the user's vagus nerve has reduced the ability to decelerate the heart rate. In this case, the user may experience symptoms such as fatigue, dizziness, memory loss, chest tightness, etc., which is a medium-risk situation for sudden death. The user can be given a medium-risk warning and corresponding suggestions, such as adjusting work and rest and taking a rest in time.

[0077] For example, when the DC value is a high-risk value, such as less than or equal to 2.5ms, it indicates that the user's vagus nerve significantly reduces the heart rate deceleration ability. In this case, the user may experience symptoms such as difficulty breathing and syncope, which is a high-risk situation for sudden death. The user can be given a high-risk warning and corresponding suggestions, such as timely medical treatment.

[0078] At present, in some implementations (implementation 1), a longer period of ECG signal is used to determine the heart rate deceleration force.

[0079] For example, record the ECG signal for 24 hours, and convert the ECG signal recorded in 24 hours into a sequence diagram with the cardiac cycle sequence as the horizontal axis and the RR value as the vertical axis.

[0080] Then, according to the sequence diagram obtained by conversion, the deceleration and acceleration cycles are determined and marked, such as by comparing the RR interval (the interval between the R wave to the R wave of the heartbeat, that is, the time interval between two consecutive R waves) of each cardiac cycle with the previous cardiac cycle, to determine whether the cycle belongs to a cardiac cycle with a decelerated heart rate or a cardiac cycle with an accelerated heart rate. Specifically, if the current cardiac cycle is longer than the previous cardiac cycle, it is determined that the current cardiac cycle belongs to a cardiac cycle with a decelerated heart rate and is marked as a black dot; on the contrary, if the current cardiac cycle is shorter than the previous cardiac cycle, it is determined that the current cardiac cycle belongs to a cardiac cycle with an accelerated heart rate and is marked as a white dot.

[0081] Next, determine the length of each heart rate segment. For example, take each deceleration point as the center and take the values ​​on both sides.

[0082] Next, the phases of each heart rate segment are arranged in order. For example, different heart rate segments are arranged in order with the selected deceleration point as the center.

[0083] Next, the signal is averaged for each heart rate segment.

[0084] Finally, by substituting the obtained signal average value into the heart rate deceleration force calculation formula, the average heart rate deceleration force corresponding to the past 24 hours can be obtained.

[0085] For specific implementation details of implementation method 1, please refer to the literature on determining heart rate deceleration force based on ECG signals, which will not be repeated here.

[0086] Therefore, based on implementation method 1, the user's heart rate deceleration force within 24 hours can be accurately predicted, and then according to the predicted heart rate deceleration force, the user can be given warning information suitable for the current situation.

[0087] However, in actual use scenarios, ECG signals need to be acquired with the help of professional equipment, such as an electrocardiograph. Since electrocardiographs are large in size, expensive, and complicated to operate, they require users to have professional knowledge and therefore cannot meet the daily needs of users.

[0088] In view of this, in some implementations (implementation 2), a photoplethysmography (PPG) signal over a longer period of time is used to determine the heart rate deceleration force.

[0089] It should be noted that PPG is a non-invasive detection method that uses photoelectric means to detect changes in blood volume in living tissue. Figure 1As shown in the figure, the user can put his finger (or other living tissue with blood vessels) in contact with the PPG optical module (PPG Optical Module, also known as: PPG sensor). In this way, the photodetector in the PPG sensor detects the intensity change of the reflected light (light emitted by a light-emitting diode (such as LED)) caused by the blood flow in the blood vessels in the finger, and can obtain the periodic waveform corresponding to the PPG signal over a period of time, such as Figure 1 WF_3 shown in FIG.

[0090] In addition, it should be noted that the intensity of the reflected light is related to the blood flow velocity, which is affected by the periodic heart rate. Therefore, the obtained WF_3 contains the pulsating change information of the blood volume during the cardiac cycle. For example, the peak value of the reflected light intensity (the highest point in a period of time) corresponds to the contraction of the heart, when the blood volume is the largest; the valley value of the reflected light intensity (the lowest point in a period of time) corresponds to the diastole of the heart, when the blood volume is the smallest. That is, WF_3 can contain the pulsating change information such as WF_1. And based on WF_1, the ECG signal waveform such as WF_2 can be estimated.

[0091] In addition, it should be noted that since PPG sensors are small, easy to carry, and cheap, predicting heart rate deceleration force based on PPG signals has become a trend.

[0092] In practical applications, PPG sensors can be integrated into wearable devices, such as Figure 2 The back of the smart watch 200 shown in the figure is the side in contact with the user's skin. In this way, when the user wears the smart watch 200 on the arm and turns on the heart rate monitoring function, the PPG sensor in the smart watch 200 can collect PPG signals while the user is wearing it, so that the smart watch 200, or other electronic devices (such as mobile phones) or servers that communicate with the smart watch 200 can predict the heart rate deceleration force based on the PPG signals.

[0093] However, in actual usage scenarios, PPG signals are greatly affected by external factors such as the wearer's wearing method, environment, and exercise status. If the PPG signal quality is poor, it will not accurately reflect the user's heart rate value at each moment, which will lead to inaccurate heart rate deceleration force predicted based on the PPG signal.

[0094] In view of this, an embodiment of the present application provides a heart rate deceleration force prediction method (implementation method 3), which aims to enhance the characteristics related to the heart rate deceleration force in the PPG signal and optimize the heart rate deceleration force output strategy, so as to facilitate users' daily use while ensuring the accuracy of the predicted heart rate deceleration force, so that the warning information based on the predicted heart rate deceleration force is more suitable for the actual situation.

[0095] Specifically in the embodiment of the present application, implementation method 3 needs to involve client equipment and server equipment.

