CN113491508A

CN113491508A - Electronic device and atrial fibrillation early warning method and medium thereof

Info

Publication number: CN113491508A
Application number: CN202010261627.8A
Authority: CN
Inventors: 贾淼; 李露平; 陈茂林; 韩羽佳
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2020-04-03
Filing date: 2020-04-03
Publication date: 2021-10-12

Abstract

The application relates to the field of communication and discloses an atrial fibrillation early warning method and medium based on electronic equipment. The atrial fibrillation early warning method comprises the following steps: the electronic device extracts a plurality of sinus cardiac intervals in photoplethysmography data of a user acquired by the electronic device; the electronic device calculating a degree of change in a relative difference value between a plurality of temporally adjacent ones of the plurality of sinus cardiac intervals; the electronic device determines whether to perform atrial fibrillation warning on the user based on the calculated degree of change. The atrial fibrillation early warning method analyzes the change degree of the relative difference values among the multiple sinus heart intervals in the photoplethysmography data of the user, and links the relationship among the multiple sinus heart intervals in the photoplethysmography data of the user, so that more accurate atrial fibrillation early warning is carried out on the user.

Description

Electronic device and atrial fibrillation early warning method and medium thereof

Technical Field

The application relates to the technical field of information processing, in particular to an atrial fibrillation early warning method and a medium thereof based on electronic equipment.

Background

Atrial Fibrillation (AF), also known as atrial Fibrillation, is a common type of arrhythmia. Long-lasting atrial fibrillation can cause serious complications such as heart failure, hypertension, and even life-threatening strokes. Therefore, it is very important to realize the early warning of atrial fibrillation in time. Whereas Heart Rate Variability (HRV), which refers to the variation of beat-to-beat cycle variability, is a valuable indicator of predicting sudden cardiac death and arrhythmic events. Clinical medicine often employs a dynamic Electrocardiography (ECG) monitoring system to perform a heart rate variability analysis on ECG waveform signals and to screen for various arrhythmia events. Namely, whether the user has atrial fibrillation is detected by analyzing waveform signal data of Electrocardiography (ECG).

In addition, photoplethysmography (PPG) is currently used to monitor and analyze cardiac rhythm abnormalities. Photoplethysmography (PPG) is an optical method for measuring tissue blood volume changes through a skin capillary bed, and general wearable electronic devices (such as a bracelet/watch and the like) have a PPG sensor. Continuous monitoring of the user by a PPG-based wearable electronic device is a viable method of atrial fibrillation screening that helps detect atrial fibrillation to reduce early intervention in stroke and other atrial fibrillation-related complications.

Disclosure of Invention

The embodiment of the application provides an atrial fibrillation early warning method based on electronic equipment and a medium thereof. The atrial fibrillation early warning method analyzes the change degree of the relative difference values among the multiple sinus heart beat intervals in the photoplethysmography data of the user, and links the relationship among the multiple sinus heart beat intervals in the photoplethysmography data of the user, so that more accurate atrial fibrillation early warning is carried out on the user. The content of the present application will be specifically described below.

In a first aspect, an embodiment of the present application provides an atrial fibrillation early warning method based on electronic equipment, including: the electronic device extracts a plurality of sinus cardiac intervals in photoplethysmography data of a user acquired by the electronic device; the electronic device calculating a degree of change in a relative difference value between a plurality of temporally adjacent ones of the plurality of sinus cardiac intervals; the electronic device determines whether to perform atrial fibrillation warning on the user based on the calculated degree of change. For example, the degree of change in the relative difference value between temporally adjacent sinus intervals of the sinus intervals may be generally used to characterize whether the user will have an atrial fibrillation event, and therefore, the user may be warned of the atrial fibrillation event by calculating the degree of change in the relative difference value between temporally adjacent sinus intervals of the sinus intervals to determine whether the user will have an atrial fibrillation event. The plurality of temporally adjacent sinus cardiac intervals refers to a plurality of groups of adjacent sinus cardiac intervals, that is, two temporally adjacent sinus cardiac intervals as a group of data.

In a possible implementation of the first aspect, the method further includes: the relative difference between temporally adjacent sinus cardiac intervals is: a ratio of a difference between two sinus intervals that are temporally adjacent to one another to a larger one of the two sinus intervals; and the degree of change in the relative difference is an average of the relative differences between a plurality of temporally adjacent sinus cardiac intervals. The relative difference may also be the ratio of the difference between adjacent sinus cardiac intervals to the smaller sinus interval of the two sinus cardiac intervals. The degree of change may be a variance of the relative difference or other value that may indicate a change in the relative difference data.

In a possible implementation of the first aspect, the method further includes: the electronic device calculates an average of relative differences between adjacent sinus intervals by:

wherein ADAN represents the mean of the deviations between adjacent sinus beat intervals，A_iRepresents the value of the ith data, max () represents the maximum value, and n represents the number of sinus cardiac intervals.

In a possible implementation of the first aspect, the method further includes:

and the electronic equipment displays prompt information for atrial fibrillation early warning to the user under the condition that the electronic equipment determines that the atrial fibrillation early warning is performed on the user. After the electronic equipment determines to perform atrial fibrillation early warning on the user, prompt information of the atrial fibrillation early warning of the user can be directly displayed on the electronic equipment. Of course, the electronic device may also remind the user directly through vibration, ringing, or the like without displaying the prompt information.

In a second aspect, an embodiment of the present application provides an atrial fibrillation early warning method based on an electronic device, including:

the method comprises the steps that a first electronic device receives photoplethysmography data, collected by a second electronic device, of a user from the second electronic device; the method comprises the steps that a first electronic device extracts a plurality of sinus heart beat intervals in acquired photoplethysmography data; calculating, by the first electronic device, a degree of change in a relative difference between a plurality of temporally adjacent ones of the plurality of sinus intervals; and the first electronic equipment determines whether to give an atrial fibrillation early warning to the user or not based on the calculated change degree. Namely, the electronic device (here, the second electronic device) in the first aspect processes the acquired photoplethysmography data, and in the second aspect, the processing can be performed by other electronic devices (namely, the first electronic device), so that parallel processing of data analysis and atrial fibrillation early warning can be realized, and further, the power consumption of the second electronic device is reduced.

In a possible implementation of the second aspect, the method further includes:

the relative difference between temporally adjacent sinus cardiac intervals is: a ratio of a difference between two sinus intervals that are temporally adjacent to one another to a larger one of the two sinus intervals; and the degree of change in the relative difference is an average of the relative differences between a plurality of temporally adjacent sinus cardiac intervals.

In a possible implementation of the second aspect, the method further includes: the first electronic device calculates an average of relative differences between adjacent sinus intervals by:

wherein ADAN represents the mean of the deviations between adjacent sinus beat intervals, A_iRepresents the value of the ith data, max () represents the maximum value, and n represents the number of sinus cardiac intervals.

In a possible implementation of the second aspect, the method further includes:

and under the condition that the first electronic equipment determines to perform atrial fibrillation early warning on the user, sending an early warning instruction to the second equipment, wherein the early warning instruction is used for instructing the second equipment to perform atrial fibrillation early warning on the user. After the first electronic device performs corresponding calculation processing on sinus heartbeat intervals in photoplethysmography data of a user, and under the condition that atrial fibrillation early warning is determined to be performed on the user, an early warning instruction is sent to the second electronic device.

In a possible implementation of the second aspect, the method further includes:

and under the condition that the first electronic equipment determines to give an atrial fibrillation early warning to the user, the first electronic equipment displays prompt information for the atrial fibrillation early warning to the user. Although the first electronic device can send the warning instruction to the second electronic device, at the same time, the first electronic device itself can also display corresponding prompt information for warning the atrial fibrillation to the user.

In a third aspect, the present application provides a training method for an atrial fibrillation early-warning model, including:

acquiring sample data of a photoplethysmography and an expected result of the corresponding sample data; extracting a plurality of sinus heart beat intervals in the acquired sample data of the photoplethysmography, and calculating the variation degree of relative difference values between a plurality of temporally adjacent sinus heart beat intervals in the plurality of sinus heart beat intervals; inputting the calculated change degree into a decision tree model, wherein nodes in the decision tree model comprise root nodes taking the calculated change degree as characteristics; and acquiring a prediction result of the decision tree model, and adjusting various parameter values in the decision tree model according to the prediction result and the expected result. Wherein the expected result refers to the labeling of the sample data of photoplethysmography (namely the sample data of photoplethysmography labeled as "early warning" and the sample data of photoplethysmography labeled as "no early warning"); the degree of variation refers to the average of the relative differences between adjacent sinus intervals. And then, taking the variation degree as a root node of the decision tree model to carry out model training. It is understood that the degree of variation may also be a leaf node of the decision tree model, and the model herein may also be other than a decision tree model.

