Sleep Quality Assessment Method and System Based on Sleep Activity Recognition

Abstract

The present invention relates to the technical field of sleep data processing, and particularly relates to a sleep quality assessment method and system based on sleep activity recognition. For each sleep, the present invention obtains multiple curve segments of each temperature curve; according to the temperature data distribution at different times on each curve segment, the final deep sleep segment is screened out; according to the sleep scores of the user in historical sleep and the distribution of the final deep sleep segments, the high-quality deep sleep coefficient of the user is obtained; according to the difference between the proportion of the duration of all the final deep sleep segments in the user's real-time sleep and the high-quality deep sleep coefficient, and the number of the final deep sleep segments, the high-quality sleep possibility of the user in real-time sleep is obtained; and the sleep quality is evaluated. By analyzing in combination with historical sleep data, the present invention obtains the possibility that the user's real-time sleep data is high-quality sleep data, thereby improving the accuracy of sleep quality assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of sleep data processing, and particularly relates to a sleep quality assessment method and system based on sleep activity recognition. Background Art

[0002] With the development of technology, the accuracy of sleep activity recognition has been continuously improved, enabling individuals to easily monitor and improve their sleep quality in daily life and avoid the occurrence of chronic health problems; good sleep quality is the basis for physical recovery and mental health, and long-term sleep problems may lead to problems such as decreased immunity, unstable emotions, and low work efficiency. Therefore, it is necessary to evaluate sleep quality.

[0003] In the prior art, sleep quality is characterized by unified body temperature characteristics and measured based on general data models or standards; however, due to differences among individuals, the existing methods cannot accurately reflect the specific sleep conditions of different people, and the evaluation of sleep quality is poor. Summary of the Invention

[0004] In order to solve the technical problem that considering the differences among individuals, the existing methods cannot accurately reflect the specific sleep conditions of different people and the evaluation of sleep quality is poor, the purpose of the present invention is to provide a sleep quality assessment method and system based on sleep activity recognition, and the specific technical solutions adopted are as follows:

[0005] The present invention proposes a sleep quality assessment method based on sleep activity recognition, and the method includes:

[0006] Obtain the temperature curve and sleep score of the user during each sleep, where the temperature curve includes temperature data at different times;

[0007] For each sleep, obtain multiple curve segments of each temperature curve; according to the temperature data distribution at different times on each curve segment, obtain the likelihood of deep sleep for each curve segment, and screen out suspected deep sleep segments; according to the change trend of temperature data in the adjacent range before and after each suspected deep sleep segment and the average likelihood of deep sleep for all curve segments, screen out the final deep sleep segments;

[0008] Obtain the high-quality deep sleep coefficient of the user according to the sleep score and the distribution of the final deep sleep segments in the historical sleep of the user; obtain the likelihood of high-quality sleep of the user during the real-time sleep according to the difference between the proportion of the duration of all the final deep sleep segments in the real-time sleep of the user and the high-quality deep sleep coefficient, and the number of the final deep sleep segments;

[0009] Evaluate the sleep quality according to the likelihood of high-quality sleep of the user during the real-time sleep.

[0010] Furthermore, the method for obtaining the deep sleep possibility includes:

[0011] For each sleep, obtain the mean value of the temperature data at all moments on each curve segment as the average temperature level;

[0012] Obtain the cumulative sum of the differences between the temperature data at different moments and the average temperature level on each curve segment;

[0013] Obtain the product of the average temperature level and the cumulative sum of the differences, and perform a negative correlation mapping as the deep sleep possibility of each curve segment.

[0014] Furthermore, the method for obtaining the suspected deep sleep segment includes:

[0015] If the deep sleep possibility of a curve segment is greater than or equal to a preset possibility threshold, take the corresponding curve segment as the initial deep sleep segment;

[0016] Merge continuously adjacent initial deep sleep segments to form multiple suspected deep sleep segments.