[0096] Exemplarily, in some embodiments of the present application, the client device includes, for example, a first electronic device (such as Figure 3 The smart watch 200 and the second electronic device (such as Figure 3 The mobile phone 100 in the example, the server device is, for example, a server (eg Figure 3 For this implementation, the smart watch 200 can collect PPG signals and transmit the PPG signals to the server 300; the server 300 can use the deployed heart rate deceleration force regression model to process the PPG signals (extract features related to the heart rate deceleration force and determine the first heart rate deceleration force), and based on the heart rate deceleration force output strategy, predict the second heart rate deceleration force, store the second heart rate deceleration force corresponding to each time period, and send the second heart rate deceleration force stored in the preset time to the mobile phone 100 when the preset time (such as 24 hours later) is reached; the mobile phone 100 can determine the average heart rate deceleration force within the preset time according to the second heart rate deceleration force in each time period, and display appropriate warning information.

[0097] It should be noted that the first heart rate deceleration force can be understood as the predicted initial heart rate deceleration force, and the second heart rate deceleration force can be understood as the heart rate deceleration force after optimizing the first heart rate deceleration force.

[0098] Exemplarily, in some other embodiments of the present application, the client device includes, for example, a first electronic device (such as Figure 3 In the smart watch 200, the server device is, for example, a server (such as Figure 3For this implementation, the smart watch 200 can collect PPG signals and transmit the PPG signals to the server 300; the server 300 can use the deployed heart rate deceleration force regression model to process the PPG signals (extract features related to the heart rate deceleration force and determine the first heart rate deceleration force), and based on the heart rate deceleration force output strategy, predict the second heart rate deceleration force, store the second heart rate deceleration force corresponding to each time period, and send the second heart rate deceleration force stored in the preset time to the smart watch 200 when the preset time (such as 24 hours later) is reached; the smart watch 200 can determine the average heart rate deceleration force in the preset time according to the second heart rate deceleration force in each time period, and display appropriate warning information.

[0099] In addition, in some other embodiments of the present application, implementation method 3 may also be performed by only the first electronic device (such as Figure 3 That is, the smart watch 200 integrates a PPG sensor, a heart rate deceleration force regression model, and a heart rate deceleration force output strategy. In this way, the smart watch 200 can process the collected PPG signal through the heart rate deceleration force regression model (extract features related to the heart rate deceleration force, determine the first heart rate deceleration force), and predict the second heart rate deceleration force based on the heart rate deceleration force output strategy, and store the second heart rate deceleration force corresponding to each time period. Accordingly, when the preset time is reached (such as 24 hours later), the smart watch 200 can determine the average heart rate deceleration force within the preset time according to the second heart rate deceleration force in each time period, and display appropriate warning information.

[0100] It should be understood that the above description is merely an example listed for a better understanding of the technical solution of this embodiment, and is not intended to be the sole limitation to this embodiment.

[0101] For the sake of convenience, the terminal involved in the implementation mode 3 in the embodiment of the present application includes Figure 3 The mobile phone 100, smart watch 200 and server 300 shown in the figure are taken as an example. In order to better understand the technical solution provided by the embodiment of the present application, before describing the technical solution of the embodiment of the present application, the hardware structure of the smart watch 200, and the hardware structure and software structure of the mobile phone 100 are first described in conjunction with the accompanying drawings.

[0102] See also Figure 4 , which exemplarily shows a hardware structure diagram of a smart watch 200.

[0103] like Figure 4As shown, the smart watch 200 may include: a processor 201, a wireless fidelity (WiFi) chip 202, a Bluetooth (BT) chip 203, a positioning (GPS) chip 204, a PPG sensor 205, a display screen 206, a memory 207, an audio module 208, a charging chip 209, a charging management module 210, a power management module 211, a battery 212, etc.

[0104] The processor 201 may include one or more processing units, and different processing units may be independent devices or integrated into one or more processors. These processing units may implement different control logics. For example, when the heart rate monitoring function is turned on, an operation control signal may be sent to the PPG sensor so that the PPG sensor can collect the PPG signal when the heart rate detection function is turned on and the user wears the smart watch 200.

[0105] Among them, the WiFi chip 202 and / or the BT chip 203 can help the smart watch 200 establish a communication connection with the mobile phone 100. In this way, the user can view the data collected by the smart watch 200, such as heart rate data, sports data, etc., through the sports health application provided by the mobile phone 100, or a dedicated heart rate monitoring application.

[0106] Among them, the WiFi chip 203 can also help the smart watch 200 establish a communication connection with the server 300. In this way, the PPG signal collected by the smart watch 200 can be transmitted to the server 300.

[0107] The GPS chip 204 is used to achieve positioning.

[0108] It should be noted that the WiFi chip 202 , the BT chip 203 and the GPS chip 204 may be independent chips or a three-in-one chip.

[0109] The PPG sensor 205 is used to collect PPG signals.

[0110] Among them, the display screen displays images, videos, etc.

[0111] The memory 207 may be used to store program codes required for the operation of the smart watch 200, as well as data created during use.

[0112] The audio module 208 may include a speaker, a receiver, a microphone, an earphone interface, and the like.

[0113] The charging chip 209 is used to receive charging input from a charger. In some implementations, the charging chip 209 may be a wireless charging chip or a universal serial bus (USB) charging interface.

[0114] The charging management module 210 is used to receive charging input, charge the battery 212 , and supply power to the smart watch 200 through the power management module 211 during charging.

[0115] It should be understood that the above description is only an example listed for a better understanding of the technical solution of this embodiment, and is not intended to be a specific limitation on the smart watch 200. In other embodiments of the present application, the smart watch 200 may include more or fewer components than shown in the figure, or combine some components, or split some components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0116] See also Figure 5 , which exemplarily shows a hardware structure diagram of a mobile phone 100.

[0117] like Figure 5 As shown, the mobile phone 100 may include: a processor 110, an external memory interface 120, an internal memory 121, a USB interface 130, a charging management module 140, a power management module 141, a battery 142, an antenna 1, an antenna 2, a mobile communication module 150, a wireless communication module 160, an audio module 170, a sensor module 180, a button 190, a motor 191, an indicator 192, a camera 193, a display screen 194, and a subscriber identification module (SIM) card interface 195, etc.

