A method, device, equipment and medium for detecting snoring behavior

Abstract

The application discloses a snoring behavior detection method, device, equipment and medium. An initial vibration signal of a user during sleep is collected through a bone conduction sensor, and an initial audio signal of the user during sleep is collected through a sound collecting device. A target time period in which the initial vibration signal and the initial audio signal overlap in a time domain is detected. An initial vibration signal of the target time period is cropped to obtain a target vibration signal, and an initial audio signal of the target time period is cropped to obtain a target audio signal. The target vibration signal and the target audio signal are input into a trained snoring detection model, and whether the user has a snoring behavior is detected through the snoring detection model to obtain a detection result output by the snoring detection model. The method can improve the accuracy of snoring behavior detection of the user, and is thus favorable for helping the user improve the sleep effect. The application can be widely applied in the field of artificial intelligence technology.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and in particular to a snoring behavior detection method, device, equipment and medium. BACKGROUND

[0002] In recent years, as people's living standards improve, more and more people pay more and more attention to health. People can recover their physical condition, relax their body cells, and enhance their body's resistance during sleep, thereby promoting physical health. It can be seen that sleep quality is closely related to the health status of the human body. Therefore, improving people's sleep is an important way to improve health. Snoring is a common behavior during sleep, which specifically refers to the vibration of the uvula (palatal) due to the narrowing of the upper respiratory tract during sleep, resulting in snoring. Snoring can cause daytime sleepiness, fatigue, lack of concentration, headache, decreased work capacity, and even severe obstructive sleep apnea, affecting physical health.

[0003] In related applications, there are applications that collect audio signals during user sleep and identify whether the user has a snoring behavior. However, because the audio signals in the surrounding environment during sleep are very complex, and in many cases there are multiple people snoring at the same time, the interference from the same person can cause the snoring recognition algorithm to misjudge, thereby reducing the accuracy of detecting the user's snoring behavior and failing to truly and effectively help the user improve sleep quality.

[0004] In summary, the technical problems existing in the related art need to be solved. SUMMARY

[0005] The present application aims to at least partially solve one of the technical problems existing in the related art.

[0006] To this end, an object of an embodiment of the present application is to provide a snoring behavior detection method.

[0007] Another object of an embodiment of the present application is to provide a snoring behavior detection device.

[0008] In order to achieve the above technical purpose, the technical solution adopted by the embodiments of the present application comprises:

[0009] In a first aspect, the embodiments of the present application provide a snoring behavior detection method, comprising the following steps:

[0010] Collecting an initial vibration signal of a user during sleep through a bone conduction sensor and collecting an initial audio signal of the user during sleep through a radio device;

[0011] Detecting a target time period in which the initial vibration signal and the initial audio signal overlap in the time domain;

[0012] clip the initial vibration signal in the target time period to obtain a target vibration signal, and clip the initial audio signal in the target time period to obtain a target audio signal;

[0013] input the target vibration signal and the target audio signal into the trained snoring detection model, detect whether the user has a snoring behavior through the snoring detection model, and obtain a detection result output by the snoring detection model.

[0014] In addition, the snoring behavior detection method according to the above-mentioned embodiments of the present application can further have the following additional technical features:

[0015] Further, in an embodiment of the present application, the target time period in which the initial vibration signal and the initial audio signal overlap in the time domain comprises:

[0016] detecting a first time region in which the signal intensity of the initial vibration signal is greater than a first preset threshold value;

[0017] detecting a second time region in which the signal intensity of the initial audio signal is greater than a second preset threshold value;

[0018] obtaining the target time period by finding the intersection of the first time region and the second time region.

[0019] Further, in an embodiment of the present application, the method further comprises:

[0020] performing low-pass filtering processing on the initial vibration signal or the initial audio signal.

[0021] Further, in an embodiment of the present application, the snoring detection model comprises a first encoder, a second encoder, a feature fusioner and a decoder; and the inputting the target vibration signal and the target audio signal into the trained snoring detection model, detecting whether the user has a snoring behavior through the snoring detection model, and obtaining the detection result output by the snoring detection model comprises:

[0022] inputting the target vibration signal into the snoring detection model, extracting features of the target vibration signal through the first encoder to obtain first feature data, and extracting features of the target audio signal through the second encoder to obtain second feature data;

[0023] performing fusion processing on the first feature data and the second feature data through the feature fusioner to obtain fusion feature data;

[0024] decoding the fusion feature data through the decoder to obtain the detection result.

