CN108338777A - A kind of pulse signal determination method and device - Google Patents
A kind of pulse signal determination method and device Download PDFInfo
- Publication number
- CN108338777A CN108338777A CN201810130423.3A CN201810130423A CN108338777A CN 108338777 A CN108338777 A CN 108338777A CN 201810130423 A CN201810130423 A CN 201810130423A CN 108338777 A CN108338777 A CN 108338777A
- Authority
- CN
- China
- Prior art keywords
- pulse signal
- signal
- individual
- feature
- classification
- 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
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Surgery (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Public Health (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Fuzzy Systems (AREA)
- Cardiology (AREA)
- Signal Processing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Psychiatry (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The embodiment of the invention discloses a kind of pulse signal determination method and device, the accuracy rate for improving pulse signal classification.Present invention method includes：Acquire at least a kind of pulse signal of user；Analyzing processing is carried out at least a kind of pulse signal, obtains the characteristic signal of at least a kind of pulse signal；Classified to the characteristic signal of at least a kind of pulse signal using classifier functions；Classification results are fed back into the user.The embodiment of the present invention additionally provides a kind of pulse signal detection and analysis device, the accuracy rate for improving pulse signal classification.
Description
Technical field
The present invention relates to technical field of medical equipment more particularly to a kind of pulse signal determination methods and device.
Background technology
Pulse signal is formed in the contraction and diastole of cardiac cycle, can be by aroused in interest in the conduction of pulse and reflection process
The hearts character factor such as period, cardiac output influences, while also by the hardness of blood vessel, diameter, blood viscosity, peripheral resistance etc.
The affect trait of blood and blood vessel, this make pulse signal contained with the relevant abundant Human Physiology of the entire circulatory system and
Pathological information.
The states such as China, India are widely used in the diagnosis of disease to pulse signal in ancient times, however since diagnosis by feeling the pulse has one
Fixed subjectivity, the personal experience of medical diagnosis on disease process dependence doctor, therefore diagnosis by feeling the pulse in recent years objectify as research hotspot
One of.Currently, having emerged in large numbers miscellaneous pulse signal acquisition device in the research process to pulse signal, in short respectively
Class pulse signal obtains system, can be divided into three classes substantially：Pressure pulse signal acquiring system, photoelectric sphyg signal acquiring system
With ultrasound pulse signal acquiring system.
Since different pulse signals has different sensitive features, if pressure pulse signal is to blood vessel elasticity and thickness of pipe wall
Degree is more sensitive, and photoelectric sphyg signal is more sensitive to blood composition, and ultrasound pulse signal is more sensitive to blood flow state and blood viscosity,
Therefore can be classified to different types of individual according to different pulse signals, and the subjective experience pair that existing basis is artificial
The method that pulse signal is classified, lacks objectivity and accuracy rate is relatively low.
Invention content
An embodiment of the present invention provides a kind of pulse signal determination method and devices, for realizing grader letter is utilized
Several Classification and Identifications to user's pulse signal feature improve recognition efficiency and accuracy rate.
First aspect of the embodiment of the present invention provides a kind of pulse signal determination method, including：
Acquire at least a kind of pulse signal of user；
Analyzing processing is carried out at least a kind of pulse signal, obtains the characteristic signal of at least a kind of pulse signal；
Classified to the characteristic signal of at least a kind of pulse signal using classifier functions；
Classification results are fed back into the user.
Preferably, before at least a kind of pulse signal of the acquisition user, the method further includes：
The pulse signal of training sample is acquired, the training sample includes first kind individual and the second class individual；
The pulse signal feature of the first kind individual and the second class individual is extracted, and the first kind is individual
The pulse signal Fusion Features with second class individual are comprehensive pulse signal feature；
By the comprehensive pulse signal feature normalization, fisrt feature is generated；
The classifier functions are determined according to the fisrt feature and machine learning algorithm.
Preferably, described that the classifier functions are determined according to the fisrt feature and machine learning algorithm, including：
Determine the slack variable and penalty factor of training sample classification；
The fisrt feature is determined in from linearly inseparable space reflection to linear separability space, linear separability spatial function
Inner product；
According to the slack variable, the penalty factor, the inner product of the linear separability spatial function and first spy
Sign determines the classifier functions.
Preferably, after the pulse signal of the acquisition training sample, the method further includes：
The pulse signal of the training sample is pre-processed, the pretreatment includes：Noise and/or baseline drift
Removal.
Preferably, described to be classified to the characteristic signal of at least a kind of pulse signal using classifier functions, it wraps
It includes：
Extract the characteristic value in the characteristic signal；
Classification and Identification is carried out to the characteristic value using the classifier functions, with the classification of the determination user.
Second aspect of the present invention provides a kind of pulse signal detection and analysis device, including：
First collecting unit, at least a kind of pulse signal for acquiring user；
Analysis and processing unit obtains at least a kind of pulse for carrying out analyzing processing at least a kind of pulse signal
The characteristic signal of signal；
Taxon, for being classified to the characteristic signal of at least a kind of pulse signal using classifier functions；
Feedback unit, for the classification of the user to be fed back to the user.
Preferably, described device further includes：
Second collecting unit, the pulse signal for acquiring training sample, the training sample include first kind individual and
Second class individual；
Integrated unit, the pulse signal feature for extracting the first kind individual and the second class individual, and will
The pulse signal Fusion Features of the first kind individual and second class individual are comprehensive pulse signal feature；
Normalization unit, for by the comprehensive pulse signal feature normalization, generating fisrt feature；
Grader determination unit, for determining the classifier functions according to the fisrt feature and machine learning algorithm.
Preferably, the grader determination unit, including：
First determining module, slack variable and penalty factor for determining training sample classification；
Second determining module, for determining the fisrt feature from linearly inseparable space reflection to linear separability space
In, the inner product of linear separability spatial function；
Third determining module, for according to the slack variable, the penalty factor, the linear separability spatial function
Inner product and the fisrt feature determine the classifier functions.
Preferably, which further includes：
Pretreatment unit is pre-processed for the pulse signal to the training sample, and the pretreatment includes：Noise
And/or the removal of baseline drift.
Preferably, taxon, including：
Extraction module, for extracting the characteristic value in the characteristic signal；
Classification and Identification module, for carrying out Classification and Identification to the characteristic value using the classifier functions, to determine
State the classification of user.
The third aspect of the embodiment of the present invention provides a kind of wearable device, includes the arteries and veins of second aspect of the embodiment of the present invention
Signal detection of fighting analytical equipment.
The embodiment of the present invention additionally provides a kind of computer installation, including processor, which is stored in processing
When computer program on reservoir, for realizing the pulse signal determination method of first aspect of the embodiment of the present invention.
The embodiment of the present invention additionally provides a kind of readable storage medium storing program for executing, is stored thereon with computer program, the computer journey
For realizing the pulse signal determination method of first aspect of the embodiment of the present invention when sequence is executed by processor.
As can be seen from the above technical solutions, the embodiment of the present invention has the following advantages：
In the embodiment of the present invention, at least a kind of pulse signal that device acquires user is tested and analyzed by pulse signal, it is right
The pulse signal carry out analyzing processing, obtain the characteristic signal of the pulse signal, using classifier functions to this feature signal into
Row classification, and classification results are fed back into user, the pulse signal detection and analysis device in the present embodiment utilizes classifier functions,
Classification and Identification, more existing human subjective's experience recognition methods, recognition efficiency higher, accuracy are carried out to the pulse signal of user
Also higher.
Description of the drawings
Fig. 1 is one embodiment schematic diagram of pulse signal detection analysis method in the embodiment of the present invention；
Fig. 2 is to be carried out to the characteristic signal of at least a kind of pulse signal of user using classifier functions in the embodiment of the present invention
One embodiment schematic diagram of classification；
Fig. 3 is one embodiment schematic diagram that classifier functions are determined in the embodiment of the present invention；
Fig. 4 is the pulse signal schematic diagram of first kind individual and the second class individual in the embodiment of the present invention；
Fig. 5 is the physical significance schematic diagram of single pulse cycle parameter attribute in the embodiment of the present invention；
Fig. 6 is a reality for determining classifier functions in the embodiment of the present invention according to fisrt feature and vectorial support machine algorithm
Illustration is applied to be intended to；
Fig. 7 is one embodiment schematic diagram of pulse signal detection analytical equipment in the embodiment of the present invention；
Fig. 8 is another embodiment schematic diagram of pulse signal detection analytical equipment in the embodiment of the present invention.
Specific implementation mode
An embodiment of the present invention provides a kind of pulse signal determination method and devices, for according to classifier functions pair
User's pulse signal feature carries out Classification and Identification, improves the recognition efficiency that pulse signal feature is identified and accuracy rate.
Since upper limb radial artery is closer apart from body surface, has compared with other superficial arteries and measure conveniently, by age, blood pressure
Etc. factors influence it is less the advantages that, be constantly subjected to the favor of medical field.Pulse signal herein is also referred to as upper limb radial artery
The pulse signal at place, but have many subjective factors during the description and diagnosis of Sphygmology, therefore objectification of pulse examination becomes close
Research hotspot over year.
Specifically, objectification of pulse examination is dedicated to the combination of diagnosis by feeling the pulse theory, sensor technology and computer technology.It is profit
Pulse signal is acquired with sensor technology, is digitized preservation in a computer, and is obtained by mode identification technology analysis
The pulse signal taken.The main process of pulse signal acquisition is that pulse signal is first converted to electric signal by sensor, by putting
Greatly, analog-to-digital conversion is eventually converted into the storable digital signal of computer.The signal that pulse signal acquisition system is measured according to it
Type can generally be divided into three classes：Pressure pulse signal acquiring system, photoelectric transfer pulse signal acquisition system and ultrasonic arteries and veins
It fights signal acquiring system.
Machine learning algorithm is in artificial intelligence field, and how research realizes the intelligence to measurement object in empirical learning
It can learn, wherein classifier functions are to realize the Accurate classification to measuring sample how according to training sample.Based on existing skill
According to human subjective's experience to the method for pulse signal Classification and Identification in art, the present invention proposes a kind of pulse signal detection and analysis
Method, for carrying out Classification and Identification to pulse signal using classifier functions.
For convenience of understanding, the pulse signal determination method in the embodiment of the present invention will be described below, referring to Fig. 1,
One embodiment of pulse signal detection analysis method in the embodiment of the present invention, including：
101, at least a kind of pulse signal of acquisition user；
Because can be by the shadow of the hearts character factor such as cardiac cycle, cardiac output in the conduction of pulse and reflection process
It rings, while also by the hardness of blood vessel, diameter, blood viscosity, the blood such as peripheral resistance and blood vessel affect trait, this makes pulse
Signal has contained and the entire circulatory system relevant abundant Human Physiology and pathological information so that the detection point to pulse signal
Analysis becomes research hotspot.
And the acquisition of pulse signal is the basis of Pulse signal analysis, accurately and efficiently obtains the pulse under specified pressure
Signal is the premise of follow-up Pulse signal analysis, in existing pulse signal acquisition other than pressure pulse signal, also two
The relatively conventional pulse signal of kind, is photoelectric sphyg signal and ultrasound pulse signal respectively, and is believed to pulse in the present embodiment
Before number detection and analysis, need to obtain at least a kind of pulse signal in above-mentioned three classes pulse signal.
102, analyzing processing is carried out at least a kind of pulse signal, obtains the feature letter of at least a kind of pulse signal
Number；
It is easily understood that by the way that often some signal was acquiring in the collected pulse signal of collecting device
Distortion is generated by various interference in journey, if directly extracting characteristic signal from these signals, it will so that extract
There is deviation between characteristic signal and true characteristic signal, to cause the error of data processed result, even mistake.
Therefore pulse signal acquisition and analysis device analyzes the pulse signal after getting at least a kind of pulse signal
Processing, to remove the interference signal in pulse signal, wherein interference common in the acquisition process of pulse signal is mainly high frequency
Noise and baseline drift,
Specifically, the mode that low-pass filtering may be used in the removal of high-frequency noise removes, and baseline drift is to be coupling in arteries and veins
The low frequency signal fought in signal, pulse signal can be made to fluctuate up and down so that the extraction of some temporal signatures points generate it is certain
Deviation, the strategy of strategy or filtering that curve matching generally may be used is removed.
103, classified to the characteristic signal of at least a kind of pulse signal using classifier functions；
After pulse signal tests and analyzes device at least a kind of pulse signal characteristic filtering, tested and analyzed using pulse signal
Classifier functions in device classify to the pulse signal, to identify the classification of user.
Specifically, to the determination process of classifier functions in this present embodiment, and according to classifier functions to pulse signal
Classification and Identification process, be described in the following embodiments.
104, classification results are fed back into the user.
In step 103, after pulse signal detection and analysis device determines the classification results of user, which is fed back
To user, specifically, classification results can be fed back to user by pulse signal detection and analysis device by the terminal screen of itself,
It can also connect the wearable device of (wired connection or wireless connection), various terminals by testing and analyzing device with pulse signal
The screen feedback of (Pad, mobile phone, computer installation) is to user.
Based on Fig. 1 the embodiment described, it is described in detail below in the embodiment of the present invention using classifier functions to user
The process that the characteristic signal of at least a kind of pulse signal is classified, referring to Fig. 2, utilizing grader letter in the embodiment of the present invention
One embodiment that several characteristic signals at least a kind of pulse signal of user are classified, including：
201, the characteristic value in characteristic signal is extracted；
After the step 102 of Fig. 1 embodiments, pulse signal tests and analyzes device and divides at least a kind of pulse signal
Analysis is handled, and after obtaining the characteristic signal of the pulse signal, further extracts the characteristic value in this feature signal.
Wherein, the characteristic value in signal characteristic refers to the periodic characteristic in each characteristic signal, specifically includes：SW、T1、
The object meaning of the combination of T1/T, T2/T, T3/T, T1/T4, h1/h and h2/h feature, specific each feature can be refering to following
Shown in table 1 in embodiment.
202, Classification and Identification is carried out to characteristic value using classifier functions, with the classification of the determination user.
Specifically, pulse signal tests and analyzes device after the characteristic value in obtaining each characteristic signal, pass through classification
Device function pair this feature value is calculated, and determines that this feature value belongs to first kind individual or the second class according to result of calculation
Body, to achieve the purpose that determining class of subscriber.
In the embodiment of the present invention, at least a kind of pulse signal that device acquires user is tested and analyzed by pulse signal, it is right
The pulse signal carry out analyzing processing, obtain the characteristic signal of the pulse signal, using classifier functions to this feature signal into
Row classification, and classification results are fed back into user, the pulse signal detection and analysis device in the present embodiment utilizes classifier functions,
It realizes according to the pulse signal of user to the Classification and Identification of user, more existing human subjective's experience recognition methods, recognition efficiency
Higher, accuracy also higher.
Based on Fig. 1 the embodiment described, the determination process of classifier functions in the embodiment of the present invention is described below, is please referred to
Fig. 3 determines one embodiment of classifier functions in the embodiment of the present invention, including：
301, the pulse signal of training sample is acquired, the training sample includes first kind individual and the second class individual；
Training sample in the embodiment of the present invention using a variety of pulse signal sources of variety classes individual as grader,
In, training sample be divided into first kind individual and the second class individual, pulse signal include pressure pulse signal, photoelectric sphyg signal and
In ultrasound pulse signal.
Specifically, the pulse signal of the first kind individual and the second class individual in the present embodiment is different, for being divided
Class is tested, and if first kind individual is male, the second class individual is women or first kind individual is diabetic, the second class
Body is healthy individuals etc., need to only meet the pulse signal difference of first kind individual and the second class individual herein, and to first
The criteria for classification of class individual and the second class individual is not particularly limited.
Assuming that the first kind individual in the present embodiment is diabetic, the second class individual is healthy individuals, wherein Fig. 4
(a) and Fig. 4 (b) be respectively first kind individual and the second class individual pulse signal schematic diagram, wherein c1 to c10 is respectively 10
The period of pulse signal.
And the pressure pulse signal in the present embodiment may be used ZMH-I type pulse signals acquisition system and be acquired, and surpass
The supersonic blood signal that sound pulse signal can directly acquire radial artery using medical ultrasonic system at radial artery is obtained
It takes, photoelectric sphyg signal can be obtained using photoelectricity volume pulse signal extraction system.It should be noted that above-mentioned pulse
The acquisition system of signal is a kind of citing to pulse signal acquisition modes, not to the acquisition modes structure of various pulse signals
At limitation.
302, the pulse signal is pre-processed, the pretreatment includes：The removal of noise and/or baseline drift；
It is easily understood that by the way that often some signal was acquiring in the collected pulse signal of collecting device
Distortion is generated by various interference in journey, if these signals directly carry out feature extraction and can make in data set comprising abnormal
Sample, to which the training of the accuracy disaggregated model of effect characteristics extraction even causes classification error.Therefore it generally requires and passes through
Preprocessing process removal distorts to improve signal quality.
Common interference is mainly high-frequency noise and baseline drift in the acquisition process of pulse signal, these interference can influence
The accuracy of feature extraction, in feature extraction, the feature for the signal being interfered has uncertainty, can influence grader
Study, such as exceptional sample can influence the position of the Optimal Separating Hyperplane of support vector machines when being classified using support vector machines
It sets.Pretreated major significance is that removal couples interference in the signal or removes containing noisy signal, to promote arteries and veins
It fights the sample quality and credibility of sample's of signal data collection so that subsequent feature extraction and classification have higher accuracy.
Specifically, the removal of high-frequency noise can both be gone mainly by way of low-pass filtering by low-pass filter
It removes, can also be removed by way of wavelet filtering.Baseline drift is the low frequency signal being coupling in pulse signal, can make arteries and veins
Signal of fighting fluctuates up and down so that the extraction of some temporal signatures points generates certain deviation, and curve matching generally may be used
Strategy or the strategy of filtering be removed, specifically, curve matching strategy is to obtain pulse signal progress period divisions respectively
Then the starting point of pulse cycle obtains a matched curve by being fitted the starting point in pulse signal each period, and by this
For curve as baseline, baseline drift can be removed by being made the difference with original pulse signal and baseline.
303, the pulse signal feature of first kind individual and the second class individual is extracted, and by first kind individual and the second class
The pulse signal Fusion Features of individual are comprehensive pulse signal feature；
After carrying out pretreatment removal noise to pulse signal, the pressure pulse letter of first kind individual can be extracted respectively
Number, single pulse cycle feature in photoelectric sphyg signal and ultrasound pulse signal, then extract the pressure pulse of the second class individual
Single pulse cycle feature in signal, photoelectric sphyg signal and ultrasound pulse signal, then by first kind individual and the second class
The single pulse signal periodic characteristic of individual is respectively combined as multiple pulse signal periodic characteristics, specific single pulse cycle letter
Number feature includes：Pulse main peak feature, pulse dicrotic pulse feature and pulse dicrotic notch feature.
Respectively by the first kind individual, the second class individual single pulse cycle Fusion Features be multiple pulse cycle features
Afterwards, by multiple pulse signal Fusion Features of first kind individual and the second class individual it is again then comprehensive pulse signal feature.
Specifically, we can indicate pulse signal spy by the following parameters in single pulse cycle feature in table 1
Sign, and the physical significance of each feature is referred to shown in Fig. 5 in table 1.
Table 1
| Feature | Meaning |
| SW | The mid-length of main peak ascending branch and decent |
| T1 | Pulse by pulse starting point to main peak time |
| T1/T | The ratio between time and pulse signal period shared by main peak ascending branch |
| T2/T | The ratio between time and pulse signal period shared by main peak decent |
| T3/T | The ratio between time and pulse signal period shared by dicrotic pulse wave crest ascending branch |
| T1/T4 | The ratio of time shared by ascending branch and decent |
| h1/h | The ratio between dicrotic notch amplitude and main peak amplitude |
| h2/h | The ratio between dicrotic wave peak amplitude and main peak amplitude |
In order to the fusion process of first kind individual and the multiple pulse signal features of the second class individual is described in detail, lift below
Example illustrates：Assuming that the number of first kind individual is n1/2, and single pressure pulse signal characteristic is p1, then p1It is one group corresponding
The combination of SW, T1, T1/T, T2/T, T3/T, T1/T4, h1/h and h2/h feature, if multiple pressure pulses letter of first kind individual
Number feature is combined as px1, then corresponding px1=[p1,p2...pn1/2]TIf the number of the second class individual is also n1/2, and single
A pressure pulse signal characteristic is p(n1/2)+1, then p(n1/2)+1Also one group of SW, T1, T1/T, T2/T, T3/T, T1/T4, h1/h are corresponded to
And the combination of h2/h features, if multiple pulse signal features of the second class individual are combined as px2, then corresponding px2=
[p(n1/2)+1,p(n1/2)+2...pn1]T, then first kind individual and the multiple pressure pulse signals of the second class individual are combined as P=[p1,
p2..., pn1]T, can similarly obtain multiple photoelectric sphyg signal characteristics of first kind individual and the second class individual is combined as L
=[l1, l2..., ln2]T, multiple ultrasound pulse signal characteristics of first kind individual and the second class individual are S=[s1, s2...,
sn3]T, the pressure of first kind individual and the second class individual, photoelectricity and ultrasound pulse signal characteristic are finally fused to comprehensive arteries and veins
It fights signal characteristic, it is assumed that comprehensive pulse signal is characterized as xn, then xn=[p1, p2..., pn1, l1, l2..., ln2, s1, s2...,
sn3]T=[τN, 1, τN, 2..., τN, m]T, so it is easy to understand that wherein n ∈ [1, N], m=n1+n2+n3。
304, by comprehensive pulse signal feature normalization, fisrt feature is generated；
It is easily understood that the purpose of data normalization is in order to which the data of separate sources are unified to a referential
Under, so that the data of separate sources are significant by comparison.
Specifically, when machine learning is applied, characteristic processing occupies most of the time in the application, therefore feature is returned
One change is the essential step in characteristic processing, because can not only accelerate gradient after normalization declines the speed for seeking optimal solution
Degree, while data precision can be improved.
After the synthesis pulse signal feature for obtaining first kind individual and the second class individual in step 203,
According to normalized formula, it is after total pulse signal feature normalization then：
It is by the fisrt feature that total pulse signal feature normalization generates then：
305, classifier functions are determined according to the fisrt feature and machine learning algorithm.
Pulse signal tests and analyzes device after obtaining fisrt feature, and fisrt feature data sample is calculated with machine learning
Method seeks optimal solution, that is, classifier functions is determined, for the pulse signal feature by the classifier functions to object to be detected
Classify, improves the accuracy of classification.
Based on Fig. 3 the embodiment described, how it is described below in detail in the embodiment of the present invention according to fisrt feature and machine
Learning algorithm determines classifier functions, referring to Fig. 6, being determined according to fisrt feature and machine learning algorithm in the embodiment of the present invention
One embodiment of classifier functions, including：
601, the slack variable and penalty factor of training sample classification are determined；
For the pulse signal of human body, wherein pressure pulse signal characteristic, photoelectric sphyg signal characteristic and ultrasonic arteries and veins
Signal characteristic of fighting is linearly inseparable, in order to during determining classifier functions so that training sample is by linearly can not
Dividing becomes linear separability, can introduce slack variable ξi, wherein i=1,2,3 ... m, m are sample numbers, and slack variable is for indicating
Fault-tolerance during sample classification.
There is sample skewness phenomenon to further prevent certain a kind of sample size in training sample too small, to give
Classification results cause relatively large deviation, can be by predefining penalty factor, and to constrain the deflection of sample size, wherein C is bigger
Show more to pay attention to the number of samples for participating in sample classification, C is smaller, then more thinks little of.
602, fisrt feature is determined in from linearly inseparable space reflection to linear separability space, linear separability spatial function
Inner product；
Because for the pulse signal of human body, wherein pressure pulse signal characteristic, photoelectric sphyg signal characteristic and super
Sound pulse signal is characterized in linearly inseparable, for convenience of calculation, needs the pulse signal Feature Conversion of linearly inseparable
For the signal characteristic of linear separability, and core sorting technique main thought is to be reflected the feature of sample from two-dimensional space by kernel function
It is mapped to higher dimensional space H, so that becoming linear separability the problem of original linearly inseparable.
Assuming that the sample characteristics observed are x, it belongs to feature space X (linearly inseparable space), by kernel function by x
Pass through mappingBe mapped to hyperspace (linear separability space) H=f | f:X → R }, and found in the H of spaceBetween linear relationship, wherein the reproducing kernel of H be K：X × X → R, kernel function, which defines, is embedded in mappingUnder it is interior
Product：
Kernel function defines feature space H, in most cases need not construction feature space H, it is only necessary to know that feature space H
Inner product, wherein the inner product of defined feature space H is：
Wherein, αi∈ R, l ∈ N, then feature space H is really the set of some functions：
Next the inner product of F is defined, ifDefinition
Then the inner product of feature space H is：
603, grader is determined according to slack variable, penalty factor, the inner product of linear separability spatial function and fisrt feature
Function.
In algorithm of support vector machine, decision function is：
sgn(f*(x)+b*) (1)
Wherein, f* and b* is the solution of following formula：
Wherein, xiFor fisrt feature, yiFor the label of training sample, wherein yi=± 1, the classification for distinguishing sample, when
yiWhen=+ 1, the first kind individual in training sample, y are indicatediWhen=- 1, indicate that the second class individual in training sample, C are to punish
Penalty factor, ξ i For slack variable,It is characterized the inner product of space H (linear separability spatial function), by fisrt feature xnValue make
For xiIn input type (2), b and f (x is determined respectivelyi) optimal value, you can to determine the value of f* and b*, and then determine in formula (1)
Decision function, also as classifier functions.
In the embodiment of the present invention, at least a kind of pulse signal that device acquires user is tested and analyzed by pulse signal, it is right
The pulse signal carry out analyzing processing, obtain the characteristic signal of the pulse signal, using classifier functions to this feature signal into
Row classification, and classification results are fed back into user, the pulse signal detection and analysis device in the present embodiment utilizes classifier functions,
It realizes according to the pulse signal of user to the Classification and Identification of user, more existing human subjective's experience recognition methods, recognition efficiency
Higher, accuracy also higher.
Secondly, the embodiment of the present invention tests and analyzes device by pulse signal and acquires training sample (first kind individual respectively
With the second class individual) pressure pulse signal characteristic, photoelectric sphyg signal characteristic and ultrasound pulse signal characteristic merged, obtain
To comprehensive pulse signal feature, then fisrt feature will be generated after comprehensive pulse signal feature normalization, then according to fisrt feature
Classifier functions are determined with machine learning algorithm, therefore the classifier functions, compared with the single pulse signal feature according to training sample
The classification accuracy higher of classifier functions obtained from training.
Described above is the pulse signal determination methods in the embodiment of the present invention, and the embodiment of the present invention will be described below
In pulse signal test and analyze device, referring to Fig. 7, in the embodiment of the present invention pulse signal detection analytical equipment a reality
Example is applied, including：
First collecting unit 701, at least a kind of pulse signal for acquiring user；
Analysis and processing unit 702 obtains at least a kind of arteries and veins for carrying out analyzing processing at least a kind of pulse signal
It fights the characteristic signal of signal；
Taxon 703, for being divided the characteristic signal of at least a kind of pulse signal using classifier functions
Class；
Feedback unit 704, for the classification of the user to be fed back to the user.
In the embodiment of the present invention, at least a kind of pulse signal of user is acquired by the first collecting unit 701, passes through analysis
Processing unit 702 carries out analyzing processing to the pulse signal, obtains the characteristic signal of the pulse signal, passes through taxon 703
Classified to this feature signal using classifier functions, and classification results are fed back into user, the pulse letter in the present embodiment
Number detection and analysis device utilizes classifier functions, realizes the Classification and Identification to user according to the pulse signal of user, more existing
Human subjective's experience recognition methods, recognition efficiency higher, accuracy also higher.
Based on Fig. 7 the embodiment described, the pulse signal detection and analysis device in the embodiment of the present invention will be described below, ask
Refering to Fig. 8, another embodiment of pulse signal detection analytical equipment in the embodiment of the present invention, including：
First collecting unit 801, at least a kind of pulse signal for acquiring user；
Analysis and processing unit 802 obtains at least a kind of arteries and veins for carrying out analyzing processing at least a kind of pulse signal
It fights the characteristic signal of signal；
Taxon 803, for being divided the characteristic signal of at least a kind of pulse signal using classifier functions
Class；
Feedback unit 804, for the classification of the user to be fed back to the user.
Preferably, described device further includes：
Second collecting unit 805, the pulse signal for acquiring training sample, the training sample include first kind individual
With the second class individual；
Integrated unit 806, the pulse signal feature for extracting the first kind individual and the second class individual, and
It is comprehensive pulse signal feature by the pulse signal Fusion Features of the first kind individual and second class individual；
Normalization unit 807, for by the comprehensive pulse signal feature normalization, generating fisrt feature；
Grader determination unit 808, for determining the grader letter according to the fisrt feature and machine learning algorithm
Number.
Preferably, the grader determination unit 808, including：
First determining module 8081, slack variable and penalty factor for determining training sample classification；
Second determining module 8082, for determining that the fisrt feature is empty from linearly inseparable space reflection to linear separability
Between in, the inner product of linear separability spatial function；
Third determining module 8083, for according to the slack variable, the penalty factor, linear separability space letter
Several inner products and the fisrt feature determine the classifier functions.
Preferably, which further includes：
Pretreatment unit 809 is pre-processed for the pulse signal to the training sample, and the pretreatment includes：
The removal of noise and/or baseline drift.
Preferably, taxon 803, including：
Extraction module 8031, for extracting the characteristic value in the characteristic signal；
Classification and Identification module 8032, for carrying out Classification and Identification to the characteristic value using the classifier functions, with true
The classification of the fixed user.
In the embodiment of the present invention, at least a kind of pulse signal of user is acquired by the first collecting unit 801, passes through analysis
Processing unit 802 carries out analyzing processing to the pulse signal, obtains the characteristic signal of the pulse signal, passes through taxon 803
Classified to this feature signal using classifier functions, and classification results are fed back into user, the pulse letter in the present embodiment
Number detection and analysis device utilizes classifier functions, realizes the Classification and Identification to user according to the pulse signal of user, more existing
Human subjective's experience recognition methods, recognition efficiency higher, accuracy also higher.
Secondly, the embodiment of the present invention acquires training sample (first kind individual and second respectively by the second collecting unit 805
Class individual) pulse signal feature merged, obtain comprehensive pulse signal feature, then will comprehensive pulse signal feature normalization
After generate fisrt feature, classifier functions are then determined according to fisrt feature and machine learning algorithm, therefore the classifier functions, compared with
The classification accuracy higher of classifier functions obtained from being trained according to the single pulse signal feature of training sample.
The embodiment of the present invention additionally provides a kind of wearable device, including pulse signal collector, processor, memory,
Power supply, wherein processor are implemented in the computer program on executing memory for realizing described in Fig. 1, Fig. 2, Fig. 3 and Fig. 4
Step in example, details are not described herein again.
The pulse signal detection and analysis device in the embodiment of the present invention is carried out from the angle of modular functionality entity above
Description, is below described the computer installation in the embodiment of the present invention from the angle of hardware handles：
The computer installation tests and analyzes the function of device, Computer of embodiment of the present invention dress for realizing pulse signal
Setting one embodiment includes：
Processor and memory；
Memory can when processor is used to execute the computer program stored in memory for storing computer program
To realize following steps：
Acquire at least a kind of pulse signal of user；
Analyzing processing is carried out at least a kind of pulse signal, obtains the characteristic signal of at least a kind of pulse signal；
Classified to the characteristic signal of at least a kind of pulse signal using classifier functions；
Classification results are fed back into the user.
In some embodiments of the invention, processor can also be specifically used for realizing following steps：
The pulse signal of training sample is acquired, the training sample includes first kind individual and the second class individual；
The pulse signal feature of the first kind individual and the second class individual is extracted, and the first kind is individual
The pulse signal Fusion Features with second class individual are comprehensive pulse signal feature；
By the comprehensive pulse signal feature normalization, fisrt feature is generated；
The classifier functions are determined according to the fisrt feature and machine learning algorithm.
In some embodiments of the invention, processor can also be specifically used for realizing following steps：
Determine the slack variable and penalty factor of training sample classification；
The fisrt feature is determined in from linearly inseparable space reflection to linear separability space, linear separability spatial function
Inner product；
According to the slack variable, the penalty factor, the inner product of the linear separability spatial function and first spy
Sign determines the classifier functions.
In some embodiments of the invention, processor can also be specifically used for realizing following steps：
The pulse signal of the training sample is pre-processed, the pretreatment includes：Noise and/or baseline drift
Removal.
In some embodiments of the invention, processor can also be specifically used for realizing following steps：
Extract the characteristic value in the characteristic signal；
Classification and Identification is carried out to the characteristic value using the classifier functions, with the classification of the determination user.
It is understood that when the processor in the computer installation of above description executes the computer program, also may be used
To realize the function of each unit in above-mentioned corresponding each device embodiment, details are not described herein again.Illustratively, the computer journey
Sequence can be divided into one or more module/units, and one or more of module/units are stored in the memory
In, and executed by the processor, to complete the present invention.One or more of module/units can be can complete it is specific
The series of computation machine program instruction section of function, the instruction segment are detected for describing the computer program in the pulse signal
Implementation procedure in analytical equipment.For example, the computer program can be divided into above-mentioned pulse signal detection and analysis device
In each unit, the concrete function of above-mentioned corresponding pulse signal detection and analysis device description such as may be implemented in each unit.
The computer installation can be that the calculating such as desktop PC, notebook, palm PC and cloud server are set
It is standby.The computer installation may include but be not limited only to processor, memory.It will be understood by those skilled in the art that processor,
Memory is only the example of computer installation, does not constitute the restriction to computer installation, may include more or fewer
Component either combines certain components or different components, such as the computer installation can also be set including input and output
Standby, network access equipment, bus etc..
The processor can be central processing unit (Central Processing Unit, CPU), can also be it
His general processor, digital signal processor (Digital Signal Processor, DSP), application-specific integrated circuit
(Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..General processor can be microprocessor or the processor can also be any conventional processor
Deng the processor is the control centre of the computer installation, utilizes various interfaces and the entire computer installation of connection
Various pieces.
The memory can be used for storing the computer program and/or module, and the processor is by running or executing
Computer program in the memory and/or module are stored, and calls the data being stored in memory, described in realization
The various functions of computer installation.The memory can include mainly storing program area and storage data field, wherein storage program
It area can storage program area, the application program etc. needed at least one function；Storage data field can store the use according to terminal
The data etc. created.In addition, memory may include high-speed random access memory, can also include non-volatile memories
Device, such as hard disk, memory, plug-in type hard disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure
Digital, SD) card, flash card (Flash Card), at least one disk memory, flush memory device or other volatibility are solid
State memory device.
The present invention also provides a kind of computer readable storage mediums, and the computer readable storage medium is for realizing pulse
The function of signal detection analytical equipment is stored thereon with computer program, when computer program is executed by processor, processor,
It can be used for executing following steps：
Acquire at least a kind of pulse signal of user；
Analyzing processing is carried out at least a kind of pulse signal, obtains the characteristic signal of at least a kind of pulse signal；
Classified to the characteristic signal of at least a kind of pulse signal using classifier functions；
Classification results are fed back into the user.
In some embodiments of the invention, the computer program of computer-readable recording medium storage is executed by processor
When, processor can also be specifically used for executing following steps：
The pulse signal of training sample is acquired, the training sample includes first kind individual and the second class individual；
The pulse signal feature of the first kind individual and the second class individual is extracted, and the first kind is individual
The pulse signal Fusion Features with second class individual are comprehensive pulse signal feature；
By the comprehensive pulse signal feature normalization, fisrt feature is generated；
The classifier functions are determined according to the fisrt feature and machine learning algorithm.
In some embodiments of the invention, the computer program of computer-readable recording medium storage is executed by processor
When, processor can also be specifically used for executing following steps：
Determine the slack variable and penalty factor of training sample classification；
The fisrt feature is determined in from linearly inseparable space reflection to linear separability space, linear separability spatial function
Inner product；
According to the slack variable, the penalty factor, the inner product of the linear separability spatial function and first spy
Sign determines the classifier functions.
In some embodiments of the invention, the computer program of computer-readable recording medium storage is executed by processor
When, processor can also be specifically used for executing following steps：
The pulse signal of the training sample is pre-processed, the pretreatment includes：Noise and/or baseline drift
Removal.
In some embodiments of the invention, the computer program of computer-readable recording medium storage is executed by processor
When, processor can also be specifically used for executing following steps：
Extract the characteristic value in the characteristic signal；
Classification and Identification is carried out to the characteristic value using the classifier functions, with the classification of the determination user.
It is understood that if the integrated unit is realized in the form of SFU software functional unit and as independent production
Product are sold or in use, can be stored in a corresponding computer read/write memory medium.Based on this understanding, this hair
The bright all or part of flow realized in above-mentioned corresponding embodiment method, can also be instructed relevant by computer program
Hardware is completed, and the computer program can be stored in a computer readable storage medium, which is being located
It manages when device executes, it can be achieved that the step of above-mentioned each embodiment of the method.Wherein, the computer program includes computer program generation
Code, the computer program code can be source code form, object identification code form, executable file or certain intermediate forms
Deng.The computer-readable medium may include：Any entity or device, record of the computer program code can be carried
Medium, USB flash disk, mobile hard disk, magnetic disc, CD, computer storage, read-only memory (ROM, Read-Only Memory), with
Machine accesses memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium etc..
It should be noted that the content that the computer-readable medium includes can be according to legislation and patent practice in jurisdiction
It is required that carrying out increase and decrease appropriate, such as in certain jurisdictions, do not wrapped according to legislation and patent practice, computer-readable medium
Include electric carrier signal and telecommunication signal.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
In several embodiments provided herein, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of division of logic function, formula that in actual implementation, there may be another division manner, such as multiple units or component
It can be combined or can be integrated into another system, or some features can be ignored or not executed.Another point, it is shown or
The mutual coupling, direct-coupling or communication connection discussed can be the indirect coupling by some interfaces, device or unit
It closes or communicates to connect, can be electrical, machinery or other forms.
The unit illustrated as separating component may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, you can be located at a place, or may be distributed over multiple
In network element.Some or all of unit therein can be selected according to the actual needs to realize the mesh of this embodiment scheme
's.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, it can also
It is that each unit physically exists alone, it can also be during two or more units be integrated in one unit.Above-mentioned integrated list
The form that hardware had both may be used in member is realized, can also be realized in the form of SFU software functional unit.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can be stored in a computer read/write memory medium.Based on this understanding, technical scheme of the present invention is substantially
The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words
It embodies, which is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or the network equipment etc.) executes the complete of each embodiment the method for the present invention
Portion or part steps.And storage medium above-mentioned includes：USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic disc or CD etc. are various can store journey
The medium of sequence code.
The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although with reference to before
Stating embodiment, invention is explained in detail, it will be understood by those of ordinary skill in the art that：It still can be to preceding
The technical solution recorded in each embodiment is stated to modify or equivalent replacement of some of the technical features；And these
Modification or replacement, the spirit and scope for various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of pulse signal determination method, which is characterized in that including：
Acquire at least a kind of pulse signal of user；
Analyzing processing is carried out at least a kind of pulse signal, obtains the characteristic signal of at least a kind of pulse signal；
Classified to the characteristic signal of at least a kind of pulse signal using classifier functions；
Classification results are fed back into the user.
2. according to the method described in claim 1, it is characterized in that, the acquisition user at least a kind of pulse signal it
Before, the method further includes：
The pulse signal of training sample is acquired, the training sample includes first kind individual and the second class individual；
Extract the pulse signal feature of the first kind individual and the second class individual, and by first kind individual and institute
The pulse signal Fusion Features for stating the second class individual are comprehensive pulse signal feature；
By the comprehensive pulse signal feature normalization, fisrt feature is generated；
The classifier functions are determined according to the fisrt feature and machine learning algorithm.
3. according to the method described in claim 2, it is characterized in that, described true according to the fisrt feature and machine learning algorithm
The fixed classifier functions, including：
Determine the slack variable and penalty factor of training sample classification；
The fisrt feature is determined in from linearly inseparable space reflection to linear separability space, linear separability spatial function it is interior
Product；
It is true according to the slack variable, the penalty factor, the inner product of the linear separability spatial function and the fisrt feature
The fixed classifier functions.
4. according to the method described in claim 2, it is characterized in that, it is described acquisition training sample pulse signal after, institute
The method of stating further includes：
The pulse signal of the training sample is pre-processed, the pretreatment includes：Noise and/or baseline drift are gone
It removes.
5. method according to claim 1 to 4, which is characterized in that described to utilize classifier functions to described
The characteristic signal of at least a kind of pulse signal is classified, including：
Extract the characteristic value in the characteristic signal；
Classification and Identification is carried out to the characteristic value using the classifier functions, with the classification of the determination user.
6. a kind of pulse signal tests and analyzes device, which is characterized in that including：
First collecting unit, at least a kind of pulse signal for acquiring user；
Analysis and processing unit obtains at least a kind of pulse signal for carrying out analyzing processing at least a kind of pulse signal
Characteristic signal；
Taxon, for being classified to the characteristic signal of at least a kind of pulse signal using classifier functions；
Feedback unit, for the classification of the user to be fed back to the user.
7. device according to claim 6, which is characterized in that described device further includes：
Second collecting unit, the pulse signal for acquiring training sample, the training sample include first kind individual and second
Class individual；
Integrated unit, the pulse signal feature for extracting the first kind individual and the second class individual, and will be described
The pulse signal Fusion Features of first kind individual and second class individual are comprehensive pulse signal feature；
Normalization unit, for by the comprehensive pulse signal feature normalization, generating fisrt feature；
Grader determination unit, for determining the classifier functions according to the fisrt feature and machine learning algorithm.
8. device according to claim 7, which is characterized in that the grader determination unit, including：
First determining module, slack variable and penalty factor for determining training sample classification；
Second determining module, for determining the fisrt feature in from linearly inseparable space reflection to linear separability space, line
The inner product of property separable space function；
Third determining module, for the inner product according to the slack variable, the penalty factor, the linear separability spatial function
And the fisrt feature determines the classifier functions.
9. a kind of wearable device, which is characterized in that analyze dress including the pulse detection described in any one of claim 6 to 8
It sets.
10. a kind of computer installation, including processor, which is characterized in that the processor is stored in processing on memory
When computer program, for realizing the pulse signal determination method described in any one of claim 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810130423.3A CN108338777A (en) | 2018-02-08 | 2018-02-08 | A kind of pulse signal determination method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810130423.3A CN108338777A (en) | 2018-02-08 | 2018-02-08 | A kind of pulse signal determination method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108338777A true CN108338777A (en) | 2018-07-31 |
Family
ID=62960060
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810130423.3A Pending CN108338777A (en) | 2018-02-08 | 2018-02-08 | A kind of pulse signal determination method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108338777A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110432885A (en) * | 2019-09-11 | 2019-11-12 | 东北大学 | A kind of photoplethysmographic noise remove method |
| WO2020237722A1 (en) * | 2019-05-27 | 2020-12-03 | 深圳六合六医疗器械有限公司 | Method and apparatus for giving reminder regarding sitting for a long time |
| CN112806973A (en) * | 2021-01-06 | 2021-05-18 | 珠海中科先进技术研究院有限公司 | Method and device for identifying pulse-taking condition |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050049470A1 (en) * | 2003-08-27 | 2005-03-03 | Terry Alvin Mark | Multi-domain motion estimation and plethysmographic recognition using fuzzy neural-nets |
| CN102824166A (en) * | 2012-07-05 | 2012-12-19 | 华东师范大学 | Device for sorting treatment of pulse wave |
| CN103153175A (en) * | 2010-05-14 | 2013-06-12 | 先进计算发展中心 | Diagnostic classifications of pulse signal waveform data |
| CN103169456A (en) * | 2013-03-29 | 2013-06-26 | 深圳职业技术学院 | Processing method and processing system for pulse wave signals |
| CN103505189A (en) * | 2013-10-21 | 2014-01-15 | 东南大学 | Pulse signal classification method based on wavelet packet conversion and hidden markov models |
| CN105286815A (en) * | 2015-11-02 | 2016-02-03 | 重庆大学 | Pulse wave signal feature point detection method based on waveform time domain features |
| CN105433916A (en) * | 2014-09-19 | 2016-03-30 | 株式会社日立制作所 | Pulse feature extraction device and method |
| CN107432736A (en) * | 2017-06-06 | 2017-12-05 | 新绎健康科技有限公司 | A kind of method for identifying pulse wave signal |
-
2018
- 2018-02-08 CN CN201810130423.3A patent/CN108338777A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050049470A1 (en) * | 2003-08-27 | 2005-03-03 | Terry Alvin Mark | Multi-domain motion estimation and plethysmographic recognition using fuzzy neural-nets |
| CN103153175A (en) * | 2010-05-14 | 2013-06-12 | 先进计算发展中心 | Diagnostic classifications of pulse signal waveform data |
| CN102824166A (en) * | 2012-07-05 | 2012-12-19 | 华东师范大学 | Device for sorting treatment of pulse wave |
| CN103169456A (en) * | 2013-03-29 | 2013-06-26 | 深圳职业技术学院 | Processing method and processing system for pulse wave signals |
| CN103505189A (en) * | 2013-10-21 | 2014-01-15 | 东南大学 | Pulse signal classification method based on wavelet packet conversion and hidden markov models |
| CN105433916A (en) * | 2014-09-19 | 2016-03-30 | 株式会社日立制作所 | Pulse feature extraction device and method |
| CN105286815A (en) * | 2015-11-02 | 2016-02-03 | 重庆大学 | Pulse wave signal feature point detection method based on waveform time domain features |
| CN107432736A (en) * | 2017-06-06 | 2017-12-05 | 新绎健康科技有限公司 | A kind of method for identifying pulse wave signal |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020237722A1 (en) * | 2019-05-27 | 2020-12-03 | 深圳六合六医疗器械有限公司 | Method and apparatus for giving reminder regarding sitting for a long time |
| CN110432885A (en) * | 2019-09-11 | 2019-11-12 | 东北大学 | A kind of photoplethysmographic noise remove method |
| CN112806973A (en) * | 2021-01-06 | 2021-05-18 | 珠海中科先进技术研究院有限公司 | Method and device for identifying pulse-taking condition |
| CN112806973B (en) * | 2021-01-06 | 2024-02-23 | 珠海中科先进技术研究院有限公司 | Guan Maimai image identification method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Heart sound classification based on scaled spectrogram and tensor decomposition | |
| CN111160139B (en) | Electrocardiosignal processing method and device and terminal equipment | |
| US20200337580A1 (en) | Time series data learning and analysis method using artificial intelligence | |
| Wang et al. | Arrhythmia classification algorithm based on multi-head self-attention mechanism | |
| Alaskar et al. | The implementation of pretrained AlexNet on PCG classification | |
| EP3692903A1 (en) | Artificial intelligence self-learning-based automatic electrocardiography analysis method and apparatus | |
| Luo et al. | A study of machine‐learning classifiers for hypertension based on radial pulse wave | |
| CN108090509B (en) | Data length self-adaptive electrocardiogram classification method | |
| CN108742697B (en) | Heart sound signal classification method and terminal equipment | |
| Nascimento et al. | Heart arrhythmia classification based on statistical moments and structural co-occurrence | |
| CN113095302B (en) | Depth model for arrhythmia classification, method and device using same | |
| CN113855038B (en) | Electrocardiosignal critical value prediction method and device based on multi-model integration | |
| CN108338777A (en) | A kind of pulse signal determination method and device | |
| Ullah et al. | An end-to-end cardiac arrhythmia recognition method with an effective densenet model on imbalanced datasets using ecg signal | |
| CN111053552B (en) | QRS wave detection method based on deep learning | |
| Chen et al. | Patient emotion recognition in human computer interaction system based on machine learning method and interactive design theory | |
| Tago et al. | Classification of TCM pulse diagnoses based on pulse and periodic features from personal health data | |
| CN107837083B (en) | J wave automatic testing method based on least square method supporting vector machine | |
| CN113796873A (en) | Wearable dynamic electrocardiosignal classification method and system | |
| Bhattacharya et al. | Synthetic data generation through statistical explosion: Improving classification accuracy of coronary artery disease using PPG | |
| Ninh et al. | An improved subject-independent stress detection model applied to consumer-grade wearable devices | |
| CN112336369A (en) | Coronary heart disease risk index evaluation system of multichannel heart sound signals | |
| Shaker et al. | Deep convolutional neural networks for ECG heartbeat classification using two-stage hierarchical method | |
| Sivanjaneyulu et al. | Cnn based ppg signal quality assessment using raw ppg signal for energy-efficient ppg analysis devices in internet of medical things | |
| CN115177260A (en) | Intelligent electrocardio-heart sound diagnosis method and device based on artificial neural network |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20190625 Address after: 518057 Yuanzheng Industrial Park, 4012 North Fifth Avenue, Bantian Street, Longgang District, Shenzhen City, Guangdong Province Applicant after: Yuanzheng Science and Technology Co., Ltd., Shenzhen City Address before: 518000 Room 201, building A, No. 1, Qian Wan Road, Qianhai Shenzhen Hong Kong cooperation zone, Shenzhen, Guangdong (Shenzhen Qianhai business secretary Co., Ltd.) Applicant before: HESVIT HEALTH TECH CO., LTD. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180731 |