[0017] Furthermore, the method for obtaining the final deep sleep segment includes:

[0018] According to the change trend of the temperature data in the adjacent ranges before and after each suspected deep sleep segment, and the average deep sleep possibility of all corresponding curve segments, obtain the deep sleep confidence of each suspected deep sleep segment;

[0019] If the deep sleep confidence of a suspected deep sleep segment is greater than or equal to a preset confidence threshold, take the corresponding suspected deep sleep segment as the final deep sleep segment.

[0020] Furthermore, the method for obtaining the deep sleep confidence includes:

[0021] Obtain the mean value of the slopes between all adjacent data points in each adjacent range corresponding to each suspected deep sleep segment as the temperature change degree;

[0022] Obtain the difference between the temperature change degrees in the adjacent range after and the adjacent range before each suspected deep sleep segment, calculate the product of the difference result and the average deep sleep possibility, and perform normalization as the deep sleep confidence of the corresponding suspected deep sleep segment.

[0023] Furthermore, the method for obtaining the high-quality deep sleep coefficient includes:

[0024] Obtain the ratio of the duration of all final deep sleep segments to the total sleep duration in each historical sleep as the local deep sleep coefficient;

[0025] If the sleep score of the user in the historical sleep is greater than the preset score threshold, the corresponding historical sleep is used as the reference sleep;

[0026] Obtain the mean value of the local deep sleep coefficients corresponding to all reference sleeps as the user's high-quality deep sleep coefficient.

[0027] Furthermore, the method for obtaining the high-quality sleep possibility includes:

[0028] Obtain the difference between the local deep sleep coefficient of the user in the real-time sleep and the high-quality deep sleep coefficient, calculate the product of the difference result and the number of all final segments of the deep sleep in the real-time sleep, and perform a negative correlation mapping as the high-quality sleep possibility of the user in the real-time sleep.

[0029] Furthermore, the evaluation of the sleep quality according to the high-quality sleep possibility of the user in the real-time sleep includes:

[0030] If the high-quality sleep possibility of the user in the real-time sleep is greater than or equal to the preset first threshold, the sleep quality is judged to be the first level;

[0031] If the high-quality sleep possibility of the user in the real-time sleep is less than the preset first threshold and greater than the preset second threshold, the sleep quality is judged to be the second level;

[0032] If the high-quality sleep possibility of the user in the real-time sleep is less than or equal to the preset second threshold, the sleep quality is judged to be the third level;

[0033] The preset first threshold is greater than the preset second threshold, and the sleep quality corresponding to the first level, the second level, and the third level decreases in turn.

[0034] Furthermore, the method for obtaining the curve segments includes:

[0035] For each sleep, the temperature curve is segmented at the same time interval to obtain multiple curve segments of each temperature curve.

[0036] The present invention also provides a sleep quality evaluation system based on sleep activity recognition, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any one of the sleep quality evaluation methods based on sleep activity recognition are implemented.

[0037] The present invention has the following beneficial effects:

[0038] For each sleep, the present invention obtains multiple curve segments of each temperature curve, and analyzes the temperature changes of each segment in more detail; according to the temperature data distribution at different times on each curve segment, obtains the likelihood of deep sleep for each curve segment, screens out suspected deep sleep segments, which helps to narrow the scope of subsequent analysis and improve the accuracy of analysis; according to the change trend of temperature data in the adjacent range before and after each suspected deep sleep segment, and the average likelihood of deep sleep corresponding to all curve segments, screens out the final deep sleep segments, which more accurately reflects the user's deep sleep situation; according to the sleep score in historical sleep and the distribution of the final deep sleep segments of the user, obtains the high-quality deep sleep coefficient of the user, which reflects the stability and change trend of the deep sleep quality of the user in different time periods; according to the difference between the proportion of the duration of all the final deep sleep segments in real-time sleep and the high-quality deep sleep coefficient of the user, and the number of the final deep sleep segments, obtains the likelihood of high-quality sleep in real-time sleep of the user, which reflects the probability of the user achieving high-quality sleep in real-time sleep; evaluates the sleep quality. By combining the analysis of historical sleep data, the present invention obtains the likelihood that the real-time sleep data of the user is high-quality sleep data, and improves the accuracy of sleep quality assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0040] Figure 1 It is a flowchart of a sleep quality assessment method based on sleep activity recognition provided by an embodiment of the present invention;

[0041] Figure 2 It is a flowchart of a method for obtaining the likelihood of deep sleep provided by an embodiment of the present invention;

[0042] Figure 3 It is a schematic diagram of an adjacent range before and after provided by an embodiment of the present invention;

[0043] Figure 4 It is a flowchart of a method for obtaining the high-quality deep sleep coefficient provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] To further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details a sleep quality assessment method and system based on sleep activity recognition proposed according to the present invention, including its specific implementation manner, structure, features, and effects. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0045] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0046] The following specifically describes the specific solution of a sleep quality assessment method and system based on sleep activity recognition provided by the present invention with reference to the accompanying drawings.

[0047] Please refer to Figure 1 , which shows a method flowchart of a sleep quality assessment method based on sleep activity recognition provided by an embodiment of the present invention, specifically including:

[0048] Step S1: Obtain the temperature curve and sleep score of the user during each sleep. The temperature curve includes temperature data at different times.

[0049] In an embodiment of the present invention, in order to more accurately reflect the unique sleep quality of each person, the distribution of temperature data during each sleep of the user is analyzed. First, a bracelet equipped with a motion sensor is used to monitor the user's movement for medical purposes, obtain the temperature data of the user during each sleep, and can judge the activity level of the user during sleep. The user scores the quality of each sleep, and the value range is 1 - 100. Obtain the temperature curve and sleep score of the user during each sleep. The temperature curve includes temperature data at different times.

[0050] It should be noted that in an embodiment of the present invention, the time interval is 1 minute, that is, the temperature data is obtained once per minute to form the temperature curve of each sleep. The abscissa is the time, and the ordinate is the temperature data. In other embodiments of the present invention, the time interval is a technical means well-known to those skilled in the art and will not be limited and elaborated herein.

[0051] Step S2: For each sleep, obtain multiple curve segments of each temperature curve; obtain the deep sleep possibility of each curve segment according to the temperature data distribution at different times on each curve segment, and screen out suspected deep sleep segments; screen out the final deep sleep segments according to the change trend of the temperature data in the adjacent range before and after each suspected deep sleep segment and the average deep sleep possibility of all curve segments.

[0052] During sleep, the user's body temperature fluctuates with the changes in sleep stages. The temperature curve is usually non-linear, containing multiple rising, falling, and stable stages. Directly analyzing the entire temperature curve may lead to information loss or misjudgment. Therefore, for each sleep, multiple curve segments of each temperature curve are obtained, and each segment can be analyzed independently to more accurately identify the temperature change patterns related to sleep stages.

[0053] It should be noted that in an embodiment of the present invention, for each sleep, the temperature curve is segmented at the same time interval to obtain multiple curve segments of each temperature curve. Among them, the size of the time interval is 5 minutes; in other embodiments of the present invention, the size of the time interval can be specifically set according to specific circumstances and will not be elaborated here.

[0054] Throughout the sleep process, the change in body temperature is closely related to sleep quality. After falling asleep, the body temperature begins to gradually decrease. When entering deep sleep, the body temperature drops to the lowest point and remains relatively stable; by analyzing the temperature data distribution at different times, the temperature change patterns related to deep sleep can be captured, and the deep sleep possibility of each curve segment can be obtained according to the temperature data distribution at different times on each curve segment.

[0055] Preferably, in an embodiment of the present invention, for the method of obtaining the deep sleep possibility, please refer to Figure 2 , which shows a flowchart of a method for obtaining the deep sleep possibility, including:

[0056] Step S201: For each sleep, obtain the mean value of the temperature data at all times on each curve segment as the average temperature level.

[0057] By taking the mean value to globally quantify the temperature data at all times on each curve segment, multiple data points within the segment can be simplified into a representative value, reflecting the overall temperature characteristics of the segment, thereby reducing data complexity and improving analysis efficiency.

[0058] Step S202: Obtain the cumulative sum of the differences between the temperature data at different times on each curve segment and the average temperature level.

[0059] The difference between the temperature data and the average temperature level reflects the temperature fluctuation within the curve segment. By calculating the cumulative sum of the differences, the temperature fluctuation degree of the segment can be quantified. The greater the difference, the greater the temperature fluctuation.

[0060] Step S203: Obtain the product of the average temperature level and the cumulative sum of the differences, and perform a negative correlation mapping as the deep sleep possibility of each curve segment.

[0061] In an embodiment of the present invention, for each sleep, the formula for the deep sleep probability is expressed as:

[0062] ;

[0063] where represents the deep sleep probability of the th curve segment; represents the average value of all temperature data of the th curve segment; represents the number of temperature data of the th curve segment; represents the rd temperature data of the th curve segment; represents the exponential function with the natural number as the base.

[0064] In the formula for the deep sleep probability, represents the cumulative sum of the differences between the temperature data at different times and the average temperature level on the th curve segment. The larger the cumulative sum of the differences, the greater the difference between the temperature data at different times and the average temperature level, the more uneven the distribution of the temperature data, and the smaller the deep sleep probability; the lower the average temperature level, the smaller the temperature data, the closer it is to the deep sleep state, and the greater the deep sleep probability.

[0065] The deep sleep probability is a quantitative index, which can convert the qualitative temperature change pattern into a quantitative probability value, can reflect the probability that a certain curve segment belongs to deep sleep, and screen out the suspected deep sleep segments.

[0066] Preferably, in an embodiment of the present invention, the method for obtaining the suspected deep sleep segments includes:

[0067] If the deep sleep probability of the curve segment is greater than or equal to the preset probability threshold, the corresponding curve segment is used as the initial deep sleep segment;

[0068] The continuously adjacent initial deep sleep segments are merged to form multiple suspected deep sleep segments.

[0069] It should be noted that the greater the deep sleep probability, the smaller the average temperature level, and the smaller the difference between the temperature data at different times and the average temperature level, that is, the lower and more stable the temperature within the corresponding curve segment, the greater the probability of deep sleep. In an embodiment of the present invention, in order to better balance the recognition accuracy and the risk of misjudgment, the preset probability threshold is set to a medium-high probability value, such as 0.6. In other embodiments of the present invention, the size of the preset probability threshold is a well-known technical means in the art and will not be limited and described herein.

[0070] After falling asleep, the body temperature begins to gradually decrease, reaches its lowest point and remains relatively stable. As the deep sleep stage ends, the body temperature begins to slowly rise. The deep sleep stage is usually accompanied by a decrease and stability in body temperature, while the light sleep or rapid eye movement (REM) sleep stage is usually accompanied by an increase in body temperature. By analyzing the trend of temperature data changes in the adjacent ranges before and after, the temperature change pattern related to deep sleep can be captured. Therefore, based on the trend of temperature data changes in the adjacent ranges before and after each suspected deep sleep segment, and the average deep sleep probability corresponding to all curve segments, the final deep sleep segments are screened out.

[0071] Preferably, in an embodiment of the present invention, the method for obtaining the final deep sleep segments includes:

[0072] Based on the trend of temperature data changes in the adjacent ranges before and after each suspected deep sleep segment, and the average deep sleep probability corresponding to all curve segments, obtain the deep sleep confidence level for each suspected deep sleep segment;

[0073] Preferably, in an embodiment of the present invention, the method for obtaining the deep sleep confidence level includes:

[0074] Obtain the mean value of the slopes between all adjacent data points in each adjacent range corresponding to each suspected deep sleep segment as the degree of temperature change;

[0075] Obtain the difference between the degrees of temperature change in the adjacent range after and the adjacent range before each suspected deep sleep segment, calculate the product of the difference result and the average deep sleep probability, and perform normalization as the deep sleep confidence level corresponding to the suspected deep sleep segment.

[0076] In an embodiment of the present invention, for any sleep, the formula for the deep sleep confidence level is expressed as:

[0077] ;

[0078] Wherein, represents the deep sleep confidence level of the th suspected deep sleep segment; represents the slope between the th and the th temperature data in the adjacent range before the th suspected deep sleep segment; represents the number of temperature data in the adjacent range before the th suspected deep sleep segment; represents the th temperature data in the adjacent range before the The slope between the th temperature data and the th temperature data; represents the number of temperature data within the adjacent range before the th suspected deep sleep segment; represents the mean of the deep sleep probabilities of all curve segments on the th suspected deep sleep segment, that is, the average deep sleep probability;

[0079] In the formula for deep sleep confidence, represents calculating the mean of the slopes between all adjacent data points within each adjacent range corresponding to each suspected deep sleep segment as the temperature change degree, and finding the difference between the temperature change degrees of the subsequent adjacent range and the previous adjacent range. The greater the difference, the greater the change in the temperature data of the subsequent adjacent range relative to the previous adjacent range, and the greater the deep sleep confidence.

[0080] If the deep sleep confidence of a suspected deep sleep segment is greater than or equal to the preset confidence threshold, the corresponding suspected deep sleep segment is taken as the final deep sleep segment.

[0081] It should be noted that in an embodiment of the present invention, the method for obtaining the slope is to calculate the ratio of the difference in the ordinates between two data points on the curve segment to the difference in the abscissas. The specific means are well-known technical means to those skilled in the art and will not be elaborated here; the method for obtaining the front and rear adjacent ranges includes, for the suspected deep sleep segment, selecting a preset number of temperature data before the suspected deep sleep segment to form the front adjacent range, and selecting a preset number of temperature data after the suspected deep sleep segment to form the rear adjacent range, where the preset number is 10, such as Figure 3 which shows a schematic diagram of a front and rear adjacent range; in other embodiments of the present invention, it can be specifically set according to specific circumstances and will not be limited and elaborated here.

[0082] It should be noted that considering that the change in the temperature data within the subsequent adjacent range is greater than that within the previous adjacent range, and the more the change within the subsequent adjacent range tends to an upward trend, the more the suspected deep sleep segment conforms to the deep sleep segment, and the greater the influence on the confidence. In an embodiment of the present invention, in order to identify most of the deep sleep segments and at the same time avoid misjudging too many non-deep sleep segments as deep sleep segments, the size of the preset confidence threshold is 0.7; in other embodiments of the present invention, the preset confidence threshold can be specifically set according to specific circumstances and will not be limited and elaborated here.

[0083] Step S3: Obtain the user's high-quality deep sleep coefficient according to the sleep score of the user in historical sleep and the distribution of the final segments of deep sleep; obtain the high-quality sleep possibility of the user in real-time sleep according to the difference between the proportion of the duration of all the final segments of deep sleep in real-time sleep and the high-quality deep sleep coefficient of the user, as well as the number of the final segments of deep sleep.

[0084] The deep sleep requirements of different users vary, and everyone's perception and experience of deep sleep also differ. Therefore, by combining the sleep scores of users after each sleep, the sleep status of each user can be more accurately evaluated; the more the distribution of the final segments of deep sleep, the relatively higher the sleep quality of the user, and then the high-quality deep sleep coefficient is evaluated; obtain the high-quality deep sleep coefficient of the user according to the sleep score of the user in historical sleep and the distribution of the final segments of deep sleep.

[0085] Preferably, in an embodiment of the present invention, for the method of obtaining the high-quality deep sleep coefficient, please refer to Figure 4 , which shows a flowchart of a method for obtaining the high-quality deep sleep coefficient, including:

[0086] Step S401: Obtain the ratio of the duration of all the final segments of deep sleep in each historical sleep to the total sleep duration as the local deep sleep coefficient.

[0087] The ratio of the duration of the final segments of deep sleep to the total sleep duration can quantify the proportion of deep sleep in the total sleep, reflecting the relative duration of deep sleep of the user in one sleep. The longer the relative duration, the larger the local deep sleep coefficient.

[0088] Step S402: If the sleep score of the user in historical sleep is greater than the preset score threshold, regard the corresponding historical sleep as the reference sleep.

[0089] The sleep score is a comprehensive index for evaluating sleep quality. By screening the historical sleep with a sleep score greater than the preset score threshold, the high-quality sleep of the user can be identified as the subsequent reference.

[0090] It should be noted that in order to ensure that the selected sleep records truly represent the high-quality sleep state of the user, a value reflecting the high standard of the user's sleep quality is selected. In an embodiment of the present invention, the preset score threshold is set to 90; in other embodiments of the present invention, it can be specifically set according to specific situations. The specific means are well-known technical means to those skilled in the art and will not be limited and elaborated here.

[0091] Step S403: Obtain the average value of the local deep sleep coefficients corresponding to all the reference sleeps as the high-quality deep sleep coefficient of the user.

[0092] In an embodiment of the present invention, the formula of the high-quality deep sleep coefficient is expressed as:

[0093] ;

[0094] Among them, represents the user's high-quality deep sleep coefficient; represents the duration of all the final segments of deep sleep corresponding to the th reference sleep; represents the total duration of the th reference sleep;

[0095] In the formula of the high-quality deep sleep coefficient, represents the ratio of the duration of all the final segments of deep sleep to the total sleep duration in the th reference sleep. The larger the ratio, the longer the duration of all the final segments of deep sleep in the

[0096] th reference sleep, and the reference sleep is high-quality deep sleep. The proportion of the duration of the final segments of deep sleep reflects the continuity and stability of deep sleep in the user's real-time sleep. A higher proportion of the duration of deep sleep usually means better deep sleep quality. The high-quality deep sleep coefficient is an indicator for comprehensively evaluating the deep sleep quality of the user in historical sleep. The difference between the proportion of the duration of the final segments of deep sleep in real-time sleep and the historical high-quality deep sleep coefficient can reflect the comparison between the user's real-time sleep state and long-term sleep quality; the number of the final segments of deep sleep reflects the dispersion degree of deep sleep in the user's real-time sleep. Too many segments may mean that the user's sleep is disturbed, resulting in discontinuous deep sleep. According to the difference between the proportion of the duration of all the final segments of deep sleep in the user's real-time sleep and the high-quality deep sleep coefficient, as well as the number of the final segments of deep sleep, the high-quality sleep possibility of the user in real-time sleep is obtained.

[0097] Preferably, in an embodiment of the present invention, the method for obtaining the high-quality sleep possibility includes:

[0098] Obtain the difference between the local deep sleep coefficient and the high-quality deep sleep coefficient of the user in real-time sleep, calculate the product of the difference result and the number of all the final segments of deep sleep in real-time sleep, and perform a negative correlation mapping as the high-quality sleep possibility of the user in real-time sleep.

[0099] In an embodiment of the present invention, the formula of the high-quality sleep possibility is expressed as:

[0100] ;

[0101] Among them, Indicates the probability of high-quality sleep for the user during real-time sleep; Indicates the high-quality deep sleep coefficient of the user; Indicates the duration of all deep sleep segments of the user during real-time sleep; Indicates the total duration of the user's real-time sleep; Indicates the number of all final deep sleep segments of the user during real-time sleep; Indicates the exponential function with the natural number as the base; Indicates the absolute value function.

[0102] In the formula for the probability of high-quality sleep, through the exponential function with the natural constant as the base, Negative correlation mapping is performed to obtain the value range ; Indicates the ratio of the duration of all deep sleep segments to the total duration of the user's real-time sleep, that is, the local deep sleep coefficient corresponding to real-time sleep. The larger the local deep sleep coefficient, the longer the duration of all deep sleep segments, and the better the sleep quality; Indicates the difference between the local deep sleep coefficient and the high-quality deep sleep coefficient of the user during real-time sleep. The larger the difference, the less close to high-quality deep sleep, and the smaller the difference, the closer to high-quality deep sleep; the more the number of all final deep sleep segments of the user during real-time sleep, the less continuous the deep sleep, and the worse the quality of deep sleep.

[0103] Step S4: Evaluate the sleep quality according to the probability of high-quality sleep of the user during real-time sleep.

[0104] Everyone's sleep needs and habits are different. The evaluation of the probability of high-quality sleep during real-time sleep is evaluated according to the individual differences of the user, which more comprehensively reflects the user's state during sleep. Evaluate the sleep quality according to the probability of high-quality sleep of the user during real-time sleep.

[0105] Preferably, in an embodiment of the present invention, evaluating the sleep quality according to the probability of high-quality sleep of the user during real-time sleep includes:

[0106] If the probability of high-quality sleep of the user during real-time sleep is greater than or equal to the preset first threshold, it is determined that the sleep quality is the first level;

[0107] If the probability of high-quality sleep of the user during real-time sleep is less than the preset first threshold and greater than the preset second threshold, it is determined that the sleep quality is the second level;

[0108] If the probability of high-quality sleep of the user during real-time sleep is less than or equal to the preset second threshold, it is determined that the sleep quality is the third level;

[0109] The preset first threshold is greater than the preset second threshold, and the sleep quality corresponding to the first level, the second level, and the third level decreases in turn.

[0110] It should be noted that the greater the possibility of high-quality sleep, the better the user's sleep quality. Therefore, the greater the possibility of high-quality sleep is greater than the preset first ratio, the better the high-quality sleep possibility. In an embodiment of the present invention, the preset first threshold is set to 0.8, and the preset second threshold is set to 0.6; that is , the sleep quality is at the first level and the quality is good; , the sleep quality is at the second level and the quality is average; , the sleep quality is at the third level and the quality is poor; In other embodiments of the present invention, the magnitudes of the preset first threshold and the preset second threshold can be specifically set according to specific situations, and are not limited and elaborated herein.

[0111] In summary, for each sleep, the present invention obtains multiple curve segments of each temperature curve; according to the temperature data distribution at different times on each curve segment, the final deep sleep segment is screened out; according to the sleep score of the user in historical sleep and the distribution of the final deep sleep segment, the high-quality deep sleep coefficient of the user is obtained; according to the difference between the duration ratio of all the final deep sleep segments in the user's real-time sleep and the high-quality deep sleep coefficient, and the number of the final deep sleep segments, the high-quality sleep possibility of the user in the real-time sleep is obtained; and the sleep quality is evaluated. By analyzing in combination with historical sleep data, the present invention obtains the possibility that the user's real-time sleep data is high-quality sleep data, and improves the accuracy of sleep quality assessment.

[0112] The present invention also proposes a sleep quality assessment system based on sleep activity recognition, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any one of the sleep quality assessment methods based on sleep activity recognition are implemented.

[0113] It should be noted that: the above-mentioned sequence of embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0114] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A sleep quality assessment method based on sleep activity recognition, characterized in that, The method comprises: Obtaining a temperature curve and a sleep score of the user during each sleep period, wherein the temperature curve includes temperature data at different times; For each sleep session, multiple curve segments of each temperature curve are obtained. Based on the temperature data distribution at different times on each curve segment, the deep sleep probability of each curve segment is obtained, and suspected deep sleep segments are screened out. Based on the change trend of the temperature data in the adjacent range before and after each suspected deep sleep segment, as well as the average deep sleep probability of all curve segments, the final deep sleep segment is screened out. The user's high-quality deep sleep coefficient is obtained based on the user's sleep score in historical sleep and the distribution of the final deep sleep segments. The user's high-quality deep sleep probability in real-time sleep is obtained based on the difference between the duration ratio of all the final deep sleep segments in real-time sleep and the high-quality deep sleep coefficient, as well as the number of final deep sleep segments. Evaluate sleep quality based on the user's likelihood of high-quality sleep in real time; The method for obtaining the final segment of deep sleep includes: The deep sleep confidence of each suspected deep sleep segment is obtained based on the change trend of the temperature data in the adjacent range before and after each suspected deep sleep segment and the average deep sleep probability of all curve segments; If the deep sleep confidence of the suspected deep sleep segment is greater than or equal to the preset confidence threshold, the corresponding suspected deep sleep segment is used as the final deep sleep segment; The method for obtaining the deep sleep confidence comprises: Obtain the average of the slopes between all adjacent data points in each adjacent range corresponding to each suspected deep sleep segment as the temperature change; The temperature variation difference between the rear adjacent range and the front adjacent range corresponding to each suspected deep sleep segment is obtained, and the product of the difference result and the average deep sleep possibility is calculated and normalized to obtain the deep sleep confidence of the corresponding suspected deep sleep segment.

2. The sleep quality assessment method based on sleep activity recognition according to claim 1, characterized in that: The method for obtaining the possibility of deep sleep includes: For each sleep period, the mean of the temperature data at all moments on each curve segment was obtained as the average temperature level; Obtain the cumulative sum of the differences between the temperature data at different moments on each curve segment and the average temperature level; The product between the mean temperature level and the cumulative sum of the differences was obtained and negatively correlated and mapped as the deep sleep probability for each curve segment.

3. The sleep quality assessment method based on sleep activity recognition according to claim 1, characterized in that, The method for obtaining the suspected deep sleep segment includes: If the deep sleep probability of the curve segment is greater than or equal to the preset probability threshold, the corresponding curve segment is used as the initial deep sleep segment; Continuous adjacent deep sleep initial segments are merged to form multiple suspected deep sleep segments.

4. The sleep quality assessment method based on sleep activity recognition according to claim 1, wherein, The method for obtaining the high-quality deep sleep coefficient includes: Obtain the ratio of the duration of all final deep sleep segments in each historical sleep period to the total sleep duration as the local deep sleep coefficient; If the user's sleep score in historical sleep is greater than the preset score threshold, the corresponding historical sleep will be used as reference sleep; The average value of the local deep sleep coefficients corresponding to all reference sleeps is obtained as the user's high-quality deep sleep coefficient.

5. The sleep quality assessment method based on sleep activity recognition according to claim 4, characterized in that, The method for obtaining the possibility of high-quality sleep includes: Obtain the difference between the local deep sleep coefficient and the high-quality deep sleep coefficient of the user during real-time sleep, calculate the product of the difference result and the number of all final segments of deep sleep during real-time sleep, and perform a negative correlation mapping, which is used as the high-quality sleep possibility of the user during real-time sleep.

6. A sleep quality assessment method based on sleep activity recognition according to claim 1, characterized in that, Evaluating the sleep quality according to the high-quality sleep possibility of the user during real-time sleep includes: If the high-quality sleep possibility of the user during real-time sleep is greater than or equal to a preset first threshold, determine that the sleep quality is the first level; If the high-quality sleep possibility of the user during real-time sleep is less than the preset first threshold and greater than the preset second threshold, determine that the sleep quality is the second level; If the high-quality sleep possibility of the user during real-time sleep is less than or equal to the preset second threshold, determine that the sleep quality is the third level; The preset first threshold is greater than the preset second threshold, and the sleep quality corresponding to the first level, the second level, and the third level decreases in turn.

7. A sleep quality assessment method based on sleep activity recognition according to claim 1, characterized in that The method for obtaining the curve segments includes: For each sleep, divide the temperature curve at the same time interval to obtain multiple curve segments of each temperature curve.

8. A sleep quality assessment system based on sleep activity recognition, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the steps of a sleep quality assessment method according to any one of claims 1 to 7.

Images