In a possible implementation of the third aspect, the method further includes:

calculating a first average value of the change degree of sample data of which the expected result is the atrial fibrillation event in the obtained sample data of the photoplethysmography, and calculating a second average value of the change degree of the sample data of which the expected result is the non-atrial fibrillation event in the obtained sample data of the photoplethysmography; and inputting the calculated first average value and the second average value of the degree of change into a decision tree model, wherein nodes in the decision tree model comprise nodes which take the calculated first average value and the calculated second average value as features respectively. That is, a first average value of the variation degree of sample data of an atrial fibrillation event and a second average value of the variation degree of sample data of a non-atrial fibrillation event may be calculated, and then the first average value or the second average value may be used as a node of the decision tree model.

In a possible implementation of the third aspect, the method further includes:

the electronic device calculates an average of relative differences between adjacent sinus intervals by:

In a fourth aspect, the present application provides a readable medium of an electronic device, where the readable medium has instructions stored thereon, and when the instructions are executed on the electronic device, the instructions may cause the electronic device to perform any one of the possible methods described above.

In a fifth aspect, an embodiment of the present application provides an electronic device, including: a memory for storing instructions for execution by one or more processors of the system, and a processor, which is one of the processors of the system, for performing any one of the possible methods described above.

In a sixth aspect, an embodiment of the present application provides an electronic device, where the electronic device has a function of implementing the search method. The functions may be implemented by hardware, or by hardware executing corresponding software. The hardware or software includes one or more units corresponding to the above functions.

Drawings

Fig. 1a illustrates a distribution of values of ADANN in non-atrial fibrillation data, where the abscissa represents the value of ADANN and the ordinate represents the number of ADANN, according to some embodiments of the present application.

Fig. 1b illustrates a distribution of values of ADANN in atrial fibrillation data, where the abscissa represents the values of ADANN and the ordinate represents the number of ADANN, according to some embodiments of the present application.

Fig. 2a is a graph illustrating distribution of values of ADANN in data of suspected atrial fibrillation patients according to some embodiments of the present application, where the abscissa represents time stamps and the ordinate represents values of ADANN.

Fig. 2b illustrates a distribution of values of ADANN across all data in a low risk population according to some embodiments of the present application, where the abscissa represents the timestamp and the ordinate represents the value of ADANN.

FIG. 3 illustrates a distribution of values for 2h ADANN before an atrial fibrillation event in a suspected patient, where the abscissa represents the value of ADANN and the ordinate represents the corresponding number of ADANN values for 2h data before an atrial fibrillation event occurs, according to some embodiments of the present disclosure.

Fig. 4a illustrates an application scenario of an atrial fibrillation warning method based on an electronic device according to some embodiments of the present application.

Fig. 4b illustrates a method for building an atrial fibrillation pre-warning model, according to some embodiments of the present application.

Fig. 4c illustrates a method for building an atrial fibrillation pre-warning model, according to some embodiments of the present application.

Fig. 5 illustrates a method of extracting RRs in PPG data, according to some embodiments of the present application.

Fig. 6a illustrates a schematic diagram of a method for atrial fibrillation pre-warning model training using a decision tree model, according to some embodiments of the present application.

Fig. 6b shows a schematic diagram of a method for atrial fibrillation pre-warning model training using a decision tree model, according to some embodiments of the present application.

FIG. 7a illustrates a method of time domain feature derivation according to some embodiments of the present application.

FIG. 7b illustrates a method of time domain feature derivation according to some embodiments of the present application.

FIG. 8 illustrates a method of base feature augmentation, according to some embodiments of the present application.

Fig. 9 illustrates a schematic diagram of a method for atrial fibrillation pre-warning model training using a decision tree model, according to some embodiments of the present application.

Fig. 10 shows a schematic interaction diagram between bracelet 100 and cell phone 200 at the time of atrial fibrillation warning, according to some embodiments of the present application.

FIG. 11a illustrates a human-machine interface, according to some embodiments of the present application.

FIG. 11b illustrates a human-machine-interaction interface, according to some embodiments of the present application.

Fig. 12 shows a schematic diagram of a method for atrial fibrillation pre-warning on a bracelet 100, according to some embodiments of the present application.

Fig. 13a shows a prior art manner in which a wearable electronic device acquires PPG data using a PPG sensor and processes the PPG data.

Fig. 13b shows characteristic values that can characterize abnormal fluctuations in heart rhythm in the prior art.

FIG. 13c is a graph showing the ranking of feature importance of feature values that may characterize atrial fibrillation events.

Fig. 14 shows the effect comparison between the atrial fibrillation early warning model and the atrial fibrillation early warning model in the prior art.

Fig. 15 shows a hardware structure diagram of an electronic device capable of implementing the function of the bracelet 100 according to some embodiments of the present application.

Fig. 16 illustrates a hardware architecture diagram of an electronic device 800 capable of implementing the functionality of the handset 200, according to some embodiments of the present application.

Fig. 17 illustrates a software architecture diagram of an electronic device 800 capable of implementing the functionality of the handset 200, according to some embodiments of the present application.

FIG. 18 illustrates a block diagram of a system capable of implementing the server 300, according to some embodiments of the present application.

Detailed Description

The technical solutions of the embodiments of the present application are described in further detail below with reference to the accompanying drawings and embodiments.

According to the technical scheme, the possibility of an Atrial Fibrillation (AF) event of a user is judged by acquiring photoplethysmography (PPG) data (or PPG data) of wearable electronic equipment worn by the user, specifically, adjacent sinus heartbeat intervals (RR) of the user are extracted from the PPG data, and the possibility of the AF event of the user is judged by calculating the fluctuation change of the RR. For example, an Average value (ADANN) of deviations of adjacent RRs is calculated by the following formula (one), and the ADANN is used as a characteristic value representing an Atrial Fibrillation (AF) event to determine whether the user may have the AF event, and when it is determined that the user may have the AF event, an atrial Fibrillation warning is given to the user.

Wherein ADAN denotes the mean value of the deviations between adjacent RRs, A_iRepresents the value of the ith data, max () represents the maximum value, and n represents the number of RRs.

In addition, the ADANN of the application can be used in an atrial fibrillation early warning model and used for atrial fibrillation early warning in conjunction with other features (such as the minimum value in RR data, the standard deviation of one difference in RR data, the skewness of RR data, and the like) for representing atrial fibrillation.

Furthermore, it is understood that ADANN may also be calculated by equation (two):

wherein ADAN denotes the mean value of the deviations between adjacent RRs, A_iA value representing the ith data, max () represents taking the maximum value, n represents the number of RRs, and max () may be replaced with min (), representing taking a small value. The application is not limited thereto.

This feature of ADANN is described in detail below.

The interval between adjacent sinus beats (RR) is used in the present application for atrial fibrillation pre-warning because RR is typically used to characterize a user's atrial fibrillation condition. For example, in sinus rhythm data, the RR is substantially smooth and consistent, but at the time of an atrial fibrillation event, the RR is chaotic, and the variance, entropy, etc. of the RR fluctuates. Therefore, by analyzing the RR, it can be determined whether the user has an atrial fibrillation event or is about to have an atrial fibrillation event at the time. According to big data statistics, when the fluctuation degree of RR data is obvious, the ratio relationship (or the degree of change) between the difference (or the relative difference) of adjacent RR data and a larger value can reflect the fluctuation condition of the heart rhythm to a certain extent, for example, the distribution condition of the Average of Deviations (ADANN) between adjacent RRs can reflect the fluctuation condition of the heart rhythm to a certain extent, so as to better perform atrial fibrillation early warning.

For example, in a particular implementation, ADANN has a significantly different distribution in atrial fibrillation/non-atrial fibrillation data when analyzing data for patients with suspected atrial fibrillation. As shown in fig. 1a and 1b, the abscissa indicates the value of ADANN and the ordinate indicates the number of ADANN, it is apparent from fig. 1a that the value of ADANN in the non-atrial fibrillation data is mostly less than 0.15, and from fig. 1b that the value of ADANN in the atrial fibrillation number is mostly greater than 0.15.

