CN109793497A - Sleep state identification method and device - Google Patents
Sleep state identification method and device Download PDFInfo
- Publication number
- CN109793497A CN109793497A CN201711146375.9A CN201711146375A CN109793497A CN 109793497 A CN109793497 A CN 109793497A CN 201711146375 A CN201711146375 A CN 201711146375A CN 109793497 A CN109793497 A CN 109793497A
- Authority
- CN
- China
- Prior art keywords
- sleep state
- time point
- observation time
- dormant
- recognizer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000007958 sleep Effects 0.000 title claims abstract description 173
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000012549 training Methods 0.000 claims abstract description 36
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 230000003252 repetitive effect Effects 0.000 claims description 10
- 238000007637 random forest analysis Methods 0.000 claims description 9
- 238000012790 confirmation Methods 0.000 claims description 4
- 239000013598 vector Substances 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 12
- 230000001133 acceleration Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 6
- 238000012360 testing method Methods 0.000 description 5
- 230000005059 dormancy Effects 0.000 description 3
- 235000013399 edible fruits Nutrition 0.000 description 3
- 238000012512 characterization method Methods 0.000 description 2
- 230000000052 comparative effect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The application discloses a sleep state identification method, which comprises the following steps: acquiring sleep data of an observation object at an observation time point; inputting sleep data into a sleep state pre-recognizer to obtain a pre-recognition result of the sleep state of an observation object at an observation time point; the sleep state pre-recognizer is generated according to the known sleep data training of the sleep state in advance; and taking a pre-recognition result of which the repetition rate exceeds a preset threshold value as the sleep state of the observation time point from a plurality of pre-recognition results of a plurality of observation time points continuously adjacent to the observation time point. The sleep state pre-recognizer is generated by training according to the known sleep data of the sleep state in advance, the sleep data are recognized, and the method does not depend on experience rules, so that the method has high accuracy and adaptability. The application also discloses a sleep state recognition device, which has the beneficial effects.
Description
Technical field
This application involves sleep detection technical field, in particular to a kind of sleep state recognition methods and device.
Background technique
With the development of sleep detection technology, many has the portable intelligent wearable device of sleep state identification function
Market is come into.
Currently, sleep state recognition methods in the prior art is mostly empirically regular, specifically added with what measurement obtained
The comparing result of the signals such as speed and preset empirical value is basis of characterization.Common Judging index includes the unit time
The number of interior observation object movement, the signal strength of observation object or signal integration area are more than default respectively in the unit time
The number etc. of threshold value.
As it can be seen that the foundation of these recognition methods is single empirical rule in the prior art, and in fact, sleep state
It is difficult to be described with simple indicator rule, it, also can be in different time sections even being in dormant observation object
The movement such as inside there is the overturning of different frequency, amplitude, acceleration and intensity, stand up, and varying with each individual, and these are difficult with simple
Empirical rule uniformly summarize, it is more difficult to suitable for all observation objects.Therefore, sleep state identification in the prior art
Method often has biggish error, and adaptability is poor.
Which kind of sleep state recognition methods and device are used as a result, are these to effectively improve accuracy and adaptability
Field technical staff technical problem urgently to be resolved.
Summary of the invention
The application's is designed to provide a kind of sleep state recognition methods and device, so as to effectively improve accuracy and
Adaptability.
In order to solve the above technical problems, the application provides a kind of sleep state recognition methods, comprising:
Observation object is obtained in the dormant data of observation time point；
The dormant data is input in sleep state preliminary recognizer, obtains the observation object in the observation time
The dormant pre-identification result of point；The sleep state preliminary recognizer is instructed previously according to dormant data known to sleep state
Practice and generates；
From with multiple pre-identification results of multiple observation time points of the observation time point continuous adjacent, by repetitive rate
Sleep state more than the pre-identification result of preset threshold as the observation time point.
Optionally, the dormant data is acceleration information.
Optionally, further includes:
The acquisition observation object after the dormant data of observation time point, described the dormant data is input to
Before in sleep state preliminary recognizer, the dormant data is filtered.
Optionally, further includes:
Sleep state preliminary recognizer is input to after being filtered to the dormant data, by the dormant data
In before, calculate separately dormant data of the observation object within the observation time point forward and backward preset duration period
Statistical characteristics；
It is described that the dormant data is input in sleep state preliminary recognizer, the observation object is obtained in the observation
The dormant pre-identification result at time point includes:
The dormant data and the statistical characteristics are input in the sleep state preliminary recognizer, the sight is obtained
Object is surveyed in the dormant pre-identification result of the observation time point.
