WO2021208656A1

WO2021208656A1 - Sleep risk prediction method and apparatus, and terminal device

Info

Publication number: WO2021208656A1
Application number: PCT/CN2021/081009
Authority: WO
Inventors: 张慧; 李靖
Original assignee: 华为技术有限公司
Priority date: 2020-04-17
Filing date: 2021-03-16
Publication date: 2021-10-21
Also published as: CN113520343A

Abstract

A sleep risk prediction method and apparatus, and a terminal device. The method comprises: during sleep, collecting a plurality of kinds of physiological data of a user to be detected (S301); separately extracting the feature information of each of the plurality of kinds of physiological data (S302); and inputting the extracted feature information into a preset classifier so as to obtain sleep risk prediction information outputted by the classifier, wherein the classifier is obtained by performing training by means of the sample physiological data of a plurality of sample users (S303). The method predicts the risk of the user to be detected suffering from various subtypes of sleep apnea syndrome, and can achieve personalized sleep quality monitoring and sleep apnea risk estimation.

Description

Sleep risk prediction method, device and terminal equipment

This application claims the priority of a Chinese patent application filed with the State Intellectual Property Office on April 17, 2020, the application number is 202010306323.9, and the application name is "Sleep Risk Prediction Method, Device and Terminal Equipment", the entire content of which is incorporated by reference In this application.

Technical field

This application relates to the field of information technology, in particular to a sleep risk prediction method, device and terminal equipment.

Background technique

Sleep Apnea Syndrome (Sleep Apnea Syndrome, SAS) is a common chronic disease of sleep disorders. According to the World Health Organization (WHO) report, 1-10% of the world’s population is affected by sleep apnea. Among the 30-60 year olds in China, more than 4% of men and 2% of women are affected by sleep apnea. The effects of sleep apnea. The incidence of sleep apnea increases with age, reaching a peak in people aged 55-60.

According to the different pathogenesis, sleep apnea syndrome can be divided into three different subtypes: obstructive type, central type and mixed type. The etiology and harm of different subtypes are also different. Among them, obstructive sleep apnea is sleep apnea caused by the upper airway obstruction and narrowing of the airway caused by the relaxation of the soft tissues near the throat; central apnea is due to the respiratory center having been damaged by strokes and traumas, and it cannot be normal. The conveyed breathing instructions cause sleep breathing skills disorders; mixed sleep apnea is a sleep disorder caused by a combination of the above reasons.

Since the occurrence and development of sleep apnea is a chronic and gradual process, early prediction of possible risks during sleep can help improve the quality of life of patients and prevent the occurrence of various complications.

Summary of the invention

The embodiments of the present application provide a sleep risk prediction method, device, and terminal equipment, which can solve the problem that sleep risk cannot be predicted simply and quickly in the prior art.

In the first aspect, the embodiments of the present application provide a method for recognizing sleep apnea syndrome, including:

Collect a variety of physiological data of the user to be tested during sleep;

Extracting feature information of each type of physiological data in the multiple types of physiological data respectively;

The extracted feature information is input into a preset classifier to obtain sleep risk prediction information output by the classifier; wherein the classifier is obtained by training on sample physiological data of a plurality of sample users.

By collecting the physiological data of the user through the wearable device, and extracting the corresponding characteristic information from it, the above-mentioned characteristic information can be input into the pre-trained classifier to predict whether the user to be detected has a certain sleep risk during sleep, and various types of sleep risk. The probability of the risk. This embodiment can simply and quickly monitor the user's sleep quality, predict the risk of potential patients with sleep apnea, remind the user to seek medical treatment in time, and prevent the further development of sleep apnea and other complications.

In a possible implementation of the first aspect, the various physiological data of the user to be detected during sleep includes at least pulse wave data, blood oxygen data, and/or sound data.

In a possible implementation of the first aspect, the photoplethysmographic feature information of the user to be detected can be extracted from the pulse wave data, such as heart rate variability feature information, respiratory wave feature information, photoplethysmographic waveform feature information, etc. ; Can extract the blood oxygen characteristic information of the user to be detected from the blood oxygen data, such as oxygen depletion index and low blood oxygen accumulation time, etc.; Can extract the sound characteristic information of the user to be detected from the sound data, such as Mel frequency cepstrum Number and Fourier spectrum feature information, etc.

In a possible implementation of the first aspect, the preset classifier may include a two-classifier. Therefore, the extracted feature information can be input to the second classifier, and by receiving the recognition result of the aforementioned feature information output by the second classifier, it can be determined whether the user has a sleep apnea event during sleep. If the recognition result of the two-classifier is that a sleep apnea event occurs, it can further predict the risk of various subtypes of sleep apnea syndrome of the user to be detected based on the collected voice data of the user to be detected.

In a possible implementation of the first aspect, when further predicting the risk of various subtypes of sleep apnea syndrome in the user to be detected based on the voice data of the user to be detected, the to-be-identified user may first be determined The duration of a sleep apnea event is to determine the duration of an apnea event. If within the duration, the user's voice data includes intermittent snoring, it can be predicted that the user’s sleep risk is the risk of obstructive sleep apnea syndrome; if within the duration, the voice data does not include snoring, It can be predicted that the user’s sleep risk is the central sleep apnea syndrome risk; if the sleep risk is predicted based on the sound data within the duration, it includes both the obstructive sleep apnea syndrome risk and the central sleep apnea syndrome risk , It can be predicted that the user’s sleep risk is the risk of mixed sleep apnea syndrome.

In a possible implementation of the first aspect, determining the duration of the sleep apnea event to be recognized can be determined according to the accumulation time of a continuous hypoxemia in the blood oxygen data, so as to eliminate a continuous hypoxemia. The cumulative time is taken as the duration of the sleep apnea event to be recognized.

In a possible implementation manner of the first aspect, the foregoing preset classifier may further include four classifiers. Therefore, the extracted feature information can be input into the four classifiers, and by receiving the identification results of the feature information output by the four classifiers, directly based on the identification results, the user's risk of various subtypes of sleep apnea syndrome can be predicted . The recognition results of the above four classifiers may include no sleep apnea events, obstructive sleep apnea events, central sleep apnea events, or mixed sleep apnea events.

In the second aspect, an embodiment of the present application provides a sleep risk prediction device, including:

The collection module is used to collect various physiological data of the user to be detected during sleep;

An extraction module for extracting feature information of each type of physiological data in the multiple types of physiological data;

The prediction module is configured to input the extracted feature information into a preset classifier to obtain sleep risk prediction information output by the classifier; wherein, the classifier is obtained by training on sample physiological data of a plurality of sample users.

In a possible implementation of the second aspect, the multiple types of physiological data include at least pulse wave data, blood oxygen data, and/or sound data.

In a possible implementation manner of the second aspect, the extraction module may specifically include the following submodules:

The photoplethysmographic feature information extraction sub-module is used to extract the photoplethysmographic feature information of the user to be detected from the pulse wave data. The photoplethysmographic feature information includes heart rate variability feature information, respiratory wave feature information, and At least one of the characteristic information of the photoplethysmography waveform; and/or,

The blood oxygen feature information extraction submodule is used to extract blood oxygen feature information of the user to be detected from the blood oxygen data, where the blood oxygen feature information includes at least one of oxygen depletion index and low blood oxygen accumulation time ;and / or,

The voice feature information extraction submodule is used to extract voice feature information of the user to be detected from the voice data, where the voice feature information includes at least one of Mel frequency cepstrum coefficient and Fourier spectrum feature information .

In a possible implementation of the second aspect, the classifier may include a two-classifier; the prediction module may specifically include the following sub-modules:

The first input submodule is used to input the extracted feature information into the second classifier;

The first receiving sub-module is configured to receive the recognition result of the feature information output by the second classifier, and the recognition result of the second classifier includes the occurrence of a sleep apnea event or the absence of a sleep apnea event;

The first prediction sub-module is used to determine the risk of various subtypes of sleep apnea syndrome in the user to be detected based on the voice data of the user to be detected if the recognition result is that a sleep apnea event occurs Make predictions.

In a possible implementation manner of the second aspect, the first prediction submodule may specifically include the following units:

The determining unit is used to determine the duration of the sleep apnea event to be recognized;

The first prediction unit is configured to predict that the sleep risk of the user to be detected is the risk of obstructive sleep apnea syndrome if the voice data of the user to be detected includes intermittent snoring within the duration;

The second prediction unit is configured to predict that the sleep risk of the user to be detected is the risk of central sleep apnea syndrome if the sound data of the user to be detected does not include snoring within the duration;

The third prediction unit is configured to predict that the sleep risk includes both the obstructive sleep apnea syndrome risk and the central sleep apnea syndrome based on the voice data of the user to be detected within the duration If the risk of symptom is detected, the sleep risk of the user to be detected is predicted to be the risk of mixed sleep apnea syndrome.

In a possible implementation manner of the second aspect, the determining unit may specifically include the following subunits:

The determining subunit is used to determine that a continuous hypoxemia accumulation time occurs in the blood oxygen data, and use the one continuous hypoxemia accumulation time as the duration of the sleep apnea event to be recognized.

In a possible implementation of the second aspect, the classifier may also include a four-classifier; the prediction module may also include the following sub-modules:

The second input submodule is used to input the feature information into the four classifier;

The second receiving sub-module is configured to receive the recognition results of the feature information output by the four classifiers. The recognition results of the four classifiers include no sleep apnea events, obstructive sleep apnea events, and Central sleep apnea event or mixed sleep apnea event;

The second prediction sub-module is used to predict the risk of various subtypes of sleep apnea syndrome in the user to be detected according to the recognition result.

In the third aspect, an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, The sleep risk prediction method according to any one of the above-mentioned first aspects is realized.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of a terminal device, any one of the above-mentioned aspects of the first aspect is implemented. The described sleep risk prediction method.

In a fifth aspect, the embodiments of the present application provide a computer program product that, when the computer program product runs on a terminal device, causes the terminal device to execute the sleep risk prediction method described in any one of the above-mentioned first aspects.