[0118] The audio module 170 may include a speaker 170A, a receiver 170B, a microphone 170C, an earphone jack 170D, and the like.

[0119] Among them, the sensor module 180 may include a pressure sensor, a gyroscope sensor, an air pressure sensor, a magnetic sensor, an acceleration sensor, a distance sensor, a proximity light sensor, a fingerprint sensor, a temperature sensor, a touch sensor, an ambient light sensor, a bone conduction sensor, etc., which are not listed here one by one and the present application does not impose any limitation on this.

[0120] The processor 110 may include one or more processing units, and different processing units may be independent devices or integrated into one or more processors.

[0121] Exemplarily, in some implementations, the processor 110 may include an application processor (AP), a modem processor (Modem), a graphics processor (GPU), an image signal processor (ISP), a controller, a video codec, a digital signal processor (DSP), a baseband processor, a neural-network processing unit (NPU), etc., which are not listed one by one here and are not limited in this application.

[0122] The controller can generate an operation control signal according to the instruction operation code and the timing signal to complete the control of instruction fetching and instruction execution.

[0123] The processor 110 may also be provided with a memory for storing instructions and data. In some implementations, the memory in the processor 110 is a cache memory. The memory may store instructions or data that the processor 110 has just used or cyclically used. If the processor 110 needs to use the instruction or data again, it may be directly called from the memory. This avoids repeated access, reduces the waiting time of the processor 110, and thus improves the efficiency of the system.

[0124] Among them, the wireless communication function of the mobile phone 100 can be realized through antenna 1, antenna 2, mobile communication module 150, wireless communication module 160, modem processor and baseband processor.

[0125] Among them, antenna 1 and antenna 2 are used to transmit and receive electromagnetic wave signals. Each antenna in mobile phone 100 can be used to cover a single or multiple communication frequency bands. Different antennas can also be reused to improve the utilization rate of the antenna. For example, antenna 1 can be reused as a diversity antenna of a wireless local area network. In some other implementations, the antenna can be used in combination with a tuning switch.

[0126] The modem processor may include a modulator and a demodulator.

[0127] Among them, the mobile communication module 150 can provide wireless communication solutions including the second generation mobile phone communication technology specifications (2-Generation wireless telephone technology, 2G), the third generation mobile communication technology (3rd-Generation, 3G), the fourth generation mobile communication technology (the 4th generation mobile communication technology, 4G), and the fifth generation mobile communication technology (5th-Generation Mobile Communication Technology, 5G) applied on the mobile phone 100.

[0128] Among them, the wireless communication module 160 can provide wireless communication solutions applied to the mobile phone 100, including wireless local area networks (WLAN) (such as wireless fidelity (Wi-Fi) networks), Bluetooth (BT), global navigation satellite system (GNSS), frequency modulation (FM), near field communication technology (NFC), infrared technology (IR), etc.

[0129] In some implementations, the antenna 1 of the mobile phone 100 is coupled to the mobile communication module 150, and the antenna 2 is coupled to the wireless communication module 160, so that the mobile phone 100 can communicate with the network and other devices through wireless communication technology. For example, it communicates with the server 300 to obtain the predicted heart rate deceleration force from the server 300, so that the mobile phone 100 can display appropriate warning information according to the heart rate deceleration force.

[0130] The camera 193 is used to capture static images or videos. In some implementations, the mobile phone 100 may include 1 or N cameras 193, where N is a positive integer greater than 1.

[0131] The display screen 194 is used to display images, videos, etc. The display screen 194 includes a display panel. In some implementations, the mobile phone 100 may include 1 or N display screens 194, where N is a positive integer greater than 1.

[0132] The external memory interface 120 can be used to connect an external memory card, such as a Micro SD card, to expand the storage capacity of the mobile phone 100. The external memory card communicates with the processor 110 through the external memory interface 120 to implement a data storage function, such as storing music, video and other files in the external memory card.

[0133] Among them, the internal memory 121 can be used to store computer executable program codes, and the executable program codes include instructions. The processor 110 executes various functional applications and data processing of the mobile phone 100 by running the instructions stored in the internal memory 121. The internal memory 121 may include a program storage area and a data storage area. Among them, the program storage area can store an operating system, an application required for at least one function, etc. The data storage area can store data created during the use of the mobile phone 100, etc. In addition, the internal memory 121 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, a universal flash storage (UFS), etc.

[0134] Specifically in the embodiment of the present application, the storage data area of ​​the internal memory 121 can store the heart rate deceleration force sent by the server 300, various data provided by the smart watch 200, and preset warning information corresponding to different heart rate deceleration forces.

[0135] It should be understood that the above description is only an example listed for better understanding of the technical solution of this embodiment, and is not intended to be a specific limitation on the mobile phone 100. In other embodiments of the present application, the mobile phone 100 may include more or fewer components than shown in the figure, or combine some components, or split some components, or arrange the components differently. The components shown in the figure may be implemented in hardware, or software, or a combination of software and hardware.

[0136] In addition, it should be noted that Figure 5 An operating system is also running on the components of the mobile phone 100, such as the iOS operating system developed by Apple, the Android open source operating system developed by Google, and the Windows operating system developed by Microsoft. These operating systems may adopt a layered architecture, an event-driven architecture, a micro-kernel architecture, a micro-service architecture, or a cloud architecture.

[0137] For ease of explanation, the embodiment of the present application takes the Android system with a layered architecture as an example to illustrate the software structure of the mobile phone 100.

[0138] It should be noted that although the embodiments of the present application are described using the Android system as an example, its basic principles are also applicable to electronic devices based on operating systems such as iOS or Windows.

[0139] See also Figure 6 , which exemplarily shows a software structure block diagram of a mobile phone 100.