Optionally, the statistical characteristics is standard deviation.
Optionally, the sleep state preliminary recognizer generates packet previously according to the training of dormant data known to sleep state
It includes:
The sleep state preliminary recognizer previously according to dormant data known to sleep state using random forests algorithm or
The training of person's radial basis function kernel support vectors machine algorithm generates.
Optionally, further includes:
It is described using repetitive rate be more than preset threshold pre-identification result as the sleep state of the observation time point it
Afterwards, according to the dormant data and the sleep state, training is finely adjusted to the sleep state identifier.
Present invention also provides a kind of sleep state identification devices, comprising:
Obtain module: for obtaining observation object in the dormant data of observation time point；
Pre-identification module: for the dormant data to be input in sleep state preliminary recognizer, the observation pair is obtained
As the dormant pre-identification result in the observation time point；The sleep state preliminary recognizer is previously according to sleep state
Known dormant data training generates；
Confirmation module: for multiple pre-identification knots from multiple observation time points with the observation time point continuous adjacent
It is more than the pre-identification result of preset threshold as the sleep state of the observation time point using repetitive rate in fruit.
Optionally, further includes:
Preprocessing module: the dormant data for getting to the acquisition module is filtered.
Optionally, further includes:
Computing module: for calculating the observation object within the observation time point forward and backward preset duration period
The statistical characteristics of dormant data；
The pre-identification module is specifically used for:
The dormant data and the statistical characteristics are input in the sleep state preliminary recognizer, the sight is obtained
Object is surveyed in the dormant pre-identification result of the observation time point.
Sleep state recognition methods provided herein includes: to obtain observation object in the sleep number of observation time point
According to；The dormant data is input in sleep state preliminary recognizer, obtains the observation object in the observation time point
Dormant pre-identification result；The sleep state preliminary recognizer is given birth to previously according to the training of dormant data known to sleep state
At；From with multiple pre-identification results of multiple observation time points of the observation time point continuous adjacent, it is more than by repetitive rate
Sleep state of the pre-identification result of preset threshold as the observation time point.
As it can be seen that compared with the prior art, in sleep state recognition methods provided herein, by previously according to sleep
The sleep state preliminary recognizer that the training of dormant data known to state generates, can obtain shape of effectively sleeping according to dormant data
State recognition result.Since the training process of sleep state preliminary recognizer is independent of empirical rule, accuracy is available
It ensures.Meanwhile according to this method, respectively different sleep state preliminary recognizers can establish to different user groups, thus
The adaptability to user's individual difference can also be improved.Above-mentioned sleep may be implemented in sleep state identification device provided herein
Dormancy state identification method equally has above-mentioned beneficial effect.
Detailed description of the invention
In order to illustrate more clearly of the technical solution in the prior art and the embodiment of the present application, below will to the prior art and
Attached drawing to be used is needed to make brief introduction in the embodiment of the present application description.Certainly, in relation to the attached drawing of the embodiment of the present application below
A part of the embodiment in only the application of description is not paying creativeness to those skilled in the art
Under the premise of labour, other attached drawings can also be obtained according to the attached drawing of offer, other accompanying drawings obtained also belong to the application
Protection scope.
Fig. 1 is a kind of flow chart of sleep state recognition methods provided by the embodiment of the present application；
Fig. 2 is a kind of structural block diagram of sleep state identification device provided by the embodiment of the present application；
Fig. 3 is the pre-identification comparative result figure of sleep state recognition methods provided by the embodiment of the present application；
Fig. 4 is the final recognition result comparison diagram of sleep state recognition methods provided by the embodiment of the present application；
Fig. 5 is provided by the embodiment of the present application using the ROC curve of the sleep state preliminary recognizer of random forests algorithm
Figure.
Specific embodiment
The core of the application is to provide a kind of sleep state recognition methods and device, so as to effectively improve accuracy and
Adaptability.
In order to which technical solutions in the embodiments of the present application is more clearly and completely described, below in conjunction with this Shen
Please attached drawing in embodiment, technical solutions in the embodiments of the present application is introduced.Obviously, described embodiment is only
Some embodiments of the present application, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall in the protection scope of this application.