Compared with the prior art, the embodiments of the present application include the following beneficial effects:

In the embodiment of this application, the user’s heart rate, heart rate variability, blood oxygen, and snoring information in the sound and other physiological data can be collected by the wearable device to predict the occurrence of sleep apnea events and subtypes of the user, and give sleep breathing The quantitative index of pause events helps to improve the accuracy of sleep apnea event prediction. Secondly, this embodiment can realize personalized sleep quality monitoring and sleep apnea risk assessment by classifying specific subtypes of sleep apnea syndrome. Third, this embodiment uses a wearable device to detect the user's sleep quality, which has the advantages of low cost, non-invasiveness, and simple operation. It can intelligently monitor the user's sleep breathing quality and remind the user to help the user better adjust sleep. Due to the portability of the wearable device, the user can wear it for a long time anytime, anywhere, and can predict the risk of potential sleep apnea patients in a timely and early manner, remind the user to seek medical treatment in time, and prevent the further development of sleep apnea and other complications. This embodiment not only improves the accuracy of sleep apnea prediction, but can also predict specific subtypes of sleep apnea, which helps to improve the quality of life of patients and prevent the occurrence of various complications.

Description of the drawings

FIG. 1 is a schematic diagram of the principle of a sleep risk prediction method provided by an embodiment of the present application;

2 is a schematic diagram of the principle of a sleep risk prediction method provided by another embodiment of the present application;

FIG. 3 is a schematic step flowchart of a sleep risk prediction method provided by an embodiment of the present application;

4 is a schematic step flowchart of a sleep risk prediction method provided by another embodiment of the present application;

FIG. 5 is a structural block diagram of a sleep risk prediction device provided by an embodiment of the present application;

FIG. 6 is a structural block diagram of a sleep risk prediction device provided by another embodiment of the present application;

FIG. 7 is a schematic structural diagram of a terminal device provided by an embodiment of the present application.

Detailed ways

Sleep apnea syndrome is a potentially fatal sleep breathing disease. Early and reasonable diagnosis and treatment of sleep apnea syndrome can significantly improve the patient’s quality of life, prevent the occurrence of various complications, and improve the patient’s Survival rate. However, because patients usually cannot know that they have sleep apnea syndrome, they need to be informed by the person next to the pillow. In today's society, more and more people live alone, and the risk of sleep apnea syndrome is extremely easy to be ignored; secondly, due to the current situation There is no relevant technology to detect the risk of sleep apnea in a sustainable and preventive manner. Given the shortage of existing medical resources, it is not realistic to implement extensive monitoring.

In the prior art, polysomnography is usually used clinically to detect sleep apnea, which is an internationally recognized gold standard for diagnosing sleep apnea syndrome. Polysomnography is to continuously detect breathing, arterial oxygen saturation, EEG, heart rate, snoring and other indicators at night to understand whether the subject has apnea, type of apnea, number of pauses, time of pause, etc. The minimum arterial blood oxygen level when the pause occurs and the degree of health impact. However, as a clinical monitoring device, polysomnography has complex equipment, many electrodes, and is expensive, requiring professional medical personnel to operate and interpret the results; secondly, polysomnographs have large weight and volume, poor portability and comfort, and are in use. Sometimes the subject has a sense of restraint, which is not conducive to the patient’s sleep, is not suitable for long-term wear and continuous monitoring, and cannot be used for large-scale screening.

At present, there are also some simplified sleep apnea syndrome detection programs, which obtain data from sensors in different locations to diagnose whether the subject has sleep apnea syndrome. For example, the pressure sensor placed on the mattress can detect the displacement and pressure change of the chest; the infrared absorption sensor placed on the nose and mouth can detect the change of carbon dioxide (CO2) in the air between the nose and mouth; by placing 50-150 cm on the head The microphone at (cm) monitors the changes of snoring sound; monitors the changes of heart rate and blood oxygen saturation through fingertip photoplethysmography. However, the above-mentioned detection scheme also has many shortcomings. Including, the detection scheme through the mattress pressure sensor is easily affected by the sleeping position, the detection scheme through the infrared sensor placed in the nose and mouth is not beautiful, the scheme of monitoring snoring through the microphone is inaccurate and easy to be disturbed by noise. The fingertip photoplethysmographic detection program will press the fingertips for a long time and make the subject uncomfortable. The most important thing is that such detection schemes use single parameters and low accuracy, and it is impossible to classify the subtypes of sleep apnea syndrome based on the test results.

Therefore, in order to solve the above problems, the core idea of the embodiments of the present application is proposed to monitor the user's heart rate (Heart rate), heart rate variability (HRV), blood oxygen (Blood oxygen), and sound Snoring information and other physiological parameters, predict sleep apnea events and their subtypes, give quantitative indicators of sleep apnea events, improve the accuracy of sleep apnea event prediction, and pass the analysis of sleep apnea syndrome Subtypes are classified to achieve personalized sleep quality monitoring and sleep apnea risk prediction.

As shown in FIG. 1, it is a schematic diagram of the principle of a sleep risk prediction method provided by an embodiment of the present application. As shown in Figure 1, when predicting the risk of various subtypes of sleep apnea syndrome, you can first use wearable devices to detect the user's pulse wave, blood oxygen, sound and other physiological data, and extract from these data Corresponding photoplethysmograph (Photoplethysmograph, PPG) characteristic information, blood oxygen characteristic information, sound characteristic information, etc. Then, based on the extracted feature information and a preset multi-parameter fusion classifier, the risk of a sleep apnea event of the user during sleep can be predicted. If it is confirmed that a sleep apnea event has occurred, the sound data can be used to further predict the user's risk of a specific subtype of sleep apnea. If the user has intermittent snoring during the sleep apnea event, it can be determined that the user is at risk of obstructive sleep apnea; if there is no snoring, it can be determined that the user is at risk of central sleep apnea; if during a sleep apnea During the process, if central sleep apnea occurs first and then obstructive sleep apnea occurs, this event can be determined as the user is at risk of mixed sleep apnea.

As shown in FIG. 2, it is a schematic diagram of the principle of a sleep risk prediction method provided by another embodiment of the present application. As shown in Figure 2, when predicting the risk of various subtypes of sleep apnea syndrome, it is also possible to first use wearable devices to detect the user's physiological data such as pulse wave, blood oxygen, and sound, and extract from these data The corresponding PPG feature information, blood oxygen feature information, and sound feature information are displayed. Then, the specific subtype of sleep apnea can be directly predicted based on the extracted feature information and the preset multi-parameter fusion classifier. That is, the classification results of the classifier in Figure 2 include four categories: no sleep apnea events, obstructive sleep apnea events, central sleep apnea events, and mixed sleep apnea events.

It should be noted that the aforementioned wearable devices include, but are not limited to, watches, bracelets and other devices. The above-mentioned PPG characteristics may include HRV characteristics, respiratory wave characteristics, PPG waveform characteristics, etc.; blood oxygen characteristics may include oxygen reduction index, low blood oxygen accumulation time and other characteristics; sound characteristics may include Mel Frequency Cepstrum Coefficient (Mel Frequency Cepstrum Coefficient, MFCC), that is, the non-linear Mel transform feature of the spectrum, the logarithm of the Mel spectrum (LogMel), the Fourier spectrum and other features are realized through the Mel formula.

The following describes the sleep risk prediction method of the present application in conjunction with specific embodiments.

In the following description, for the purpose of illustration rather than limitation, specific details such as a specific system structure and technology are proposed for a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from obstructing the description of this application.

The terms used in the following embodiments are only for the purpose of describing specific embodiments, and are not intended to limit the application. As used in the specification and appended claims of this application, the singular expressions "a", "an", "said", "above", "the" and "this" are intended to also This includes expressions such as "one or more" unless the context clearly indicates to the contrary. It should also be understood that in the embodiments of the present application, "one or more" refers to one, two or more than two; "and/or" describes the association relationship of the associated objects, indicating that there may be three relationships; for example, A and/or B can mean the situation where A exists alone, A and B exist at the same time, 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.

Referring to FIG. 3, a schematic step flowchart of a sleep risk prediction method provided by an embodiment of the present application is shown. As an example and not a limitation, the method can be applied to a terminal device, and the method can specifically include the following steps:

S301. Collect various physiological data of the user to be detected during sleep;

It should be noted that this method can be applied to terminal devices. The above-mentioned terminal devices can be wearable devices such as smart watches and smart bracelets; or, the terminal devices can also be other devices that have the function of collecting and processing physiological data of users. For smart devices, this embodiment does not limit the specific types of terminal devices.

Taking the terminal device as a smart bracelet as an example, after the user wears the smart bracelet, the user can collect physiological data of the user through a variety of sensors integrated in the bracelet, including various physiological data of the user in the sleep state. Such as pulse wave data, blood oxygen data, sound data, etc., this embodiment does not limit the specific types of physiological data that can be collected.

S302: Extract characteristic information of each type of physiological data in the multiple types of physiological data respectively;

In the embodiments of the present application, for different physiological data, characteristic information that can be used for subsequent data analysis and processing can be extracted according to its characteristics. Through the corresponding feature information, it is analyzed whether the user to be detected has an apnea event during sleep, and the specific subtype of the time.

For example, for pulse wave data, PPG feature information of the user to be detected can be extracted from it. The PPG feature information may include one or more of HRV feature information, respiratory wave feature information, and PPG waveform feature information.

For blood oxygen data, the corresponding blood oxygen characteristic information can be extracted from it, such as oxygen depletion index, low blood oxygen accumulation time, and so on.

For sound data, corresponding sound feature information can also be extracted from it, such as Mel frequency cepstrum coefficient, Mel frequency logarithm, Fourier spectrum and other feature information.

Of course, the above description is only an example of this embodiment. According to actual needs, a variety of different feature information can be collected for different physiological data. This embodiment is about how to extract corresponding feature information from the collected physiological data. , And what kind of feature information to extract is not limited.

S303. Input the extracted feature information into a preset classifier to obtain sleep risk prediction information output by the classifier; wherein the classifier is obtained by training on sample physiological data of a plurality of sample users.

In the embodiment of the present application, the preset classifier may be obtained based on a supervised machine learning model training. That is, by collecting the physiological data of multiple sample users, and performing model training based on these physiological data, the corresponding multi-parameter fusion classifier is obtained.

In a specific implementation, the sleep process of sample users can be monitored, and various physiological data of each sample user's sleep can be collected, and at the same time, whether each sample user has an apnea event. Then, the corresponding feature information is extracted from the collected physiological data, and the respective feature information of the sample users with and without apnea events are input into each classifier for model training, and then the output can be obtained. A multi-parameter fusion classifier with two classification results. The above two classification results include the occurrence of apnea events or the absence of apnea events.

It should be noted that when classifier training is performed, multiple classifiers can be used separately, and the classifier with the best training effect is selected from the classifier as the target classifier for recognizing sleep apnea syndrome in the wearable device.

Since the output result of the two-category classifier only includes whether there is an apnea event, other data can be used to further predict the occurrence of various subtypes of sleep apnea syndrome for the user to be detected who has an apnea event. risks of. For example, the probability of occurrence of a specific subtype of sleep apnea syndrome can be predicted based on the sound data in the physiological data.

In specific implementation, if the user to be detected has intermittent snoring during a sleep apnea event, it can be considered that this event may be caused by obstructive sleep apnea, so it can be predicted that the user will appear during sleep The risk of obstructive sleep apnea is greater; if there is no snoring, it can be considered that this event was caused by central sleep apnea, so it can be predicted that the user will have a greater risk of central sleep apnea during sleep; If during an apnea process, central apnea occurs first, followed by obstructive apnea, it can be predicted that the user may be at risk of mixed sleep apnea.

Of course, when collecting the physiological data of the sample user, the specific subtype to which the apnea syndrome of the sample user belongs can also be directly marked according to the monitoring results, so that in the subsequent classifier training, a four-class classifier can be trained The recognition result of the above-mentioned four-category classifier may include no sleep apnea event, obstructive sleep apnea event, central sleep apnea event, or mixed sleep apnea event.

Therefore, after collecting various physiological data of the user to be detected, and extracting the characteristic information of each physiological data, the above-mentioned characteristic information can be input into the four-classifier obtained by training, and by receiving the characteristic information output by the four-classifier According to the recognition result, the risk of various subtypes of sleep apnea syndrome can be directly predicted based on the recognition result.

In the embodiments of the present application, the user’s heart rate, heart rate variability, blood oxygen, and snoring information in the sound can be collected by the wearable device to predict the occurrence of sleep apnea events and their subtypes. The quantitative indicators of sleep apnea events can help improve the accuracy of sleep apnea event prediction. Secondly, this embodiment can realize personalized sleep quality monitoring and sleep apnea risk assessment by classifying specific subtypes of sleep apnea syndrome. Third, this embodiment uses a wearable device to detect the user's sleep quality, which has the advantages of low cost, non-invasiveness, and simple operation. It can intelligently monitor the user's sleep breathing quality and remind the user to help the user better adjust sleep. Due to the portability of the wearable device, the user can wear it for a long time anytime, anywhere, and can predict the risk of potential sleep apnea patients in a timely and early manner, remind the user to seek medical treatment in time, and prevent the further development of sleep apnea and other complications. This embodiment not only improves the accuracy of sleep apnea prediction, but can also predict specific subtypes of sleep apnea, which helps improve the patient's quality of life and prevent the occurrence of various complications.

Referring to FIG. 4, a schematic step flowchart of a sleep risk prediction method provided by another embodiment of the present application is shown. The method may specifically include the following steps:

S401. Collect a variety of physiological data of the user to be detected during sleep, and extract characteristic information of each of the multiple types of physiological data.

It should be noted that S401 in this embodiment is similar to S301-302 in the previous embodiment, and can be referred to each other, which is not repeated in this embodiment.

S402. Input the extracted characteristic information into the second classifier, and receive a recognition result of the characteristic information output by the second classifier, where the recognition result of the second classifier includes the occurrence of a sleep apnea event;

In the embodiment of the present application, for the extracted feature information of various physiological data of the user to be detected, the feature information can be input into a preset multi-parameter fusion classifier to obtain a corresponding classification result.

The multi-parameter fusion classifier in this embodiment may be a two-classifier, that is, the classification results of the two classifiers include two types.

In a specific implementation, supervised model training can be performed by collecting physiological data of multiple sample users, so as to obtain a two-classifier that can output whether a sleep apnea event occurs.

In the embodiment of the present application, after the characteristic information of the user to be detected is input into the above-mentioned two classifier, the corresponding classification result can be obtained. That is, a sleep apnea event occurred during the sleep of the user to be detected, or no sleep apnea event occurred.

For a user to be detected who has a sleep apnea event, the risk of various subtypes of sleep apnea syndrome can be further predicted based on the user's voice data.

S403: Determine the duration of the sleep apnea event to be recognized;

In the embodiments of the present application, the prediction of specific subtypes of apnea syndrome can be performed based on an apnea event. That is, in the process of an apnea event, predict the probability that the event belongs to a specific subtype.

Therefore, the duration of the sleep apnea event to be recognized may refer to the duration of an apnea event.

In the embodiment of the present application, it is possible to determine that a continuous hypoxemia accumulation time occurs in the blood oxygen data, and then use a continuous hypoxemia accumulation time as the duration of the sleep apnea event to be identified.

Usually, when an apnea event occurs, it is often accompanied by physiological conditions such as hypoxemia. Therefore, a hypoxemia accumulation time can be used as a criterion for determining the occurrence of an apnea event.

Of course, in hypoxemia, physiological conditions such as decreased oxygen depletion index may also occur. Therefore, it is also possible to determine whether an apnea event has occurred in combination with the two indicators of low blood oxygen accumulation time and oxygen depletion index, which is not limited in this embodiment.

S404. If the voice data of the user to be detected includes intermittent snoring within the duration, predict that the sleep risk of the user to be detected is the risk of obstructive sleep apnea syndrome;

In the embodiment of the present application, if intermittent snoring of the user is detected during an apnea event, it can be considered that the user has a greater risk of developing obstructive sleep apnea syndrome.

S405. If snoring is not included in the voice data of the user to be detected within the duration, predict that the sleep risk of the user to be detected is a central sleep apnea syndrome risk;

If the user's intermittent snoring is not detected during an apnea event, it can be considered that the user is at greater risk of developing central sleep apnea syndrome.

S406. If, within the duration, the sleep risk is predicted to include both the obstructive sleep apnea syndrome risk and the central sleep apnea syndrome risk based on the voice data of the user to be detected, predict The sleep risk of the user to be detected is the risk of mixed sleep apnea syndrome.

If the user's intermittent snoring is detected during a certain period of time during an apnea event, but no snoring is detected in another period of time, then it can be considered that the user has simultaneous obstructive sleep breathing during an apnea event. Pauses and central sleep apnea are at greater risk. For such situations, the risk of sleep apnea syndrome of the user can be predicted as the risk of mixed sleep apnea syndrome.

In the embodiment of the present application, the physiological data of the user during sleep is collected through the wearable device, and whether the user has an apnea event is detected based on a binary classifier. If an apnea event occurs, it can be based on the sound data in the physiological data , Predict the user's risk of various subtypes of sleep apnea syndrome. This embodiment realizes the risk prediction of apnea syndrome subtypes based on the physiological data collected by the wearable device. The operation is simple and convenient, which helps to detect potential sleep apnea patients in time and early, and realizes personalized sleep quality monitoring.

It should be understood that the size of the sequence number of each step in the foregoing embodiment does not mean the order of execution. The execution sequence of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

5, there is shown a structural block diagram of a sleep risk prediction device provided by an embodiment of the present application. The device can be used in wearable devices such as smart watches and smart bracelets, and can also be used to collect physiological data of users. In other terminal equipment with processing functions, the device may include a signal acquisition unit, a signal processing unit, and other modules.

The above-mentioned signal acquisition unit can integrate a heart rate detection module, a blood oxygen detection module, a snoring detection module, a sleep and signal detection module, and so on. Among them, the heart rate detection module can be used to detect heart rate, heart rate variability and other information, related sensors can include but not limited to PPG sensors, electrocardiogram (Electrocardiogram, ECG) sensors, etc.; the blood oxygen detection module can be used to obtain signals related to blood oxygen detection Related sensors can include but are not limited to infrared and red light PPG sensors; the snoring detection module can be used to detect sound signals and extract snoring information, and the related sensors can include but are not limited to microphones; sleep and signal quality detection modules can be used to detect the user’s Judging the time to fall asleep and the user's action range, assist in the correction of calculated blood oxygen data, and predict the accuracy of blood oxygen and sleep apnea prediction results. Related sensors may include but are not limited to acceleration (ACC) sensors, Gyroscope, etc.

The above-mentioned central processing unit may be a control unit for receiving and processing data and instructions from other modules, calculating heart rate, heart rate variability, blood oxygen, and snoring information related characteristics, and presenting various subtypes through preset models The risk of sleep apnea is predicted.

The above-mentioned other modules may include a display module, a notification module, a communication module, and so on.

Corresponding to the sleep risk prediction method described in the above embodiment, FIG. 6 shows a structural block diagram of a sleep risk prediction device provided by another embodiment of the present application. For ease of description, only the information related to the embodiment of the present application is shown part.

Referring to Figure 6, the device can be applied to terminal devices such as wearable devices, and specifically can include the following modules:

The collection module 601 is used to collect various physiological data of the user to be detected during sleep;

The extraction module 602 is configured to extract feature information of each type of physiological data in the multiple types of physiological data;

The prediction module 603 is configured to input the extracted feature information into a preset classifier to obtain sleep risk prediction information output by the classifier; wherein, the classifier is obtained by training on the sample physiological data of a plurality of sample users .

In the embodiment of the present application, the multiple types of physiological data include at least pulse wave data, blood oxygen data, and/or sound data.

In the embodiment of the present application, the extraction module 602 may specifically include the following sub-modules:

In the embodiment of the present application, the classifier may include a two-classifier; the prediction module 603 may specifically include the following sub-modules:

In the embodiment of the present application, the first prediction submodule may specifically include the following units:

In the embodiment of the present application, the determining unit may specifically include the following subunits:

In the embodiment of the present application, the classifier may also include four classifiers; the prediction module 603 may also include the following sub-modules:

As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for related parts, please refer to the description of the method embodiment part.

Referring to FIG. 7, a schematic diagram of a terminal device according to an embodiment of the present application is shown. As shown in FIG. 7, the terminal device 700 of this embodiment includes a processor 710, a memory 720, and a computer program 721 that is stored in the memory 720 and can run on the processor 710. When the processor 710 executes the computer program 721, the steps in each embodiment of the sleep risk prediction method described above are implemented, for example, steps S301 to S303 shown in FIG. 3. Alternatively, when the processor 710 executes the computer program 721, the functions of the modules/units in the foregoing device embodiments, for example, the functions of the modules 601 to 603 shown in FIG. 6 are realized.

Exemplarily, the computer program 721 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 720 and executed by the processor 710 to complete This application. The one or more modules/units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments may be used to describe the execution process of the computer program 721 in the terminal device 700. For example, the computer program 721 can be divided into an acquisition module, an extraction module, and a prediction module, and the specific functions of each module are as follows:

The terminal device 700 may be a computing device such as a smart watch or a smart bracelet. The terminal device 700 may include, but is not limited to, a processor 710 and a memory 720. Those skilled in the art can understand that FIG. 7 is only an example of the terminal device 700, and does not constitute a limitation on the terminal device 700. It may include more or less components than those shown in the figure, or combine certain components, or different components. For example, the terminal device 700 may also include input and output devices, network access devices, buses, and so on.

The processor 710 may be a central processing unit (Central Processing Unit, CPU), other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (ASIC), Ready-made programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.

The memory 720 may be an internal storage unit of the terminal device 700, such as a hard disk or a memory of the terminal device 700. The memory 720 may also be an external storage device of the terminal device 700, such as a plug-in hard disk equipped on the terminal device 700, a smart memory card (Smart Media Card, SMC), and a Secure Digital (SD) Card, Flash Card, etc. Further, the memory 720 may also include both an internal storage unit of the terminal device 700 and an external storage device. The memory 720 is used to store the computer program 721 and other programs and data required by the terminal device 700. The memory 720 can also be used to temporarily store data that has been output or will be output.

The embodiment of the present application also discloses a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the aforementioned sleep risk prediction method can be implemented.

In the above-mentioned embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail or recorded in an embodiment, reference may be made to related descriptions of other embodiments.

A person of ordinary skill in the art may realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software depends on the specific application and design constraint conditions of the technical solution. Professionals and technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered beyond the scope of this application.

In the embodiments provided in this application, it should be understood that the disclosed sleep risk prediction method, device, and terminal device can be implemented in other ways. For example, the division of the modules or units is only a logical function division, and there may be other divisions in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored. Or not. In addition, the displayed or discussed mutual coupling or direct coupling or communication connection may be indirect coupling or communication connection through some interfaces, devices or units, and may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solutions of the embodiments.

In addition, the functional units in the various embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or software functional unit.

If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the implementation of all or part of the processes in the above-mentioned embodiment methods in the present application can be accomplished by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When executed by the processor, the steps of the foregoing method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file, or some intermediate forms. The computer-readable medium may at least include: any entity or device capable of carrying the computer program code to the sleep risk prediction device or terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random memory Take memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. For example, U disk, mobile hard disk, floppy disk or CD-ROM, etc. In some jurisdictions, according to legislation and patent practices, computer-readable media cannot be electrical carrier signals and telecommunication signals.

Finally, it should be noted that the above are only specific implementations of this application, but the scope of protection of this application is not limited to this. Any changes or substitutions within the technical scope disclosed in this application shall be covered by this application. Within the scope of protection applied for. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

A method for predicting sleep risk, which is characterized in that it includes:

Collect a variety of physiological data of the user to be tested during sleep;

Extracting feature information of each type of physiological data in the multiple types of physiological data respectively;

The extracted feature information is input into a preset classifier to obtain sleep risk prediction information output by the classifier; wherein the classifier is obtained by training on sample physiological data of a plurality of sample users.
The method according to claim 1, wherein the multiple types of physiological data include at least pulse wave data, blood oxygen data, and/or sound data.
The method according to claim 2, wherein said extracting characteristic information of each type of physiological data in said multiple types of physiological data respectively comprises:

Extract photoplethysmography feature information of the user to be detected from the pulse wave data, where the photoplethysmography feature information includes at least one of heart rate variability feature information, respiratory wave feature information, and photoplethysmography waveform feature information ;and / or,

Extracting blood oxygen characteristic information of the user to be detected from the blood oxygen data, where the blood oxygen characteristic information includes at least one of an oxygen reduction index and a low blood oxygen accumulation time; and/or,

The voice feature information of the user to be detected is extracted from the voice data, where the voice feature information includes at least one of Mel frequency cepstrum coefficient and Fourier spectrum feature information.
The method according to any one of claims 1-3, wherein the classifier comprises a two-classifier; the extracted feature information is input into a preset classifier to obtain the output of the classifier Sleep risk prediction information, including:

Input the extracted feature information into the second classifier;

Receiving a recognition result for the feature information output by the second classifier, where the recognition result of the second classifier includes occurrence of a sleep apnea event or no occurrence of a sleep apnea event;

If the recognition result is that a sleep apnea event occurs, the risk of various subtypes of sleep apnea syndrome of the user to be detected is predicted based on the voice data of the user to be detected.
The method according to claim 4, wherein the predicting the risk of various subtypes of sleep apnea syndrome in the user to be detected based on the voice data of the user to be detected comprises:

Determine the duration of the sleep apnea event to be recognized;

If within the duration, the voice data of the user to be detected includes intermittent snoring, predict that the sleep risk of the user to be detected is the risk of obstructive sleep apnea syndrome;

If the voice data of the user to be detected does not include snoring within the duration, predict that the sleep risk of the user to be detected is the risk of central sleep apnea syndrome;

If, within the duration, the sleep risk is predicted to include both the obstructive sleep apnea syndrome risk and the central sleep apnea syndrome risk based on the voice data of the user to be detected, then predict the The sleep risk of the user to be tested is the risk of mixed sleep apnea syndrome.
The method according to claim 5, wherein the determining the duration of the sleep apnea event to be recognized comprises:

It is determined that a continuous accumulation time of hypoxia occurs in the blood oxygen data, and the continuous accumulation time of hypoxia is used as the duration of the sleep apnea event to be recognized.
The method according to any one of claims 1 to 3, wherein the classifier includes four classifiers; the extracted feature information is input into a preset classifier to obtain the output of the classifier Sleep risk prediction information, including:

Input the extracted feature information into the four classifier;

Receive the recognition result of the feature information output by the four-classifier, the recognition result of the four-classifier includes the occurrence of no sleep apnea event, occurrence of obstructive sleep apnea event, occurrence of central type sleep apnea event, or occurrence Mixed sleep apnea events;

According to the recognition result, the risk of various subtypes of sleep apnea syndrome of the user to be detected is predicted.
A terminal device includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the following steps when the processor executes the computer program:

Collect a variety of physiological data of the user to be tested during sleep;

Extracting feature information of each type of physiological data in the multiple types of physiological data respectively;

The extracted feature information is input into a preset classifier to obtain sleep risk prediction information output by the classifier; wherein the classifier is obtained by training on sample physiological data of a plurality of sample users.
The terminal device according to claim 8, wherein the multiple types of physiological data include at least pulse wave data, blood oxygen data, and/or sound data, and the processor further implements the following when executing the computer program step:

Extract photoplethysmography feature information of the user to be detected from the pulse wave data, where the photoplethysmography feature information includes at least one of heart rate variability feature information, respiratory wave feature information, and photoplethysmography waveform feature information ;and / or,

Extracting blood oxygen characteristic information of the user to be detected from the blood oxygen data, where the blood oxygen characteristic information includes at least one of an oxygen reduction index and a low blood oxygen accumulation time; and/or,

The voice feature information of the user to be detected is extracted from the voice data, where the voice feature information includes at least one of Mel frequency cepstrum coefficient and Fourier spectrum feature information.
The terminal device according to claim 8 or 9, wherein the classifier comprises a two-classifier; the processor further implements the following steps when executing the computer program:

Input the extracted feature information into the second classifier;

Receiving a recognition result for the feature information output by the second classifier, where the recognition result of the second classifier includes occurrence of a sleep apnea event or no occurrence of a sleep apnea event;

If the recognition result is that a sleep apnea event occurs, the risk of various subtypes of sleep apnea syndrome of the user to be detected is predicted based on the voice data of the user to be detected.
The terminal device according to claim 10, wherein the predicting the risk of various subtypes of sleep apnea syndrome in the user to be detected based on the voice data of the user to be detected comprises:

Determine the duration of the sleep apnea event to be recognized;

If within the duration, the voice data of the user to be detected includes intermittent snoring, predict that the sleep risk of the user to be detected is the risk of obstructive sleep apnea syndrome;

If the voice data of the user to be detected does not include snoring within the duration, predict that the sleep risk of the user to be detected is the risk of central sleep apnea syndrome;

If, within the duration, the sleep risk is predicted to include both the obstructive sleep apnea syndrome risk and the central sleep apnea syndrome risk based on the voice data of the user to be detected, then predict the The sleep risk of the user to be tested is the risk of mixed sleep apnea syndrome.
The terminal device according to claim 8 or 9, wherein the classifier includes a four classifier; the processor further implements the following steps when executing the computer program:

Input the extracted feature information into the four classifier;

Receive the recognition result of the feature information output by the four-classifier, the recognition result of the four-classifier includes the occurrence of no sleep apnea event, occurrence of obstructive sleep apnea event, occurrence of central type sleep apnea event, or occurrence Mixed sleep apnea events;

According to the recognition result, the risk of various subtypes of sleep apnea syndrome of the user to be detected is predicted.
A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program implements the sleep risk prediction method according to any one of claims 1 to 7 when the computer program is executed by a processor .