For another example, in a specific implementation, ADANN also has a significantly different distribution in data of patients with suspected atrial fibrillation and data of healthy people (or low risk people). As shown in fig. 2a and 2b, where each point represents an ADANN data, the abscissa represents a time stamp, and the ordinate represents the value of ADANN, it can be seen from fig. 2a that the value of ADANN in data of suspected atrial fibrillation patients is mostly greater than 0.15 and less than 0.25. As can be seen from fig. 2b, the values of ADANN in the data of healthy population (or low risk population) are mostly less than 0.15.

For another example, as shown in fig. 3, the abscissa represents the ADANN value, and the ordinate represents the corresponding number of ADANN values of the data 2h before the occurrence of atrial fibrillation event, it can be seen that 99% of the data 2h before the occurrence of atrial fibrillation event has abnormal change of the ADANN value, and further, most of the ADANN values 2h before the occurrence of atrial fibrillation event are greater than 0.15.

Therefore, the ADANN value has strong distinguishability on whether atrial fibrillation occurs or not, and can be used as a sign of the occurrence of atrial fibrillation to predict the occurrence of the atrial fibrillation and carry out atrial fibrillation early warning on a corresponding user.

Fig. 4a shows an atrial fibrillation warning system 10, according to some embodiments of the present application, which performs atrial fibrillation warning based on the ADANN mentioned above. Specifically, the atrial fibrillation warning system 10 includes a wearable electronic device 100, an electronic device 200, and an electronic device 300. The wearable electronic device 100 is mainly used for acquiring PPG data of a user, performing atrial fibrillation judgment based on the atrial fibrillation early warning model and the acquired PPG data, and performing early warning on the user when the user is judged to have an atrial fibrillation event or will have the atrial fibrillation event. Wearable electronic device 100 may wirelessly communicate with other electronic devices in various wireless manners, for example, wirelessly communicate with electronic device 200 or electronic device 300.

It is understood that in some embodiments of the present application, wearable electronic device 100 may be a variety of devices, including but not limited to a watch, a bracelet or glasses, a helmet, a headband, and like wearable electronic devices, medical detection instruments, and the like. In the following description, for simplicity of explanation, the electronic device 100 is taken as a bracelet as an example to explain the technical solution of the present application.

Electronic device 200 may include but not limited to mobile terminals such as cell-phone, tablet computer, when electronic device 200 has the early warning model of atrial fibrillation loaded, electronic device 200 can judge that atrial fibrillation is performed according to the PPG data that wearable electronic device 100 obtained, and when judging that the user takes place or will take place the atrial fibrillation event, send the judged result to wearable electronic device 10, make wearable electronic device 100 carry out the early warning of atrial fibrillation to the user, or directly carry out the early warning of atrial fibrillation to the user by electronic device 200. In the following description, for simplicity of explanation, the electronic device 200 is a mobile phone, and the technical solution of the present application is described in detail.

The electronic device 300 may include, but is not limited to, a laptop computer, a desktop computer, a tablet computer, a server, and so forth. The electronic device 300 is mainly used for training the atrial fibrillation early warning model and sending the atrial fibrillation early warning model to the wearable electronic device 100 or the electronic device 200. In addition, the electronic device 300 may also perform atrial fibrillation determination according to the PPG data acquired by the wearable electronic device 100, and send a determination result to the wearable electronic device 100 when determining that the user has atrial fibrillation symptoms, so that the wearable electronic device 100 performs atrial fibrillation warning to the user. In the following description, the electronic device 300 is taken as an example to explain the technical solution of the present application.

The process of server 300 for establishing an atrial fibrillation pre-warning model is described in detail below with reference to fig. 4b-4 c.

Fig. 4b shows a process of building an atrial fibrillation pre-warning model according to an embodiment of the present application. Specifically, as shown in fig. 4b, the method includes:

401: sample data acquisition and preprocessing

The training data of the server 300 for training the atrial fibrillation early warning model may be existing PPG sample data, such as PPG sample data used by a developer of the bracelet 100, or may be PPG sample data acquired specially for training the atrial fibrillation early warning model of the present application. The PPG sample data includes a large amount of PPG sample data under "atrial Fibrillation event (AF)" and "Non-atrial Fibrillation event (NAF)", as well as PPG sample data of healthy people (or low risk people) and PPG sample data of suspected patients with atrial Fibrillation.

The server 300 may also pre-process the acquired PPG sample data and extract the interval between adjacent sinus heart beats (RR) in the acquired PPG sample data. In addition, in the process of establishing an atrial fibrillation early warning model, the quality of the PPG sample data is guaranteed to be effective under the condition that sufficient PPG sample data is guaranteed, so that the PPG sample data needs to be preprocessed. For example, some PPG sample data may be obtained when the user starts to use the bracelet 100, the quality of the PPG sample data at this time may be poor due to the initialization of the device state, or the quality of the PPG sample data may be poor when the user wears the bracelet incorrectly (for example, too loose or too tight), and at this time, the PPG sample data may be removed from the obtained PPG sample data, and only the remaining PPG sample data with better quality is analyzed.

Neighboring sinus heart beat (RR) is then extracted from the pre-processed PPG sample data. Specifically, a group RR: r1(A1, A2, A3. ANG. An), R2(An +1, An +2, An + 3. ANG. A2n), R3(A2n +1, A2n +2, A2n + 3. ANG. A3 n. ANG. Ra (A (a-1) n +1, A (a-1) n + 2. ANG. Aan).

For example, as shown in fig. 5, the server 300 can extract 3 sets of RR, R1(721, 671, 66, 1680, 766, 890 ·), R2(928, 610, 777, 546, 740, 715 ·), R3(960, 760, 880, 720, 1200, 760 ·) from PPG sample data.

It is understood that the PPG data includes data that may reflect the physical health of the user, such as heart rate data, pulse data, and the like, which is not limited in this application.

402: computing RR-based features

For the RR of the PPG sample data extracted above, ADANN may be calculated using the above formula (one) or (two).

It is understood that in some embodiments of the present application, the basic features that can be used for performing atrial fibrillation warning include, but are not limited to, ADANN, the minimum value min _ RR in RR, the standard deviation SDNN of RR, the standard deviation RMSSD of the first difference of RR, the skewness sk _ RR of RR, and the like.

The minimum value Min _ RR in the extracted RRs can be calculated through a Min () formula (three), wherein Min () represents the minimum value;

calculating standard deviation SDNN of RR by formula (IV)

Wherein SDNN represents the standard deviation of RR, n represents the number of RR, A_iA value representing the ith data is calculated,

represents the average value of RR;

calculating the standard deviation RMSSD of the first difference of the RR data by the formula (V)

Wherein RMSSD represents the standard deviation of the first difference of RR, n represents the number of RR, A_iA value representing the ith data;

calculating skewness sk _ RR of RR data by formula (VI)

Wherein sk _ RR represents the skewness of RR,

mean values of the RR are indicated, and SDNN standard deviation of the RR.

403: decision tree model for training atrial fibrillation early warning model

And establishing an atrial fibrillation early warning model based on the calculated features.

In some embodiments of the present application, the ADANN may be used as a root node of the decision tree model to determine whether to perform atrial fibrillation warning, or the ADANN and other features mentioned above (for example, the minimum value min _ RR in the RR data, the standard deviation SDNN of the RR data, the standard deviation RMSSD of the first difference of the RR data, the skewness sk _ RR of the RR data, etc.) may be used as nodes of the decision tree model respectively.

For example, in some embodiments, as shown in fig. 6a, taking ADANN as the root node of the decision tree model, the training process of the decision tree model includes:

601: the PPG sample data of the marked suspected atrial fibrillation patient is input to the decision tree model, and ADANN is calculated for a plurality of 45s (or other durations, such as 50s, 60s, 100s, etc.) of PPG sample data in each PPG sample data. For example, as shown in fig. 6b, it is known that atrial fibrillation events occur at time t1 and time t2, and PPG sample data corresponding to each time t1-t2 is marked. If the time between t0 and t2 is longer than a preset early warning duration (for example, 2h, 4h, 8h), marking all the PPG sample data corresponding to the time t1-t0 as "no early warning" (i.e., no atrial fibrillation event occurs between t1-t 0), and marking the PPG sample data at the time t0-t2 as "early warning" (i.e., an atrial fibrillation event occurs between t0-t 2). ADANN for a plurality of segments of 45sPPG sample data in each PPG sample data is then calculated and compared to a first threshold and a second threshold. If the ADANN is greater than the first threshold, 602 is entered, and if the ADANN is less than the second threshold, 603 is entered. For example, assuming that the first threshold and the second threshold are both 0.15 obtained through the big data statistics, 602 is performed when ADANN calculated by one PPG sample data is greater than 0.15, and 603 is performed when ADANN calculated by one PPG sample data is less than 0.15. It is to be understood that 0.15 is merely exemplary for convenience of explanation, and various thresholds used may be set to make the determination in practice.

602: determining that the input PPG sample data is sample data needing to be subjected to early warning on a user, comparing the output result with a mark when the PPG sample data is input, if the deviation between the output result and the mark when the PPG sample data is input is smaller than a preset threshold value or is consistent with the preset threshold value, considering that the training of the atrial fibrillation early warning model is finished, otherwise, continuously modifying the first threshold value, and repeating the steps until the deviation between the output result and the mark of the input PPG sample data is smaller than a preset threshold value or is consistent with the preset threshold value.

603: and determining that the input PPG sample data is sample data which does not need early warning, comparing the output result with a mark when the PPG sample data is input, if the output result is consistent with the mark when the PPG sample data is input, considering that the training of the atrial fibrillation early warning model is finished, otherwise, continuously modifying the second threshold, and repeating the steps until the deviation of the output result and the mark of the input PPG sample data is smaller than a certain preset threshold or consistent.

The output result is consistent with the input label of the PPG sample data, which may be that the output result is the same as the input PPG sample data or that the deviation between the output result and the input PPG sample data is smaller than a certain preset threshold. Alternatively, the deviation between the output result and the input label of the PPG sample data may be used as a loss function, and when the loss function is close to 0, the output result may be considered to be consistent with the input label of the PPG sample data. The first threshold and the second threshold may be the same or different.

In addition, in other embodiments of the present application, the decision tree model in the atrial fibrillation warning model may use other basic features besides ADANN as nodes, for example, the minimum value min _ RR in the RR data mentioned above, the standard deviation SDNN of the RR data, the standard deviation RMSSD of the first difference of the RR data, the skewness sk _ RR of the RR data, and the like. The ADANN may be a root node, and the other basic features may be leaf nodes, or the other basic features (e.g., SDNN) may be root nodes, and the ADANN and other basic features may be nodes. In a specific training process, the training of the decision tree model may be performed according to a deviation between an output result of the decision tree model and a label of the PPG sample data, which is similar to the above example and is not limited herein.

Fig. 4c shows another atrial fibrillation pre-warning model building process according to an embodiment of the present application. Specifically, as shown in fig. 4c, the method includes:

steps

401, 403 are the same as the

previous steps

401, 402, 403, and are not described again here.

402 a: ADANN-based feature time domain derivation

In the above process, the server 300 extracts a group RR from the PPG sample data: r1(A1, A2, A3. ANG. An), R2(An + I, An +2, An + 3. ANG. A2n), R3(A2n + I, A2n +2, A2n + 3. ANG. A3 n. ANG. Ra (A (a-1) n +1, A (a-1) n + 2. ANG. Aan), and based on the extracted RR, a group ADANN value is calculated. The ADANN-based characteristic time-domain derivation is described below with reference to figures 7a and 7b, taking the ADANN value as an example:

(1) and calibrating a historical data window: extracting N (for example, N-50) ADANN values before the current time as the window data for feature fusion.And acquiring and framing the window data, and labeling the value of ADANN, i.e. labeling whether the value of ADANN in the window corresponds to the occurrence of atrial fibrillation event (i.e. whether ADANN is the one in which atrial fibrillation occurs). For example, the ADANN value in the window data is labeled as ADANN under "atrial fibrillation event_AF1、ADANN_AF2、ADANN_AF3···ADANN_AFbAnd ADANN under "non-atrial fibrillation events_NAF1、ADANN_NAF2、ADANN_NAF3、···ADANN_NAFcAnd b + c ═ N.

(2) As shown in FIG. 7b, the ADANN values corresponding to "atrial fibrillation events" and "non-atrial fibrillation events" in the window data are averaged according to the formula (VII) to obtain the average values

And

two-dimensional time-domain derived new features corresponding to ADANN values under atrial fibrillation events and non-atrial fibrillation events

Wherein,

represents the average of all data, n represents the number of data, A_iIndicating the value of the ith data.

402 b: basic feature extension based on RR

As shown in fig. 8, based on the ADANN feature value obtained from the RR extracted from the PPG sample data, the minimum value min _ RR in the RR data, the standard deviation SDNN of the RR data, the standard deviation RMSSD of the first difference of the RR data, and the skewness sk _ RR of the RR data, the new feature derived in the 2(N +1) dimension is extended by the method of "402 a" described above.

For example, the historical data window method may be specified as in (1): extracting previous to the current timeN (e.g., N ═ 50) min _ RR values are used as the window data for feature fusion. And acquiring and framing window data, and labeling the min _ RR value, i.e. labeling whether the min _ RR value in the window corresponds to the occurrence of an atrial fibrillation event (i.e. whether the min _ RR value is the min _ RR value when atrial fibrillation occurs). For example, marking the min _ RR value in windowed data as min-RR at "atrial fibrillation events_AF1、min_RR_AF2、min_RR_AF3···min_RR_AFbAnd min _ RR at "non-atrial fibrillation events_NAF1、min_RR_NAF2、min_RR_NAF3、···min-RR_NAFcAnd b + c ═ N.

According to the method (2), the min _ RR values corresponding to atrial fibrillation events and non-atrial fibrillation events in the window data are respectively averaged according to the formula (seven) to respectively obtain

And

i.e. the two-dimensional time-domain derived new feature corresponding to the min _ RR value under "atrial fibrillation events" and "non-atrial fibrillation events".

The way of extending other basic features to corresponding two-dimensional time domain derived new features is consistent with the above process, and is not repeated here. Expanding all the (N +1) -dimensional features (namely ADANN, min _ RR, SDNN, RMSSD and sk _ RR) according to the method to obtain 2(N +1) -dimensional new features

403: decision tree model for training atrial fibrillation early warning model

And training a decision tree model in the atrial fibrillation early warning model based on the calculated features. Wherein, the blockThe nodes of the policy tree model may be ADANN and its derived features (i.e., the

) It can also be all the basic features and their derivatives (i.e. ADANN, min _ RR, SDNN, RMSSD, sk _ RR, and

)。

for example, as shown in fig. 9, ADANN (or other features described above) is used as the root node of the decision tree model. The other features are then used as leaf nodes of the decision tree model. For example, SDNN, ADANN and

as leaf nodes of the decision tree model. The method specifically comprises the following steps:

901: inputting the marked PPG sample data of the suspected atrial fibrillation patient into the decision tree model, and calculating the ADANN of multiple 45s (or other time lengths, such as 50s, 60s, 100s and the like) of PPG sample data in each piece of PPG sample data.

Comparing the ADANN with a third threshold, a fourth threshold and a fifth threshold respectively, and if the ADANN is larger than the third threshold, entering 902; if ADANN is less than the fourth threshold, then 903 is entered; if the ADAN is less than the fifth threshold then 904 is entered.

902: comparing the SDNN with a sixth threshold value and a seventh threshold value respectively, and if the SDNN is larger than the sixth threshold value, entering 905; if less than the seventh threshold, then 906 is entered.

903: the mean value of ADANN in atrial fibrillation events was calculated. Comparing the average value of the ADANN in the atrial fibrillation event with an eighth threshold value and a ninth threshold value respectively, and entering 905 if the average value of the ADANN in the atrial fibrillation event is larger than the eighth threshold value; if the average value of ADANN in atrial fibrillation events is less than the ninth threshold, then 906 is entered.

904: the mean value of ADANN in non-atrial fibrillation events was calculated. Comparing the average value of the ADANN in the non-atrial fibrillation event with a tenth threshold and an eleventh threshold respectively, and if the average value of the ADANN in the non-atrial fibrillation event is larger than the tenth threshold, entering 905; if the average value of ADANN in the non-atrial fibrillation event is less than the eleventh threshold, then 906 is entered.

905: and determining the input PPG sample data as data needing early warning.

906: and determining the input PPG sample data as data which does not need early warning.

And then judging whether the output result is consistent with the mark when the PPG sample data is input, if not, modifying a related threshold (for example, a third threshold, a fourth threshold, a fifth threshold and the like) in the decision tree model, and repeating the steps until the output result is consistent with the mark when the PPG sample data is input or whether the deviation between the output result and the mark when the PPG sample data is input is smaller than a preset threshold. And when the output result is consistent with the mark when the PPG sample data is input or the deviation between the output result and the mark when the PPG sample data is input is less than a certain preset threshold value, the training of the atrial fibrillation early warning model is considered to be finished. The third threshold, the fourth threshold, and the fifth threshold are different from each other, the sixth threshold and the seventh threshold may be the same or different from each other, the eighth threshold and the ninth threshold may be the same or different from each other, and the tenth threshold and the eleventh threshold may be the same or different from each other.

In addition, in other embodiments of the present application, the decision tree model in the atrial fibrillation warning model may use ADANN-derived features other than ADANN and other basic features and their derived features as nodes, such as the minimum value min _ RR in the RR data mentioned above, the standard deviation SDNN of the RR data, the standard deviation RMSSD of the first difference of the RR data, the skewness sk _ RR of the RR data, or derived features of these basic features (i.e. the deviation sk _ RR of the RR data)

) And the like. Wherein, ADAN and its derivativesThe feature may be a root node, and the other basic feature and its corresponding derived feature are leaf nodes, or the other basic feature (e.g., SDNN) and its corresponding derived feature are root nodes, and the ADANN and its derived feature and the other basic feature and its corresponding derived feature are leaf nodes. In a specific training process, the training of the decision tree model may be performed according to a deviation between an output result of the decision tree model and a label of the PPG sample data, which is similar to the above example and is not limited herein.

After the atrial fibrillation early warning model is trained, the server 300 can transplant the model to the mobile phone 200 or the bracelet 100, and then judge atrial fibrillation of the user in real time. Fig. 10 shows a process of performing atrial fibrillation warning on a user wearing the bracelet 100 after the atrial fibrillation warning model is transplanted to the mobile phone 200. As shown in fig. 10, includes:

1000: after receiving the atrial fibrillation early warning model established by the server 300, the mobile phone 200 installs the atrial fibrillation early warning model. For example, an Android project may be established, the model is read and analyzed through a model reading interface in the project, and then an Android Application Package (APK) file is generated by compiling and installed in the mobile phone 200.

1002: the mobile phone 200 actively or passively sends an instruction for acquiring PPG data to the bracelet 100. For example, the mobile phone 200 may receive a user initiated instruction of "obtaining PPG data" through the human-computer interface, and then send the instruction of "obtaining PPG data" to the bracelet 100 according to the user instruction. Of course, the handset 200 may also automatically initiate an instruction to the bracelet 100 to "acquire PPG data" at a specific moment. For example, the user may be working during the day and may be more prone to fatigue, so the cell phone 200 may be configured to acquire PPG data during the day "6: 00-18: 00". Alternatively, the user may be more dangerous with atrial fibrillation occurring in the evening, so the cell phone 200 may be equipped to acquire PPG data for "22: 00-5: 00" in the evening.

1004: after receiving the instruction of "acquiring PPG data" from the mobile phone 200, the bracelet 100 acquires PPG data for T seconds at intervals of T seconds. Specifically, when the bracelet 100 acquires PPG data at "6: 00-18: 00" in the daytime, the interval time T and the acquisition time T of the bracelet 100 may be larger because the user is in the working state, the active state or the wearing manner of the bracelet may be unstable, for example, the user may take down the bracelet when participating in a conference. For example, the bracelet 100 acquires half an hour of PPG data every 1 hour. For another example, the user may want to know his own physical condition at a certain time, such as after a strenuous exercise, at which time, the user may set the bracelet 100 through the mobile phone 200 to acquire PPG data for 45s every 100 s.

1006: the bracelet 100 sends the acquired PPG data to the cell phone 200.

1008: the mobile phone 200 analyzes the acquired PPG data by using the model and obtains an atrial fibrillation early warning result

1010: the mobile phone 200 performs atrial fibrillation early warning on the user based on the atrial fibrillation early warning result.

For example, the bracelet 100 acquires PPG data of a user for T seconds at intervals of T seconds according to an instruction sent by the mobile phone 200, and then sends the acquired PPG data to the mobile phone 200, and performs atrial fibrillation early warning on the user by using an atrial fibrillation early warning model. Specifically, the user can independently select the time period for atrial fibrillation early warning through the mobile phone 200, then send an instruction to the bracelet 100, make the bracelet 100 acquire the PPG data of this time period, then send the acquired PPG data to the mobile phone 200, and utilize the atrial fibrillation early warning model to carry out atrial fibrillation early warning on the user. For example, even a person who is usually exercising may have atrial fibrillation after a certain strenuous exercise and may have severe sudden death. Therefore, the user can select to acquire PPG data while exercising through the mobile phone 200. For example, at night (22:00-23:00), after the user just exercises, the bracelet 100 may acquire PPG data of the user for 45s every 100s, and send the acquired PPG data to the mobile phone 200, and then the mobile phone 200 outputs an atrial fibrillation early warning result according to the received PPG data by using an atrial fibrillation early warning model, and displays the atrial fibrillation early warning result of the user on the mobile phone 200 and performs atrial fibrillation on the user at the same time through the mobile phone 200. For example, as shown in fig. 11a, the mobile phone 200 displays the atrial fibrillation warning result of the user on the screen of the mobile phone 200 in the form of a message (for example, the mobile phone 200 interface displays "you currently have risk of atrial fibrillation"), and then as shown in fig. 11b, the user can hide the message by clicking "know la" and then return to the mobile phone 200 interface.

It can be understood that, it may also be performed through bracelet 100 to perform atrial fibrillation early warning on the user, for example, bracelet 100 may remind the physical condition of the user at this moment through vibration, or bracelet 100 may also display the atrial fibrillation early warning result of the user at this moment on the interface to perform atrial fibrillation early warning on the user. The application is not limited thereto.

In addition, it can be understood that the bracelet 100 may also be equipped with the aforementioned atrial fibrillation pre-warning model, and then independently execute the data acquisition and the atrial fibrillation pre-warning process of the aforementioned mobile phone 200, and the specific process is as follows:

1201: the atrial fibrillation early warning model is installed on the bracelet 100 side.

1202: each time the bracelet 100 is spaced for T seconds, PPG data for T seconds is acquired.

1203: the bracelet 100 analyzes the acquired PPG data by using a model and obtains an atrial fibrillation early warning result.

1204: the mobile phone 200 performs atrial fibrillation early warning on the user based on the atrial fibrillation early warning result.

The way of specifically acquiring the PPG data and the way of acquiring the PPG data by using model analysis are consistent with the above-mentioned ways, and are not described herein again.

Referring now to fig. 13-14, the technical effects that can be achieved by the present application are illustrated.

As shown in fig. 13a, typically a wearable electronic device may acquire 45s PPG sensor data at intervals with the PPG sensor when the user is at rest, and then extract the interval between adjacent sinus beats (RR) for each single PPG segment acquired. Because the RR sequences are basically smooth and consistent in sinus rhythm data, and the variance, entropy and the like of the RR sequences fluctuate in atrial fibrillation attacks, whether a user is under an atrial fibrillation event or is about to occur can be judged by analyzing the fluctuation of the RR sequences of the heart rhythm. However, in the prior art, only the deviation degree of the RR from the average value and the deviation degree between adjacent RRs are concerned in the current 45s signal segment, and only the RR in the current 45s is concerned, and the multi-dimensional fusion of continuous multiple pieces of PPG data is lacked. The ADANN provided by the application not only pays attention to the deviation degree between adjacent RRs, but also fuses the RR change conditions in multiple sections of PPG data, so that the result of atrial fibrillation early warning on a user through the ADANN is more accurate.

Next, as shown in fig. 13b, the original features that can characterize atrial fibrillation events are: the number of RR and average RR is more than 50 milliseconds, the percentage of the total number of RR, the standard deviation of all RR and the ratio of the standard deviation to the average value of all RR, wherein the number of RR and average RR is more than 50 milliseconds. As can be seen from FIG. 13c, the original ranking in the feature importance ranking is later, but the reference number f corresponding to ADANN in this application₁₅And ranks first 12 in all feature importance orderings, while the other 2(N +1) feature values proposed in this application correspond to the reference number f₁₈-f₂₇The top 10 in all feature importance rankings.

Moreover, as shown in fig. 14, under the test data of the suspected atrial fibrillation user of as many as 70000 persons, the effect of the atrial fibrillation early warning model using only the original features and the ADANN and 2(N +1) feature values provided by the present application has a significant difference:

the early warning sensitivity of ADANN and the atrial fibrillation early warning model of 2(N +1) eigenvalues that this application provided is 81.72%, compares original model early warning sensitivity 62% and has promoted 19.72%, and the atrial fibrillation early warning model of this application is up to 97.27% in the early warning specificity, compares original model early warning specificity 91.45% and has promoted 5.82%, finally, the early warning total correct rate of the atrial fibrillation early warning model of this application is 93.15%, compares 83.66% of original model, has promoted 9.49%.

Fig. 15 is a schematic diagram illustrating a hardware structure of an electronic device implementing the bracelet 100 according to some embodiments of the present application. The bracelet may include a bracelet body 100. In one embodiment of the present application, the main body of the bracelet 100 may include a touch screen 101 (also referred to as a touch panel), a display screen 102, a housing including a front case (not shown in fig. 15) and a bottom case (not shown in fig. 15), and a processor 103, a Micro Control Unit (MCU) 104, a memory 105, a wireless communication unit 106, a PPG sensor 107, a power supply 111, a power management system 112, and the like.

The following describes each functional component of the bracelet 100:

the touch screen 101, which may also be a touch panel, may collect touch operations of the user on the bracelet (for example, operations of the user on or near the touch panel using any suitable object or accessory such as a finger or a stylus pen), and drive a responsive connection device according to a preset program.

The display screen 102 may be used for displaying information input by a user or prompt information provided to the user and various menus on the bracelet, and further, the touch panel 101 may cover the display screen 102, and when the touch panel 101 detects a touch operation on or near the touch panel 101, the touch panel transmits the touch operation to the processor 103 to determine the type of the touch event, and then the processor 103 provides a corresponding visual output on the display screen 102 according to the type of the touch event. For example, in some embodiments of the present application, when the mobile phone 200 uses an atrial fibrillation pre-warning model, the user may be pre-warned of atrial fibrillation by using the bracelet 100, and pre-warning information is directly displayed on the display screen 102.

The processor 103 is used for system scheduling, the touch screen 101, the control display screen 102, bluetooth 106 processing support and the like.

A micro control unit 104 for controlling the sensors, communicating with the processor 103, etc. The sensor may include a photoplethysmography (PPG) sensor 107 or other sensors such as an acceleration sensor and a motion sensor (not shown in fig. 15). For example, in some embodiments of the present application, the micro control unit 104 controls the PPG sensor to acquire PPG data.

The memory 105 is used for storing software programs and data, and the processor 103 executes various functional applications and data processing of the bracelet by running the software programs and data stored in the memory 105. For example, in some embodiments of the present application, the memory 105 may store PPG data acquired by a PPG sensor. Meanwhile, the memory may also store registration information, login information, and the like of the user.

The wireless communication module 106, the bracelet 100 can interact with other electronic devices (such as a mobile phone, a tablet computer, etc.) through the wireless communication module, and can be connected to a network or a server through the electronic devices. The wireless communication module may provide a solution for wireless communication applied to the bracelet 100, including Wireless Local Area Networks (WLANs), (such as wireless fidelity (Wi-Fi) networks), bluetooth (blue tooth, BT), Global Navigation Satellite System (GNSS), Frequency Modulation (FM), Near Field Communication (NFC), Infrared (IR), and the like.

For example, the bracelet 100 may communicate with the mobile phone 200 through bluetooth, receive an instruction of the mobile phone 200, communicate with the server 300 through wireless communication, and send the acquired PPG data to the server 300.

It is understood that the structure shown in fig. 15 is only one specific structure for realizing the function of the bracelet 100 in the technical solution of the present application, and the bracelet 100 having other structures and realizing similar functions is also applicable to the technical solution of the present application, and is not limited herein.

Fig. 16 is a schematic diagram illustrating a hardware structure of an electronic device 800 capable of implementing the mobile phone 200 according to some embodiments of the present application. Specifically, as shown in fig. 16, the electronic device 800 may include a processor 810, an external memory interface 820, an internal memory 821, a Universal Serial Bus (USB) interface 830, a charging management module 840, a power management module 841, a battery 842, an antenna 1, an antenna 2, a mobile communication module 850, a wireless communication module 860, an audio module 870, a speaker 870A, a receiver 870B, a microphone 870C, a headset interface 870D, a sensor module 880, a key 890, a motor 898, an indicator 892, a camera 893, a display 894, and a Subscriber Identification Module (SIM) card interface 895, and the like. The sensor module 880 may include a pressure sensor 880A, a gyroscope sensor 880B, an air pressure sensor 880C, a magnetic sensor 880D, an acceleration sensor 880E, a distance sensor 880F, a proximity light sensor 880G, a fingerprint sensor 880H, a temperature sensor 880J, a touch sensor 880K, an ambient light sensor 880L, a bone conduction sensor 880M, and the like.

It is to be understood that the illustrated structure of the embodiments of the invention is not to be construed as a specific limitation to the electronic device 800. In other embodiments of the present application, the electronic device 800 may include more or fewer components than illustrated, or combine certain components, or split certain components, or a different arrangement of components. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

Processor 810 may include one or more processing units, such as: the processor 810 may include an Application Processor (AP), a modem processor, a Graphics Processing Unit (GPU), an Image Signal Processor (ISP), a controller, a video codec, a Digital Signal Processor (DSP), a baseband processor, and/or a neural-Network Processing Unit (NPU), etc. The different processing units may be separate devices or may be integrated into one or more processors. For example, the processor 810 may perform preprocessing and specific analysis and calculation on the PPG sample data sent by the bracelet 100 and the PPG data acquired during the atrial fibrillation early warning process.

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

A memory may also be provided in processor 810 for storing instructions and data. In some embodiments, the memory in processor 810 is a cache memory. The memory may hold instructions or data that have just been used or recycled by the processor 810. If the processor 810 needs to use the instruction or data again, it can be called directly from the memory. Avoiding repeated accesses reduces the latency of the processor 810, thereby increasing the efficiency of the system. Meanwhile, the processor 810 may also store the PPG data of the user sent by the bracelet 100 received by the electronic device 800, extract the RR in the PPG data, and calculate the ADANN value.

In some embodiments, processor 810 may include one or more interfaces. The interface may include an integrated circuit (12C) interface, an integrated circuit built-in audio (12S) interface, a Pulse Code Modulation (PCM) interface, a universal asynchronous receiver/transmitter (UART) interface, a Mobile Industry Processor Interface (MIPI), a general-purpose input/output (GPIO) interface, a Subscriber Identity Module (SIM) interface, and/or a Universal Serial Bus (USB) interface, etc.

Micro USB interface, USB Type C interface etc.. The USB interface 830 may be used to connect a charger to charge the electronic device 800, and may also be used to transmit data between the electronic device 800 and a peripheral device. And the earphone can also be used for connecting an earphone and playing audio through the earphone. The interface may also be used to connect other electronic devices, such as AR devices and the like.

It should be understood that the connection relationship between the modules according to the embodiment of the present invention is only illustrative, and is not limited to the structure of the electronic device 800. In other embodiments of the present application, the electronic device 800 may also adopt different interface connection manners or a combination of multiple interface connection manners in the above embodiments.

The charging management module 840 is configured to receive charging input from a charger. The power management module 848 is used to connect the battery 842, the charge management module 840 and the processor 880. The power management module 848 receives input from the battery 842 and/or the charge management module 840 and provides power to the processor 880, the internal memory 821, the display 894, the camera 893, and the wireless communication module 860, among other things. The power management module 848 may also be used to monitor battery capacity, battery cycle count, battery state of health (leakage, impedance), etc. In some other embodiments, the power management module 841 may also be disposed in the processor 880. In other embodiments, the power management module 841 and the charging management module 840 may be disposed in the same device.

The wireless communication function of the electronic device 800 may be implemented by the antenna 1, the antenna 2, the mobile communication module 850, the wireless communication module 860, the modem processor, the baseband processor, and the like.

The antennas 1 and 2 are used for transmitting and receiving electromagnetic wave signals. Each antenna in the electronic device 800 may be used to cover a single or multiple communication bands. Different antennas can also be multiplexed to improve the utilization of the antennas. For example: the antenna 1 may be multiplexed as a diversity antenna of a wireless local area network. In other embodiments, the antenna may be used in conjunction with a tuning switch.

The mobile communication module 850 may provide a solution including 2G/3G/4G/5G wireless communication applied on the electronic device 800. The wireless communication module 860 may provide solutions for wireless communication applied to the electronic device 800, including Wireless Local Area Networks (WLANs), such as wireless fidelity (Wi-Fi) networks, Bluetooth (BT), Global Navigation Satellite Systems (GNSS), Frequency Modulation (FM), Near Field Communication (NFC), Infrared (IR), and the like. The wireless communication module 860 may be one or more devices that integrate at least one communication processing module. The wireless communication module 860 receives electromagnetic waves via the antenna 2, performs frequency modulation and filtering processing on electromagnetic wave signals, and transmits the processed signals to the processor 810. The wireless communication module 860 may also receive signals to be transmitted from the processor 810, frequency modulate them, amplify them, and convert them into electromagnetic waves via the antenna 2 to radiate them.

In some embodiments, electronic device 800 is capable of communicative connection with wristband 100 through mobile communication module 850 or wireless communication module 860.

In some embodiments, antenna 1 of electronic device 800 is coupled to mobile communication module 850 and antenna 2 is coupled to wireless communication module 860, such that electronic device 800 may communicate with networks and other devices via wireless communication techniques. The wireless communication technology may include global system for mobile communications (GSM), General Packet Radio Service (GPRS), code division multiple access (code division multiple access, CDMA), Wideband Code Division Multiple Access (WCDMA), time-division code division multiple access (time-division code division multiple access, TD-SCDMA), Long Term Evolution (LTE), LTE, BT, GNSS, WLAN, NFC, FM, and/or IR technologies, etc. The GNSS may include a Global Positioning System (GPS), a global navigation satellite system (GLONASS), a beidou navigation satellite system (BDS), a quasi-zenith satellite system (QZSS), and/or a Satellite Based Augmentation System (SBAS).

The electronic device 800 implements display functions via the GPU, the display screen 894, and the application processor, among other things. The GPU is a microprocessor for image processing, and is connected to a display screen 894 and an application processor. The GPU is used to perform mathematical and geometric calculations for graphics rendering. Processor 810 may include one or more GPUs that execute program instructions to generate or alter display information. For example, in some embodiments of the present application, a prompt message for warning the user of atrial fibrillation may be displayed on the display screen 894, and a human-computer interaction function may be implemented.

The electronic device 800 may implement a shooting function through the ISP, the camera 893, the video codec, the GPU, the display screen 894, and the application processor, etc.

The external memory interface 820 may be used to connect an external memory card, such as a Micro SD card, to extend the memory capability of the electronic device 800. The external memory card communicates with the processor 810 through the external memory interface 820 to implement data storage functions. For example, files such as music, video, etc. are saved in an external memory card.

The internal memory 821 may be used to store computer-executable program code, which includes instructions. The internal memory 821 may include a program storage area and a data storage area. The storage program area may store an operating system, an application program (such as a sound playing function, an image playing function, etc.) required by at least one function, and the like. The data storage area may store data (e.g., audio data, phone book, etc.) created during use of the electronic device 800, and the like. In addition, the internal memory 821 may include a high-speed random access memory, and may further include a nonvolatile memory, such as at least one magnetic disk storage device, a flash memory device, a universal flash memory (UFS), and the like. The processor 810 performs various functional applications and data processing of the electronic device 800 by executing instructions stored in the internal memory 821 and/or instructions stored in a memory provided in the processor.

Electronic device 800 may implement audio functionality via audio module 870, speaker 870A, receiver 870B, microphone 870C, headset interface 870D, and an application processor, among other things. Such as music playing, recording, etc.

The keys 890 include a power-on key, a volume key, and the like. The keys 890 may be mechanical keys. Or may be touch keys. The electronic device 800 may receive a key input, generate a key signal input related to user settings and function control of the electronic device 800.

The motor 891 may generate a vibration cue. The motor 891 may be used for incoming call vibration prompts, as well as for touch vibration feedback. For example, touch operations applied to different applications (e.g., photographing, audio playing, etc.) may correspond to different vibration feedback effects. The motor 891 may also respond to different vibration feedback effects for touch operations applied to different areas of the display screen 894. Different application scenes (such as time reminding, receiving information, alarm clock, game and the like) can also correspond to different vibration feedback effects. The touch vibration feedback effect may also support customization. For example, in some embodiments of the present application, cell phone 200 may alert the user of atrial fibrillation by vibrating.

Indicator 892 may be an indicator light that may be used to indicate a state of charge, a change in charge, or a message, missed call, notification, etc.

The SIM card interface 895 is used to connect a SIM card.

Referring now to fig. 17, the software system of the electronic device 800 may employ a hierarchical architecture, an event-driven architecture, a microkernel architecture, a microservice architecture, or a cloud architecture. The embodiment of the invention takes an Android system with a layered architecture as an example, and exemplarily illustrates a software structure of a terminal device. Fig. 17 is a block diagram of a software configuration of a terminal device according to an embodiment of the present invention.

The layered architecture divides the software into several layers, each layer having a clear role and division of labor. The layers communicate with each other through a software interface. In some embodiments, the Android system is divided into four layers, an application layer, an application framework layer, an Android runtime (Android runtime) and system library, and a kernel layer from top to bottom.

The application layer may include a series of application packages.

As shown in fig. 17, the application package may include phone, camera, gallery, calendar, talk, map, navigation, WLAN, bluetooth, music, video, short message, etc. applications.

The application framework layer provides an Application Programming Interface (API) and a programming framework for the application program of the application layer. The application framework layer includes a number of predefined functions.

As shown in FIG. 17, the application framework layers may include a window manager, content provider, view system, phone manager, resource manager, notification manager, and the like.

The window manager is used for managing window programs. The window manager can obtain the size of the display screen, judge whether a status bar exists, lock the screen, intercept the screen and the like.

The content provider is used to store and retrieve data and make it accessible to applications. The data may include video, images, audio, calls made and received, browsing history and bookmarks, phone books, etc.

The view system includes visual controls such as controls to display text, controls to display pictures, and the like. The view system may be used to build applications. The display interface may be composed of one or more views. For example, the display interface including the short message notification icon may include a view for displaying text and a view for displaying pictures.

The telephone manager is used for providing a communication function of the terminal equipment. Such as management of call status (including on, off, etc.).

The resource manager provides various resources for the application, such as localized strings, icons, pictures, layout files, video files, and the like.

The notification manager enables the application to display notification information in the status bar, can be used to convey notification-type messages, can disappear automatically after a short dwell, and does not require user interaction. Such as a notification manager used to inform download completion, message alerts, etc. The notification manager may also be a notification that appears in the form of a chart or scroll bar text at the top status bar of the system, such as a notification of a background running application, or a notification that appears on the screen in the form of a dialog window. For example, text information is prompted in the status bar, a prompt tone is given, the terminal device vibrates, an indicator light flickers, and the like.

The Android Runtime comprises a core library and a virtual machine. The Android runtime is responsible for scheduling and managing an Android system.

The core library comprises two parts: one part is a function which needs to be called by java language, and the other part is a core library of android.

The application layer and the application framework layer run in a virtual machine. And executing java files of the application program layer and the application program framework layer into a binary file by the virtual machine. The virtual machine is used for performing the functions of object life cycle management, stack management, thread management, safety and exception management, garbage collection and the like.

The system library may include a plurality of functional modules. For example: surface managers (surface managers), Media Libraries (Media Libraries), three-dimensional graphics processing Libraries (e.g., OpenGL ES), 2D graphics engines (e.g., SGL), and the like.

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

The media library supports a variety of commonly used audio, video format playback and recording, and still image files, among others. The media library may support a variety of audio-video encoding formats, such as: MPEG4, H.264, MP3, AAC, AMR, JPG, PNG, etc.

The three-dimensional graphic processing library is used for realizing three-dimensional graphic drawing, image rendering, synthesis, layer processing and the like.

The 2D graphics engine is a drawing engine for 2D drawing.

The kernel layer is a layer between hardware and software. The inner core layer at least comprises a display driver, a camera driver, an audio driver and a sensor driver.

Referring now to FIG. 18, shown is a block diagram of a system 1800 capable of implementing the functionality of the electronic device 300 in accordance with one embodiment of the present application. Fig. 18 schematically illustrates an example system 1800 in accordance with various embodiments. In one embodiment, system 1800 may include one or more processors 1804, system control logic 1808 coupled to at least one of processors 1804, system Memory 1812 coupled to system control logic 1808, Non-volatile Memory 1816 coupled to system control logic 1808, and network interface 1820 coupled to system control logic 1808.

In some embodiments, processor 1804 may include one or more single-core or multi-core processors. In some embodiments, the processor 1804 may include any combination of general-purpose processors and special-purpose processors (e.g., graphics processors, application processors, baseband processors, etc.). In embodiments in which system 1800 employs an enhanced Node B (eNB) or a Radio Access Network (RAN) controller, processor 1804 may be configured to perform various consistent embodiments, e.g., as one or more of the various embodiments shown in fig. 1-10.

In some embodiments, the system control logic 1808 may include any suitable interface controller to provide any suitable interface to at least one of the processors 1804 and/or to any suitable device or component in communication with the system control logic 1808.

In some embodiments, system control logic 1808 may include one or more memory controllers to provide an interface to system memory 1812. System memory 1812 may be used to load and store data and/or instructions. The Memory 1812 of the system 1800 may include any suitable volatile Memory, such as suitable Dynamic Random-Access Memory (DRAM), in some embodiments.

The NVM/memory 1816 may include one or more tangible, non-transitory computer-readable media for storing data and/or instructions. In some embodiments, the NVM/memory 1816 may include any suitable non-volatile memory, such as flash memory, and/or any suitable non-volatile storage device, such as at least one of a Hard Disk Drive (HDD), Compact Disc (CD) Drive, and Digital Versatile Disc (DVD) Drive.

The NVM/memory 1816 may include a portion of the storage resources on the device on which the system 1800 is installed, or it may be accessible by, but not necessarily a part of, the device. For example, the NVM/storage 1816 may be accessed over a network via the network interface 1820.

In particular, system memory 1812 and NVM/storage 1816 may include: a temporary copy and a permanent copy of instructions 1824. The instructions 1824 may include: instructions that when executed by at least one of the processors 1804 cause the system 1800 to perform a method as shown in fig. 1-10. In some embodiments, the instructions 1824, hardware, firmware, and/or software components thereof may additionally/alternatively be disposed in the system control logic 1808, the network interface 1820, and/or the processor 1804.

Network interface 1820 may include a transceiver to provide a radio interface for system 1800 to communicate with any other suitable device (e.g., front end module, antenna, etc.) over one or more networks. In some embodiments, network interface 1820 may be integrated with other components of system 1800. For example, the network interface 1820 may be integrated with at least one of the processor 1804, the system memory 1812, the NVM/storage 1816, and a firmware device (not shown) having instructions that, when executed by at least one of the processors 1804, implement the methods shown in fig. 1-10.

The network interface 1820 may further include any suitable hardware and/or firmware to provide a multiple-input multiple-output radio interface. For example, network interface 1820 may be a network adapter, a wireless network adapter, a telephone modem, and/or a wireless modem.

In one embodiment, at least one of the processors 1804 may be packaged together with logic for one or more controllers of the System control logic 1808 to form a System In a Package (SiP). In one embodiment, at least one of the processors 1804 may be integrated on the same die with logic for one or more controllers of the System control logic 1808 to form a System on Chip (SoC).

System 1800 may further include: input/output (I/O) devices 1832. I/O device 1832 may include a user interface to enable a user to interact with system 1800; the design of the peripheral component interface enables peripheral components to also interact with the system 1800. In some embodiments, system 1800 further includes sensors for determining at least one of environmental conditions and location information associated with system 1800.

Claims

1. An atrial fibrillation early warning method based on electronic equipment is characterized by comprising the following steps:

an electronic device extracting a plurality of sinus cardiac intervals in photoplethysmography data of a user acquired by the electronic device;

the electronic device calculating a degree of change in a relative difference between a plurality of temporally adjacent sinus intervals of the plurality of sinus intervals;

and the electronic equipment determines whether to give an atrial fibrillation early warning to the user or not based on the calculated change degree.

2. The method of claim 1, comprising:

the relative difference between the temporally adjacent sinus cardiac intervals is:

a ratio of a difference between two sinus intervals that are temporally adjacent to one another to a larger one of the two sinus intervals; and is

The degree of change in the relative difference is an average of the relative differences between the plurality of temporally adjacent sinus cardiac intervals.

3. The method of claim 2, comprising:

the electronic device calculates an average of relative differences between the adjacent sinus intervals by:

4. The method of claim 1, comprising:

and the electronic equipment displays prompt information for atrial fibrillation early warning to the user under the condition that the electronic equipment determines that the atrial fibrillation early warning is performed on the user.

5. An atrial fibrillation early warning method based on electronic equipment is characterized by comprising the following steps:

the method comprises the steps that a first electronic device receives photoplethysmography data, collected by a second electronic device, of a user from the second electronic device;

the first electronic device extracts a plurality of sinus inter-cardiac intervals in the acquired photoplethysmography data;

the first electronic device calculating a degree of change in a relative difference between a plurality of temporally adjacent sinus intervals of the plurality of sinus intervals;

and the first electronic equipment determines whether to give an atrial fibrillation early warning to the user or not based on the calculated change degree.

6. The method of claim 5, comprising:

the relative difference between the temporally adjacent sinus cardiac intervals is: a ratio of a difference between two sinus intervals that are temporally adjacent to one another to a larger one of the two sinus intervals; and is

7. The method of claim 6, comprising:

the first electronic device calculates an average of relative differences between the adjacent sinus intervals by:

8. The method of claim 5, further comprising:

and sending an early warning instruction to the second equipment under the condition that the first electronic equipment determines to carry out atrial fibrillation early warning on the user, wherein the early warning instruction is used for instructing the second equipment to carry out atrial fibrillation early warning on the user.

9. The method of claim 5, further comprising:

and under the condition that the first electronic equipment determines to carry out atrial fibrillation early warning on the user, the first electronic equipment displays prompt information for the atrial fibrillation early warning on the user.

10. A training method of an atrial fibrillation early warning model is characterized by comprising the following steps:

acquiring sample data of photoplethysmography and an expected result corresponding to the sample data;

extracting a plurality of sinus cardiac intervals in the acquired sample data of the photoplethysmography, and calculating the variation degree of the relative difference value between a plurality of temporally adjacent sinus cardiac intervals in the plurality of sinus cardiac intervals;

inputting the calculated degree of change into a decision tree model, wherein nodes in the decision tree model comprise nodes characterized by the calculated degree of change;

and acquiring a prediction result of the decision tree model, and adjusting various parameter values in the decision tree model according to the prediction result and the expected result.

11. The method of claim 10, further comprising:

calculating a first average value of the degree of change of sample data of which the expected result is an atrial fibrillation event in the acquired sample data of the photoplethysmography, and calculating a second average value of the degree of change of sample data of which the expected result is a non-atrial fibrillation event in the acquired sample data of the photoplethysmography;

inputting the calculated first average value and the second average value of the degree of change into the decision tree model, wherein nodes in the decision tree model comprise nodes which take the calculated first average value and the calculated second average value as features respectively.

12. The method of claim 10, wherein the relative difference between temporally adjacent sinus cardiac intervals is:

13. The method of claim 12, wherein the electronic device calculates an average of the relative difference between the adjacent sinus intervals by:

14. A readable medium of an electronic device, characterized in that the readable medium has stored thereon instructions which, when executed on the electronic device, cause the electronic device to perform the method of any of claims 1 to 13.

15. An electronic device, comprising: a memory for storing instructions for execution by one or more processors of the system, and the processor, being one of the processors of the electronic device, for performing the method of any one of claims 1 to 13.