[0140] like Figure 6 As shown, the layered architecture of the mobile phone 100 divides the software into several layers, each layer has a clear role and division of labor. The layers communicate with each other through software interfaces. In some embodiments of the present application, the Android system is divided into four layers, which are the application layer, the application framework layer, the Android runtime (Android runtime) and the system library, and the kernel layer from top to bottom.

[0141] The application layer may include a series of application packages. Figure 6 As shown, the application package may include sports health, Bluetooth, WiFi, settings, and other applications.

[0142] The application framework layer provides application programming interfaces (APIs) and programming frameworks for applications in the application layer. In some implementations, these programming interfaces and programming frameworks can be described as functions. Figure 6 As shown, the application framework layer may include a content provider, a view system, a window manager, and the like.

[0143] Among them, Android Runtime includes core libraries and virtual machines. Android Runtime is responsible for the scheduling and management of the Android system.

[0144] The core library consists of two parts: one part is the function that needs to be called by the Java language, and the other part is the Android core library.

[0145] The application layer and the application framework layer run in a virtual machine. The virtual machine executes the Java files of the application layer and the application framework layer as binary files. The virtual machine is used to perform functions such as object life cycle management, stack management, thread management, security and exception management, and garbage collection.

[0146] The system library can include multiple functional modules, such as surface manager, media library, 3D graphics processing library (such as OpenGL ES), 2D graphics engine (such as SGL), etc.

[0147] The surface manager is used to manage the display subsystem and provide the fusion of 2D and 3D layers for multiple applications.

[0148] The media library supports a variety of commonly used audio and video format playback and recording, as well as static image files, etc. The media library can support a variety of audio and video encoding formats, such as: MPEG4, H.264, MP3, AAC, AMR, JPG, PNG, etc.

[0149] The 3D graphics processing library is used to implement 3D graphics drawing, image rendering, synthesis, and layer processing.

[0150] It can be understood that the 2D graphics engine mentioned above is a drawing engine for 2D drawing.

[0151] The kernel layer is the layer between hardware and software, including various hardware drivers and hardware management. Figure 6 As shown, the kernel layer may include display driver, audio driver, sensor driver, etc.

[0152] It should be understood that the above description is only an example listed for a better understanding of the technical solution of this embodiment, and is not a specific limitation on the operating system of the mobile phone 100. In other embodiments of the present application, the operating system of the mobile phone 100 may include more or fewer layers than shown in the figure, and each layer may include more or fewer components, which is not limited in the present application.

[0153] The following Figure 3 Taking the system for implementing the heart rate deceleration force prediction method provided in the embodiment of the present application as an example, the heart rate deceleration force prediction method provided in the embodiment of the present application is specifically described from the stages of training the heart rate deceleration force regression model (heart rate deceleration force regression model training stage), optimizing the heart rate deceleration force output strategy (heart rate deceleration force output strategy optimization stage), and predicting the heart rate deceleration force based on the heart rate deceleration force regression model and the optimized heart rate deceleration force output strategy (heart rate deceleration force prediction stage).

[0154] The following is the content of the heart rate deceleration force regression model training stage

[0155] It should be noted that according to the learning strategy of contrastive learning, in visual tasks, different perspectives of the same individual should share important features. Figure 7As shown in (1) to (4), for the same cuboid and cylinder, at different viewing angles (such as different brightness, different chromaticity, different depth, different optical flow, etc.), the shape, size, positional relationship and other features of the cuboid and cylinder are the same (that is, such features are shared features). Based on this principle, it is applied to the heart rate monitoring scenario. It can be seen that for the same user, although the ECG signal at the same time reflects the potential transmission and changes of the heart, and the PPG signal reflects the changes in blood volume during the heart's cardiac cycle, at this moment, the ECG signal and the PPG signal share the key physiological information features of the user. That is, the physiological state of the same individual, at different viewing angles (ECG signal perspective and PPG signal perspective) at the same time, the physiological information features are unchanged (or the same).

[0156] Therefore, in the embodiment of the present application, the training phase of the heart rate deceleration force regression model can be completed with the help of ECG signals and PPG signals through the characteristic of physiological invariance. Specifically, based on physiological invariance, a learning strategy of contrastive learning can be adopted, the PPG signal is used as a query sample (query sample), and the ECG signal is used as a verification sample (key sample), and then an end-to-end network architecture based on a twin network encoder is adopted to determine the features to be compared in the PGG signal and the EGG signal. Then, the loss function used in the training process is defined based on the obtained features. Finally, based on the heart rate deceleration force algorithm, the PPG signal processed by the network architecture and loss function of the twin network encoder is processed to obtain the first heart rate deceleration force. The entire structure for obtaining the first heart rate deceleration force can be regarded as the heart rate deceleration force prediction stage, which is a heart rate deceleration force regression model for predicting the first heart rate deceleration force.

[0157] In addition, it should be noted that the end-to-end twin network encoder-based network architecture used in the embodiments of the present application is different from the network structure of the traditional twin network encoder.

[0158] Specifically, the network structure of the traditional twin network encoder can be as follows Figure 8 As shown, each input object (such as and ), although they have passed through their own convolutional neural networks, all weights (all parameters) of the two convolutional neural networks are shared. That is, the two convolutional neural networks share weights .

[0159] In the embodiment of the present application, the network architecture of the twin network encoder does not share the weights used by the two encoders, that is, the parameters loaded by the two encoders are different when they work. In order to distinguish it from the network structure of the traditional twin network encoder, the following embodiment refers to the network architecture of the twin network encoder involved in the embodiment of the present application as the network architecture of the twin network encoder with non-shared parameters. For example Fig. 9 As shown, in the network architecture of the twin network encoder with non-shared parameters, a first encoder and a second encoder may be included. The network structures of the first encoder and the second encoder may be the same, for example, both are residual network structures, and both include a method for receiving an input object (such as and ), the maximum pooling layer connected to the convolutional layer, and multiple convolutional layers after the maximum pooling layer. However, the weights (parameters) loaded by the first encoder and the second encoder are different when they work. For example, the first encoder loads , the second encoder records .