[0025] Further, in an embodiment of the present application, the feature extraction of the target vibration signal by the first encoder includes:

[0026] determining spectral feature data of the target vibration signal; the spectral feature data includes at least one of amplitude or power ratio;

[0027] inputting the spectral feature data into the first encoder for encoding to obtain the first feature data.

[0028] Further, in an embodiment of the present application, the fusion processing of the first feature data and the second feature data by the feature fusion device includes:

[0029] performing weighted processing or splicing processing on the first feature data and the second feature data by the feature fusion device to obtain the fusion feature data.

[0030] Further, in an embodiment of the present application, the determination of the spectral feature data of the target vibration signal includes:

[0031] performing fast Fourier transform on the target vibration signal to obtain spectral data of the target vibration signal;

[0032] determining the spectral feature data of the target vibration signal according to the spectral data.

[0033] In a second aspect, an embodiment of the present application provides a snoring behavior detection device, including:

[0034] a collection unit configured to collect an initial vibration signal of a user during sleep through a bone conduction sensor and collect an initial audio signal of the user during sleep through a sound collecting device;

[0035] a detection unit configured to detect a target time period in which the initial vibration signal and the initial audio signal overlap in a time domain;

[0036] a clipping unit configured to clip the initial vibration signal of the target time period to obtain a target vibration signal and clip the initial audio signal of the target time period to obtain a target audio signal;

[0037] a processing unit configured to input the target vibration signal and the target audio signal into a trained snoring detection model, detect whether the user has a snoring behavior through the snoring detection model, and obtain a detection result output by the snoring detection model.

[0038] In a third aspect, an embodiment of the present application provides a computer device, comprising:

[0039] at least one processor;

[0040] at least one memory for storing at least one program;

[0041] When the at least one program is executed by the at least one processor, the at least one processor implements the snoring behavior detection method in the first aspect.

[0042] In a fourth aspect, an embodiment of the present application further provides a computer readable storage medium, wherein a processor executable program is stored, and the processor executable program is used for implementing the snoring behavior detection method in the first aspect when executed by a processor.

[0043] The advantages and beneficial effects of the present application will be partially given in the following description, partially will become obvious from the following description, or will be learned by the practice of the present application:

[0044] In an embodiment of the present application, a snoring behavior detection method is provided. The method collects initial vibration signals of a user during sleep through a bone conduction sensor, collects initial audio signals of the user during sleep through a sound collecting device, detects a target time period in which the initial vibration signals and the initial audio signals overlap in time domain, trims the initial vibration signals in the target time period to obtain target vibration signals, and trims the initial audio signals in the target time period to obtain target audio signals, inputs the target vibration signals and the target audio signals into a trained snoring detection model, detects whether the user has a snoring behavior through the snoring detection model, and obtains a detection result output by the snoring detection model. The method detects whether the user has a snoring behavior based on bone conduction vibration signals and audio signals generated during sleep of the user, and through a snoring detection model trained by a machine learning algorithm, can reduce interference caused by environmental noise and snoring behaviors of other persons on detection, and improve accuracy of the obtained detection result. Moreover, in the method of the present application, the target vibration signals and the target audio signals are trimmed based on the target time period in which the vibration signals and the audio signals overlap, which can remove invalid signals, on the one hand, can reduce data processing pressure of the model, and on the other hand, can improve prediction accuracy of the model. Therefore, the method can improve accuracy of detection of the snoring behavior of the user, and is thus beneficial to helping the user improve sleep effect. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following introduces the drawings of the related technical solutions in the embodiments of the present application or the prior art. It should be understood that the drawings in the following introduction are only for facilitating the clear description of part of the embodiments of the technical solutions in the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0046] Figure 1 An implementation environment schematic diagram of a snoring behavior detection method provided in an embodiment of the present application;

[0047] Figure 2 A flowchart schematic diagram of a snoring behavior detection method provided in an embodiment of the present application;

[0048] Figure 3 A structure schematic diagram of a snoring behavior detection device provided in an embodiment of the present application;

[0049] Figure 4 A structure schematic diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the drawings, wherein the same or similar notations represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are only used to explain the present application, and cannot be understood as a limitation of the present application. For the step numbers in the following embodiments, they are only set for facilitating the description and explanation, and the order between the steps is not limited in any way, and the execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0051] Before the embodiments of the present application are described in detail, first, some nouns and terms involved in the embodiments of the present application are explained, and the nouns and terms involved in the embodiments of the present application are applicable to the following explanations.

