Physiological characteristic index measuring method and related device

Abstract

The embodiment of the invention provides a physiological feature index measuring method and a related device, and relates to the technical field of terminals. The method comprises the steps that first information of a first user is acquired, the first information comprises physiological feature information of the first user, and when the first information is acquired, a first model is preset in the electronic equipment; measuring physiological feature indexes of the first user based on the updated first model; wherein the first model is updated in the following mode that the first information is input into the second model, output of the second model is obtained, and the output of the second model is multiple pieces of physiological feature information of the first user and physiological feature indexes corresponding to the multiple pieces of physiological feature information; and taking the output of the second model as a physiological feature information sample of the first user, and updating the first model to obtain an updated first model. In this way, the updated first model can identify the physiological feature index of the first user more correctly, and the measurement accuracy is improved.

Description

Technical Field

[0001] This application relates to the technical field of terminals, and in particular, to a method for measuring physiological characteristic indicators and related devices. Background Art

[0002] Currently, wearable devices can have the function of measuring blood pressure. However, wearable devices are prone to inaccurate blood pressure measurement. Summary of the Invention

[0003] Embodiments of this application provide a method for measuring physiological characteristic indicators and related devices, which are applied to the technical field of terminals and are conducive to improving the accuracy of blood pressure measurement.

[0004] In a first aspect, embodiments of this application propose a method for measuring physiological characteristic indicators, which can be applied to an electronic device. The method includes: obtaining first information of a first user, where the first information includes physiological characteristic information of the first user, and when obtaining the first information, a first model is preset in the electronic device; measuring physiological characteristic indicators of the first user based on the updated first model, where the physiological characteristic indicators of the first user include indicators for reflecting the physiological functions of the first user; where the first model is updated in the following manner: inputting the first information into a second model to obtain the output of the second model, and the output of the second model is multiple physiological characteristic information of the first user and physiological characteristic indicators corresponding to the multiple physiological characteristic information; the first information in the second model is a constraint condition; using the output of the second model as a physiological characteristic information sample of the first user to update the first model to obtain the updated first model.

[0005] The method provided by embodiments of this application can be applied to various scenarios, for example, blood pressure measurement scenarios and electroencephalogram recognition scenarios, etc. The physiological characteristic information of the first user can vary according to different scenarios.

[0006] In some implementations, in a blood pressure measurement scenario, the physiological characteristic information of the first user may include the first user's photoplethysmography (PPG) signal. In other implementations, the physiological characteristic information of the first user may include the first user's electroencephalogram signal.

[0007] The first information may include physiological characteristic information of the first user. In some scenarios, in addition to including physiological characteristic information of the first user, the first information may further include one or more pieces of information that affect physiological characteristic information, such as the gender, age, or body mass index (BMI) of the first user.

[0008] The first model can measure physiological characteristic indicators reflected by physiological characteristic information. For example, in a blood pressure measurement scenario, the first model can measure blood pressure values based on PPG signals.

[0009] The second model can be understood as a generative model that can generate multiple physiological characteristic information of the first user and the physiological characteristic indicators corresponding to the multiple physiological characteristic information according to the first information of the first user. It can be understood that these multiple physiological characteristic information of the first user and the physiological characteristic indicators corresponding to the multiple physiological characteristic information are not the information of the first user actually collected, but predicted by the second model.

[0010] In the embodiments of the present application, in Figure 4 In the example shown, the first model can refer to a blood pressure prediction meta-model, the updated model can refer to a calibration model, the second model can refer to a Diffusion model, and the first information can refer to personalized information.

[0011] The output of the second model can be used as a sample of the physiological characteristic information of the first user to update the first model, and an updated first model is obtained. The updated first model is used to measure the physiological characteristic indicators of the first user, so that the updated first model can more accurately identify the corresponding physiological characteristic indicators according to the physiological characteristic information of the first user, which is beneficial to improving the measurement accuracy.

[0012] In a possible implementation manner, the second model includes a reverse process and a forward process. The reverse process of the second model is used to generate multiple physiological characteristic information of the first user and the physiological characteristic indicators corresponding to the multiple physiological characteristic information with the first information as a constraint condition; the forward process of the second model is used to obtain the information for training the reverse process of the second model.

[0013] The structure of the second model can include two parts. One part is the forward process, which is used to generate the information for training the reverse process of the second model, and the second part is the reverse process, which can be used to generate the required samples. In this way, it is beneficial to realize the requirement of generating samples.

[0014] In a possible implementation manner, the reverse process of the second model is used to perform different noise reduction processes on the first region and the second region of the first information to generate multiple physiological characteristic information of the first user and the physiological characteristic indicators corresponding to the multiple physiological characteristic information. The second region is the region other than the first region, and the correlation degree between the first region and the physiological characteristic indicator is greater than the correlation degree between the second region and the physiological characteristic indicator.

[0015] In the embodiments of the present application, since the correlation degree between the first region and the physiological characteristic indicator is greater than the correlation degree between the second region and the physiological characteristic indicator, the first region can also be called the key region, and the second region can also be called the non-key region.

[0016] In a blood pressure measurement scenario, the degree of correlation between the first region and blood pressure is greater than that between the second region and blood pressure.

[0017] Performing different noise reduction processes on the first region and the second region can generate information related to physiological characteristic indicators through the first region and information related to time series through the second region. For example, in a blood pressure measurement scenario, an electronic device can generate waveform information related to blood pressure values through the first region and waveform information related to time series through the second region, and then generate the required PPG signal.

[0018] In this way, the information generated through the first region includes information related to physiological characteristic indicators, and the information generated through the second region is complete, which is beneficial to improving the accuracy of the generated information.

[0019] In a possible implementation manner, the forward process of the second model includes a first stage and a second stage. In the first stage, noise is added to the input information, and in the second stage, noise is added to the first region of the input information. The first stage and the second stage are executed in parallel, and the information for the backward process of training the second model is obtained through the first stage and the second stage. The correlation intensity between the first region and the physiological characteristic indicator corresponding to the input information is greater than that between the second region and the physiological characteristic indicator corresponding to the input information, and the second region is the region other than the first region.

[0020] The forward process of the second model can generate the information for the backward process of training the second model. In some implementations, the forward process of the second model can include a first stage and a second stage. The first stage can add noise to the whole of the input information, and the second stage can add noise to the first region of the input information.

[0021] In the embodiments of the present application, the forward process of the second model can refer to Figure 5 the forward process.

[0022] In this way, different noises can be added to the first region and the second region, so that when training the backward process of the second model according to this information later, it can be realized that the backward process of the second model can perform different processes on the first region and the second region.

[0023] In a possible implementation manner, during the process of adding noise to the input information in the first stage, the first noise-added information at time t is obtained by adding noise to the input information t times, and the first noise-added information satisfies the following formula:

[0024]

[0025]

[0026] Among them, x t is the first noise-added information, x0 is the input information, ε is the noise, and q(x t |x0) represents the conditional probability distribution of the forward process, is the mean value, is the variance, is a preset constant, and I is the identity matrix.

[0027] In some implementations, ε can be random noise. In this way, it is beneficial to add noise to the overall input information.

[0028] In a possible implementation manner, noise is added to the input information in the second stage. The second noise-added information at time t is obtained by adding noise to the first region of the input information t times. The second noise-added information at time t satisfies the following formula:

[0029]

[0030] Among them, is the second noise-added information, x0 is the input information, m is the number of the first regions included in the input information, μ is the range of the first region, ε is the noise, and q(x t |x0) represents the conditional probability distribution of the forward process, is the mean value, is the variance, is a preset constant, and I is the identity matrix.

[0031] In some implementations, ε can be random noise. In this way, it is beneficial to add noise to the first region.

[0032] In a possible implementation manner, the reverse process of the second model is trained in the following way: taking the output of the forward process of the second model as the input of the reverse process of the source model; performing different noise reduction processes on the first region and the second region of the input of the reverse process of the source model to obtain the output of the reverse process of the source model; obtaining the reverse process of the second model when the loss function converges, and the loss function is the difference between the output of the reverse process of the source model and the input of the forward process of the second model;

[0033] Among them, the input of the forward process of the second model includes the physiological characteristic information of multiple users and the physiological characteristic indicators corresponding to the physiological characteristic information of multiple users; the loss function is related to the loss corresponding to the first region, the loss corresponding to the second region, and the physiological characteristic information of multiple users.

[0034] In the embodiments of the present application, the reverse process of the second model can refer to Figure 5 the reverse process of

[0035] The loss function can include two parts. The first part is to denoise the first region, and the loss is concentrated on the parameters of the first region. The purpose is to learn the fine-grained information of this region and enhance the information of this region. The second part is to denoise the second region, and the purpose is to capture the global temporal information of the entire information. The loss function uses the physiological characteristic information of multiple users as a constraint, which is beneficial to learning the unique information of different users. In this way, it is beneficial to improve the accuracy of the information generated in the reverse process of the second model.

[0036] In a possible implementation, the loss function satisfies the following formula:

[0037]

[0038] where, is the loss function, E[*] is the expectation of *, λ1 is the weight of the loss corresponding to the first region, λ2 is the weight of the loss corresponding to the second region, λ1 and λ2 are preset constants, μ is the range of the first region, ε is the noise, is the noise of the first region predicted by the source model, is the noise of the second region predicted by the source model, m is the number of the first regions included in the physiological characteristic information of multiple users, x t is the first noise-added information at time t in the first stage of the forward process, is the second noise-added information at time t in the second stage of the forward process, and c is the physiological characteristic information of multiple users. In this way, it is beneficial to implement the training of the reverse process of the second model.

[0039] In a possible implementation, the first information further includes one or more of the gender of the first user, the age of the first user, or the body mass index BMI of the first user.

[0040] Information such as the gender of the first user, the age of the first user, or the body mass index BMI of the first user will all affect the physiological characteristic information of the first user. Taking these information as constraint conditions is beneficial to enabling the second model to generate more accurate information.

[0041] Optionally, the above first information is input into the second model to obtain the output of the second model, which may include: encoding the first information to obtain the encoded first information, and inputting the encoded first information into the second model to obtain the output of the second model.

[0042] The present application embodiment does not limit the encoding method. In the present application embodiment, the encoding method may refer to Figure 4 the Transformer network structure shown, and encode the first information through the Transformer network structure.

[0043] The first information may include text and / or numbers. Encoding the text in the first information can convert the text into an array, facilitating subsequent input into the second model. Encoding the numbers in the first information can convert the numbers into numerical values in a unified dimension, facilitating subsequent input into the second model.

[0044] In a possible implementation, the method further includes: sending the first information to the server; receiving the updated first model from the server. The server may be deployed with the second model. After receiving the first information, the first information can be input into the second model to obtain the output of the second model. The output of the second model is used as the physiological characteristic information sample of the first user to update the first model, obtaining the updated first model. In this way, there is no need for the electronic device to update the first model, which helps to save the power consumption of the first model.

[0045] In a possible implementation, the physiological characteristic information of the first user includes the photoplethysmography (PPG) signal, and the physiological characteristic index of the first user includes the blood pressure value. In this way, the embodiments of the present application are applicable to the blood pressure measurement scenario.

[0046] In a second aspect, an embodiment of the present application provides a measuring device for physiological characteristic indexes. The measuring device for physiological characteristic indexes may be an electronic device, or a chip or a chip system within the electronic device. The measuring device for physiological characteristic indexes may include an acquisition unit and a processing unit. When the measuring device for physiological characteristic indexes is an electronic device, the processing unit may be a processor. The measuring device for physiological characteristic indexes may further include a storage unit, and the storage unit may be a memory. The storage unit is used to store instructions, and the processing unit executes the instructions stored in the storage unit to enable the electronic device to implement a method for measuring a physiological characteristic index described in the first aspect or any possible implementation manner of the first aspect. When the measuring device for physiological characteristic indexes is a chip or a chip system within the electronic device, the processing unit may be a processor. The processing unit executes the instructions stored in the storage unit to enable the electronic device to implement a method for measuring a physiological characteristic index described in the first aspect or any possible implementation manner of the first aspect. The storage unit may be a storage unit within the chip (e.g., register, cache, etc.), or a storage unit outside the chip within the electronic device (e.g., read-only memory, random access memory, etc.).

[0047] Exemplarily, an acquisition unit is configured to acquire first information of a first user, where the first information includes physiological characteristic information of the first user. When acquiring the first information, a first model is preset in the electronic device; a processing unit is configured to measure physiological characteristic indicators of the first user based on the updated first model, and the physiological characteristic indicators of the first user include indicators for reflecting the physiological functions of the first user; where the first model is updated in the following manner: inputting the first information into a second model to obtain an output of the second model, and the output of the second model is multiple physiological characteristic information of the first user and physiological characteristic indicators corresponding to the multiple physiological characteristic information; the first information in the second model is a constraint condition; using the output of the second model as a physiological characteristic information sample of the first user to update the first model to obtain the updated first model.

[0048] In a possible implementation manner, the second model includes a reverse process and a forward process. The reverse process of the second model is configured to generate multiple physiological characteristic information of the first user and physiological characteristic indicators corresponding to the multiple physiological characteristic information with the first information as a constraint condition; the forward process of the second model is configured to obtain information for training the reverse process of the second model.

[0049] In a possible implementation manner, the reverse process of the second model is configured to perform different noise reduction processes on a first region and a second region of the first information to generate multiple physiological characteristic information of the first user and physiological characteristic indicators corresponding to the multiple physiological characteristic information. The second region is a region other than the first region, and the correlation degree between the first region and the physiological characteristic indicators is greater than the correlation degree between the second region and the physiological characteristic indicators.

[0050] In a possible implementation manner, the forward process of the second model includes a first stage and a second stage. Noise is added to the input information in the first stage, and noise is added to the first region of the input information in the second stage. The first stage and the second stage are executed in parallel, and information for training the reverse process of the second model is obtained through the first stage and the second stage. The correlation intensity between the first region and the physiological characteristic indicators corresponding to the input information is greater than the correlation intensity between the second region and the physiological characteristic indicators corresponding to the input information. The second region is a region other than the first region.

[0051] In a possible implementation manner, in the process of adding noise to the input information in the first stage, the first noise-added information at time t is obtained by adding noise to the input information t times, and the first noise-added information satisfies the following formula:

[0052]

[0053]

[0054] where, x tis the first noisy information, x0 is the input information, ε is the noise, q(x t |x0) represents the conditional probability distribution of the forward process, is the mean, is the variance, is a pre-set constant, and I is the identity matrix.

[0055] In a possible implementation, noise is added to the input information in the second stage. The second noisy information at time t is obtained by adding noise to the first region of the input information t times. The second noisy information at time t satisfies the following formula:

[0056]

[0057]

[0058] where, is the second noisy information, x0 is the input information, m is the number of the first regions included in the input information, μ is the range of the first region, ε is the noise, q(x t |x0) represents the conditional probability distribution of the forward process, is the mean, is the variance, is a pre-set constant, and I is the identity matrix.

[0059] In a possible implementation, the reverse process of the second model is trained in the following way: taking the output of the forward process of the second model as the input of the reverse process of the source model; performing different noise reduction processes on the first region and the second region of the input of the reverse process of the source model to obtain the output of the reverse process of the source model; when the loss function converges, obtaining the reverse process of the second model, and the loss function is the difference between the output of the reverse process of the source model and the input of the forward process of the second model;

[0060] where, the input of the forward process of the second model includes the physiological characteristic information of multiple users and the physiological characteristic indexes corresponding to the physiological characteristic information of multiple users; the loss function is related to the loss corresponding to the first region, the loss corresponding to the second region, and the physiological characteristic information of multiple users.

[0061] In a possible implementation, the loss function satisfies the following formula:

[0062]

[0063] where, Let \(L\) be the loss function, \(E[\cdot]\) be the expectation of \(\cdot\), \(\lambda_1\) be the weight of the loss corresponding to the first region, \(\lambda_2\) be the weight of the loss corresponding to the second region, \(\lambda_1\) and \(\lambda_2\) be preset constants, \(\mu\) be the range of the first region, \(\epsilon\) be the noise, is the noise of the first region predicted by the source model, is the noise of the second region predicted by the source model, \(m\) is the number of physiological characteristic information of multiple users that includes the first region, \(x\) t is the first noise addition information at time \(t\) in the first stage of the forward process, is the second noise addition information at time \(t\) in the second stage of the forward process, \(c\) is the physiological characteristic information of multiple users.

[0064] In a possible implementation manner, the first information further includes one or more of the gender of the first user, the age of the first user, or the body mass index BMI of the first user.

[0065] In a possible implementation manner, the above device further includes a transceiver unit, and the transceiver unit is configured to: send the first information to the server; receive the updated first model from the server.

[0066] In a possible implementation manner, the physiological characteristic information of the first user includes a photoplethysmography PPG signal, and the physiological characteristic index of the first user includes a blood pressure value.

[0067] In a third aspect, an embodiment of the present application provides an electronic device, including a processor and a memory, the memory is used to store code instructions, and the processor is used to run the code instructions to execute the method described in the first aspect or any possible implementation manner of the first aspect.

[0068] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program or instruction is stored. When the computer program or instruction runs on a computer, the computer is caused to execute the method described in the first aspect or any possible implementation manner of the first aspect.

[0069] In a fifth aspect, an embodiment of the present application provides a computer program product including a computer program. When the computer program runs on a computer, the computer is caused to execute the method described in the first aspect or any possible implementation manner of the first aspect.

[0070] Sixth aspect, the present application provides a chip or a chip system, which includes at least one processor and a communication interface. The communication interface and the at least one processor are interconnected by a line. The at least one processor is configured to run a computer program or instruction to execute the method described in the first aspect or any possible implementation manner of the first aspect. Among them, the communication interface in the chip can be an input / output interface, a pin, a circuit, etc.

[0071] In a possible implementation, the chip or chip system described above in the present application further includes at least one memory, and instructions are stored in the at least one memory. The memory can be a storage unit inside the chip, such as a register, a cache, etc., or it can be a storage unit of the chip (such as a read-only memory, a random access memory, etc.).

[0072] It should be understood that the second aspect to the sixth aspect of the present application correspond to the technical solutions of the first aspect of the present application, and the beneficial effects obtained by each aspect and the corresponding feasible implementation manners are similar, and will not be elaborated herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 It is a schematic diagram for a PPG module to obtain a PPG signal;

[0074] Figure 2 It is a schematic diagram of the waveform of a PPG signal;

[0075] Figure 3 It is a schematic diagram of the hardware structure of a wearable device provided by an embodiment of the present application;

[0076] Figure 4 It is a schematic diagram of a method for measuring a physiological characteristic index provided by an embodiment of the present application;

[0077] Figure 5 It is a schematic diagram of training a Diffusion model provided by an embodiment of the present application;

[0078] Figure 6 It is a schematic diagram of calibrating a model provided by an embodiment of the present application;

[0079] Figure 7 It is a schematic diagram of an electroencephalogram recognition scenario provided by an embodiment of the present application;

[0080] Figure 8 It is a schematic diagram of another method for measuring a physiological characteristic index provided by an embodiment of the present application;

[0081] Figure 9 It is a schematic diagram of the structure of a chip provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0082] To facilitate a clear description of the technical solutions of the embodiments of the present application, the following explanations are provided first:

[0083] In the embodiments of the present application, terms such as "first" and "second" are used to distinguish identical or similar items with basically the same functions and roles. For example, the first model and the second model are only used to distinguish different models, and do not limit their sequence. Those skilled in the art can understand that terms such as "first" and "second" do not limit the quantity and execution order, and "first", "second", etc. do not necessarily mean different.

[0084] It should be noted that in the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "exemplary" or "for example" is intended to present relevant concepts in a specific manner.

[0085] In the embodiments of the present application, "at least one" means one or more, and "a plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (item)" or its similar expression refers to any combination of these items, including any combination of single item (item) or plural items (items). For example, at least one (item) of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, c can be single or multiple.

[0086] Currently, wearable devices can have blood pressure measurement functions. However, wearable devices are prone to inaccurate blood pressure measurement.

[0087] To better understand the blood pressure measurement function of wearable devices, the principle of blood pressure measurement by wearable devices is introduced below.

[0088] A wearable device can include a photoplethysmography (PPG) module. The PPG module is used to collect the user's PPG signal, and the user's PPG signal is input into a blood pressure prediction model to obtain the output of the blood pressure prediction model, and the output of the blood pressure prediction model is the user's blood pressure value. In this way, the measurement of the user's blood pressure is realized.

[0089] Since the PPG module is portable and inexpensive, it is a common method for measuring blood pressure by using the PPG module to collect the user's PPG signal and measure the user's blood pressure.

[0090] The following will introduce the principle of obtaining the PPG signal by the PPG module in conjunction with Figure 1 the following content.

[0091] Exemplarily, Figure 1 a schematic diagram of a PPG module for obtaining a PPG signal is shown. As Figure 1 shown, the PPG module may include a light emitting diode (LED) and a photo diode (PD). When the user's finger presses the PPG module, the PPG module can use the LED to irradiate the human skin, and then use the PD to measure the change in the intensity of the reflected light caused by blood flow to obtain a periodic waveform, that is, the PPG signal. This PPG signal can contain information about the pulsatile change in blood volume during the cardiac cycle. When the heart contracts, the blood volume is the largest, and when the heart relaxes, the diastolic volume is the smallest. Therefore, the systolic blood pressure and diastolic blood pressure can be obtained through the relationship between the volume pulse blood flow information and blood pressure.

[0092] Figure 2 A schematic diagram of the waveform of a PPG signal is shown. As Figure 2 shown, the peak of the PPG signal corresponds to heart contraction, and the trough of the PPG signal corresponds to heart relaxation. The wearable device can calculate the systolic blood pressure and diastolic blood pressure based on the peaks and troughs of the user's PPG signal and output the user's blood pressure value.

[0093] Currently, the PPG signal is greatly affected by human differences, which will cause the blood pressure prediction model in the wearable device to be unable to accurately identify the user's blood pressure based on the user's blood pressure, resulting in inaccurate blood pressure measurement.