[0160] The following combination Fig.10 , the process of training and obtaining the features related to the heart rate deceleration force in the PPG signal during the heart rate deceleration force regression model training phase is specifically described.

[0161] See also Fig.10 For example, during the training process, a mini-batch can be given, that is, a smaller data set is selected from the training data set. The model uses each mini-batch of samples for forward propagation, loss calculation and back propagation, and then updates the model parameters according to the gradients of these samples.

[0162] It should be noted that this model can include Fig. 9 The first encoder and the second encoder mentioned in , and the fully connected layers corresponding to the first encoder and the second encoder respectively.

[0163] Continue to see Fig.10 , for example, during the training process, a PPG signal and an ECG signal in a mini-batch can be input each time. , and Input the first encoder and convert the corresponding , and ECG signals such as , and Input the second encoder. In this way, after the PPG signal in the same mini-batch passes through the first encoder and the fully connected layer after the first encoder, the features output by the fully connected layer are normalized by the Euclidean norm (L2 norm), and the final features to be compared can be obtained, such as , and After the ECG signals in the same mini-batch pass through the second encoder and the fully connected layer after the second encoder, the features output by the fully connected layer are normalized by L2 norm to obtain the final features to be compared, such as , and .

[0164] Continue to see Fig.10 , for example, in the obtained , and ,as well as , and After that, the individual discrimination probabilities of the positive sample pairs and the negative sample pairs can be calculated respectively, which can be specifically calculated according to the following formulas (1) and (2).

[0165] (1)

[0166] (2)

[0167] in, It is a PPG signal The features obtained after the first encoder, the fully connected layer after the first encoder, and L2 norm normalization, It is the ECG signal The features obtained after the second encoder, the fully connected layer after the second encoder, and L2norm normalization, is the temperature coefficient constant.

[0168] According to formula (1) and formula (2), based on the assumption that each category is independent, the negative log-likelihood loss function of the joint probability of all categories can be obtained, as follows:

[0169]

[0170] Next, the obtained loss function is used to process the pre-constructed positive sample pairs and negative sample pairs, and the relationship between the features of the positive sample pairs and the negative sample pairs can be obtained.

[0171] It should be noted that the positive sample pair can be understood as the sample to be classified, and the negative sample pair is the sample used to distinguish the positive sample pair. Specifically in the embodiment of the present application, the positive sample pair is composed of a PPG signal and a corresponding ECG signal. For example, and the ECG signal of the corresponding time period A positive sample pair can be formed. A negative sample pair is composed of the PPG signal and other PPG signals in the same mini-batch, as well as the ECG signals corresponding to other PPG signals. For example, and in the same mini-batch , , ,as well as , forming 4 pairs of negative sample pairs respectively. In this way, after being processed by the loss function obtained above, the relationship between the features of the positive sample pairs and the negative sample pairs can be obtained.

[0172] For ease of understanding, the present embodiment describes the relationship between the features of the positive sample pair and the negative sample pair by the distance. Fig.10 ,Exemplarily, for the features corresponding to the positive sample pairs (same color), the distance is closer; for the features corresponding to the negative sample pairs (different colors), the distance is farther.

[0173] Specifically in the embodiment of the present application, the feature of closer distance can be understood as a feature related to the heart rate deceleration force. Based on this, by processing the PPG signal that meets the requirements (including the PPG signal with features related to the heart rate deceleration force) based on the heart rate deceleration force algorithm, the first heart rate deceleration force can be obtained.

[0174] Regarding the heart rate deceleration force algorithm, the calculation process of calculating the first heart rate deceleration force can be as follows:

[0175] First, the RR interval stacks that meet the screening conditions (such as motion state judgment) are calculated to obtain a stacking matrix, such as .

[0176] Next, calculate the average of the RR interval stacking matrix, and then use the calculation formula based on the center point's nearest neighbor to perform the calculation.

[0177] The implementation details of this part are similar to the process of determining the heart rate deceleration force of the ECG signal. For details, please refer to relevant literature and will not be repeated here.

[0178] Therefore, by combining the part of the PPG signal that meets the requirements and the heart rate deceleration force algorithm part into one model, a heart rate deceleration force regression model can be obtained.

[0179] The following is the content of the heart rate deceleration force output strategy optimization stage

[0180] It should be noted that the heart rate deceleration force is a quantitative assessment of the level of the user's vagus nerve through an overall trend analysis over a long period of time (such as 24 hours), and then early warning information is provided. Based on the long-term measurement and analysis requirements required for the characteristic of the heart rate deceleration force, the embodiment of the present application provides a momentum-based post-processing heart rate deceleration force output strategy. Specifically, a more accurate output is achieved by integrating historical heart rate deceleration forces (inputting the heart rate deceleration force corresponding to the current time period). In order to achieve this effect, in the embodiment of the present application, by dividing the PPG signal within 24 hours into multiple small time segments, the iterative formula for the heart rate deceleration force updated with the time segment is derived, and then the function expression of the momentum update is sorted out. Finally, based on the function expression of the momentum update, a more accurate output of the heart rate deceleration force corresponding to the current time period can be achieved.

[0181] In addition, it should be noted that the above momentum-based post-processing heart rate deceleration force output strategy is also applicable to the scenario of determining the heart rate deceleration force based on ECG signals. In this scenario, the ECG signal within 24 hours is divided into multiple small time segments, and the iterative formula for updating the heart rate deceleration force with the time segment is derived, and then the function expression of momentum update is sorted out.

[0182] The following describes the derivation process of the function expression corresponding to the momentum update after the PPG signal or ECG signal is divided into multiple small time segments.

[0183] Specifically, set The RR interval stacking matrix at time is , The RR interval stacking matrix at time is , then the momentum update function expression is as follows:

[0184]

[0185] in, for The RR interval stacking matrix at time The number of heart rate segments included in for The The value corresponding to each heart rate segment is for The RR interval stacking matrix at time The number of heart rate segments included in for The The value corresponding to each heart rate segment.

[0186] Therefore, based on the above momentum update function expression, we can get The heart rate deceleration force at the moment, that is, the second heart rate deceleration force in the above embodiment .because The heart rate deceleration force of the moment is combined with The heart rate deceleration force before the moment, so it can more accurately reflect the user's overall The heart rate deceleration force at all times makes the warning information more reasonable and reduces false warnings caused by interference from environmental factors, wearing methods, exercise status, etc.

[0187] The following is the content of the heart rate deceleration force prediction stage

[0188] Specifically, after completing the training of the heart rate deceleration force regression model, obtaining the heart rate deceleration force regression model (obtained by unsupervised comparative learning training) required for implementing the heart rate deceleration force prediction method provided in the embodiment of the present application, and completing the operation of the heart rate deceleration force output strategy optimization phase, and obtaining the momentum update function expression required for implementing the heart rate deceleration force prediction method provided in the embodiment of the present application, the obtained heart rate deceleration force regression model and momentum update function expression can be deployed on the server 300. In this way, after the user wears a wearable device with an integrated PPG sensor, such as a smart watch 200, and turns on the heart rate deceleration force monitoring function, the smart watch 200 can collect the PPG signal corresponding to the user in real time during the wearing period, and transmit the collected PPG signal to the server 300.

[0189] Exemplarily, in some implementations, the server 300 may pre-process the received PPG signal.

[0190] For example, the PPG signal is timestamped and synchronized, packet loss is detected based on the timestamps of adjacent data packets of the PPG signal, and the PPG signal is filtered and denoised.

[0191] For example, according to the preset small time segment length, the received PPG signal is processed into a mini-batch of PPG signals of the preset small time segment length in time order. Then, the PPG signal in the mini-batch that meets the preset small time segment length is input into the heart rate deceleration force regression model, such as Fig.11 shown.

[0192] Continue to see Fig.11 For example, after being processed by the heart rate deceleration force regression model, the heart rate deceleration force corresponding to the set small time segment can be obtained.

[0193] It should be noted that the heart rate deceleration force obtained after being processed by the heart rate deceleration force regression model can be understood as the first heart rate deceleration force corresponding to the current small time segment.

[0194] Continue to see Fig.11 For example, the first heart rate deceleration force obtained after processing by the heart rate deceleration force regression model needs to be updated through the momentum update function expression to obtain the optimized heart rate deceleration force, that is, the second heart rate deceleration force mentioned above.

[0195] Continue to see Fig.11 For example, in some implementations, the server 300 may store the second heart rate deceleration force corresponding to each set small time segment for a long period of time, such as within 24 hours. Fig.11 In, you can store The second heart rate deceleration force corresponding to the moment , The second heart rate deceleration force corresponding to the moment , The second heart rate deceleration force corresponding to the moment After 24 hours have passed, the second heart rate deceleration force corresponding to each small time segment stored in the 24 hours is sent to the mobile phone 100 (or the smart watch 200). That is, the server 300 stores the second heart rate deceleration force.

[0196] In addition, it should be noted that, in order to reduce the memory usage of the server 300, the stored second deceleration force may be cleared after the second deceleration force of the set duration is issued. For example, after the second heart rate deceleration force corresponding to all the small time segments recorded within 24 hours is issued to the mobile phone 100 (or smart watch 200) every 24 hours, the second heart rate deceleration force within 24 hours that has been issued is cleared.

[0197] For example, in some other implementations, the server 300 may also directly send the obtained second heart rate deceleration force to the mobile phone 100 (or smart watch 200) after obtaining the second heart rate deceleration force of a small time segment each time. That is, the server 300 does not store the second heart rate deceleration force.

[0198] It should be understood that the above description is merely an example listed for a better understanding of the technical solution of this embodiment, and is not intended to be the sole limitation to this embodiment.

[0199] The relationship between the PPG signal input into the heart rate deceleration force regression model, the first heart rate deceleration force, and the second heart rate deceleration force can be as follows: Fig.12 shown.

[0200] See also Fig.12 , for example, in The PPG signal corresponding to the moment, such as After inputting into the heart rate deceleration force regression model, we can obtain The first heart rate deceleration force corresponding to the moment . The first heart rate deceleration force corresponding to the moment Processed by the momentum updated function expression (for example, combined with The second heart rate deceleration force corresponding to the moment , not shown in the figure), we can get The second heart rate deceleration force corresponding to the moment .

[0201] Continue to see Fig.12 , for example, in The PPG signal corresponding to the moment, such as After inputting into the heart rate deceleration force regression model, we can obtain The first heart rate deceleration force corresponding to the moment . The first heart rate deceleration force corresponding to the moment Processed by the momentum updated function expression (for example, combined with The second heart rate deceleration force corresponding to the moment ) then we can get The second heart rate deceleration force corresponding to the moment .

[0202] Continue to see Fig.12 , for example, in The PPG signal corresponding to the moment, such as After inputting into the heart rate deceleration force regression model, we can obtain The first heart rate deceleration force corresponding to the moment . The first heart rate deceleration force corresponding to the moment Processed by the momentum updated function expression (for example, combined with The second heart rate deceleration force corresponding to the moment ) then we can get The second heart rate deceleration force corresponding to the moment .

[0203] from Fig.12 It can be seen that due to Second heart rate deceleration force at the moment , which combines The second heart rate deceleration force before the moment, such as The second heart rate deceleration force corresponding to the moment .therefore, Second heart rate deceleration force at the moment , not only integrating and The second heart rate deceleration force corresponding to the moment , and also incorporates The second heart rate deceleration force corresponding to the moment Based on this heart rate deceleration force prediction method, the second heart rate deceleration force of the user in each small time period over a long period of time can be more accurately reflected overall, making the warning information more reasonable and reducing false warnings caused by interference from environmental factors, wearing methods, exercise status, etc.

[0204] See also Fig.13 , exemplarily showing a user interface diagram of warning information related to heart rate deceleration force displayed on a mobile phone 100.

[0205] See also Fig.13 In (1), for example, in this scenario, the user can start the sports health application by clicking the sports health application icon 10a-1 on the interface 10a. That is, the user interface of the mobile phone changes from display interface 10a to display Fig.13 Interface 10b shown in (2).

[0206] See also Fig.13 In (2), for example, in some implementations, the interface 10b may be the first interface of a sports health application. In the interface 10b, an option to view heart health may be displayed, such as option 10b-1.

[0207] For example, after the user clicks option 10b-1, the mobile phone responds to the user operation and the user interface of the mobile phone may display Fig.13 Interface 10c shown in (3).

[0208] See also Fig.13 In (3), exemplarily, in some implementations, the interface 10c may include a first set time length, such as the average heart rate deceleration force of 24 hours (all day) (as shown in 3.4 in the figure); a second set time length, such as the average heart rate deceleration force of 7 consecutive days (as shown in 4.8 in the figure), and the heart rate deceleration force corresponding to different small time segments in the first set time length.

[0209] For example, in some implementations, the small time segments in the first duration can be divided according to the user's awake time and sleep time. For example, as shown in the figure, the average heart rate deceleration force corresponding to the awake time in the first set duration is 5.1, and the average heart rate deceleration force corresponding to the sleep time is 2.4.

[0210] Continue to see Fig.13In (3), for example, in order to enable the user to better understand the risks indicated by different DC values, the interface 10c may also display the division of DC values ​​into normal, medium-risk, high-risk, etc. interval values, as well as the relationship between DC during sleep and DC during wakefulness.

[0211] Continue to see Fig.13 (3), illustratively, in some implementations, the interface 10c may also display warning information that matches the user's average heart rate deceleration power in different time periods within 24 hours or throughout the day. For example, when the user's average heart rate deceleration power throughout the day is at a medium-risk value, and the relationship between the DC during sleep and the DC during wakefulness does not conform to the normal relationship, the user may be reminded in the interface 10c that there is an abnormality in the body, and the user may be advised to take a rest in time, seek medical treatment, and other information, such as Fig.13 (3) in the figure states “Your body is showing signs of abnormality. Please take a rest and adjust your work and rest schedule. If necessary, please go to the hospital for medical treatment.”

[0212] Exemplarily, in other implementations, when the average heart rate deceleration force for the whole day, or the heart rate deceleration force corresponding to multiple small time segments is obtained, and the obtained heart rate deceleration force is not a normal value, a warning message can be popped up in the notification center by way of notification.

[0213] For example, in other implementations, when the warning information pops up in the notification center, the user may be reminded by vibration, ringing, etc. so that the user can check the warning information in time.

[0214] It should be understood that the above description is merely an example listed for a better understanding of the technical solution of this embodiment, and is not intended to be the sole limitation to this embodiment.

[0215] Therefore, based on the heart rate deceleration force prediction method provided in the embodiment of the present application, it is possible to conveniently and accurately predict the heart rate deceleration force over a long period of time based on the PPG signal, thereby better monitoring the user's heart health.

[0216] In addition, it should be noted that the implementation logic of the heart rate deceleration force prediction method provided in the embodiment of the present application can also be applied to other scenarios, such as EEG scenarios and blood concentration monitoring scenarios.

[0217] Understandably, when applied to other scenarios, the training data used in the model training phase is different, but it is also based on physiological invariance. For example, in an EEG scenario, based on physiological invariance, EEG signals, such as steady-state visual evoked potential (SSVEP) signals, can be feature extracted, classified and judged, and then the corresponding control instructions can be predicted. Then, based on the characteristics of momentum update, the function expression of momentum update suitable for EEG scenarios is derived through the control instructions obtained at adjacent moments. In this way, based on physiological invariance, the model obtained by unsupervised contrastive learning training can be used to obtain the first control instructions corresponding to different SSVEPs, and the first control instructions can be optimized based on the function expression of momentum update to obtain the second control instructions for controlling target objects, such as wheelchairs, remote controls, etc., so as to facilitate users to use brain-computer interface products.

[0218] In addition, it is understandable that, in order to implement the above functions, the electronic device includes hardware and / or software modules corresponding to the execution of each function. In combination with the algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application in combination with the embodiments, but such implementation should not be considered to be beyond the scope of the present application.

[0219] In addition, it should be noted that in actual application scenarios, the heart rate deceleration force prediction method provided by the above embodiments implemented by an electronic device can also be executed by a chip system included in the electronic device. Among them, the chip system may include a processor. The chip system can be coupled to a memory so that when the chip system is running, it calls a computer program stored in the memory to implement the steps executed by the above electronic device. Among them, the processor in the chip system can be an application processor or a processor other than an application processor.

[0220] In addition, an embodiment of the present application further provides a computer-readable storage medium, in which computer instructions are stored. When the computer instructions are executed on an electronic device, the electronic device executes the above-mentioned related method steps to implement the method in the above-mentioned embodiment.

[0221] In addition, an embodiment of the present application further provides a computer program product. When the computer program product is run on an electronic device, the electronic device executes the above-mentioned related steps to implement the method in the above-mentioned embodiment.

[0222] In addition, an embodiment of the present application further provides a device. The device may be a chip, a component or a module. The device may include a connected processor and a memory; wherein the memory is used to store computer-executable instructions, and when the device is running, the processor may execute the computer-executable instructions stored in the memory, so that the chip executes the methods in the above-mentioned method embodiments.

[0223] In addition, it can be seen from the above description that the electronic device, computer-readable storage medium, computer program product or chip provided in the embodiments of the present application are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0224] In addition, through the description of the above embodiments, those skilled in the art can understand that for the convenience and simplicity of description, only the division of the above functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed. That is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above.

[0225] In addition, it is understandable that in several embodiments provided in the present application, the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0226] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for predicting heart rate deceleration force, characterized in that: The heart rate deceleration force prediction method comprises: Within a first set time period, receiving a continuous photoplethysmogram (PPG) signal collected by a first electronic device; After each PPG signal of a preset time segment is received, a first heart rate deceleration force corresponding to the preset time segment is determined according to a heart rate deceleration force regression model and the PPG signal of the preset time segment, wherein the heart rate deceleration force regression model is obtained by training a twin network with non-shared parameters based on physiological invariance and unsupervised comparative learning with the help of sample PPG signals and sample ECG signals, wherein the physiological invariance refers to the physiological state of the same individual, and the physiological characteristics are unchanged from the perspective of ECG signals and PPG signals at the same time; wherein training the heart rate deceleration force regression model comprises: obtaining a sample PPG signal and a sample ECG signal, wherein the sample PPG signal and the sample ECG signal are collected from the same user and in the same time period PPG signals and corresponding ECG signals; dividing the sample PPG signals and the corresponding sample ECG signals into multiple small batch data sets; constructing positive sample pairs and negative sample pairs according to the sample PPG signals and the sample ECG signals in the same small batch data set; based on physiological invariance, using the twin network with unshared parameters through unsupervised comparative learning, comparatively learning the positive sample pairs and the negative sample pairs, and extracting physiological features that meet physiological invariance in the sample PPG signals and the sample ECG signals; defining the loss function used in the training process according to the physiological features; constructing the heart rate deceleration force regression model according to the twin network with unshared parameters, the loss function and the heart rate deceleration force algorithm; Determine a second heart rate deceleration force corresponding to the preset time segment according to a function expression of momentum update and a first heart rate deceleration force corresponding to the preset time segment, wherein the function expression of momentum update is determined according to a calculation formula of the RR interval corresponding to each preset time segment within the first set time length, and the second heart rate deceleration force is used to generate warning information; Wherein, determining the second heart rate deceleration force corresponding to the preset time segment according to the momentum updated function expression and the first heart rate deceleration force corresponding to the preset time segment includes: Based on the second heart rate deceleration force corresponding to the first time segment and the first heart rate deceleration force corresponding to the preset time segment, the second heart rate deceleration force corresponding to the preset time segment is determined, the first time segment is a preset time segment before the preset time segment, and the duration of the first time segment is the same as the duration of the preset time segment.

2. The heart rate deceleration force prediction method according to claim 1, characterized in that: Each of the small batch data sets includes a first PPG signal, a second PPG signal, a first ECG signal corresponding to the first PPG signal, and a second ECG signal corresponding to the second PPG signal; The constructing positive sample pairs and negative sample pairs according to the sample PPG signals and the sample ECG signals in the same mini-batch data set includes: According to the first PPG signal and the first ECG signal, and the second PPG signal and the second ECG signal in the same mini-batch data set, respectively construct the positive sample pairs; The negative sample pairs are constructed respectively according to the first PPG signal and the second PPG signal, the first PPG signal and the second ECG signal, and the second PPG signal and the first ECG signal in the same mini-batch data set.

3. The heart rate deceleration force prediction method according to claim 1, characterized in that: The heart rate deceleration force prediction method further includes: Within the first set time length, storing the second heart rate deceleration force corresponding to each preset time segment; After reaching the first set time, all second heart rate deceleration forces stored in the first set time are sent to a second electronic device, and the second electronic device is used to generate a warning message according to the received second heart rate deceleration force, and display the warning message and the received second heart rate deceleration force.

4. The heart rate deceleration force prediction method according to claim 1, characterized in that: The heart rate deceleration force prediction method further includes: For each of the preset time segments, after determining the second heart rate deceleration force corresponding to the preset time segment, the second heart rate deceleration force corresponding to the preset time segment is sent to a second electronic device, and the second electronic device is used to generate a warning message based on the received second heart rate deceleration force, and display the warning message and the received second heart rate deceleration force.

5. The heart rate deceleration force prediction method according to claim 3 or 4, characterized in that: The first electronic device is communicatively connected to the second electronic device, and the second electronic device is further configured to send the warning information and the corresponding second heart rate deceleration force to the first electronic device.

6. The heart rate deceleration force prediction method according to claim 1, characterized in that: The heart rate deceleration force prediction method further includes: Within the first set time length, storing the second heart rate deceleration force corresponding to each preset time segment; After reaching the first set time, all second heart rate deceleration forces stored in the first set time are sent to the first electronic device, and the first electronic device is used to generate warning information according to the received second heart rate deceleration force, and display the warning information and the received second heart rate deceleration force.

7. The heart rate deceleration force prediction method according to claim 1, characterized in that: The heart rate deceleration force prediction method further includes: For each of the preset time segments, after determining the second heart rate deceleration force corresponding to the preset time segment, the second heart rate deceleration force corresponding to the preset time segment is sent to the first electronic device, and the first electronic device is used to generate a warning message based on the received second heart rate deceleration force, and display the warning message and the received second heart rate deceleration force.

8. The heart rate deceleration force prediction method according to claim 1, characterized in that: The first electronic device is a wearable device integrated with a PPG sensor.

9. A terminal device, characterized in that: The terminal device includes: a memory and a processor, the memory and the processor are coupled; the memory stores program instructions, and when the program instructions are executed by the processor, the terminal device executes the heart rate deceleration force prediction method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: It comprises a computer program, which, when running on a terminal device, enables the terminal device to execute the heart rate deceleration force prediction method as described in any one of claims 1 to 8.

Images