Referring to FIG. 1, Fig. 1 is a kind of flow chart of sleep state recognition methods provided by the embodiment of the present application, mainly
The following steps are included:
Step 1: obtaining observation object in the dormant data of observation time point.
Specifically, dormant data is generally preferred to the acceleration information got by accelerometer, and preferably three axis
Acceleration；It can certainly be other data such as collected heart rate data.Its frequency acquisition, i.e., two neighboring observation time point
Between interval time inverse, can also voluntarily select to design and Implement by those skilled in the art, the application to this not into
Row limits；But it preferably, can be selected within the scope of 20~30Hz.
Step 2: dormant data being input in sleep state preliminary recognizer, observation object sleeping in observation time point is obtained
The pre-identification result of dormancy state.
Wherein, sleep state preliminary recognizer is generated previously according to the training of dormant data known to sleep state.
Specifically, basis of characterization used by the application is dormant data known to sleep state, utilizes letter known to these
Breath can establish a sleep state preliminary recognizer using software approach by certain arithmetic programming training, allow to pair
The dormant data of input carries out pre-identification, and exports dormant pre-identification result.So-called pre-identification result includes two kinds
State, i.e., " sleep " and " awake ", and can specifically use digital signal " 1 " and " 0 " expression.As it can be seen that sleep state preliminary recognizer
In establishment process, the required information used only is output and input, and is not needed to be explicitly described out and be accorded between the two
The specific rules of conjunction, because without by rule limited, as long as obtaining enough training datas, sleep state preliminary recognizer
There can be preferable precision.
Step 3: from multiple pre-identification results of multiple observation time points of observation time point continuous adjacent, will repeat
Rate is more than sleep state of the pre-identification result of preset threshold as observation time point.
It specifically, can be with after the pre-identification result for obtaining each observation time point using sleep state preliminary recognizer
Pre-identification result is further processed and obtains final output result.Observation object when due to sleep state is not yet
Be it is completely motionless, especially within a period of time just fallen asleep, therefore, the sleep number got at this time by accelerometer etc.
According to sleep state may being enabled to judge by accident.In order to reduce error, can according to current observation time point continuous adjacent to be identified
Multiple pre-identifications of multiple observation time points are as a result, to judge final recognition result.It specifically, can will be described more
Repetitive rate is more than sleep of the pre-identification result of preset threshold as current observation time point to be identified in a pre-identification result
State.Certainly, which is generally 50%, for example, when having 501 in 1000 pre-identification results is " sleep ", just
It can be determined that sleep state is " sleep ".
Multiple observation time points mentioned here with current observation time point continuous adjacent to be identified, recommend but unlimited
In centered on current time point to be observed, that is, it is evenly distributed in front of and after observation time point, and its number can also be by ability
Field technique personnel voluntarily select.Generally, it can choose all in the 10-20 minute centered on current time point to be observed
The pre-identification result of observation time point.
As it can be seen that in sleep state recognition methods provided by the embodiment of the present application, by previously according to known to sleep state
The sleep state preliminary recognizer that generates of dormant data training, effective sleep state identification knot can be obtained according to dormant data
Fruit.Since the training process of sleep state preliminary recognizer is independent of empirical rule, the available guarantee of accuracy.Together
When, according to this method, it can establish respectively different sleep state preliminary recognizers to different user groups, to can also mention
Adaptability of the height to user's individual difference.
Sleep state recognition methods provided herein, on the basis of the above embodiments:
Optionally, further includes:
Obtain observation object after the dormant data of observation time point, dormant data be input to sleep state know in advance
Before in other device, dormant data is filtered.
Specifically, can be using arithmetic mean of instantaneous value filter method, middle position value filtering method or gaussian filtering method etc., it can also be straight
It is down-sampled to tap into row.Those skilled in the art voluntarily can select and be arranged, and this embodiment of the present application is not limited thereto.
For example, observation time point sum per minute is 25*60 i.e. 1500 point when sample frequency is 25Hz.If adopting
It is filtered with down-sampled mode, and specifically down-sampled with the frequency progress of 1/60Hz, then it can be from per minute 1500
It is chosen in point and retains a point.Certainly, which is specifically as follows the 751st point per minute, or first point,
100 points or the 1500th point etc., those skilled in the art voluntarily can select and be arranged, the embodiment of the present application to this simultaneously
Without limiting.