[0052] Artificial Intelligence (AI) is the use of digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology of computer science, which attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines, so that machines have the functions of perception, reasoning and decision-making. Artificial intelligence technology is a comprehensive discipline, involving a wide range of fields, both hardware and software technologies. Artificial intelligence basic technologies generally include sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction system, mechatronics and other technologies. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology and machine learning / deep learning and other major directions.

[0053] Machine Learning (ML) is a multi-disciplinary subject involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, etc. It is dedicated to studying how computers simulate or implement human learning behavior to acquire new knowledge or skills, and reorganize existing knowledge structure to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental approach to making computers intelligent. Its applications are widespread in various fields of artificial intelligence. Machine learning (deep learning) usually includes artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and rule-based learning.

[0054] In recent years, as people's living standards improve, more and more people pay more and more attention to health. During sleep, people can restore their physical condition, relax their cells, and enhance their body's resistance, thereby promoting physical health. It can be seen that sleep quality is closely related to the health status of the human body. Therefore, improving people's sleep is an important way to improve health. Snoring is a common behavior during sleep, which is specifically the vibration of the uvula (palatal) caused by the narrowing of the upper respiratory tract during sleep, resulting in snoring. Snoring can cause daytime sleepiness, fatigue, lack of concentration, headache, decreased work capacity, and even cause obstructive sleep apnea, affecting physical health.

[0055] In related applications, there are applications for collecting audio signals during the user's sleep process to identify whether the user has snoring behavior. However, because the audio signals in the surrounding environment during sleep are very complex, and in many cases there are multiple people snoring at the same time, interference from the companion can cause the snoring sound recognition algorithm to misjudge, thereby causing the detection accuracy of the user's snoring behavior to be low, and unable to truly and effectively help the user improve the sleep effect.

[0056] Therefore, in the embodiments of the present application, a snoring behavior detection method is provided, which is based on the bone conduction vibration signals and audio signals generated during the user's sleep process, and a snoring detection model trained by a machine learning algorithm is used to detect whether the user has snoring behavior, which can reduce the interference caused by environmental noise and other people's snoring behavior on detection, and improve the accuracy of the obtained detection result. Moreover, in the embodiments of the present application, the vibration signals and audio signals are cropped based on the overlapping target time period, which can remove invalid signals, on the one hand, the data processing pressure of the model can be reduced, and on the other hand, the prediction accuracy of the model can be improved. Therefore, the method can improve the accuracy of the detection of the user's snoring behavior, and thus help the user improve the sleep effect.

[0057] First, please refer to Figure 1 , Figure 1 is an implementation environment schematic diagram of a snoring behavior detection method provided by the embodiments of the present application. Referring to Figure 1 , the main body of the implementation environment mainly includes a sleep mattress 101, a server 102 and a terminal device 103, the sleep mattress 101 is in communication connection with the server 102, and the terminal device 103 is also in communication connection with the server 102. The snoring behavior detection method can be executed through the interaction among the sleep mattress 101, the server 102 and the terminal device 103, and specific selection can be made according to actual application conditions, which is not specifically limited in the embodiments.

[0058] In some embodiments, the sleep mattress 101 side in the present application can be provided with a bone conduction sensor and a sound collecting device. When the user rests on the sleep mattress 101, the bone conduction sensor can collect vibration signals during the user's sleep process, and the sound collecting device can collect initial audio signals during the user's sleep process. After the bone conduction sensor and the sound collecting device collect the relevant data, the data can be sent to the server 102 or the terminal device 103. The server 102 or the terminal device 103 can obtain the final detection result by analyzing and processing the data, and then feed back the detection result to the user. For example, the detection result obtained on the server 102 side can be sent to the terminal device 103 for display; or the terminal device 103 directly displays the detection result obtained by local analysis.

[0059] In some embodiments, the terminal device 103 can include, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted terminal, and the like. The server 102 can be a standalone physical server, a server cluster composed of multiple physical servers, or a distributed system, and can also be a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and basic cloud computing services such as big data and artificial intelligence platforms. The sleep mattress 101 and the server 102, and the terminal device 103 and the server 102 can establish a communication connection through a wireless network or a wired network using standard communication technologies and / or protocols. The network can be set as the Internet, or any other network, for example, including but not limited to a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a private network or any combination of a virtual private network.

[0060] Next, in conjunction with the implementation environment shown in Figure 1 The sleep snoring behavior detection method provided in the embodiments of the present application is introduced and described.

[0061] Please refer to Figure 2 , Figure 2 is a schematic diagram of a sleep snoring behavior detection method provided by the embodiments of the present application. The sleep snoring behavior detection method includes but is not limited to:

[0062] Step 110, collecting initial vibration signals in the user's sleep process through a bone conduction sensor, and collecting initial audio signals in the user's sleep process through a sound collecting device;

[0063] In this step, when detecting the snoring behavior of the user, the vibration signals in the user's sleep process can be collected through a bone conduction sensor, denoted as initial vibration signals, and the audio signals in the user's sleep process can be collected through a sound collecting device, denoted as initial audio signals. Specifically, here, the bone conduction sensor is a component that collects vibration signals through bone conduction, which can be in contact with the user's ear area, thereby facilitating the collection of vibration signals generated by the user due to snoring. The sound collecting device can use existing components such as microphones, and the present application does not limit this.

[0064] It should be noted that due to the possibility of sudden failure or unstable operation of a single component, the initial vibration signal and the initial audio signal collected may have low reliability. Therefore, in some embodiments, the bone conduction sensor and the sound collecting device of the present application can be provided in multiple, for example, 6 bone conduction sensors + 1 microphone can be provided on the user side. Of course, the specific setting should also be considered comprehensively in combination with the component cost, and the present application does not make any limitation.

[0065] It should be noted that in the embodiments of the present application, for the same type of signals collected by multiple components, for example, the initial vibration signals collected by 6 bone conduction sensors, they can be processed as 6 independent signals respectively, or they can be weighted and summed to integrate into one signal for processing, and the present application does not make any limitation.

[0066] Step 120, detecting a target time period in which the initial vibration signal and the initial audio signal overlap in time domain;

[0067] In this step, after collecting the initial vibration signal and the initial audio signal, the target time period in which the two types of signals overlap in time domain can be detected. It can be understood that during the user's sleep, there may be environmental noise or vibration caused by the user's turning over, and these vibration signals and audio signals are not caused by the user's snoring behavior. If all the initial vibration signals and the initial audio signals are input into the prediction model for detection, it will obviously greatly increase the data processing pressure of the model, and may interfere with the prediction result. Therefore, in the embodiments of the present application, the initial vibration signal and the initial audio signal can be optimized. Specifically, it is easy to understand that in some cases, for example, when environmental noise is generated, the user may not be in a snoring state, and there is no corresponding action, at this time only the initial audio signal collected by the sound collecting device. Similarly, when the user adjusts the position in the sleep state and generates a vibration signal, the sound triggered is small, at this time the sound collecting device is likely to collect no audio signal and only the initial vibration signal collected by the bone conduction sensor. When the user is in a snoring state, obviously the bone conduction sensor can collect the initial vibration signal, and the sound collecting device can also collect the initial audio signal, that is, when both types of signals exist at the same time, the user may be in a snoring state; on the contrary, when only one type of signal exists, the user is most likely not in a snoring state. Therefore, in the embodiments of the present application, the time region in which the initial vibration signal and the initial audio signal overlap in time domain can be detected, which is recorded as the target time period.

[0068] It should be noted that in the embodiments of the present application, the number of target time periods can be one or multiple, and the present application does not make any specific limitation.

[0069] Step 130, clipping the initial vibration signal of the target time period to obtain a target vibration signal, and clipping the initial audio signal of the target time period to obtain a target audio signal;

[0070] In this step, as described above, when only one type of signal exists in the initial vibration signal and the initial audio signal, the user is most likely not in a snoring state. Therefore, in order to reduce the model data processing pressure caused by noise data, and as much as possible to exclude the adverse interference of noise data on the prediction result, in the embodiments of the present application, the initial vibration signal and the initial audio signal can be clipped according to the target time period. Specifically, the initial vibration signal of the target time period can be clipped to obtain a target vibration signal, and the initial audio signal of the target time period can be clipped to obtain a target audio signal.

[0071] Step 140, inputting the target vibration signal and the target audio signal into the trained snoring detection model, detecting whether the user has a snoring behavior through the snoring detection model, and obtaining a detection result output by the snoring detection model.

[0072] In this step, after clipping the target vibration signal and the target audio signal, they can be input into the trained snoring detection model to detect whether the user has a snoring behavior. Here, the snoring detection model can be built by using any machine learning algorithm, and the present application does not limit this.

[0073] Specifically, when detecting whether the user has a snoring behavior based on the target vibration signal and the target audio signal, feature data of the signals can be extracted, and then classification prediction is performed through the feature data to determine whether the user has a snoring behavior or does not have a snoring behavior. For example, in some cases, taking the target audio signal as an example, some time-frequency points can be selected from the spectral data of the target audio signal according to a predetermined rule, and encoded into a digital sequence, which can be used as feature data of the target audio data, denoted as second feature data in the embodiments of the present application. Of course, the feature extraction algorithm used in the embodiments of the present application can be optional as needed. For example, in another embodiment, the target audio data can be divided into frames and windowed according to a certain frame stacking to obtain a plurality of audio frames. Specifically, frame division here refers to cutting the entire audio data into multiple segments according to a predetermined rule, and each segment is an audio frame. Windowing refers to using a predetermined window function to intercept each audio frame, so that the audio data after frame division is more coherent and exhibits better periodic function characteristics. The window function can be any one of a rectangular window, a Hamming window, or a Gaussian window. Then, the obtained audio frames are subjected to fast Fourier transform (FFT, Fast Fourier Transformation) to obtain the spectrum of each audio frame, and the peak signal in each frame spectrum is selected as the digital feature of the spectrum to construct the second feature data of the target audio data. In some embodiments, after obtaining the spectrum of the audio frame, the audio frame can also be divided into multiple subbands in the frequency domain, and then the difference value of the energy difference between any two adjacent subbands is calculated, and the second feature data of the target audio data is constructed based on the obtained difference value.

[0074] In some embodiments, for example, for the target vibration signal, its spectral feature data can be determined. Specifically, the target vibration signal can be subjected to fast Fourier transform first to obtain its corresponding spectral data, and then the amplitude or power ratio data of the target vibration signal is determined based on the spectral data, and at least one of the amplitude or power ratio data is used as the spectral feature data thereof. Then, the spectral feature data can be encoded to obtain the feature data corresponding to the target vibration signal, denoted as first feature data in the embodiments of the present application. When encoding specifically, the digital feature can be converted into a hash sequence, that is, the hash sequence of the spectral feature data is used as the first feature data of the target vibration signal.

[0075] Specifically, for example, in some embodiments, an optional structure of the snoring detection model can include a first encoder, a second encoder, a feature fusioner and a decoder. In use of the model, the target vibration signal can be input to the first encoder, so as to encode the target vibration signal to obtain first feature data. Similarly, the target audio signal can be encoded by the second encoder to obtain second feature data, and then the first feature data and the second feature data can be fused by the feature fusioner to obtain fused feature data. Here, the fusion manner can include any one of weighting or splicing, which is not limited in the present application. Then, the obtained fused feature data can be input to the decoder for decoding to obtain the output detection result. Specifically, the first encoder and the second encoder can encode the target vibration signal and the target audio signal by using the foregoing implementation manners, which will not be described herein again.

[0076] In some embodiments, the target time period in which the initial vibration signal and the initial audio signal overlap in time domain is detected, including:

[0077] detecting a first time region in which the signal intensity of the initial vibration signal is greater than a first preset threshold value;

[0078] detecting a second time region in which the signal intensity of the initial audio signal is greater than a second preset threshold value;

[0079] obtaining the target time period by taking the intersection of the first time region and the second time region.

[0080] In the embodiments of the present application, when determining the target time period, the initial vibration signal can be detected first. It can be understood that the bone conduction sensor can detect various vibration signals triggered by the user, and some noise may exist in these vibration signals. Therefore, in the embodiments of the present application, the initial vibration signal can be subjected to low-pass filtering to filter out part of the noise. And after filtering, part of the signal may still be in a slight vibration state, which is not caused by the user's snoring behavior. Therefore, in the embodiments of the present application, a signal intensity threshold value, denoted as a first preset threshold value, can be set. The threshold value can be a signal amplitude value, etc., which is not limited in the present application. Then, the initial vibration signal is compared with the first preset threshold value, and a time region in which the initial vibration signal is greater than the first preset threshold value is determined as a first time region. The first time region can be considered as a time region that is more likely to be caused by the user's snoring behavior.

[0081] Similarly, in the embodiments of the present application, the initial audio signal can also be subjected to low-pass filtering processing, and a second preset threshold can be set for the initial audio signal. Then, the initial vibration signal and the first preset threshold can be compared, and a time region in which the initial vibration signal is greater than the first preset threshold is determined as the first time region.

[0082] It should be noted that in the embodiments of the present application, the numerical values of the first preset threshold and the second preset threshold can be flexibly set as needed, and the present application does not make specific limitations thereto.

[0083] In the embodiments of the present application, after obtaining the first time region and the second time region, the intersection of the two time regions can be calculated to obtain the target time period. It should be noted that in the embodiments of the present application, the first time region or the second time region can be discrete, and therefore the target time period obtained can also be discrete, and the specific number is not limited by the present application.

[0084] In the embodiments of the present application, before the snoring detection model is put into use, it needs to be trained to adjust its internal parameters, so as to achieve better prediction effect. Specifically, when training the model, a batch of sample data can be obtained, wherein each sample data includes a sample vibration signal and a sample audio signal corresponding to a sample user, and the processing processes of the two signals are the same as those of the target vibration signal and the target audio signal described above, which will not be repeated here. The sample data also includes a corresponding sample label, which is used to record whether the sample user has snoring behavior in the target time period. Each sample data and its corresponding sample label can be used as a set of training data, the input data of the model is the sample vibration signal and the sample audio signal, and the output data of the model is the prediction result of whether the sample user has snoring behavior in the target time period. After obtaining the prediction result output by the model, the accuracy of the model prediction can be evaluated according to the prediction result and the sample label, so as to update the parameters of the model.

[0085] For a machine learning model, the accuracy of the model prediction result can be measured by a loss function, which is defined on a single training data and is used to measure the prediction error of a training data, specifically by determining the loss value of the training data through the label of the single training data and the prediction result of the model on the training data. In actual training, a training data set has many training data, so a cost function is generally used to measure the overall error of the training data set, which is defined on the entire training data set and is used to calculate the average of the prediction errors of all training data, which can better measure the prediction effect of the model. For a general machine learning model, based on the aforementioned cost function, plus a regular term that measures the complexity of the model, it can be used as the target function of training, and based on the target function, the loss value of the entire training data set can be obtained. There are many kinds of commonly used loss functions, such as 0-1 loss function, square loss function, absolute loss function, logarithmic loss function, cross-entropy loss function, etc. can be used as the loss function of the machine learning model, which will not be described one by one. In the embodiments of the present application, any one of the loss functions can be selected to determine the loss value of the training, for example, the cross-entropy loss function. Based on the loss value of the training, the parameters of the model are updated using the back propagation algorithm, and after several iterations, a trained snoring detection model can be obtained. The specific number of iterations can be pre-set, or the training can be considered complete when the test set reaches the accuracy requirement.

[0086] A snoring behavior detection device is described below according to an embodiment of the present application.

[0087] Referring to Figure 3 A snoring behavior detection device is described below according to an embodiment of the present application.

[0088] The acquisition unit 201 is configured to acquire an initial vibration signal of a user during sleep through a bone conduction sensor and acquire an initial audio signal of the user during sleep through a sound collecting device.

[0089] The detection unit 202 is configured to detect a target time period in which the initial vibration signal and the initial audio signal overlap in the time domain.

[0090] The cropping unit 203 is configured to crop the initial vibration signal of the target time period to obtain a target vibration signal, and crop the initial audio signal of the target time period to obtain a target audio signal.

[0091] The processing unit 204 is configured to input the target vibration signal and the target audio signal into a trained snoring detection model, detect whether the user has a snoring behavior through the snoring detection model, and obtain a detection result output by the snoring detection model.

[0092] It can be understood that the contents in the above method embodiments are applicable to the present device embodiments, the present device embodiments specifically implement the functions same as those of the above method embodiments, and achieve the beneficial effects same as those of the above method embodiments.

[0093] With reference to Figure 4 The present application provides a computer device, comprising:

[0094] at least one processor 301;

[0095] at least one memory 302, configured to store at least one program;

[0096] When the at least one program is executed by the at least one processor 301, the at least one processor 301 implements a snoring behavior detection method.

[0097] Similarly, the contents in the above method embodiments are applicable to the present computer device embodiments, the present computer device embodiments specifically implement the functions same as those of the above method embodiments, and achieve the beneficial effects same as those of the above method embodiments.

[0098] The present application further provides a computer readable storage medium, wherein the computer readable storage medium stores a program executable by the processor 301, and the program executable by the processor 301 is used for executing the above-mentioned snoring behavior detection method when executed by the processor 301.

[0099] Similarly, the contents in the above method embodiments are applicable to the present computer readable storage medium embodiments, the present computer readable storage medium embodiments specifically implement the functions same as those of the above method embodiments, and achieve the beneficial effects same as those of the above method embodiments.

[0100] In some alternative embodiments, the functions / operations mentioned in the block diagram can not occur in the order mentioned in the operation diagram. For example, depending on the functions / operations involved, two blocks shown in succession can actually be executed substantially simultaneously or the blocks can sometimes be executed in reverse order. In addition, the embodiments presented and described in the flowcharts of the present application are provided by way of example, with the purpose of providing a more comprehensive understanding of the technology. The disclosed method is not limited to the operations and logical flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and in which sub-operations described as part of larger operations are independently executed.

[0101] Furthermore, although the present application is described in the context of functional modules, it is understood that one or more of the functions and / or features can be integrated in a single physical device and / or software module, or one or more functions and / or features can be implemented in separate physical devices or software modules. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary to an understanding of the present application. Rather, the actual implementation is within the routine skill of engineers familiar with the property, function and internal relationships of the various functional modules disclosed herein. Accordingly, the present application is not limited to the specific details of the functional modules described herein. Rather, it is understood that one of ordinary skill in the art is able to practice the application as claimed without undue experimentation having regard to the property, function and internal relationships of the various functional modules disclosed herein. It is also understood that the specific concepts disclosed are merely illustrative and that the scope of the present application is determined by the appended claims and their equivalents.

[0102] If the functions are implemented in software, the functions can be stored in or implemented as one or more computer program products. The computer program product can be stored in a computer readable medium, which can include, but is not limited to, RAM, ROM, electrically programmable ROM (EPROM or EEPROM), flash memory, or a magnetic or optical card, or any suitable device used for storing a computer program. Furthermore, the computer program product can be implemented as at least one program that runs on a computer, which can be a personal computer, a server, a network device, or any suitable device capable of executing a program.

[0103] The logic and / or steps represented in the flowcharts and / or otherwise described herein, for example, can be embodied in non-transitory computer-readable media, which can be executed by an instruction execution system, apparatus, or device such as a computer-based system, processor, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. In the context of this specification, a "computer-readable medium" can be any means that can contain, store, communicate, propagate or transport the program for use by or in connection with the instruction execution system, apparatus, or device.

[0104] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example, via optical scanning of the paper or other medium, then compiled, interpreted, or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.

[0105] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the above embodiments, various steps or methods can be implemented in software or firmware which are stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any of the following technologies, known in the art, or their combinations can be used: discrete logic circuitry having logic gates for implementing logic functions upon an application of data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and the like.

[0106] In the above description of the present specification, the description referring to the terms "one embodiment", "another embodiment", or "certain embodiments" or the like means that a specific feature, structure, material or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. The illustrative expressions of the above terms do not necessarily refer to the same embodiment or example in the present specification. Also, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in an appropriate manner.

[0107] Although the embodiments of the present application have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, alternatives and variations to these embodiments can be made without departing from the principles and spirit of the application, the scope of which is defined by the claims and their equivalents.

[0108] The above is a specific description of the preferred embodiments of the present application, but the present application is not limited to the embodiments, and those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present application, and these equivalent modifications or substitutions are included in the scope defined by the claims of the present application.

Claims

1. A method for detecting snoring behavior, characterized in that: include: collecting an initial vibration signal of the user during sleep through a bone conduction sensor, and collecting an initial audio signal of the user during sleep through a sound receiving device; detecting a target time period in which the initial vibration signal and the initial audio signal overlap in the time domain; cutting the initial vibration signal of the target time period to obtain a target vibration signal, and cutting the initial audio signal of the target time period to obtain a target audio signal; The target vibration signal and the target audio signal are input into a trained snoring detection model, and the snoring detection model is used to detect whether the user has snoring behavior, thereby obtaining a detection result output by the snoring detection model.

2. The method for detecting snoring behavior according to claim 1, characterized in that: The detecting a target time period in which the initial vibration signal and the initial audio signal overlap in the time domain includes: Detecting a first time period during which the signal strength of the initial vibration signal is greater than a first preset threshold; Detecting a second time period during which the signal strength of the initial audio signal is greater than a second preset threshold; An intersection of the first time zone and the second time zone is obtained to obtain the target time period.

3. The method for detecting snoring behavior according to claim 1 or 2, characterized in that: The method further comprises: Perform low-pass filtering on the initial vibration signal or the initial audio signal.

4. The method for detecting snoring behavior according to claim 1, characterized in that: The snoring detection model includes a first encoder, a second encoder, a feature fusion unit, and a decoder. The target vibration signal and the target audio signal are input into the trained snoring detection model, and the snoring detection model is used to detect whether the user is snoring. The detection result output by the snoring detection model includes: Inputting the target vibration signal into the snoring detection model, performing feature extraction on the target vibration signal using the first encoder to obtain first feature data, and performing feature extraction on the target audio signal using the second encoder to obtain second feature data; fusing the first feature data and the second feature data by the feature fuser to obtain fused feature data; The fused feature data is decoded by the decoder to obtain the detection result.

5. The method for detecting snoring behavior according to claim 4, characterized in that: The extracting features of the target vibration signal by the first encoder to obtain first feature data includes: Determining frequency spectrum characteristic data of the target vibration signal; the frequency spectrum characteristic data includes at least one of amplitude or power ratio; The spectrum feature data is input into the first encoder for encoding to obtain the first feature data.

6. The method for detecting snoring behavior according to claim 4, characterized in that: The fusing the first feature data and the second feature data by the feature fuser to obtain fused feature data includes: The feature fusion device performs weighted processing or splicing processing on the first feature data and the second feature data to obtain the fused feature data.

7. The method for detecting snoring behavior according to claim 5, characterized in that: The determining of the frequency spectrum characteristic data of the target vibration signal includes: Performing a fast Fourier transform on the target vibration signal to obtain frequency spectrum data of the target vibration signal; Spectrum characteristic data of the target vibration signal is determined according to the spectrum data.

8. A snoring behavior detection device, characterized in that: include: a collection unit, configured to collect an initial vibration signal of the user during sleep through a bone conduction sensor, and collect an initial audio signal of the user during sleep through a sound receiving device; a detection unit, configured to detect a target time period in which the initial vibration signal and the initial audio signal overlap in the time domain; a cutting unit, configured to cut the initial vibration signal of the target time period to obtain a target vibration signal, and to cut the initial audio signal of the target time period to obtain a target audio signal; The processing unit is configured to input the target vibration signal and the target audio signal into a trained snoring detection model, detect whether the user has snoring behavior through the snoring detection model, and obtain a detection result output by the snoring detection model.

9. A computer device, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method for detecting snoring behavior according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a program executable by a processor, characterized in that: The processor-executable program is used to implement a snoring behavior detection method according to any one of claims 1 to 7 when executed by the processor.

Images