[0094] If the PPG signal and blood pressure value of the user during the use of the wearable device are collected as training samples to calibrate the blood pressure prediction model, the time is relatively long, and during the collection process, the obtained blood pressure value is inaccurate, which will affect the accuracy of calibrating the blood pressure prediction model.

[0095] In view of this, embodiments of the present application provide a method for measuring physiological characteristic indexes and related devices, which can generate a PPG signal related to a user and a blood pressure value corresponding to the PPG signal based on the user's personalized information, and use these PPG signals and the blood pressure values corresponding to the PPG signals as training samples to calibrate a blood pressure prediction model to obtain a calibrated model. In this way, the blood pressure prediction model can be adjusted according to individual differences, and the calibrated model is used to measure the blood pressure value of the user, which is beneficial to improving the accuracy of blood pressure measurement. Among them, the user's personalized information may include the user's PPG signal, gender, age, body mass index (BMI), exercise intensity, altitude, and other information.

[0096] The method provided by the embodiments of the present application can be applied to an electronic device with a blood pressure measurement function. The electronic device may include a wearable device, such as a smart watch, a smart bracelet, or an augmented reality (AR) wearable device, etc. Among them, the wearable device can also be called a wearable intelligent device, which is the general term for devices developed by applying wearable technology to intelligent design of daily wear, such as glasses, gloves, watches, clothing, shoes, etc. A wearable device is a portable device that is directly worn on the body or integrated into the user's clothes or accessories. A wearable device is not only a hardware device, but also realizes powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable intelligent devices include those with complete functions and large sizes that can realize complete or partial functions without relying on a smart phone, such as smart watches or smart glasses, etc., and those that only focus on a certain type of application function and need to cooperate with other devices such as smart phones, such as various smart bracelets and smart jewelry for physical sign monitoring.

[0097] To better understand the embodiments of the present application, the hardware structure of the electronic device will be described first.

[0098] Exemplarily, Figure 3 FIG. shows a schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present application. As Figure 3 shown, the electronic device may include a main control module 110, a pressure sensor 120, a PPG module 130, a gyroscope 140, a touch sensor 150, a display screen 160, a communication module 170, a storage module 180, and an acceleration sensor 190. Each module can be connected through a bus or other means, and the embodiments of the present application do not limit this.

[0099] It can be understood that the structure illustrated in the embodiments of the present invention does not constitute a specific limitation on the electronic device. In some other embodiments of the present application, the electronic device may further include more or fewer components than those illustrated, or combine certain components, or split certain components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0100] The main control module 110 may include a blood pressure calculation module 111, and the blood pressure calculation module 111 may be used to identify the blood pressure value based on the user's PPG signal. The main control module 110 may further include one or more processors for supporting other functions of the electronic device.

[0101] The pressure sensor 120 may be used to collect the pressure data of the electronic device. In one possible implementation, the pressure sensor 120 may be used to collect the pressure value of a finger pressing the electronic device.

[0102] The PPG module 130 may include a PD module 131 and an LED module 132. The PD module 131 may include one or more PDs, which are not limited in the embodiments of the present application. The PD may be used to receive an optical signal and process the optical signal into an electrical signal. For example, in the scenario of measuring blood pressure, the PD may receive the optical signal reflected back through the fingertip tissue and process the signal into an electrical signal. The LED module 132 may include one or more LEDs, which may be used as a measurement light source in the scenario of measuring blood pressure.

[0103] The gyroscope 140 may be used to obtain the angular velocity data of the electronic device.

[0104] The touch sensor 150 may be referred to as a "touch control device". The touch sensor 150 may be disposed in the display screen 160, and the touch sensor 150 and the display screen 160 together form a touch screen, which may also be referred to as a "touch control screen". The touch sensor 150 may be used to detect a touch operation acting thereon or nearby. The display screen 160 may be used to display images, videos, controls, text information, and so on.

[0105] The communication module 170 may include a Bluetooth communication module and a WLAN communication module. In some embodiments, the WLAN communication module may be integrated with other communication modules (for example, the Bluetooth communication module). The WLAN communication module and / or the Bluetooth module may transmit signals to detect and scan devices near the electronic device, such as a terminal device or a server, so that the electronic device can establish a wireless communication connection with the terminal device or the server through one or more wireless communication technologies such as WLAN and Bluetooth, and perform data transmission and data reception based on the wireless communication connection.

[0106] The storage module 180 is coupled to the main control module 110 and is used to store various software programs and / or multiple sets of instructions. In a specific implementation, the storage module 180 may include a volatile memory, such as a random access memory (RAM); it may also include a non-volatile memory, such as a ROM, a flash memory, a hard disk drive (HDD), or a solid-state drive (SSD); the storage module 180 may also include a combination of the above types of memories. The storage module 180 may store some program codes so that the main control module 110 can call the program codes stored in the storage module 180 to implement the method of the embodiment of the present application.

[0107] The acceleration sensing module 190 may be an acceleration sensor and can be used to detect the magnitude of the acceleration of the electronic device in various directions (generally three axes).

[0108] To save the power consumption of the electronic device, the method provided by the embodiment of the present application may also be executed by a server. The server may obtain the personalized information of the user from the electronic device, and based on the personalized information of the user, generate a PPG signal related to the user and the blood pressure value corresponding to the PPG signal, use these PPG signals and the blood pressure values corresponding to the PPG signals as training samples to calibrate the blood pressure prediction model, obtain a calibrated model, and send the calibrated model to the electronic device. The electronic device may use the calibrated model to measure the blood pressure value of the user.

[0109] Among them, there are various possible implementation manners for the server to obtain the personalized information of the user.

[0110] In some examples, the server may instruct the electronic device to upload the personalized information of the user. The electronic device may, based on the instruction of the server, collect the PPG signal of the user for a certain period of time (for example, 30 milliseconds or 1 minute), and obtain one or more pieces of information such as the gender, age, BMI, exercise intensity, and altitude of the user through the human-computer interaction interface, and upload them to the server.

[0111] In this way, the server can control the timing of obtaining the personalized information of the user, with stronger flexibility.

[0112] In some other examples, when the user first starts the device or first uses the device to measure the PPG signal, the electronic device may collect the user's PPG signal within a certain period of time (for example, 30 milliseconds or 1 minute), and obtain one or more pieces of information including the user's gender, age, BMI, exercise intensity, and altitude through the human-computer interaction interface, and upload them to the server.

[0113] In this way, the electronic device can actively upload personalized information, which is beneficial to reducing the information interaction between the server and the electronic device.

[0114] Next, taking the electronic device as the execution subject, the method for measuring physiological characteristic indexes provided in the embodiments of the present application will be introduced.

[0115] Exemplarily, Figure 4 shows a schematic diagram of a method for measuring physiological characteristic indexes provided in an embodiment of the present application. As Figure 4 shown, the method may include the following steps:

[0116] S401. Obtain the user's personalized information.

[0117] There may be various situations for the user's personalized information.

[0118] In some examples, the user's personalized information may include the user's PPG signal, gender, age, and BMI and other information.

[0119] In some other examples, the user's personalized information may include the user's PPG signal, gender, age, BMI, and exercise intensity and other information.

[0120] In still some other examples, the user's personalized information may include the user's PPG signal, gender, age, BMI, exercise intensity, and altitude and other information.

[0121] The PPG signal is greatly affected by individual differences and is easily affected by gender, age, BMI, exercise intensity, and altitude. These information may vary due to different users, so these information can be called the user's personalized information.

[0122] In some implementations, the electronic device may collect the user's PPG signal within a certain period of time (for example, 30 milliseconds or 1 minute), and obtain one or more pieces of information including the user's gender, age, BMI, exercise intensity, and altitude through the human-computer interaction interface.

[0123] S402. Encode the user's personalized information to obtain the encoded personalized information.

[0124] Encode the personalized information of the user to obtain the encoded personalized information, which can be understood as converting the personalized information of the user into encoded information.

[0125] Exemplarily, as Figure 4 shown, the electronic device can use the Transformer network structure to convert the personalized information into encoded information.

[0126] The personalized information of the user can include text and / or numbers. Encoding the text in the personalized information of the user can convert the text into an array, which is convenient for subsequent input into the model; encoding the numbers in the personalized information of the user can convert the numbers into numerical values in a unified dimension, which is convenient for subsequent input into the model.

[0127] S403. Input the encoded personalized information and random noise into the Diffusion model to obtain the output of the Diffusion model. The output of the Diffusion model includes multiple PPG signals of the user and the blood pressure values corresponding to the multiple PPG signals.

[0128] The role of the Diffusion model is to generate a large number of required PPG signals and the blood pressure values corresponding to the PPG signals given random noise and conditions. Among them, the random noise has no meaning.

[0129] As Figure 4 shown, in the embodiment of the present application, the input of the Diffusion model includes the encoded personalized information and random noise. The Diffusion model can use the encoded personalized information as a constraint condition and perform noise reduction T times based on the random noise using the reverse process to generate multiple PPG signals related to the user and the blood pressure values corresponding to the multiple PPG signals.

[0130] S404. Use the output of the Diffusion model as a training sample to train the blood pressure prediction meta-model to obtain a calibration model.

[0131] The blood pressure prediction meta-model can also be called a blood pressure prediction model, and the embodiment of the present application does not limit this. The output of the Diffusion model includes multiple PPG signals of the user and the blood pressure values corresponding to the multiple PPG signals. The electronic device can input the multiple PPG signals of the user into the blood pressure prediction meta-model to obtain the output of the blood pressure prediction meta-model, calculate the loss function through the output of the blood pressure prediction meta-model and the blood pressure values corresponding to the multiple PPG signals, and obtain the calibration model when the loss function converges.

[0132] The measurement method of physiological characteristic indexes provided by the embodiments of the present application can quickly generate a PPG signal related to the user and the blood pressure value corresponding to the PPG signal through the user's personalized information, without actually collecting the PPG signal and blood pressure value of the user, which can shorten the collection duration. At the same time, using the generated PPG signals and the blood pressure values corresponding to the PPG signals as training samples to calibrate the blood pressure prediction model to obtain a calibrated model. In this way, the blood pressure prediction model can be adjusted according to individual differences, and measuring the blood pressure value of the user through the calibrated model is beneficial to improving the accuracy of blood pressure measurement.

[0133] In the above Figure 4 In the method shown, the role of the Diffusion model is to generate a large number of required PPG signals and the blood pressure values corresponding to the PPG signals given random noise and conditions. The training process of the Diffusion model will be described in detail below.

[0134] Exemplarily, Figure 5 A schematic diagram showing the training of a Diffusion model is shown. The model before training can be called the source model, and the model after training can be called the Diffusion model. As Figure 5 shown, the forward process of the source model is to gradually add noise to the input signal along with the Markov process, and the reverse process of the source model is to gradually denoise the signal with added noise. As Figure 5 shown, the forward process of the source model can add noise T times, and the reverse process can denoise T times.

[0135] In the embodiments of the present application, in the forward process of the source model, the input original PPG signals can include PPG signals at various ages, different genders, different BMIs, different exercise intensities, and various altitudes. These original PPG signals can all correspond to blood pressure values, and can also correspond to one or more pieces of information such as age, gender, BMI, exercise intensity, and altitude. It can be understood that Figure 5 The processing process of the PPG signal is mainly shown, and the processing process of the corresponding information is similar.

[0136] The forward process of the source model is divided into two stages. The first stage can gradually add noise to the entire original PPG signal and its corresponding information (such as blood pressure value, age, gender, and BMI, etc.), and the second stage can divide the original PPG signal beat by beat to find the key areas in the PPG signal related to blood pressure and add noise to the key areas. For example, in Figure 5 the area framed by the rectangular wireframe is used to represent the key area.

[0137] In some implementations, the electronic device can find the key regions by calculating the R peaks, and the key regions can include the systolic peak, dicrotic notch, and diastolic peak region, etc. in the PPG signal.

[0138] The forward process of the source model can gradually add noise with the Markov process through these two stages to obtain the PPG signal with added noise.

[0139] As Figure 5 shown, in the first stage of the forward process, x0 can represent the original PPG signal, ε can represent the noise. If ε is random noise, then ε satisfies N(0, I). The forward process continuously adds Gaussian noise, and x t can represent the PPG signal with added noise at time t, and x t satisfies the following formula:

[0140]

[0141]

[0142] where q(x t |x0) represents the conditional probability distribution of the forward (noise-added) process, which can be a Gaussian distribution, and the mean can be and the variance is which are pre-set constants, and I is the identity matrix, and the dimension of I is not limited.

[0143] The electronic device can calculate x t from the original PPG signal x0 according to the Gaussian distribution q(x t .

[0144] In the second stage of the forward process, m can represent the number of segments obtained by dividing the entire original PPG signal beat by beat, and μ can represent an indicator function used to specify the key region range. The forward process continuously adds Gaussian noise, can represent the PPG signal with added noise at time t, satisfies the following formula:

[0145]

[0146] The regions in the original PPG signal other than the key regions can all be called non-key regions. In the reverse process of the source model, the PPG signal with added noise is used as the input of the reverse process of the source model.

[0147] During the reverse process of the source model, according to the divided key regions and non-key regions, progressive noise reduction is performed on the key regions and non-key regions respectively to obtain the output of the reverse process of the source model. The output of the reverse process of the source model is the PPG signal generated based on the input signal and the corresponding blood pressure value. It can be understood that if the noise addition methods for the key regions and non-key regions are different, the noise reduction methods can also be different.

[0148] The electronic device can calculate the loss function based on the output of the reverse process of the source model, the original PPG signal, and its corresponding information. When the loss function converges, the training of the reverse process of the source model is completed, and the Diffusion model is obtained. The probability distribution parameters at each step in the reverse process of the Diffusion model are all trained, and these parameters can be used for the reverse process to generate the PPG signal.

[0149] As Figure 5 shown, during the reverse process, the loss function satisfies the following formula:

[0150]

[0151] Among them, is the loss function, c is used to represent the constraint condition. In the embodiments of the present application, c is the user's personalized information, including the original PPG signal and its corresponding information, x t is the noisy PPG signal at time t in the first stage of the forward process, μ is used to specify the key region range, is the noisy PPG signal at time t in the second stage of the forward process, ε is the noise added during the forward process, ε θ is the noise of the key region predicted by the model, is the noise of the non-key region predicted by the model, λ1 is the loss coefficient of the key region, λ2 is the loss coefficient of the non-key region, λ1 and λ2 are preset constants, and E[*] is the expectation of *.

[0152] It can be understood that the loss function can include two parts. The first part is to reduce noise in the key regions, and the loss is concentrated on the key region parameters, aiming to learn the fine-grained information of this region and enhance the waveform of this region. The second part is to reduce noise in the non-key regions, aiming to capture the global timing information of the entire PPG signal.

[0153] In this way, the loss function can be related to the loss of the key regions and the loss of the non-key regions, and the weight of the loss of the key regions can be greater than the loss of the non-key regions. In this way, better learning can be performed on the key regions, which is beneficial to making the generated PPG signal more accurate.

[0154] In the embodiment of the present application, the key area focuses on a small part of the entire PPG signal waveform, which is strongly correlated with the blood pressure value, that is, emphasizing details (called fine-grained information), which can enhance the waveform of the key area. The non-key area accounts for a large proportion and runs through the entire PPG signal, corresponding to the timing waveform information of the PPG signal. In this way, different degrees of noise addition and noise reduction training are performed on the key area and the non-key area, which is conducive to generating a PPG signal with an accurate blood pressure value label, that is, it can generate local key information that is strongly correlated with the blood pressure value, and can generate the global timing information of the PPG signal.

[0155] When training the Diffusion model, in order to improve the training efficiency, representative PPG signals can be selected through Gaussian distribution, and these representative PPG signals can be used as training samples to train the Diffusion model. In this way, the number of training samples is relatively small and representative, which can improve the training efficiency while taking into account the accuracy of the Diffusion model.

[0156] In the embodiment of the present application, the Gaussian distribution is used to select a representative PPG signal because the statistical laws of nature consider the Gaussian distribution to be the most common, and a large amount of natural data is also most likely to obey the Gaussian distribution.

[0157] Therefore, the method for measuring physiological characteristic indicators provided in the embodiment of the present application may include: the electronic device obtains personalized information of different users, selects representative samples through Gaussian distribution, and uses these representative samples as training samples (or training sets) to train the Diffusion model. The electronic device can also collect the personalized information of the user, input the personalized information of the user into the Diffusion model, obtain multiple PPG signals related to the user and corresponding blood pressure values, and train the blood pressure prediction metamodel based on these multiple PPG signals related to the user and corresponding blood pressure values ​​to obtain a calibration model. The calibration model can predict the blood pressure of the user relatively accurately.

[0158] For example, Figure 6 A schematic diagram of calibrating the model is shown. Figure 6 As shown in FIG. 1 , individual test data is used to train the Diffusion model. Statistical data is used to represent personalized information of different users. The Gaussian distribution can be used to obtain representative samples from the statistical data, and these samples can be used to train the Diffusion model.

[0159] Individual 1 may represent User 1, Individual 2 may represent User 2, and Individual N may represent User N. Each electronic device used by Individuals 1 to N can utilize the measurement method of physiological characteristic indexes provided in the embodiments of the present application to obtain a calibration model for identifying the blood pressure value of this individual.

[0160] As Figure 6 shown, the personalized data of Individual 1 is input into the Diffusion model, and multi-interval blood pressure data of Individual 1 can be obtained. Among them, the personalized data of Individual 1 can also be referred to as the personalized information of Individual 1, and the embodiments of the present application do not limit this. The multi-interval blood pressure data of Individual 1 is used to represent multiple PPG signals related to Individual 1 and the corresponding blood pressure values.

[0161] Based on the multi-interval blood pressure data of Individual 1, the blood pressure prediction meta-model is calibrated to obtain the calibration model of Individual 1, and this calibration model can identify the blood pressure value of Individual 1 based on the PPG signal of Individual 1.

[0162] The personalized data of Individual N is input into the Diffusion model, and multi-interval blood pressure data of Individual N can be obtained. Among them, the personalized data of Individual N can also be referred to as the personalized information of Individual N, and the embodiments of the present application do not limit this. The multi-interval blood pressure data of Individual N is used to represent multiple PPG signals related to Individual N and the corresponding blood pressure values.

[0163] Based on the multi-interval blood pressure data of Individual N, the blood pressure prediction meta-model is calibrated to obtain the calibration model of Individual N, and this calibration model can identify the blood pressure value of Individual N based on the PPG signal of Individual N.

[0164] It can be seen from this that the models for users to identify blood pressure values used by different individuals are different, which is beneficial to improving the accuracy of blood pressure identification.

[0165] It should be noted that the above is described by taking the electronic device as the execution subject as an example. In Figure 6 this, if the execution subject is a server, the server can obtain the personalized information of different users, select representative samples through Gaussian distribution, and use these representative samples as training samples to train the Diffusion model. The server can receive the personalized information of the user collected by the electronic device used by Individual 1 or Individual N, input the personalized information of the user into the Diffusion model, obtain multiple PPG signals related to the user and the corresponding blood pressure values, and train the blood pressure prediction meta-model based on these multiple PPG signals related to the user and the corresponding blood pressure values to obtain a calibration model, and send the calibration model to the electronic device used by Individual 1 or Individual N. This calibration model can be used to predict the blood pressure of this user.

[0166] The measurement method of physiological characteristic indexes provided by the embodiments of the present application is applicable not only to blood pressure prediction scenarios, but also to other scenarios. For example, it can be applicable to other scenarios that require personalized customization of models for physiological indexes or human-computer interaction, etc., for user individual signals, and scenarios with model migration or fine-tuning requirements, such as brain-computer interface (BCI), which realizes direct interaction between the brain and external devices through brain signals emitted by individuals.

[0167] The following takes the BCI scenario as an example for illustration.

[0168] Figure 7 Shows a schematic diagram of an electroencephalogram recognition scenario. As Figure 7 shown, 1) The electroencephalogram acquisition device can acquire electroencephalogram information, and the electroencephalogram acquisition device can transmit the electroencephalogram signal to the electroencephalogram processing device. 2) The electroencephalogram processing device can obtain the electroencephalogram signal from the electroencephalogram acquisition device and process the electroencephalogram signal. 3) The electroencephalogram processing device can extract features from the electroencephalogram signal. 4) The electroencephalogram processing device can also classify the electroencephalogram signal based on the extracted features. 5) The classification result of the electroencephalogram signal by the electroencephalogram processing device can be used in scenarios such as movement assistance, communication and interaction, entertainment / game control, etc. 6) Whether it is implemented in the scenario can generate feedback to the user, and the user's electroencephalogram signal can change based on the feedback.

[0169] In this scenario, the electroencephalogram processing device can use the electroencephalogram recognition model to extract features from the electroencephalogram signal and classify the electroencephalogram signal based on the extracted features. However, the electroencephalogram signal is greatly affected by human differences. If the electroencephalogram device is not calibrated for different users, the electroencephalogram recognition accuracy will be low.

[0170] The measurement method of physiological characteristic indexes provided by the embodiments of the present application can acquire the electroencephalogram signal of a user, input the electroencephalogram signal of this user into the Diffusion model, obtain multiple electroencephalogram signals and recognition results related to this user, and use these multiple electroencephalogram signals and recognition results related to this user as training samples to calibrate the electroencephalogram recognition model, obtaining a calibrated model, which is beneficial to improving the electroencephalogram recognition accuracy.

[0171] It should be noted that in the embodiments of the present application, the Diffusion model is taken as an example for illustration, and other generative models can also be used, such as autoregressive models, variational autoencoders, generative adversarial networks, etc. The embodiments of the present application do not make limitations in this regard.

[0172] The method provided by the embodiments of the present application is introduced above in combination with application scenarios. Next, the method provided by the embodiments of the present application will be introduced in combination with the step flow.

[0173] Exemplarily, Figure 8A schematic flowchart of a method for measuring a physiological characteristic index is shown. This method can be applied to an electronic device. As Figure 8 shown, the method may include the following steps:

[0174] S801. Obtain first information of a first user, where the first information includes physiological characteristic information of the first user. When obtaining the first information, a first model is preset in the electronic device.

[0175] S802. Measure the physiological characteristic index of the first user based on the updated first model. The physiological characteristic index of the first user includes an index for reflecting the physiological function of the first user. Wherein, the first model is updated in the following manner: input the first information into a second model to obtain the output of the second model. The output of the second model is multiple physiological characteristic information of the first user and physiological characteristic indexes corresponding to the multiple physiological characteristic information; the first information in the second model is a constraint condition; use the output of the second model as a physiological characteristic information sample of the first user to update the first model to obtain the updated first model.

[0176] The output of the second model can be used as a physiological characteristic information sample of the first user to update the first model to obtain the updated first model, and use the updated first model to measure the physiological characteristic index of the first user. This can make the updated first model more accurately identify the corresponding physiological characteristic index according to the physiological characteristic information of the first user, which is beneficial to improving the measurement accuracy.

[0177] Optionally, the second model includes a reverse process and a forward process. The reverse process of the second model is used to generate multiple physiological characteristic information of the first user and physiological characteristic indexes corresponding to the multiple physiological characteristic information with the first information as a constraint condition; the forward process of the second model is used to obtain information for training the reverse process of the second model. In this way, it is beneficial to meet the requirement of generating samples.

[0178] Optionally, the reverse process of the second model is used to perform different noise reduction processes on a first region of the first information and a second region of the first information to generate multiple physiological characteristic information of the first user and physiological characteristic indexes corresponding to the multiple physiological characteristic information. The second region is a region other than the first region, and the degree of correlation between the first region and the physiological characteristic index is greater than the degree of correlation between the second region and the physiological characteristic index. In this way, the information generated through the first region includes information related to the physiological characteristic index, and the complete information is generated through the second region, which is beneficial to improving the accuracy of the generated information.

[0179] Optionally, the forward process of the second model includes a first stage and a second stage. In the first stage, noise is added to the input information. In the second stage, noise is added to a first region of the input information. The first stage and the second stage are executed in parallel. Information for the reverse process of training the second model is obtained through the first stage and the second stage. The correlation strength between the first region and the physiological characteristic index corresponding to the input information is greater than the correlation strength between the second region and the physiological characteristic index corresponding to the input information. The second region is the region other than the first region.

[0180] In this way, different noises can be added to the first region and the second region, so that when training the reverse process of the second model based on this information subsequently, the reverse process of the second model can perform different processes on the first region and the second region.

[0181] Optionally, during the process of adding noise to the input information in the first stage, the first noise-added information at time t is obtained by adding noise to the input information t times. The first noise-added information satisfies the following formula:

[0182]

[0183]

[0184] where x t is the first noise-added information, x0 is the input information, ε is the noise, and q(x t |x0) represents the conditional probability distribution of the forward process. is the mean, is the variance, is a preset constant, and I is the identity matrix.

[0185] In some implementations, ε can be random noise. In this way, it is beneficial to add noise to the entire input information.

[0186] Optionally, when adding noise to the input information in the second stage, the second noise-added information at time t is obtained by adding noise to the first region of the input information t times. The second noise-added information at time t satisfies the following formula:

[0187]

[0188]

[0189] where is the second noise-added information, x0 is the input information, m is the number of first regions included in the input information, μ is the range of the first region, ε is the noise, and q(x t |x0) represents the conditional probability distribution of the forward process. is the mean, is the variance, is a preset constant, and I is the identity matrix. In this way, it is beneficial to add noise to the first region.

[0190] Optionally, the reverse process of the second model is trained as follows: taking the output of the forward process of the second model as the input of the reverse process of the source model; performing different noise reduction processes on the first region and the second region of the input of the reverse process of the source model to obtain the output of the reverse process of the source model; when the loss function converges, obtaining the reverse process of the second model, where the loss function is the difference between the output of the reverse process of the source model and the input of the forward process of the second model;

[0191] Among them, the input of the forward process of the second model includes the physiological characteristic information of multiple users and the physiological characteristic indicators corresponding to the physiological characteristic information of multiple users; the loss function is related to the loss corresponding to the first region, the loss corresponding to the second region, and the physiological characteristic information of multiple users. In this way, it is beneficial to improve the accuracy of the information generated by the reverse process of the second model.

[0192] Optionally, the loss function satisfies the following formula:

[0193]

[0194] Among them, is the loss function, E[*] is the expectation of *, λ1 is the weight of the loss corresponding to the first region, λ2 is the weight of the loss corresponding to the second region, λ1 and λ2 are preset constants, μ is the range of the first region, ε is the noise, is the noise of the first region predicted by the source model, is the noise of the second region predicted by the source model, m is the number of the first regions included in the physiological characteristic information of multiple users, x t is the first noise addition information at time t in the first stage of the forward process, is the second noise addition information at time t in the second stage of the forward process, and c is the physiological characteristic information of multiple users. In this way, it is beneficial to realize the training of the reverse process of the second model.

[0195] Optionally, the first information further includes one or more of the gender of the first user, the age of the first user, or the body mass index BMI of the first user. In this way, it is beneficial to make the second model generate more accurate information.

[0196] Optionally, the method further includes: sending the first information to the server; receiving the updated first model from the server. In this way, there is no need for the electronic device to update the first model, which is beneficial to saving the power consumption of the first model.

[0197] Optionally, the physiological characteristic information of the first user includes a photoplethysmogram (PPG) signal, and the physiological characteristic index of the first user includes a blood pressure value. In this way, the embodiments of the present application are applicable to the blood pressure measurement scenario.

[0198] It should be noted that the module names involved in the embodiments of the present application can all be defined as other names, as long as the functions of each module can be realized, and no specific restrictions are imposed on the module names.

[0199] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the embodiments of the present application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or refuse.

[0200] The measurement method of the physiological characteristic index in the embodiments of the present application has been described above. Next, the measurement device for the physiological characteristic index that executes the above method provided by the embodiments of the present application will be described. Those skilled in the art can understand that the method and the device can be combined and referenced with each other, and the relevant device provided by the embodiments of the present application can execute the steps in the above measurement method of the physiological characteristic index.

[0201] The embodiments of the present application provide a chip. Exemplarily, Figure 9 is a schematic structural diagram of a chip provided by the embodiments of the present application. As Figure 9 shown, the chip 90 includes one or more than two (including two) processors 901, a communication line 902, a communication interface 903, and a memory 904.

[0202] In some embodiments, the memory 904 stores the following elements: executable modules or data structures, or subsets thereof, or extended sets thereof.

[0203] The measurement method of the physiological characteristic index described in the embodiments of the present application can be applied to or implemented by the processor 901. The processor 901 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the measurement method of the physiological characteristic index can be completed by the integrated logic circuit of the hardware in the processor 901 or the instructions in the form of software. The above-mentioned processor 901 may be a general-purpose processor (e.g., a microprocessor or a conventional processor), a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates, transistor logic devices or discrete hardware components. The processor 901 can implement or execute various processing-related methods, steps and logic block diagrams disclosed in the embodiments of the present application.

[0204] The steps of the measurement method of the physiological characteristic index disclosed in combination with the embodiments of the present application can be directly implemented by the hardware decoding processor, or implemented by the combination of the hardware and software modules in the decoding processor. Among them, the software module can be located in a mature storage medium in the art such as a random access memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable read-only memory (EEPROM). This storage medium is located in the memory 904, and the processor 901 reads the information in the memory 904 and combines its hardware to complete the steps of the above method.

[0205] Communication can be carried out between the processor 901, the memory 904 and the communication interface 903 through the communication line 902.

[0206] In the above embodiments, the instructions stored in the memory for the processor to execute can be implemented in the form of a computer program product. Among them, the computer program product can be pre-written in the memory in advance, or downloaded and installed in the memory in the form of software.

[0207] The measurement method of the physiological characteristic index provided by the embodiments of the present application can be applied to an electronic device with processing functions. For the specific device form of the electronic device, etc., reference can be made to the above relevant descriptions, which will not be elaborated here.

[0208] An embodiment of the present application provides a terminal device, which includes a processor and a memory; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, so that the terminal device executes the above method.

[0209] An embodiment of the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the above method is implemented. The methods described in the above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. If implemented in software, the functions can be stored as one or more instructions or codes on a computer-readable medium or transmitted on a computer-readable medium. The computer-readable medium may include a computer storage medium and a communication medium, and may also include any medium that can transfer a computer program from one place to another. The storage medium can be any target medium accessible by a computer.

[0210] In a possible implementation, the computer-readable medium may include RAM, ROM, a compact disc read-only memory (CD-ROM), or other optical disc memories, magnetic disk memories, or other magnetic storage devices, or any other medium targeted at carrying or storing the required program code in the form of instructions or data structures and accessible by a computer. Moreover, any connection is properly referred to as a computer-readable medium. For example, if software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of the medium. As used herein, disk and optical disc include optical disc, laser disc, optical disc, digital versatile disc (DVD), floppy disk, and Blu-ray disc, where disks typically reproduce data magnetically, while optical discs use lasers to optically reproduce data. The above combinations should also be included within the scope of the computer-readable medium.

[0211] An embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is run, it causes the computer to execute the above method.

[0212] Embodiments of the present application are described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processing unit of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable devices to generate a machine, such that the instructions executed by the processing unit of the computer or other programmable data processing device generate means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0213] In the above specific embodiments, the purpose, technical solution, and beneficial effects of the present invention have been further described in detail. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solution of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for measuring a physiological characteristic index, characterized in that, Applied to an electronic device, the method includes: Obtain first information of a first user, where the first information includes physiological characteristic information of the first user. When obtaining the first information, a first model is preset in the electronic device; Measure physiological characteristic indicators of the first user based on the updated first model, where the physiological characteristic indicators of the first user include indicators for reflecting the physiological functions of the first user; Wherein, the first model is updated in the following manner: Input the first information into a second model to obtain the output of the second model, and the output of the second model is multiple physiological characteristic information of the first user and the physiological characteristic indicators corresponding to the multiple physiological characteristic information; the first information in the second model is a constraint condition; Use the output of the second model as a physiological characteristic information sample of the first user to update the first model to obtain the updated first model.

2. The method according to claim 1, wherein The second model includes a reverse process and a forward process. The reverse process of the second model is used to generate multiple physiological characteristic information of the first user and the physiological characteristic indicators corresponding to the multiple physiological characteristic information with the first information as a constraint condition; The forward process of the second model is used to obtain information for training the reverse process of the second model.

3. The method according to claim 2, wherein The reverse process of the second model is used to perform different noise reduction processing on a first region of the first information and a second region of the first information to generate multiple physiological characteristic information of the first user and the physiological characteristic indicators corresponding to the multiple physiological characteristic information. The second region is the region other than the first region, and the correlation degree of the first region with the physiological characteristic indicators is greater than the correlation degree of the second region with the physiological characteristic indicators.

4. The method according to claim 2 or 3, characterized in that, The forward process of the second model includes a first stage and a second stage. Noise is added to the input information in the first stage, and noise is added to the first region of the input information in the second stage. The first stage and the second stage are executed in parallel. Information for training the reverse process of the second model is obtained through the first stage and the second stage. The correlation intensity of the first region with the physiological characteristic indicators corresponding to the input information is greater than the correlation intensity of the second region with the physiological characteristic indicators corresponding to the input information. The second region is the region other than the first region.

5. The method according to claim 4, characterized in that, During the process of adding noise to the input information in the first stage, the first noise-added information at time t is obtained by adding noise to the input information t times, and the first noise-added information satisfies the following formula: where x t is the first noise-added information, x0 is the input information, ε is the noise, and q(x t |x0) represents the conditional probability distribution of the forward process, is the mean, is the variance, is a preset constant, and I is the identity matrix.

6. The method according to claim 4 or 5, characterized in that, Noise is added to the input information in the second stage, and the second noise-added information at time t is obtained by adding noise to the first region of the input information t times. The second noise-added information at time t satisfies the following formula: Among them, is the second noise addition information, x0 is the input information, m is the number of the first regions included in the input information, μ is the range of the first regions, ε is the noise, q(x t |x0) represents the conditional probability distribution of the forward process, is the mean value, is the variance, is a preset constant, and I is the identity matrix.

7. The method according to any one of claims 4 to 6, characterized in that The reverse process of the second model is trained in the following manner: Use the output of the forward process of the second model as the input of the reverse process of the source model; Perform different noise reduction processing on the first region and the second region of the input of the reverse process of the source model to obtain the output of the reverse process of the source model; In the case where the loss function converges, obtain the reverse process of the second model, where the loss function is the difference between the output of the reverse process of the source model and the input of the forward process of the second model; wherein, the input of the forward process of the second model includes physiological characteristic information of multiple users and physiological characteristic indicators corresponding to the physiological characteristic information of the multiple users; the loss function is related to the loss corresponding to the first region, the loss corresponding to the second region, and the physiological characteristic information of the multiple users.

8. The method according to claim 7, wherein The loss function satisfies the following formula: wherein, is the loss function, E[*] is the expectation of *, λ1 is the weight of the loss corresponding to the first region, λ2 is the weight of the loss corresponding to the second region, λ1 and λ2 are preset constants, μ is the range of the first region, ε is the noise, is the noise of the first region predicted by the source model, is the noise of the second region predicted by the source model, m is the number of the physiological characteristic information of the multiple users that includes the first region, x t is the first noise addition information at time t in the first stage of the forward process, is the second noise addition information at time t in the second stage of the forward process, and c is the physiological characteristic information of the multiple users.

9. The method according to any one of claims 1 to 8, characterized in that The first information further includes one or more of the gender of the first user, the age of the first user, or the body mass index (BMI) of the first user.

10. The method according to any one of claims 1 to 9, characterized in that, The method further includes: sending the first information to a server; receiving the updated first model from the server.

11. The method according to any one of claims 1 to 10, characterized in that, The physiological characteristic information of the first user includes a photoplethysmogram (PPG) signal, and the physiological characteristic indicator of the first user includes a blood pressure value.

12. An electronic device, characterized in that, including: a processor and a memory; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, so that the electronic device executes the method according to any one of claims 1-11.

13. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method according to any one of claims 1-11.

14. A chip system, characterized in that, including at least one processor and a communication interface, the communication interface and the at least one processor are interconnected by a line, and the at least one processor is configured to run a computer program or instructions to execute the method according to any one of claims 1-11.

15. A computer program product, characterized in that, including a computer program, when the computer program is run, it causes a computer to execute the method according to any one of claims 1-11.

Images