Optionally, further includes:
Before being input to after being filtered to dormant data, by dormant data in sleep state preliminary recognizer,
Calculate separately the statistical characteristics of dormant data of the observation object within the observation time point forward and backward preset duration period；
Dormant data is input in sleep state preliminary recognizer, obtains observation object in the sleep state of observation time point
Pre-identification result include:
Dormant data and statistical characteristics are input in sleep state preliminary recognizer, obtain observation object in observation time
The dormant pre-identification result of point.
Specifically, in order to further increase the accuracy of recognition result, sleep state pre-identification can also be further increased
The input data type of device, i.e., necessary not only for dormant data, it is also necessary to the statistical characteristics of dormant data.Common statistics is special
Value indicative includes variance, standard deviation, average value etc., wherein preferably standard deviation, standard deviation are to measure the dispersion degree of one group of data
Statistical characteristics, the dispersion degree of dormant data before and after Current observation time point can be reacted well.It is mentioned here pre-
If duration, voluntarily it can be selected and be arranged by those skilled in the art, recommended but be not limited to 2~4 minutes.
For example, for current observation time point T, selecting preset duration when dormant data is 3-axis acceleration data
It is 3 minutes, then needing the data being input in sleep state preliminary recognizer just includes that three axis corresponding to observation time point T accelerate
Spend ax、ayAnd azAnd all dormant data moulds are long in 3 minutes before observation time point TStandard deviation STDpreWith
All dormant data moulds are long in 3 minutes after observation time point TStandard deviation STDpos.
Referring to FIG. 3, Fig. 3 is the pre-identification Comparative result of sleep state recognition methods provided by the embodiment of the present application
Figure.
In Fig. 3, from top to bottom, the 1st column is the pre-identification knot using the sleep state preliminary recognizer of random forests algorithm
Fruit timing diagram, the 2nd column are the x-axis acceleration a gotxTiming diagram, the 3rd column is the y-axis acceleration a gotyTiming
Figure, the 4th column is the z-axis acceleration a gotzTiming diagram, the 5th column is mark in 3 minutes before calculated observation time point
Quasi- difference STDpreTiming diagram, the 6th column is standard deviation STD in 3 minutes after calculated observation time pointposTiming diagram, the 7th
Column is the true sleep state timing diagram of the subject of record, specifically can be according to subject's individual memory or other people observation
To determine.
From figure 3, it can be seen that the pre-identification result of sleep state preliminary recognizer output and the true sleep state of subject
It compares, there are points isolated on a small quantity, the i.e. point of false judgment.
Referring to FIG. 4, Fig. 4 is the final recognition result comparison of sleep state recognition methods provided by the embodiment of the present application
Figure.
In Fig. 4, from top to bottom, the 1st column is the true dormant timing diagram of the subject of record, and the 2nd column is to adopt
With the pre-identification result timing diagram of the sleep state preliminary recognizer of random forests algorithm, the 3rd column be output it is dormant most
Whole recognition result timing diagram.
Figure 4, it is seen that compared to the pre-identification of sleep state preliminary recognizer output as a result, defeated by step 3 institute
Dormant final recognition result out has preferable precision very close to the true sleep state of subject, can be right
False judgment during pre-identification is effectively corrected.
Optionally, sleep state preliminary recognizer is generated previously according to the training of dormant data known to sleep state includes:
Sleep state preliminary recognizer uses random forests algorithm or diameter previously according to dormant data known to sleep state
It is generated to the training of basic function kernel support vectors machine algorithm.
Specifically, random forests algorithm is less prone to over-fitting because of the introducing of randomness, and has anti-well
The adaptability of noise immune and data set, therefore there is preferable recognition result.In addition, radial basis function kernel support vectors machine
Algorithm is a kind of algorithm for being usually used in carrying out pattern-recognition, and is had radial basis function as the support vector machines of kernel function more
Good is of overall importance, therefore also has preferable recognition effect.
Optionally, further includes:
Be more than using repetitive rate preset threshold pre-identification result as the sleep state of observation time point after, according to sleeping
Dormancy data and sleep state are finely adjusted training to sleep state identifier.
Specifically, sleep state identifier can also be further increased after completing identification by fine tuning training accurate
Property.After end of identification, it is contemplated that the sleep state obtained at this time has higher confidence level, therefore, can be from having obtained
Select a part as training data into the data of sleep state recognition result, and pre- to sleep state according to the training data
Identifier re-starts training, to improve its accuracy in time.In training process, it is preferred to use grid data service selects most
Excellent parameter.
Optionally, the sample of identical quantity can be selected from the dormant data that recognition result is " sleep " and " awake " respectively
As training data.Recommend but is not limited to select 80% data in total sample size as training data, and other 20%
Sample be then also used as test data, to fine tuning training after sleep state preliminary recognizer test.Also, training and
Test can repeat repeatedly, for example, if selecting 20% sample for test data every time, recommend but be not limited to repeatedly into
Row 5 times, to ensure the fine tuning training precision to sleep state preliminary recognizer.And, so it is easy to understand that it is used for ensure
Dormant data and its corresponding sleep state for finely tuning training are representative, can be from obtaining the sleep of sleep state recognition result
Select the dormant data in the continuous 12 hour time as total sample size in data, then from total sample size select training data and
Test data.
Referring to FIG. 5, Fig. 5 is the sleep state preliminary recognizer provided by the embodiment of the present application using random forests algorithm
ROC curve (Receiver Operating Characteristic Curve, experience linearity curve) figure.
ROC curve is used to characterize the classifying quality of binary classification model, be using true positive rate i.e. sensitivity as ordinate, with
False positive rate, that is, 1- specificity is the curve of abscissa, and coordinate range is between 0~1 in length and breadth；ROC curve is closer to upper left
Angle, i.e. area under a curve are bigger, illustrate that the accuracy of the disaggregated model of tested person is higher.From fig. 5, it can be seen that this Shen
Please be provided by embodiment using the sleep state preliminary recognizer of random forests algorithm, the area under ROC curve is close to 1,
With more accurate recognition effect.
Sleep state identification device provided by the embodiment of the present application is introduced below.
Referring to Fig. 2, Fig. 2 is a kind of structural block diagram of sleep state identification device provided herein；Including obtaining
Module 1, pre-identification module 2 and confirmation module 3；
Module 1 is obtained for obtaining observation object in the dormant data of observation time point；
Pre-identification module 2 obtains observation object and is observing for dormant data to be input in sleep state preliminary recognizer
The dormant pre-identification result at time point；Sleep state preliminary recognizer is instructed previously according to dormant data known to sleep state
Practice and generates；
Confirmation module 3 is used for multiple pre-identification results from multiple observation time points with observation time point continuous adjacent
In, it is more than the pre-identification result of preset threshold as the sleep state of observation time point using repetitive rate.
As it can be seen that sleep state identification device provided herein, by previously according to number of sleeping known to sleep state
According to the sleep state preliminary recognizer that training generates, effective sleep state recognition result can be obtained according to dormant data.Due to
The training process of sleep state preliminary recognizer is independent of empirical rule, therefore the available guarantee of accuracy.Meanwhile according to this
Method can establish respectively different sleep state preliminary recognizers to different user groups, to can also improve to user
The adaptability of individual difference.
Sleep state identification device provided herein, on the basis of the above embodiments:
As a kind of preferred embodiment, further includes:
Preprocessing module: the dormant data for getting to acquisition module 1 is filtered.
As a kind of preferred embodiment, further includes:
Computing module: for dormant data of the calculating observation object within the observation time point forward and backward preset duration period
Statistical characteristics；
Pre-identification module 2 is specifically used for:
Dormant data and statistical characteristics are input in sleep state preliminary recognizer, obtain observation object in observation time
The dormant pre-identification result of point.
The specific embodiment of sleep state identification device provided herein and sleep state as described above are known
Other method can correspond to each other reference, just repeat no more here.
Each embodiment is described in a progressive manner in the application, the highlights of each of the examples are with other realities
The difference of example is applied, the same or similar parts in each embodiment may refer to each other.For device disclosed in embodiment
Speech, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is referring to method part illustration
?.
Technical solution provided herein is described in detail above.Specific case used herein is to this Shen
Principle and embodiment please is expounded, the present processes that the above embodiments are only used to help understand and its
Core concept.It should be pointed out that for those skilled in the art, in the premise for not departing from the application principle
Under, can also to the application, some improvement and modification can also be carried out, these improvement and modification also fall into the protection of the claim of this application
In range.
Claims (10)
1. a kind of sleep state recognition methods characterized by comprising
Observation object is obtained in the dormant data of observation time point；
The dormant data is input in sleep state preliminary recognizer, obtains the observation object in the observation time point
Dormant pre-identification result；The sleep state preliminary recognizer is given birth to previously according to the training of dormant data known to sleep state
At；
From with multiple pre-identification results of multiple observation time points of the observation time point continuous adjacent, it is more than by repetitive rate
Sleep state of the pre-identification result of preset threshold as the observation time point.
2. sleep state recognition methods according to claim 1, which is characterized in that the dormant data is to accelerate degree
According to.
3. sleep state recognition methods according to claim 1, which is characterized in that further include:
The acquisition observation object after the dormant data of observation time point, described the dormant data is input to sleep
Before in state preliminary recognizer, the dormant data is filtered.
4. sleep state recognition methods according to claim 3, which is characterized in that further include:
It in sleep state preliminary recognizer is input to after being filtered to the dormant data, by the dormant data
Before, calculate separately the statistics of dormant data of the observation object within the observation time point forward and backward preset duration period
Characteristic value；
It is described that the dormant data is input in sleep state preliminary recognizer, the observation object is obtained in the observation time
Point dormant pre-identification result include:
The dormant data and the statistical characteristics are input in the sleep state preliminary recognizer, the observation pair is obtained
As the dormant pre-identification result in the observation time point.
5. sleep state recognition methods according to claim 4, which is characterized in that the statistical characteristics is standard deviation.
6. sleep state recognition methods according to any one of claims 1 to 5, which is characterized in that the sleep state is pre-
Identifier is generated previously according to the training of dormant data known to sleep state
The sleep state preliminary recognizer uses random forests algorithm or diameter previously according to dormant data known to sleep state
It is generated to the training of basic function kernel support vectors machine algorithm.
7. sleep state recognition methods according to claim 6, which is characterized in that further include:
It is described using repetitive rate be more than preset threshold pre-identification result as the sleep state of the observation time point after, root
According to the dormant data and the sleep state, training is finely adjusted to the sleep state identifier.
8. a kind of sleep state identification device characterized by comprising
Obtain module: for obtaining observation object in the dormant data of observation time point；
Pre-identification module: it for the dormant data to be input in sleep state preliminary recognizer, obtains the observation object and exists
The dormant pre-identification result of the observation time point；The sleep state preliminary recognizer is previously according to known to sleep state
Dormant data training generate；
Confirmation module: for multiple pre-identification results from multiple observation time points with the observation time point continuous adjacent
In, it is more than the pre-identification result of preset threshold as the sleep state of the observation time point using repetitive rate.
9. sleep state identification device according to claim 8, which is characterized in that further include:
Preprocessing module: the dormant data for getting to the acquisition module is filtered.
10. according to sleep state identification device described in claim 8 or 9, which is characterized in that further include:
Computing module: for calculating sleep of the observation object within the observation time point forward and backward preset duration period
The statistical characteristics of data；
The pre-identification module is specifically used for:
The dormant data and the statistical characteristics are input in the sleep state preliminary recognizer, the observation pair is obtained
As the dormant pre-identification result in the observation time point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711146375.9A CN109793497B (en) | 2017-11-17 | 2017-11-17 | Sleep state identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711146375.9A CN109793497B (en) | 2017-11-17 | 2017-11-17 | Sleep state identification method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109793497A true CN109793497A (en) | 2019-05-24 |
| CN109793497B CN109793497B (en) | 2022-08-19 |
Family
ID=66554795
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711146375.9A Active CN109793497B (en) | 2017-11-17 | 2017-11-17 | Sleep state identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109793497B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110742580A (en) * | 2019-09-18 | 2020-02-04 | 华为技术有限公司 | Sleep state identification method and device |
| CN111012132A (en) * | 2019-12-17 | 2020-04-17 | 珠海格力电器股份有限公司 | Sleep state adjusting method and device based on pillow and intelligent pillow |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050043652A1 (en) * | 2003-08-18 | 2005-02-24 | Lovett Eric G. | Sleep state classification |
| CN103961109A (en) * | 2014-05-05 | 2014-08-06 | 北京航空航天大学 | Human body posture detection device based on acceleration signals and angular speed signals |
| CN104138260A (en) * | 2014-07-02 | 2014-11-12 | 中山大学 | Sleeping posture multi-classifying identifying method utilizing SVM classifier |
| CN104905795A (en) * | 2015-06-15 | 2015-09-16 | 深圳市奋达科技股份有限公司 | Method for monitoring Bluetooth low energy (BLE) networking sleeping |
| CN105496362A (en) * | 2015-12-15 | 2016-04-20 | 成都麦杰康科技有限公司 | Sleep state monitoring system and method |
| CN105678300A (en) * | 2015-12-30 | 2016-06-15 | 成都数联铭品科技有限公司 | Complex image and text sequence identification method |
| CN106419893A (en) * | 2016-09-18 | 2017-02-22 | 广州视源电子科技股份有限公司 | Method and device for detecting sleep state |
-
2017
- 2017-11-17 CN CN201711146375.9A patent/CN109793497B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050043652A1 (en) * | 2003-08-18 | 2005-02-24 | Lovett Eric G. | Sleep state classification |
| CN103961109A (en) * | 2014-05-05 | 2014-08-06 | 北京航空航天大学 | Human body posture detection device based on acceleration signals and angular speed signals |
| CN104138260A (en) * | 2014-07-02 | 2014-11-12 | 中山大学 | Sleeping posture multi-classifying identifying method utilizing SVM classifier |
| CN104905795A (en) * | 2015-06-15 | 2015-09-16 | 深圳市奋达科技股份有限公司 | Method for monitoring Bluetooth low energy (BLE) networking sleeping |
| CN105496362A (en) * | 2015-12-15 | 2016-04-20 | 成都麦杰康科技有限公司 | Sleep state monitoring system and method |
| CN105678300A (en) * | 2015-12-30 | 2016-06-15 | 成都数联铭品科技有限公司 | Complex image and text sequence identification method |
| CN106419893A (en) * | 2016-09-18 | 2017-02-22 | 广州视源电子科技股份有限公司 | Method and device for detecting sleep state |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110742580A (en) * | 2019-09-18 | 2020-02-04 | 华为技术有限公司 | Sleep state identification method and device |
| CN111012132A (en) * | 2019-12-17 | 2020-04-17 | 珠海格力电器股份有限公司 | Sleep state adjusting method and device based on pillow and intelligent pillow |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109793497B (en) | 2022-08-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107341473B (en) | Palm characteristic recognition method, palm characteristic identificating equipment and storage medium | |
| CN108960182B (en) | P300 event related potential classification identification method based on deep learning | |
| CN105877766B (en) | A kind of state of mind detection system and method based on the fusion of more physiological signals | |
| CN104484644B (en) | A kind of gesture identification method and device | |
| CN107890339B (en) | Sleep staging detection method and wearable sleep staging detection device | |
| CN103927851B (en) | A kind of individualized multi thresholds fall detection method and system | |
| CN103955270B (en) | Character high-speed input method of brain-computer interface system based on P300 | |
| CN111288986B (en) | Motion recognition method and motion recognition device | |
| CN105224104B (en) | Pedestrian movement's state identification method based on smart mobile phone grip mode | |
| CN106175754B (en) | Waking state detection device in sleep state analysis | |
| CN110532850B (en) | Fall detection method based on video joint points and hybrid classifier | |
| CN107609477A (en) | It is a kind of that detection method is fallen down with what Intelligent bracelet was combined based on deep learning | |
| CN107358248B (en) | Method for improving falling detection system precision | |
| CN109793497A (en) | Sleep state identification method and device | |
| CN106503643A (en) | Tumble detection method for human body | |
| JP2007229218A (en) | Apparatus, system, and method for vigilance estimation | |
| CN108008151A (en) | A kind of moving state identification method and system based on 3-axis acceleration sensor | |
| CN112883802A (en) | Method for identifying destructive event of pipeline optical fiber vibration safety early warning system | |
| CN106333676B (en) | The annotation equipment of EEG signals data type under waking state | |
| CN107169334B (en) | The user authen method based on straight punch motion detection for hand wearable device | |
| CN106901689A (en) | Cognitive and motor function detecting system and data processing method based on interaction in kind | |
| CN107123126A (en) | A kind of stream of people's moving scene temperature method of estimation | |
| CN110693510A (en) | Attention deficit hyperactivity disorder auxiliary diagnosis device and using method thereof | |
| KR101793934B1 (en) | Method and apparatus for automatically classifying types of weight training workouts | |
| CN106361327B (en) | Waking state detection method and system in sleep state analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |