CN104346369B - A kind of method for establishing heartbeat shock wave form feature database - Google Patents
A kind of method for establishing heartbeat shock wave form feature database Download PDFInfo
- Publication number
- CN104346369B CN104346369B CN201310327012.0A CN201310327012A CN104346369B CN 104346369 B CN104346369 B CN 104346369B CN 201310327012 A CN201310327012 A CN 201310327012A CN 104346369 B CN104346369 B CN 104346369B
- Authority
- CN
- China
- Prior art keywords
- heartbeat
- vector
- shock wave
- wave form
- sample
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
- G06F2218/10—Feature extraction by analysing the shape of a waveform, e.g. extracting parameters relating to peaks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Signal Processing (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The present invention provides a kind of method for establishing heartbeat shock wave form feature database, including：Select and add master sample vector；New feature vector is added by way of machine learning；Correct edge data.Preferably, the step for selecting and adding master sample vector includes：By measuring bed vibrations collection people's heartbeat shockwave information；Sample waveform is selected from the heartbeat shockwave information collected；Characteristic point is selected for the sample waveform；The sample waveform is converted into sample vector；The sample vector is added to the heartbeat shock wave form feature database.As described above, the method for establishing heartbeat shock wave form feature database of the present invention, has the advantages that：Establish heartbeat shock wave form feature database by way of machine learning, efficiency high, accuracy is good.And according to preferable mode, the sample vector and characteristic vector of waveform are simple, and amount of calculation is small.
Description
Technical field
The present invention relates to a kind of method for establishing heartbeat shock wave form feature database, and engineering is utilized more particularly to one kind
The method that the mode of habit establishes heartbeat shock wave form feature database.
Background technology
Medical science, health etc. numerous areas widespread need measurement human heartbeat's feature, one kind measurement human heartbeat's feature
Mode be people lie on a bed it is static after, with shock sensor collection human heart beat when to vibration data caused by bed so as to
To human heartbeat's feature（Call heartbeat shock wave morphological feature in the following text）Although this mode is convenient and small on testee's influence,
The data of collection can be because affected by various factors and accuracy is poor.The factor of influence includes but is not limited to：People lies in
Posture, around bed someone on bed are passed by etc..So heartbeat shock wave characteristic pattern data needs gathered in this way
By screening or correcting.This heartbeat shock wave characteristic pattern data obtained by measuring bed to shake is sieved at present
Then the mode of choosing or amendment will measure obtained heartbeat shock wave form spy first to establish heartbeat shock wave form feature database
Data are levied compared with the data in storehouse, if the gap for measuring data in obtained data and storehouse is more than pre-determined threshold,
Give up the data.The core of this method is the heartbeat shock wave form feature database of high quality to be established.But establish at present this kind of
The method efficiency of heartbeat shock wave form feature database is low, accuracy is poor and computationally intensive.
The content of the invention
In view of the above the shortcomings that prior art, it is an object of the invention to provide one kind to establish heartbeat shock wave form
The method of feature database, for solving, prior art foundation jump shock wave form feature database efficiency is low, accuracy is poor and computationally intensive
The problem of.
In order to achieve the above objects and other related objects, the present invention provides a kind of heartbeat shock wave form feature database established
Method, including：Select and add master sample vector；New feature vector is added by way of machine learning；Correct number of edges
According to.
Preferably, the step for selecting and adding master sample vector includes：By measuring the bed vibrations collection popular feeling
Jump shockwave information；Sample waveform is selected from the heartbeat shockwave information collected；Selected for the sample waveform special
Sign point；The sample waveform is converted into sample vector；The sample vector is added to the heartbeat shock wave morphological feature
Storehouse.
As described above, the method for establishing heartbeat shock wave form feature database of the present invention, has the advantages that：Pass through
The mode of machine learning establishes heartbeat shock wave form feature database, efficiency high, and accuracy is good.And according to preferable mode, ripple
The sample vector and characteristic vector of shape are simple, and amount of calculation is small.
Brief description of the drawings
Fig. 1 is shown as the flow chart for establishing heartbeat shock wave form feature database method of the present invention；
Fig. 2 is shown as the bed shocking waveshape figure that the static prostrate of the measured measures when on bed；
Fig. 3 is shown as the spy generated according to one embodiment for establishing heartbeat shock wave form feature database method of the present invention
DUAL PROBLEMS OF VECTOR MAPPING is levied to the design sketch of two-dimensional space.
Embodiment
Illustrate embodiments of the present invention below by way of specific instantiation, those skilled in the art can be by this specification
Disclosed content understands other advantages and effect of the present invention easily.The present invention can also pass through specific realities different in addition
The mode of applying is embodied or practiced, the various details in this specification can also be based on different viewpoints with application, without departing from
Various modifications or alterations are carried out under the spirit of the present invention.
Refer to accompanying drawing.It should be noted that only explanation is of the invention in a schematic way for the diagram provided in the present embodiment
Basic conception, then in schema only display with relevant component in the present invention rather than according to component count during actual implement, shape
Shape and size are drawn, and kenel, quantity and the ratio of each component can be a kind of random change during its actual implementation, and its component cloth
Office's kenel may also be increasingly complex.
Referring next to Fig. 1, Fig. 1 show the flow chart that the present invention establishes the method for heartbeat shock wave form feature database.This
Invention employs the method for machine learning to establish heartbeat shock wave form feature database, can substantially be divided into three steps, with step S1
~S3 is represented, each step is described in detail below.
Step S1 represents to select and add master sample.Selection or addition master sample can be manually performed, can also
It is the data directly obtained using other approach.In this example, by the way of artificial selection and adding, specific mode is：Allow
People to be measured is tranquil to lie on a bed, and the heartbeat shock wave form signals that record sensor collection arrives, analysis obtains heartbeat each time
Waveform, this waveform is changed into sample vector, one or more groups of sample vectors of selection standard the most are as master sample.This
Just complete the input for the heartbeat shock wave form master sample that lies low.
Illustrate to obtain the specific method of sample vector by heartbeat surge waveform in this example referring next to Fig. 2-3, for the side of explanation
Just, only lift here exemplified by lying down.Fig. 2 displays behave static prostrate when on bed, the bed vibrations number that shock sensor measures
According to including heartbeat surge waveform in the data.The characteristic point in waveform is selected first.The maximum point in waveform is chosen in this example
J and minimum K are as characteristic point.Then characteristic point is converted into sample vector, can be with before characteristic feature point during conversion
The amount of position or mathematical relationship constructs sample vector.This example only for convenience of explanation, the transformation rule used for：Take between a=J and K
Distance, the slope of b=J and 2 lines of K, then bivector（a,b）As sample vector.During actual use, in order that result
More confidence level, hyperspace is can use, such as：C=J and the abscissa quadratic sum ... of K abscissas variance, d=J and K, then can obtain
It is designated as to multidimensional sample vector（a,b,c,d,……）.
Similarly, according to the above method, can complete to sit up, the input for the sample that left/right is lain on one's side, for the purpose of simplifying the description, this example
In, only input is lieed down, sat up, lying on the left side and the sample of the sleeping four kinds of actions in right side inputs, and is each acted and only established a sample
This input.Those skilled in the art can also analyze other actions according to same method as needed, for example go to bed, leave the bed,
More sample inputs can also be established to each action.It is more convenient to increase data accuracy and data being utilized from now on,
In this example, by the identity information of measured personnel and action classification information（I.e. the shocking waveshape be measure people prostrate, sit up,
Which action in crouching with right side or go to bed or leave the bed of lying on the left side obtains）Also heartbeat shock wave form feature database is added to together.This
Sample, when utilizing the heartbeat shock wave form feature database if desired from now on, for example, it is necessary to screen the heartbeat of some action of the personnel
During data, it is possible to the heartbeat shock wave form character separation of the action classification of personnel when establishing this feature storehouse is come out,
Further increase screening accuracy.In Fig. 3, the point of shade represents lie down, sit up, lying on the left side respectively from left to right, from top to bottom
Mapping of the sample vector to be crouched with right side under four kinds of action classifications in two-dimensional space.
It should be noted that according to being actually needed, in the above method, those skilled in the art can use other manner to select
Characteristic point is selected, or waveform is converted into by sample vector using other manner.Such as, but not limited to：The waveform of corresponding diagram 2, only chooses more
More or less characteristic point, accordingly using other polynomial functions generation sample vector；Or become using Fourier transformation, small echo
The mode such as change and waveform is changed into characteristic vector.
Because different people lies on a bed, caused heartbeat shock wave form will not be identical, on the other hand, can be each
Individual builds a Sample Storehouse, and everyone Sample Storehouse is then concluded into whole big heartbeat shock wave form feature database again
In.So as to more accurately recognize the heartbeat shock wave form of different people.
Step S2 represents to add new feature vector by way of machine learning.Machine learning（Machine Learning）
It is exactly to simulate or realize the learning behavior of the mankind using concluding, by the way of synthesis using computer, reorganize existing knowledge
Structure is allowed to constantly improve the performance of itself.During machine learning, for the waveform newly collected, using calculating itself and sample ripple
Shape is also translated into characteristic vector.After obtaining the characteristic vector of the waveform, by the two-dimensional space shown in the DUAL PROBLEMS OF VECTOR MAPPING to Fig. 3,
Then check that new feature vector is nearest from the sample vector in which block region, then it is assumed that be caused by the action representated by the region
Heart shock wave form signals.Then the similarity of this feature vector and area sample vector is calculated again, if similarity is higher than
Some threshold value, then this feature vector is added into heartbeat shock wave form feature database.
In this example, for simplify illustrate, only to lie down, sit up, lie on the left side and right side crouch four kinds of actions respectively establish a sample
This input.For the heartbeat shock wave waveform newly collected, characteristic point will be used as by point corresponding with J, K on sample waveform thereon,
And the heartbeat shock wave waveform newly collected is converted into characteristic vector according to mode same in step S1.Judging characteristic vector
During with the similarity of sample vector, this example calculates the distance of characteristic vector and sample orientation by the way of vector distance（Such as sample
This vector is multiple, then can calculate average distance, characteristic vector and the sample vector positioned at center or the sample positioned at edge
Distance of vector etc.）If vector distance is less than certain threshold value, represent that similarity is high, then it is assumed that add new feature vector
Heartbeat shock wave form feature database.The like, it can not stop to grow the sample inside feature database in a manner of by machine learning
This.Preferably, in this example, when new feature vector is added by way of machine learning, also by measured's identity information and action
Classification information is added in heartbeat shock wave form feature database.
It is highly preferred that in step s 2, machine can be set differently to learn in the different periods, such as,
In 0~t1Period, the distance of new feature vector and that sample vector in its region is only calculated, if distance is less than
Threshold T1, then new feature vector is added into heartbeat shock wave form feature database；And in t1~t2Period, then by 0~t1When
Between the characteristic vector that newly adds in section be also considered as sample vector, t1~t2In period freshly harvested characteristic vector need to calculate with
The average distance of all sample vectors in its region, if the average distance is less than thresholding T2, then by new feature vector
Heartbeat shock wave form feature database is added, by that analogy.According to the present embodiment, after machine learning a period of time, heartbeat impact
The mapping effect of characteristic vector and sample vector in two-dimensional space in wave morphology characteristic storehouse is as shown in Figure 3.
It is highly preferred that in step 2, SVMs can be used（Support Vector Machine, abbreviation SVM）Learn
Habit mode：Weight is set to each latitude value of sample vector first；Then using some sample vector data tests and adjust
The weight of each vector；Finally judge which region new feature vector should belong to using SVM（For example lie down, sit up, lie on the left side
Any class vibration data in being crouched with right side）Or whether should be rejected.
Step S3 represents amendment edge data.Here edge data, refer to that some signal is in four regions shown in Fig. 3
Middle region and the data of region intersection.Because some ambiguous heart surge waveforms occur unavoidably in the machine learning stage
State signal, it is possible to belong to heart shock wave form signals when lying low, heart shock wave form letter during sitting may be belonged to again
Number, that just can not be correctly divided into correct Sample Storehouse in machine-learning process, then needed in the light of actual conditions to this
A little data are modified.Amendment can use various ways, can directly give up, or using threshold discrimination, can also be artificial
Identification etc..In this example, by the way of manual identified, the composition that manually participates in may be needed relatively more in early stage, with rear
Phase Sample Storehouse increases, and judgement is more and more accurate, and the composition that people participates in will be reduced rapidly.
By step S1~S3, that is, complete the foundation of heartbeat shock wave form feature database.
In summary, the present invention establishes heartbeat shock wave form feature database by way of machine learning, efficiency high, accurately
Property is good.And according to preferable mode, the sample vector and characteristic vector of waveform are simple, and amount of calculation is small.So the present invention is effectively
Overcome various shortcoming of the prior art and have high industrial utilization.
The above-described embodiments merely illustrate the principles and effects of the present invention, not for the limitation present invention.It is any ripe
Know the personage of this technology all can carry out modifications and changes under the spirit and scope without prejudice to the present invention to above-described embodiment.Cause
This, those of ordinary skill in the art is complete without departing from disclosed spirit and institute under technological thought such as
Into all equivalent modifications or change, should by the present invention claim be covered.
Claims (8)
- A kind of 1. method for establishing heartbeat shock wave form feature database, it is characterised in that including：Select and add master sample vector, including：By measuring bed vibrations collection people's heartbeat shockwave information；Adopted from described Sample waveform is selected in the heartbeat shockwave information collected；Characteristic point is selected for the sample waveform；According between the characteristic point Position relationship or mathematical relationship generation sample vector；The sample vector is added to the heartbeat shock wave morphological feature Storehouse；New feature vector is added by way of machine learning, including：The heartbeat surge waveform newly collected conversion is characterized Vector；The similarity of the characteristic vector and the sample vector is judged, if the similarity is higher than pre-determined threshold, by institute State characteristic vector and be added to the heartbeat shock wave form feature database；Otherwise the characteristic vector is given up；Correct edge data.
- 2. the method according to claim 1 for establishing heartbeat shock wave form feature database, it is characterised in that described to pass through machine The mode of device study is added in the step of new feature vector, and new feature vector is added using SVM.
- 3. the method according to claim 1 for establishing heartbeat shock wave form feature database, it is characterised in that according to the spy The step of position relationship or mathematical relationship between sign point generate sample vector includes：The first dimension values using the distance between described characteristic point as the sample vector；The second dimension values using the slope of the characteristic point line as the sample vector.
- 4. the method according to claim 1 for establishing heartbeat shock wave form feature database, it is characterised in that the similarity Represented with the distance of the characteristic vector and the sample vector.
- 5. the method according to claim 3 for establishing heartbeat shock wave form feature database, it is characterised in that the sample to Measure as the multi-C vector more than two dimension；The third dimension angle value of the sample vector is the characteristic point horizontal stroke/ordinate variance, or Person, the third dimension angle value of the sample vector is the characteristic point horizontal stroke/ordinate quadratic sum.
- 6. the method according to claim 4 for establishing heartbeat shock wave form feature database, it is characterised in that the selection is simultaneously In the step for adding master sample vector, the sample vector comprises at least two.
- 7. the method according to claim 1 for establishing heartbeat shock wave form feature database, it is characterised in that described to pass through survey In the step of measuring bed vibrations collection people's heartbeat shockwave information, including the static prostrate of collection people in bed, static sit up in bed It is upper, static lie on the left side in bed, static right side it is sleeping in bed when heartbeat shockwave information and people bed when getting in and out of bed vibrations Information.
