CN110236558A - Infant development situation prediction technique, device, storage medium and electronic equipment - Google Patents
Infant development situation prediction technique, device, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN110236558A CN110236558A CN201910346364.8A CN201910346364A CN110236558A CN 110236558 A CN110236558 A CN 110236558A CN 201910346364 A CN201910346364 A CN 201910346364A CN 110236558 A CN110236558 A CN 110236558A
- Authority
- CN
- China
- Prior art keywords
- data
- sample
- limb motion
- tested
- single limb
- 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.)
- Pending
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
- A61B5/1113—Local tracking of patients, e.g. in a hospital or private home
- A61B5/1114—Tracking parts of the body
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7271—Specific aspects of physiological measurement analysis
- A61B5/7275—Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B2503/00—Evaluating a particular growth phase or type of persons or animals
- A61B2503/04—Babies, e.g. for SIDS detection
Abstract
A kind of infant development situation prediction technique, device, storage medium and electronic equipment are disclosed, technical field of computer programs is belonged to.This method comprises: obtaining the single limb motion data and corresponding monthly age data of sample to be tested；Data processing is carried out, the characteristic value for sample to be tested single limb motion corresponding with monthly age data is obtained；The characteristic value of the single limb motion corresponding with sample to be tested of pre-determined number and corresponding monthly age data are brought into the mathematical model of stack extreme learning machine, logical operation is carried out；The output result of the logical operation of extreme learning machine develops normal or hypoevolutism foundation as sample to be tested in a stacked, obtains the prediction result of infant development situation.The device, storage medium and electronic equipment are implemented for this method.Developmental state prediction can be carried out to baby by it, a possibility that infant development is slow is further obtained, hereby it is possible to provide reliable foundation for the slow early intervention of infant development.
Description
Technical field
The present invention relates to technical field of computer programs, more particularly to a kind of infant development situation prediction technique, device,
Storage medium and electronic equipment.
Background technique
The prior art generally be intended to just to make a definite diagnosis after 2 years old until infant infant whether hypoevolutism.However, up to
Just make a definite diagnosis that infant development is slow at this time, relatively late to the intervention of infant, the infant for being unfavorable for hypoevolutism is extensive
It is multiple.It is therefore desirable to which a kind of can be for the method that infant development situation is predicted, so as to the baby children for hypoevolutism
The recovery of youngster, which intervene, provides foundation.
Summary of the invention
In view of this, the present invention provides a kind of infant development situation prediction technique, device, storage medium and electronics to set
It is standby, developmental state prediction can be carried out to baby by it, further obtain a possibility that infant development is slow, hereby it is possible to
Reliable foundation is provided for the slow early intervention of infant development, thus more suitable for practical.
In order to reach above-mentioned first purpose, the technical solution of infant development situation prediction technique provided by the invention is such as
Under:
Infant development situation prediction technique provided by the invention the following steps are included:
Obtain the single limb motion data and corresponding monthly age data of sample to be tested；
Data processing is carried out for the single limb motion data of the sample to be tested, is obtained corresponding with monthly age data
For the characteristic value of the sample to be tested single limb motion；
By the characteristic value of the single limb motion corresponding with the sample to be tested of pre-determined number and corresponding monthly age number
According to bringing into the mathematical model of stack extreme learning machine, logical operation is carried out；
Developed using the output result of the logical operation of the stack extreme learning machine as the sample to be tested it is normal or
The foundation of person's hypoevolutism obtains the prediction result of the infant development situation.
Infant development situation prediction technique provided by the invention also can be used following technical measures and further realize.
Preferably, the construction method of the mathematical model of the stack extreme learning machine the following steps are included:
Obtain the single limb motion data and corresponding monthly age data of training sample, wherein the training sample is
Know the sample of development condition；
Data processing is carried out for the single limb motion data of the training sample, is obtained corresponding with monthly age data
For the characteristic value of the training sample single limb motion；
The characteristic value for the training sample single limb motion corresponding with monthly age data is brought into patrolling
Collect operational formula:When the training sample is multiple, obtain multiple based on institute
State the expression formula of logical operation formula；
Wherein,
L- is the number of nodes of neural networks with single hidden layer,
G (x)-is activation primitive,
wi=[Wi1,wi2,...,win]TIt is the input weight of i-th of Hidden unit,
Bi- is the biasing of i-th of Hidden unit,
βi=[βi1,βi2,...,βim]TIt is the output weight of i-th of Hidden unit,
Oj- be for training sample, carry out sorted logical operation output using stack extreme learning machine as a result, its
In, the logical operation output result includes developing normal and 2 major class of hypoevolutism；
According to the multiple expression formula based on the logical operation formula, the expression of the activation primitive g (x) is determined
Formula, the input weight wi of i-th of Hidden unit, i-th of Hidden unit biasing bi and i-th of hidden layer list
The output weight beta i of member；
Again by the expression formula of the activation primitive g (x) after determination, the input weight wi of i-th of Hidden unit, described
The biasing bi of i Hidden unit and the output weight beta i band of i-th of Hidden unit are back to the logical operation formula, obtain
To the mathematical model of the stack extreme learning machine.
Preferably, the single limb motion data include:
The moment beginning T of single limb motioni0, last moment Ti1, average acceleration av, peak accelerator am, left leg move class
Type SL, right leg type of sports SR；
Wherein, single motion duration tiIt is single motion last moment and single motion beginning moment Ti0Difference Ti1-
Ti0。
Preferably, the single limb motion data for the sample to be tested carry out data processing, obtain and monthly age number
It is special that single argument feature selecting, recurrence are selected from according to the method for the corresponding characteristic value for the sample to be tested single limb motion
Sign selection, gradually one of feature selecting, the single limb motion data for the sample to be tested carry out at data
Reason, obtains the characteristic value for the sample to be tested single limb motion corresponding with monthly age data and specifically includes:
The single argument feature selecting passes through the statistical measures side to the unitary variant in the single limb motion data
Method, selection obtain the characteristic value as the single limb motion；Or
The recursive feature selection is by being normalized at data each variable in the single limb motion data
Reason, using obtained data as the characteristic value of the single limb motion；Or
The gradually feature selecting passes through the unitary variant chosen in the single limb motion data one by one, to the list
The spy that the data that data processing obtains successively are used as single limb motion is normalized in each variable in secondary limb motion data
Value indicative.
Preferably, in logical operation output result,
According to the mathematical model of the stack extreme learning machine, different classifications is normally carried out to development；
According to the mathematical model of the stack extreme learning machine, different classifications is carried out to hypoevolutism.
In order to reach above-mentioned second purpose, the technical solution of infant development situation prediction meanss provided by the invention is such as
Under:
Infant development situation prediction meanss provided by the invention include:
Data capture unit, for obtaining the single limb motion data and corresponding monthly age data of sample to be tested；
Data processing unit carries out data processing for the single limb motion data for the sample to be tested, obtains
The characteristic value for the sample to be tested single limb motion corresponding with monthly age data；
Arithmetic element, for by treated the limb motion data corresponding with the sample to be tested of pre-determined number and
Corresponding monthly age data are brought into the mathematical model of stack extreme learning machine, and logical operation is carried out；
Prediction result output unit, for using the output result of the logical operation of the stack extreme learning machine as institute
It states sample to be tested and develops normal or hypoevolutism foundation, obtain the prediction result of the infant development situation.
In order to reach above-mentioned third purpose, the technical solution of storage medium provided by the invention is as follows:
Infant development situation Prediction program, the infant development situation prediction are stored on storage medium provided by the invention
The step of infant development situation prediction technique provided by the invention is realized when program is executed by processor.
In order to reach above-mentioned 4th purpose, the technical solution of electronic equipment provided by the invention is as follows:
Electronic equipment provided by the invention includes motion sensor, processor, memory and is stored on the memory
And the infant development situation Prediction program that can be run on the processor, wherein
The motion sensor, for obtaining the single limb motion data of sample to be tested；
The infant development situation Prediction program realizes infant development feelings provided by the invention when being executed by the processor
The step of condition prediction technique.
Infant development situation prediction technique, device, storage medium and electronic equipment provided by the invention are in the stack limit
Study and mathematical model known in situation, obtain the single limb motion data and corresponding monthly age data of sample to be tested,
This feature value is substituting to known stack by the characteristic value that sample to be tested single limb motion can be obtained by data processing
The limit study and mathematical model, carry out logical operation, output result can be obtained, wherein output result include hypoevolutism
With normal 2 major class of development, wherein since the date of birth that the monthly age data of sample to be tested can be shown according to birth certificate is obtained
, that is to say, that in infant development situation prediction technique provided by the invention, data to be obtained are only single limb motion
Data, in such a case, it is possible to be directly obtained by motion sensor, therefore, using infant development feelings provided by the invention
Condition prediction technique can more easily predict infant development situation, in addition, in infant development feelings provided by the invention
In condition prediction technique, the stack limit study and mathematical model during building, the baby at all monthly ages can be covered,
Therefore, even if the monthly age of sample to be tested is smaller, can also by the stack limit learn and mathematical model aligned
Therefore true output is as a result, can provide reliable foundation for the slow early intervention of infant development.
Detailed description of the invention
By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field
Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
Fig. 1 is that the infant development situation for the hardware running environment that the embodiment of the present invention is related to predicts device structure signal
Figure；
Fig. 2 is the step flow chart for the infant development situation prediction technique that the embodiment of the present invention is related to；
The stack limit study applied in the infant development situation prediction technique that Fig. 3 is related to for the embodiment of the present invention
The step flow chart of the construction method of the mathematical model of machine；
Signal in the infant development situation prediction meanss that Fig. 4 is related to for the embodiment of the present invention between each functional module
Flow to relation schematic diagram；
The stack limit study applied in the infant development situation prediction technique that Fig. 5 is related to for the embodiment of the present invention
The schematic illustration of machine.
Specific embodiment
The present invention in order to solve the problems existing in the prior art, provides a kind of infant development situation prediction technique, device, storage
Medium and electronic equipment can carry out developmental state prediction to baby by it, further obtain the slow possibility of infant development
Property, hereby it is possible to reliable foundation be provided for the slow early intervention of infant development, thus more suitable for practical.
It is of the invention to reach the technical means and efficacy that predetermined goal of the invention is taken further to illustrate, below in conjunction with
Attached drawing and preferred embodiment set infant development situation prediction technique, device, storage medium and electronics proposed according to the present invention
Standby, specific embodiment, structure, feature and its effect, detailed description is as follows.In the following description, a different " implementation
What example " or " embodiment " referred to is not necessarily the same embodiment.In addition, the feature, structure or feature in one or more embodiments can
It is combined by any suitable form.
The terms "and/or", only a kind of incidence relation for describing affiliated partner, indicates that there may be three kinds of passes
System, it is specific to understand for example, A and/or B are as follows: it can simultaneously include A and B, can be with individualism A, it can also be with individualism
B can have above-mentioned three kinds of any case.
Referring to Fig.1, Fig. 1 is the pre- measurement equipment of infant development situation for the hardware running environment that the embodiment of the present invention is related to
Structural schematic diagram.
As shown in Figure 1, the pre- measurement equipment of infant development situation may include: processor 1001, such as central processing unit
(CentralProcessingUnit, CPU), communication bus 1002, user interface 1003, network interface 1004, memory
1005.Wherein, communication bus 1002 is for realizing the connection communication between these components.User interface 1003 may include display
Shield (Display), input unit such as keyboard (Keyboard), optional user interface 1003 can also include that the wired of standard connects
Mouth, wireless interface.Network interface 1004 optionally may include standard wireline interface and wireless interface (such as Wireless Fidelity
(WIreless-FIdelity, WI-FI) interface).Memory 1005 can be the random access memory of high speed
(RandomAccessMemory, RAM) memory is also possible to stable nonvolatile memory (Non-
VolatileMemory, NVM), such as magnetic disk storage.Memory 1005 optionally can also be independently of aforementioned processor
1001 storage device.
It will be understood by those skilled in the art that structure shown in Fig. 1 is not constituted to the pre- measurement equipment of infant development situation
Restriction, may include perhaps combining certain components or different component layouts than illustrating more or fewer components.
As shown in Figure 1, as may include operating system, data storage mould in a kind of memory 1005 of storage medium
Block, network communication module, Subscriber Interface Module SIM and infant development situation Prediction program.
In the pre- measurement equipment of infant development situation shown in Fig. 1, network interface 1004 be mainly used for and network server into
Row data communication；User interface 1003 is mainly used for carrying out data interaction with user；The pre- measurement equipment of infant development situation of the present invention
In processor 1001, memory 1005 can be set in the pre- measurement equipment of infant development situation, infant development situation prediction set
It is standby that the infant development situation Prediction program stored in memory 1005 is called by processor 1001, and execute the embodiment of the present invention
The infant development situation prediction technique of offer.
Embodiment one
Referring to attached drawing 1, infant development situation prediction technique that the embodiment of the present invention one provides the following steps are included:
Step 101: obtaining the single limb motion data and corresponding monthly age data of sample to be tested.
Specifically, obtaining the limb motion data of sample to be tested here by sensor.Wherein, limb motion data
Including single motion duration, single motion average acceleration av, single motion peak accelerator am, left leg type of sports SL、
Right leg type of sports SR.Wherein, single motion duration tiIt is single motion last moment Ti1With single motion beginning moment Ti0's
Difference, that is, Ti1-Ti0.Wherein, left leg type of sports SL, right leg type of sports SRIt can be shown using the mode of verbal description, example
Such as, it can qualitatively describe to kick, squat down, walk, creeping, being bent, a symbol can also be defined to each type of sports, such as
Kicking=Q1, squat down=Q2, walking=Q3, creep=Q4, bending=Q5, then, it is described in a manner of symbol.In addition, by
When infant is especially in 12 months in 24 months, limb action ability development is more rapid, sometimes, even if differing only by number
It, the limb action of infant can have greatly changed, and therefore, the monthly age, data Y needed to be accurate to day, for example, monthly age number
According to can be 2m+1,8m+10 etc..
Step 102: carrying out data processing for the single limb motion data of sample to be tested, obtain opposite with monthly age data
The characteristic value for sample to be tested single limb motion answered.
Specifically, may include for the mode of the single limb motion data progress data processing of the sample to be tested
Single argument feature selecting, recursive feature selection, the gradually modes such as feature selecting, wherein
Single argument feature selecting is referred to by being chosen based on some univariate statistically independents to as to test sample
The characteristic value X of this single limb motionj.At this point, first embodiment can be Xj={ Ti1-Ti0, m }；Second embodiment can be
Xj={ av, Y }；3rd embodiment can be Xj={ am, Y }；Fourth embodiment can be Xj={ XL, XR, Y }.
Recursive feature selection refers to by the way that data processing is normalized to above-mentioned each variable, using obtained data as
The characteristic value X of sample to be tested single limb motionj.For example, X at this timej={ y1(Ti1-Ti0)+y2av+y3am, XL, XR, Y }, wherein
y1For (Ti1-Ti0) characteristic coefficient, y2For avCharacteristic coefficient, y3For amCharacteristic coefficient.In the present embodiment, the normalizing taken
Changing data processing method is linear normalization data processing, can also be directed to above-mentioned (T according to actual needsi1-Ti0)、av、amThis
Certain carry out power operations in three features can also be directed to above-mentioned (T so that increasing it influences gradei1-Ti0)、av、amThis
Certain carry out extracting operations in three features, to reduce influence grade.
Gradually feature selecting refers to chooses X for the first timej={ Ti1-Ti0, Y }, second of selection Xj={ av, Y }, third
Secondary selection Xj={ am, Y }, the 4th selection Xj={ XL, XR, Y }；It is then also possible to the 5th selection Xj={ y1(Ti1-Ti0)+
y2av+y3am, XL, XR, Y }.
Step 103: by the characteristic value of the single limb motion corresponding with sample to be tested of pre-determined number and the corresponding moon
Age data are brought into the mathematical model of stack extreme learning machine, and logical operation is carried out.
Specifically, by the characteristic value X of above-mentioned sample to be tested single limb motionjIt is substituting to stack extreme learning machine
Mathematical model is
Wherein,
L- is the number of nodes of neural networks with single hidden layer,
G (x)-is activation primitive,
wi=[Wi1,wi2,...,win]TIt is the input weight of i-th of Hidden unit,
Bi- is the biasing of i-th of Hidden unit,
βi=[βi1,βi2,...,βim]TIt is the output weight of i-th of Hidden unit.
In the mathematical model of trained stack extreme learning machine, the expression formula of activation primitive g (x), described i-th
The output power of the input weight wi of a Hidden unit, the biasing bi of i-th of Hidden unit and i-th of Hidden unit
Therefore known to being, by above-mentioned logical operation, output result oj can be obtained comprising develop normal and development in weight β i
Slow 2 major class.
Step 104: in a stacked the output result of the logical operation of extreme learning machine as sample to be tested development it is normal or
The foundation of person's hypoevolutism obtains the prediction result of infant development situation.
Herein it should be noted that within the same period, the process of the exercise data for same sample to be tested is obtained
In, it at least needs to obtain 5-10 group limb motion data and is just judged, which is because, if in the limbs for obtaining sample to be tested
When exercise data, the group number of the limb motion data of acquisition is very few, it is likely that due to sample to be tested since accidentalia causes
Movement is excessively fierce or movement excessively slowly causes to judge incorrectly, still, if within the same period, at least acquisition 5-10
When group limb motion data are just judged, then the misjudgment as caused by accidentalia can be reduced to the greatest extent.If obtaining
Limb motion group number it is excessive, then need to expend the more time, actually or to judging that the normal development of infant's development is slow
Slow meaning is also little.If sample to be tested by 5-10 group limb motion data judge the result is that hypoevolutism, can be again
Increase and obtain 1-3 times of limb motion group number acquisition quantity, more limb motion group numbers are to making a definite diagnosis whether infant develops late
It is slow to have the function of making a definite diagnosis.
The embodiment of the present invention one provide infant development situation prediction technique the stack limit learn and mathematical model
In known situation, the single limb motion data and corresponding monthly age data of sample to be tested are obtained, data processing can be passed through
Obtain the characteristic value of sample to be tested single limb motion, by this feature value be substituting to known stack limit study and mathematics
Model carries out logical operation, output result can be obtained, wherein output result includes hypoevolutism and develops normal 2 big
Class, wherein since the monthly age data of sample to be tested can be obtained according to the date of birth that birth certificate is shown, that is to say, that
In infant development situation prediction technique provided by the invention, data to be obtained are only single limb motion data, in this feelings
Under condition, it can be directly obtained by motion sensor, it therefore, can using infant development situation prediction technique provided by the invention
More easily infant development situation is predicted, in addition, in infant development situation prediction technique provided by the invention, heap
The stacked limit study and mathematical model during building, the baby at all monthly ages can be covered, therefore, even if to test sample
This monthly age is smaller, can also be learnt by the stack limit and mathematical model obtain relatively accurate output as a result, because
This, can provide reliable foundation for the slow early intervention of infant development.
Referring to attached drawing 3 and attached drawing 5, the construction method of the mathematical model of stack extreme learning machine the following steps are included:
Step 201: obtaining the single limb motion data and corresponding monthly age data of training sample, wherein training sample
For the sample of known development condition.
Specifically, obtaining the limb motion data of training sample here by sensor.Wherein, limb motion data
Including single motion duration, single motion average acceleration av, single motion peak accelerator am, left leg type of sports SL、
Right leg type of sports SR.Wherein, single motion duration tiIt is single motion last moment Ti1With single motion beginning moment Ti0's
Difference, that is, Ti1-Ti0.Wherein, left leg type of sports SL, right leg type of sports SRIt can be shown using the mode of verbal description, example
Such as, it can qualitatively describe to kick, squat down, walk, creeping, being bent, a symbol can also be defined to each type of sports, such as
Kicking=Q1, squat down=Q2, walking=Q3, creep=Q4, bending=Q5, then, it is described in a manner of symbol.In addition, by
When infant is especially in 12 months in 24 months, limb action ability development is more rapid, sometimes, even if differing only by number
It, the limb action of infant can have greatly changed, and therefore, the monthly age, data Y needed to be accurate to day, for example, monthly age number
According to can be 2m+1,8m+10 etc..
Step 202: carrying out data processing for the single limb motion data of training sample, obtain opposite with monthly age data
The characteristic value for training sample single limb motion answered.
Specifically, may include for the mode of the single limb motion data progress data processing of the training sample
Single argument feature selecting, recursive feature selection, the gradually modes such as feature selecting, the single limbs for the sample to be tested
Exercise data carries out data processing, obtains the feature for the sample to be tested single limb motion corresponding with monthly age data
Value specifically includes:
Single argument feature selecting is referred to by being chosen based on some univariate statistically independents to as training sample
The characteristic value X of this single limb motionj.At this point, first embodiment can be Xj={ Ti1-Ti0, m }；Second embodiment can be
Xj={ av, Y }；3rd embodiment can be Xj={ am, Y }；Fourth embodiment can be Xj={ XL, XR, Y }.Or
Recursive feature selection refers to by the way that data processing is normalized to above-mentioned each variable, using obtained data as
The characteristic value X of training sample single limb motionj.For example, X at this timej={ y1(Ti1-Ti0)+y2av+y3am, XL, XR, Y }, wherein
y1For (Ti1-Ti0) characteristic coefficient, y2For avCharacteristic coefficient, y3For amCharacteristic coefficient.In the present embodiment, the normalizing taken
Changing data processing method is linear normalization data processing, can also be directed to above-mentioned (T according to actual needsi1-Ti0)、av、amThis
Certain carry out power operations in three features can also be directed to above-mentioned (T so that increasing it influences gradei1-Ti0)、av、amThis
Certain carry out extracting operations in three features, to reduce influence grade.Or
Gradually feature selecting refers to chooses X for the first timej={ Ti1-Ti0, Y }, second of selection Xj={ av, Y }, third
Secondary selection Xj={ am, Y }, the 4th selection Xj={ XL, XR, Y }；It is then also possible to the 5th selection Xj={ y1(Ti1-Ti0)+
y2av+y3am, XL, XR, Y }.
Step 203: the characteristic value for training sample single limb motion corresponding with monthly age data is brought into patrolling
Collect operational formula:When training sample is multiple, multiple logic-based fortune are obtained
Calculate the expression formula of formula；
Wherein,
L- is the number of nodes of neural networks with single hidden layer,
G (x)-is activation primitive,
wi=[Wi1,wi2,...,win]TIt is the input weight of i-th of Hidden unit,
Bi- is the biasing of i-th of Hidden unit,
βi=[βi1,βi2,...,βim]TIt is the output weight of i-th of Hidden unit,
Oj- be for training sample, carry out sorted logical operation output using stack extreme learning machine as a result, its
In, it includes developing normal and 2 major class of hypoevolutism that logical operation, which exports result,.
Herein it should be noted that in the case of training sample, the developmental state of training sample is known, that is,
It says, in the case of training sample, oj is known.Under normal conditions, oj includes hypoevolutism or normal 2 points of development
Class,
Step 204: according to the expression formula of multiple logic-based operational formulas, determine activation primitive g (x) expression formula, i-th
The input weight wi of a Hidden unit, the biasing bi of i-th Hidden unit and the output weight beta i of i-th of Hidden unit.
In this case, it after the completion of the stack limit learns, obtains one and only needs to stack trace learning machine
Input the characteristic value X of sample to be tested single limb motionjAnd after determining the number of nodes L of neural networks with single hidden layer, it can output
The logical operation of oj.That is, in this case, it is only necessary to obtain the characteristic value X of sample to be tested single limb motionj,
It can judge the development condition of sample to be tested.
Step 205: again by the expression formula of the activation primitive g (x) after determination, the input weight wi of i-th Hidden unit,
The biasing bi of i Hidden unit and the output weight beta i band of i-th of Hidden unit are back to logical operation formula, obtain stack
The mathematical model of extreme learning machine.
Specifically, in such a case, it is possible to obtaining the mathematical model of stack extreme learning machine i.e.
Wherein,
The number of nodes of L- neural networks with single hidden layer,
G (x)-activation primitive,
wi=[Wi1,wi2,...,win]TThe input weight of-i-th Hidden unit,
The biasing of i-th of Hidden unit of bi-,
βi=[βi1,βi2,...,βim]TThe output weight of-i-th Hidden unit.
Wherein, for sample to be tested, output, can be according to the table of known activation primitive g (x) the result is that unknown
Up to formula, the output power of the input weight wi of i-th Hidden unit, the biasing bi of i-th Hidden unit and i-th of Hidden unit
The number of nodes L of weight β i and neural networks with single hidden layer, export oj by logical operation.
Wherein, single limb motion data include: the moment beginning T of single limb motioni0, last moment Ti1, average acceleration
av, peak accelerator am, left leg type of sports SL, right leg type of sports SR.Wherein, single motion duration tiIt is single motion
Last moment and single motion beginning moment Ti0Difference Ti1-Ti0。
Herein it is to be understood why choosing data of these indexs as single limb motion, consider first
Be motion sensor data availability.Then, it is also contemplated that the relevance of data and infant development situation, wherein average
Acceleration av, peak accelerator am, left leg type of sports SL, right leg type of sports SRThese indexs of z can indicate that baby's is big
Locomitivity, therefore, using them as acquisition single limb motion characteristic value XjBasic data.Wherein, the list of sample to be tested
Secondary limb motion data, the single limb motion data of training sample are measured by same motion sensor, alternatively, at least having
What the sensor of the same model of same manufacturer's production obtained.Same sensor why is selected to be at least the phase of same manufacturer's production
It is because of such feelings for selecting to can reduce the error in judgement as caused by sensor detection error with the sensor of model
Condition.
Wherein, data processing is carried out for the single limb motion data of sample to be tested, obtained corresponding with monthly age data
The characteristic value for sample to be tested single limb motion method be selected from single argument feature selecting, recursive feature selection, gradually
One of feature selecting.Wherein:
Single argument feature selecting is chosen by the statistically independent to the unitary variant in single limb motion data
To the characteristic value as single limb motion.
Specifically, in this case, since the feature of selection is unitary variant, computational efficiency is high, still,
Since the impact factor that unitary variant considers is usually less, it is also relatively large that a possibility that error occurs in judging result.
Recursive feature is selected by the way that data processing is normalized to each variable in single limb motion data, to obtain
Characteristic value of the data as single limb motion.
Specifically, in this case, since recursive feature selection has comprehensively considered unitary variant, and also by drawing
With the mode of characteristic coefficient, it is contemplated that the association between each unitary variant, therefore, can reduce that error occurs in judging result can
It can property.
Gradually feature selecting passes through the unitary variant chosen in single limb motion data one by one, to single limb motion number
The characteristic value that the data that data processing obtains successively are used as single limb motion is normalized in each variable in.
Specifically, in this case, combining unitary variant and normalization variable carrying out comprehensive descision, Neng Gougeng
It is further reduced judging result and a possibility that error occurs.
Wherein, in logical operation output result,
According to the mathematical model of stack extreme learning machine, different classifications is normally carried out to development；
According to the mathematical model of stack extreme learning machine, different classifications is carried out to hypoevolutism.
Specifically, in the case where the specific subdivision of some needs, it can also be normal for hypoevolutism and development respectively
Classification is more specifically classified, such as slow -1 grade, and slow -2 grades, slow -3 grades, wherein slow -1 grade of slow degree
Most light, slow -3 grades of slow degree is most heavy；Normally-excellent, it is normal-good, it is normal-medium.Specifically, special in above-mentioned point variable
Sign selection, recursive feature selection on the basis of gradually the judging result of feature selecting exports oj, can be combined with (Ti1-Ti0)、
av、amThe value and X of these three detection datasL, XRThe classification of the two data, to normal or slow further progress point
Grade, wherein when oj is normal, (Ti1-Ti0)、av、amThese three value datas are bigger, illustrate in normotrophic situation more
It is outstanding, at this point, can also be further directed to (Ti1-Ti0)、av、amThese three data settings are excellent, it is good, in corresponding threshold value；
XL, XRClassification it is more, illustrate it is more outstanding in normotrophic situation, at this point, can also be further directed to XL, XRClassification
Be arranged it is excellent, good, in corresponding threshold value.When oj is slow, (Ti1-Ti0)、av、amThese three value datas are smaller, explanation
It is more serious in the case where hypoevolutism, at this point, can also be further directed to (Ti1-Ti0)、av、am1 grade of these three data settings,
2 grades, 3 grades of corresponding threshold values；XL, XRClassification it is fewer, illustrate it is more serious in the case where hypoevolutism, at this point, can be with
Further directed to XL, XRClassification be arranged 1 grade, 2 grades, 3 grades of corresponding threshold values.
Embodiment two
Referring to attached drawing 4, infant development situation prediction meanss provided by Embodiment 2 of the present invention include:
Data capture unit 301, for obtaining the single limb motion data and corresponding monthly age data of sample to be tested.
Specifically, obtaining the limb motion data of sample to be tested here by sensor.Wherein, limb motion data
Including single motion duration, single motion average acceleration av, single motion peak accelerator am, left leg type of sports SL、
Right leg type of sports SR.Wherein, single motion duration tiIt is single motion last moment Ti1With single motion beginning moment Ti0's
Difference, that is, Ti1-Ti0.Wherein, left leg type of sports SL, right leg type of sports SRIt can be shown using the mode of verbal description, example
Such as, it can qualitatively describe to kick, squat down, walk, creeping, being bent, a symbol can also be defined to each type of sports, such as
Kicking=Q1, squat down=Q2, walking=Q3, creep=Q4, bending=Q5, then, it is described in a manner of symbol.In addition, by
When infant is especially in 12 months in 24 months, limb action ability development is more rapid, sometimes, even if differing only by number
It, the limb action of infant can have greatly changed, and therefore, the monthly age, data Y needed to be accurate to day, for example, monthly age number
According to can be 2m+1,8m+10 etc..
Data processing unit 302, for for sample to be tested single limb motion data carry out data processing, obtain with
The corresponding characteristic value for sample to be tested single limb motion of monthly age data.
Specifically, may include for the mode of the single limb motion data progress data processing of the sample to be tested
Single argument feature selecting, recursive feature selection, the gradually modes such as feature selecting, wherein
Single argument feature selecting is referred to by being chosen based on some univariate statistically independents to as to test sample
The characteristic value X of this single limb motionj.At this point, first embodiment can be Xj={ Ti1-Ti0, m }；Second embodiment can be
Xj={ av, Y }；3rd embodiment can be Xj={ am, Y }；Fourth embodiment can be Xj={ XL, XR, Y }；
Recursive feature selection refers to by the way that data processing is normalized to above-mentioned each variable, using obtained data as
The characteristic value X of sample to be tested single limb motionj.For example, X at this timej={ y1(Ti1-Ti0)+y2av+y3am, XL, XR, Y }, wherein
y1For (Ti1-Ti0) characteristic coefficient, y2For avCharacteristic coefficient, y3For amCharacteristic coefficient.In the present embodiment, the normalizing taken
Changing data processing method is linear normalization data processing, can also be directed to above-mentioned (T according to actual needsi1-Ti0)、av、amThis
Certain carry out power operations in three features can also be directed to above-mentioned (T so that increasing it influences gradei1-Ti0)、av、amThis
Certain carry out extracting operations in three features, to reduce influence grade.
Gradually feature selecting refers to chooses X for the first timej={ Ti1-Ti0, Y }, second of selection Xj={ av, Y }, third
Secondary selection Xj={ am, Y }, the 4th selection Xj={ XL, XR, Y }；It is then also possible to the 5th selection Xj={ y1(Ti1-Ti0)+
y2av+y3am, XL, XR, Y }.
Arithmetic element 303, for by treated the limb motion data corresponding with sample to be tested of pre-determined number and
Corresponding monthly age data are brought into the mathematical model of stack extreme learning machine, and logical operation is carried out.
Specifically, by the characteristic value X of above-mentioned sample to be tested single limb motionjIt is substituting to stack extreme learning machine
Mathematical model is
Wherein,
L- is the number of nodes of neural networks with single hidden layer,
G (x)-is activation primitive,
wi=[Wi1,wi2,...,win]TIt is the input weight of i-th of Hidden unit,
Bi- is the biasing of i-th of Hidden unit,
βi=[βi1,βi2,...,βim]TIt is the output weight of i-th of Hidden unit.
In the mathematical model of trained stack extreme learning machine, the expression formula of activation primitive g (x), described i-th
The output power of the input weight wi of a Hidden unit, the biasing bi of i-th of Hidden unit and i-th of Hidden unit
Therefore known to being, by above-mentioned logical operation, output result oj can be obtained comprising develop normal and development in weight β i
Slow 2 major class.
Prediction result output unit 304, the output result of the logical operation for extreme learning machine in a stacked be used as to
Test sample this development is normal or the foundation of hypoevolutism, obtains the prediction result of infant development situation.
Herein it should be noted that within the same period, the process of the exercise data for same sample to be tested is obtained
In, it at least needs to obtain 5-10 group limb motion data and is just judged, which is because, if in the limbs for obtaining sample to be tested
When exercise data, the group number of the limb motion data of acquisition is very few, it is likely that due to sample to be tested since accidentalia causes
Movement is excessively fierce or movement excessively slowly causes to judge incorrectly, still, if within the same period, at least acquisition 5-10
When group limb motion data are just judged, then the misjudgment as caused by accidentalia can be reduced to the greatest extent.If obtaining
Limb motion group number it is excessive, then need to expend the more time, actually or to judging that the normal development of infant's development is slow
Slow meaning is also little.If sample to be tested by 5-10 group limb motion data judge the result is that hypoevolutism, can be again
Increase and obtain 1-3 times of limb motion group number acquisition quantity, more limb motion group numbers are to making a definite diagnosis whether infant develops late
It is slow to have the function of making a definite diagnosis.
Infant development situation prediction meanss provided by Embodiment 2 of the present invention the stack limit learn and mathematical model
In known situation, the single limb motion data and corresponding monthly age data of sample to be tested are obtained, data processing can be passed through
Obtain the characteristic value of sample to be tested single limb motion, by this feature value be substituting to known stack limit study and mathematics
Model carries out logical operation, output result can be obtained, wherein output result includes hypoevolutism and develops normal 2 big
Class, wherein since the monthly age data of sample to be tested can be obtained according to the date of birth that birth certificate is shown, that is to say, that
In infant development situation prediction technique provided by the invention, data to be obtained are only single limb motion data, in this feelings
Under condition, it can be directly obtained by motion sensor, it therefore, can using infant development situation prediction technique provided by the invention
More easily infant development situation is predicted, in addition, in infant development situation prediction technique provided by the invention, heap
The stacked limit study and mathematical model during building, the baby at all monthly ages can be covered, therefore, even if to test sample
This monthly age is smaller, can also be learnt by the stack limit and mathematical model obtain relatively accurate output as a result, because
This, can provide reliable foundation for the slow early intervention of infant development.
Embodiment three
Infant development situation Prediction program, infant development situation are stored on the storage medium that the embodiment of the present invention three provides
The step of infant development situation prediction technique provided by the invention is realized when Prediction program is executed by processor.
The storage medium that the embodiment of the present invention three provides the stack limit learn and mathematical model known in situation,
The single limb motion data and corresponding monthly age data for obtaining sample to be tested, can obtain sample to be tested list by data processing
The characteristic value of secondary limb motion, by this feature value be substituting to known stack limit study and mathematical model, carry out logic
Output result can be obtained in operation, wherein output result includes hypoevolutism and develops normal 2 major class, wherein due to
The monthly age data of test sample sheet can be obtained according to the date of birth that birth certificate is shown, that is to say, that in baby provided by the invention
In youngster's developmental state prediction technique, data to be obtained are only single limb motion data, in such a case, it is possible to pass through fortune
Dynamic sensor directly obtains, therefore, can be more easily to baby using infant development situation prediction technique provided by the invention
Youngster's development condition predicts, in addition, in infant development situation prediction technique provided by the invention, the study of the stack limit and
Mathematical model during building, the baby at all monthly ages can be covered, therefore, even if the monthly age of sample to be tested is smaller,
Can also by the stack limit learn and mathematical model obtain relatively accurate output as a result, can be baby therefore
The early intervention of youngster's hypoevolutism provides reliable foundation.
Example IV
The electronic equipment that the embodiment of the present invention four provides includes motion sensor, processor, memory and is stored in storage
On device and the infant development situation Prediction program that can run on a processor, wherein
Motion sensor, for obtaining the single limb motion data of sample to be tested；
The infant development feelings that the embodiment of the present invention one provides are realized when infant development situation Prediction program is executed by processor
The step of condition prediction technique.
The electronic equipment that the embodiment of the present invention four provides the stack limit learn and mathematical model known in situation,
The single limb motion data and corresponding monthly age data for obtaining sample to be tested, can obtain sample to be tested list by data processing
The characteristic value of secondary limb motion, by this feature value be substituting to known stack limit study and mathematical model, carry out logic
Output result can be obtained in operation, wherein output result includes hypoevolutism and develops normal 2 major class, wherein due to
The monthly age data of test sample sheet can be obtained according to the date of birth that birth certificate is shown, that is to say, that in baby provided by the invention
In youngster's developmental state prediction technique, data to be obtained are only single limb motion data, in such a case, it is possible to pass through fortune
Dynamic sensor directly obtains, therefore, can be more easily to baby using infant development situation prediction technique provided by the invention
Youngster's development condition predicts, in addition, in infant development situation prediction technique provided by the invention, the study of the stack limit and
Mathematical model during building, the baby at all monthly ages can be covered, therefore, even if the monthly age of sample to be tested is smaller,
Can also by the stack limit learn and mathematical model obtain relatively accurate output as a result, can be baby therefore
The early intervention of youngster's hypoevolutism provides reliable foundation.
Although preferred embodiments of the present invention have been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the scope of the invention.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (8)
1. a kind of infant development situation prediction technique, which comprises the following steps:
Obtain the single limb motion data and corresponding monthly age data of sample to be tested；
Data processing is carried out for the single limb motion data of the sample to be tested, obtains be directed to corresponding with monthly age data
The characteristic value of the sample to be tested single limb motion；
By the characteristic value of the single limb motion corresponding with the sample to be tested of pre-determined number and corresponding monthly age data band
Enter to the mathematical model of stack extreme learning machine, carries out logical operation；
Normal or hair is developed using the output result of the logical operation of the stack extreme learning machine as the sample to be tested
Slow foundation is educated, the prediction result of the infant development situation is obtained.
2. infant development situation prediction technique according to claim 1, which is characterized in that the stack extreme learning machine
Mathematical model construction method the following steps are included:
Obtain the single limb motion data and corresponding monthly age data of training sample, wherein the training sample is known hair
Educate the sample of situation；
Data processing is carried out for the single limb motion data of the training sample, obtains be directed to corresponding with monthly age data
The characteristic value of the training sample single limb motion；
The characteristic value for the training sample single limb motion corresponding with monthly age data is brought into logic and is transported
Calculate formula:When the training sample is multiple, obtain patrolling described in multiple be based on
Collect the expression formula of operational formula；
Wherein,
L is the number of nodes of neural networks with single hidden layer,
G (x) is activation primitive,
wi=[Wi1,wi2,...,win]TFor the input weight of i-th of Hidden unit,
Bi is the biasing of i-th of Hidden unit,
βi=[βi1,βi2,...,βim]TFor the output weight of i-th of Hidden unit,
Oj is to carry out sorted logical operation for training sample using stack extreme learning machine and export result, wherein institute
Stating logical operation output result includes developing normal and 2 major class of hypoevolutism；
According to the multiple expression formula based on the logical operation formula, expression formula, the institute of the activation primitive g (x) are determined
State the input weight wi of i-th of Hidden unit, the biasing bi of i-th of Hidden unit and i-th of Hidden unit
Export weight beta i；
Again by the expression formula of the activation primitive g (x) after determination, the input weight wi, i-th described of i-th of Hidden unit
The biasing bi of Hidden unit and the output weight beta i band of i-th of Hidden unit are back to the logical operation formula, obtain
The mathematical model of the stack extreme learning machine.
3. infant development situation prediction technique according to claim 1 or 2, which is characterized in that the single limb motion
Data include:
The moment beginning T of single limb motioni0, last moment Ti1, average acceleration av, peak accelerator am, left leg type of sports SL、
Right leg type of sports SR；
Wherein, single motion duration tiIt is single motion last moment and single motion beginning moment Ti0Difference Ti1-Ti0。
4. infant development situation prediction technique according to claim 1 or 2, which is characterized in that be directed to the sample to be tested
Single limb motion data carry out data processing, obtain corresponding with monthly age data for the sample to be tested single limbs
The method of motion characteristics value is selected from single argument feature selecting, recursive feature selection, gradually one of feature selecting, the needle
Data processing is carried out to the single limb motion data of the sample to be tested, obtain it is corresponding with monthly age data for it is described to
The characteristic value of test sample this single limb motion specifically includes:
The single argument feature selecting passes through the statistically independent to the unitary variant in the single limb motion data, choosing
Obtain the characteristic value as the single limb motion；Or
Recursive feature selection by the way that data processing is normalized to each variable in the single limb motion data, with
Characteristic value of the obtained data as the single limb motion；Or
The gradually feature selecting passes through the unitary variant chosen in the single limb motion data one by one, to the single limb
The characteristic value that the data that data processing obtains successively are used as single limb motion is normalized in each variable in body exercise data.
5. infant development situation prediction technique according to claim 2, which is characterized in that export and tie in the logical operation
In fruit,
According to the mathematical model of the stack extreme learning machine, different classifications is normally carried out to development；
According to the mathematical model of the stack extreme learning machine, different classifications is carried out to hypoevolutism.
6. a kind of infant development situation prediction meanss characterized by comprising
Data capture unit, for obtaining the single limb motion data and corresponding monthly age data of sample to be tested；
Data processing unit carries out data processing for the single limb motion data for the sample to be tested, obtains and the moon
The corresponding characteristic value for the sample to be tested single limb motion of age data；
Arithmetic element, for by corresponding treated limb motion data of pre-determined number and the sample to be tested and corresponding
Monthly age data bring into the mathematical model of stack extreme learning machine, carry out logical operation；
Prediction result output unit, for using the output result of the logical operation of the stack extreme learning machine as it is described to
Test sample this development is normal or the foundation of hypoevolutism, obtains the prediction result of the infant development situation.
7. a kind of storage medium, which is characterized in that be stored with infant development situation Prediction program, the baby on the storage medium
Any infant development situation prediction in Claims 1 to 5 is realized when youngster's developmental state Prediction program is executed by processor
The step of method.
8. a kind of electronic equipment, which is characterized in that including motion sensor, processor, memory and be stored in the memory
Infant development situation Prediction program that is upper and can running on the processor, wherein
The motion sensor, for obtaining the single limb motion data of sample to be tested；
Any baby in Claims 1 to 5 is realized when the infant development situation Prediction program is executed by the processor
The step of youngster's developmental state prediction technique.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910346364.8A CN110236558A (en) | 2019-04-26 | 2019-04-26 | Infant development situation prediction technique, device, storage medium and electronic equipment |
| SG11202008420UA SG11202008420UA (en) | 2019-04-26 | 2019-08-28 | Machine learning based method and device for infant development prediction, storage medium, and electronic device |
| PCT/CN2019/103148 WO2020215566A1 (en) | 2019-04-26 | 2019-08-28 | Machine learning-based infant developmental condition predicting method and apparatus, storage medium, and electronic device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910346364.8A CN110236558A (en) | 2019-04-26 | 2019-04-26 | Infant development situation prediction technique, device, storage medium and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110236558A true CN110236558A (en) | 2019-09-17 |
Family
ID=67883539
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910346364.8A Pending CN110236558A (en) | 2019-04-26 | 2019-04-26 | Infant development situation prediction technique, device, storage medium and electronic equipment |
Country Status (3)
| Country | Link |
|---|---|
| CN (1) | CN110236558A (en) |
| SG (1) | SG11202008420UA (en) |
| WO (1) | WO2020215566A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114469072A (en) * | 2021-12-08 | 2022-05-13 | 四川大学华西第二医院 | Method for automatically predicting baby psychological development by using camera |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070016109A1 (en) * | 2003-10-31 | 2007-01-18 | Tokyo University Of Agriculture And Technology Tlo | Infant movement analysis system and infant movement analysis method |
| US20120059283A1 (en) * | 2010-08-26 | 2012-03-08 | University Of California | System for evaluating infant movement |
| US20140066780A1 (en) * | 2010-08-26 | 2014-03-06 | University Of California | System for evaluating infant movement using gesture recognition |
| CN105310695A (en) * | 2015-11-03 | 2016-02-10 | 苏州景昱医疗器械有限公司 | Dyskinesia assessment equipment |
| CN106388831A (en) * | 2016-11-04 | 2017-02-15 | 郑州航空工业管理学院 | Method for detecting falling actions based on sample weighting algorithm |
| CN107851356A (en) * | 2015-04-05 | 2018-03-27 | 斯米拉布莱斯有限公司 | Determine the posture of infant and wearable the infant's monitoring device and system of motion |
| CN109009143A (en) * | 2018-07-12 | 2018-12-18 | 杭州电子科技大学 | A method of ecg information is predicted by body gait |
| CN109620244A (en) * | 2018-12-07 | 2019-04-16 | 吉林大学 | The Infants With Abnormal behavioral value method of confrontation network and SVM is generated based on condition |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| ES2891327T3 (en) * | 2012-08-25 | 2022-01-27 | Owlet Baby Care Inc | Wireless Child Health Monitor |
| US9277870B2 (en) * | 2013-09-12 | 2016-03-08 | Sproutling, Inc. | Infant monitoring system and methods |
| CN104765959A (en) * | 2015-03-30 | 2015-07-08 | 燕山大学 | Computer vision based evaluation method for general movement of baby |
| CN107924643B (en) * | 2015-04-05 | 2021-05-18 | 斯米拉布莱斯有限公司 | Infant development analysis method and system |
| WO2017027709A1 (en) * | 2015-08-11 | 2017-02-16 | Cognoa, Inc. | Methods and apparatus to determine developmental progress with artificial intelligence and user input |
-
2019
- 2019-04-26 CN CN201910346364.8A patent/CN110236558A/en active Pending
- 2019-08-28 WO PCT/CN2019/103148 patent/WO2020215566A1/en active Application Filing
- 2019-08-28 SG SG11202008420UA patent/SG11202008420UA/en unknown
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070016109A1 (en) * | 2003-10-31 | 2007-01-18 | Tokyo University Of Agriculture And Technology Tlo | Infant movement analysis system and infant movement analysis method |
| US20120059283A1 (en) * | 2010-08-26 | 2012-03-08 | University Of California | System for evaluating infant movement |
| US20140066780A1 (en) * | 2010-08-26 | 2014-03-06 | University Of California | System for evaluating infant movement using gesture recognition |
| CN107851356A (en) * | 2015-04-05 | 2018-03-27 | 斯米拉布莱斯有限公司 | Determine the posture of infant and wearable the infant's monitoring device and system of motion |
| CN105310695A (en) * | 2015-11-03 | 2016-02-10 | 苏州景昱医疗器械有限公司 | Dyskinesia assessment equipment |
| CN106388831A (en) * | 2016-11-04 | 2017-02-15 | 郑州航空工业管理学院 | Method for detecting falling actions based on sample weighting algorithm |
| CN109009143A (en) * | 2018-07-12 | 2018-12-18 | 杭州电子科技大学 | A method of ecg information is predicted by body gait |
| CN109620244A (en) * | 2018-12-07 | 2019-04-16 | 吉林大学 | The Infants With Abnormal behavioral value method of confrontation network and SVM is generated based on condition |
Non-Patent Citations (6)
| Title |
|---|
| DAVID GOODFELLOW 等: "《Predicting infant motor development status using day long movement data from wearable sensors》", 《ARXIV.ORG》 * |
| DAVID GOODFELLOW 等: "《Predicting infant motor development status using day long movement data from wearable sensors》", 《ARXIV.ORG》, 14 October 2018 (2018-10-14), pages 1 - 5 * |
| 崔金铎: "《基于近红外光谱的堆叠极限学习机算法及其应用研究》", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 崔金铎: "《基于近红外光谱的堆叠极限学习机算法及其应用研究》", 《中国优秀硕士学位论文全文数据库信息科技辑》, 15 January 2019 (2019-01-15), pages 8 - 19 * |
| 邵肖梅: "《胎儿和新生儿脑损伤 第2版》", 31 October 2017 * |
| 马良: "《扭动运动阶段全身运动细化评估信度效度研究以及在高危儿随访中的应用》", 《中国博士学位论文全文数据库医药卫生科技辑》, 15 February 2019 (2019-02-15) * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114469072A (en) * | 2021-12-08 | 2022-05-13 | 四川大学华西第二医院 | Method for automatically predicting baby psychological development by using camera |
| CN114469072B (en) * | 2021-12-08 | 2023-08-08 | 四川大学华西第二医院 | Method for automatically predicting psychological development of infants by using camera |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020215566A1 (en) | 2020-10-29 |
| SG11202008420UA (en) | 2020-11-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | A comprehensive active learning method for multiclass imbalanced data streams with concept drift | |
| Minku et al. | The impact of diversity on online ensemble learning in the presence of concept drift | |
| US10706332B2 (en) | Analog circuit fault mode classification method | |
| CN103632168B (en) | Classifier integration method for machine learning | |
| US11200461B2 (en) | Methods and arrangements to identify feature contributions to erroneous predictions | |
| CN108228684A (en) | Training method, device, electronic equipment and the computer storage media of Clustering Model | |
| CN110236558A (en) | Infant development situation prediction technique, device, storage medium and electronic equipment | |
| US20210216845A1 (en) | Synthetic clickstream testing using a neural network | |
| Waegeman et al. | On the scalability of ordered multi-class ROC analysis | |
| CN106776757A (en) | User completes the indicating means and device of Net silver operation | |
| Messner | From black box to clear box: A hypothesis testing framework for scalar regression problems using deep artificial neural networks | |
| CN114386688A (en) | User intention prediction method and system based on multi-data fusion | |
| Wu et al. | Calibrate-Extrapolate: Rethinking Prevalence Estimation with Black Box Classifiers | |
| Nouwou Mindom et al. | A comparison of reinforcement learning frameworks for software testing tasks | |
| Juselius | Time to reject the privileging of economic theory over empirical evidence? A Reply to Lawson (2009) | |
| Fontes et al. | Automated support for unit test generation: a tutorial book chapter | |
| US20230419104A1 (en) | High dimensional dense tensor representation for log data | |
| Huang et al. | Machine learning and its applications in test | |
| Zhang et al. | An improved Conv-LSTM method for gear fault detection | |
| CN116413587B (en) | Method and device for selecting rollback path | |
| Yin et al. | Fault diagnosis method based on CWGAN-GP-1DCNN | |
| EP4287198A1 (en) | Method and system for determining which stage a user performance belongs to | |
| CN116052907A (en) | Inquiry method and device and electronic equipment | |
| US20230113750A1 (en) | Reinforcement learning based group testing | |
| Zhang | P&A: Make Wordle Game Better |
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 |