- 8. the method for establishing heartbeat shock wave form feature database according to any one of claim 1-7, its feature exist In, it is described to select and add in the step for adding new feature vector described in master sample vector sum by way of machine learning, Also include the identity information of measured and measured's action classification information being added in the heartbeat shock wave form feature database.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310327012.0A CN104346369B (en) | 2013-07-30 | 2013-07-30 | A kind of method for establishing heartbeat shock wave form feature database |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310327012.0A CN104346369B (en) | 2013-07-30 | 2013-07-30 | A kind of method for establishing heartbeat shock wave form feature database |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104346369A CN104346369A (en) | 2015-02-11 |
| CN104346369B true CN104346369B (en) | 2018-03-23 |
Family
ID=52501992
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310327012.0A Active CN104346369B (en) | 2013-07-30 | 2013-07-30 | A kind of method for establishing heartbeat shock wave form feature database |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104346369B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108542381B (en) * | 2018-04-11 | 2021-08-31 | 京东方科技集团股份有限公司 | Data processing method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101216884A (en) * | 2007-12-29 | 2008-07-09 | 北京中星微电子有限公司 | A method and system for face authentication |
| CN102397064A (en) * | 2011-12-14 | 2012-04-04 | 中国航天员科研训练中心 | Continuous blood pressure measuring device |
| CN102929872A (en) * | 2011-08-08 | 2013-02-13 | 阿里巴巴集团控股有限公司 | Computer-implemented message filtering method, message filtering device and system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7644361B2 (en) * | 2002-12-23 | 2010-01-05 | Canon Kabushiki Kaisha | Method of using recommendations to visually create new views of data across heterogeneous sources |
-
2013
- 2013-07-30 CN CN201310327012.0A patent/CN104346369B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101216884A (en) * | 2007-12-29 | 2008-07-09 | 北京中星微电子有限公司 | A method and system for face authentication |
| CN102929872A (en) * | 2011-08-08 | 2013-02-13 | 阿里巴巴集团控股有限公司 | Computer-implemented message filtering method, message filtering device and system |
| CN102397064A (en) * | 2011-12-14 | 2012-04-04 | 中国航天员科研训练中心 | Continuous blood pressure measuring device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104346369A (en) | 2015-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7160932B2 (en) | Generating prescriptive analytics using motion identification and motion information | |
| CN109522966A (en) | A kind of object detection method based on intensive connection convolutional neural networks | |
| CN108229355A (en) | Activity recognition method and apparatus, electronic equipment, computer storage media, program | |
| CN108830826A (en) | A kind of system and method detecting Lung neoplasm | |
| CN106096561A (en) | Infrared pedestrian detection method based on image block degree of depth learning characteristic | |
| CN108319957A (en) | A kind of large-scale point cloud semantic segmentation method based on overtrick figure | |
| CN107330949A (en) | A kind of artifact correction method and system | |
| CN107249455A (en) | Gradually gesture recognition device and vegetative nervous function information acquisition device, methods and procedures | |
| CN103971361B (en) | Image processing device and method | |
| CN104751185B (en) | SAR image change detection based on average drifting genetic cluster | |
| CN104282026B (en) | Distributing homogeneity appraisal procedure based on watershed algorithm and minimum spanning tree | |
| CN103218832B (en) | Based on the vision significance algorithm of global color contrast and spatial distribution in image | |
| CN110501278A (en) | A kind of method for cell count based on YOLOv3 and density estimation | |
| CN107016233A (en) | The association analysis method and system of motor behavior and cognitive ability | |
| CN109711401A (en) | A kind of Method for text detection in natural scene image based on Faster Rcnn | |
| CN106031631A (en) | Method, device, and system for heart rate detection | |
| CN105912991A (en) | Behavior identification method based on 3D point cloud and key bone nodes | |
| CN105046281A (en) | Human body behavior detection method based on Kinect | |
| CN109948637A (en) | Object test equipment, method for checking object and computer-readable medium | |
| CN103034778A (en) | Method for extracting brain function network of individual based on analysis of multiple tested brain function data | |
| CN107392085A (en) | The method for visualizing convolutional neural networks | |
| TW201123074A (en) | Facial animation system and production method | |
| CN107341440A (en) | Indoor RGB D scene image recognition methods based on multitask measurement Multiple Kernel Learning | |
| CN106133789A (en) | For splitting image processing apparatus and the method for area-of-interest | |
| CN106780422A (en) | A kind of notable figure fusion method based on Choquet integrations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |